Gastroenterologic Endoscopy (Vickey1214) VRG

WILLIAM S. HAUBRICH, M.D.
JAMES M. EDMONSON, PH.D.

The word endoscopy is derived from the Greek by combining the prefix endo-, meaning within, and the
verb skopein, meaning to view or observe. The result is an apt term for the procedure of peering into
the recesses of the living body. But there is more to the term than that. Skopein means not to merely
look at something but rather to view with a purpose, to observe with intent, to monitor. The ancient
Greeks had no equivalent for endoscopy, but being clever and wise, they would have understood its
modern meaning and likely would have admired its choice.
Inception of Gastrointestinal Endoscopy

The original concept of using a tube to peer into the hidden cavities of the living human body is easily
understood, but the impediments that confronted early experimenters may be difficult for today's
endoscopists to comprehend. First was the matter of design. Few channels and cavities in the living
body lend themselves to inspection by a simple, straight, rigid, hollow probe. Second was a lack of
materials suited to the construction of a proper endoscope. In the early 19th century, a variety of
metals was available, but machines to fashion metals of high tensile strength were relatively crude.
Rubber was known, but in short supply, and the process of vulcanizationwhereby rubber is rendered
strong and elasticwas not discovered until 1839. Plastics as we know them were unheard of. Third,
and the most serious impediment, was lack of adaptable light. Before the invention of the incandescent
bulb, the only convenient source of artificial light was a burning candle, the wick of an oil or gas lamp,
or an exposed, glowing, platinum wire. None of these was adequate. This was the setting in which
gastrointestinal endoscopy had its humble beginnings.1(1)
A footnote to these former times concerns the famous clinician William Osler who attended the August
1879 meeting of the American Association for the Advancement of Science where he encountered
Thomas Alva Edison. The prolific inventor had just demonstrated the marvel of his newly devised
incandescent bulb. Osler remarked that Edison, in conversation, expressed his belief that now "it would
be possible to illumine the interior of the body by passing a small electric burner into the stomach."2(2)
Osler may have given this only passing thought, but Edison's casual comment would prove prescient.
The earliest recorded attempt at endoscopy was by Phillip Bozzini3(3) of Mainz and Frankfurt who, in
1806, devised a tin tube illuminated by a wax candle fitted with a mirror. With this instrument, which he
called einer Lichtleiter (a light conductor), he tried to peer into the urinary tract. On hearing of this, the
medical faculty at Vienna derisively dismissed the preposterous idea of "a magic lantern in the human
body."
Much of the subsequent development of the endoscope, before the advent of gastroscopy, took place
in France. In 1826, Pierre Salomon-Segalas introduced a "urethro-cystic speculum," the first
endoscope of any practical use in diagnosis.4(4) Unfortunately, faulty illumination hampered the utility
of Segalas' endoscope. Jean-Pierre Bonnafont, in the 1830s, offered technical changes that partially
solved this problem, and his designs were incorporated in the endoscope that Antonin-Jean
Desmoreaux devised in the 1850s.5,6(5)
Early Attempts at Gastroscopy

Adolf Kussmaul, a German physician who lived from 1822 to 1902 and who is perhaps more often
remembered for his description of air-hunger as a symptom of diabetic acidosis, is generally credited
with fashioning and employing in 1868 what might be hailed as the first gastroscope.7(6) In a plethora
of publications related to the gastrointestinal tract, it is curious that he hardly mentioned his efforts to
devise a gastroscope. For an account of Kussmaul's contribution to endoscopy, one must turn to the
W.B. Saunders Company items and derived items copyright 1999 by W.B. Saunders Company.
recollections of others.8(7)
Kussmaul's instrument was a straight, rigid, metal tube, passed over a previously inserted flexible
obturator (Figure 11A). The light source was a Desmoreaux lamp (Figure 11B) that burned a mixture
of alcohol and turpentine, illumination being concentrated by a reflector and lens. The first subjects for
experimentation were recruited from the ranks of sword-swallowers who performed at country fairs.
Legend has it that one of the first recruits balked when he saw the prototype instrument, exclaiming,
"I'll swallow a sword anytime, but I'll be damned if I'll swallow a trumpet!" Needless to say, Kussmaul's
tube, although marking an epoch, was hardly practical.
(8)Figure 11. A, Kussmaul's original simple tubular endoscope with obturator in place
(1868). B, Desmoreaux's lamp that provided endoscopic illumination, of a sort, before the
advent of the incandescent electric bulb. C, Mikulicz's rigid, angled gastroscope (1881); note
the provision of a rubber bulb for insufflation of air. (A-C, From Walk L. The history of
gastroscopy. Clio Med 1965; 1:20922.)
Kussmaul's interest in the gastroscope, despite his inability to produce a practical instrument, had a
direct and lasting influence on the design of early instruments. Joseph Leiter, the Viennese instrument
maker responsible for producing several variants of the esophagoscope and gastroscope, visited
Kussmaul at Strasbourg in 1876 and again in 1880 to confer on technical matters. Leiter returned to
Vienna impressed by Kussmaul's demonstration that a straight tube was preferable. When later he
collaborated with Mikulicz in the design of a gastroscope, Leiter recommended the straight form.7(9)
The first clinician to seriously consider the special requirements of a workable gastroscope was
Johann von Mikulicz-Radecki (18501905), a Polish surgeon more remembered for a variety of
innovative operations. Mikulicz understood that the longitudinal axis of the esophagus is not the same
as that of the stomach. Therefore, in 1881, he reported his design of a tube, 65 cm long and 14 mm in
diameter. With a slight angle in its distal fourth, this instrument employed an optical system and was
equipped to insufflate air (see Figure 11C).9(10) Illumination was originally supplied by an exposed,
electrically activated, glowing, platinum wire, but this was soon replaced by a miniature incandescent
globe (then called, by an odd coupling of French and German, a Mignon Lampchen; devised by Max
Nitze, developer of the first useful cystoscope). It was with this instrument, later refined by
othersnotably Rosenheim,10(11) Kelling,11(12) and Elsner12(13) (Figure 12A)that the first
informative views of the intact, living esophagus and stomach were obtained.
(14)Figure 12. Early gastroscopes devised by Elsner in 1911 (A), and modified by Schindler
in 1922 (B); the diameter of both was 11 mm. C, Schindler's semiflexible gastroscope
introduced in 1932. The original model had a sponge rubber ball at the tip, which Schindler
later found to be a hazard; it was replaced by a tapered, finger-like rubber tip. (A-C, From
Walk L. The history of gastroscopy. Clio Med 1965, 1:20922.)
A Cleveland surgeon, F. C. Herrick, in 1911 proposed intraoperative gastroscopyinserting a modified
cystoscope through a small gastrotomyas a means of locating a site of bleeding in the
stomach.13(15) Among the early accounts of peroral esophagogastroscopy in the United States are
those by Chevalier Jackson14(16) of Philadelphia and by Janeway and Green15(17) of New York.
To understand the compulsion of early investigators to find a means, however crude, of diagnosing
diseases of the stomach, one must remember that the earliest efforts to develop gastroscopy preceded
by decades the advent of gastrointestinal fluoroscopy and radiography. The remarkable power of
x-rays was discovered by Wilhelm Roentgen in 1895.
At the juncture of the 19th and 20th centuries, the most notable accomplishments in endoscopy were
by German workers. The reason was the technical supremacy in optics and instrument fabrication by
German artisans. However, these skills were soon mastered by American craftsmen, among the
foremost being Reinhold Wappler of New York, who later organized the firm of American Cystoscope
Makers, Incorporated, now better known by the acronym ACMI.
The Remarkable Career of Rudolf Schindler

Beginning with the use of instruments such as those devised earlier by Mikulicz and by Elsner, Rudolf
Schindler embarked on his career of lifelong dedication to the advancement of gastroscopy. Why is the
career of Schindler so pivotal? Because this one man, almost single-handedly, provided an impetus for
the method when the ranks of erstwhile proponents were in disarray. Moreover, this same man
sparked the promotion of gastroscopy throughout Europe and, later, in the Americas. One can argue
that others, sooner or later, would have played the same role. True, but also beside the point. Schindler
was the man.
Rudolf Schindler was born in Berlin on May 10, 1888, the son of a banker and an artistically gifted
mother who encouraged his early interest in classical music, poetry, and natural history. Fascination
with marine biology led to collection of sea shells, a lifelong hobby. Young Schindler's decision to study
medicine, we are told by Audrey Davis16(18) (whose perceptive account of Schindler's career is
recommended to the reader, along with that by Martin Gordon and Joseph Kirsner17(19)), derived
from his respect and admiration of his maternal uncle Richard Simon, a Berlin ophthalmologist.
On graduation from the University of Berlin, Schindler undertook a rural practice that was soon
interrupted by service in the German army during World War I. In this bleak setting, he observed the
prevalence of alimentary ailments in both the military and the civilian populations, and he became
convinced that much digestive disability could be attributed to morbid changes in the gastric mucosa,
undetectable by conventional methods of diagnosis. An answer to the problem, Schindler concluded,
would emerge only from direct observation of the internal milieu of the living stomachin short, by
gastroscopy.
On his return to practice, Schindler attended patients at the Munich-Schwabing Hospital where, in
1920, he acquired a rigid Elsner gastroscope that had been discarded by someone whose interest had
flagged. With this instrument and a later model modified according to Schindler's own suggestions (see
Figure 12A and 12B)both constructed by the Munich firm of Reiniger, Gebbert, and
SchallSchindler performed hundreds of gastroscopic examinations, carefully documenting each
procedure. Often gastroscopy was repeated in the same subject so as to observe the evolution of
mucosal lesions. Nor did Schindler neglect to scrutinize the normal stomach. He is said to have
examined on a number of occasions the stomach of his surprisingly willing housekeeper. This early
work, conducted with meticulous care and often in an atmosphere of hostility toward the procedure,
culminated in the publication in 1923 of Schindler's monumental "Lehrbuch und Atlas der
Gastroskopie."18(20) This exposition, like none before it, elicited serious interest in gastroscopy by
clinicians the world over.
In the era before intragastric photography was feasible, gastroscopic images were vividly rendered by
colored drawings or paintings, usually by professional artists who were allowed glances through the
endoscope, but sometimes by the examining physicians themselves (Figure 13). Both Schindler's
1923 classic and the atlas compiled in 1937 by Kurt Gutzeit and Heinrich Teitge19(21) can still be
perused for delight and instruction, especially for their depiction of stomach lesions now seldom
seen.20(22)
(23)Figure 13. Color rendition of endoscopic views by an artist thought to have been
employed by Dr. Rudolf Schindler. A, Cricopharyngeal sphincter in open and closed

positions. B, Gastric ulcers. (A and B, Courtesy of Dr. Eric R. Lee.)
Meanwhile, Schindler was far from satisfied with the efficacy and safety of the rigid gastroscope. He
devoted himself to the design of an instrument that could be inserted with less discomfort and risk to
the patient and that would provide clearer, more extensive images for the examiner. Schindler was
joined in this effort by Georg Wolf, a skilled Berlin instrument maker. The collaboration of Schindler
and Wolf, even though marked by occasional dissension, is an early and prime example of the alliance
between clinician and engineer needed to bring to fruition new and useful ideas in the advancement of
diagnostic and therapeutic instrumentation.
The idea shared by Wolf and Schindler was to construct an optical gastroscope wherein the light rays
that conveyed the image could be made to follow a flexible arc. The optical principle of articulated
prisms and lenses within a bendable tube had been established by Michael Hoffmann21(24) in 1911,
but its incorporation in a workable endoscope proved difficult. Wolf had contrived a prototype offered
for use in the inspection of condensing pipes in steam engines. (Flexible endoscopes are used to this
day to inspect otherwise inaccessible internal mechanisms of various types of machinery.) Wolf first
proposed and constructed an optical gastroscope that was flexible throughout its length, but Schindler
had the better idea of combining a rigid proximal half with a flexible distal half, thus producing a more
wieldy instrument that conveyed a brighter, clearer image. The result was the famous Wolf-Schindler
semiflexible gastroscope, first produced in 1932 (see Figure 12C). In retrospect, what was considered
"semiflexible" by the examiner more than likely was felt to be "semirigid" by the examinee.
With this instrument, Schindler22(25) was able to announce to the medical world the advent of an
acceptably safe, workable gastroscope capable of conveying informative images of the stomach's
interior to the eye of the examining physician (Figure 14). Schindler's achievement was readily
recognized, and clinicians flocked to Munich to observe and learn the new technique. Of these,
Schindler recognized the following workers as influential in propagating the method: Francois Moutier
of France (who published his own "Traite de gastroscopie et de pathologie endoscopique de
l'estomac"23(26) in 1935), Norbert Henning of Leipzig, and Kurt Gutzeit of Breslau. Among Schindler's
American students were Samuel Weiss, a New York gastroenterologist, and Edward Benedict, a
Boston ear-nose-and-throat surgeon. It should be mentioned that in certain prominent centers, such as
the University of Pennsylvania and the Mayo Clinic, peroral endoscopy in those early days was
considered to be solely in the province of the specialty then known as bronchoesophagology.
(27)Figure 14. A, Schindler's semiflexible gastroscope with accessories. B, Proximal

viewing end with insufflation bulb and electrical connection. C, Distal end of semiflexible
gastroscope. Note side-viewing optical design and tapered rubber tip. (C, Courtesy of Dr. Eric
R. Lee.)
With the darkening clouds of Nazism hovering over Germany, Rudolf Schindler, whose father was
Jewish, was subjected to increasing persecution by the anti-Semitic regime. He was befriended by two
American colleagues, Dr. Marie Ortmayer and Dr. Walter Lincoln Palmer of Chicago, who had visited
Schindler's clinic in the 1920s. Hearing of his plight, they invited Schindler to come to the University of
Chicago as a visiting professor. In 1934, Schindler managed his escape to the United States, where he
found an opportunity to practice and teach his endoscopic method at the Billings Hospital. His
publications, now in English, proliferated. Walter Palmer provided unstinting assistance to Schindler in
publishing his textbook Gastroscopy in 1937. This, with subsequently revised editions24(28) appearing
in 1950 and 1966, was the standard work on gastroscopy for a generation of clinicians. To be noted is
that Schindler subtitled his book The Endoscopic Study of Gastric Pathology. Thus, Schindler made
clear his belief in the endoscopic method as an approach to an understanding of diseases and to the
care of patients, not as an end in itself. Schindler always considered himself to be first a physician and
only secondarily one who was skilled in technology.

On settling in Chicago, Schindler ever sought to improve the design and construction of the
gastroscope. He allied himself with William J. Cameron, whose Cameron Surgical Specialty (on
Instruments) Company became the world's largest supplier of illuminated instruments. On Cameron's
staff was a talented instrument maker, Louis Streifeneder, with whom Schindler developed a close
rapport. With the supply of German equipment cut off during World War II, the Cameron Omni-angle
gastroscope, modeled closely after the original Wolf-Schindler design, became the standard
instrument used in many clinics throughout the United States. Streifeneder later formed his own firm,
the Eder Instrument Company, that continues a tradition of meticulous workmanship, now mainly
directed to the production of laparoscopes.
In 1943, Schindler departed Chicago to reside in Los Angeles, where he taught at the College of
Medical Evangelists (now Loma Linda University). Later, he served as consultant at the Long Beach
Veterans Hospital, where his principal coworkers were Dr. Stephen Stempien and Dr. Angelo Dagradi.
At age 70, Schindler learned Portuguese and accepted appointment to the faculty of the University of
Minas Gerais at Belo Horizonte, Brazil. His wife's ill health necessitated their return to the United States
in 1960.
No account of Schindler's career would be complete without mention of his wife, Gabrielle (ne
Winkler), whom he wed in 1922. The use of the rigid and semiflexible gastroscopes required a capable
assistant who could hold and manipulate the patient's head. Moreover, it was the assistant who closely
monitored the patient's reaction to examination and who sought to soothe the patient's apprehension
and discomfort. Although untrained as a nurse, Gabrielle learned to perform the function of
gastroscopic assistant to perfection (Figure 15). It is said that on the rare occasion when Gabrielle
was unavailable to assist him, Schindler declined to perform gastroscopy.
(29)Figure 15. The Schindler team at work at the University of Chicago Clinics about 1940.
Rudolf Schindler peers through his semiflexible gastroscope as Gabrielle Schindler exhibits
her art of head-holding. Note that her left index finger aids the drainage of oropharyngeal
secretions. (Courtesy of Dr. Martin Gordon. From Gordon ME, Kirsner JB. Rudolf Schindler,
pioneer gastroscopist; glimpses of the man and his work. Gastroenterology 1979, 77:35461.)
From 1960 to 1965, Schindler again served as consultant at the Long Beach Veterans Hospital. On
him in 1962 was bestowed the first Schindler Award by the American Society for Gastrointestinal
Endoscopy (ASGE). That occasion was his last attendance at an annual meeting of the group he had
originally founded as the American Gastroscopic Club in 1941.25(30)
Schindler's beloved Gabrielle died in 1964. He then married Marie Koch, a friend of long standing from
Munich, and in 1965, he returned to that city where much of his early work had been accomplished. He
occupied himself by preparing an atlas based on his lifetime performance of 10,300 gastroscopic
examinations. Also, he found time to enjoy his music and his shell collection. Rudolf Schindler died of
heart disease in Munich in 1968.26(31)
Gastroscopy As It Was Performed During the Era of Rudolf Schindler

Perhaps the reader will find of interest a brief account of what it was like to perform gastroscopy in the
era before the advent of fully flexible instruments. In most centers, gastroscopy was perceived as an
extraordinary step to be undertaken only when all other diagnostic means had been found inadequate
or inconclusive. Properly trained and experienced gastroscopists were relatively few. The patient was
carefully selected, the need for gastroscopy always being predicated on a previous barium meal
radiographic examination. There was no talk of gastroscopy being a first-line venture in those days,
with the possible exception of the "vigorous diagnostic approach" to the bleeding patient as advocated
by Dr. Eddy Palmer.27(32)

Premedication usually consisted of an intramuscular injection of a short-acting barbiturate. Atropine
was often added to reduce gastric secretion and motility. The lens system gastroscope was not
equipped with a suction channel, and for this reason, the stomach was first routinely emptied by use of
an Ewald tube. If one followed Schindler's carefully specified method, the patient's throat was
anesthetized by the injection of a numbing solution through a 25-cm rubber tube with side openings at
its distally tapered end. This tube, when fully inserted, extended through the upper esophageal
sphincter. This maneuver had the further advantage of slightly dilating the sphincter, thus facilitating
passage of the gastroscope.
The patient lay in the left lateral decubitus position, with an experienced and sturdy nurse-assistant
firmly gripping the sides of the patient's head. The operator's index and middle fingers guided the
tapered, soft rubber tip of the gastroscope into the proximal esophagus. The assistant then firmly and
fully extended the patient's head so as to provide a straight mouth-throat-esophageal channel, allowing
smooth passage of the endoscope into the stomach. Needless to say, one hoped for an edentulous
patient whose mouth would more easily accommodate the instrument and whose absent teeth would
not grate on the operator's fingers or the proximal steel rod of the fully inserted gastroscope.
Even the more advanced semiflexible gastroscopes (Figure 16) employed a side-viewing objective;
therefore, the examiner had no view of the esophagus on either insertion or withdrawal.
Esophagoscopy required a quite different instrument (Figure 17) and was an altogether separate
procedure.
(33)Figure 16. A, Eder-Palmer semiflexible gastroscope. Note side-viewing optics. B,

Eder-Palmer instrument with tip deflected.
(34)Figure 17. A, Eder-Hufford rigid esophagoscope. Note, from top to bottom, the
obturator, lens assembly, examination tube, and illumination rod. B, Eder-Hufford rigid
esophagoscope assembled for use.
Once the gastroscope was within the stomach, air was insufflated. The distal half of the instrument,
with its lens and prisms in a rubber sheath, was partially flexible to accommodate the contour of the
stomach, but its flexion was uncontrollable by the examiner. One could, however, precisely manipulate
torque; the external handle, carrying electric current to the small (and sometimes unreliable)
incandescent light bulb at the distal end, always pointed in the direction of the objective. Thus,
orientation of the view in the stomach was exact. A problem was that inspection of the stomach lining
was somewhat limited. The pylorus was usually seen but never transgressed. The lesser curve of the
antrum was usually a "blind area," as was the area underlying the tip of the instrument.
The examiner had to be content with relatively fleeting glimpses. Discomfort to the patient, whose neck
was rigidly extended, hardly encouraged a leisurely look. The gastroscopist's own visual impression
usually had to suffice. Photographic documentation was managed by only a few accomplished
operators.28(35) There was no provision for insertion of a biopsy forceps, except in the special
"operating gastroscope" devised by Dr. Edward Benedict,29(36) an instrument found too cumbersome
by most gastroscopists.
Thus it was that gastroscopy was performed in the decades from the 1930s to the 1960s. Despite the
formidability of the instruments then available, complications were relatively rare,30(37) doubtless
because of the caution sternly inculcated by preceptors in those who would aspire to be practitioners of
the art. But such constraints may explain why, as late as 1966, there were only 268 members of the
ASGE.
The Brief but Illuminating Era of the Gastrocamera

An impediment to the appreciation of early gastroscopy was that views of the internal milieu of the
stomachand such lesions as might be seenwere limited to the visual cognizance of the examiner
or to such fleeting glimpses as might be permitted an observer peering over the examiner's shoulder.
In a way, this was an advantage to the endoscopist, who could wax eloquently about what was seen,
and none could dispute it. As might be expected, skeptics abounded, especially when the topic was
gastritis.
The first challenge to endoscopic interpretation was the introduction in the late 1940s of the gastric
mucosal biopsy tube devised by Ian Wood and his Australian associates.31(38) Biopsy required a
separate procedure, but then the pathologist had a section of tissue with which to counter the
endoscopist's mental imageoften imperfectly recollected.
The 19th-century invention of photography provided a means of recording a permanent image, but
because of obvious technical inadequacies, it was long before photography could be adapted to
endoscopy. Lange and Meltzing,32(39) as early as 1898, actually tried to take black-and-white
photographs at gastroscopy, but lack of sufficient light and the slowness of the then-available film
emulsions discouraged further use of the method.
Another idea was to construct a miniature camera, small enough to be swallowed by the patient and
attached to a tube through which the camera could be actuated and then retrieved (Figure 18). This
concept was ideally suited to implementation by the Japanese. Thus was developed a workable
gastrocamera by Tatsuno Uji33(40) in collaboration with his engineering colleagues at the Olympus
Optical Company (Figures 19 and 110).
(41)Figure 18. Diagram of the disassembled distal tip of the gastrocamera apparatus; at the
proximal end of the connecting tube were a control unit and a rubber bulb for air insufflation
of the stomach. (From Perna G, Honda T, Morrissey JF. Gastrocamera photography. Arch
Intern Med 1965; 116:43441. Copyright 1965, American Medical Association.)
(42)Figure 19. Early model gastrocamera manufactured by the Olympus Optical Company.
(Courtesy of Olympus Optical Company, Ltd.)
(43)Figure 110. Later gastrocamera model GTF-A manufactured by the Olympus Optical
Company. Note small 5-mm film cartridge placed near the distal tip of the instrument.
(Courtesy of Olympus Optical Company, Ltd.)
Dr. John Morrissey has related the story34(44) of how Dr. Yoshio Hara brought the first gastrocamera
to the University of Wisconsin in 1962. Hara had come to the Madison campus to study cancer
chemotherapy. He happened to bring along a set of vivid, full-color photographs of lesions as they
appeared in the living stomach. He brought, too, the tiny camera that had captured them. Morrissey
was quick to recognize the utility of the instrument and became adept in its use. Soon scores of
clinicians journeyed to Madison, eager to become gastrophotographers. Gastrocameras,
gastroprojectors, and gastroscreens proliferated across the land (although never as many as in Japan,
where according to Morrissey, in 1966, there were 10,000 gastrocameras taking pictures in half a
million subjects annually).

Although no finer photographs of the intact stomach have ever been obtained by any other means (the
reader is invited to inspect the color plates in Morrissey's 1965 article35(45)), the early gastrocameras
had a major disadvantage: The operator could not see what was photographed until the film was
developed. Pictures (32 exposures on a roll of 5 mm film) were taken in rote sequence (Figure
111).36(46) If all went well, the result was a fairly complete survey of the stomach's interior. However,
a gastroscope was still needed to provide a visual image. The solution: Attach the gastrocamera to a
gastroscope. This was done in 1963.
(47)Figure 111. Scheme of the prescribed sequence in which gastrocamera photographs

were taken. The tip of the tube containing the miniature camera was first placed in the gastric
antrum. N indicates the neutral position; U1, U2, D1, and D2 indicate flexion of the tip up or
down, 15 degrees or 35 degrees, respectively. The sequence was repeated as the tube was
moved proximally. (From Perna G, Honda T, Morrissey JF. Gastrocamera photography. Arch
Intern Med 1965; 116:43441. Copyright 1965, American Medical Association.)
Meanwhile, the acuity of the fiberoptic endoscope was rapidly improved, as were illumination and film,
and soon it was much easier for most endoscopists to simply attach an external 35 mm camera to the
eye piece of the gastroscope and thereby record the image conveyed by the fiber bundle. Insertion of
the miniature camera into the colon was suggested,37(48) but then came the fiberoptic colonoscope.
Ease of use prevailed. The gastrocamera, marvelous as it was, became obsolete.
The Advent of Fiberoptics

A Transatlantic Coalition
The era of fiberoptic gastrointestinal endoscopy might be said to have dawned on a wintry day in Ann
Arbor, Michigan, when Basil Hirschowitz picked up the January 1954 issue of Nature. What caught his
eye were two articles on the optical properties of fine glass fibers. One of them in particular, by
Hopkins and Kapany38(49) of the Imperial College of Science and Technology in England, fired his
imagination of how an endoscopic image might be transmitted by a coherent bundle of fully flexible
glass fibers from the alimentary tract of a patient to the eye of an examiner. The ensuing events have
been described in fascinating accounts by Hirschowitz39(50) and by Hopkins.40(51)
That light would follow the curved path of a stream of water pouring from a tank was first demonstrated
by John Tyndall, a British physicist, in 1870. The idea of using flexible glass fibers to propagate light
was proposed in patent applications by J. L. Baird of England in 1927 and by C. W. Hansell of the
United States in 1930. Schindler credited Heinrich Lamm41(52) with being the first to recommend the
adaptation of fiberoptics to gastroscopy. Unfortunately, Schindler gave no further thought to the matter,
as he was then preoccupied with developing his own system of lenses and prisms that he knew to be
feasible.
Light-carrying bundles of glass fibers are of two types: incoherent and coherent. Incoherent or random
bundles intended only to transmit illumination are easily constructed. Quite another matter is
construction of a coherent bundle intended to convey an image, wherein many thousands of hair-thin
fibers must be in precisely the same array at both ends of the bundle. Any appreciable leakage of light
from one fiber to another defeats the purpose of a coherent bundle. This principal impediment to the
development of a fiberscope was overcome by Lawrence Curtiss, then an undergraduate physics
student at the University of Michigan. Curtiss devised a process whereby an individual fiber of optical
glass could be clad in a layer of glass of lower refractive index that then served as an insulating coat.
Hirschowitz, with whom Curtiss collaborated, credits this as the single most important innovation in the
advent of fiberoptic endoscopy.39(53)
Development of the Fiberoptic Endoscope

A prototype fiberscope, adapted for gastroscopy, was finally constructed at Ann Arbor in January 1957
(Figure 112). A photograph of an image transmitted by this prototype is illustrated in Figure 113. The
photograph was made with an Exacta 35 mm camera that is still used by Dr. Hirschowitz (Figure
114). Hirschowitz manfully undertook to be the first to swallow this oversized worm. Unscathed by the
experience, within a few days at the University of Michigan Hospital he performed the first fiberoptic
gastroscopy, the examinee being a young woman who harbored a duodenal ulcer. Hirschowitz then
demonstrated his instrument to a small, polite, but less than enthusiastic audience later that year at the
annual session of the American Gastroenterological Association.42,43(54)
(55)Figure 112. First prototype flexible fiberoptic endoscope. (Courtesy of Dr. Basil
Hirschowitz.)
(56)Figure 113. Photograph of image seen with the first prototype fiberscope constructed by
Basil I. Hirschowitz, Larry Curtiss, and C. Wilbur Peters at the University of Michigan in
January 1957. (Courtesy of Dr. Basil Hirschowitz.)
(57)Figure 114. Exacta 35-mm camera attached to prototype flexible fiberoptic endoscope.
(Courtesy of Dr. Basil Hirschowitz.)
It was not until the latter 1960s that meticulous technicians, both Japanese and American, succeeded
in refining construction of the first workaday fiberoptic endoscopes. Notably improved were optical
clarity, wieldability, and control of the distal tip. Channels were provided for biopsy and therapeutic
maneuvers. The Japanese were especially intent on this work, largely because of their need to
precisely diagnose gastric cancer, so prevalent in their country.
The first fiberoptic instrument to gain widespread use in the United States was the esophagoscope with
a working length of 75 cm. This had the enormous advantage of an end-viewing objective rather than
the traditional side-viewing objective with which Hirschowitz' early fiberscope was equipped. To
clinicians previously obliged to attempt endoscopy of the esophagus by means of a straight, rigid, steel
tube, the fiberoptic esophagoscope was a marvel. Even more pleased with the new instrument, one
can be sure, were the patients. Soon thereafter, the coherent fiberoptic bundle was extended to a
working length of 110 cm, thus providing an easily insertable instrument with which the esophagus,
stomach, and duodenum could be thoroughly and unhurriedly examined, all at the same sitting.
Improvements in endoscope design were so numerous and so rapidly introduced during the early
1970s that one could hardly purchase a new instrument and become acquainted with its use before it
was rendered obsolete by a new model. The ungainly device that had its inauspicious beginning in the
corner of a physics laboratory in Ann Arbor was transformed into an invaluable instrument. In the
hands of a host of clinicians, for a full generation, fiberoptic endoscopy notably advanced the diagnosis
and management of patients with gastrointestinal disease.
The New Era of Video Endoscopy

The rapid pace of advancing technology is further illustrated by the observation that the same
generation that so benefited by the advent of fiberoptics also witnessed its evanescence. Fiberoptics
dominated the field for barely a quarter century. The burgeoning technology of video endoscopy has
now largely superseded use of the coherent, image transmitting, fiberoptic bundle for most
gastroenterologic applications.
Vividly accurate documentation of a clear image has been a goal long sought by gastrointestinal
endoscopists. Earlier mention was made of determined efforts to capture by still or cine photography
what could otherwise be witnessed only by the endoscopist.28,44(58) The advent of television
cameras and monitors naturally suggested their adaptation to endoscopy. Black-and-white televised
images of bronchoscopy were reported in 1957.45(59) Rider and Hirschowitz collaborated in a
demonstration of televised images of upper gastrointestinal endoscopy at the 1963 meeting of the
ASGE. Television cameras were devised that could be attached to the eye piece of standard fiberoptic
gastrointestinal endoscopes,46(60) but the cameras were unwieldy and heavy (about 7 lbs) and had to
be suspended from a ceiling. What was needed was a means of incorporating within the endoscope
itself an electronic device capable of registering and transmitting a vivid image. This became feasible
with the innovation of the charge-coupled device.
The first video endoscope was introduced by Welch Allyn Inc. of Skaneateles, NY, a long-established
maker of illuminated diagnostic instruments. The earliest report of clinical experience with prototype
video endoscopes is that by Sivak and Fleischer47(61) in the United States, this being followed in short
order by those of Classen and Phillip,48(62) and Demling and Hagel49(63) in Europe.
Proctosigmoidoscopy and Colonoscopy

The success of the fiberoptic endoscope introduced through the mouth naturally led to its adaptation
for insertion at the nether end of the alimentary tract. Specula for examination of the anus and rectum
had been in wide use since the late 1800s. For practical purposes, the maximum length of a straight,
hollow, rigid proctosigmoidoscope had been found to be 25 cm. An instrument similar to the
semiflexible gastroscope would have been handy, but the tortuosity of the sigmoid colon was far
beyond the turning capacity of a lens-and-prism endoscope. Even the early fiberscope could not be
properly advanced in a retrograde manner into the colon, although a few venturesome examiners
tacitly tried to use the fiberoptic esophagoscope in this manner from time to time. For the development
of a workable fiberoptic colonoscope, we return to the University of Michigan.
According to Bergein F. Overholt,50(64) who pioneered development of fiberoptic colonoscopy, the
work was stimulated by an unusually disagreeable proctosigmoidoscopic examination suffered by the
person of Dr. J. Howard Gowan. In 1961, Overholt was an intern at the University Hospital in Ann
Arbor. As part of his application for a U.S. Public Heath Service fellowship, Overholt was interviewed by
Gowan, who had just undergone a somewhat trying physical checkup at the hospital. Overholt, being
well acquainted with the principles and promise of fiberoptics, commented on the prospect of a more
comfortable sigmoidoscopic procedure.
Meanwhile, others had been tempted to set aside esthetics and insert the fiberoptic gastroscope in the
anus, a procedure later reported.51(65) Also, all manner of contrivance had been suggested to pull
through the colon a fiberscope hooked onto the end of a swallowed string. But, just as the peculiar
anatomy of the proximal alimentary tract had posed a problem for Kussmaul and the earliest would-be
gastroscopists, so did the serpentine anatomy of the lower bowel present a problem to would-be
colonoscopists.
Overholt, with the support of his mentor, Dr. H. Marvin Pollard,52(66) strived to overcome this difficulty.
From a silicone-rubber cast, he devised a lifelike model of the human distal colon, thus enabling the
necessary adjustments in torque and control of a fiberoptic endoscope that resulted in a prototype
instrument first employed clinically in 1963 (Figures 115 and 116). Further refinement was required,
and it was not until the 1967 meeting of the ASGE that Overholt described, to a somewhat skeptical
audience, his experience in examining the first 40 patients.53(67) By coincidence, this event took place
at the same locale, the Broadmoor Hotel in Colorado Springs, where Hirschowitz, also representing the
University of Michigan, had first described fiberoptic gastroscopy exactly 10 years earlier. Meanwhile,
Japanese workers were busily engaged in improving the colonoscope and advancing its clinical
application.54,55(68)
(69)Figure 115. Early prototype fiberoptic sigmoidoscope developed at the University of

Michigan in conjunction with the Illinois Institute of Technology Research. (Courtesy of Dr.
Bergein Overholt.)
(70)Figure 116. Prototype fiberoptic sigmoidoscope made by the Eder Instrument Company.
(Courtesy of Dr. Bergein Overholt.)
Colonoscopy gained rapid acceptance by clinicians who had been recently introduced to the marvel of
fiberoptic gastroscopy, although most operators soon found that successful colonoscopy was much
more tedious and demanding than examination of the upper gastrointestinal tract. In England,
Christopher Williams,56(71) and in the United States, Hiromi Shinya,57(72) and Jerome Waye58(73)
and their colleagues were among the early proponents and most proficient teachers of the new art of
fiberoptic colonoscopy. Thereby, an additional, clinically important segment of the alimentary tract was
placed within the purview of gastrointestinal endoscopists.
Endoscopic Retrograde Cholangiopancreatography

The intricacies of the bile and pancreatic ducts had long intrigued physicians. The fascinating story of
visualizing the biliary tree begins in 1924 with the demonstration by surgeons Evarts Graham and
Warren Cole that intravenously administered iodinated phenolphthalein was selectively excreted in bile.
This served as the prototype of a contrast agent whereby the configuration of the biliary tree could be
recorded radiographically. However, even with the later use of improved agents, the radiographic
image was often indistinct, and the pancreatic duct defied demonstration. In 1955 Doubilet et al.59(74)
described intraoperative pancreatography, but what was needed was a means of obtaining a clear
image of the pancreatic duct without surgical intervention. In 1965, Ravinov and Simon60(75) reported
cannulation of the duodenal ampulla by a swallowed catheter manipulated under fluoroscopic
guidance. Their painstaking procedure was successful in only one of eight patients.
Why not guide a cannula into the ampulla under endoscopic control? This was first accomplished by
William S. McCune et al.61(76) at George Washington University in 1968. Again, it remained for the
Japanese to perfect the technique that was described by Oi et al.62,63(77) in 1970 and by Kasugai et
al.64(78) in 1972. Foremost among proponents of endoscopic retrograde cholangiopancreatography in
the United States were Jack Vennes et al.65(79) at the Minneapolis Veterans Administration Hospital.
For his masterful teaching of the method, Vennes received the 1978 Schindler Award of the
ASGE.66(80)
Choledochoscopy
Just as early clinicians were stimulated to develop gastroscopy because of dissatisfaction with existing
means of diagnosis, so surgeons have sought a better means of detecting lesions within the biliary
ducts. To view inside the living common bile duct was first attempted in 1923 by Bakes,67(81) a
German surgeon better known for devising the graduated probes commonly used at operation to dilate
the bile ducts and the ampulla of Vater. A right-angled telescope specifically constructed for the
common bile duct was reported by McIver68(82) in 1941, and improved instruments were described 20
years later by Wildegans69(83) and Berci.70(84) Predictably, fiberoptic technology was soon applied in
choledochoscopy,71(85) but there was then a return to the use of a rigid instrument equipped with
prisms and lenses,72(86) because it was more wieldy and conveyed a brighter, clearer image. This is
one of the few instances in the development of endoscopy wherein the rigid lens system prevailed, for
a while, over the flexible fiberoptic system. A more recent innovation, however, has been adaptation of
a video urethroscope for the purpose of both diagnostic and therapeutic choledochoscopy under
laparoscopic guidance.73(87)
With the demonstrated feasibility of endoscopically guided cannulation of the ampulla of Vater with a
catheter for the purpose of contrast radiography, it was inevitable that substitution of a small-caliber
fiberoptic endoscope would be proposed. Indeed, Nakajima et al.74(88) reported this remarkable feat
in 1978. They modified a duodenoscope (the "mother 'scope") by enlarging to 2.8 mm the bore of its
catheter channel through which they passed a 2.3-mm fiberoptic endoscope (the "baby 'scope") that, in
turn, could be guided under direct vision into either the common bile duct or the pancreatic duct.
Because the exceedingly fine caliber of the probing endoscope could accommodate only 3000 fibers
and because the view was through a fluid medium, the image lacked the clarity associated with other
endoscopic techniques. Peroral cholangiopancreatography has not gained wide use; nevertheless, its
accomplishment must be regarded as a tour de force.
Laparoscopy
Laparoscopy has been given almost as many names as there were early promoters. This method, by
which the contents of the abdomen and pelvic cavities can be visualized telescopically, has also been
termed peritoneoscopy, organoscopy, and coelioscopyall meaning the same thing. Laparoscopy has
the advantage of historical precedence and, largely because of adoption by gynecologists and
surgeons, this term has now achieved almost universal currency.
In 1902, Georg Kelling75(89) reported at Hamburgto the 73rd meeting of a group known as German
Natural Scientists and Physicianshis observations of the abdominal viscera of a dog by means of a
Nitze cystoscope. In the same year, Dimitri Ott,76(90) a Russian gynecologist, deliberately introduced
a speculum into the pelvic cavity of a female patient through an incision in the posterior vaginal fornix,
a procedure he referred to as ventroscopy. Working independently in Sweden, H. C.
Jacobaeus77,78(91) reported his percutaneous endoscopic examinations of the abdominal and
thoracic cavities, and in 1912, he proposed the term laparoscopy. Thereafter, Kelling and Jacobaeus
long disputed the priority of discovery. In what may have been one of the earliest efforts at cost
containment, Kelling,79(92) in 1923, wrote that he used what he called coelioscopy as a means of
sparing his poor German patients, then in an abyss of economic depression, the expense of surgical
laparotomy.
Use of the procedure in the United States was first reported in 1911 by Bertram Bernheim,80(93) a
surgeon at the Johns Hopkins Hospital in Baltimore. He had inserted a proctoscope with a -inch bore
through the abdominal wall of a jaundiced patient to confirm the presence of a Courvoisier gallbladder.
Benjamin Orndorff81(94) of Chicago used a similar device and published the first report of an
extensive experience with what he called peritoneoscopy. Meanwhile, the procedure became widely
applied in Europe where Korbsch82(95) published his "Lehrbuch und Atlas der Laparound
Thorakoskopie" in 1923. Heinz Kalk, a renowned German proponent of laparoscopy, culminated a
series of publications with his own textbook and atlas, written in collaboration with Egmont Wildhirt in
1962,83(96) that emphasized the additional benefit of endoscopically guided liver biopsy.
John Ruddock84(97) of Los Angeles was almost a lone champion of the procedure in the United
States during the 1930s. Ruddock used a slightly modified cystoscope in hundreds of cases and
proposed what was then an astonishing number of therapeutic applicationsproposals that were
greeted with derision by mainstream physicians and surgeons. After World War II, with resumed
importation of refined German instruments and the immigration of physicians trained in their use,
laparoscopy won new adherents. Quick to perceive the value of laparoscopy were gynecologists whose
specialty now encompasses a majority of laparoscopists in the United States. Prominent among
endoscopically oriented gynecologists was Peter Steptoe of England, who found laparoscopy essential
in his development of the technique that resulted in a "test-tube baby" boom.85(98)
Very recent years have seen an astonishing adaptation of the laparoscopic approach to numerous
intra-abdominal manipulations, notably those directed to treatment of calculus biliary tract disease.
Pioneered in France,86,87(99) laparoscopic cholecystectomy has come, in many centers, to largely
supersede open laparotomy and has caused to be rewritten entire chapters dealing with the surgical
management of gallstones.
Therapeutic Endoscopy
As the science and art of endoscopy evolved, its proponents well understood the meaning of Mark
Twain's complaint, "Everybody talks about the weather, but nobody does anything about it." In former
years, endoscopists could regale their colleagues with descriptions of what they saw but were relatively
powerless to employ their skill in beneficial intervention. The prospect was soon to brighten.88(100)
True, the early bronchoesophagologists were adept in extracting foreign objects inadvertently or
deliberately choked down by misguided patients. For years, there was an astonishing display of
trophies of incredible variety, extracted from various recesses, that lined the walls of the Chevalier
Jackson Clinic at the Graduate Hospital in Philadelphia. Endoscopists no longer so decorate their
workplaces (the array of foreign bodies extracted by Jackson and his associates is now part of the
Mutter Museum collection at the College of Physicians of Philadelphia), but the ingenious application of
a variety of techniques employing gastrointestinal endoscopy became widely promulgated.89(101)
The therapeutic procedure currently and most widely used by gastrointestinal endoscopists, and
probably that of benefit to the greatest number of patients, is polypectomy. In Miami at the 1971
session of the American Gastroenterological Association, Shinya and Wolf showed motion pictures
depicting the innovative technique of colonoscopic polypectomy.57(102) Viewers were enthralled as
they watched an improvised snare approach the polyp. The audience erupted in a spontaneous cheer
when the severed polyp was seen to topple from its pedicle. The number of patients who have since
been saved from carcinoma by having benign polyps extirpated from segments of the gastrointestinal
tract accessible to endoscopy is now becoming evident.
Control of active bleeding from lesions within the range of endoscopyby electrocoagulation, heater
probe, or laserwas accomplished as a result of intense investigation during the decade following
1970. Endoscopically guided percutaneous gastrotomy was first performed in 1979, using an
improvised appliance.90(103) No longer are strictures in the alimentary tract merely observed by the
endoscopist; they are now effectively dilated under endoscopic guidance. Obstructing neoplasms can
be obliterated under direct vision by endoscopically directed laser beams or electrodessicators. Even
more dramatic is endoscopic intervention in cases of calculus disease in the biliary and pancreatic
ducts, previously cited in these pages. Many of the pioneering and most knowledgeable investigators in
this field present their work in chapters to be found in this book.
Sometimes what looms on the horizon and may seem new turns out not to be new at all. Being aware
of what has gone before can improve one's perspective. This is exemplified by endoscopic
sclerotherapy of esophageal varices, thought by some to be a relatively recent innovation. It is
notinterest in the procedure has simply been renewed. This means of treatment was first reported by
Clarence Crafoord and Paul Frenckner91(104) in Europe in 1939, and a few years later, in the United
States by Herman Moersch92(105) and by Cecil Patterson and Milford Rouse.93(106)
Epilogue
What can one learn from the history of endoscopy? Several recurring themes stand out.
First, almost without exception, advances in endoscopy have come about by virtue of close
collaboration between clinician and artisan; neither could have succeeded alone. One need only recall
the names of Mikulicz and Leiter, Schindler and Wolf, and Palmer or McCune and Streifeneder. It can
happen, too, that a notable advance issues from a merging of basically unrelated technologies. An
example is the advent of endoscopic ultrasonography reported in the mid-1980s94,95(107) and soon
after hailed by Sivak.96(108)
Second, those clinicians who have notably contributed to the advance of endoscopy were invariably
seeking an answer to a broader medical problem; none thought of himself as merely a technician
pursuing technology as a goal in itself. A case in point is the career of Rudolf Schindler.
Third, innovative instrumentation often has been prompted by technical advances made far afield of
the recognized domain of gastrointestinal endoscopy or even beyond the purview of medicine
generally. Instructive examples are the adaptation of fiberoptic technology and the charge-coupled
device to endoscopy.
Fourth, the progress of instrumentation and technique does not always proceed in a smoothly linear or
logical fashion, despite a semblance of order suggested by hindsight. Development often occurred in
repeated fits and starts. The pattern is familiar: astonishment at the first announcement, skepticism of
early enthusiastic reports, then a sorting out, with acceptance or rejection depending on the cautious
assessment by respected clinicians and one's own personal experience.
History is looked on as a humanistic discipline, but it has an essential role, too, in the understanding of
science. To know the history of endoscopy is to pay tribute to those who have cleared the path we now
tread, to recognize the impediments that have been overcome, to appreciate more fully the facility we
now enjoy, and to point to the prospect of still more marvelous advances to come.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Walk L. The history of gastroscopy. Clio Med 1965;1:20922.

Cushing H. The Life of Sir William Osler. Vol 1. Oxford: Clarendon, 1925;174.
Bozzini PH. Lichtleiter, eine Erfindung zur Anshauung innere Teile und Krankheiten. J Prakt
Heilk 1806;24:207.
Segal A. Pierre-Salomen Sgalas d'Etchepare, prcurseur de l'endoscopie moderne. Bull
Acad Natl Med 1978;162:70914.
Segal A. Le mdicin principal de Ire classe Jean-Pierre Bonnafont (18051891): Sa place
prponderante dans l'histoire de l'endoscopie au XIXe sicle. Histoire des sciences
mdicales. 1983.
Hillemand P, Gilbrin E. Antonin-Jean Dsmoreaux (18151894), le crateur de l'endoscopie.
Bull Acad Natl Med 1976;160:95100.
Kluge F, Seidler E. Zur Erstanwendung der sophago- und Gastroskopie: Briefe von Adolf
Kussmaul und seinen Mitarbeitern. Med Hist J 1986;21:288307.
Killian G. Zur Geschichte der Oesophago- und Gastroskopie. Dtsch Z Chir 1901;58:499512.
Mikulicz J. ber Gastroskopie und Oesophagoskopie. Wien Med Presse 1881;52:1629.
Rosenheim T. ber die Beschtigung der Kardia nebst Bemerkungen ber Gastroskopie.
Dtsch Med Wochenschr 1985;21:7404.
Kelling G. Endoskopie fr Speiserohre und Magen. Mnch Med Wochenschr 1987;44:9347.
Elsner HD. Ein Gastroskop. Klin Wochenschr 1910;47:5935.
Herrick FC. Profuse recurrent gastric hemorrhage, with report of cases and description of an
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
instrument for viewing the gastric interior at operation. Cleve Med J 1911;10:96976.
Jackson C. Tracheobronchoscopy, Esophagoscopy, and Gastroscopy. St Louis:
Laryngoscope. 1907.
Janeway HH, Green N. Esophagoscopy and gastroscopy. Surg Gynecol Obstet
1911;13:24553.
Davis AB. Rudolf Schindler's role in the development of gastroscopy. Bull Hist Med
1972;46:15070.
Gordon ME, Kirsner JB. Rudolf Schindler, pioneer gastroscopist: Glimpses of the man and his
work. Gastroenterology 1979;77:35461.
Schindler R. Lehrbuch und Atlas der Gastroskopie. Munich: IF Lehmann. 1923.
Gutzeit K, Teitge H. Die Gastroskopie; Lehrbuch und Atlas. Berlin: Urban and
Schwartzenberg. 1937.
Palmer ED. Gutzeit and Teitge revisited. Gastrointest Endosc 1977;23:244.
Hoffmann M. Optische Instrumente mit beweglicher Achse und ihre Verwendlung fr die
Gastroskopie. Mnch Med Wochenschr 1911;58:24468.
Schindler R. Ein vollig ungefahrliches flexibles Gastroskop. Mnch Med Wochenschr
1932;79:12689.
Moutier F. Traite de Gastroscopie et de Pathologie Endoscopique de l'Estomac. Paris:
Masson. 1935.
Schindler R. Gastroscopy: The Endoscopic Study of Gastric Pathology. Chicago: University of
Chicago Press, 1937; revised, 1950; 2d ed. New York: Hafner. 1966.
Gerstner P. The A/S/G/E; A history. Gastrointest Endosc 1991;37(suppl):126.
Dagradi AE, Stempien SJ. In memoriam: Rudolf Schindler. Gastrointest Endosc 1968;15:121.
Palmer ED. Clinical Gastroenterology, 2d ed. New York: Hoeber Medical Divison, Harper &
Row. 1963;2334.
Nelson RS. Routine gastroscopic photography. Gastroenterology 1956;30:6618.
Benedict EB. An operating gastroscope. Gastroenterology 1948;11:2813.
Schindler R. Results of the questionnaire on fatalities in gastroscopy. Am J Dig Dis
1940;7:2935.
Wood IJ, Doig RK, Motteram R, et al. Gastric biopsy. Lancet 1949;1:1821.
Lange FM, Meltzing D. Die Photographis des Magenintern. Mnch Med Wochenschr
1898;45:15858.
Uji T. The gastrocamera. Tokyo Med J 1952;61:1358.
Morrissey JF. Gastrointestinal endoscopy: 20 years of progress. Gastrointest Endosc
1983;29:536.
Morrissey JF. The use of the gastrocamera for the diagnosis of gastric ulcer.
Gastroenterology 1965;48:7117.
Perna G, Honda T, Morrissey JF. Gastrocamera photography. Arch Intern Med
1965;116:43441.
Matsunaga F, Isushima H, Kuboto T. Photography of the colon. Gastroenterol Endosc (Japan)
1959;1:5863.
Hopkins HH, Kapany NS. A flexible fiberscope, using static scanning. Nature 1954;173:3941.
Hirschowitz BI. A personal history of the fiberscope. Gastroenterology 1979;76:8649.
Hopkins HH. The 1991 Lister Oration, delivered at the Royal College of Surgeons, London.
Audiotape on file in the A/S/G/E Archives, Cleveland Health Sciences Museum. 1991.
Lamm H. Biegsame optische Gerate. Z Instrumentenk 1930;50:579.
Curtiss LE, Hirschowitz BI, Peters CW. A long fiberscope for internal medical examinations. J
Am Optical Soc 1956;46:1030.
Hirschowitz BI, Curtiss LE, Peters CW, et al. Demonstration of a new gastroscope, the
"fiberscope". Gastroenterology 1958;35:503.
Segal HL, Watson JS Jr. Color photography through the flexible gastroscope.
Soulas A. Bronchoscopy and television. Dis Chest 1957;31:580.
46. Rider JA, Puletti EJ, Rider RD, et al. Color television gastroscopy: A critical analysis.
Gastrointest Endosc 1971;18:668.
47. Sivak MV, Fleischer DE. Colonoscopy with a videoendoscope: Preliminary experience.
48. Classen M, Phillip J. Electronic endoscopy of the gastrointestinal tract: Initial experience with
a new type of endoscope that has no fiberoptic bundle for imaging. Endoscopy 1984;16:169.
49.
Demling L, Hagel HJ. Video endoscopy: Fundamentals and problems. Endoscopy
1985;5:1679.
50.
Overholt BF. The history of colonoscopy. In Hunt RH, Waye JD (eds). Colonoscopy: Clinical
Practice and Colour Atlas. London: Chapman Hall. 1981.
51.
Lemire S, Cocco AE. Visualization of the left colon with the fiberoptic gastroduodenoscope.
Gastrintest Endosc 1966;13:2930.
52.
Pollard HM. Presentation of 1975 Schindler Award to B. F. Overholt. Gastrointest Endosc
1975;22:623.
53. Overholt BF. Clinical experience with the fibersigmoidoscope. Gastrointest Endosc
1968;15:27.
54.
Oshiba S, Watanabe A. Endoscopy of the colon. Gastroenterol Endosc (Japan)
1965;7:4402.
55.
Niwa H. Endoscopy of the colon. Gastroenterol Endosc (Japan) 1965;7:4028.
56. Williams CB, Muto T. Examination of the whole colon with the fiberoptic colonoscope. Br Med
J 1972;3:27881.
57. Wolff WI, Shinya H. Colonofiberoscopy. JAMA 1971;217:150912.
58. Waye JD. Colonoscopy. Surg Clin North Am 1972;52:101324.
59.
Doubilet H, Poppel MH, Mulholland JH. Pancreatography: Technics, principles, and
observations. Radiology 1955;64:32539.
60. Ravinov KR, Simon M. Peroral cannulation of the ampulla of Vater for direct cholangiography
and pancreatography: Preliminary report of a new method. Radiology 1965;85:6937.
61. McCune WS, Shorb PE, Moscovitz H. Endoscopic cannulation of the ampulla of Vater: A
preliminary report. Ann Surg 1968;167:7526.
62.
Oi I, Kobayashi S, Kondo T. Endoscopic pancreatocholangiography. Endoscopy 1970;2:103.
63. Oi I. Fiberduodenoscopy and endoscopic pancreatocholangiography. Gastrointest Endosc
1970;17:5962.
64. Kasugai T, Kuno N, Kobayashi S, et al. Endoscopic pancreatocholangiography. I. The normal
endoscopic pancreatocholangiogram. Gastroenterology 1972;63:21726.
65.
Vennes JA, Silvis SE. Endoscopic visualization of the bile and pancreatic ducts. Gastrointest
Endosc 1972;18:14752.
66.
Silvis SE. Presentation of 1978 Schindler Award to Jack Vennes. Gastrointest Endosc
1978;24:263.
67.
Bakes J. Die Choledochopapilloskopie nebst Bemerkungen ber Hepaticusdrainage und
Dilatation de Papill. Arch Klin Chir 1923;126:47383.
68.
McIver MA. An instrument for visualizing the interior of the common duct at operation. Surgery
1941;9:1124.
69.
Wildegans H. Die operative Gallengangendoskopie. Munich: Urban & Schwartzenberg. 1960.
70.
Berci G. Choledochoscopy. The exploration of the extrahepatic biliary system under visual
control; Preliminary report. Med J Aust 1961;2:8623.
71.
Shore JM, Lippman HN. Operative endoscopy of the biliary tract. Ann Surg 1962;156:9515.
72.
Shore JM, Berci G. The clinical importance of cholangiography. Endoscopy 1970;2:11720.
73. Sackier J, Berci G, Paz-Partlow M. Transcystic laparoscopic choledocholithotomy. Am Surg
1991;57:3236.
74. Nakajima M, Akasaka Y, Yamaguchi K, et al. Direct endoscopic visualization of the bile and
pancreatic duct systems by peroral cholangiopancreatoscopy (PCPS). Gastrointest Endosc
1978;24:1415.
75.
Kelling G. ber Oesophagoskopie, Gastroskopie, und Klioskopie. Mnch Med Wochenschr
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
1902;49:21.
Ott DO. Die Beleuchtung der Bauchhhle Ventroskopie als Methode bei vaginaler Kliotomie.
Zentralbl Gynakol 1902;31:81720.
Jacobaeus HC. ber Laparo- und Thorakoskopie bei Untersuchung seroser Hohlungen
anzuwenden. Mnch Med Wochenschr 1910;58:20902.
Jacobaeus HC. ber Laparo- und Thorakoscopie. Beitr Klin Erforsch Tuberk Lungenkr
1912;25:183.
Kelling G. Klioskopie und gastroskopie. Arch Klin Chir 1923;136:2268.
Bernheim BM. Organoscopy: Cystoscopy of the abdominal cavity. Ann Surg 1911;53:7647.
Orndorff BH. The peritoneoscope in diagnosis of diseases of the abdomen. J Radiol
1920;1:3057.
Korbsch R. Lehrbuch und Atlas der Laparo- und Thorakoskopie. Munich: IF Lehmann. 1927.
Kalk H, Wildhirt E. Lehrbuch und Atlas der Laparoskopie und Leberpunktion. Stuttgart: Georg
Thieme. 1962.
Ruddock JC. Peritoneoscopy. Surg Gynecol Obstet 1937;65:62939.
Steptoe PC. Laparoscopy in Gynecology. Edinburgh: E & S Livingstone. 1967.
Dubois F, Berthelots G, Levard H. Cholecystectomy par coelioscopie. Presse Med
1989;18:9808.
Perissat J, Belliard R, Collet DC, et al. Cholecystectomy par laparoscopie. J Chir (Paris)
1990;123:34750.
Soergel KH, Hogan WJ. Therapeutic endoscopy. Hosp Pract 1983;18:8192.
Vizcarrondo FJ, Brady PG, Nord HJ. Foreign bodies of the upper gastrointestinal tract.
Gauderer MWL, Ponsky JL, Izant RJ. Gastrostomy without laparotomy: A percutaneous
endoscopic technique. J Pediatr Surg 1980;15:8728.
Crafoord C, Frenckner P. Nonsurgical treatment of varicose veins in the esophagus. Acta
Otolaryngol 1939;27:4229.
Moersch HJ. Further studies on the treatment of esophageal varices by injection of a
sclerosing solution. Ann Otol Rhinol Laryngol 1941;50:123344.
Patterson CO, Rouse MO. Injection treatment of esophageal varices. JAMA 1946;130:3846.
Tio TL, Tytgat GN. Endoscopic ultrasonography of normal and pathologic upper
gastrointestinal wall structure: Comparison of studies in vivo and in vitro with histology. Scand
J Gastroenterol 1986;123(suppl):2733.
Cotton PB, Shorvon PJ, Lees WR. Endoscopic ultrasonography: A new look from within. BMJ
(Clin Res Ed) 1985;290:13734.
Sivak MV Jr. Is there an ultrasonographic endoscope in your future?. Gastrointest Endosc
1988;34:645..
Chapter 2 Flexible Endoscope Technology: The Fiberoptic

Endoscope
(109)
ICHIZO KAWAHARA, M.D.(HON.)
HIROSHI ICHIKAWA, B.S.(MECH. ENG.)
The flexible endoscope is a highly refined and complex instrument that may utilize one of two different
imaging systems: fiberoptic technology and semiconductor technology. A fiberoptic endoscope
(fiberscope) employs a glass fiber bundle that transmits images from the objective lens at the distal
end of the instrument to the ocular lens in the eye piece. A video endoscope (videoscope) uses a
charge-coupled device (CCD) to electronically transmit images from the endoscope to a television
monitor.
Mechanical Construction of the Flexible Endoscope

The major parts and controls of a typical fiberscope and videoscope are shown (Figure 21). The
mechanical components of these two types of endoscopes are basically the same. In this chapter,
aspects of the internal construction that are common to both types of instruments are discussed in
detail.
(110)Figure 21. Nomenclature of the various controls and components of a fiberscope (left)
and the control section of a videoscope (right).
Distal Tip
An end view and a cross section of the distal tip of a typical fiberscope and videoscope are shown in
Figure 22. On the faces of the tip, the following can be seen: (1) the channel opening for suction and
passage of accessories; (2) the complex illumination lens system, which distributes light from the light
guide (LG) fiber bundle into wide-angle, even illumination of the visual field; (3) the objective lens
system, which focuses an image of the mucosa onto the face of the image guide (IG) bundle; and (4)
an air/water nozzle, which provides air for insufflation of the organ being observed and water that
washes across the lens to remove substances such as secretions, mucus, and blood.
(111)Figure 22. End view and cross section of the distal tip of a fiberscope and a
videoscope. I.G.image guide; L.G.light guide; CCDcharge-coupled device.
Two cross sections of the distal tip of a side-viewing flexible endoscope (fiberscope) are illustrated in
Figure 23. A roof prism is used to produce a 90-degree change in direction of view. A forceps raiser is
also necessary to deflect the tip of various accessories passed through the channel to bring them
within the field of view.
(112)Figure 23. Cross section of the distal tip of a side-viewing flexible endoscope in two
different planes. L.G.light guide; I.G.image guide.
Bending Section and Angulation System

The design of the bending section, the portion that produces controlled deflection of the distal tip, was
originally developed in Japan for the gastrocamera. Four-way tip deflection was first introduced with the
Olympus Model CF-MB/LB colonoscope in 1970. Since then, a variety of different mechanisms have
been developed to control tip deflection.
A typical angulation system in current use is illustrated in Figure 24. The endoscope used for
gastrointestinal purposes generally has a built-in resistance or adjustable braking system to fix the
deflection of the tip, allowing the operator to remove the hand from the control knob and still maintain
tip deflection. To produce tip deflection, the operator rotates the control knob that is connected to a
sprocket within the control section. This sprocket moves a chain that, in turn, pulls on various wires
running the length of the insertion tube. As a wire is pulled, it produces tip deflection in that direction.
The bending section is constructed from interlocking metal rings. The pivot points between these rings
alternate by 90 degrees, thereby giving the bending section the ability to bend in any direction. A
cross-sectional view of the bending section is shown in Figure 25.
(113)Figure 24. Bending section and angulation system of a flexible endoscope.
(114)Figure 25. Cross-sectional view of the bending section of a flexible endoscope.

L.G.light guide; I.G.image guide.
During an endoscopic procedure, the tip of the instrument is deflected many times. A tracking of total
movement during a procedure is shown in Figure 26. To withstand such repeated, complex
movements, the deflecting mechanism must be extremely durable and tested extensively to guarantee
that it is safe and reliable.
(115)Figure 26. Data showing the degree and direction of tip deflection during a typical
endoscopic procedure.
Insertion Tube
Although the insertion tube, excluding the bending section, is not capable of controlled deflection, its
carefully calculated flexibility and torque-free construction are of major importance in endoscope
design.
The basic construction of the insertion tube is shown in Figure 27. Helical steel bands form the
supporting structure of the tube and give it its round shape. These bands are covered with a layer of
stainless steel wire mesh. Together, these components prevent the insertion tube from twisting or
stretching along its axis and also help shield the glass fiber bundles from damaging x-radiation.
(116)Figure 27. Internal components and construction of the insertion tube of the flexible
endoscope. L.G.light guide; I.G.image guide.
The outer plastic covering over the metal structure of the instrument also helps prevent twisting of the
insertion tube and compression along its axis. It is important that this final covering be waterproof and
able to withstand a variety of chemical agents from gastric acid to corrosive disinfectants. The
specifications for each component of the insertion tube are carefully chosen to ensure that the final
tube has the proper flexibility and elasticity to recover from repeated bending.
Control Section
The control section of the endoscope is designed to be held entirely by the left hand, leaving the right
hand free to hold and manipulate the insertion tube. The second, third, and fourth fingers grip the
instrument against the palm, leaving the left thumb free to control the up/down angulation knob and the
left index finger free to operate the air/water and suction valves. The right hand is used to torque and
advance the insertion tube, insert and operate accessories, control the right/left angulation knob,
operate the camera, adjust the diopter of the ocular lens, and focus the objective lens as necessary.
The technique of endoscopy is described in Chapter 38: Technique of Upper Gastrointestinal
Endoscopy; Chapter 57: Techniques of Endoscopic Retrograde Cholangiopancreatography; and
Chapter 81: Technique of Colonoscopy.
Air, Water, and Suction System

A cross section of the entire instrument (Figure 28) illustrates the various internal channels for air,
water, and suction. This type of design for automatic air, water, and suction originated in 1968 with the
Olympus Model EF esophagoscope and has remained basically the same through succeeding
generations of instruments.
(117)Figure 28. Air, water, suction, and CO2 systems of a typical colonoscope.
Air supplied by a pump within the light source is emitted from the nozzle on the distal tip when the
opening in the air/water valve is covered. Air is used to insufflate the organ under observation and to
blow water off the objective lens. When the air/water valve is depressed, water is forced from the
pressurized water container through the endoscope and out the nozzle on the distal tip.
Aspiration of either air or fluid through the endoscope is accomplished by depressing the suction valve.
An external suction pump and collection bottle are connected to the endoscope and provide the
required negative pressure. When performing electrosurgery, the endoscopist may wish to introduce
CO2, as an inert gas, to reduce the risk of accidental explosion in the bowel, particularly the colon. This
is done by depressing the CO2 gas valve, which allows gas from an externally regulated tank to be
insufflated via the endoscope.
Valves are provided on all of these channels to check the backward flow of material from the various
openings on the distal tip into the air, water, and gas chan-nels within the endoscope. However, since
the endoscope is used in a pressurized environment owing to organ insufflation and because the air
within the channels of the endoscope is compressible, it is possible for small amounts of debris to work
their way slowly into the interior air, water, and gas tubing of the endoscope. To prevent occlusion of
these small tubes and for reasons of infection control, it is important to flush and disinfect these lines
when the instrument is cleaned.
Light Guide Connector

A cross section of the LG connector of the endoscope is shown in Figure 29. This portion of the
instrument connects to the light source and provides light and pressurized air. It also has connections
for a water container, a suction pump, and an electrical cord to safely return any electrosurgical
leakage current to the generator. A vent on the connector allows the interior of an airtight and fluid-tight
endoscope to be vented before the instrument is placed in an evacuated chamber for gas sterilization.
(118)Figure 29. Cross section of the light guide (L.G.) connector of an endoscope.
Principles of Fiberoptics
As described previously, one of two imaging technologies can be used in the flexible endoscope:
fiberoptic and CCD (video) technology. CCD technology is described in Chapter 3: Flexible Endoscope
Technology: The Video Image Endoscope; fiberoptic technology is discussed in this chapter.
Principle of Total Internal Reflection

The heart of the flexible fiberoptic endoscope (fiberscope) is the image-carrying fiber bundle, which
transmits the image through the instrument. This bundle contains tens of thousands of individual
ultra-thin glass fibers. The ability of an individual fiber to efficiently transmit light entering the distal end
of the fiberwithout substantial loss of brightness, change in color, or leakage of light to adjacent
fibersis basic to the development of fiberoptic instruments.
The phenomenon of refraction and reflection as shown in Figure 210 explains why light can be
transmitted through flexible glass fibers. This figure shows two transparent substances. The lower
substance can be assumed to be glass with a refractive index of n. The upper substance may be either
air or glass of a different composition, possessing a lower refractive index of n. Line L represents the
boundary separating the two media of index n and n. A ray of light traveling through the lower medium
and hitting the boundary surface with incident angle A (Figure 210A) will be refracted and travel
through the upper transparent medium at angle B. The relationship between angle A and angle B is
given by the the equation:
n sin A = n' sin B
(119)Figure 210. A-C, Phenomenon of refraction and reflection of light by a plane surface.
See text for explanation.
As the angle of incidence A is increased, the angle of the refracted ray B also increases according to
the relationship given in the previous equation. When A equals AC (known as the critical angle of
incidence), the refracted ray will travel along the boundary surface (Figure 210B). The critical angle
for any two substances is found by setting B = 90 (sin B = 1). In this case,
If the angle of incidence is increased further to A0, A0 being greater than AC, the ray is totally reflected
at the boundary surface back into the lower medium (Figure 210C). The angle of reflection always
equals the angle of incidence. It is this condition of total internal reflection that enables glass fibers to
transmit light. Total internal reflection can occur only when the ray is incident on a medium whose
index is less than that of the medium in which the ray is traveling.
Optical Fibers
A long, cylindrical glass fiber will transmit light if its surface is clean and it is surrounded by a medium
with a lower refractive index (Figure 211). However, any debris on the surface of the fiber or any
contact of the fiber with adjacent fibers or other objects will disturb the boundary condition and prevent
total internal reflection from occurring at that point on the surface. These conditions will result in
leakage of light from the fiber, a loss in transmission, and the transfer of light from the fiber to other
objects with which it comes in contact. To prevent this and to ensure a proper boundary surface, all
glass fibers used for fiberoptics are clad with a very thin layer of glass (Figure 211B). The cladding
glass has a lower index than the core glass, a condition that guarantees total internal reflection of all
rays traveling through the core glass.
(120)Figure 211. Path of light ray through an unclad (A) and a clad (B) glass fiber.
A ray incident on the face of the fiber at angle 0 (Figure 212) will be refracted and travel through the
core at angle A0. At point P1 on the boundary surface, it undergoes total internal reflection. It
undergoes a second reflection at point P2, and continues onward down the fiber, reflecting each time it
hits the boundary surface. As the angle of incidence 0 on the face of the fiber is increased, the angle
of internal reflection decreases until A0 reaches AC, which is called the critical angle for internal
reflection (Figure 212). The angle representing the maximum incidence angle for total internal
reflection is referred to as the maximum acceptance angle C. AC is a function of n and n, the indices
of the glass used in the core and the cladding, respectively:
(121)Figure 212. Total internal reflection and maximum acceptance angle of incidence, C.
See text for explanation.
A ray incident on the boundary at an angle less than AC will not undergo reflection but will be refracted
through the cladding, out through the side of the fiber, and will be lost.
A calculation of the value of C will show that it is dependent on the indices of the core, the cladding,
and the medium at the face of the fiber:
The value of N.A. is designated as the numerical aperture of the fiber and equals sin C when the face
of the fiber is in contact with air (n0 = 1 for air). The greater the difference between n and n, the
greater the numerical aperture and the larger the cone of light that the fiber will accept and transmit.
Because of production limitations on the glasses used for the core and cladding, the numerical
aperture of optical fibers is generally limited to 0.52. By calculation with the preceding formula, C (the
maximum acceptance angle) is 31 degrees for a fiber with this numerical aperture. Light entering at
angles greater than this will not undergo internal reflection, but will pass out through the side of the
fiber.
In practice, not all light entering the face of the fiber at incident angles less than the maximum
acceptance angle of incidence (C) will be transmitted and exit the other end of the fiber. Light
transmission is also decreased by the following factors:
1. There is some absorption of light by the core glass. This loss is proportional to both the length of
the fiber and the length of the path light takes within the fiber.
2. Although the preceding discussion implies that at angles greater than AC the light will be totally
reflected, in practice this reflection is not 100%. The small amount of refraction or scattering that
takes place is insignificant for one reflection. However, because the light may be reflected tens of
thousands of times in traveling 1 m, a small loss at each reflection point results in a measurable
loss at the end of the fiber. Fibers of small diameter and long length suffer the greatest loss of this
type.
3. Some of the light falling on the surface of the fiber will be reflected by the surface rather than
entering the fiber. The amount of loss is approximately 4% at each end.
Fiberoptic Bundles
An individual fiber cannot transmit an image. If one observes the end of an illuminated fiber, only a spot
of light of a certain color and intensity is seen. To create an image, a large number of fibers must be
grouped together. The pattern formed by the color and intensity of the individual fibers is perceived by
the observer as an image. For the image at one end of the bundle to duplicate the image at the other
end, it is necessary that the ends of each individual fiber occupy the same relative position in both ends
of the bundle. A bundle that is organized in this manner is called a coherent bundle (Figure 213). Only
coherent bundles are capable of producing an image. Therefore, they are also referred to as image
guide fiber bundles (IG bundles).
(122)Figure 213. Alignment of fibers on the face of a coherent fiber bundle.

An important property of an IG bundle is its resolving power, that is, the amount of image detail the
bundle can convey. A bundle's resolving power depends on the diameter of the fiber core, the
thickness of the cladding, and the alignment and orderliness of the packing of the fibers within the
bundle faces. The smaller the fiber and the thinner the cladding, the greater the image resolution. In
practice, the thickness of the fiber cladding cannot be made less than 1.2 m for visible light because
of production limitations as well as physical optical theory.
The ratio of the total area occupied by the individual fiber cores to the total area of the fiber bundle is
referred to as the packing fraction. Since only the cores transmit light, it is advantageous to make the
packing fraction as large as possible. However, since there is a limit to the thickness of the cladding,
as the diameter of the core is reduced to improve resolution, a certain point is reached where the
packing fraction becomes unacceptably low. The high proportion of cladding area to core area results
in a very dark image. Owing to this limitation, the smallest practical fiber diameter (including cladding)
is generally approximately 6 m. A typical packing fraction is 60%.
The alignment of the fibers in the face of the bundle greatly affects the quality and resolution of the
image. Imperfect alignment will result in distorted images and annoying flaws or dark areas in the
image.
The length and number of fibers within an IG vary greatly depending on the type and size of the
endoscope. The number of individual fibers in the IG bundle usually ranges between 5000 and 40,000.
The diameter of the bundle may vary between 0.5 mm and 3 mm.
Basic Optical System of a Fiberscope

A schematic of the type of optical system used in a fiberscope is shown in Figure 214. The objective
lens at the tip of the fiberscope forms an image of the object in view X0 on the distal face of the IG X0.
This miniature image is limited by the size of the fiber bundle. The light representing this image is
transmitted through the IG and a duplicate image is formed on the proximal face of the bundle near the
eye piece. Since the objective lens produces an inverted image on the distal face of the bundle, the
bundle must be twisted 180 degrees to produce an upright image at the proximal face. The ocular lens
(fiberscope eye piece) functions as a simple magnifying glass and creates an enlarged virtual image
representing the image resting on the tiny tip of the bundle.
(123)Figure 214. A basic fiberscope optical system. I.G.image guide.

Magnification of the viewed object is determined by several factors. Magnification produced by the
objective lens is variable and is determined by the distance between the objective lens (fiberscope tip)
and the object being viewed:
where M0 is usually much less than 1.

Magnification of the tiny image on the proximal face of the IG bundle is determined by the focal length
of the ocular lens:
where M1 typically ranges between 15 and 30.

The visual magnification of the fiberscope's entire optical system M is the product of M0 and M1:
M = M0 X M1
The field of view () is given by:
where f is the focal length of the objective lens.

Progress in modern lens design technology has made it possible to obtain a maximum up to 140
degrees.
All fiberscopes have a diopter adjustment on the eye piece. This allows focusing of the ocular lens to
compensate for the diopter of the individual endoscopist's eyesight.
Illumination System
An optical system used for endoscopic illumination is shown in Figure 215. An incoherent light guide
fiber bundle (LG bundle) is used to transmit light through the endoscope. Special lens systems on both
ends of the bundle are needed to effectively capture the maximum amount of light from the light source
lamp and to produce a wide angle for even illumination from the tip. Fibers used in the LG bundle are
designed for high a numerical aperture (N.A.) and a high transmission ratio. To produce the highest
possible packing fraction, LG fibers are made as large as possible without compromising their flexibility
and durability. At present, 30-m fibers are commonly used.
(124)Figure 215. Flexible endoscope illumination system. L.G.light guide.

Owing to the intense heat produced by the high-intensity light source used in endoscopy, dichroic
coatings and heat-absorbing filters are employed to cut out nonvisible radiation from the lamp output.
In addition, heat sinks and forced-air cooling systems in the light source prevent the LG bundle from
overheating.
Production and Quality of the Image Guide

The following is a general description of an automated process that produces high-quality round
bundles, which is the method used to make the IG bundles found in the majority of fiberscopes today.
Production
The individual fibers start out as single glass rods consisting of a core of high-quality optical glass, a
cladding of glass with a refractive index lower than that of the optical glass, and a second cladding of
acid-leachable glass (Figure 216). This original glass rod is many times larger in diameter than the
final flexible glass fiber.
(125)Figure 216. Composition of a glass rod that, after heating and elongation, will become
an individual glass fiber.
The necessary number of these single glass rods (usually tens of thousands) are perfectly aligned
within a larger cylinder made of acid-leachable glass. This large coherent bundle is called the master
IG (Figure 217).
(126)Figure 217. Master IG bundle.

The master IG is then placed in an electric furnace and pulled (Figure 218). Owing to the heat and
tension, the bundle elongates and becomes thinner. This process is repeated many times until the
individual fibers are reduced to the desired size. A cross section of the bundle at this stage is shown in
Figure 219. Because of the tension of the pulling process, the acid-leachable glass is forced into a
hexagonal honeycomb pattern, fusing the individual glass rods into a single rigid bundle. The perfect
packing and coherence of the original master IG are maintained.
(127)Figure 218. Process of pulling a master IG bundle to reduce its diameter.
(128)Figure 219. Cross section of the fiber bundle after the pulling process, showing the
fusion of the acid-leachable glass.
Protective holders are placed over both ends of the bundle, which is then soaked in an acid solution
(Figure 220). Only the acid-leachable glass that fuses the fibers is affected by the acid. As this glass
is slowly dissolved, the fibers become free and flexible (Figure 221). At this point, the individual fibers
are freed from one another to convert the fragile, rigid bundle into a durable, flexible bundle.
(129)Figure 220. Rigid fiber bundle is soaked in an acid solution that dissolves the
acid-leachable glass and frees the individual fibers.
(130)Figure 221. Cross section of the fiber bundle after the acid leaching process.
Quality
Several factors can affect the quality of the final IG bundle. Some of these are shown in Figure 222
and include black dots caused by broken fibers; half-opaque (gray) dots caused by poor cladding or an
air bubble in the fiber; and disorders in fiber alignment or dust in the fiber inherent in the manufacturing
process.
(131)Figure 222. Example of various forms of fiber bundle imperfections.
Ancillary Equipment
Endoscopic Light Sources
A variety of endoscopic light sources are available, ranging from simple low-power halogen light
sources to sophisticated high-intensity xenon units. The larger, more advanced light sources generally
have multiple features such as automatic flash photography, automatic brightness control for television
and video endoscopy, and the capability to accept rigid telescopes. A block diagram of the function of
the various components of a high-intensity xenon light source is shown in Figure 223).
(132)Figure 223. Block diagram showing the function of various components of a xenon
light source.
Electrosurgical Generators
Only solid-state generators having isolated outputs should be used for endoscopic procedures.
Although almost any generator that operates within the proper power range can be adapted for
endoscopic use, electrosurgical units designed specifically for use with endoscopes have several
important additional safety features. One of these is a separate terminal to connect a safety return cord
(S-cord) from the endoscope (Figure 224). This S-cord returns any capacitance-induced "leakage"
current in the endoscope directly to the generator, thereby avoiding potentially dangerous circuiting
through the patient or physician. If proper electrical connections have not been made, a warning
system built into the generator will alert the operator and will prevent the unit from operating until these
connections are correct. A second safety feature is that the current returning through the S-cord is
continuously monitored by the generator and compared with the total output. If a significant amount of
current is detected in the S-cord, the generator automatically decreases power output to a substantial
degree in order to prevent possible injury. The principles of electrosurgery are discussed in Chapter 9:
Principles of Electrosurgery.
(133)Figure 224. Schematic diagram of endoscopic electrosurgery with a safety return cord
(S-cord).
For safety, it is important that endoscopic electrosurgery be attempted only through the endoscope
specifically designed for this application. The endoscopes that have exposed metal bands on the
insertion tube, noninsulated distal tips, and nongrounded construction should never be used.
Imaging Apparatus
Various types of imaging apparatus (Figure 225) can be used with a fiberscope by connecting these
devices to the eye piece of the endoscope. These include full-automatic still camera (16-mm or 35-mm
film), instant still camera (Polaroid), motion picture camera (8-mm or 16-mm film), miniaturized
television camera, and video converter (connects to the video processor and allows the fiberscope to
be used as a video endoscope).
(134)Figure 225. Endoscopic system chart. A complete endoscopic system includes not only
a variety of fiberscopes but also a complete line of accessories and ancillary equipment.
Total Endoscopic System

A large variety of accessories and ancillary equipment are required for modern endoscopy. This total
endoscopic system is depicted schematically in Figure 225.
Cleaning, Sterilization, and Maintenance of the Endoscope and

Accessories
Meticulous cleaning of the flexible endoscope immediately after use is important to keep the instrument
free of accumulated organic debris and in good mechanical working order. An improperly cleaned
instrument may also compromise subsequent disinfection or sterilization procedures, since the surface
layer of debris may shield underlying microorganisms from the biocidal action of the disinfectant or
sterilization process. All detachable parts such as distal hoods, forcep valves, air/water and suction
valves, water containers, and tubes should be removed from the instrument, cleaned, and sterilized
separately. The majority of modern endoscopes are completely watertight and can be totally immersed.
The entire exterior surface and internal channels can be easily washed with a sponge or soft brush.
The most common disinfectant used for endoscopes is glutaraldehydes. Many different brands are
available. It is suggested that the user consult the endoscope manufacturer concerning the suitability of
particular disinfecting agents before using them to disinfect equipment. Instrument manufacturers can
suggest products that they have tested and have found not to damage their instruments after long-term
use. It is important that the instrument be thoroughly rinsed after using disinfectant solution and that all
interior channels be flushed with clean water and air-dried. Many disinfectants are irritating to body
tissue and must be completely rinsed from the instrument before its use on patients. It is also important
that the instrument be properly dried before storage. Moisture in the interior channels of the endoscope
can promote the growth of microorganisms, and water drops allowed to dry will leave a film or stain on
electrical contacts and lens surfaces.
Ethylene oxide gas (ETO) sterilization is suitable for sterilizing endoscopes. The endoscope must be
completely dry before ETO sterilization. The venting valve on airtight and fluid-tight fiberoptic
instruments must be opened before placing them in a gas sterilizer. This allows the pressure inside the
instrument to equilibrate with the chamber pressure. If the instrument is not vented, as the vacuum is
being drawn in the sterilization chamber the atmospheric pressure inside the endoscope will cause
expansion and bursting of the rubber sheath covering the bending section of the instrument. Generally,
videoscopes do not require this type of venting.
After exposure to ETO, the endoscope must be aerated to remove all residual ETO that has been
absorbed by the plastic and rubber parts of the instrument. The endoscope manufacturer should be
consulted concerning the maximum temperature, pressure, or vacuum the instrument can withstand
during the ETO sterilization and aeration cycles.
Although the flexible endoscope cannot be placed in an autoclave, many accessories such as biopsy
forceps, polypectomy snares, cytology brushes, water containers, and mouthpieces can be sterilized by
this method, which makes use of steam under pressure.
An automatic washing and disinfecting machine may be desirable in a busy endoscopy suite, since it
saves time in disinfecting endoscopes.
Safety
Mechanical Safety
As with any medical instrument, the endoscope must be manufactured according to specific national
standards such as the Good Manufacturing Practice standards of the U.S. Food and Drug
Administration. However, for the ultimate in safety, an endoscope manufacturer must consider all
possible types of malfunction and design an instrument such that it will not cause harm to the patient
should a malfunction occur. An example of this type of design philosophy would be an angulation
system that cannot jam or malfunction with locking of the tip of the instrument in a deflected position,
thereby preventing the removal of the instrument from the patient. Another example would be an
insufflation system that cannot malfunction in any manner that would produce a potentially dangerous
overinsufflation. A final example is the use of a braided wire consisting of numerous strands of
extremely fine wire in the angulation system of the instrument. A braided wire with only a few, thicker
wires could be used. However, if one of the strands became frayed or cut, a thick wire would be strong
enough to puncture the outer sheath of the insertion tube and injure the patient, whereas a thinner wire
would not.
With many mechanical devices, it is easy to increase the durability by simply increasing the size and
weight of critical components, thereby strengthening them and preventing future failure. It is quite
possible to substantially increase the durability of the endoscope by also strengthening or redesigning
certain components. However, this would severely compromise the safety and effectiveness of the
instrument. For example, using larger, thicker wires in the angulation system would reduce the amount
of stretching that might occur. As explained earlier, this would compromise its mechanical safety.
Other changes to improve durability might be adding a thicker protective covering over delicate fiber
bundles to prevent fiber breakage, using a heavier supporting structure in the insertion tube to prevent
damage due to patient bites, reducing the amount of tip deflection to reduce stress on the internal
components, and increasing the thickness of the rubber sheath to eliminate water leakage owing to pin
holes. These and many more changes could be made to produce an instrument that would require little
or no repair during its lifetime. However, the results of these changes would be an endoscope with a
large, stiff insertion tube, limited tip deflection, reduced accessory channel size, and a larger, heavier
control section. Since these changes are unacceptable, modern endoscope design represents a
delicate balance between the various features the physician requires and the best design in terms of
ultimate durability.
Another important design factor is that the endoscope must "yield" before the patient does. The
flexibility of the instrument and the tactile feedback of the control knobs allow the physician to gauge
the force the instrument is exerting on the patient. This required "feel" of the instrument also puts
important limitations on the type of materials used in the endoscope and the manner in which the
instrument is constructed.
Electrical Safety
Underwriters Laboratories and International Electrotechnical Commission standards apply to the
ancillary electrical equipment used with an endoscope. The basic intent of these standards is to
prevent patient or operator shock due to leakage of line voltage, and to maintain electrical safety of the
equipment should some internal component fail or malfunction.
Endoscopes are now designed so that they are electrically isolated from ground. This is done to
prevent current from flowing through the endoscope and patient should a fault occur in attached
ancillary equipment. This also prevents the endoscope from acting as the return electrode should the
return electrode become dislodged or disconnected when performing electrosurgery with a
ground-referenced electrosurgical generator.
Owing to the nature of high-frequency current, it is impossible to totally eliminate capacitive coupling
between the endoscope and the electrosurgical accessory passed through it. Therefore, it is essential
that there be no unnecessary exposed metal parts on the insertion tube that could possibly result in an
electrosurgical burn to the patient. The use of a protective eye shield for the electrically nonisolated
metal eye pieces is also recommended. Despite all the safety features designed into modern
instruments, owing to the potential hazards of electrosurgery, the therapeutic endoscopist should be
thoroughly trained in electrosurgical procedures and aware of all possible complications (see Chapter
9: Principles of Electrosurgery).
Chapter 3 Flexible Endoscope Technology: The Video Image

Endoscope
(135)
DAVID E. BARLOW, B.S., M.S., PH.D.
The video image endoscope, referred to as "the most fundamental change in endoscope design since
the introduction of the fiberscope in 1957,"1(136) represents a new era for endoscopy. The clinical
advantages of the video image endoscope, or videoscope, are well documented.27(137) These
include the ability to perform procedures with a natural, relaxed posture while observing a video
monitor; simultaneous viewing by the physician, trainee, endoscopy assistant, and even the patient;
enhancement of training through observation of procedures and videotape; and when the endoscope is
integrated with a computer, the assembly of a database of high-quality, easily accessible endoscopic
images for documentation, patient follow-up, and research.
Many of these advantages are not unique to the video image endoscope. In fact, by 1983, the year
Welch Allyn introduced the first commercial videoscope, many endoscopists were routinely performing
"video endoscopy" with small video cameras attached to fiberscopes. These external cameras
provided many of the benefits of the videoscope, including the provision for multiple observers,
enhanced training, and documentation. Despite these advantages, however, external cameras were
both bulky and heavy, and they had two insurmountable problems: (1) image resolution on the monitor
was limited by the resolution of the fiber bundle and (2) the image obtained through the fiber bundle
was further deteriorated by moir*(138) interference.
Moir interference is created when two regularly repeating patterns are superimposed (Figure 31).
When the honeycomb fiber bundle pattern of a fiberscope (see Chapter 2: Flexible Endoscope
Technology: The Fiberoptic Endoscope) is projected onto the image sensor of an external camera, a
disturbing moir pattern is likely to result. To prevent this, endoscopic camera designers typically
diffuse or slightly defocus the endoscopic image. The video image endoscope avoids this problem by
eliminating the fiber bundle and positioning the image sensor at the distal end of the instrument.
(139)Figure 31. A moir pattern is produced by the interference of two regularly repeating
patterns, such as the scan lines of a video camera and the fiber pattern of a fiberoptic bundle.
Charge-Coupled Device
The video image endoscope became a reality because of improvements in charge-coupled device
technology. The charge-coupled device (CCD or "chip") was invented in 1969 at Bell Laboratories by
W. S. Boyle and G. E. Smith. The CCD was first used in commercial cameras in 1981. Since then, an
array of remarkable machines has been built on the technology of the solid-state image sensor,
including facsimile machines, optical character readers, department store checkout wands,
surveillance cameras, and home video systems. The CCD also provides the "eyes" for robotic vision
systems and can be integrated with computers to automatically count parts, perform critical
measurements, and spot defects in manufacturing processes.
Solid-state image sensors have several important advantages over older technologies such as vacuum
tube imaging systems. They require much less power, are less susceptible to mechanical vibration and
shock, and are not damaged by bright illumination. They also do not "lag", that is, produce comet-like
tails trailing behind highlights when an image is moving.
Solid-state CCD image sensors have additional characteristics that make them particularly well suited
for certain unique applications. CCDs are commonly used in astronomy because their sensitivity to
light is more than 100 times greater than photographic film. They also have a dynamic range that is
more than 50 times greater than photographic film or tube cameras. They can be made to "stare" into
space for long periods of time to image faint celestial bodies that would otherwise be obscured by
bright nearby objects if they were recorded using conventional photographic techniques. At the
opposite extreme, the CCD can be electronically shuttered and operated at very short exposures to
capture images of fast-moving objects in as little as 1/10,000 second. For consumer applications, the
CCD's low power consumption, high resolution, and good color reproduction make it the sensor of
choice for home video cameras. For many of these same reasons, CCDs are found in most broadcast
video cameras as well.
Despite all of these inherent advantages as an imaging device, the CCD could not be used for
endoscopic applications until solid-state technology evolved to the point that extremely small CCDs
could be manufactured to fit in the limited space at the distal tip of the endoscope.
Solid-State Imaging Technology

CCDs are solid-state image sensors made of silicon semiconductor material. The silicon on the
surface of the sensor is responsive to light and exhibits the photoelectric effect. When a photon of light
strikes this photosensitive surface, it displaces an electron from a silicon atom. This produces a free,
negatively charged electron and a positively charged "hole" created by the absence of the electron in
the regular crystalline structure of the silicon.
Although a single photosensitive element can be used to measure the brightness of the light falling on
the device (e.g., a light meter), it cannot reproduce an image. To do so, the photosensitive surface
must be divided into a matrix of small, independent photosites. When an image is focused on the
surface of such a sensor, the brightness of each discrete point in the image can be measured for later
reproduction.
The surface of a CCD is divided into an array of discrete photosites (Figure 32), individually referred
to as picture elements, or pixels. For simplicity, Figure 32 illustrates a pixel array of 10 rows by 10
columns (a total of 100 pixels). An actual endoscopic CCD may contain up to several hundred
thousand pixels.
(140)Figure 32. The photosensitive surface of a charge-coupled device (CCD) image sensor
is divided into a matrix of individual photosites called pixels.
In the video image endoscope, the CCD is located within the distal tip directly behind the objective lens
(Figure 33). The objective lens focuses a miniature picture of the observed area on the surface of the
CCD. (In some endoscopes, a prism reflects the image onto a CCD mounted at 90 degrees to the
optical axis of the instrument.) The pattern of light falling on the CCD is instantly converted into an
array of electrical charges owing to the photoelectric effect. Because charges created in each pixel are
isolated from neighboring charges, the CCD transforms the optical image into a faithful electrical
representation (Figure 34). The greatest number of charges is generated in those areas that are
brightest (Figure 34C).
(141)Figure 33. The components within a video endoscope's distal tip include a fiberoptic
light guide bundle for bringing light into the gastrointestinal tract, an objective lens for
focusing an image of the mucosa on the face of the image sensor, and a CCD image sensor
that captures the image for display on a video monitor.
(142)Figure 34. A, The photosensitive surface of a CCD is divided into electrically isolated
pixels. B, An endoscopic image is projected onto the CCD surface. C, A representation of the
image is created in the form of electrical charges. A detailed look at individual pixels indicates
that the charge developed is proportional to the brightness of the light falling on each pixel.
Each pixel can develop any amount of charge in a continuum from some minimum to some maximum
level, depending on the brightness of the incident light. The basic physics of the process is simple
and linear. Doubling the number of photons of light falling on a pixel doubles the number of charges
generated until the potential well (pixel) storing the electrons is finally full. Because of this linear
relationship, the charge developed at each photosite is the product of the brightness of the incident
light and the exposure time.

After the exposure period, the charges developed in the CCD must be "read out" in an orderly manner
and processed to reproduce the image. The several methods used to transfer charges within and
finally out of a CCD include line, frame, and interline transmission systems.
Line Transfer CCD

The line transmission system is schematically illustrated in Figure 35. The charges within each
pixel are controlled and shifted via electrodes adjacent to each pixel (not shown in the figure). By
varying the voltages applied to these electrodes, the electrons within individual pixels are transferred
as "charge packets" from one pixel to another. Sequential voltage changes on these electrodes march
the charge packets toward one end of the CCD and into a horizontal shift register.
(143)Figure 35. Schematic representation showing how charges are transferred and read out
of a CCD employing the line transmission system.
Figure 35 illustrates how the charges in the bottom row of the CCD are first transferred into the
horizontal shift register, leaving the bottom row empty. This starts the orderly transfer of charges from
each row to the row below it, starting at the bottom row and feeding from the top. After completing one
series of transfers (see Figure 35, Step 1 to Step 2), the electrical representation of the image has
shifted down by one row of pixels, with the charge packets from the bottom row of pixels now residing
in the horizontal shift register.
The quantity of charge in each charge packet within the horizontal shift register is then measured as
the packets progress one step at a time toward the amplifier at the left of the horizontal shift register.
The amplifier generates an output signal corresponding to the quantity of charges contained in each
charge packet. This output signal is later used to re-create the variation in brightness of the bottom row
of pixels.
Once the horizontal shift register is read out and emptied (see Figure 35, Step 3) the process repeats.
The charges in the bottom row of pixels are transferred into the horizontal shift register and the entire
image replica is shifted down one more row (see Figure 35, Step 4). The shift register is read out
again (see Figure 35, Step 5), and the process continues until the entire image has been read out and
the CCD surface is cleared for another exposure.
The "charge-coupling" processthe transfer of charges as packets from potential well to potential
wellgives the CCD its name. The charges in the furthermost corners of the CCD are actually moved
sequentially through several hundred potential wells before they finally reach the output of the
horizontal shift register. In current video image endoscopes, the CCD is exposed, read out, and
reexposed 60 to 90 times each second. To maintain image integrity during these repetitive transfers, it
is essential that these charge packets remain intact with no loss or gain in charge quantity as they
undergo hundreds of thousands of transfers per second as the CCD is read out.
Frame Transfer CCD

Some videoscopes utilize a frame transfer CCD (Figure 36). A frame transmission system uses two
separate pixel arrays. The first is a sensor array on which the endoscopic image is focused. This array
generates the charges during exposure. After exposure, the charges quickly transfer (as indicated by
the long vertical arrows in Figure 36) to a second storage array of identical size that is shielded from
light. The charges are held in this storage array until they are read out and processed. In the
meantime, the sensor array is reexposed to capture another image.
(144)Figure 36. Schematic representation showing how charges are transferred through a
CCD employing the frame transmission system.
The processing of charges within the storage array is similar to that for the line transmission system.
Step-by-step (as indicated by the short arrows in Figure 36), each row of charge packets in the
storage array is transferred into the horizontal shift register and read out to the amplifier to generate
the output signal.
Because a frame transfer CCD contains both sensor and storage arrays, frame transfer CCDs tend to
be physically larger than line transfer CCDs. Although frame transfer CCDs have high light sensitivity,
their larger size is a disadvantage in many endoscopic applications.
The sensor arrays in both line and frame transfer CCDs generate charges whenever they are
illuminated. During the time period in which charges are transferred from the photosensitive area,
additional charges are generated by light still incident on the sensor pixels. These charges are added
to those being transferred through the sensor array. As a result, as the image is shifted through the
array, it becomes blurred by the incoming light, with especially bright areas of the scene producing
prominent vertical bands across the image. The solution to this problem is to prevent unnecessary light
from falling on the CCD while charges are being transferred out of the sensor array. Therefore, all
videoscopes using line or frame transfer CCDs must have a means of strobing the endoscope's
illumination so that light is present only during exposure and is absent during charge transfer.
Interline Transfer CCD

One type of CCD operates with nonstrobed, continuous illumination (Figure 37). Known as the
interline transmission system, this CCD has a vertical shift register adjacent to each column of
photosites. Immediately after exposure, charges developed at the photosites are transferred in one
quick step to the storage sites in the adjacent vertical shift registers. Owing to the rapid, one-step
transfer of these charges, illumination of the CCD need not be interrupted and the CCD can continue
to collect light. In the meantime, charges transferred to the storage sites in the vertical shift registers
move step-by-step down the vertical registers into the horizontal shift register, and then to the output
amplifier. The vertical shift registers are shielded from incident light, allowing them to be emptied while
the sensor array is exposed. When the vertical shift registers are empty, a new group of charges,
representing a new image, is transferred from the photosites.
(145)Figure 37. Schematic representation showing how charges are transferred through a
CCD employing the interline transmission system.
All three types of CCDs described previously have been used in commercial video image endoscopes.
Each type has its own advantages and disadvantages in terms of physical size, circuit complexity, light
sensitivity, and illumination requirements.
Shape of the Displayed Image

All endoscopes emit a conical beam of light from their distal tip. Likewise, the objective lens at the
distal end of the endoscope captures a round image. Because of these geometric considerations,
modern fiberoptic endoscopes typically use round image fiber bundles and present a round image in
the eye piece of the instrument. By contrast, commercially available videoscopes produce a wide
variety of image shapes on the observation monitor.
Because all CCDs use a matrix of rows and columns to transfer charges, the photosensitive surface of
all CCDs is either square or rectangular. Because the objective lens of the endoscope projects a
circular image onto the photosensitive surface of the CCD, the videoscope engineer must make the
best use of the mismatch between the round image and the square sensor. The shape of the image on
the monitor screen is ultimately determined by the relative size of the projected endoscopic image on
the image sensor (Figure 38). If the endoscopic image is reduced to fit inside the borders of the
sensor, the CCD captures the entire illuminated area and a round image is displayed by the monitor. A
large number of pixels in the corners of the CCD, however, are not illuminated and therefore are not
used (Figure 38A).
(146)Figure 38. The size of the endoscopic image as projected on the CCD sensor
determines the ultimate shape of the endoscopic image on the monitor. A, The entire
endoscopic image is reduced to fit inside the CCD sensor, producing a round image. B, The
image is enlarged to cover the entire CCD surface, resulting in a square image. C, An
intermediate condition producing an eight-sided image.
If the projected image is enlarged to cover the corners of the CCD, the entire CCD is illuminated and a
square or rectangular image is displayed by the monitor. Since all pixels are illuminated, the CCD
sensor is used efficiently. A large portion of the endoscopic image, however, falls off the photosensitive
area of the CCD and is not detected and not displayed (see Figure 38B). The angle of view of the
displayed image is greatly reduced, and the endoscopist sees only the center portion of the illuminated
tissue.
A compromise between these two extremes is to project an intermediate-sized image on the CCD to
minimize the number of unused pixels while capturing as much of the endoscopic image as possible.
This results in an eight-sided image shape (see Figure 38C). All three image shapes are used by
videoscope manufacturers.
Reproduction of Color
All solid-state image sensors are inherently monochromatic devices, that is, black-and-white (B/W)
sensors. Because the silicon photosites respond only to the intensity (brightness) of the light falling on
them, they cannot distinguish the actual wavelengths (color) of the incident light, and therefore cannot
provide direct information on the hue and saturation of each point in the image (HSI color space).
For the endoscope to reproduce colora capability essential for medical useit is necessary to further
electronically analyze the incident light to determine its color attributes.
To understand the process of color reproduction, it is first necessary to understand how the human eye
and brain perceive color. All photographic and electronic imaging systems attempt to mimic the
biologic imager, the eye. The sensitivity of the human eye to light varies with wavelength (Figure 39).
The eye is most sensitive to green, and less sensitive to reds, blues, and other colors. The CCD has a
similar sensitivity response curve, but it responds to a broader range of wavelengths, including infrared
light (wavelengths greater than 780 nm) and ultraviolet light (wavelengths less than 380 nm) that fall
outside the visible portion of the spectrum to which the eye responds.
(147)Figure 39. Comparison of the relative sensitivities of the human eye and the CCD to
various wavelengths of light.
Anyone who has mixed paints has observed that when two or more colors are mixed together, we no
longer perceive two independent colors, but instead see a single, newly created, third color. In any
mixture of colors, it is not possible for the eye to detect the exact component colors. The eye differs
from the ear in this respect. If two notes of different pitch are sounded simultaneously, the ear senses
not a single intermediate tone but two clearly distinguishable notes. Our sense of hearing is analytical,
whereas our sense of color is not.
Trichromatic Vision
Nearly any color to which the human eye is sensitive can be matched by mixing light of three
colorsred, green, and blue (RGB). These are often referred to as the additive primary colors. If
three light projectors are fitted with red, green, and blue filters, and light from each is slightly
overlapped and projected simultaneously, the resulting image is similar to that shown in Figure 310.
Light reflected from the area where the red and green projectors overlap produces a mental sensation
indistinguishable from that produced by monochromatic yellow light. Likewise, overlapping red and blue
light produces magenta; blue and green light produces cyan. Where all three primary colors overlap at
the center, the eye perceives the combined reflected light as pure white with no hint of the component
colors. Thus, white light results from the mixture of all colors and is not the absence of color. By
varying the light intensities of the three individual projectors, it is possible to reproduce nearly any
spectral color.
(148)Figure 310. Additive primary colorsred, green, and bluecombine to form pure
white.
The concept of human trichromatic vision was first proposed in the early 1800s by Thomas Young
based on his experiments with red, green, and blue light projectors. His experiments, and those that
followed, caused scientists to postulate that humans perceive color through the stimulation of three
different types of neural cells (cones) within the retina. These are presumed to have the approximate
sensitivity response curves shown in Figure 311.
(149)Figure 311. Hypothetical relative sensitivities of the suggested three types of retinal
cones to various wavelengths of light.
In many ways, it is fortunate that our eyes see color in the manner described. The additive and
subtractive properties of color make it possible for a printer to produce a full-color brochure using inks
of only three colors, for a chemist to manufacture color film using dyes of only three colors, and for
engineers to design color monitors that reproduce an unlimited variety of color using phosphors of only
three colors painted on the face of the picture tube.
Theory of Color Video

Video color reproduction is based on the concepts of trichromatic vision and additive color (RGB color
space). All video images are composed from the three component colors of red, green, and blue. As
mentioned, these are also the colors used in the phosphors for re-creating the video image on the face
of the color television (TV) monitor.
Three other colors commonly encountered in TV systems are the subtractive primary colors: yellow,
magenta, and cyan. Each of these is produced by mixing equal amounts of two additive primary colors
(see Figure 310). They are referred to as subtractive primaries because they are the colors that result
when an additive primary is subtracted from white. For example, filtering red light from white light
leaves cyan, that is, blue plus green. Since the additive and subtractive primary colors are basic to all
video reproduction, they commonly appear in the "color bar" displays used for adjusting video monitors
during setup.
RGB Sequential Imaging Technology

All videoscopes may be categorized by the type of color imaging system they employ. Currently, two
types of videoscopes are commercially available: the RGB sequential videoscope and the color-chip
videoscope. Each has its own characteristics, advantages, and disadvantages.
The VideoEndoscope (the first commercial video image endoscope) introduced by Welch Allyn in
1983, as well as many videoscopes available today, employed an RGB sequential imaging system.
The components of the RGB sequential imaging system are shown in Figure 312. A monochromatic
CCD is mounted in the endoscope immediately behind the distal objective lens. The lens focuses a
miniature image of the endoscopic field on the CCD's photosensitive surface. Similar to the fiberoptic
endoscope, the video image endoscope employs a fiber bundle to bring light for illumination into the
gastrointestinal tract. However, in the case of an RGB sequential videoscope, the illumination is not the
ordinary white light as used for fiberoptic viewing; it is a pulsed or strobed, colored light.
(150)Figure 312. Schematic of an RGB sequential imaging system. A spinning filter wheel
within the video processor produces pulses of red, green, and blue light. Images captured
under this variously colored illumination are temporarily stored in the video processor prior to
simultaneous display on a monitor.
A xenon lamp within the video processor unit emits a conventional broad spectrum of white light.
However, positioned between the lamp and the fiber bundle of the endoscope is a rotating filter wheel
containing red, green, and blue filter segments (see Figure 312). As the filter wheel rotates,
alternating flashes of red, green, and blue light are emitted at the distal end of the endoscope. Owing
to the speed of filter wheel rotation, typically 20 to 30 revolutions/second (rps), these bursts of colored
light appear to the eye as a flickering white light, the phenomenon illustrated in Figure 310.
The imaging system creates three separate monochromatic images. During the fraction of a second in
which the red filter is in the light path, the CCD image sensor records an intensity (B/W) profile of the
endoscopic image as it appears under red illumination. Areas that appear reddish under natural white
light reflect heavily under red illumination and are therefore bright in the monochromatic image. Areas
of the image that are predominantly blue or green, or combinations thereof, appear dark under red
illumination.
After an intensity profile is obtained under red illumination, the filter wheel rotates to one of the opaque
segments of the wheel that separate the filter components. This prevents light from entering the
endoscope and instantly causes the interior of the gastrointestinal tract to become dark. During this
instant, the image on the CCD is read out and directed through a switching circuit for storage in one of
the three solid-state memories within the video processor (e.g., the image information obtained under
red illumination is directed to the memory area reserved for "red images").
A second exposure is made under green illumination and the resulting B/W image replica is sent to the
"green image" memory in the video processor. Finally, a third image is obtained under blue illumination
and stored in the "blue image" memory. This sequence of capturing a separate set of images for each
of the three primary colors is repeated 20 to 30 times each second; with the exact rate being
determined by the video processor.
The face of a color video monitor is actually composed of an organized matrix of red, green, and blue
phosphor dots (circles or rectangles) (see Figure 312). The monitor also contains three electron guns,
each of which scans the face of the picture tube. The beam from the "red" gun hits and activates only
the red phosphor dots. The "green" and "blue" electron guns illuminate only the green and blue
phosphor dots, respectively.
By feeding the signal stored in the red memory of the video processor to the red electron gun, the
monitor reproduces the endoscopic image as it appears under red illumination. Likewise, feeding the
signals stored in the green and blue memory arrays to the green and blue electron guns reproduces
the green and blue components of the image. (Although three guns are described here, some monitors
achieve the same results using a single electron gun.)
It is a phenomenon of human vision that when two or more sources of color are placed in close
proximity, but not overlapping, and are viewed at a sufficient distance, the colors blend together to form
a third color. This phenomenon, referred to as the juxtaposition of color sources, causes an observer to
perceive the three superimposed images from the red, green, and blue electron guns as blending
together to form a single, full-colored, natural-looking image, rather than a confusing collection of
multicolored dots. The RGB sequential imaging process is summarized in Figure 313.
(151)Figure 313. Illustration of color reproduction using additive primary color images.
Three monochrome (intensity profile) images are captured while the objects are illuminated
sequentially with red, green, and blue light. Note that all of these monochrome images are
different, indicating that the various objects illuminated absorb and reflect light differently,
depending on the wavelength of the incident light. To reproduce the image, the monochrome
images are converted into red, green, and blue component images. Superimposing these three
component images on a video monitor results in a final image of natural color.
Color-Chip Video Imaging Technology

A "color-chip" CCD is a B/W image sensor with a custom-fabricated, multicolored filter bonded to its
surface. This filter allows the CCD to directly and simultaneously resolve the color components of the
image. Often the term instantaneous single-plate color imaging system is used to emphasize that
all three color components are obtained concurrently by a single plate, or CCD.
There are many different ways to construct a color CCD. One of the simpler methods is to cover the
CCD with a precisely aligned RGB-striped filter (Figure 314A). When an image is projected onto the
photosensitive surface behind the striped filter (Figure 314B), and the information from all of the red
columns of pixels is combined, a basic red component image can be constructed. Likewise, combining
the information from pixels behind the green and blue filter stripes enables the reconstruction of the
green and blue component images, respectively.
(152)Figure 314. A, CCD with an RGB-striped filter. B, Combining the brightness

information of the individual red, green, and blue columns of pixels will allow re-creation of
the red, green, and blue image components, respectively.
In many respects, the basic components of the RGB sequential system described herein and the basic
components of an RGB-striped CCD are the same, except that the filters have changed size and
location. The filters that were previously located in the rotating filter wheel of the light source are now
miniaturized, cut into thin strips, and bonded to the surface of the CCD itself in the color-chip design.
Although RGB-striped CCDs (see Figure 314) are sometimes used in miniature TV cameras, they are
not commonly used in endoscopes. Rather, color-chip videoscopes typically employ a mosaic filter
arrangement of nonprimary colors owing to certain advantages that will be explained later.
When a mosaic filter is designed, a number of different filter patterns and color choices are available,
each with advantages and disadvantages. One of the more popular mosaic filter designs is shown in
Figure 315. In this example, the colors chosen for the filter elements are yellow, cyan, and white (no
filter). These segments are arranged in a 2 2 pixel pattern, regularly repeated over the face of the
CCD. Since the final output signals sent to the monitor must be the standard red, green, and blue
component signals, the image produced behind this yellow/cyan/white mosaic filter must be converted
into its primary red, green, and blue components. A processing algorithm for doing this is also
illustrated in Figure 315.
(153)Figure 315. Mosaic filter composed of yellow, cyan, and white filter elements. An
analysis of a representative block of four pixels is shown. The relative intensities of the final
RGB components for this block of four pixels can be calculated using the processing
algorithm illustrated on the right.
A yellow filter element absorbs blue light but passes red and green light (see Figure 310). Therefore,
pixels behind all yellow elements receive both red and green image information. Likewise, pixels
behind the cyan filter elements receive light covering both the blue and the green portions of the color
spectrum (see Figure 310). The filter-free white pixels receive light covering all three primary colors.
Thus, in a representative block of four pixels (one yellow, one cyan, and two white), three pixels
receive red component information, four pixels receive green information, and three pixels receive blue
information (see Figure 315). By adding or subtracting the information obtained from adjacent pixels
using an appropriate algorithm (see Figure 315), it is possible to derive the individual red, green, and
blue component values for each block of (four) pixels.
A yellow/cyan/white mosaic filter (see Figure 315) has a significant brightness advantage over an
RGB stripe filter (see Figure 314). When red-, green-, and blue-striped filters are used, each pixel is
filtered to receive only one of the three primary colors. A cyan-filtered pixel, on the other hand, is
exposed to both blue and green light. It is therefore more heavily illuminated than a pure blue or pure
green pixel. Likewise, pixels behind a yellow filter (red plus green) or white filter (no filtration) receive
more photons (light) than pixels behind a pure red, green, or blue filter.
Because of the increase in light passing through the yellow/cyan/white mosaic filter, a CCD with this
construction exhibits greater light sensitivity than an RGB-striped CCD. The increase in brightness
achieved by using non-primary-colored filters is obtained at the expense of the additional processing
required to later separate the mixed color signals into their primary components. However, owing to
their increased light sensitivity, color mosaic CCDs allow the videoscope designer to construct an
endoscope with a smaller light guide fiber bundle, to maximize its angle of view, and to increase the
optical system's depth of field. All of these characteristics improve optical performance. Because of
these advantages, all commercial color-chip videoscopes use color mosaic rather than RGB-striped
CCDs.
Reproduction of Motion
The color-chip videoscope has an inherent advantage over the RGB sequential videoscope in
reproducing motion. The filter wheel in current RGB sequential video processors typically rotates at 20
to 30 rps. Since each of the color component images is captured individually in sequence, it takes 1/30
second (with a 30-rps filter wheel) to capture the three component images that make up a single video
image (Figure 316). If there is relative motion between the endoscope and the object being viewed,
as often occurs during endoscopy, the three component images may differ slightly with respect to
object size and position. When these three images are subsequently superimposed on the video
monitor, it is likely that they will be misaligned.
(154)Figure 316. RGB sequential videoscope requires 1/30 second to capture all three RGB
components.
This misalignment is clearly visible if the endoscopist happens to freeze the image when it is moving
rapidly. However, this color separation is continuously present, to a greater or lesser extent, throughout
the entire examination. It gives the images an unnatural, colorful, stroboscopic appearance when there
is rapid motion of the endoscope or the object being viewed, or both. This type of color separation is
especially apparent when water droplets are sprayed across the lens to clean it.
The color-chip videoscope, on the other hand, has an inherent advantage in imaging moving objects.
Because it captures all three color components of the image simultaneously, there is never any color
separation with either moving or "frozen" images. Since illumination is continuous and unstrobed, and
since the frame rate is consistent with customary TV standards, reproduction of moving images is
always smooth and looks natural.
Another unique advantage of the color-chip videoscope is that the effective shutter speed for capturing
still images can be shortened to increase the sharpness of frozen images. The color-chip system
normally captures a new video field every 1/60 second (Figure 317). Even though this time is
relatively short, quickly moving subjects, when frozen for observation or documentation, may appear to
be slightly blurred (but without color separation) owing to movement during this 1/60-second capture
period. However, this 1/60-second capture period can be electronically shortened to a fraction of its
normal time. Although the videoscope designer can choose any appropriate time period, a typical
"fast-shutter" exposure is 1/250 second. At this speed, sharp images of the fastest-moving endoscopic
subjects are obtainable.
(155)Figure 317. Color chip videoscope captures all color components simultaneously.
Capture time can be shortened to 1/250 second to eliminate blurring of fast-moving subjects.
As in traditional film photography, the shorter the exposure period, the more brightly the subject must
be illuminated to prevent underexposure. Therefore, the fast-shutter mode may not provide enough
light for distant panoramic images. However, in situations in which it is truly needed (e.g., close-up
viewing when there is likely to be a large amount of relative motion between the endoscope and the
mucosa), the fast-shutter capture mode is very effective at producing bright, sharp, frozen images.
Inherent Advantages of a Color-Chip Videoscope

The color-chip videoscope has several inherent advantages over the RGB sequential system. Those
discussed previously include: (1) smooth, natural reproduction of motion; (2) the absence of color
separation on frozen images; and (3) a fast-shutter mode that prevents image blur of even the fastest
moving subject.
A fourth advantage of the color-chip CCD is compatibility with standard (nonstrobing) xenon light
sources. This allows use of the same light source with both fiberoptic and video image endoscopes.
Increased transillumination is a fifth advantage. Since the color-chip videoscope operates with the
same intense white light used in fiberoptic instrument systems, transabdominal illumination is usually
possible without increasing the brightness of the light source. If additional light is needed, the light
source can be set to maximum. This may cause the video image to be overly bright, but there will be
no loss of color. With the RGB sequential system, transillumination is problematic, since its strobed
light output is substantially weaker than that of a fiberscope system. Therefore, many RGB sequential
systems have a means for temporarily removing the spinning filter wheel from the light path. This
produces a constant, intense white light for transillumination. Once the filter wheel is removed,
however, the image is lost, for in most cases the illumination is so intense that it saturates the CCD,
producing a totally white screen. Even if an image is visible, it will be in B/W, since the filter wheel must
be in proper position to reproduce color.
A sixth advantage of the color-chip system is superior performance when used in endoscopic laser
therapy procedures (see later).
Laser Therapy Via the Videoscope

No video image endoscope can be effectively used with lasers that operate within the visible spectrum.
For example, it is impossible to use an argon-ion laser operating at approximately 500 nm with a video
image endoscope without distorting the image. Since the blue-green argon-ion beam is much more
intense than the background illumination of the video system, laser light completely saturates the CCD
whenever the laser is fired. The result is severe blooming and often a totally white screen with
complete loss of image detail. If the endoscope were constructed with a distal filter to eliminate the
blue-green laser light, similar blue-green image information would be removed from the endoscopic
images, resulting in abnormally colored mucosa on the monitor.
It is possible to efficiently adapt video image endoscopes for use with lasers that operate outside the
visible spectrum. For example, the neodymium:yttrium-aluminum-garnet (Nd:YAG) laser, which
produces near-infrared light at 1060 nm, is compatible with modified videoscopes. Since the Nd:YAG
laser output is outside the visible range, manufacturers commonly protect the endoscope CCD by
covering it with a filter that transmits visible light (the image) but heavily absorbs the reflected laser light
(near-infrared light). Whenever the laser is fired within the endoscopic field, the filter prevents the laser
output from reaching the CCD and leaves the image undisturbed.
Despite the addition of a filter over the CCD, RGB sequential videoscopes have two disadvantages
when used with an Nd:YAG laser. The first is a loss of the true color of the aiming beam (a conical
beam of visible light that approximates the geometry of the surgical laser beam and allows the
endoscopist to aim the invisible Nd:YAG laser at the target site). The helium-neon (He-Ne) aiming
beam used in almost all Nd:YAG lasers appears as a red spot when observed with a fiberscope or
color-chip videoscope. However, when observed by an RGB sequential videoscope, the beam is
always white. This is because the red aiming beam is illuminated continuously and therefore appears
equally bright to the CCD during all portions of the RGB imaging cycle. As a result, the RGB sequential
video processor interprets the light as being white. Because of this color loss, the bright, artificially
white aiming beam displayed on the monitor obscures the tissue effects produced by the Nd:YAG laser
and impairs observation of the laser's action. This loss of aiming beam color can be avoided by
strobing the aiming beam in synchrony with the light source filter wheel. This modification, however, is
complex and costly.
A second disadvantage of the RGB sequential endoscope for laser therapy is the relatively low
brightness of its strobed illumination. This causes two problems: (1) the intensity of the laser aiming
beam must be reduced because medical lasers have been traditionally designed to work with the
intense illumination produced by fiberoptic endoscopes, and (2) during periods of concentrated
treatment, tissue may glow at the point of laser impact. Since the burning tissue may be brighter than
the videoscope's background illumination, the glowing tissue also may cause the CCD to bloom and
mask the local tissue effects of the laser.
In contrast, the color-chip videoscope uses intense, nonstrobed white light illumination, similar to that
used with fiberscopes. The intensity of the laser aiming beam, therefore, is usually not a problem. Also,
the aiming beam always retains its red color. The result is a view similar to that seen through
fiberscopes. Because of these factors, the color-chip videoscope is usually the better choice for
endoscopic laser therapy.
Advantages of the RGB Sequential Videoscope

Having considered the advantages of the color-chip videoscope, it is now appropriate to review the
advantages of the RGB sequential videoscope. These can also be exploited when designing a video
image endoscope.
One of the major advantages of the RGB sequential videoscope is the opportunity for increased
resolution. In simple terms, resolution relates directly to the number of pixels in an image. The
color-chip system requires that information from several different pixels be combined (added to or
subtracted from each other) to obtain the red, green, and blue component values for a single point
within the image (Figure 315). In the RGB sequential system, each pixel is illuminated by red, green,
and blue light in sequence and each provides information on each of the three color components. If the
number of pixels were unlimited, the resolution of the two technologies could be similar, but in
endoscopic applications the physical size of the CCD is restricted to the space available within the
distal end of the endoscope. This places limits on the size of the CCD and the number of pixels it can
contain. Since the color-chip CCD uses several pixels to provide the same information obtained from a
single pixel in the RGB system, the latter system can theoretically produce the greatest image
resolution, assuming pixel size and number are comparable.
Because the RGB sequential videoscope uses primary-colored filters, and since the color components
are captured and processed separately within the videoprocessor, this type of videoscope can provide
very accurate color information. Although both systems produce natural-looking images, the RGB
sequential system can theoretically produce a truer color signal, which may make it more suitable for
image analysis research.
Recent Advances in RGB Sequential Videoscopes

Recently introduced RGB sequential videoscope systems are engineered to overcome many of the
limitations of previous designs. To prevent the previously described problem of color separation of
frozen images, some new RGB sequential video processors incorporate an anti-color-slip circuit to
analyze images in real time and freeze the image at the moment when color separation is at a
minimum (Figure 318). In early RGB systems, the processor "froze" the image displayed on the
monitor at the exact instant when the image capture function was activated. Instead of capturing this
initial image, activation of the freeze function of newer RGB video processors triggers a special capture
circuit that analyzes the stream of images for the next 0.25-second, that is, the next 5 rotations of the
filter wheel (Figure 318). From these 5 complete images (a total of 15 RGB component images), the
circuit selects the set of RGB component images that exhibits the least amount of color separation. In
the example shown in Figure 318, the capture circuit found that relative motion was minimal during
RGB component images 8, 9, and 10, these being obtained during the third and fourth rotations of the
filter wheel. The circuit therefore holds these three RGB component images in memory as the best
possible still image of the subject, and displays this frame on the monitor as the best frozen image.
Even though the system is able to choose only from the images obtained within a 0.25-second period,
the circuit is remarkably effective in reducing color separation.
(156)Figure 318. EVIS-200 anti-color-slip capture circuit analyzes images obtained during a
0.25-second time period. RGB components 8, 9, and 10 exhibit the least amount of relative
motion and hence the best frozen image.
Designing for Compatibility

When endoscopy entered the video age, it left behind the compatibility advantages of fiberoptic
instruments. When designing a fiberscope, the designer can select a fiber bundle from a wide range of
sizes in order to make an optimum compromise among fiber bundle diameter, image resolution,
accessory channel diameter, and insertion tube diameter. Thus, a manufacturer's line of fiberscopes
frequently contains a wide range of optical systems with various levels of image resolution. All of these
instruments are fully compatible with a large number of light sources, still cameras, and ancillary
equipment.
Videoscopes, on the other hand, greatly restrict optical design flexibility. As much as one half of the
circuitry in a video processor is designed around the specific CCD with which the processor is intended
to operate. Until recently, no manufacturer had designed a video processor that would work with
several different CCDs. This meant that within a single videoscope system, all instrumentsfrom the
largest colonoscope to the slimmest gastroscopeused identical CCDs (although some
manufacturers modified the image display to hide this fact). Because video processors were tied to a
single CCD, the number of available videoscope models was limited when contrasted with the wide
range of comparable fiberoptic models.
The Olympus EVIS-200 system was the first video image endoscopy system designed around a family
of compatible CCDs. The CV-200 video processor drives three distinct CCDs, each of which differs in
size and image resolution capability. An advantage of this strategy is that a family of compatible CCDs
allows for development of a wide range of videoscopesincluding large-diameter high-resolution
instruments, specialty endoscopes, thin pediatric endoscopes, and large-channeled therapeutic
instrumentsall of which are compatible with the same videoprocessor. For a single video processor
to drive several different CCDs, the various CCDs usually must have similar electrical and physical
characteristics. In particular, the video processor must specifically compensate for differences in pixel
number, illumination requirements, data transfer rates, and drive circuitry.
Functions of a Typical Video Processor

Figure 319 schematically illustrates the functions of the various electrical circuits within a typical RGB
sequential video processor. The CCD image sensor (in the distal end of the endoscope) receives both
power and drive (timing) signals from the video processor. The timing signals control the readout of the
CCD and the transfer of charges to the horizontal shift register (see earlier discussion). From the shift
register, the image signal is fed through an amplifier on the CCD and into the preprocess circuitry of
the video processor.
(157)Figure 319. Simplified schematic illustrating some of the basic functions of an RGB
sequential video processor. A/Danalog to digital; D/Adigital to analog.
The preprocess circuitry is responsible for electrically isolating the patient from the potentially
dangerous high-voltage circuitry of the videoprocessor, for initiating automatic brightness control,
and for adjusting the chroma (color) and white balance of the image. The preprocess circuitry further
amplifies the signal and often performs additional image processing functions such as edge
enhancement.
The signal then passes through an analog-to-digital (A/D) converter that changes the signal from
analog to digital form. The digitized image is then directed through a switching circuit for storage in
one of the red, green, or blue segments of the image memory array. Images from the digital memories
are next passed through a size-correction circuit to adjust the relative size of the endoscopic image for
presentation on the video monitor. A second circuit then adjusts the relative position of the endoscopic
image and sizes and positions any subscreen image that may be added to the screen along with the
main image. An example of a typical subscreen image is a reduced display of the real-time endoscopic
image that is added to the screen whenever the main image is frozen.
At this point, the developing monitor image is still in digital form, but it has been adjusted in size and
position, with subscreen images (if any) added. The image then passes through a digital-to-analog
(D/A) converter to change the image back to analog form. The developing monitor image then passes
through a masking circuit that adds a mask (typically black) around the image(s) to provide a uniform
background color. Finally, the image passes through postprocessing circuitry to make the video signal
conform to a recognized video signal standard so that the image can be displayed on a standard video
monitor.
Figure 319 also schematically illustrates some of the mechanical components of the video processor
and light source. Light from the source lamp first passes through an infrared absorption filter to remove
nonvisible heat rays. The light then passes through a lens that focuses it on the tip of the fiberoptic light
guide bundle within the umbilical cord of the videoscope. An iris in the light path controls the brightness
of the light transmitted to the endoscope, and a filter wheel modifies the color of the light (described
previously). The motor rotating the filter wheel is regulated by a control circuit to ensure that the wheel
spins at the precise speed required by the video processor. Detectors placed adjacent to the filter
wheel identify at all times which filter segment (red, green, or blue) is in the light path.
Synchronization circuitry within the video processor ensures that all functions of the video processor
and light source are coordinated with the video output signal. These functions include synchronization
of filter wheel rotation, exposure and readout of the CCD, memory transfers within the video processor,
and image freeze control.
Video processors also generally require endoscope discrimination circuitry to identify the model (or
type) of endoscope connected to the processor. Endoscope discrimination allows the system to
compensate for differences in CCD type and endoscope length.
Although Figure 319 summarizes many of the basic functions of a video processor, for simplicity
some typical functions are not illustrated. Excluded are the connections to an external keyboard for text
input; circuitry for superimposing text on the monitor image; and circuitry for communicating with
external computers, printers, and image documentation devices.
Universal Light Source

The first commercial videoscope systems combined the video processor and light source in a single
unit. This was convenient given the close interaction required between the two devices. Recently,
however, "universal" light sources have been introduced. The concept of an all-purpose light source
originated from a desire to increase the flexibility of endoscopy systems. True universal light sources
are compatible with a wide variety of endoscopes, both fiberoptic and video, and also make it easier to
upgrade the endoscopic system as a whole. For example, the Olympus CLV-U20 universal light source
is compatible with fiberscopes (OES series), RGB sequential videoscopes (EVIS-200 series), and
color-chip videoscopes (EVIS-100 series). An effective universal light source has automatic brightness
control for both video and fiberoptic instruments, automatic flash photography capabilities for
fiberscopes, and the ability to remove the rotating RGB filter wheel from the light path when the light
source is used with color-chip and fiberoptic instruments.
Video Standards
All cathode ray tube (CRT) monitors, whether used for computer or video applications, "paint" an
image on a screen with a scanning electron beam. This beam typically starts in the upper left corner
and scans the monitor face, line by line, from top to bottom. The energy of this electron beam causes
the phosphors applied to the back of the screen to glow briefly, thus creating an image. The left-to-right
movement of the beam is referred to as horizontal scanning; top to bottom movement is referred to as
vertical scanning.
Figure 320 illustrates two scanning methods. One common protocol requires that the beam start at
the top of the monitor and scan the image downward on the screen, one line at a time, until it reaches
the bottom. Figure 320A illustrates the state of the monitor in the middle of such a process. (The red
arrow in the figure points to the last line scanned just before interrupting the process.) By the time the
beam reaches the bottom of the screen, the lines scanned first (those at the top) have begun to fade
away (Figure 320B). The monitor face is then repainted, starting again at the top and working down. If
each new image is slightly different from its predecessor, the system depicts motion, each newly
scanned image being the next frame in a motion sequence.
(158)Figure 320. Comparison between the scanning protocols for non-interlaced (A and B)
and interlaced (C and D) video monitors.
A disadvantage of this simple scanning protocol is that the phosphors must decay very rapidly in order
that the screen just ahead of the scanning beam be continuously black for repainting. The quickly
fading phosphors cause an annoying rolling or flickering of the screen as a dark horizontal band
continuously precedes the area of the screen being repainted.
To address the problem of flickering, an interlaced scanning protocol was developed. An interlaced
scan starts at the top, scans line 1, skips line 2, scans lines 3, 5, 7, and so on, until all of the odd lines
are scanned (see Figure 320C). The beam then returns to the top and begins filling in the even lines.
By the time the even lines are completed, the odd lines have begun to fade away (see Figure 320D).
The odd lines are then refreshed as the entire process repeats. The advantage of this system is that
older lines, which are fading and being repainted, are interlaced with a full screen of brightly glowing,
recently scanned lines. Screen interlace, therefore, noticeably reduces screen flicker.
The TV signal for creating an interlaced screen must alternate between segments of video information
for scanning the odd lines and information for scanning the even lines. The odd lines represent the first
field of the image; the even lines represent the second field. Together, the two fields paint the entire
screen once and create a single video frame. The process then starts over again, first one field, then
the other, until a second complete picture or video frame is displayed.
The specifications of the interlace system described were standardized by the Electronic Institute of
America (EIA) and adopted by the U. S. Federal Communications Commission in 1941. Under the EIA
RS-170A standard, used for all broadcast and closed-circuit TV systems in the United States, a video
frame consists of 525 horizontal scan lines, of which 483 are actually observable on the screen.
Furthermore, the RS-170A standard requires that 60 fields (30 frames) be displayed during each
second of live video. Since a frame is composed of 525 lines, a total of 15,750 lines of information (30
525) are processed by the video system each second.
In 1953, the National Television System Committee (NTSC) proposed a color encoding method that
expanded on the RS-170A standard to become the first B/W-compatible, simultaneous color system
used for public broadcasting. The NTSC standard for color TV signals is the universal color encoding
method used in the United States, Canada, and Japan. In contrast to the NTSC standard, which is
based on 525 lines/frame and 30 frames/second, most European countries conform to the
incompatible PAL or SECAM color TV standards that specify 625 lines/frame and 25 frames/second.
Figure 321 displays a typical portion of an NTSC video signal for a single, active horizontal scan line.
The word active is emphasized, since out of the specified 525 scan lines required per frame,
approximately 35 do not contain any picture information. Instead, these are used by the inactive
electron beam as it travels in zigzag fashion from the bottom of the screen at the end of each field back
to the top. In addition to this loss, a portion of each horizontal scan line is also used for purposes other
than picture information. This nonpicture portion of the signal is used for synchronization and color
reference signals. Altogether, approximately 30% of the video signal used in the NTSC process is not
available for transmitting true picture information. The reference and synchronization segments of the
signal are essential, however, for reproducing true color and synchronizing the monitor scan that is
reconstructing the image with the device (e.g., video processor) that is generating the image.
(159)Figure 321. Representation of a portion of an NTSC composite video signal for

scanning a single horizontal line.
Each horizontal line of the video signal is divided into two parts (see Figure 321): the active video
period, and a horizontal blanking (H-blanking) period. During the active video period, the electron
beam paints one horizontal line of the image. During the horizontal blanking and retrace interval, the
beam returns invisibly from the right to the left side of the screen to begin another line. The total time
allotted to the scan of a single horizontal line is approximately 63.6 sec, of which approximately 52
sec is spent generating a single horizontal line of glowing phosphor dots on the face of the picture
tube. The H-blanking period contains a horizontal sync pulse and a 3.58-MHz color burst signal. This
color burst signal establishes the reference phase and frequency of the color signal and serves as a
reference to the circuit that is decoding the color information.
The active video portion of the signal contains both a luminance (brightness) signal and a
chrominance (color) signal. The luminance signal describes the brightness of each point in the scan
line and is the only part of the signal used by B/W TV monitors. The luminance signal, the
low-frequency component of the active video signal, ranges between the signal levels representing
pure black and pure white. Superimposed on the luminance signal is the higher-frequency
chrominance signal. Whereas the luminance signal determines the brightness of each point in the
horizontal scan line, the chrominance signal establishes the colorspecifically the hue and saturation
of each point along the scan line. At any instant, the color saturation level is determined by the
amplitude of the chrominance signal, and the hue is determined by the phase difference between the
chrominance signal and the reference color burst. The chrominance signal and reference color burst
may be up to 360 degrees out of phase, all colors being encoded in this 0- to 360-degree phase range.
A signal similar to that represented in Figure 321 is repeated until 245.5 (262.5 - 17) lines are traced,
at which point the electron beam is at the bottom of the screen. Then a vertical blanking and retrace
(V-blanking) signal hides the beam as it zigzags from the bottom of the screen to the top. Once the
beam returns to the top, blanking stops and a horizontal scan signal (see Figure 321) returns to
create the first line of the next field.
The NTSC video signal (see Figure 321) is a composite video signalthat is, luminance
(brightness) and chrominance (color) information are combined in one signal. Although a composite
signal is convenient for broadcasting, image quality deteriorates to some extent as the color
information is encoded into and decoded from the brightness signal. Most videoscopes today also have
a means of outputting the image in a luminance/ chrominance (Y/C) video format. Y/C video
connections maintain the brightness and chrominance information on two separate wires. Monitors,
videocassette recorders (VCRs), and other peripheral equipment communicating with Y/C signals
provide better image reproduction than composite video equipment.
A third method of interconnecting video equipment is to use red, green, blue, sync (RGBS) cables. This
video communication method is the best means of preserving video image quality, since the
information for the red, green, and blue image components and synchronization is carried separately
over four different wires, so that the information is never mixed. Most videoscopes provide the option of
using any or all of these three video connection methods (NTSC composite, Y/C, or RGBS).
White Balance
The various lamps used in endoscopic light sources emit light ranging from a reddish tint (halogen
lamps operating at a color temperature of 3700 K) to white (xenon lamps operating at 5500 K) to a
bluish tint (metal halide lamps operating at 6000 K). In addition to the type of lamp used, the color of
the light emitted at the distal end of an endoscope is further affected by the length and optical
characteristics of the fiberoptic bundle carrying the light from the lamp. To reproduce white objects as
pure white on the monitor, regardless of the tint of the illuminating light, the video system must be
white balanced. This process produces a signal with equal parts of red, green, and blue, regardless of
the color temperature of the illumination. If, for example, the illumination system produces more red
light than blue or green, the video processor must proportionally reduce the red output signal to equal
the blue and green signals. A videoscope can be white balanced by pointing the endoscope at a white
object and activating the automatic white balance circuitry. During white balance, the video processor
automatically adjusts its internal processing circuits to compensate for the color temperature of the
illumination system. Once this adjustment is made, the system is "white balanced." All colors will
remain true as long as the same endoscope and illumination source are used.
Although it was stated previously that white balance attempts to achieve equal amounts of red, green,
and blue light, this should be clarified. Video systems are designed to mimic human vision. The eye is
most sensitive to green and least sensitive to blue light (see Figure 311), a fact that must be
accounted for in the final video signal. When the RGB color information is combined into the final
composite signal in the NTSC system, the composite brightness signal is not the sum of equal parts of
the individual red, green, and blue brightness signals. Rather, the NTSC luminance signal equals 30%
of the red, plus 59% of the green, plus 11% of the blue luminance values (1 lumen white = 0.30 lumen
red + 0.59 lumen green + 0.11 lumen blue). When red, green, and blue video primary colors are mixed
in these proportions, the result appears to the eye to be pure white.
Field Versus Frame Capture

The distinction between a video field and a video frame is significant when images of rapidly moving
subjects are captured. In extreme conditions, the image may shift significantly during the time it takes
(1/30 second) to capture both fields of the video image. Because the fields are captured sequentially,
the image captured by the first video field may be slightly displaced from the image captured by the
second field. Owing to this offset, the edges of objects within the image may appear ragged when the
two images are interlaced. The faster an object is moving, the greater the difference between the
images in the two fields.
When these two frozen offset fields are displayed together, there may also be a disturbing flicker in the
image as the monitor alternately reproduces the two displaced fields. Flicker can be reduced in a
"frozen image" by switching the documentation device from the frame mode to the field capture mode.
In field mode, every other line is displayed (the second field is removed from the display), resulting in a
single, crisp, nonflickering image from the first field. However, because the displayed image lacks the
second field, vertical image resolution drops accordingly. It is often better to operate the videoscope in
frame mode and, if flickering occurs, to take a second photo while endoscope movement is minimized.
Video Resolution
The term resolution refers to the smallest element of an image visible to the human eye. In video,
resolution measurements are based on what a viewer with normal vision can distinguish when
positioned at a distance from the monitor that is approximately eight times the height of the picture on
the monitor screen. Vertical resolution is expressed as the maximum number of horizontally oriented
lines, from the top to the bottom of the screen or vice versa, that can be made visible to the observer.
NTSC TV images have an aspect ratio of 4:3. This indicates that the image is 4 units wide and 3 units
high. Horizontal resolution is specified in a manner comparable with that of vertical resolution by
determining the maximum number of vertically oriented lines that the viewer can distinguish over three
fourths of the TV screen width. Resolution therefore equals the maximum number of horizontal or
vertical lines that are distinguishable in a square area of the monitor covering 75% of the monitor's
width (Figure 322). (Although this description outlines the basic theory on which video resolution
specifications are based, manufacturers typically use standardized test patterns and electronic
instruments to determine video resolution.)
(160)Figure 322. Vertical resolution is determined by the maximum number of horizontal

lines the system can reproduce from the top of the screen to the bottom. Owing to the 4:3
aspect ratio of an NTSC monitor, horizontal resolution is determined by the maximum number
of vertical lines reproducible within three quarters of the monitor's width.
Other Advances in Video Image Endoscopes

Video image endoscope technology has improved significantly since its introduction in 1983. Recent
advances in CCD manufacturing have reduced the physical size and increased the resolution of CCDs.
Smaller chip size allows the videoscope designer to reduce the physical dimensions of the videoscope
and thereby improve the handling and maneuverability of the instrument. The large size of early CCDs
necessitated larger insertion tube diameters or smaller working channels, or both. In addition, the rigid
distal tip housing the CCD in early videoscopes was typically quite long, making the instrument less
maneuverable than fiberoptic models. This was particularly true of video image duodenoscopes.
However, advances in CCD technology have allowed most current video image endoscopes to rival
their fiberoptic counterparts in all respects.
In contrast to first-generation videoscopes, current instruments offer simultaneous display of real-time
and frozen images, plus a wide range of options for displaying text and system information on the
monitor screen. In general, image brightness, dynamic range, angle of view, and image resolution have
all been improved. In addition, remote switches on the control body of current videoscopes provide the
endoscopist with direct operation of ancillary equipment.
Factors to Consider When Evaluating a Video Image Endoscope

The video image endoscope is a technologically advanced and complex clinical tool. Comparisons of
various models have been reported in the literature.810(161) It is difficult, however, to identify any
single design criterion as the deciding factor in selecting the best videoscope for a particular clinical
application. Each of the following criteria must be considered:
1. Image quality (wide angle of view and good depth of field, high image resolution, wide dynamic
range, a high signal-to-noise ratio, good image contrast, minimum blooming, accurate color,
and excellent frozen image quality).
2. Illumination characteristics (adequate image brightness under all clinical conditionseven
illumination from center to edge of the imageand automatic brightness adjustment as viewing
distances change).
3. Basic endoscope functions (responsive handling and appropriate insertion tube characteristics;
smooth tip angulation; a control section of appropriate shape and weight; well-positioned
angulation knobs and valves; and good suction, insufflation, and lens washing performance).
4. Basic specifications (appropriate insertion tube diameter, sufficient accessory channel capacity,
and availability of a full range of instrument models).
5. Suitability for special therapeutic procedures (good image protection from electrosurgical noise
and minimum laser interference).
6. System features (easy-to-understand video processor controls; adequate keyboard functions;
sufficient image and color controls; accessible endoscope switches for control of remote devices;
and acceptable equipment size, weight, and transportability).
7. System expansion and integration (full capability for interfacing the videoscope system with hard
copy devices, video tape recorders, and computers; the ability to combine endoscopic images
with database information; and compatibility with existing fiberoptic instrumentation).
Summary
Since the early 1990s, the video image endoscope has supplanted the fiberscope as the preferred
instrument for most clinical applications. The availability of two distinct video imaging technologies
allows the endoscopist a choice of systems according to clinical requirements. Recent improvements
in RGB sequential systems have increased image resolution, reduced the problem of color separation
in frozen images, and expanded the range of videoscope models. Of the two imaging technologies, the
RGB sequential system has the potential for highest image resolution and purest color information
(perhaps useful in the future for computer analysis of endoscopic images). In contrast, the color-chip
system has inherent advantages in image reproduction when the subject is moving, offers a fast
shutter mode for sharp frozen images, produces better transillumination, provides improved
performance with Nd:YAG laser therapy, and is compatible with standard xenon light sources.
Glossary
Additive primary colors
The colors red, green, and blue. These can be combined in

various proportions to reproduce all other spectral colors.
Aspect ratio
The ratio of width to height of the video image. In NTSC

systems, the video image is 4 units in width by 3 units in
height.
Automatic brightness control
A combination of functions that continuously monitors and

automatically adjusts the light source and video signal
levels to maintain optimum image brightness on the video
monitor. Since the level of mucosal illumination is
inversely proportional to the square of the distance between
the illumination source (distal end of the endoscope) and
the mucosa, endoscopic images darken as the endoscope
moves away from an object or the gut wall. Automatic
brightness control compensates for this by (1) automatically
adjusting the brightness of the light source and (2) using an
automatic gain control circuit within the video processor to
electronically increase the amplitude of the video signal if
the image is below average brightness.
Blanking
The process of cutting off the electron beam in the picture

tube during the retrace period to make the retrace of the
beam invisible on the monitor.
Blooming
A condition in which image highlights become blurred or

defocused on the monitor owing to overexposure of a
pixel(s) to light (pixel saturation). The charges produced in
the saturated pixel(s) spill over into adjacent pixels,
enlarging the representation of the highlight on the monitor.
CCD
An abbreviation for charge-coupled device. A solid-state

device that, through the photoelectric effect, transforms
light (images) into finite quantities of electrical charges and
transfers the charges as packets through charge-coupling to
a processing circuit.
Chrominance signal
The portion of a composite video signal that contains the

color information.
Color burst
The portion of a composite video signal that establishes a

reference for decoding the chrominance signal. The color
burst consists of several cycles of a sine wave at the

chrominance subcarrier frequency.
Color-chip videoscope
A common term for describing a video image endoscope

that employs an instantaneous single-plate imaging system.
Composite video
A video signal in which the elements of luminance

(brightness) and chrominance (color) are encoded into a
single signal. The luminance component of a composite
video signal is compatible with B/W video transmissions.
Contrast
The range of light and dark values within a picture, or the

ratio between maximum and minimum brightness values.
Low-contrast images appear mainly as shades of gray, and
high-contrast images appear primarily as blacks and whites
with little gray.
Digital
Information quantified in discrete values. In contrast to

analog systems that handle continuously varying signals,
microprocessors and logic circuits require digital
information. Whereas the transmission and recording of
analog information (e.g., long-playing records, analog audio
tapes) is subject to deterioration and noise, digital-based
information (e.g., computer files, compact discs) can be
transmitted and duplicated completely free of degradation.
Although standard video monitors are analog devices, much
of the image processing that is performed in the endoscope
video processor is accomplished after the image has been
digitized. Endoscopic documentation systems that archive
endoscopic images via computer also store the image in a
digital format.
Dynamic range
The range in signal strength from a minimum level (dark

signal) determined by noise to a maximum level (bright
signal) beyond which the CCD does not output an increase
in signal level with a corresponding increase in incident
light. A videoscope with wide dynamic range produces
good image detail in both dark and brightly illuminated
regions of the image.
Field
One of the two equal parts into which a television frame is

divided in an interlaced system of scanning. In the NTSC
system, a field consists of 262.5 horizontal scan lines.
Frame
The total amount of instantaneous picture information

perceived by the viewer, consisting of two interlaced odd
and even fields. An NTSC video frame consists of 525
horizontal lines.
Frame transmission system
A CCD charge transfer system in which one frame of image

information is captured in a sensor array and is then
transmitted at one time to a separate storage array on the
CCD itself. The disadvantage of this design is that these
CCDs tend to be larger than other types of CCDs. Frame
transmission CCDs typically require that the endoscope has
a longer rigid distal tip or a larger diameter.
Freeze
To display a single still image from an ongoing sequence.

When frozen, either one video field (field capture) or one
video frame (frame capture) is stored in digital memory or
displayed on the monitor.
Highlights
Areas of maximum brightness within an image, occurring

in regions of highest illumination or greatest reflection.
Horizontal blanking and retrace
H-blanking) The blanking signal produced at the end of

each horizontal line while the electron beam invisibly
returns from the right to the left side of the monitor screen.
HSI color space
A method of analyzing and representing color based on the

attributes of hue, saturation, and intensity. (See also RGB
color space.)
Hue
A color attribute that describes a pure color, such as red,

yellow, green, purple, etc. Hue is typically what is meant by
the term "color." (See also HSI color space, Intensity, and
Saturation.)
Incident light
Light falling directly on an object.
Instantaneous single-plate
A type of imaging system used in home video cameras
color imaging system
and some video image endoscopes. The system uses

standard white light for illumination. Light reflected from
the object under observation is passed through a colored
mosaic filter bonded to the CCD surface. The advantages of
this type of imaging system include smooth reproduction of
rapidly moving images without strobing or color separation
and compatability with standard white light sources. (See
also RGB sequential imaging system.)
Intensity
A color attribute referring to the gray level (black and white

value) or relative brightness or darkness of the color.
Interlaced scanning
A scanning protocol in which the monitor generates a

complete image by first scanning all odd-numbered lines
(first field), followed by a second scan interleaving the
even-numbered lines (second field).
Interline transmission system
A commonly used CCD charge transfer system. Unlike the
frame transmission system, in which one frame of

information is transmitted all at once, the interline system
transmits signals each time the horizontal direction of the
chip is scanned. The CCD structure is more complex, but
the size of the CCD is smaller than chips that use the frame
transmission system.
Line transmission system
A CCD charge transfer system that requires a smaller

number of input/output connections than other CCD
designs having an equal number of pixels. A line
transmission CCD allows miniaturization of electronic
circuitry, resulting in smaller endoscope diameters and
shorter rigid distal tip lengths.
Luminance signal
The portion of a composite video signal that represents the

brightness of the scan line.
Moir
A wavy or satiny effect produced by the convergence of

lines or patterns. Moir is often seen when a regularly
repeating pattern (such as a honeycomb fiberbundle pattern)
is projected on an imaging device (such as the pixel matrix
in a CCD).
Monochromatic
Having only one color. Or alternately, a "black-and-white"

representation of the brightness profile of a color image.
Noise
A signal disturbance that obscures or reduces image clarity.

Noise can appear to be either random and incoherent or
fixed to the image.
NTSC
An abbreviation for the National Television System

Committee. In 1953, the color encoding method proposed
by this committee was adopted by the Federal
Communications Commission as the first B/W-compatible,
simultaneous color system for public broadcast. The United
States, Canada, and Japan use the NTSC system for
broadcast television. (See also PAL and SECAM.)
PAL
An abbreviation for phase alternating line. The color video

transmission system used by the majority of European
countries. The PAL television standard uses more scan
lines (625 lines/frame versus 525 lines/frame) but a lower
frame rate (25 frames/second versus 30 frames/second)
than the NTSC system used in the United States and Japan.
Pixels
An abbreviation for picture elements, the smallest unit

component of an image. The term also refers to the discrete
sensor elements that make up a solid-state image sensor.
The number of pixels on a CCD is one factor that
determines the resolution of an imaging system.

Resolution
A measure of a video system's ability to reproduce detail.

Image resolution is determined subjectively by the number
of lines that can be recognized on a television monitor
displaying an Electronic Institute of America (EIA)
Resolution Chart. Resolution is measured separately in the
horizontal and the vertical direction.
RGB color space
A method of representing color based on the relative

amounts of the three additive primary colors (red, green,
and blue) contained in the color under analysis. When
complex color image processing is performed, it is often
easier to convert the RGB representation of color into HSI
(hue, saturation, intensity) values. (See also HSI color
space.)
RGB sequential imaging system
The type of imaging system used by the majority of video

image endoscope manufacturers. The system requires a
strobed light source outfitted with three primary color
filters. The endoscopic field under observation is
sequentially imaged under red, green, and blue
illumination. The three sequentially obtained images are
then processed and recombined on the viewing monitor.
Color reproduction is excellent with this system; relative
motion, however, results in color separation in the
re-created image. RGB sequential imaging systems are also
able to use CCDs with fewer input/output connections,
which results in videoscopes with smaller diameters and
shorter rigid distal tips.
RGB sequential videoscope
A video image endoscope employing an RGB sequential

imaging system.
RS-170A
A video scanning standard developed by the EIA. It

specifies a video format of 525 total lines, a 2:1 interlaced
frame rate of 30 Hz, and a 15.734-KHz horizontal line scan
frequency.
Saturation
A color attribute that describes color purity, or the degree to

which a pure color is diluted with white. A high-saturated
color has a low white content (e.g., vivid red), whereas a
low-saturated color has a high white content (e.g., pink). In
composite video signals, color saturation is determined by
the amplitude of the chrominance signal.
Scanning
The process in which a two-dimensional image is analyzed

and converted to line information. Scanning also refers to
the opposite process in which lines of information are

reassembled to form an image. Monitor scanning is
performed from left to right, top to bottom. As the beam
scans across the face of the monitor, it causes the phosphors
coating the back of the monitor glass to glow in intensity
proportional to the intensity of the electron beam hitting the
phosphors.
SECAM
An abbreviation for sequential color with memory, from the

French "sequential couleur avec mmoire." A color
broadcast system used in several parts of Eastern Europe.
The system, which is based on 625 lines/frame and 25
frames/second, features strong resistance to color changes
and good immunity to transmission interference, but
performs poorly in the presence of heavy electrical noise.
Sensitivity
A measure of a video system's ability to produce usable

pictures in low light. The more sensitive a system is, the
less light it requires.
Signal-to-noise ratio
The ratio, expressed in logarithmic terms, between a useful

video signal and the unwanted noise or interference
superimposed on the signal. The higher the ratio, the
"cleaner" the picture. The signal-to-noise ratio is an
indicator of how fine details appear relative to background
noise. Noise typically appears as a spotty or grainy texture
within the image.
Subtractive primary colors
The colors yellow, magenta, and cyan. These colors are

referred to as subtractive primaries because they are the
colors that result when an additive primary color is
subtracted (filtered out) from white light.
Synchronization (sync)
Electronic pulses that are inserted in the video signal for the
purpose of timing and correctly assembling the picture
information. These timing signals synchronize the transfer
of image information from the camera to the processing
circuitry and video monitor.
Trichromatic vision
A theory that there are three unique types of receptor cells

(cones) in the eye, each of which has selective sensitivity to
photons of red, green, or blue wavelengths.
Vertical blanking and retrace

(V-blanking)
The invisible return of the electron beam to the top of the picture tube at the
completion of a field scan.
White balance
A procedure for adjusting the circuitry of a video camera or

video processor to compensate for the (nonwhite) color of
the incident light falling on the object being imaged.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Morrissey JF. Thoughts on the VideoEndoscope. Gastrointest Endosc 1984;30:43.

Sivak MV Jr, Fleischer DE. Colonoscopy with a video endoscope. Preliminary experience.
Schapiro M. Electronic video endoscopy. A comprehensive review of the newest technology
and techniques. Pract Gastroenterol 1986;10:818.
Graham DY, Smith JL, Schwartz JT. Endoscopic television: Traditional and video endoscopy.
Demling L, Hazel HJ. Video endoscopy: Fundamentals and problems. Endoscopy
1985;5:1679.
Sivak MV. Video endoscopy. Clin Gastroenterol 1986;15:20534.
Schapiro M, Auslander M, Schapiro MB. The electronic video endoscope: Clinical experience
with 1200 diagnostic and therapeutic cases in the community hospital. Gastrointest Endosc
1987;33:638.
Knyrim K, Seidlitz H, Hagenmuller F, Classen M. Videoendoscopes in comparison with
fiberscopes: Quantitative measurement of optical resolution. Endoscopy 1987;4:1569.
Knyrim K, Seidlitz H, Vakil N, et al. Optical performance of electronic imaging systems for the
colon. Gastroenterology 1989;96:77682.
Video colonoscope systems. Health Devices 1994;23:151205.
Chapter 4 The Endoscopy Unit

(162)
(163)
MICHAEL V. SIVAK, JR., M.D.
RUSSELL MANOY, M.SC., F.C.S.D.
MARTIN E. RICH, A.I.A.
This chapter does not attempt to define the ideal endoscopy unit. The operational characteristics of
every unit are so unique that there can never be a single design plan that would be suitable in every
case. Rather, our purpose is to define principles of design and function that can be applied in most
instances. By considering the endoscopy unit as a concept rather than as a place, certain general rules
and guidelines can be evolved that are applicable in many situations.
Once a decision has been made to construct a new unit or to modify an existing facility, there is a
natural compulsion to think in terms of diagrams and blueprints, rooms and walls. This is the wrong
place to begin. Instead, the first and most important step is to define very clear concepts as to how a
unit functions. In fact, there are general principles of function that are common to all units. But there
are also operational characteristics that are unique, so that the second step in the process of building a
unit, also vitally important, is a careful and detailed assessment of the requirements that have
determined that a new unit should be built. This has two aspects: an analysis of historical data
concerning endoscopy operations and a forecast of future requirements. Past experience is,
unfortunately, the least reliable frame of reference for the future. Its value, however, lies in the
discovery of simple principles that shape action and in the measure of confidence it inspires in the
existence of logical sequences of cause and effect.
Construction of an endoscopy unit frequently involves great expenditures of time, money, and effort.
Mistakes can lead to long-term inefficiency, high operational costs, loss of revenue, frustration on the
part of those who work in the unit, and dissatisfied patients. It is therefore imperative to remember two
axioms: As far as possible, form must fit function, and the fundamental purpose of the endoscopy unit
is to care for patients.
Evolution of the Endoscopy Unit

It is germane to ask why we have endoscopy units and how they come about.
The practice of dedicating a specific area of a hospital, an outpatient clinic, or even a private office to
the performance of endoscopic procedures is relatively recent. Although many factors have made
allocation of dedicated space for gastrointestinal (GI) endoscopy inevitable, the driving force, operative
in virtually every health care facility, has been growth in the demand for endoscopic services.
In its early phases, the development of endoscopy was limited by a lack of technically adequate
means. With the advent of the coherent fiberoptic bundle and its incorporation into a flexible
endoscope, the stage was set for growth of truly exponential proportions. Not only were risk and
requisite levels of skill reduced by flexible instruments, but sustained technical improvement also
steadily expanded diagnostic capability. As a result of remarkable improvements in GI endoscopes
since the early 1970s, certain diagnostic procedures such as upper GI endoscopy have become
technically less difficult and are thus performed by greater numbers of physicians. The ready
availability of such procedures combined with their superior diagnostic accuracy has naturally led to
greater utilization.
GI diseases themselves have added impetus to the rapid evolution of the endoscopy unit. For
example, GI endoscopy developed rapidly in a country such as Japan where gastric carcinoma is a
significant problem. Colon cancer and the recognition of its relation to the colon polyp have promoted
the development of colonoscopy and flexible sigmoidoscopy in the United States and other countries in
which the prevalence of this disease is high. GI diseases such as these have thus contributed to
steady growth year by year in the number of endoscopic procedures performed.
Therapeutic endoscopy is a somewhat more recent invention. Colonoscopic polypectomy is without
doubt the prototypic therapeutic maneuver for the modern era of therapeutic endoscopy. Previously,
certain therapeutic maneuvers existed, such as the extraction of foreign bodies and sclerotherapy.
However, these were not performed in numbers that could remotely approximate the potential for
colonoscopic polypectomy. The further development of side-viewing endoscopes in the early 1970s
provided an important new approach to the investigation of diseases of the duodenal papilla, bile duct,
and pancreas. Although diagnostic endoscopic retrograde cholangiopancreatography (ERCP) has
been supplanted to a certain extent by other, less invasive imaging methods such as computed
tomography, the addition of an electrosurgical wire to the standard cannula provided endoscopic
access to the biliary system and pancreas, which became the basis for a variety of more complicated
therapeutic maneuvers. There are now numerous examples of complex endoscopic treatment
methods that require sophisticated instruments, multitudes of ancillary equipment and accessories,
and specially trained personnel.
At the onset of the fiberoptic era, endoscopic procedures were performed in almost any convenient
place. Descriptions of these "procedure rooms" suggest that most amounted to little more than closets.
In the course of time, two closets were required in order to meet demand. Eventually, the need for
additional ancillary and support spaces became obvious, and this collection of rooms and space began
to take on a functionality of its own. In time, it became customary to refer to this place as the
endoscopy unit.
The endoscopy unit concept came into its own when specific areas within health care facilities were
reserved for the performance of endoscopic procedures. This event was a long time in coming for
many institutions, but this was fortunate in one respect: a better understanding of the necessary
facilities could be deduced from the inadequate nature and shortcomings of existing rooms and space.
Somewhat later, those who had participated in several cycles of unit development began to derive
principles of unit design and function that could be applied to the construction of any unit. Because of
size and cost, the development of new units began to require higher levels of thought and careful
planning. Many endoscopy units are now designed in conjunction with professional architects and with
careful attention to the unique nature and purpose of the proposed facility. Fortunately, there are now a
few excellent publications that are concerned with the construction and operation of endoscopy
units.16(164)
The percentage of procedures performed on an ambulatory basis has steadily increased since the
mid-1980s. This shift toward outpatient-based endoscopy has been a nearly universal trend in most
countries. Although therapeutic endoscopy has relatively low, albeit definite, morbidity and mortality
rates, many interventional procedures as well as diagnostic endoscopy in general can be performed
safely and expeditiously on an outpatient basis. In the United States, this trend has been motivated by
cost containment and efforts to avoid hospitalization whenever possible. Moreover, the great majority
of patients clearly regard the outpatient setting as preferable and more convenient. This emphasis
influences the design and operation of an endoscopy unit; more space must be available for waiting
rooms, dressing, preprocedure holding, and postprocedure recovery and additional personnel are
required.
The advent of the electronic endoscope has been another seminal event in the history of the
endoscopy unit. This occurred during the early part of the 1980s at a point when the fiberscope had
reached its apogee of technical sophistication. The introduction of this new technology had obvious
ramifications for clinical practice and teaching. Less evident at that time was the effect it would have on
the endoscopy unit. Few, if any, units were designed to accommodate this new instrument. It was
common, for example, to find television monitors in the most improbable and inefficient places, often
the only available location in an otherwise crowded room. Wires were strewn about the floor and ceiling
to such an extent that they became a physical danger for the unwary.
Although the physical adaptation of the endoscopy unit to the "new electronics" has been problematic,
of greater consequence for the unit has been the transition to the use of computers to manipulate and
store images that was triggered by the electronic endoscope.
Whether considering a large tertiary care center or a small community hospital, the two attributes of
growth in numbers of procedures and expanding complexity are common to virtually every endoscopy
unit. To date, these two elements have been the driving force in the development of the endoscopy
unit.
Current Status of the Endoscopy Unit

It would be gratifying to believe that the development of endoscopy units around the world has
proceeded in an orderly, logical fashion based on a thorough knowledge of well-defined general
principles of unit operations, and guided in every case by a thoughtful needs assessment.
Undoubtedly, this cannot be the case. However, it would be valuable to have some knowledge of the
actual status worldwide of the endoscopy unit. Given the tremendous numbers of procedures
performed and the importance of endoscopy in the diagnosis and treatment of digestive diseases, it is
important to inquire whether existing units approximate reasonable standards of patient care.
It is extremely difficult, perhaps impossible, to obtain an accurate assessment of the status of
endoscopy units throughout the world or even in a particular country. Seifert and Weismuller7(165)
stated in 1986 that a completely satisfactory GI endoscopy unit probably did not exist in the Federal
Republic of Germany. Based on a survey of 31 hospitals, this statement was remarkable, coming from
a country with high standards of endoscopic practice. The British Society of Gastroenterology has
published an excellent series of monographs that provide detailed analysis on the design, staffing, and
operation of endoscopy units.810(166)
The Olympus Optical Company (Tokyo, Japan) and KeyMed Ltd. (Southend-on-Sea, Essex, England)
in 1988 commissioned one of us (RM) to conduct an Ergonomics and Design Study that surveyed a
total of 41 endoscopic facilities in five countries.11(167) Included in this study were tertiary care
institutions (n = 17), general hospitals (n = 18), and private offices (n = 6). Among the facilities visited,
13 were located in the United States, 14 in Japan, 3 in Germany, 7 in the United Kingdom and 4 in
France.
The objective of the Ergonomics and Design Study was to gather information concerning the manner
whereby the work of an endoscopy unit is accomplished and the extent to which the infrastructure and
design of a unit either facilitate or interfere with the tasks to be accomplished. These data were
analyzed from several perspectives such as layout of procedure rooms, the overall design of the unit,
and the utilization of space and equipment. The survey was not intended to assess level of endoscopic
skill and technique, nor to evaluate the appropriateness of procedures, safety, diagnostic accuracy, or
effectiveness of endoscopic treatment. It did, however, assess the quality of patient care in terms of
the emotional and physical comforts, or the lack thereof, provided for the patient. Inasmuch as data
were collected by an individual with no medical background or training, the survey also provides a
unique impression of endoscopy and the endoscopy unit obtained, in essence, from the vantage of a
patient. Lastly, the survey was undertaken to assess future requirements and expectations for
endoscopy and the endoscopy unit.
During survey visits, data were collected by means of a questionnaire. When permitted, still
photographs and videotapes were made throughout the unit (videotaping was permitted in 26 units).
Photographs were assembled into storyboards for each facility. In addition to providing a simple
mechanism for understanding the layout of a unit and its procedure room(s), the storyboard also gave
a general impression of the decor and amenities as well as the types of equipment in use. Videotapes
were especially useful in providing an assessment of the way in which individual physicians and
gastrointestinal assistants (GIAs) performed their functions within the procedure room. In this fashion,
approximately 100 physicians were observed in the performance of both upper and lower GI
procedures. Diagrams were then drawn to illustrate the design of procedure rooms (Figure 41), types
of equipment present, placement of this equipment, and the range of activities performed by physicians
and GIAs within the room.
(168)Figure 41. Movement patterns that determine endoscopist and gastrointestinal assistant
(GIA) sectors within a procedure room with video endoscopy systems. The diagram is based
on a compilation of survey data from KeyMed Endoscopy Unit Ergonomics and Design Study.
Note the extreme variability in the positions of the light source, the videoprocessor, and the
television monitor.
Units were rated according to assessment parameters on a scale of effectiveness that ranged from
poor to fair, average, good, and very good. On this admittedly subjective scale, 26 units were rated as
below average or unacceptable, 9 were considered average/acceptable, and 14 were rated as above
average. It is of interest that there were few differences in quality of equipment between the highest
and the lowest rated units. Crowding and a basic disregard for the dignity of the patient were invariably
observed in the poorest rated units.
Although it is not possible or appropriate to present all of the data contained in the KeyMed survey,
some of which remain proprietary, certain salient points are of interest.
The equipment in use in the great majority of units was of high quality, relatively new, and in good
condition. Roentgenographic equipment (including C-arm machines) was utilized for endoscopic
procedures on a frequent and regular basis in 31 units. In 15 institutions, this equipment was available
within the endoscopy unit. It was necessary that endoscopic equipment be transported to the radiology
department in the remaining 16 institutions, and in many of these facilities usage was sufficient to
justify the presence of an x-ray machine within the endoscopy unit. Lasers were present in 13 units and
available within the institution in a further 2. Endoscopic ultrasound equipment was found in 15 units
surveyed. Fourteen units had a computer or computer system. This was rated as reasonably
sophisticated in 7 institutions.
At the time of the survey, many units were in various stages of transition from the exclusive use of
fiberoptic instruments to video endoscopy systems. A failure to recognize the special requirements for
incorporation of such systems was evident in the layout of many units. Video processors, television
monitors, and associated cabling were found at virtually any location within the procedure room (see
Figure 41).
Among the endoscopy units surveyed, 21 had special or separate areas designated for cleaning and
disinfection; 18 were equipped with automatic washing machines. In 5 units, washing machines were
placed within procedure rooms. In many cases, the areas set aside for endoscope/accessory cleaning
and disinfection were crude, badly designed, and poorly organized.
For patients undergoing procedures, storage lockers or cabinets for clothes and changing facilities
varied considerably from poor or nonexistent to very good. The design and decor of some but not all
units incorporated features and amenities that provided for the comfort of patients, both physical and
emotional, and preserved their privacy and dignity.
Comparisons of the individual units that composed the KeyMed survey would be inappropriate and
unnecessary. Furthermore, no single unit could be considered as perfect; each had flaws as well as
attractive features. However, a number of general conclusions can be drawn from the collective data.
The most immediate and obvious conclusion is the existence of an enormous range in the quality of
endoscopy units in terms of layout and efficiency. Some were clearly well organized and efficient;
others were characterized by chaos and disarray. Marked differences were found between as well as
within countries. Although it is often assumed that limited space results in poor design, large size was
not necessarily synonymous with a satisfactory arrangement of rooms and spaces (layout). In fact, the
size of many units was comparatively small, but the layout was relatively good, whereas in some of the
large units, the layout was very poor and could be attributed only to bad planning.
In the planning and construction of the units surveyed, it was generally apparent that relatively little
attention was given to cleaning and disinfection functions, including handling of soiled instruments.
This aspect of the endoscopy unit stood out clearly as the area most in need of immediate
improvements.
The comfort and dignity of patients were not a high priority in the design of many units. In this regard,
there was again a wide range in ratings for the units surveyed. Data assembled in the survey identified
this as another major area in need of substantial improvement.
Although difficult to measure, there was a definite correlation between good design, emphasis placed
on nursing skill, and attention to the comfort of patients. Conversely, when nursing skill was at a low
level, patients were treated poorly, as if objects, and the entire operation of the unit reflected this
unacceptable treatment. Units that exhibited high levels of nursing skill were generally rated high in
terms of the attention given to patient comfort. When these elements were present, the unit was
usually extremely well organized with efficient layout and use of available space. Invariably, these
achievements were the result of careful planning.
Endoscopy Unit Function

A large endoscopy unit (25 to 35 or more procedures per day) is a complicated place in which a great
number of different procedures are performed. It is no small task to organize the use of available
space, numerous pieces of equipment, and the activities of various staff members to accomplish a
heavy schedule of procedures with a minimum of confusion. By the same token, the operation of such
a unit becomes difficult to comprehend as the numbers and complexity of procedures increase.
Planning for future needs and growth becomes even more perplexing. Given the seeming complexity,
problem solving as well as establishing and achieving goals can be an arduous process that
sometimes amounts to a series of educated guesses. From the standpoint of daily operations, problem
solving, and future goals, it is therefore essential to clearly understand the way in which a unit
functions. To this end, certain management tools can be used. These are actually abstract concepts,
but they provide insight into function as well as a workable apparatus for analysis and problem solving.
These devices are the procedure unit and the weighted scale.
The Procedure Unit

Most endoscopists regard an endoscopic procedure only in terms of the actual use of an endoscope in
a patient. From an operational standpoint, a procedure is more complicated. Nevertheless, any
endoscopic procedure can be reduced to a series of basic steps (Table 41) that can be
subcategorized into three components: preprocedural, procedural, and postprocedural. The basic
procedure components taken together are termed a procedure unit. From a managerial viewpoint, the
business of an endoscopy unit is to produce endoscopy procedure units.
TABLE 41
Procedure Unit
PREPROCEDURE
Scheduling
Patient check-in
Patient instruction, interview
Patient preparation
Premedication
Room/equipment preparation
PROCEDURE
Examination/procedure
(Handling of biopsy and other specimens)
POSTPROCEDURE
Patient recovery
Monitoring
Postprocedure instructions/scheduling
Discharge
Room/equipment cleaning, turn around
Charting/report generation*
Written chart notation
Report dictation
Typing report
Review/signature of typed report
Processing of report to chart
File copy of report
Data processing
Billing/collections
* Non-computer-based system.
Variable activity. May not be performed in some
endoscopy units.
The actual endoscopic (procedural) component of the procedure unit is but one part of a relatively
complex process. The length of time required for each component varies according to the type of
endoscopic procedure, although the endoscopy itself may require only a short length of time relative to
that needed for the other steps. Diagnostic esophagogastroduodenoscopy (EGD), for example,
requires only 5 to 10 min. But the related components of scheduling, preparing the patient for the
procedure, preparing the procedure room and equipment, administering sedative drugs,
cleaning/disinfecting, and report generating require the same amounts of time as those for a
colonoscopy, even though the time required to perform a colonoscopy versus that for an EGD may be
longer by a factor of three or four.
Weighted Scale
Although all procedure units are identical, actual endoscopic procedures differ in many respects. From
the managerial point of view, the fundamental variables are the time required, number of personnel,
and items of equipment. The procedure unit concept is useless unless it is adjusted for these factors.
It is impossible to assess the true amount of work performed in an endoscopy unit simply by counting
the number of procedures performed. This "raw count" does not take into consideration complicated
procedures that require more time, equipment, and personnel. One way to deal with this variability is to
develop a weighted scale for procedures, essentially an equivalent value system that adjusts total
procedure count for degree of complexity. The number assigned in the scale to each specific type of
procedure relates the complexity of that procedure to an arbitrarily selected reference procedure. The
best reference procedure for a weighted scale is diagnostic EGD (without biopsy or other ancillary
procedures) because this procedure should be the least time consuming and least complex in terms of
personnel and equipment required. Diagnostic EGD may be assigned a value of 1.0 in a weighted
scale. Giving the procedure unit an arbitrary reference value in the weighted scale facilitates
manipulation and analysis of data. One procedure unit therefore becomes equivalent to a simple
diagnostic EGD.
The next step is to determine an equivalent time value for the unit used in the weighted scale. For
example, a weighted scale value of 1.0 may be given a time value of 30 min. The time value assigned
must take into account all of the steps in the procedure unit. Although it may require only 10 min to
perform the endoscopic component of a diagnostic EGD, a large number of other tasks must be
accomplished before the procedure room can be "turned around" for the next examination. A
suggested equivalent value table for a hypothetical endoscopy unit is shown in Table 42.
TABLE 42
Suggested Weighted Scale of Endoscopic
Procedures
EGD
1.00
EGD with heat probe or Bicap probe
2.00
Other procedures with EGD
1.50*
Sclerotherapy
1.75
Variceal banding
1.75
Esophageal dilation (bougie)
0.50
Colonoscopy
2.00
Limited colonoscopy
1.50
Polypectomy
(add)
0.25
Multiple polypectomies (710 polyps)
3.00
Colonoscopy with sequential biopsy for dysplasia
2.50
Colonoscopy with heat probe or Bicap probe
3.00
Laparoscopy
3.00
ERCP
3.00
Sphincterotomy
4.00
ERCP, spincterotomy, stent placement
4.50
Sphincter of Oddi manometry
1.50
Laser, all types
2.75
PEG
2.00
EUS
2.00
TABLE 42
Suggested Weighted Scale of Endoscopic
Procedures
EUS with FNA cytology
2.50
Any major complication, any procedure (e.g.,
(add)
3.00
respiratory arrest, hemorrhage)
Any procedure done on hospital floor (except laser
(add)
0.50
and ERCP)
EGDesophagogastroduodenoscopy; ERCPendoscopic retrograde
cholangiopancreatography;PEGpercutaneous endoscopic gastrostomy;
EUSendoscopic ultrasonography; FNAfine-needle aspiration.
* For example, polypectomy, dilation over guidewire, insertion of endoprosthesis.
Each endoscopy unit must develop its own weighted scale for procedures if this device is to be utilized
effectively. This is because there are unique aspects in the way each unit functions that cannot be
accounted for in a universal scale. For example, it may be necessary in some units that lack x-ray
equipment to perform ERCP procedures in a radiology department. In this case, transport time and
additional time to place necessary equipment on carts must be calculated into the procedure
equivalent value for ERCP. In units that provide instruction to trainees, equivalent values should also
reflect the fact that teaching is almost always inefficient. As a rule, supervised instruction increases the
time required for a procedure by a factor of about one third, and this added time must be factored into
the scale. In a nonteaching institution, the time equivalent for a weighted scale value of 1.0 might be
reduced to 25 or 20 min.
The procedure unit/weighted scale methodology has many uses with respect to analysis, planning, and
problem solving. It can be used to accurately gauge the work output and performance of a unit. For
example, the average weighted value for all procedures performed during specific periods of time can
be calculated. An average value that increases over the course of time indicates increasing complexity
of procedures and workload. This may occur with or without a change in the total number of
procedures performed. If little or no increase in the total count of procedures is recorded, the overall
output of work has nevertheless increased if the average weighted value for procedures has increased.
Also, an estimate of total work is obtained by multiplying the average weighted value by the total
number of procedures performed. In effect, this converts the work output to an equivalent number of
diagnostic EGDs.
The weighted scale methodology can also be used to obtain an actual measurement of the time and
effort required of the GIAs within a unit. To do so, the GIAs record a weighted value for each procedure
performed according to the guidelines in Table 42. Adjustments are permitted for especially lengthy
and difficult procedures or those that require an unusual number of items of equipment. For any given
period of time, a total of the weighted values becomes a fairly accurate representation of the actual
output of work. Dividing this total by the number of procedures performed provides a precise measure
of the acuity of procedures performed within the unit.
The weighted scale method of analysis of procedure activity can be used in planning for acquisition of
space, equipment, and staff. The theoretical maximum capacity of a given endoscopy unit can be
assessed in terms of procedure units once the average weighted value and the time value for one
procedure unit are established. This calculation would be as follows:
where N is the number of procedure rooms available, Hrs is the number of working hours per day,
Days refers to the number of working days per year, T is the time value for 1.0 procedure unit (or a
weighted scale value of 1.0), Avg is the average procedure equivalent value, and Eff is the expected
efficiency of the unit. The last value, expected efficiency, is somewhat arbitrary, but if this is not
included, the calculation assumes that every room in the unit operates at 100% efficiency, something
that is not attainable. A more reasonable efficiency factor would be about 70%. It is also easy to
calculate the actual efficiency of a unit by comparing the weighted procedure count with the theoretical
maximum of 100% efficiency.
The weighted scale methodology can be used to assess the level of GIA staffing needed in the unit.
Although an increase (or decrease) in numbers of procedures is easy to demonstrate, this "raw count"
alone might not be enough to justify additional personnel. However, in addition to increasing numbers
of procedures, there may also be a corresponding increase in procedure complexity that is not
reflected in the count. In essence, the workload of the GIAs may be increasing to a much greater
extent than a simple count of procedures would suggest.
The Endoscopy Unit

Acquisition of Space
Until recently, few endoscopy units were specifically designed for GI endoscopy. Rather, portions of a
hospital or clinic were converted for this purpose. To a certain extent, this means that function must fit
form; that is, the design of the unit must conform to the physical dimensions of the space allocated.
Fitting the basic components of the unit into available space frequently leads to a less than ideal design
with respect to the flow of activities within the unit. All too often, sufficient time is not allocated for
planning and problem solving at the outset, so that end users are frequently disappointed with the final
result.
Another liability inherent in the renovation of existing space is a lack of flexibility. The space available is
usually subdivided according to specific functions, and rooms are constructed for highly specific
purposes. However, it often becomes necessary to change the function of a specific room or rooms as
the demand for endoscopic services changes and new procedures and newly developed equipment
are added. Shifts in emphasis will be possible to the degree that foresight is exercised in designing
adaptability into the unit. It is therefore critical that the planners or design team allow sufficient time to
research current trends and future requirements before final acceptance of the proposed scheme.
An endoscopy unit is usually developed in response to problems that have arisen in meeting an
increasing demand for services. The unswerving desire of administrators to economize wherever
possible often leads to construction of a unit that is adequate for current needs, but poorly positioned
for future growth. Existing problems are solved to a large extent, but there is little anticipation of future
needs and potential problems. This leads to an obsolescence cycle in which a series of units will have
been constructed that meet the needs of the moment or at best the immediate future. This will prove to
be a false economy, as the subsequent efforts to accommodate future growth and change invariably
consume resources over and above those that would have been needed at the outset. Unfortunately,
the capacity for growth in numbers and complexity of procedures is often seen as a distinct luxury, and
built-in provisions for growth are seldom encountered. Rather, it is more usual that growth is
compromised in a fashion that is, based on past performance, predictable.
Principles of Design
All endoscopic procedures are alike in the sense that the same basic steps are required to complete
one procedure unit (see Table 41). From this standpoint, endoscopic procedures can be considered
uniform and interchangeable when designing an endoscopy unit. From another perspective, they are
dissimilar. In relation to the design and function of the unit, the differences in procedures relate to
requirements for diverse items of equipment and variation in procedure room size.
The Procedure Room
At least two persons are required for every endoscopic procedure (in addition to a patient): an
endoscopist and a GIA. The procedure table should be considered the central focus of the room. It so
happens that for most procedures, the GIA and the endoscopist take different positions in relation to
the table as they perform their respective tasks. The basic rule of specific places within the room is
thus established by the way in which endoscopic procedures are performed.
It is necessary to have access to the procedure table from at least three and preferably four sides, so
that it is almost always placed at or near the center of the room. The next logical step conceptually is to
divide the room into sectors or work zones for the GIA and the endoscopist, based on their relationship
to the procedure table and the tasks they perform. Although the location of the endoscopist sector is
relatively fixed, GIA sectors vary with the type of procedure being performed. A well-designed
procedure room takes greatest advantage of this simple but valuable principle of dividing the room into
three basic sectors according to the position of the patient (table) and the activities of the GIA and the
endoscopist.
The sectors within the procedure room are interactive. Considerable foresight and planning are
therefore needed before the layout of a room is finalized. In some cases, the size and location of the
sectors for the endoscopist and GIA may be dependent on the size and shape of the room where these
parameters have been defined by other considerations, as for example the need to construct a unit
within an existing structure.
The relatively fixed work sector for the endoscopist is smaller in terms of actual space than that for the
GIA. Obviously, this sector includes the position taken by the endoscopist for performance of the
procedure. The area immediately surrounding the endoscopist within his or her sector should provide
storage for commonplace procedural items such as lubricating jelly, gloves, and gauze pads. A second
satellite sector may be provided within the procedure room for the endoscopist for charting, dictating,
reviewing x-rays, accessing computer data, reviewing other types of records, and making telephone
calls. In larger units, a separate area outside the procedure room but in close proximity may be
provided for these functions.
The organizational routine of the GIA requires the most careful analysis in order that a procedure room
layout be designed that allows the assistant to work with maximum effectiveness. The GIA work sector
is far more complex than that of the endoscopist. It has three functional parts. The first is a work
surface immediately within the primary work sector of the GIA. Generally placed near the head of the
patient, this surface is designated for storage of endoscopic accessories, paraphernalia for handling
tissue specimens, gloves, and any other items that are likely to be needed during the procedure. The
second is a satellite area, close at hand, that includes storage space and an additional work surface.
Accessories, all of the drugs that are likely to be required, and other equipment that is less likely to be
needed during the procedure can be kept in this area. The third component to the GIA work sector is
designated for storage. It usually consists of cupboards and drawers as well as an extra work surface
when possible.
The procedure room as well as the endoscopist and GIA will come into contact with a wide array of
body fluids and tissues during endoscopic procedures. It is essential, therefore, that the distribution of
potentially infectious tissues and fluids within the procedure room be controlled as rigidly as possible.
For example, tissue samples should be handled in an area that is separate from that designated for
sedative and other drugs that may be administered to the patient. Endoscopes and accessories used
in a procedure should never be placed in an area designated as clean. There should be adherence to
this principle of strict separation of clean and dirty areas throughout the unit. A sink basin should be
provided in each procedure room for hand washing.
Two variables that must be considered in the design of a procedure room are size and shape. Opinion
varies somewhat as to an acceptable minimum size for a procedure room. Our recommendations are:
172 ft2 (16 m2) (RM); 192 ft2 (17.8 m2) (MER); 300 ft2 (27.9 m2) (MVS). These recommendations
refer to gross as opposed to net size, which takes into account installed cabinets. They pertain to the
floor space required for a basic procedure room. The recommendations at the smaller end of the range
are satisfactory for the basic standard procedures such as EGD and colonoscopy, whereas the larger
recommendation provides contingency space for additional items of equipment such as a laser or an
endoscopic ultrasonography (EUS) unit. Rooms designated for these special procedures should have
a minimum of 200 net ft2. Larger rooms tend to be more adaptable and can be used for more than one
special procedure or for the development of new and highly technical procedures. Furthermore, rooms
of larger dimension are more suitable in teaching units where there are likely to be additional personnel
within the procedure room.
The shape of the room is another consideration. Unfortunately, this is sometimes dictated by the
overall dimensions of the building, placement of stairwells, elevators, plumbing, ducting, and a host of
other factors. When possible, however, a somewhat re 20 ft (4.5 6 m). In a room of rectangular
proportions, the work surfaces, cupboards, and storage facilities should be within convenient and
comfortable reach for the endoscopist and GIA. Diagrams of several hypothetical procedure rooms
based on these concepts are shown in Figure 42.
(169)Figure 42. Schematic diagrams of hypothetical procedure rooms. A and B, Rooms

designed for routine procedures such as esophagogastroduodenoscopy (EGD) and
colonoscopy. C and D, Rooms designed for endosonography and laser treatment. E, Procedure
room design with C-arm fluoroscope suitable for endoscopic retrograde
cholangiopancreatography (ERCP) and other procedures that require fluoroscopy.
Ooxygen; Vvacuum (suction); Aair (medical).
Division of Floor Space
The physical layout of an endoscopy unit must accommodate the basic steps in the procedure unit
(see Table 41). Space must be allotted to each component whether the endoscopy unit consists of a
single procedure room or has multiple rooms. Thus, any floor plan must subdivide available space
according to the fundamental functions listed in Table 43. In assigning available space(s) to specific
function(s), some areas may serve dual purposes. Areas can be set aside in the procedure room for
instrument storage and charting; scheduling, secretarial, and reception functions can be located in a
single area.
Basic Functions That

Require Allocation of Space in the Endoscopy
Unit
TABLE 43
Scheduling
Patient reception
Patient interview room(s)
Patient dressing/recovery
Procedure room(s)
Cleaning
Storage
Charting/dictation
Secretarial
With maximum consolidation, the simplest possible endoscopy unit must consist of five room groups or
areas designated as follows: (1) patient reception, (2) scheduling/secretarial, (3) procedure room (with
areas for instrument storage and charting/report generation), (4) cleaning and disinfection of
instruments, and (5) patient dressing/recovery room or area. At the opposite extreme is the tertiary
center unit that is substantially larger than the simplest model, there being much greater floor space
and many more rooms. Although available space is much greater in such a unit, it is nevertheless
allocated according to the same basic functional plan (see Table 43). In a unit in a sizable referral
center, separate, larger areas, even actual rooms, are provided for the basic functions listed
previously.
Independent and Interdependent Procedure Rooms
The degree to which the procedure rooms of a large unit will function independently of one another is
an important design decision.
In the simplest unit model (or single-room unit), the procedure room can serve several functions. To
the maximum possible extent, the elements of the procedure unit (see Table 41) are contained and
accomplished within this room. With this approach, the procedure room functions in a
semiautonomous fashion that is relatively independent of the other procedure rooms and support
areas. Even in a larger unit, this design can be preserved by placing as many elements of the
procedure unit as possible in the procedure room. For the large unit, however, such a system is usually
less efficient and is likely to increase the operational cost of running the unit.
The opposite approach is to accomplish as many procedure unit elements (see Table 41) away from
the procedure room as possible. The only activity performed in the procedure room is therefore the
endoscopic procedure itself. Because the procedure room represents a relatively large fraction of the
cost of development of a unit, it is usually more cost effective to restrict its use to the performance of
procedures. Other areas and rooms in the unit are provided and designated for other components of
the procedure unit. A single room or area for cleaning/disinfection and one for storage could serve
many or all procedure rooms. Centralized patterns for cleaning/disinfection and for storage versus a
more dispersed scheme with more than one area or room for cleaning/disinfection and for storage
(e.g., between two procedure rooms) is one of the main differences in design among units.
Centralization of storage and cleaning/disinfection functions provides flexibility in the use of procedure
rooms because the same equipment can be used in more than one room. Administration of sedative
drugs can also be accomplished outside the procedure room with the use of a cart exchange system.
In a completely interdependent system, the sole purpose of the procedure room will be the procedure
itself.
A unit built on the interdependent design concept is inherently more complex but potentially more
efficient from an operational standpoint. For example, procedure scheduling becomes more intricate
because care must be taken to avoid simultaneous procedures that require the same item of
equipment. The efficient operation of such a unit demands tight organization and teamwork on the part
of all personnel. The arrangement of the physical counterparts of the various procedure unit elements
also has a bearing on staffing. For example, with centralized cleaning/disinfection and storage, it may
be efficient to assign a GIA to this task alone, perhaps on a daily rotating basis.
Although the interdependent design concept offers the greatest prospect for efficiency and cost
effectiveness, extremely careful planning is required to realize this potential. By the same token, a
single design error or miscalculation becomes magnified throughout the entire unit and any gain in
efficiency is thereby eliminated. In particular, the size and functional capacity of support areas, such as
recovery and cleaning/disinfection, must be adequate to handle the maximum output from procedure
rooms. Otherwise, these ancillary and support functions can adversely affect efficiency and thereby
limit the overall productivity of the unit.
Multifunctional Versus Dedicated Procedure Rooms
An endoscopy unit takes on its unique character as various purposes are assigned to procedure
rooms. The balance between multifunctional versus dedicated procedure rooms is another of the major
differences between units.
In the simplest model (one procedure room), all pro-cedures in the repertoire are necessarily
performed in one room. As the number of procedure rooms increases, an independent room
philosophy can be maintained; however, if carried to its logical conclusion, this results in a group of
extraordinarily well-equipped, functionally autonomous procedure rooms, any one of which could serve
alone as a respectable endoscopy unit. This is, of course, exaggerated, and in practice, such a unit
would be highly unusual. In fact, as procedure rooms become more numerous, they are diversified
according to various functions, each room being designed, equipped, and arranged with an emphasis
on one or a few types of procedures.
The way in which different procedures are diversified among available procedure rooms is another
major difference between endoscopy units. To a large extent, this reflects the pattern of procedures
performed in the unit, which in turn depends on referral patterns, the demand for particular procedures,
and the interests and expertise of the professional staff. Thus, in one unit, an emphasis on laser
procedures might require a room designated for this purpose. A center with a large referral base for
ERCP would logically require a separate room for this purpose. It is essential to bear in mind that none
of the factors that influence the pattern of the types of procedures performed in a unit is static. Rather,
the pattern always changes over the course of time, especially as new techniques are developed. It is
probably safe to predict, for example, that the demand for colonoscopy procedures will continue to
accelerate in coming years. At the same time, the demand for laser procedures has declined, although
this may change again with the further development of photodynamic therapy.
A tradeoff necessarily occurs when a room is designated for a precise function. When altered for a
specific purpose by the way it is scheduled and equipped, perhaps even with respect to physical
changes in layout, the room becomes less suitable for other procedures. Essentially, the unit as a
whole becomes less flexible and less adaptable to changes in the pattern of procedures. A unit in
which rules for room utilization are relatively inflexible tends to be inefficient. Certain functions such as
scheduling are simplified, but procedure rooms may stand idle. If demand for endoscopic services is
high, this may translate to limitations on access, delays in scheduling procedures, dissatisfied patients,
and disgruntled referring physicians. At the other extreme is the concept that there should be the
capability to perform virtually any procedure in any room. The difficulty here is frequently the need to
move equipment from room to room. Scheduling actually may become more complex, and there is a
greater tendency toward disorganization in the management of the daily activities unless each room is
so well equipped that procedures can be scheduled ad lib in any room without concern for availability of
GIAs and equipment. The problem with this approach is the high cost for duplicate staff and
equipment.
Special and Routine Procedures
One method of establishing a balance in a unit between dedicated and multifunctional procedure
rooms is to divide procedures into "routine" and "special" types. Routine procedures have the following
characteristics:
1.
2.
3.
4.
5.
Usually but not always high volume

Performed by most endoscopists working in the unit
Usually diagnostic
Require fairly standard, relatively unsophisticated equipment that is readily available in the unit
Significant day to day (even hour to hour) variation in demand
Special procedures tend to have the following attributes:

1. Performed by a relatively small number of endoscopists
2. Usually require special, often expensive equipment that is frequently a one-of-a-kind item within
the unit
3. Lower volume relative to that for routine procedures
4. Frequently more time consuming
5. Often require more than one GIA
It can be useful to divide routine procedures into high and low volume. The definition of a high-volume
procedure is arbitrary, but this label generally applies to any procedure that constitutes 20% or more of
the total volume of procedures.
The division of procedures into routine and special categories is unique in each unit. Although EGD
and colonoscopy would typically be categorized as routine in most units, large units might include
ERCP and even EUS. This depends on the pattern and numbers of procedures performed and on the
availability of equipment. It is unusual to find the capability to perform ERCP in more than one room in
any given unit, so that by default ERCP becomes a special procedure with a designated room. In the
unit at University Hospitals of Cleveland, ERCP (diagnostic and therapeutic), laser endoscopy, EUS,
certain dilation procedures, and esophageal prosthesis placement are special, whereas all other
procedures are routine and can be performed in any room. EGD, colonoscopy, polypectomy, some
types of dilation, and most methods for control of bleeding including sclerotherapy and variceal banding
can be performed in any room.
Achieving Balance
As a general rule, specific rooms should be designated for special category procedures. Usually this
follows without much deliberation because the availability of special equipment is limited. However, as
far as possible, routine procedures should not be restricted to specific procedure rooms. For maximum
efficiency, it should be feasible to conduct routine procedures in any room, even rooms designated for
special procedures. A large referral center unit should have adequate equipment to carry out routine,
high-volume procedures simultaneously in most or all procedure rooms, including those designated for
special purposes.
Because by definition certain routine procedures are high volume, it might seem logical to assign
specific rooms for such procedures while excluding other activities from these rooms. Depending on
circumstances, this may be efficient or restrictive. If the number of routine procedures of one type,
EGD, for example, is sufficient to reasonably expect that a room will be fully utilized in the performance
of this one procedure, then it may be efficient to designate one or more procedure rooms for this
purpose alone. However, another characteristic of the routine procedure is that the demand for this
type of service varies a great deal on a day-to-day basis (even hour to hour in a highly active unit).
Over longer periods of time, there will be changes in the numbers and patterns of procedures
performed.
In deciding the mix of multifunctional and dedicated rooms within a unit, it is important to analyze
efficiency and the effect that limiting room usage has on the scheduling process.
A retrospective approach to assessing efficiency is to determine the average number of hours per day
that a unifunctional, designated room is not in use in relation to the desired overall efficiency of room
utilization in the unit. If, for example, this is greater than about 30%, then the use of the room for a
single specific function may be inefficient; that is, the procedure activity in question does not justify
restrictions on the use of a room. Conversely, utilization of the room at greater than targeted efficiency
may mean that demand exceeds capacity. If, at the same time, other rooms designated specifically for
other procedures are not in use more than about 30% of the time, the unit overall is not functioning at
maximum efficiency.
There is a further element in balancing dedicated versus multifunctional room usage that is potentially
more damaging to the orderly function of the endoscopy unit. Rigid adherence to specific room
designations may restrict patient access. One way to determine this is to track overall scheduling
activity of the unit with special attention to schedule lag. This is the length of time required to complete
a procedure unit beginning with the moment of contact between the endoscopy unit and the patient or
the referring physician. It is important to consider all types of procedures performed in the unit because
the time required for completion of one type may be within expectations, whereas that for another may
be unsatisfactory. The many possible reasons for such an imbalance include limited staff availability.
However, in some cases, it may be due entirely to or at least compounded by an inflexible adherence
to a policy of dedicated room usage.
The uncontemplated "trigger" reaction to apparently excessive demand for a routine procedure is to
acquire another designated room. A more discerning approach is to first determine whether the way in
which the unit functions is the problem. This emphasizes the basic importance of trying to think about
the endoscopy unit from an abstract conceptual point of view rather than as a physical cluster of
rooms.
The blend of unifunctional, special procedure rooms and multifunctional, routine procedure rooms
should never become absolute and unchangeable. If past experience is any guide, it can be assumed
that the pattern, numbers, complexity, and general mix of procedures will continue to change.
Therefore, it is important that a unit be constructed in such a way that the function of any room can be
altered when necessary. Designing a "few small rooms" for EGD, for example, is very imprudent. A unit
cannot be set up today that will serve all the foreseeable and unforeseeable needs of the next decade.
The only answer to this problem is flexibility in design.
Ultra-High-Volume Procedures
Certain procedures may be performed in such numbers that the sheer volume is overwhelming to the
point that the functionality and efficiency of the unit are compromised. It usually becomes necessary to
consider such ultra-high-volume procedures as separate categories in relation to other endoscopic
procedures. The only procedure that potentially qualifies in this respect at present is screening flexible
sigmoidoscopy.
Flexible sigmoidoscopy does not fit the procedure unit concept; that is, it differs significantly from other
endoscopic procedures with respect to scheduling, patient preparation and recovery, and the amount
of time required. These differences in relation to other procedures would not be so important were it
not for the high volume of procedures. In great numbers, the special characteristics of an
ultra-high-volume procedure tend to dominate the function of the unit. It would seem to be not only
efficient but also responsive to patient needs to design a specific room or rooms in the endoscopy unit
for this purpose. It may also be sensible to divorce the truly ultra-high-volume procedure physically and
functionally from the rest of the endoscopy unit.
The dissociation of flexible sigmoidoscopy from the rest of the endoscopy unit need not be absolute.
For example, scheduling might be done in the same place and by the same personnel for this and
other procedures. Staff can rotate from one unit to the other. However, at any point in time, the function
of the flexible sigmoidoscopy unit should be independent of the endoscopy unit. Patient flow patterns
should not intersect nor should patients use the same dressing facilities. To some degree, toilet
facilities might be shared in a well-designed unit, but in this case, care must be taken that the
requirements for flexible sigmoidoscopy preparation do not overwhelm the capacity to prepare patients
for colonoscopy.
Ultra-high-volume could become problematic with respect to EGD in relation to the growing utilization
of this examination as a primary diagnostic procedure in place of the upper GI x-ray series. From a
conceptual and operational point of view, it may be reasonable to separate facilities for primary upper
GI diagnostic endoscopy from the other procedure activities. Such a separation can be physical or
functional to greater or lesser degrees. Another possible concept is to separate special (frequently
therapeutic) procedures from routine high-volume diagnostic procedures. Thus, laser, ERCP (with a
high volume of therapeutic endoscopy), EUS, endoscopic methods of hemostasis, and other
therapeutic procedures might be performed in one area of the unit, while EGD and colonoscopy are
performed in another. Operationally, these two areas would work at different paces and require
different methods of scheduling, different types of equipment, and different levels of staffing.
The Multidisciplinary Unit
It is possible to dispense with specific GI endoscopy scheduling/secretarial, reception, and patient
dressing/ recovery areas in a multidisciplinary unit in which a variety of other nongastrointestinal
diagnostic and therapeutic procedures are performed. This is not germane here, since a GI endoscopy
unit as an entity does not exist in this type of unit. However, this plan has the advantage of cost
containment in that some equipment, staff, and spaces are shared among disciplines, but it can be
extremely restrictive in terms of growth, in both numbers and complexity of procedures. Furthermore,
the similarity between different types of endoscopy is superficial with regard to the instruments required
and to the actual procedures, so that cross training of assistants for more than one discipline has its
limits. Shared units frequently lead to scheduling conflicts and apportionment of time and room
availability to the various specialties that use the unit, a practice that in effect means that each
specialty has its own unit but that it is operated at something less than full time.
Certain expensive pieces of equipment, a C-arm fluoroscope or neodymium:yttrium-aluminum-garnet
laser, for example, might be shared between two or more specialties if the volume of cases for all
users is small and the demand for use of a machine is expected to remain static. However, when two
or more departments must combine their patient activity to justify the acquisition of an expensive
device, a laser, for example, there is reason to ask whether this represents an improvement in patient
care, because the number of procedures in any category of usage is by definition small. Nevertheless,
cost is a significant factor in the function of a unit. Except where the volume of procedures is very high,
ERCP is usually performed away from the endoscopy unit in a radiology suite because of the cost of
x-ray equipment. In general, however, the efficiency of an endoscopy unit decreases in proportion to
the number of procedures performed away from the unit.
Aesthetics
Another design principle, often overlooked and sometimes even ignored, is that the endoscopy unit
exists first and foremost for the care of patients. Too often a unit will have been designed to suit the
needs of physicians or nurses with little attention to the physical and emotional well-being of the
patients. A unit may be a model of factory-like efficiency, while patients find it detestable. Although it is
difficult for patients to accurately assess medical expertise, they nonetheless make judgments,
consciously and unconsciously, about the quality of care they receive. These are based on such purely
human considerations as cleanliness, attractiveness of surroundings, privacy, courtesy, respect,
efficiency, availability, and physical comforts.
The best endoscopy units are based on a total design approach, that is, one that accounts for the
well-being of the patient, GIA, and endoscopist. In the planning process, the GIA and endoscopist are
usually well represented, whereas the patient may not have an advocate. As the process proceeds, it is
therefore advisable to take time to look at the developing plan from the viewpoint of the patient. From
this perspective, two factors are especially important: preservation of privacy and dignity and the
aesthetics of the unit.
When a unit is aesthetically pleasing, patients feel more comfortable and relaxed. This is beneficial for
the patient about to undergo a procedure, but it also facilitates the tasks to be performed by the GIA
and endoscopist. Furthermore, an attractive decor also has a positive effect on all who work in the unit
in terms of morale, motivation, teamwork, and general performance.
Location
The location of an endoscopy unit in the hospital setting is not inconsequential. Although options as to
site are often not available (see "Acquisition of Space," earlier in this chapter), placement of the unit in
relation to other hospital departments has a direct bearing on efficiency. The hospital-based endoscopy
unit has a natural relation with many other areas of a hospital, including the emergency room, radiology
department, and intensive care units, because it is frequently necessary to perform endoscopic
procedures in these areas. At the same time, the unit must be convenient to hospital wards. When
certain beds are reserved for specific types of disease or for specific hospital departments, placement
of the endoscopy unit near wards set aside for patients with GI diseases is advantageous.
Because more and more endoscopic procedures are being performed on an outpatient basis,
convenient outpatient access becomes a necessity. In hospital units built at a time when there was less
emphasis on outpatient procedures, patients must endure a series of signs, maps, and conflicting
directions that begin at the hospital front door, a symbolic portal that is not necessarily "patient
friendly."
A location for a unit that is desirable in every way is an impossibility that would require that the hospital
be built around the unit. Fortunately, close spatial relationships within the hospital have relative
degrees of importance that are determined largely by practice methods. As a general rule, however,
proximity to the radiology suite and easy outpatient access should be given precedence in determining
unit location. In addition to the fact that ERCP and related procedures may be performed in the
radiology suite, many radiologic and endoscopic procedures are complementary. This facilitates patient
evaluation in that patients may undergo a series of procedures in each of the two units without
returning to a hospital ward between diagnostic studies. Proximity also facilitates a cooperative
professional relationship between radiologist and endoscopist that enhances patient care.
Number of Procedure Rooms

The number of procedure rooms that will be required is a very basic consideration in planning an
endoscopy unit. It is a simple matter to calculate the needs of the moment, but the future, be it further
growth in numbers and complexity or contraction of growth, is more difficult to predict. Past experience
helps to some extent, and in this respect records of unit activity are extremely helpful. But there are no
formulas for future activity and forecasts are at best educated guesses. Some variables in this
calculation are common to all units, such as the socioeconomic factors and other trends described
later. However, each unit faces unique questions peculiar to its environment. Such an analysis can be
far-reaching and can include factors as diverse as the quality of competing organizations and the
demographic characteristics of the population that the unit serves. In the final analysis, however,
growth is strongly related to the quality of the unit. Is it innovative in a critical way? Is it responsive to
the needs of patients and referring physicians? Is the proficiency of its staff high? Does it function
efficiently?
The theoretical maximum capacity of an endoscopy room can be calculated in procedure units using
the weighted scale. It is first necessary to calculate the number of working days per year
(approximately 256 days based on a 5-day work week and five or six holidays per year) and the
working hours of the unit (usually 7 hr, with time being permitted for lunch breaks for GIAs). The time
corresponding to a 1.0 procedural equivalent is a relative value that must be estimated for each
individual unit (1.0 equals about 30 min in our unit). Thus, the theoretical maximum number of
procedure units (diagnostic EGD equivalents) that can be performed in one room during 1 year is 3584
(256 2.0 7). However, this is a theoretical value that assumes that all procedures performed are
diagnostic EGD. To arrive at a more realistic number, it is necessary to allow for the average acuity
(degree of difficulty) of the procedures performed. For example, if the average weighted acuity value of
all procedures is 1.53, the theoretical maximum number of procedures that a single room can
accommodate is 2342 (3584 1.53). This, however, assumes 100% efficiency in the use of the room,
a level that is not possible. A more attainable efficiency for use of a room would be about 60 to 70%, in
which case the capacity of one procedure room is about 1500 procedure units per year.
Support Rooms
Although the procedure room(s) is the nucleus of every endoscopy unit, many other kinds of rooms or
designated areas are also necessary. A relatively complete list of possibilities is given in Table 44. It
is doubtful that the majority of units are so complete. However, such a complement of rooms is a
reasonable goal for a large tertiary referral unit.
Endoscopy Unit Rooms

(Full Complement)
TABLE 44
Reception
Procedure room(s)
Appointment desk
Dressing room
Toilet facilities (women/men)
Interview room(s)
Endoscopy Unit Rooms

(Full Complement)
TABLE 44
Offices
Director of unit
Medical staff
Head nurse
Secretaries
Word processing
Computer facilities
Records
Staff changing, toilet facilities, locker room
Nurse base area, utility room
Conference roomstaff, students
Library
Storage room(s)
Dressing Rooms
The term dressing room is preferred for the preprocedure holding and postprocedure recovery area for
patients, although in some units this is termed the recovery room. In most units, a single dressing room
serves all procedure rooms. This arrangement is most common because it allows for the most efficient
use of space and staff. However, other plans are possible such as a dressing/recovery area for each
procedure room, although this is usually impractical. There are other potentially useful modifications of
the basic arrangement, provided space and equipment are adequate. One would be separate areas or
even separate rooms for preprocedure holding and postprocedure recovery (inefficient for most units
with respect to utilization of staff). Another would be separate areas for inpatients and outpatients.
Hospitalized patients do not require extensive dressing facilities, although they often require closer
postprocedure monitoring while awaiting transport to a hospital ward. Conversely, outpatients generally
require more elaborate dressing facilities. Furthermore, it is somewhat inconsiderate to mix relatively
healthy patients with those who are seriously ill. Separation of inpatients and outpatients in the
dressing room does require some duplication of space because the ratio of inpatient to outpatient
procedures will fluctuate from day to day (in our unit, this may change by as much as 30 to 40% from
one day to the next).
The major concern with respect to the dressing room is its size; that is, what is the required ratio of bed
spaces to procedure rooms. There are various formulas but these can be deceptive because an ideal
ratio depends greatly on methods of practice. For example, the types and dosages of sedative drugs
used may necessitate more prolonged observation. Another major factor is the number of outpatient
therapeutic procedures that require protracted postprocedure observation. Just as there is a trend
toward performance of routine diagnostic procedures on an outpatient basis, there is also a trend in
favor of performing therapeutic procedures on an ambulatory basis. Thus, any formula for dressing
room bed space must be modified according to individual circumstances.
One basis for determining the number of beds needed is the average time a patient spends in the
dressing room. This depends in part on how the room is used. Usually, patients are brought to the
dressing room and assigned a space to be utilized throughout the procedure. Because ambulatory
patients require a place for personal belongings and to gown for the procedure, assignment of a
dressing room space on arrival is most satisfactory. Clothes and other personal belongings can be
placed in a lockable cabinet, the key to which the patient retains throughout the procedure (e.g., on a
bracelet). The result of this method, however, is that a certain number of bed spaces are reserved but
unoccupied while patients are in the procedure room. It is possible to circumvent this by simply not
assigning a space until after the procedure. Theoretically, this would be more efficient and allow a
reduction in the ratio of dressing room bed spaces to procedure rooms. However, in a large unit, the
flow of patients to and from multiple procedure rooms is always so unpredictable that this method
allows for only a slight reduction at best in the ratio of recovery beds to procedure rooms. Furthermore,
it makes changing clothes and handling personal belongings more difficult, and it usually deprives the
patient of some privacy.
The average time required for recovery after procedure can be determined by actual observation (e.g.,
a dressing room log). For many routine procedures, however, the recovery time is about equal to the
length of the procedure. For routine diagnostic EGD, for example, this is usually about 30 to 45 min.
For maximum efficiency, at least two dressing room spaces should be available for each procedure
room; a patient should be ready and waiting in one bed while the second bed awaits the patient
undergoing the procedure. After the first procedure has been performed, one bed will be occupied by a
recovering patient while the other contains a patient who is awaiting a procedure. One procedure room
at 100% efficiency can sustain about 14.0 procedure units per day. At 70% efficiency, this decreases to
9.8 procedure units per day. Because two bed spaces are always occupied (by a recovering patient
and a patient awaiting a procedure), the number of required bed spaces is approximately twice the
number of procedure units that will be performed in the procedure room. This would mean that under
ideal circumstances about 19.6 hr of dressing room time would be required to service one procedure
room that generates 9.8 procedure units per day. Based on an 8-hr working day, about 2.45 beds
would therefore be required to support one procedure room. However, the average weighted value for
procedures performed in a unit is never 1.0, since all procedures other than basic EGD are by
definition longer or more complex and therefore given a higher value in the weighted scale. If the
average weighted value for procedures in a unit were 1.5, then the actual number of procedures
performed in one room in 1 day could be as low as 6.5. Theoretically, only 1.6 dressing room spaces
would be needed for this level of activity, although this does not take into account the length and
difficulty of the procedure or the fact that more extended periods of recovery are required for some
procedures. Thus, the ideal ratio of beds for a unit operating rooms at 70% efficiency with a average
weighted scale value of 1.5 would be between 1.6 and 2.45 beds per procedure room. A unit with four
procedure rooms would therefore require between 8 and 10 recovery beds.
A cart exchange system has been adopted in many units. This requires the use of mobile beds or carts
(trolleys or stretchers, if you prefer). When properly gowned, the patient is placed on a cart in the
dressing room. The cart is wheeled to the procedure room where the examination takes place with the
patient on the same cart. On completion of the procedure, the cart is moved back to the dressing
room, where the recovery component of the procedure unit is completed. This very efficient method of
moving patients about in the unit also adds a measure of safety, in that it is not necessary to transfer
patients to and from a procedure table. Suitable carts are available from a number of manufacturers.
An essential feature is the ability to alter the height of the cart to accommodate short as well as tall
endoscopists. In many units, the patient's clothes and personal belongings are placed underneath the
cart (although some patients may find this less desirable compared with placing clothes and personal
effects in a locked wardrobe). This system works well for virtually all types of procedures with the
exception of ERCP where it is still necessary to transfer the patient to the x-ray table.
Lavatory Facilities
The number of toilets required depends in part on methods of preparation for colonoscopy. Colon
lavage is the preferred method of preparation in many units. In some units, the lavage is performed
within the unit under the supervision of a GIA. This markedly reduces the number of procedures
terminated because of poor preparation. Of all possible methods, however, this requires the maximum
lavatory facilities. About 3 to 4 hr is required to complete this type of preparation. It is mandatory that
one toilet be available during this period of time for each patient undergoing colon lavage. There must
be duplicate facilities for men and women. Because this approach is the most labor intensive and
requires installation of a large number of toilets within the endoscopy unit, most units rely on
self-preparation by patients. Nevertheless, it is extremely unwise to stint on toilet facilities in designing
a unit where large numbers of colonoscopy and sigmoidoscopy procedures will be performed.
Interview Rooms
The requirements for preprocedure and postprocedure interview rooms relate for the most part to the
level of outpatient activity in the unit. When this is high (e.g., 70% of procedures), more rooms will be
needed. In a unit with this level of outpatient procedure activity, a ratio of about one interview room to
every two procedure rooms is usually satisfactory.
Other Rooms
Requirements for other types of rooms shown in Table 44 depend on size and the range of functions
of the unit.
The reception area or room generally serves as a waiting room for family and friends. It may also serve
as a step-down recovery unit following certain procedures when extended close monitoring is not
necessary but it is still desirable that a patient remain in the unit for several hours.
As the endoscopy unit increases in size, it usually becomes necessary to separate the appointment
desk from other functions such as reception. In a large unit, scheduling can be so complex and
essential to orderly operations that a separate area for appointment secretaries is needed.
With increasing size comes increasing administrative complexity and a more intricate administrative
structure that requires a director, ideally a physician, and a chief GIA who should be a registered nurse,
preferably with managerial experience. These individuals make up the nucleus of a management team,
and their offices should be near, if not within, the endoscopy unit. It may be prudent to allocate office
space for other physicians within the unit, depending on the amount of time they spend in endoscopic
activities. Consultation/examination rooms must also be provided in this case.
A conference room, although not absolutely essential, is of great value, especially in teaching units.
Many different meetings may be scheduled in an endoscopy unit, such as demonstrations of new
equipment, management conferences, problem-solving discussions, and teaching conferences for
attending physicians, trainees, and GIAs. A pleasantly appointed conference room not only provides a
convenient, time-saving meeting place but also helps to promote esprit de corps simply by virtue of the
fact that there is a common ground.
A library is essential where endoscopy is taught. This is not a library in the ordinary sense, although
some books on endoscopy should be provided. Rather it should contain audiovisual materials such as
television tapes that demonstrate and illustrate endoscopic procedures and findings.
Photodocumentation changed dramatically with the inception of video endoscopy. Whereas 35 mm
slide photographs obtained with a camera attached to the ocular of a fiberoptic endoscope were once
the preferred method for photodocumentation, the advent of video endoscopy systems has made
videotape the standard for retention of endoscopic images. Computer systems have made retention of
images fast and easy. However, these static digital images are relatively unsatisfactory compared with
a videotaped segment of a procedure. Static EUS photographs, whether obtained using instant film or
a thermal printer, are especially unsatisfactory for review of findings inasmuch as EUS diagnosis is
based on real-time continuous observations. In the past, 35 mm slides were problematic because of
the need for handling, labeling, filing, and storing the slides or filmstrips. Extensive use of videotape
presents virtually the same problems. Because these are compounded over time, it is usually prudent
to set aside a small area in the unit for handling, storing, and labeling of videotapes when there is a
need to retain this type of information.
Report generation and record keeping are integral to the operation of any unit, large or small. These
basic functions can be either paper-based or computer-based; some combination of these is frequently
found within an endoscopy unit. The least sophisticated method is the handwritten procedure note.
This is sometimes combined with some form of computerized record-keeping system. Other units may
use a central dictating system accessed via a telephone. The least labor-intensive approach to report
generation and record keeping is an interactive computer system. Given their potential for speed,
accuracy, and savings in terms of cost and labor, computer-based systems are becoming the norm.
Whatever the reporting and record-archiving system, provision must be made in the design of a unit to
accommodate these functions. The exact requirements depend on the type of system adopted. For
example, it might be necessary to consider inclusion of a secretarial area. If a computer network
system is envisioned, it may be necessary to include conduits for wiring.
Unusable Space
In designing an endoscopy unit, it is important to realize that all the floor space allocated for the unit
cannot be used for the various rooms that one wishes to construct. About 30 to 35% of the available
space will be taken up by corridors, walls, building supports, stairwells, ductwork, and the like.
Floor Plan
The last step in designing an endoscopy unit is the arrangement of the desired rooms in a floor plan. In
all probability, these will fall into one of the following configurations: procedure rooms grouped in one
area, procedure rooms arranged around the dressing/recovery area, procedure rooms placed on
opposite sides of a central dressing/recovery area, or procedure rooms arranged around a central
cleaning/disinfection room. Often an ideal arrangement is prevented by the configuration of a building.
Even when floor space is adequate, the shape of the building may not accommodate the desired plan.
The compromises made at this stage determine what is good and bad about the layout of a unit, and
every unit has a little of each.
The arrangement of rooms should be made with traffic flow patterns in mind. For an ambulatory
patient, for example, the major flow is from reception to dressing room to procedure room and back
again to dressing room and then reception. Thus, the dressing room should be relatively close to the
procedure rooms. If at all possible, there should be separate routes to and from the procedure room. If
a separate room(s) for cleaning and storage of equipment is used, it should be located near the
procedure rooms. As a general principle, it is highly desirable that heavy traffic flow patterns not
intersect.
Procedure rooms are often grouped in one area of the unit. This is not absolutely necessary. Arranging
the procedure rooms around the dressing room or at opposite sides of a central dressing room can
have advantages. However, this usually makes it difficult to place cleaning and storage functions
centrally.
Support roomsother than the dressing, cleaning, and storage roomscan usually be located away
from the more active areas of the unit. However, personnel, equipment, and rooms for report
generation should be clustered if at all possible. Appointment and reception areas should also be close
together because communication between these areas is often needed.
Endoscopy Unit Profiles

The authors have engaged in an exercise that illustrates the principles and process of endoscopy unit
design presented in this chapter. One of us (MVS) developed imaginary profiles for four units. Using
the information provided in these fictional scenarios, another of the authors (MER), an architect with
extensive experience in the design of endoscopy units, developed schematic diagrams to fit these
hypothetical profiles. These are arranged as Figures 43, 44, 45 and 46. Each figure consists of a
profile (A), a design layout (B), and an inventory of rooms (C). The inventory for each profile unit lists
room category (description), dimensions, size (in square feet), numbers of such rooms, number of
persons that the room can accommodate, and the total area within the unit occupied by rooms of each
description. Numbers within the descriptive notes refer to a table of notes (Table 45). The design
layouts (B) represent design concepts and not actual architectural plans.
TABLE 45
Descriptive Notes for Endoscopy Unit Profiles
TABLE 45
Descriptive Notes for Endoscopy Unit Profiles
1. Remote monitors on procedure imaging network. Cables that run in ceiling areas
should be housed in metal piping or conduits, to shield video signals from
interference caused by fluorescent light ballasts, motors, or other potential
disruptions.
2. Walls that surround the procedure rooms, as well as doors to these rooms, should
have acoustic ratings to contain procedure sounds. Walls should have a minimum 55
sound transmission coefficient (STC) rating and doors should have 40 STC.
3. The procedure rooms should be designed with two modes of lighting. In a two-mode
system, fluorescent fixtures should provide the required level of work light (100
foot-candles), and adjustable low-voltage or mini-spotlights should be positioned to
provide low levels of light during procedures.
4. A double monitor system should be planned for the procedure rooms. In a
two-monitor system, both physician and assistant can have a comfortable view of the
procedure image.
5. Procedure spaces should be equipped with a means of exhaust and ventilation to
remove odors. A total air change capability should be designed. Potentially irritating
chemical fumes from washing disinfectants must be mechanically removed from the
cleaning/disinfection room. Provide 15 air changes per hour with direct exhaust of
used air to outside.
6. In cleaning rooms, exhaust grills should be located near the floor and at counter
height.
7. Endoscope washing rooms should be located in clusters with pairs of procedure
rooms.
8. Oxygen, suction, and medical air should be located in procedure rooms, recovery
areas, etc.
9. Emergency call button.
10. Lead lining in walls and doors. Anesthesia capability. Open hanging wall storage for
snares, etc.
11. Locked cabinets for drugs.
12. Emergency communication system.
13. Smoke exhaust system at laser procedure room.
14. Emergency (crash) cart.
(170)Figure 43. (For Parts A and C, see below Figure 43A and Figure 43C) Endoscopy
procedure suite: Endoscopy procedure suite: Profile ITertiary center. B, Design layout
(overall dimensions 126.5 85.6 feet; scale 1/16 inch = 1 foot). HMOhealth maintenance
organization;
EGDesophagogastroduodenoscopy;
ERCPendoscopic
retrograde
cholangiopancreatography; EUSendoscopic ultrasonography; FNAfine-needle aspiration;
PEGpercutaneous
endoscopic
gastrostomy;
GIgastrointestinal;
Ttoilet;
M/Fmale/female; Jan.janitor's; Mmen's; Wwomen's; sfsquare feet; Gengeneral.
Endoscopy procedure suite:
Figure 4-3A.
Profile ITertiary center. (See table below) A, Profile. HMOhealth maintenance

organization;
ERCPendoscopic
retrograde
PEGpercutaneous
endoscopic
gastrostomy;
GIgastrointestinal;
Ttoilet;

Figure 43 Part A
LOCATION: Large, growing metropolitan area, population approximately 3 million.
CATEGORY: Unit is based in a tertiary level (teaching) hospital of 950 beds. Hospital is the central
element in a large integrated health care system that includes numerous primary care providers
throughout the community, three smaller peripherally located hospitals, and several free-standing
clinics that provide both primary and specialty care. By design, tertiary level care is provided only at the
central hospital. Two thousand five hundred physicians work within the system. System is medical
school affiliated.
Care for about 25% of the population is governed by managed care contracts. This percentage is
increasing at 7% per year. About 35% of the population is Medicare eligible; this group is also
increasing by 7% per year. Fee-for-service insurance contracts cover about 15% of care provided
(expected to remain stable). The integrated system has its own insurance plan and an HMO. About
15% of patients who undergo procedures in the endoscopy unit are referred from outside the system
specifically for endoscopy. About 10% of care is provided to indigent persons.
No. of Procedures per Year
EGD
COLON
SIGMOIDOSCOPY
ERCP
EUS
PEG
NONENDOSCOPIC
7000 Case Acuity: 1.58

3000 (10% pediatric cases; 75 laser cases)
2000 (10% pediatric cases)
900 (mostly screening)
900 (80% therapeutic)
200 (includes FNA)
200
400
PROFILE: The unit is staffed by a closed faculty of gastroenterologists (includes three pediatric
gastroenterologists) and surgeons. Two physicians devote most of their professional effort to
endoscopy. Nine other physicians perform various endoscopic procedures. Nonendoscopic GI
procedures (e.g., esophageal manometry, liver biopsy) are also performed in the unit. There is a GI
fellowship program; surgery residents and pediatric GI fellows also receive endoscopic training.
Between two and five fellows are working in the unit at any given time. One is an advanced trainee in
endoscopy. A range of research projects are conducted in the unit including the evaluation of prototype
endoscopic equipment. Other types of endoscopic procedures (e.g., bronchoscopy) are not performed
in the unit. Seventy-five percent of procedures are performed on an ambulatory basis. The volume of
cases has increased 13% per year over the last 3 years.
The unit manager is a nurse administrator who reports to the GI Division Chief. There is also a
physician/endoscopist section head for endoscopy. There is a well-developed and functional matrix
management system. The unit operates at 75% efficiency.
Figure 4-3C.
(See table above) Profile ITertiary center. C, Inventory of rooms (numbers within
Descriptive Notes refer to Table 45). HMOhealth maintenance organization;
ERCPendoscopic
retrograde
PEGpercutaneous
endoscopic
gastrostomy;
GIgastrointestinal;
Ttoilet;
Figure 43 Part C
Room Name
Room Size
Area
Qty
Persons
Descriptive Notes
Room Name
Room Size
Area
Qty
Persons
Descriptive Notes
15 35
525 sf
30
Individual seating for 35 persons
8 9
72 sf
Administration Areas
Waiting room
Patient interview area
Discussion, examination, desk,

sink
Toilet
6 8
48 sf
Disabled accessible
Reception station
7 12
84 sf
Scheduling and appointments
Computer station
6 8
48 sf
Billing area
Files and records
8 10
80 sf
Copier/fax
6 8
48 sf
Offices
8 12
96 sf
Head nurses office
8 10
80 sf
Computer-server, imaging
12 18
216 sf
Computer image manager (1)
Conference room
16 26
416 sf
30
Videomonitor
Fellows work area
8 16
128 sf
Computer, coat storage, desk

space
6 6
36 sf
Male, female
Locker area
6 12
72 sf
Male, female
Patient toilets
6 6
36 sf
Endoscopy [general]
17 17
289 sf
Gen, EGD, colonoscopy

(2,3,4,5,8,12)
Endoscopy [manometry]
12 15
180 sf
Gen, motility (2,3,4,5,8,12)
Endoscopy [fluoroscopy]
14 20
280 sf
ERCP, dilation (2,3,4,5,8,10,12)
Endoscopy [laser]
17 17
289 sf
Gen, laser (2,3,4,5,8,12,13)
Endoscopy [ultrasound]
17 17
289 sf
Gen, ultrasonography (2,3,4,5,
High-density filing
Monitors (1)
Procedure Areas
Dressing
areas/preparationM/F
8,12)
Nurses station
8 8
64 sf
Equipment storage
10 17
170 sf
Dictation area
7 12
84 sf
8 15
120 sf
5 12
60 sf
6 12
72 sf
16
10
Curtain cubicles (8,
Isolation holding area
6 12
72 sf
Acoustic separation (8,9)
Toilets
6 6
36 sf
Nurses station
10 15
150 sf
Scope washing
Scope storage
Recovery
Handwashing sink
Mobile endoscopy cart, crash ca

lasers
Washing closet
12 sf
1
3 4
Automatic scope washer

(5,6,7,8,11)
Centrally located, good view (14

Bedpan hopper
Room Name
Room Size
Area
Qty
Persons
Descriptive Notes
Staff roomM/F
12 14
168 sf
12
Refrigerator, microwave, etc.
Staff toilet
6 6
36 sf
Doctors and staff lockers
6 8
48 sf
Storage supplies/utility
10 12
120 sf
Soiled area
8 10
80 sf
Janitors closet
4 4
16 sf
Support Services
Space Total
Shower stalls
18
70
Bulk style carts
Subtotal area
Circulation factor 35% (corridors, shafts, columns, etc.)

TOTAL NET AREA
Measurement excludes space occupied by stairs, elevators, electrical closets, and mechanical spaces
(171)Figure 44. Endoscopy procedure suite: Profile IICommunity hospital. A, Profile

(click to see Part A below). B, Design layout (overall dimensions 85 70 feet; scale 1/16 inch
= 1 foot). HMOhealth maintenance organization; EGDesophagogastroduodenoscopy;
ERCPendoscopic retrograde cholangiopancreatography; PEGpercutaneous endoscopic
gastrostomy; GIgastrointestinal; Ttoilet; M/Fmale/female; Jan.janitor's; sfsquare
feet; Gengeneral; Broncbronchoscopy. Endoscopy procedure suite: Profile
IICommunity hospital. C, (click to see Part C below)
Figure 44 Part A
LOCATION: Medium-sized city with a population of about 500,000.
CATEGORY: Unit is in a hospital of 300 beds. Hospital is affiliated with two other hospitals, and this
group competes with another system that is slightly smaller in size.
Health care for about 10% of the population is governed by managed care contracts. This percentage
is increasing at around 2% per year. About 40% of the population is Medicare eligible, and this is
increasing by 10% per year. About 20% of care is covered by fee-for-service insurance contracts.
There is a large indigent population which accounts for about 15% of care provided. Both of the
competition belong to HMOs. About 20% of the population belong to HMOs. The hospital operates an
extended care facility as well as a long term nursing home.
No. of Procedures per Year:
EGD
COLON
SIGMOIDOSCOPY
ERCP

75 (40% therapeutic)
PEG
NONENDOSCOPIC
100
65 (includes bronchoscopy)
PROFILE: Procedures are performed in the unit by various physicians ans physician groups who have
been credentialed by the hospital. Twenty-two physicians have credentials to perform endoscopy.
Eleven of these 22 physicians perform procedures on a regular basis. Nonendoscopic GI procedures
(e.g., esophageal manometry, liver biopsy, paracentesis) are performed in the unit. The hospital offers
training in family practice as part of its affiliation with a medical school. Family practice residents
receive training in sigmoidoscopy. There are no GI fellows. Research is not part of the mission
statement of the unit. The unit is multidisciplinary with pulmonary medicine; bronchoscopy and
pulmonary function studies are performed in the unit. Eighty-five percent of procedures are performed
on an ambulatory basis. The volume of procedures has been relatively steady over the last 3 years.
The unit manager is a nurse administrator who reports to the Director of Nursing. There is a physician
advisor. The unit operates at 60% of efficiency.
Figure 4-4C.
(Ssee table below) C, Inventory of rooms (numbers within Descriptive Notes refer to Table
45). HMOhealth maintenance organization; EGDesophagogastroduodenoscopy;
ERCPendoscopic retrograde cholangiopancreatography; PEGpercutaneous endoscopic
gastrostomy; GIgastrointestinal; Ttoilet; M/Fmale/female; Jan.janitor's; sfsquare
feet; Gengeneral; Broncbronchoscopy.
Figure 44 Part C
Room Name
Room Size
Area
Qty
Persons
Descriptive Note
15 30
450 sf
21
Individual seating
Consultation room
8 9
72 sf
Discussion, exam
Toilet
6 8
48 sf
Disabled accessib
7 12
84 sf
Computer station
6 8
48 sf
Files and records
8 10
80 sf
8 12
96 sf
Monitors (1)
Preparation area
6 6
36 sf
Male, female
Locker area
6 12
72 sf
Male, female
Patient toilet
Waiting room
Reception station
Nurses office
Scheduling and a
Procedure Areas
6 6
36 sf
Endoscopy [general]
17 17
289 sf
Gen, EGD, colono
Endoscopy [manometry]
12 12
144 sf
Gen, motility (2,3,
Endoscopy [fluoroscopy]
14 20
280 sf
Bronc, ERCP, dila

(2,3,4,5,8,10,12
Dictation area
7 12
84 sf
Nurses station
7 10
70 sf
12 15
180 sf
5 10
50 sf
6 10
60 sf
10
Scope washing
Scope storage
Recovery
Automatic scope
Curtain cubicles (
Room Name
Room Size
Area
Qty
Persons
Descriptive Note
Holding area
6 10
60 sf
Acoustic separati
Toilets
5 6
30 sf
8 8
64 sf
Centrally located,
12 10
120 sf
12
Refrigerator, micr
Staff toilet
6 6
36 sf
Doctors and staff lockers
10 10
100 sf
8 12
96 sf
Clean area
6 8
48 sf
Soiled area
8 8
64 sf
Janitors closet
4 4
16 sf
Nurses station
Support Services
Staff roomM/F
Storage supplies
Space Total
Shower stall
18
43
Bulk style carts
Subtotal area

TOTAL NET AREA
Measurement excludes space occupied by stairs, elevators, electrical closets, and mechanical spaces.
(172)Figure 45. Endoscopy procedure suite: Profile IIIAmbulatory center. A, Profile

(click to see Part A below). B, Design layout (overall dimensions 157 54 feet; scale 1/16
inch
=
1
foot).
HMOhealth
maintenance
organization;
ERCPendoscopic
retrograde
cholangiopancreatography; PEGpercutaneous endoscopic gastrostomy; EUSendoscopic
ultrasonography; Ttoilet; M/Fmale/female; Jan.janitor's; sfsquare feet;
Gengeneral. Part C, (click to se Part C below).
Figure 45 Part A
LOCATION: Large, growing metropolitan area, population approximately 3 million people.
CATEGORY: Unit is free-standing in a multispecialty ambulatory care facility. The facility is part of a
large integrated health care system that includes a 950-bed tertiary level (teaching) hospital, three
smaller hospitals, and numerous primary care providers throughout the community. By design, tertiary
level care is provided only at the central hospital. Two thousand five hundred physicians work within
the system. System is medical school affiliated.
Care for about 25% of the population is governed by managed care contracts. This percentage is
increasing at 7% per year. About 35% of the population is Medicare eligible; this group is also
increasing by 7% per year. Fee-for-service insurance contracts cover about 20% of care provided, and
this is expected to remain stable. Facility is in an affluent suburb; there is no indigent care. The
integrated system has its own insurance plan and an HMO.
EGD
COLON
SIGMOIDOSCOPY
ERCP
PEG
NONENDOSCOPIC

1100 (no pediatric cases)
0
0
50
PROFILE: The unit is staffed by closed faculty of three gastroenterologists. Nonendoscopic

procedures (e.g., esophageal manometry, liver biopsy) are also performed in the unit. ERCP, EUS,
and PEG procedures are referred to the central hospital. No endoscopic training or research is
conducted in the unit. Other types of endoscopic procedures (e.g., bronchoscopy) are performed in the
unit. All procedures are performed on an ambulatory basis. The volume of cases has increased 20%
per year over the last 3 years. the unit manager is a nurse administrator who reports to the
Administrator (nonphysician). the unit operates at 80% efficiency.
Figure 45C.
(See table below). Endoscopy procedure suite: Profile IIIAmbulatory center. C, Inventory
of rooms (numbers within Descriptive Notes refer to Table 45). HMOhealth maintenance
organization;
ERCPendoscopic
retrograde
cholangiopancreatography; PEGpercutaneous endoscopic gastrostomy; EUSendoscopic
ultrasonography; Ttoilet; M/Fmale/female; Jan.janitor's; sfsquare feet;
Gengeneral.
Figure 45 Part C
Room Name
Room Size
Area
Qty
Persons
Descriptive Notes
20 30
600 sf
30
Individual seating for 30 persons
Toilet
6 8
48 sf
Patient education area
7 9
63 sf
8 12
96 sf
Scheduling and appointments
Computer station
6 8
48 sf
Billing area
Files and records
12 20
240 sf
Copier/fax
6 8
48 sf
Offices, physicians
12 15
180 sf
Nurses office
8 10
80 sf
Computer server
12 18
216 sf
Computer image manager (1)
Conference room
13 16
208 sf
10
Video monitor
Examination rooms
8 10
80 sf
Patient toilets
6 6
36 sf
6 8
48 sf
Waiting room
Reception station
Disabled accessible
High-density filing
Monitors (1)
Discussion, examination, desk,

sink
Procedure Areas
Dressing
areas/preparationM/F
Male, female
Locker area
6 12
72 sf
Patient toilets
6 6
36 sf
Waiting
10 14
140 sf
Nurses desk
6 10
60 sf
15 18
270 sf
Gen, EGD, colonoscopy

(2,3,4,5,8,12)
Nurses station
8 8
64 sf
Handwashing sink
Utility/storage
12 17
204 sf
Dictation area
7 12
84 sf
8 15
120 sf
5 12
60 sf
Recovery
6 12
72 sf
Toilet
5 6
30 sf
Nurses station
8 12
96 sf
Centrally located, good view (14)
12 14
168 sf
10
Refrigerator, microwave, etc.
Staff toilet
6 6
36 sf
Staff lockers
10 12
120 sf
Storage supplies
8 12
96 sf
Clean area
6 10
60 sf
Soiled area
8 10
80 sf
Janitors closet
4 4
16 sf
Endoscopy [general]
Scope washing
Scope storage
Male, female
Automatic scope washer

(5,6,7,8,11)
Curtain cubicles (8,9)
Support Services
Staff roomM/F
Space Total
60
Shower stall
10
Bulk style carts
Subtotal area

TOTAL NET AREA
(173)Figure 46. Endoscopy procedure suite: Profile IVPrivate office. A, Profile (Click to
see Part A below). B, Design layout (overall dimensions 77 60 feet; scale 1/16 inch = 1
foot). HMOhealth maintenance organization; EGDesophagogastroduodenoscopy;
GIgastrointestinal; CEOchief executive officer; Ttoilet; sfsquare feet; Gengeneral.
C, (Click to see Part C below).
Figure 46 Part A
LOCATION: Medium-sized city with a population of about 1 million.

CATEGORY: Unit is within a physician office complex. It is owned by a group of physicians that
includes four gastroenterologists. Physicians attend at several different hospitals.
Health care for about 10% of the population is governed by managed care contracts. This percentage
is increasing at around 5% per year. About 40% of the population is Medicare eligible, and this is
increasing by 10% per year. About 20% of care is covered by fee-for-service insurance contracts.
About 20% of the population belongs to HMOs.
EGD
COLON
SIGMOIDOSCOPY
NONENDOSCOPIC

100 (includes bronchoscopy)
PROFILE: Procedures are performed in the unit by four gastroenterologists. Nonendoscopic GI

procedures (e.g., esophageal manometry) are performed in the unit. There is no training or research.
The unit is multidisciplinary with pulmonary medicine; bronchoscopy and pulmonary function studies
are performed in the unit. All procedures are performed on an ambulatory basis. The volume of
procedures has been increasing at 10% per year over the last three years.
The unit manager is a nurse administrator who reports to the Facility Manager (nurse) who reports to
the CEO of the corporation (physician). The unit operates at 70% efficiency.
Figure 46C
Endoscopy procedure suite: Profile IVPrivate office. C, Inventory of rooms (numbers

within Descriptive Notes refer to Table 45). HMOhealth maintenance organization;
EGDesophagogastroduodenoscopy; GIgastrointestinal; CEOchief executive officer;
Ttoilet; sfsquare feet; Gengeneral.
Figure 46 Part C
Room Name
Room Size
Area
Qty
Persons
Descriptive Note
13 20
260 sf
15
Individual seating
Waiting room
sons
Toilet
6 8
48 sf
Disabled accessib
Reception station
8 12
96 sf
Scheduling and a
Computer station
Billing area
6 8
48 sf
Files and records
12 15
180 sf
Copier/fax
6 6
36 sf
Physicians consultation
12 12
144 sf
Patient education
7 12
84 sf
Nurse manager
12 13
156 sf
Toilet
6 6
36 sf
6 10
60 sf
High-density filing
Monitors (1)
room
Procedure areas
Preparation/recovery/
dressing
Male, female
Room Name
Room Size
Area
Qty
Persons
Patient toilet
6 6
36 sf
Examination rooms
8 10
80 sf
Descriptive Note
Discussion, exam
sink
Endoscopy [general]
12 15
180 sf
Gen, EGD, colono

5,8,12)
Nurses station
3 10
30 sf
Storage
10 10
100 sf
Scope washing
8 15
120 sf
Handwashing sin
Automatic scope
8,11)
Scope storage
5 5
25 sf
Staff lounge
9 12
108 sf
Storage supplies
6 10
60 sf
Clean area
6 10
60 sf
Janitors closet
4 4
16 sf
Support Services
Space Total
39
Subtotal area

TOTAL NET AREA
Endoscopy Unit Management

Capital Equipment
The number of endoscopes necessary for routine procedures depends on how the unit functions. It has
been suggested that at least three upper GI endoscopes should be available for each room in which
EGD is performed (one being cleaned, one as backup, and one in use).12(174) This ratio is too high
and one backup instrument for every room is unnecessary, unless a simultaneous malfunction of every
instrument is expected. Two upper GI endoscopes for each multifunctional procedure room in which
EGD is performed should be adequate.
In a large unit (5000 or more procedures per year), repair costs are a significant operational cost item.
The actual cost of repairs will again depend on the skill and care exercised in using the instruments.
With careful record keeping, it is possible to cost-account major repairs.
The life expectancy of a GI endoscope might be expressed in number of procedures. This information
would be of great value, but unfortunately, there are no such data. The number of times an endoscope
can be used depends entirely on the manner in which it is used.13(175) In units where there are
numerous endoscopists with varying levels of skill, and where endoscopic procedures are taught, an
endoscope may not last very long. If, however, an endoscopy unit has an efficient record-keeping
system that tracks the usage of an instrument as well as its repair record, it is not too difficult to
determine the average lifespan of an endoscope within the unit.13(176)
Systems
Systematic methods for directing daily activities are crucial to orderly operations. Such systems are
also based on management concepts. Two of the most important systems are scheduling and record
keeping.
Scheduling
If the procedure room is the heart of the endoscopy unit, the daily schedule is its nervous system.
When 25 or more procedures requiring numerous different pieces of equipment and procedural skills
are being performed daily by several physicians, obviously an enormous potential for disorder and
confusion exists. The extent to which this potential remains latent and unrealized is a direct measure of
managerial skill in designing a scheduling system.
The primary task in devising a method of scheduling is to determine which procedure component will
be most restrictive. There may be an abundance of procedure rooms but limited numbers of
endoscopists, GIAs, or instruments. Some deficiencies and imbalances are easier to correct than
others. When there are adequate numbers of staff, the overriding factors that control the schedule are
availability of procedure rooms and, to a lesser extent, availability of equipment for special procedures.
These two factors necessarily govern the schedule.
The endoscopy schedule also reflects the basic philosophy for procedure room utilization. The
efficiency and capacity of the unit increase in proportion to the number of procedure rooms that are
considered multifunctional, although greater numbers of multifunctional rooms increase the complexity
of the schedule. When there is dogmatic restriction of procedure types to specific rooms, scheduling is
uncomplicated, but there is a loss of flexibility in responding to day-by-day fluctuations in the demand
for procedures.
The first step in devising a schedule is to set up a timetable for each available procedure room. This
timetable should be divided into units of time that equal one procedure unit (e.g., 1.0 procedure unit =
30 min). Special procedures (see definition earlier in this chapter) should then be restricted to specific
rooms to the extent that necessary items of equipment are limited or unique in the unit. Thus, one
room might be designated for ERCP (more desirable would be the simple designation fluoroscopy),
another for laser endoscopy, and perhaps another for EUS. These designated rooms, however, should
not stand idle in the absence of special procedure activity. If properly designed and equipped, a
laser-designated room or EUS-designated room will serve nicely for EGD and colonoscopy and any
number of other procedures. Even an ERCP room can be used for other procedures when necessary.
A multifunctional-room scheduling scheme requires careful thought and judgment by the appointment
secretary. In effect, this individual must have a basic insight as to the function of the unit. Such an
individual must be a tactician with the ability to effectively deploy and coordinate staff, equipment, and
rooms. In a highly multifunctional room system, close cooperation between the appointment secretary
and the GIA in charge is essential for success. With such cooperation, it is entirely possible to shift and
adjust staff, rooms, and equipment on an hour-to-hour basis to achieve a highly flexible system that
adapts quickly to delays in one or more procedure rooms, urgent requests for examinations, equipment
breakdown, and a variety of other foreseen and unforeseen circumstances.
The work habits of physicians can be one of the more difficult problems in scheduling. Although most
physicians are not only hard working but also productive, their schedules and work habits may be at
odds with those of the unit as a whole. Therefore, an important stratagem is to induce the endoscopists
to think in terms of overall productivity and goals of the endoscopy unit. One step in this direction is to
persuade them to accept assigned, regularly scheduled work periods in the unit. In medical
communities where primary diagnostic endoscopy is accepted, certain procedures can be scheduled
with any endoscopist working in the unit at the time requested by the patient or referring physician.
Records, Statistics, and Computers
The administration and, hence, the operation of an endoscopy unit are facilitated by reliable and readily
available statistical data. In a large unit, such data are critical to the decision-making process.
Certain types of information (e.g., complications) are relevant to the maintenance of high standards of
patient care. When the volume of cases is high and numerous endoscopists and GIAs are performing
procedures, a pattern of recurring complications as a result of a breakdown in technique may become
evident only by reference to data gathered over an extended period. Accurate methods of endoscopy
record keeping can also be used to support and promote scientific inquiry.
A record-keeping system is a management tool. Data and statistics may seem mundane and irrelevant
in a small unit where standards of care are readily assessed, scientific inquiry has a low priority, and
management is a relatively simple, part-time business. However, recall and subjective assessment are
crude administrative techniques that are poorly suited, not only for recognizing and dealing with
problems but also for guiding the unit toward its goals. When problems such as low staff morale,
confusion and delays in scheduling, poor condition of instruments, high cost, low productivity, obsolete
equipment, inadequate numbers of or poorly trained GIAs, excessive delays between procedures,
shortages of supplies, declining referrals, dissatisfied referring physicians, dissatisfied patients, an
insufficient number of procedure rooms, crowded working conditions, and incompetent personnel
arise, the first query is whether fault lies in the system of management or perhaps the lack of a system.
When problems are numerous and interrelated, there may be a basic inability to define and understand
the difficulty. This is often due to a lack of pertinent information.
Computers
A computer can perform three operations in the endoscopy unit: report generation, image
management, and data management. The relative importance of each will depend on the objectives of
the individual endoscopy unit. For almost all facilities, however, a computer-generated report is likely to
be of benefit. A glance at Table 41 and a rudimentary knowledge of computers lead to the inevitable
conclusion that the procedure unit could be streamlined by delegating as many "secretarial/clerical"
tasks as possible to the computer. With proper software and a relatively inexpensive computer, it is
entirely possible to dispense with handwritten chart notes, report dictation and typing, and all file
copies. Rather, the endoscopist interacts with the computer by responding to a series of questions.
Based on these responses, the computer can produce an accurate, "hard-copy," English language (or
any language) report that can be placed in the patient's chart within minutes. Best of all, the patient's
record then becomes part of a computerized endoscopy database (eliminating the file copy) that can
be accessed to provide information that has a bearing on the management of the endoscopy unit.
There can be no doubt that computer technology has had, and will continue to have, a profound effect
on virtually all aspects of modern society. However, evidence of this is not to be found in the average
endoscopy unit. Many unit operations could be accomplished by computers with greater efficiency and
at lower cost. However, the transition to such automated systems has been slow to develop by
comparison with virtually any other field in which large amounts of data are processed. In fact, the
technology required to transform even a large endoscopy unit into a fully automated, state-of-the-art,
data-processing machine is relatively mundane by modern standards. The missing factor has been
appropriate and acceptable software. The potential and the problems entailed in the transition of the
endoscopy unit to a computer-based operation are discussed in detail in Chapter 6: Medical
Informatics and Chapter 7: Electronic Image Management.
The electronic endoscope is a significant driving force behind the effort to bring the computer into the
endoscopy unit. It is logical to believe that the combination of computers and electronic endoscopes
can provide great benefit to the operation of endoscopy units. Whether this will occur is difficult to
predict. The lack of a truly well-designed computer/software system with the flexibility to meet the
objectives of diverse endoscopy units at reasonable cost may remain a limiting factor for some time to
come. However, GI endoscopy has become a huge and growing field of medicine, and size alone may
be the technologic imperative of greatest significance to the realization of the potential benefits of the
computer.
Staffing Levels
Endoscopists
From a managerial viewpoint, it is necessary to know how many endoscopists work in an endoscopy
unit. This is not the same as the number of physicians who perform procedures because most will not
be full-time endoscopists. However, it is useful to calculate the equivalent number of full-time
endoscopists using the part-time contributions of each of the staff members, that is, to express the
availability of endoscopists in terms of full-time equivalents (FTEs).
It is easy to estimate the number of procedures (using the weighted scale method to allow for more
complex procedures) that one full-time endoscopist can perform in 1 year based on 256 working days
per year and a 7-hr day. This number should then be modified by taking into account scheduled staff
meetings, conferences, vacation and meeting time, and perhaps a day or two for illness. Furthermore,
time is inevitably consumed in a variety of procedure-related activities such as review of x-ray films,
chart maintenance, discussions with patients and family members, and unavoidable interruptions. An
estimate of the efficiency of the unit should be added because the endoscopist cannot be more
efficient than the least efficient component of the system. This works out to about 1800 procedure
units. This computation should not be used out of context because it is based on a hypothetical
institution that has three or four conferences per week, 30 days vacation and meeting days per year for
a tired, middle-aged endoscopist, an expected endoscopy unit efficiency of 70%, and other
assumptions. The actual maximum output of a full-time endoscopist in terms of procedure units can,
however, be calculated for any institution. The actual number of procedures will also be less because
the endoscopist usually performs a variety of procedures that are more complex, that is, have a higher
weighted scale value, than a simple EGD.
It is important to note that a calculation of the number of procedure units that one FTE endoscopist
performs will always exceed the maximum theoretical capacity of one procedure room, usually by
about 300 to 500 procedures per year.
The purpose of these computations is to determine whether there are sufficient numbers of procedure
rooms, items of equipment, and GIAs in relation to the number of endoscopists. The staff of few if any
units is made up by full-time endoscopists. However, the part-time contributions of the actual members
of the staff can be converted to the number of FTEs. The easiest method is to determine how much
time each endoscopist spends in the unit during an average week. Because there are 10 half days in a
week, a full-time individual would have a score of 10. Often, an endoscopist will perform procedures on
a half-day basis. A physician who performs procedures 3 mornings per week would have a score of 3;
someone working 3 mornings and 2 afternoons would score 5. Totaling the scores for all endoscopists
and dividing by 10 yields the number of FTEs performing endoscopy. This analysis assumes that the
unit is fully operational and that procedures are performed throughout the day (idle rooms and GIAs
are very expensive). There are other ways to use this method of analysis. For example, when it
appears that limited space is restricting procedural activity and acquisition of additional rooms or even
construction of a new unit is contemplated, a similar calculation can be made based on the amount of
time that each endoscopist thinks she or he needs to perform procedures.
Gastrointestinal Assistants
The number of GIAs needed in a unit is a very important determination. Of all the types of personnel
that contribute to one procedure unit, the GIA has the most functions and duties. The maximum
capacity of 1 GIA in terms of procedure units per year will be lower than that of the endoscopist or the
procedure room itself. Applying the methods of analysis used throughout this chapter results in a figure
of about 1000 procedure units per year. This therefore translates to a ratio of about 1.5 GIAs per
procedure room or 1.8 GIAs per FTE endoscopist.
The role of the GIA in the endoscopy unit is discussed in Chapter 5: The Gastrointestinal Assistant. In
terms of management of the unit, there are many ways to deploy GIAs. These methods are also
considered in Chapter 5.
Other Personnel
There are many other important personnel in a well-organized endoscopy unit. A process of
redelegating responsibilities in relation to the procedure unit becomes more necessary as the volume
and complexity of procedures increase. The easiest way to conceptualize this is to think in terms of
one procedure unit (see Table 41) and the various types of personnel needed to accomplish one unit
of work. Early in the development of the unit, one employee may be responsible for more than one
component. With growth, more employees will be required and each will be responsible for a smaller
number of components in the procedure unit. For example, in a small unit (not more than 1000
procedure units per year), one secretary might handle appointments, report generation, statistics, and
filing. However, as volume increases, the number of employees required to accomplish each of the
steps in the procedure unit increases, and each employee must focus on a smaller number of
components.
After the head nurse and director of the unit, one of the most important members of the team is the
appointment secretary, especially in a unit that has a flexible method of scheduling as described
previously. At a level of about 2500 to 3000 procedure units per year, an endoscopy unit should have
one person assigned to this job.
Quality and Efficiency

As a rule, it is not possible to eliminate any of the basic elements in the procedure unit (see Table 41).
To assess the quality and efficiency of an endoscopy unit, it is only necessary to study the way in which
each of the basic steps in the procedure unit are accomplished. There are numerous definitions of
efficiency, all of which depend on overall objectives.
Quality and Standards of Patient Care
A high standard of patient care should be an obvious goal. Generally, this pertains to certain aspects of
the endoscopy procedure itself that can be assessed by reviewing factors such as appropriateness of
indications, complication rates, and numbers of unsuccessful procedures. For example: How often are
colonoscopy procedures canceled because of poor preparation? Is there a high incidence of sepsis in
association with ERCP? What percentage of colonoscopy procedures are not accomplished to the
cecum? Additional, less tangible but no less consequential considerations are the qualifications,
training, and morale of the staff, including the GIA.
Efficiency and the Business of Endoscopy
From the administrative standpoint, two interrelated factors determine efficiency: time and cost.
It is very much in vogue in the United States to believe that solutions to health care problems are to be
found in disciplines that are remote to the practice of medicine. It is thought, for example, that the
application of principles and methods from fields such as business, finance, marketing, and
management will encourage the delivery of care that is more effective and less costly. These solutions
become conflicts, however, when applied in an overly rigorous fashion to health care. Nevertheless,
there is much to be gained in trying to look at medical practice from new perspectives.
For better or worse, the modern endoscopy unit is in many respects a business, not by analogy but in
point of fact. In simple terms, resources (personnel, supplies, physical plant, and equipment) are used
to produce a product (a procedure unit). This means that there are costs involved (payroll, capital
equipment, repairs, and the like) and that the services provided to health care consumers (sometimes
referred to as patients) generate income. The operation of any endoscopy unit, therefore, can be
expressed as an operating statement, that is, a breakdown of cost versus revenue. It would be
intriguing, perhaps provocative, to develop this theme of the endoscopy unit as a business, but it is
sufficient for the subsequent discussion if the veracity of this concept is accepted, even if only
conditionally.
In the remaining years before the 21st century and no doubt beyond, it is no great feat to predict that
the overriding consideration in health care will be cost. Any description of the endoscopy unit of the
future is ill conceived if it does not take this into consideration.
From the standpoint of the endoscopy unit as a business, it is appropriate to inquire as to the costs of
doing business. After the physical plant itself (in reality, a fixed asset that depreciates long term), there
are four major cost factors: supplies, capital equipment (also a fixed asset), repairs, and personnel. For
purposes of cost analysis, the first of these, supplies, gains in significance as the numbers of
procedures increase, especially those with a high complexity (weighted) value. Capital costs are
obviously significant, especially if maintenance of equipment is included, and cannot be totally ignored.
However, the cost of equipment is relatively small compared with the cost of labor. From a long-range
accounting point of view, the financial commitment to a staff member in terms of salary and benefits far
exceeds that of almost any piece of equipment. Even the most expensive laser or fluoroscope
represents a one-time cost that can be amortized over several years, even with allowances for periodic
repair. With proper care, many endoscopes can be used for 1000 or more procedures before a major
overhaul is necessary. A member of the endoscopy unit staff, however, represents a repeated cost
year to year that will only increase with periodic salary increases and seniority. Control of labor costs is
therefore axiomatic.
The efficiency and productivity of any endoscopy unit can be estimated quickly by comparing the
number of procedures performed (with allowances for complexity) with the number of people
(nonphysicians) required to generate these procedures. Complex procedures that consume many
resources may not be very productive from the business point of view. Furthermore, such procedures
are often difficult to perform and, therefore, most likely to require extended periods of time and
additional supplies that, from the business standpoint, amount to a cost overrun.
Every endoscopy procedure requires a certain minimum number of support personnel below which the
quality and safety of the examination are compromised. This would suggest that labor costs in the
endoscopy unit have a relatively fixed relation to the number of procedures performed. It is often
argued, therefore, that the only way to control cost is to reduce service. However, a better answer to
this dilemma is to maximize the productivity of the unit, that is, greater numbers of procedure units in
relation to the number of personnel available. The productivity of a unit can be defined as the ratio of
the number of procedures performed per number of personnel multiplied by a constant that reflects the
overall mix of procedures in terms of complexity.
How much time is required to accomplish one procedure unit? This also has a direct bearing on patient
care. If the time required is excessive, delays in scheduling are the usual result. This can be extremely
costly for hospitalized patients. Delays in scheduling can be a major restricting factor in the growth of
an endoscopy unit. In some cases, growth is restricted by limitations on space or staff. But in others, it
may seem that space and staffing should be more than adequate for continued growth. When such
growth is not apparent, there are many possible reasons, including a declining referral base and
poor-quality patient services. However, another factor, often more difficult to define, is inefficient use of
existing resources.
Productivity is influenced to a substantial extent by the design of an endoscopy unit. Design determines
functionality, that is, the ease with which required tasks are accomplished. Numerous small errors in
the organization of procedure rooms or the overall layout of a unit increase the time and effort required
for a procedure. With growth, the "cost" of such small errors is magnified. In order to achieve the
highest possible levels of productivity, efficiency must be one of the guiding principles of design in the
endoscopy unit of the future.
The Endoscopy Unit of the Future

Design
The endoscopy unit of the 21st century will be larger than its predecessors. The present growth in
numbers of procedures will be sustained over the next several decades. Although new endoscopic
methods will be developed and applied to clinical problems, most of the growth will occur in procedures
that are already well established. Barium contrast x-ray procedures will be supplanted, for most
purposes, by endoscopy. The most dramatic increase in numbers of procedures will be noted for
colonoscopy. In many countries, large portions of the endoscopy unit will be reserved for colon cancer
screening and the care of patients with colon cancer and polyps.
Socioeconomic Factors
It is essential to recognize that endoscopy alone, in the sense of technologic advancements, will no
longer be the single, dominant factor shaping the endoscopy unit. Rather, other forces will define the
unit and dictate to a greater extent the course of development of endoscopy; a subtle but nevertheless
profound shift in emphasis will occur. These new elements, not medical in the sense of individual
patient care, will usually be socioeconomic in nature. They differ greatly from country to country, so that
attempts at generalizations become problematic. Nonetheless, it cannot be denied that GI endoscopy
and thus the GI endoscopy unit of the 21st century will be influenced to a marked degree by new,
largely external forces.
When systems of health care are centrally planned by government agencies, it is common that
complex procedures requiring high levels of skill and expensive equipment are performed by a
relatively small number of individuals in a few centers. Economic issues are a major driving force in
such a system, so that the allocation of funds for staff and equipment determines the pattern of
medical practice.
Economic forces are becoming increasingly important in health care delivery in the United States and
most other countries, although these forces operate differently. In the United States, the trend is toward
a competitive, marketplace strategy in which health care delivery systems must vie for fixed or even
reduced health care dollars. To an ever-increasing extent, the health care dollar is spent by
organizations, the so-called third-party payers, rather than individuals. The third party has traditionally
been an insurance company or government agency, but increasingly, other third parties such as
corporations will negotiate for health care benefits, frequently with other for-profit corporations.
Economic factors such as these favor larger, efficient, flexible, and responsive institutions that have an
economy of scale. The large, for-profit health care company is a derivative of these changes. Another
result is the consolidation of smaller health care institutions to achieve an economy of scale. Thus,
there are hospital mergers and acquisitions, as well as affiliations and amalgamations as "networks," to
provide total health care systems. The essence of competition is that over the course of time there are
fewer competitors.
It is impossible to know the long-range consequences of these socioeconomic forces with regard to an
endoscopy unit. One net result may be that within any health care systema network of hospitals, for
exampleendoscopic services may also be consolidated. It is logical to expect that specialized
endoscopic procedures will be concentrated at one or a few locations within an organization. The end
result may be a system that more closely resembles those in which governmental agencies are
responsible for centralized planning of health care; that is, complex procedures that require high levels
of skill and expensive equipment will be performed by a relatively small number of individuals in a few
centers.
The need to control costs will lead to better designed and more efficient units that require fewer
support personnel. Much of this will be accomplished by application of technology, especially
computers. Most of the clerical functions such as scheduling, report generation, recall of patients, data
processing, record keeping, image management, supply, and many managerial functions related to the
operation of the unit will be streamlined by computer programs designed specifically for endoscopy.
As manufacturers recognize the need for efficiency and cost control, they will develop complete
endoscopy procedure rooms that can be purchased as modular units. These will comprise
interchangeable components that can be arranged in different patterns to fit into available space. The
modular procedure room will be designed for optimum efficiency with particular emphasis on the tasks
performed by the GIA.
Ambulatory Endoscopy Services

The endoscopy unit of the next century will continue to provide services to hospitalized patients, but the
number of such patients will continue to decline in relation to the number of ambulatory patients
undergoing procedures. In most respects, the endoscopy unit will be an outpatient facility. As such,
more space will be allocated to patient waiting and recovery areas, and greater attention to the comfort
and convenience of the patient will be evident throughout the unit.
Primary Diagnostic Endoscopy

Another emerging trend is that of providing endoscopic services on request. In the future, a substantial
portion of relatively simple diagnostic procedures such as EGD will be performed at the request of a
referring physician without prior consultation, much the same as a radiologist performs an upper GI
x-ray series. This practice is referred to by a variety of names, such as primary diagnostic endoscopy,
primary panendoscopy, and open access endoscopy. Frequently controversial, this concept offers
certain advantages but also has potential problems (see Chapter 37: Indications, Contraindications,
and Complications of Upper Gastrointestinal Endoscopy). Such a system presently works best in a
closed staff environment in which there is an established referral pattern of patients for endoscopy and
in which the endoscopist has an existing professional relationship as well as direct communication with
physicians who request endoscopy.
New Technology
GI endoscopy since the early 1970s has been a masterpiece built on two complementary themes:
technical development and the clinical application of technology. To a significant extent, the endoscopy
unit has evolved to accommodate this progress. Its existence is in fact secondary to the more
fundamental purpose of performing procedures of ever-increasing sophistication in greater numbers of
patients. If the future could be modeled on the past, then the endoscopy unit of the 21st century would
be a truly remarkable place, a technologic tour de force. Unfortunately, the old formulas will not be
quite so reliable in the future.
Over the next decade, there will be relatively few technologic developments in the field of endoscopy
that are truly innovative with respect to endoscopic diagnosis and therapy. New devices and methods
will tend to be more costly in terms of the time required for development and their complexity. These
will address a narrower range of clinical problems and will be less applicable to large groups of
patients, so that the overall effect on numbers of procedures and the functional nature of the
endoscopy unit will be relatively small.
Developmental advances in the field of endoscopy since the mid-1960s have come at relatively low
cost. Future technologic gains that require computers, videosystems, lasers, specialized endoscopes,
and other instruments will have a higher monetary cost. Currently, EUS is the best example of this
trend. It is doubtful, therefore, that every hospital will be able to afford a state-of-the-art endoscopy unit
in the future.
Computers
The endoscopy unit of the 21st century will be all electronic. In addition to new methods of diagnosis,
the electronic instrument will link the procedure itself to the computer that runs the unit. Computer
systems for image management will be commonplace, as will electronic communication between units
for consultation and patient referral. In systems of health care that comprise multiple facilities, the
paper record will be a relic of the past. Because patients may be seen in more than one facility within
the system, the patient record that consists of pieces of paper must be replaced by electronic systems
of computer databases and networks. The computer system that serves the endoscopy unit will be but
one part of a much larger network of computers.
Expertise
It is virtually certain that a full range of endoscopic services will not be found in every unit. Decisions for
new technology will be more difficult because of cost and uncertainty as to the clinical applications of
new devices. In many cases, choices will have to be made as to a range of technical capability, and it
will be difficult or impossible to offer full endoscopic services. In the coming years, the numerous
institutions that make up the delivery of health care will continue to consolidate. There will be smaller
numbers of health care systems, but these will consist of multiple hospitals and other facilities. Within
each system, a state-of-the-art endoscopy unit for each component facility will not be cost effective.
The net result of this line of reasoning is the separation of endoscopy units into two categories, tertiary
and primary.
Regulation and Oversight

It would be unrealistic to deny that the policies of government regulatory agencies will continue to
influence endoscopy and, therefore, the endoscopy unit, at least in the United States. It is impossible to
explore in this essay the impact of each and every regulatory body on the endoscopy unit of the future.
Whether this will enhance the quality of care and reduce costs is a matter of some speculation.
Furthermore, there is a real concern that regulatory policy can adversely affect the development of new
technology.
The Consumer Movement

There is a growing demand for objective measurements of the quality of medical care. This is driven in
part by concerns over the cost of care, but also by more fundamental questions as to the true value of
care. This movement has many facets and ramifications, but for the endoscopy unit, it will mean more
rigorous standards of practice in terms of appropriate use, outcome, and safety of procedures. There
will be a great need for accurate data in these three areas. Added emphasis will be given to proper
training and credentials, for physicians as well as GIAs.
Greater emphasis will be given to assessing the benefit to patients of new technology. That a new
device or method can be demonstrated to work as declared will be insufficient justification for
acceptance as standard practice. Rather, every invention and new application of existing technology
must be shown to improve outcomes for patients. The time required for the development and
implementation of new techniques will therefore be longer.
Acknowledgment
The authors gratefully acknowledge the assistance of Mr. S. M. Greengrass, Research and
Development Director, KeyMed, Ltd. (Southend-on-Sea, Essex, England), for his assistance in
providing data from the KeyMed study for this chapter.
REFERENCES
1.
2.
3.
4.
Overholt BF, Chobanian SJ. Office Endoscopy. Baltimore: Williams & Wilkins. 1990.
Waye JD, Rich ME. Planning an endoscopy suite for office and hospital. New York,
Igaku-Shoin. 1990.
Gostout CJ, Ott BJ, Burton D, DiMagno EP. Design of the endoscopy procedure room.
Gastrointest Endosc Clin N Am 1993;3:50924.
Burton D, Ott BJ, Gostout CJ, DiMagno EP. Approach to designing a gastrointestinal
endoscopy unit. Gastrointest Endosc Clin N Am 1993;3:52540.
5.
Gostout CJ. Unit management. Gastrointest Endosc Clin N Am 1993;3:54147.

The American Society for Gastrointestinal Endoscopy. Guidelines for establishment of
gastrointestinal endoscopy areas. Gastrointest Endosc 1988;34(suppl 3):3S.
7. Seifert E, Weismuller J. How to run an endoscopy unit? Experience in the Federal Republic of
Germany. Results of a survey of 31 centers. Endoscopy 1986;18:204.
8.
The British Society of Gastroenterology. Design of gastrointestinal endoscopy units. The
British Society of Gastroenterology, 1984;128.
9.
The British Society of Gastroenterology. Report of a working party on the staffing of
endoscopy units. The British Society of Gastroenterology, 1987;139.
10. The British Society of Gastroenterology. Provision of gastrointestinal endoscopy and related
services for a district general hospital. Working Party of the Clinical Services Committee of
the British Society of Gastroenterology. Gut 1991;32:95105.
11.
Manoy R. Ergonomics and Design Study. Personal communication. 1989.
12. Larson DE, Ott BJ. The structure and function of the outpatient endoscopy unit. Gastrointest
Endosc 1986;32:104.
13. Urayama S, Kozarek R, Raltz S. Evaluation of per-procedure equipment costs in an outpatient
endoscopy center. Gastrointest Endosc 1996;44:12932.
6.
Chapter 5 The Gastrointestinal Assistant

(177)
(178)
JOANNE ZEROSKE, B.S.N., M.B.A.
Evolution of the Gastrointestinal Assistant Concept

The comfort and safety of a patient undergoing gastrointestinal endoscopy in the era of the rigid
endoscope depended largely on the skill and attention of the endoscopy assistant. Just as rigid
endoscopy is no longer widely practiced, few gastrointestinal assistants (GIAs) today are cognizant of
the role of the "head-holder." This is unfortunate because the present-day GIA has little awareness of
the background and early development of this form of nursing. Although an appreciation of the role of
the assistant from a historical perspective is not essential, it adds greatly to the sense of profession
and of participation in the development of gastrointestinal endoscopy.
Although much has changed for the endoscopist and GIA, the fundamental relationship remains the
same. During the early days of endoscopy, the assistant had two basic functions, which have remained
essentially unaltered: custody and preparation of endoscopic instruments and a role in the comfort and
safety of the patient. The specific ways in which these functions are carried out, however, have
changed substantially. Responsibility for only the endoscopic instruments now encompasses a large
endoscopy unit and numerous items of equipment. Patients today undergo many types of procedures
that require numerous and different items of equipment; each procedure has unique potential
complications, and each presents special difficulties for patients. Most procedures are now performed
on an ambulatory basis, so that patient education and assessment of the patient's status before
discharge are integral to the role of the GIA.
The remarkable increase in the number of procedures performed presents new problems for the GIA
with respect to patient comfort and safety. Heretofore, the assistant might have been responsible for
the postprocedure recovery of one or a few patients, whereas the GIA now has responsibility for
numerous patients with various disorders who have undergone any number of different procedures.
Safety in rigid endoscopy once depended on holding the patient's head properly and monitoring
respiratory and cardiac function. Although exact positioning of the patient is perhaps less important
now, attention to the patient's vital functions is no less essential. When large numbers of procedures
are performed, they become "routine." This plus the inherent low complication rate for most
endoscopic procedures can lead to a false sense of security. In a busy unit, the many tasks to be
performed can distract attention from the most critical function of patient monitoring.
The GIA's custody of the instruments also has a direct bearing on patient safety. The endoscopist and
the patient both expect that all instruments will be in working order and that the success and safety of
the procedure will not be compromised by poor condition of the equipment or by disorganization. This
has ramifications ranging from instrument malfunction during a critical stage of a procedure to disease
transmission by inadequately cleaned instruments. A high volume of procedures can also compound
problems with respect to proper cleaning and maintenance of equipment.
The avoidance of pain, discomfort, and injury to the patient as a direct result of a procedure is
essentially the responsibility of the endoscopist. In many respects, however, the GIA shares in this
responsibility. For example, most patients approach a procedure with trepidation. The size and intense
activity of a large endoscopy unit can increase this anxiety by virtue of a seeming lack of individual
attention. These factors make the role of the GIA even more crucial with respect to the emotional
preparation of the patient and require highly developed skills.
The relationship of nursing to gastrointestinal endoscopy in the past was usually that of a part-time
assistant, since the volume and variety of procedures performed did not fully occupy a nurse in most
institutions. Growth and evolution of gastrointestinal endoscopy have changed this relationship to the
extent that the role of the GIA has become a highly specialized area within the field of nursing. It is
appropriate that this full-time profession with its special requirements and qualifications be given its
own designation; hence, the term gastrointestinal assistant. This specialty has grown to the point that
the Society of Gastroenterology Nurses and Associates now offers professional certification to both
nurses and GIAs.
Qualifications
Certain fundamental qualifications are required for an individual to become a GIA. The candidate
should be trained in the basic aspects of patient care; prior experience involving contact with patients is
desirable. Most gastrointestinal endoscopy procedures require administration of conscious sedation.
The GIA must be familiar with the essential aspects of cardiovascular and respiratory physiology and
must be able to quickly and accurately assess the cardiopulmonary status of any patient. Certification
in cardiopulmonary resuscitation (CPR) should be required, including a thorough knowledge of all
drugs and equipment used in CPR.
Medications are administered in most endoscopic procedures, so that the GIA must be knowledgeable
about dosages, methods of administration, side effects, and interactions of the various drugs used.
Since the medications used in endoscopy are often controlled substances, the GIA must be licensed
and qualified to handle these agents.
For the GIA with no previous experience, it is helpful to understand that gastrointestinal endoscopy has
a large technical component and that a certain amount of manual dexterity and mechanical ability are
necessary. The amount of equipment that must be put in place, maintained, cleaned, and inventoried
is far more than that found in most other patient care situations. Candidates with experience in the
operating room, emergency room, cardiac catheterization laboratory, intensive care areas, or other
procedurally oriented settings may acclimate to gastrointestinal endoscopy more easily than those
whose background is general medical surgical nursing.
An endoscopic procedure encompasses many tasks that must be accomplished, sometimes
simultaneously. These include monitoring and reassuring the patient, assisting with the administration
of medications, processing specimens, and keeping records and documentation, in addition to working
with the endoscopy equipment.
The GIA must be flexible and able to adapt. Abrupt changes in the schedule of procedures are
common. Level scheduling is problematic in most active units, busy periods alternate with lighter ones,
and frequently the daily schedule of procedures does not conclude on time. The GIA must also be able
to adapt to changing needs for equipment. As a procedure progresses, accessories and instruments
may be required that could not be anticipated at the beginning of the procedure.
Of all the special attributes required of the GIA, attention to detail is perhaps the most important. The
difference between an adequate and an excellent GIA is the quality of thoroughness and punctilious
attention to the numerous small aspects of patient care, procedures, and equipment. By virtue of their
education, nurses are uniquely qualified for this position.
Functions of the GIA

The Procedure Unit
The endoscopy "procedure unit" is described in Chapter 4: The Endoscopy Unit. Conceptionally, the
essential elements derive from all endoscopic procedures being alike in many respects, since all have
certain steps in common, and the actual endoscopic component of the procedure unit being only one
part of a more complex process (Table 51). The GIA has responsibility for more steps and for the
accomplishment of more tasks within the endoscopy procedure unit than any of the other personnel,
including the endoscopist (Table 51).
Functions of the
Gastrointestinal Assistant in the Procedure
Unit
TABLE 51
PREPROCEDURE
Scheduling
Patient check-in*
Patient instruction, interview*
Patient preparation*
Premedication*
Room/equipment preparation*
PROCEDURE
Examination/procedure*
Handling of biopsy and other specimens
POSTPROCEDURE
Patient recovery
Monitoring*
Postprocedure instructions/scheduling*
Discharge*
Room/equipment cleaning, turn-around*
Charting/report generation
Written chart notation
Report dictation
Report typing
Review/signature of typed report
Process report to chart
File copy of report
Data processing
Billing
* Functions normally performed by GIA.
Functions of the
Gastrointestinal Assistant in the Procedure
Unit
TABLE 51
Functions that may or may not be performed by GIA,

depending on unit.
The length of time required for the endoscopic component of the procedure unit is highly variable and
depends on the type of the procedure, whether any abnormalities are encountered, and the skill of the
endoscopist. Thus, a routine diagnostic upper gastrointestinal endoscopy can be accomplished in 5 to
10 minutes, whereas a complex colonoscopic polypectomy may require a significantly longer period of
time. As a general rule, there is less variation in the time required for the steps in the procedure unit
that are accomplished by the GIA. The time necessary for preparation of instruments and the
procedure room, cleaning, and postprocedure monitoring of the patient is relatively fixed regardless of
the type of procedure performed.
Preprocedure Functions
The principal duties of the GIA prior to a procedure include preparation of the patient, proper
arrangement of the procedure room, and preparation of the endoscopic equipment. In some units, the
GIA may also have a partial role in the scheduling of procedures, a function that is usually the
responsibility of the GIA supervisor or scheduling personnel. The supervisor's role is discussed later in
this chapter. Preparing the patient for endoscopy has several facets that may differ according to the
type of procedure to be performed.
First Contact
If the procedure is to be performed on an ambulatory or outpatient basis, the patient must be brought
to the dressing room and shown the place he or she will occupy before and after the procedure. In
some units, the patient's belongings are placed on a tray beneath the cart or gurney that the patient
occupies before, during, and after the procedure. This practice effectively saves time by eliminating the
need for the patient to move from cart to cart after sedation is given. It also keeps clothes and other
personal belongings with the patient (Figure 51). The GIA can explain what must be done in
preparation for the procedure, such as disrobing and gowning (some patients require assistance).
Although these details are seemingly mundane and unimportant, they offer the GIA an ideal
opportunity to meet with the patient and to establish a rapport.
(179)Figure 51. Preprocedure preparation area and recovery area (Cleveland Clinic). Carts
are equipped with storage trays underneath to hold the patient's belongings. Monitoring
equipment, "sharps" containers, gloves, tissue, basins, and other frequently needed supplies
are found between bedspaces. Side rails on carts can be positioned to provide a flat surface for
starting intravenous lines. A rocking chair, toys, and books are available for pediatric patients.
Whether a patient is an inpatient or an outpatient, it should be verified that all the necessary steps in
the preparation process have been accomplished. For example, when preparation for colonoscopy has
been carried out by the patient, it is useful for the GIA to review the process with the patient to confirm
that cleansing of the colon is adequate for the examination. It is not unusual to discover that the patient
has encountered difficulty with some aspect of the preparation. An alert GIA can sometimes salvage
this situation with additional enemas and thereby avoid loss of valuable time.
Medical History
It is of value to inquire about any recent change in the patient's overall medical status. As a general
rule, this line of questioning can be based on chart notes and should focus on the patient's
cardiovascular and pulmonary status, any change in gastrointestinal symptoms, and the use of
medication. Recent changes in gastrointestinal symptoms such as the onset of protracted vomiting,
worsening or prolonged abdominal pain, and the appearance of bloody diarrhea sometimes
necessitate changes in the investigative plan.
Medications
The use of medications prior to a procedure is a common area of confusion for patients. Essential
drugs may have been discontinued in the mistaken belief that this was a necessary part of fasting.
Conversely, the patient may have continued the use of an agent that she or he did not consider to be
medication, for example, a liquid antacid. Some patients continue to take iron preparations orally
during the standard preparation for colonoscopy, which often compromises the preparation of the
colon.
The GIA should be particularly careful with the insulin-dependent diabetic patient. In our units, we
usually ask the diabetic patient to administer half the morning dose of insulin at the usual time.
Procedures are scheduled early in the day for these patients, and the second half of the usual insulin
dose is given along with a morning meal after the patient recovers from the procedure. However, errors
can still occur with this simple plan, especially in patients on a stable dosage who have a consistent
routine for insulin administration.
The GIA should also ask about the use of sedative and hypnotic drugs, as this may have a bearing on
the selection of medications and their dosages for the procedure. The anxiety-prone patient may find it
necessary to take such a drug before coming to the endoscopy unit. Chronic users of this type of
medication may also have a tolerance to some agents. Finally, it is essential to ask about any drug
allergies, especially if antibiotic prophylaxis will be used for the procedure.
Description of the Endoscopic Procedure
In our units, all patients receive a written description of the procedure they are to undergo. Ideally, they
receive this far enough in advance so that they can express any concerns or questions that arise after
studying the written outline. The printed material contains information on the goals of the study, steps
the patient must take to prepare for the procedure, a general description of the way in which the
endoscopy will be performed, and some basic facts concerning complications including signs and
symptoms along with specific instructions should the patient encounter symptoms of a complication
after discharge. This type of instruction serves as a point of reference for the patient and often fills in
gaps in the patient's memory regarding the physician's instructions. Although the printed description is
necessary and valuable, it is impersonal and may strike some patients as legalistic. It seldom offers the
patient a sense of confidence or reassurance. Regardless of how well it is written, it is unrealistic to
believe that all the information it contains will be correctly interpreted by every patient. Therefore, it can
never totally substitute for an informed consent or serve as the sole vehicle for conveying instructions
and information.
In some units, the responsibility for the formal explanation of the endoscopic procedure is delegated to
the GIA. However, the primary responsibility for advising the patient of the indication(s), methodology,
and possible complications of the endoscopic procedure remains with the endoscopist. Patients,
however, are sometimes reluctant to discuss with the physician concerns and questions that they may
consider trivial or unimportant, and often questions arise in their minds some time after contact with the
physician. For some patients, it may be necessary for signifiant points in the description to be repeated,
especially when the information has become distorted by anxiety. The GIA plays an essential role in
this process. The GIA is usually the first person with a medical background that a patient encounters
immediately before a procedure. At this time, the sense of uneasiness is greatest and the patient is
most likely to ask questions. Furthermore, the patient often recognizes the GIA not only as someone
with knowledge of the endoscopy procedure but also as someone to relate to in a more personal and
direct way than may occur in the traditional doctor-patient relationship.
The patient's most common questions and predictable comments can be anticipated by the
experienced GIA"Will it hurt?" "Is it dangerous?" "Will I gag?" "Will I be able to breathe?" "I don't like
needles!" "How long will it take?" The GIA should answer questions forthrightly. However, because of
the natural human tendency to offer support to a person in emotional distress, the GIA must be careful
not to offer excessive reassurance. The GIA should never attempt to completely dismiss from the
patient's mind any of the potentially unpleasant parts or aspects of a procedure and, equally important,
should not overly stress these factors.
Psychologic Preparation
Almost all patients are apprehensive to some degree. Many patients control this well, but in others
there is overt fear and aversion. The way this is dealt with can influence the course and outcome of the
procedure. If the patient is at ease, the procedure will be better tolerated, the endoscopist will have
adequate time to complete it, and smaller amounts of sedative drugs will be required. Although the
value of this psychologic preparation in addition to the pharmacologic preparation is easily recognized,
the emotional needs of patients may be difficult to resolve. Success is in some measure dependent on
the natural attributes and personality of the GIA. The patient would not permit the procedure under
normal circumstances if she or he did not regard the physician as a benefactor. But sometimes the
patient is not always at ease and comfortable with the physician as the "authority figure" during the
procedure. In contrast, the GIA is in a less formal, more personal position to offer the type of
reassurance that leads to the patient's emotional confidence in all the personnel involved in the
procedure. The patient needs to know that her or his welfare is first and foremost. The simple phrase "I
am ... , I will be with you through the procedure" can be as effective as any drug, the key words being I,
with, and you.
Premedication
Sedative drugs are usually administered before endoscopic procedures, although this varies in extent
from country to country. The two most widespread classes in use are narcotic and benzodiazepine
agents. These may be given in a variety of ways, although intravenous administration is probably
preferred. Anticholinergic drugs are rarely used for most routine endoscopic procedures with the
exception of laparoscopy. In almost all cases, the dosage of medication is calculated according to the
patient's age, weight, and general medical condition, and it is administered slowly while the patient's
response is observed. Endoscopists also differ with respect to methods of intravenous administration.
For a relatively short diagnostic procedure with a motivated patient, some endos-copists administer no
sedation at all or only small dose(s) by direct intravenous injection. For a more prolonged procedure in
which some discomfort is expected, the endoscopist may prefer to maintain constant intravenous
access, so that additional amounts of sedative drugs may be administered. Generally, a heparin lock is
adequate for this purpose, although a slowly running intravenous infusion is sometimes requested,
especially for procedures that involve some risk such as bleeding. In our units, direct intravenous
injection of drugs is performed only by a physician or by a registered nurse under the direct supervision
and observation of a physician.
In some units, sedative and other drugs are administered prior to bringing the patient to the endoscopy
room. Generally, this requires a cart-exchange system between the dressing room and the procedure
room; that is, the patient remains on one cart during all steps of the procedure. In most units,
intravenous sedation is given in the procedure room immediately before a procedure.
The GIA has primary responsibility for the drugs used in the endoscopy unit. This includes the
maintenance of a proper inventory of agents commonly used and the handling of controlled substances
as prescribed by law. In addition, the GIA is also responsible for preparing correct dosages of drugs for
use during procedures.
Preparation of the Procedure Room and Equipment
An important duty of the GIA is to prepare the procedure room for each endoscopy. The room must be
maintained in a clean and orderly fashion. Surfaces that may be contaminated are cleaned with an
appropriate disinfecting agent between cases. The need for various items of equipment is anticipated.
Our scheduling systems are based on room availability. The master schedule sheet can be easily
separated into individual sections that outline activity within each room on a daily basis. A copy of this
is placed in each procedure room and the dressing room station. A master schedule is kept by the GIA
supervisor. Using this schedule, the GIA working in each room can plan ahead with respect to the
equipment needed for each procedure and, since the physician's name is also listed on the schedule,
can also anticipate the special preferences of each endoscopist. When appropriate, marginal notes
and remarks are included on the schedule to indicate special requirements, especially with respect to
therapeutic endoscopy.
The ability to organize all equipment and to keep the procedure room and unit functioning smoothly is
based on a thorough understanding of endoscopic procedures, and this can only be learned by actual
practice. Thus, it is impractical to list all the items of equipment needed for the many different
diagnostic and therapeutic endoscopic procedures. Moreover, there are differences among institutions
in the way procedures are performed, and these modifications often necessitate additional accessories
and pieces of equipment.
Functions During Endoscopy
It is impossible to address every variation in the activity of the GIA that may be encountered during
every kind of procedure. However, it is appropriate to comment on the more prominent aspects of the
functions of the GIA during the main types of endoscopic procedures.
Esophagogastroduodenoscopy
It is necessary for an ambulatory patient to only partially disrobe for esophagogastroduodenoscopy
(EGD). It is ordinarily sufficient to remove outer garments above the waist, but not undergarments, and
then put on a hospital gown. Eye glasses and dentures should be removed.
Some but not all endoscopists prefer topical pharyngeal anesthesia (see Chapter 38: Technique of
Upper Gastrointestinal Endoscopy). Several methods of accomplishing this include the GIA's
administering an anesthetic spray or having the patient gargle with a viscous anesthetic agent. It is
advisable to inquire again about any drug allergies the patient may have before any medication is
administered in the procedure room.
The patient is then positioned on the examination table. One or two pillows should be available. The
patient usually first assumes a supine position if sedative drugs are to be given, although she or he
may be placed directly in the left decubitus position for endoscopy if there is convenient intravenous
access. Once the medication is given, the GIA places the patient in the proper position for the
procedure. For upper gastrointestinal endoscopy, the patient is generally on the left side, knees flexed,
left arm under pillow, right arm on the right side, with head and chin tilted slightly downward toward the
chest.
Except in the case of edentulous patients, a mouthguard is placed between the patient's teeth to
protect both the instrument and the teeth during the examination. One of the GIA's responsibilities is to
ensure that the mouthguard remains in proper position. It may fall out or be pushed out by the patient
during the procedure. Some endoscopists prefer to lubricate the insertion tube of the endoscope, and a
supply of lubricant should be at hand. Because the insertion tube also becomes coated with secretions,
gastric juice, and mucus during the procedure, it may become slippery and difficult to handle.
Therefore, a few gauze sponges are kept on a washcloth on the pillow near the patient's head for the
endoscopist to use to grasp the instrument.
Most endoscopists pass the endoscope under direct observation without inserting a finger into the
patient's mouth (see Chapter 38: Technique of Upper Gastrointestial Endoscopy). During direct vision
passage of the endoscope through the pharynx and cricopharyngeus, the GIA must keep the patient's
head in proper position, which is a slight flexion of the neck. There is a natural tendency for patients to
extend the neck and to move away from the instrument. Gentle words of reassurance and
encouragement offered by the GIA are frequently helpful as the instrument is being passed. Just as the
instrument enters the esophagus, a disposable oral suction catheter can be placed in the left side of
the patient's mouth next to the mouthguard. This has a blunt tip and can be bent to fit comfortably in
the corner of the mouth. This will allow removal of any regurgitated fluid and oral secretions.
During the endoscopic procedure, the GIA must keep the patient in proper position on the examination
table and must monitor cardiopulmonary status, especially respiration. Some endoscopists prefer that
the lights be dimmed in the room, others do not. The room should never be so dark, however, that it
becomes difficult to monitor the patient. The GIA must also ensure that the mouthguard and suction
device remain in proper position and must be alert for unexpected vomiting, as aspiration is a serious
potential complication of EGD.
The GIA has several functions with regard to the various types of specimens collected during the
procedure. These include tissues (biopsies), cells (cytology specimens), and secretions or exudate
(e.g., potassium hydroxide preparation). The GIA should prepare the necessary accessories and
equipment and assist in the actual collection of specimens. The GIA usually takes care of the handling
of the specimens including proper labeling of containers and preparation of the necessary forms that
must accompany specimens to the laboratory. The GIA also cleans certain accessories.
When the endoscope is set out for use (or replaced in storage), its appropriate set of accessories
should always be included. It should never be necessary to interrupt a procedure to search for the
correct accessory. In particular, although all endoscopic biopsy forceps have the same basic design,
there are significant differences. The most important are the differences in diameter necessitated by
variations in accessory channel diameter according to the type and manufacture of an endoscope. This
can be a problem, albeit minor, in a large unit where there are many endoscopes. This can be avoided
if each endoscope and its related accessories are stored in the same place. Biopsy forceps may also
differ slightly in some other respects; they may have open (fenestrated) or closed forceps cups, or
there may be a bayonet spike enclosed within the cups. The GIA must determine the type of forceps
preferred by the endoscopist.
Most often, the endoscopist directs the GIA to open and close the biopsy forceps. It is important that
the GIA and endoscopist agree on specific and simple directions for operation of the biopsy forceps
(e.g., "open" and "close"). The forceps should always be checked to be certain that it works properly
before it is given to the endoscopist. The forceps can be broken if it is opened or closed too forcefully.
The endoscope may be damaged as well as the forceps if the latter is opened within the accessory
channel or if it is withdrawn into the channel while it is open. An alert and experienced GIA can
recognize that the forceps is partially or completely within the channel if an unusual opposing force is
encountered when attempting to open the device. This may save expensive repair and replacement
costs and is particularly important when trainees obtain biopsies. Even an experienced endoscopist
may attempt to open the forceps when it is partially within the channel if she or he encounters difficulty
in obtaining a specimen and must work close to a lesion. Closure of the forceps should not be abrupt
or forceful. Brusque closure sometimes causes the forceps cups to recoil from a firm lesion without
obtaining a specimen. An experienced GIA can often appreciate that a lesion is unusually firm or hard
as the forceps is closed; this tactile perception is frequently associated with malignancy. As the forceps
is removed from the accessory channel of the instrument, simultaneous wiping with a gauze pad will
prevent splashing and dripping of fluid. This also cleans and dries the coil somewhat and makes it
easier to handle.
The GIA is also responsible for proper handling of tissue and cytologic specimens. The methods of
doing this should be specified by the pathology and cytology laboratories. These requirements vary
from hospital to hospital. In the Cleveland Clinic endoscopy unit, biopsy specimens are gently teased
out of the forceps cup and on to a small disk of filter paper (one paper disk per biopsy) using a
toothpick. The paper and specimen are then placed in a bottle containing Hollande's fixative solution.
No attempt is made to orient the specimen, as most endoscopic biopsy specimens are small and
attempts to rearrange the specimen on the paper damage the tissue. By contrast, pathologists at
University Hospitals of Cleveland prefer that the specimen for biopsy be simply dropped into a small
bottle of fixative, which is accomplished by opening the forceps within the solution and gently moving it
back and forth several times.
Proper procedures should always be observed in handling bodily tissues and fluids, as these are
potentially infectious.
Probably the second most commonly used accessory is the sheathed cytology brush. In some cases,
this may be used to obtain specimens such as exudate from the esophageal surface to determine the
presence of Candida albicans. However, it is most commonly used to obtain cells from the surface of a
lesion for cytologic study. Since this accessory cannot be adequately cleaned and disinfected, a new
cytology brush should be used for each procedure and then discarded, Furthermore, reuse may permit
malignant cells to be transferred from one patient's sample to another's. Commercially available
disposable cytology brushes are relatively inexpensive.
Cytologic specimens obtained by brushing are placed on a microscope slide by simply brushing the
specimen on to the glass. With a good specimen, there is usually enough material for two slides. The
slide is then placed in a fixative solution (e.g., Carnoy's solution).
The GIA is usually responsible for proper identification of the biopsy specimens. In some units, the
endoscopist may direct that the GIA complete the pathology request forms. This obligation should
never be taken lightly, since mislabeling or confusing specimens from one patient with those from
another may have serious consequences. Therefore, the procedures and methods for handling
specimens should be firmly established and rigidly adhered to in every unit.
Sclerotherapy of esophageal varices is a therapeutic procedure commonly performed in conjunction
with EGD. Injection of the sclerosant solution is almost always performed by the GIA and requires
precise communication between the GIA and the endoscopist. The directions given by the endoscopist
should be unambiguous, and a single set of terms should be used. The GIA reports the volume of
sclerosant injected in increments of 0.5 ml to allow the endoscopist to control the rate and volume of
the injection. Injections may be intravariceal or paravariceal, according to the technique of the
endoscopist (see Chapter 32: Technique of Endoscopic Sclerotherapy). When intravariceal technique
is employed, it is usually possible for an experienced GIA to recognize that the needle is not within a
vessel or that the vessel is thrombosed, according to the degree of resistance encountered while
injecting. A 10-ml syringe is the easiest to use, since it contains enough sclerosant for several
injections and it allows for relatively precise control of the rate and volume of injections. Since most
sclerosants are injurious to the eyes, care must be taken that the syringe not be abruptly disconnected
from the injector needle. If this should occur, the sclerosant may be sprayed into the eyes of the
physician, patient, or GIA. To prevent this, it is advisable to wrap a gauze sponge around the
connection while injecting. The total volume used during sclerotherapy and the number of injections
are recorded by the GIA.
Certain complications are possible with sclerotherapy. Those that may occur suddenly during the
procedure include the vomiting and aspiration of blood and gastric contents during acute hemorrhage
and the onset of active bleeding during the procedure. In rare instances, balloon tamponade has been
necessary for control of acute bleeding. Therefore, it is advisable to keep the necessary equipment at
hand.
Colonoscopy
The functions of the GIA during colonoscopy differ somewhat from those attendant to EGD. It may be
necessary, for example, to assist the patient in changing position during the procedure. Additional
lubricant may also be needed to be applied to the insertion tube near the patient's anus intermittently
during the procedure (e.g., a small amount of lubricant may be placed on a gauze pad or other
methods may be used). In addition to the usual monitoring of the patient, it is essential in this
procedure to observe the patient for signs of excess discomfort or pain and evidence of abdominal
distention.
The GIA should be aware that there are substantial variations in colonoscopy technique among
endoscopists (see Chapter 81: Technique of Colonoscopy). The endoscopist may request that the GIA
hold the instrument at the anus as the endoscopist shifts his or her hand from the insertion tube to the
lateral deflection knob. As a general rule, most expert colonoscopists do not permit the GIA to advance
or withdraw the colonoscope. Many endoscopists use external counterpressure on the patient's
abdominal wall during the procedure. Generally, pressure is applied to the mobile segments of the
colon that are on the mesentery (see Chapter 81: Technique of Colonoscopy). There are many
individual variations for this part of the procedure. If external counterpressure is requested, the GIA
should ask the endoscopist to indicate the place and direction to apply pressure on the abdominal wall,
especially if the GIA is not familiar with the endoscopist's methods. The patient should be forewarned
of this maneuver and should be observed for evidence of pain or excessive discomfort. The
suctionaccessory channel of the colonoscope sometimes becomes clogged during colonoscopy. This
can often be cleared by forcing water or air, or both, back through the channel. During this maneuver, it
is necessary that the endoscopist depress the suction valve.
The GIA usually has several important functions in the process of polypectomy. Modern electrically
isolated electrosurgical generators provide a margin of safety for endoscopic polypectomy (see
Chapter 9: Principles of Electrosurgery). In addition, the patient is almost always awake, albeit sedated,
during the procedure, so that it is unlikely that an inadvertent burn will lead to tissue necrosis. But these
built-in safety factors do not reduce the responsibility of the GIA for the overall safety of this procedure.
All connections between the active electrode (snare) and the generator, as well as those for the
dispersive electrode (plate), must be made correctly. This equipment must be inspected periodically to
be sure that the wires and connections are not worn or broken. In addition, the GIA must be certain at
all times that there is a good contact between the plate and the patient, especially if the patient's
position has been changed between successive polypectomies. Braided polypectomy snare wires
should be checked carefully for any small broken wires in the braid. A broken strand can result in stray
electrosurgical currents. These are usually difficult to see but can be found by gently running a gauze
sponge over the open snare loop.
The generator output is usually set by the GIA. The settings should always be verified before
proceeding with polypectomy and never assumed to be correct, especially if the generator is used for a
variety of procedures by different endoscopists. Even if the GIA is thoroughly familiar with an
endoscopist's particular technique, it is always wise to announce the settings before proceeding. It is
best also that the GIA have an understanding with every endoscopist that polypectomy must not begin
until the GIA announces that everything is ready.
Closure of the snare loop is a critical phase of endoscopic polypectomy. Since the GIA and the
endoscopist must cooperate closely in this maneuver, it is essential that the directions that the
endoscopist will use be established and clearly understood by the GIA in advance. As with biopsy
technique, simple directions work best: for example, "close" to close the loop, "open," "open slightly,"
"open full," "open half," to open the loop. The endoscopist may repeat the direction to "close" until the
loop is against the stalk of the polyp. Electrosurgical polypectomy encompasses several variations in
technique (see Chapter 9: Principles of Electrosurgery, and Chapter 85: Polyps and Tumors of the
Colon) that determine how tightly the snare loop must be closed. Excessive force may guillotine the
polyp and result in bleeding, or it may inhibit the sparking necessary to produce electrosurgical cutting.
Inadequate closure may also prevent adequate coagulation or cutting. It is virtually impossible to
describe the correct amount of tension to be applied to the snare wire; the GIA can learn this only
through experience, and this learning process is based entirely on close cooperation and
communication with the endoscopist. As the snare is closed, the GIA can often appreciate that viable
tissue is still present within the loop because the effort to close meets with a certain resistance. It is
also possible to sense the point that the snare cuts through tissue. The degree of closure of the loop
can also be gauged by observing the control handle.
The tasks associated with endoscopic retrograde cholangiopancreatography (ERCP) are similar to
those for EGD with some additions, such as the injection of contrast medium. Therapeutic ERCP
usually requires a greater variety of accessories than most other procedures. Certain ERCP
procedures have a higher risk of complications, depending on the nature of the disease. In some
clinical situations, there is a higher risk of sepsis.
As with biopsy, polypectomy, sclerotherapy, and any cooperative therapeutic procedure, injection of
radiographic contrast material requires precise communication between endoscopist and assistant.
Excessive force with an increased rate and volume of injection usually causes discomfort and in the
pancreatic duct may contribute to postprocedure pancreatitis. Generally, the endoscopist controls the
rate of injection by reference to the fluoroscopic image. The GIA should also watch the filling of the
ductal systems fluoroscopically, although instructions from the endoscopist should primarily direct the
filling procedure. Some special types of ERCP catheters may be used such as those with a tapered tip
and those with a small metal tip for cannulation of a stenotic papilla or the minor papilla (see Chapter
57: Technique of Endoscopic Retrograde Cholangiopancreatography).
Aseptic technique is mandatory in ERCP, since in some clinical situations, contrast medium may be
injected into a closed space such as a pancreatic pseudocyst or beyond a tight stricture in the bile duct.
The introduction of bacteria into a closed space without proper drainage can lead to septicemia and
abscess formation. One of the major sources of contamination is the light source water bottle. Because
of this danger, water bottles should be sterilized and changed between procedures. All catheters and
other accessories are gas sterilized. Although it is not possible to make ERCP a sterile procedure,
measures such as these are known to reduce the incidence of serious infection (see Chapter 8:
Disinfection of Endoscopes and Accessories).
Endoscopic sphincterotomy (ES), along with its ancillary therapeutic maneuvers in the bile and
pancreatic ducts, is one of the most demanding of all therapeutic procedures for both the endoscopist
and the GIA. The papillotome is usually operated by the GIA in response to established instructions. It
is necessary that the GIA be thoroughly familiar with the technique of sphincterotomy. There are also a
number of variations of ES, most of which require special papillotomy devices (see Chapter 60:
Endoscopic Papillotomy). Extraction of bile duct stones may require different devices during a
procedure including various baskets, mechanical lithotriptors, and balloons. Breakage of certain
accessories commonly occurs, so backup equipment must always be available. Biliary stenting
requires that a number of different types of stents be available in different lengths and diameters.
Insertion of a large-diameter prosthesis requires not only an endoscope with a large accessory channel
but also several types of catheters and guidewires (see Chapter 64: Diagnosis and Management of
Malignant Biliary Obstruction). Many other pieces of equipment may be needed during any given
procedure. The GIA must know the names and understand the use of each of these items. As virtually
any accessory may be called for quickly and unexpectedly, the entire complement of devices and
accessories must be well organized and within easy reach.
Pediatric Endoscopy
There are some important differences in the role of the GIA in pediatric endoscopy and in endoscopy in
adults (see Chapter 53: Esophagogastroduodenoscopy in the Pediatric Patient, Chapter 79:
Esophagogastroduodenoscopy in Infants and Children, and Chapter 90: Colonoscopy in the Pediatric
Patient). For example, in most children, endoscopy requires two assistants. One performs those
functions normally associated with adult procedures while the other holds and comforts the child.
Generally, the endoscopes used are the same as those for adults, except that in small children,
smaller-diameter instruments must be used. Dosages of medications differ in some cases.
Certain additional supplies are required when pediatric endoscopy is performed in an endoscopy unit.
These include diapers, small gowns, and toys. CPR also differs slightly in infants and children. For
example, certain items of equipment must be available in smaller sizes.
Charting
The requirements for charting have increased over recent years. In general, baseline vital signs,
allergies, relevant medical history, and current medications should be documented preprocedurally.
Intraprocedurally, vital signs are recorded at least every 15 minutes, medications are recorded as they
are given, and changes in patient status are documented. The use of blood or blood products is
recorded according to institutional policy. Postprocedure documentation includes vital signs, any
unusual event(s) or complication(s), intervention(s), and patient response, as well as patient teaching,
taking into account the patient's level of understanding. All discharge criteria must be met and
documented.
Postprocedure Functions
The GIA has three main tasks after endoscopy: monitoring the patient's recovery, educating the
patient, and cleaning the instruments.
Patient Monitoring
The length of time that a patient must remain under observation depends on the type and dosage of
the sedative drugs administered and the nature of the procedure. Even after a patient recovers from
the medication, it may be necessary to continue observation after therapeutic procedures that carry a
risk of complications. Generally, proper postprocedure monitoring requires determination of vital signs
and observation of the patient for any signs of pain or discomfort. In our units, the GIA is permitted to
discharge patients from the recovery area after specific discharge criteria are met (Table 52).
TABLE 52
Discharge Criteria
1. Patient alert and oriented to time and place or mentation equal to that at admission.
Mobility similar to admission status.
2. Vital signs are obtained prior to the procedure, immediately after the procedure, and
15 and 30 minutes after the procedure completion in the recovery room.
a. Discharge blood pressure
Should not be less than 90 mm Hg, unless accompanied by a physician's note of
explanation.
Should not be more than 30 mm Hg below the preprocedure or baseline* systolic
blood pressure.
b. Discharge pulse rate
Should not be less than 50 bpm or greater than 120 bpm.
Should not be more than 30 bpm different from the preprocedure or baseline*
value.
c. Respiration rate
Should not be less than 12 or greater than 26 per minute.
Should not be more than 8 per minute different from the preprocedure or
baseline* value.
3. Discomfort, nausea, vomiting and dizziness at a minimum. Swallow, cough, and gag
reflexes present.
4. Verbal instructions pertinent to the procedure given to the patient along with a written
copy. Prescriptions given to the patient are reviewed, and the presence of a
responsible adult to accompany the patient is confirmed.
* Baseline value recorded in the hospital record within 3 months of procedure
performance.
Patient Education
Discharge instructions are given by the GIA in many units; this can include a printed sheet that
provides information on possible delayed complications and instructions for the patient should a
complication develop. Whenever possible, another person, such as the patient's spouse, should also
receive instructions.
Cleaning and Disinfection
In addition to cleaning and preparing the room, following procedures the GIA must clean the
endoscope and other equipment used during the procedure. Any disinfecting or sterilization process
will be ineffective if mechanical cleaning of the instrument has not been thorough.
Water and then air should be flushed through the air/water channel of the endoscope immediately after
the procedure, using the special flush valve. Clean water should also be drawn through the suction
channel, followed by room air. Use of transparent suction tubing makes it possible to see the color and
the contents of the suctioned water. The cleaning brush is then passed through the accessory suction
channel(s) of the endoscope and through the universal cord suction channel.
Proper cleaning requires the use of a sink with a fairly large basin. In addition, a rack on the wall above
the sink on which the instrument may be hung is extremely useful. This allows the insertion tube of the
instrument to hang down into the basin (see "Procedure Room Organization," later in this chapter).
Gloves and an appropriate gown or jacket should always be worn during the cleaning procedure.
Debris and secretions should be removed from the opening of the accessory channel valve with a
cotton-tipped applicator. The control section of the endoscope should then be cleaned, especially
around the suction and air/water valves, deflection locks, and deflection knobs. A gauze sponge
moistened with enzymatic detergent is usually adequate for this purpose.
Disinfection of endoscopes and accessories is discussed in Chapter 2: Flexible Endoscope
Technology: The Fiberoptic Endoscope, and Chapter 8: Disinfection of Endoscopes and Accessories.
GIA Productivity and Staffing Level

Procedure Unit/Weighted Scale
Although the procedure unit concept considers endoscopic procedures in terms of their similarities, in
fact from the GIA's perspective, endoscopic procedures differ significantly. From a systems
management viewpoint, the major differences among procedures relate to the number of personnel
required, the time required for the endoscopic portion of the procedure, and the time required for
assembling and cleaning needed items of equipment.
The concept of a rating system for procedures based on a weighted scale is presented in Chapter 4:
The Endoscopy Unit. All procedures are converted to a weighted number of simple diagnostic EGDs.
This procedure unit/weighted scale system has many applications with regard to analysis, planning,
and problem solving, but it was developed mainly as a method for assessing productivity.
The output of work in an endoscopy unit (as an entity including all personnel and resources) may be
substantial, but this may not be totally reflected in a simple count of procedures performed over a given
period of time. If an increase in the complexity of procedures has occurred, productivity may actually
be at a much higher level than is suggested by a simple count. The workload of the GIA may have
therefore increased to a greater degree than can be determined from existing data. The subjective
impression by the unit's staff of an increasing workload may be correct, but this can be difficult to
substantiate for those who do not understand the function of an endoscopy unit and the nature of
endoscopic procedures. Therefore, an objective method of analysis more accurately reflects the actual
performance and productivity of a unit.
The GIA plays a central role in the procedure unit/weighted scale system's methodology. As noted, the
GIA has responsibility for more steps in the procedure unit than any other participant, including the
endos-copist. In developing the weighted procedure scale, the time and effort of the GIA were primary
considerations. At the conclusion of every procedure, a weighted scale value is calculated by the GIA
using a set of guidelines (see Chapter 4: The Endoscopy Unit). The assistant can also adjust the value
within certain limits for uncommonly long and difficult procedures or those that require an unusual
amount of equipment.
Staffing
From the standpoint of productivity, especially cost effectiveness, it is of great importance to know the
actual number of GIAs that an endoscopy unit requires. Although equipment can be expensive, the
greater cost factor over the course of time is usually the payroll. The basic problem in determining the
number of GIAs required is an imbalance in the amount of time required of the GIA, the endoscopist,
and the procedure room to accomplish one procedure unit. In terms of procedure units, one GIA is not
equivalent to one endoscopist or one procedure room.
The maximum number of procedure units that one GIA can produce in a given time can be calculated.
This will always be fewer than the maximum capability of one endoscopist, and it is always less than
the maximum capacity of one procedure room, this being due principally to the need to monitor
patients after procedures. A GIA who assists with the endoscopic portion of a procedure may not be
the one who performs postprocedure functions such as cleaning and patient monitoring. Nevertheless,
these tasks require GIA time, a factor that must be considered. Other adjustments must be included in
calculating maximum productivity such as vacation days, conferences, and the expected efficiency for
the unit as a whole.
Expected efficiency is difficult to analyze. Individuals almost always assume that they function at a level
approaching 100%, but it is difficult to exceed the maximum possible efficiency of the endoscopy unit
as a whole. No unit functions at 100% efficiency, since this requires that every procedure be performed
with virtually no delay and that all equipment and personnel be fully occupied at all times. A more
attainable goal would be about 70% of the theoretical maximum capacity. This would also, therefore,
be a reasonable expectation for GIA productivity. Taking into account these factors, a reasonable
expectation for the productivity of one GIA is 1000 procedure units per year.
To determine the required level of GIA staffing for a unit, however, it is also necessary to know the
maximum capacity (in terms of procedure units) of the available procedure rooms and the number of
full-time equivalent endoscopists working in the unit along with the maximum number of procedure
units that each full-time equivalent endoscopist can produce. It is possible to calculate the equivalent
number of full-time endoscopists based on estimates of the part-time (or in a few cases the full-time)
contributions of each staff member. This is usually not equivalent to the number of physicians who
perform endoscopy, since most physicians are not full-time endoscopists and perform procedures only
during a portion of a working day or week. These various calculations indicate that there should be a
ratio of about 1.5 GIAs per procedure room or 1.8 GIAs per equivalent full-time endoscopist.
The final step in calculating the level of GIA staffing required is to determine whether the number of
endos-copists or number of procedure rooms will be the limiting factor on productivity. If the number of
full-time endoscopist equivalents exceeds the capacity of all of the unit's procedure rooms, then GIA
staffing should be based on the number of rooms available. Conversely, if the limiting factor is the
number of full-time equivalent endoscopists, GIA staffing should be based on the number of equivalent
full-time endoscopists available. In practice, other factors must also be considered, such as the
demand for procedures.
Many small, seemingly unimportant tasks must be performed in an endoscopy unit. Many of these
tasks pertain to supply and logistics (see later in this chapter). Time and effort are required to keep a
unit clean and orderly. The morale of the GIA staff requires that time be set aside for discussions,
in-service meetings pertaining to new procedures and equipment, and continuing education. Rigid
calculations of the required staffing level often ignore these less tangible, less noticeable, but
nevertheless important activities of the GIA. In the long run, the efficiency and the quality of the work
produced by a unit will be greater if some estimate of the time and personnel needed for these tasks is
included when determining the level of staffing appropriate for a unit.
In almost any endoscopy unit, it is necessary that some procedures be performed away from the unit.
Some of the reasons for this are the type of practice (e.g., a large number of procedures for acute
hemorrhage), the need to share expensive equipment such as a laser with other departments, and the
need to have the services of a radiologist and an x-ray suite available for certain procedures such as
ERCP. The types and quantity of procedures performed away from the endoscopy unit have a bearing
on the number of assistants needed. Transporting equipment to and from the endoscopy unit is always
time-consuming. In a large medical center, travel time may be significant. When two or more off-unit
sites are used routinely, travel time during a day can be equivalent to two or three procedure units. For
complex procedures, such as therapeutic biliary endoscopy, the variety and amount of equipment in
itself would prohibit transportnot just the fact that x-ray equipment is often available only in a
radiology department. Although a full complement of equipment could be maintained in the radiology
suite, this would provide only a partial solution. Certain useful pieces of equipment, such as a light
source, would be unemployed for significant periods of time. Consequently, the GIA has the problems
of cleaning, storing, and restocking supplies as they are consumed and refurbishing worn equipment at
these additional sites. From a managerial viewpoint, this approach means in essence that any
part-time procedure rooms accompanied by a large number of procedures performed away from the
unit require additional GIA staffing.
Another factor to consider when staffing a unit is the need to match the procedure to the skill level of
the GIA. Nurse Practice Acts vary from state to state, and it is therefore important to be familiar with
restrictions for licensed practical nurses and other nonnursing personnel. Temporary or "float"
personnel who work in the unit infrequently are best utilized in relatively simple diagnostic procedures.
The GIA Supervisor

Just as the growth of gastrointestinal endoscopy has brought forth the position of GIA, so the growth of
the endoscopy unit has brought about the position of GIA supervisor. Sustained increases in the
numbers and complexity of procedures have also resulted in a substantial increase in the staff of most
units. During the early development of endoscopy, a single individual usually accomplished most of the
GIA tasks entailed in a single procedure unit. In the modern unit, there is often a substantial division of
labor because of the greater numbers of procedures. Thus, one GIA may assist with the endoscopic
portion of a procedure, another is responsible for instrument cleaning and preparation, and still another
is concerned only with postprocedure patient monitoring. Various assignments are often rotated among
the GIA staff. Because of the intricacy of new therapeutic techniques, more than one GIA is often
needed for certain procedures. An operation of such complexity has a high potential for disorder and
confusion, a point strongly in favor of the need for a specific supervisor of GIA personnel.
Extensive experience in gastrointestinal endoscopy and a thorough knowledge of all procedures are
major qualifications for the position of GIA supervisor. However, a comprehensive background in
endoscopy is only one of the required qualifications. Such an individual occupies a pivotal position in
the organizational structure of the unit. In this role, the GIA supervisor is directly and most immediately
involved in the allocation of the resources of personnel, equipment, and procedure rooms. This central
role obligates the supervisor to participate in the day-to-day scheduling of procedures. The GIA
supervisor also shares responsibility for meeting the endoscopy unit's standards of quality, for
maintaining plant and equipment, for training GIAs, for developing and evaluating new procedures and
equipment, and for tracking consumable supplies. The GIA supervisor directs the activities of the GIA
staff throughout the working day and also interacts with many people who support the operation of the
unit but who are not staff members, such as manufacturer's representatives, biomedical engineers,
radiation safety officers, and pharmacists. Thus, it is highly desirable that the GIA supervisor have
managerial and organizational skills in addition to a total grasp of the field of gastrointestinal
endoscopy.
The GIA supervisor is frequently the representative of the unit and must interact with many other
groups within an institution. The supervisor must remain abreast of changes in standards of practice
through organizations such as the Society of Gastroenterology Nurses and Associates, Association of
Practitioners of Infection Control, and American Operating Room Nurses' Association. The GIA
supervisor should play an active role in the quality improvement program for the endoscopy unit and
has significant responsibility for the implementation of guidelines and standards as promulgated by the
Joint Commission on Accreditation of Healthcare Organizations.
It is mandatory that the importance of the GIA supervisor's many functions be clearly recognized and
that these functions be considered apart from the activities of other GIA personnel. The GIA supervisor
and the physician head of the unit form the nucleus of the unit's management team. It is unrealistic to
expect that a single individual can accomplish these managerial tasks and still perform the more usual
duties of a GIA. Paradoxically, therefore, the GIA supervisor must become less involved with the actual
steps of the procedure unit.
Scheduling
Endoscopy unit efficiency requires a systematic approach to the scheduling of procedures. Certain
aspects of this are discussed in Chapter 4: The Endoscopy Unit. As a guiding principle, the design of
any system should be based on the most restrictive factor. This may be any one or more of the
following: procedure rooms, endos-copists, assistants, or instruments. In our units, the schedule is
based on the procedure rooms. The scheduling system will also reflect a basic philosophy of room
utilizationthat is, to what extent a room can accommodate various procedures. The complexity of a
scheduling process increases in proportion to the degree to which procedure rooms are
multifunctional. Multifunctional rooms increase efficiency and flexibility, but they also demand more
sophisticated scheduling methods. There is usually a mixture of multipurpose and dedicated or
semidedicated rooms in every unit.
Virtually no scheduling system can anticipate the minute-to-minute changes in the daily activity of a
large endoscopy unit unless a superficial orderliness is achieved at the expense of productivity. Ways
in which the daily activities can be disrupted include unanticipated delays in patient preparation,
last-minute cancellations, mistakes in scheduling, procedures prolonged beyond the allotted time,
complications, equipment breakdown, urgent and emergency requests for service, absence of staff
members owing to illness, and delays in starting procedures for any number of reasons. When there is
a close working relationship and communication between an appointment secretary and the GIA
supervisor, the resources of the unit can be quickly redeployed to adjust the schedule to changing
circumstances on an hour-to-hour basis, the result of which is a flexible, efficient, and productive
system.
Logistics
An incredible amount and variety of equipment are required for the efficient function of an endoscopy
unit. The capital equipment (e.g., endoscopes, light sources, lasers, x-ray machines) is the most
obvious. However, a less evident but no less important category of equipment is that of consumable
supplies, a seemingly endless listfor example, antibiotics, arm boards, aspirin, bandages, batteries,
biliary stents, biopsy channel valves, cleaning solutions of all types, coffee, colon lavage solutions,
cytology brushes, dental suction devices, diapers, emesis basins, enemas, facial tissues, gauze
sponges, glass slides, hospital gowns, intravenous tubing and solutions, laboratory requisition forms,
light source lamps, lubricants, nasobiliary tubes, needles, paper towels, photographic film (various
types), sedative drugs, sheets, silicon lubricants, specimen bottles, suction tubing, syringes, tape,
tongue depressors, toothpicks, toys, videotapes, washcloths, and x-ray monitoring badges. Another
category of equipment that must be replaced periodically is endoscopic accessories, including items
such as snare wires, papillotomes, biopsy forceps, multipolar cautery and heat probes, and stone
extraction balloons. Although much of this seems mundane, it must not be taken for granted. Every
endoscopy unit, if it is to function smoothly, must have an inventory-supply system, which should also
be the responsibility of the GIA supervisor. The establishment of par levels makes maintenance of
appropriate levels of critical items manageable.
Deployment of GIA Staff

Once the number of GIAs needed for a specific unit's function and design is known, it is necessary to
deploy them effectively. The physical design of an endoscopy unit must accommodate the basic steps
of the procedure unit. Since space must be allocated for each fundamental component of the
procedure unit, the floor plan of the unit also locates the work stations of the GIAs.
It is highly inefficient for a GIA who assists at the endoscopic segment of the procedure unit to also
perform postprocedure monitoring, since this would result in a significant lapse of time between
procedures. Furthermore, the simplest possible endoscopy unit must be divided into four rooms or
areas: patient reception, scheduling/secretarial, and procedure room (with cleaning, storage, charting
areas), and patient dressing/recovery area. These factors dictate that, at a minimum, all but the
smallest units with a single procedure room will require at least two GIAs.
As the size of an endoscopy unit increases, many more questions of design must be answered. The
degree to which procedure rooms function independently of one another must be resolved. Even in a
unit with many procedure rooms and a large amount of floor space, the individual procedure rooms can
remain semiautonomous by retaining many of the procedure unit steps. In practical terms, this means
that the procedure room must contain facilities for cleaning and storage as well as all drugs,
accessories, and other paraphernalia required for a variety of procedures. This endoscopy unit
"philosophy" also demands some duplication of equipment. In essence, however, each procedure
room in such a scheme will have its own daily schedule of cases. In general, this system requires one
GIA for each active procedure room. When procedure room utilization approaches 75%, at least one
GIA will be required for each procedure room in the unit.
A design that utilizes an interdependent room concept is another option. In such a system, as many
elements of the procedure unit as possible are removed from the procedure room and performed in
other areas provided for these purposes. A unit design might use a single space for cleaning and
storage of instruments that is central to all of the procedure rooms. To utilize such a plan to maximum
advantage, it is necessary to assign at least one GIA to the cleaning and preparation of endoscopes.
Consolidation of the cleaning and storage functions for groups of two or more procedure rooms within
the unit is another possibility. The advantages of this are space savings and a simplified interchange of
equipment between the associated rooms. Since cleaning and storage for the unit as a whole are not
truly centralized in such a plan, however, it is usually impractical to assign an individual exclusively to
these duties unless the unit is exceptionally large and a GIA can be fully occupied at each of the
cleaning-storage substations.
Preparation, cleaning, and storage of endoscopes and accessories are relatively uncomplicated tasks.
It is possible, therefore, to hire a less highly trained individual to perform only these functions. In such
cases, this activity is usually regarded as a separate job description that is not filled by a GIA. This has
certain advantages and disadvantages. One view is that it is desirable that a GIA be familiar with all
aspects of endoscopy, that expensive equipment is better cared for by individuals who understand and
respect its purposes, and that these tasks will be performed with greater care by someone more
directly involved with the patients for whom the equipment is intended. Although it is essential that this
type of work be performed in a careful and expeditious manner, it is nevertheless relatively routine and
perhaps unexciting. If the activity of a fully qualified GIA is to be limited to this work, then it is best to
rotate this assignment among the GIA staff. The opposite view is that hiring a person with lesser
qualifications for this position is economical and also frees GIAs for activities that are more concerned
with direct patient contact.
The basic task elements of the procedure unit can be subdivided in other ways. Patient preparation for
procedures can also be accomplished outside the procedure room. This method usually requires a cart
exchange system between the procedure rooms and dressing room. Since sedative drugs are
administered before the patient reaches the procedure room, the patient must be attended by a GIA. It
can be efficient in a large unit that uses such a system to assign a GIA exclusively to the patient
exchange function. This function could also include an assisting role in postprocedure observation in
the recovery area in conjunction with another GIA assigned only to monitoring. The patient exchange
GIA could also be responsible for preparation of the procedure roomexcept for instruments and
equipmentbetween cases.
The number of GIAs required in the recovery room depends on the number of patients that can be
accommodated, which in turn should relate to the number of procedure rooms (see Chapter 4: The
Endoscopy Unit). The number of GIAs also depends on the types of procedures being performed. The
length and intensity of postprocedure monitoring are less if a high percentage of procedures are simple
diagnostic EGDs. If, however, there are many outpatient therapeutic procedures that require closer
observation for more extended periods of time, then the workload of the GIA will be greater.
Various types of procedures are performed in most endoscopy units. Many of these procedures will be
considered "routine," whereas others might be termed "special" procedures. Those in the special
category tend to be therapeutic and are usually performed less frequently relative to the number of
routine examinations. This mix will be reflected in the design of the unit, specifically with respect to the
ratio of dedicated rooms for specific procedures to multifunctional rooms. From a managerial
viewpoint, it is important that individual GIAs not become identified with certain procedures. Although it
can be argued that the result of restricting certain procedures to one or a few GIAs improves the quality
of the procedure, this tends to segregate the GIAs into groups. The implication of this is a hierarchy
within the GIA staffthat some members have greater importance than others. This managerial
stancethat only certain individuals are capable of performing particular procedurescan be
inefficient and may be a source of friction within the GIA staff. In this same context, it is essential that
GIA assignments be rotated among the endoscopists who work in a unit. Allowing an endoscopist to
identify those GIAs she or he will work with, or permitting GIAs to select endoscopists to work with
according to their preferences, makes it impossible to develop and maintain a team approach. Such a
predicament may be inadvertent and its consequences may only become evident as it impedes growth
of the unit. It must always be the primary goal of the GIA supervisor that the GIA staff function as a
unit.
The attributes of "routine" and "special" procedures are discussed in Chapter 4: The Endoscopy Unit.
Special category procedures are usually therapeutic and often require distinctive, expensive,
one-of-a-kind equipment within the unit. Some special procedures require relatively large amounts of
equipment. These factors increase the magnitude of preparation for the procedure as well as the
amount of postprocedure work of the GIA. Depending on the skill of the endoscopist, special
procedures can be more time-consuming. In addition, they are intrinsically more complex for both the
endoscopist and the GIA. The assistant must often perform a number of tasks, including patient
monitoring, at the same time. Although one GIA is the minimum requirement for a special procedure, in
some cases the difficulty of the procedure, patient safety requirements, and the need to perform a
number of tasks simultaneously justify the presence of a second and sometimes a third GIA. ES,
especially with additional therapeutic maneuvers such as insertion of a biliary stent, is an example of
such a procedure.
GIA Training
In-Service Conferences
Associated with the increase in the scope and complexity of endoscopic procedures is a corresponding
increase in the quantity and intricacy of items of endoscopic equipment. New procedures are
constantly being developed. Another aspect of this change and growth is that the GIA must acquire
new knowledge and skill concurrently. There is considerable value, therefore, in regularly scheduled
in-service conferences.
Method of Training
Much of the present knowledge and skills that the GIA requires is based on practical experience. A
formal preceptor program works well in this setting. Learning to be a GIA has a large psychomotor
component, and therefore, guided hands-on experience is essential.
The fact that the knowledge and skills that qualify a nurse as a GIA reside with relatively few individuals
is inauspicious for the future development of gastrointestinal endoscopy. The method of training GIAs
is essentially that of passing information and experience from one individual to another. Even the
methods and techniques used in training new GIAs, as well as the actual experience of teaching those
who undertake this activity, are known to relatively few nurses. Thus, in many large units, the training
of GIAs is the responsibility of one nurse with a number of years of service and experience. The loss of
this individualthrough retirement, for exampleis equivalent to a substantial loss of ability to train
new GIAs.
The training of GIAs is a stepwise process. It is first necessary to provide an overview, through
observation, of the way in which endoscopy is performed. The student GIA must learn the primary
parts of a gastrointestinal endoscope along with the correct methods of handling this type of
instrument, especially with respect to proper techniques for cleaning and storage.
In our experience, GIA training is best conducted by teaching one procedure at a timeEGD serving
as the initial, basic procedure. This method of instruction is similar to systems for training
gastrointestinal endos-copists. The use of one fundamental procedure as a prototype in instruction
simplifies the learning process. Rather than confronting the trainee with many different endoscopes,
accessories, and procedures, the subject matter is broken down into modules that fit into a natural
progression in which the information and knowledge acquired in earlier stages of training can be built
on and modified as the GIA is introduced to colonoscopy and, later, ERCP. In this respect, it is also
best to introduce therapeutic procedures only after the assistant is thoroughly familiar with basic
diagnostic endoscopy.
In order to properly assist at endoscopy, it is essential that a GIA have a basic understanding of the
way in which endoscopic procedures are performed. This knowledge of the procedure itself should
include the indications and contraindications and, in particular, its major complications and the signs
and symptoms of untoward events. Familiarity with the procedure allows the GIA to be alert to possible
patient discomfort and to recognize potential hazards and difficulties for the endoscopist. Furthermore,
it also allows the GIA to anticipate the need for accessories, other items of equipment, and medication.
A well-designed procedure room supports the performance of endoscopic procedures in a number of
waysfor example, items of equipment, the examination table, work surfaces, storage spaces, and
sink are located according to the role they play in the procedure unit (see Chapter 4: The Endoscopy
Unit). The organization of the procedure room should be presented early in the course of training at a
time when the new GIA is learning about the endoscopy procedure itself. An understanding of the
objectives and actual performance of an endoscopic procedure makes it relatively easy to comprehend
the seemingly complex organization of a procedure room. As the relationship between form and
function of the procedure room becomes more apparent, the new GIA is also better able to understand
and recognize his or her role in the procedure unit.
Equipment
The GIA supervisor plays an essential role in maintaining equipment within the endoscopy unit. The
quality of the procedures performed in the unit, the safety of the patients, and the overall efficiency of
the unit depend substantially on the cleanliness and condition of the equipment. The growth of
endoscopy has also fostered growth in the commercial sector, so that many manufacturers now sell
endoscopic equipment. Although competition is desirable in many respects, it also leads to differences
in price, quality, durability, design, safety, and availability. Consideration must also be given to
provisions for maintenance and compatibility with existing equipment. The choice of equipment can,
therefore, be difficult; proper selection requires a thorough familiarity with available items.
Manufacturers also make improvements in their products, so that equipment must be continuously
reevaluated. The GIA supervisor plays an essential role in this process and collects and synthesizes
the opinions and experience of the unit's staff.
New endoscopic procedures are constantly being developed. Each new technique requires further
skills and knowledge on the part of the GIA. In a large unit, one GIA, often the GIA supervisor but not
necessarily the same individual in each case, can be assigned to assist in the development of new
procedures and to evaluate new equipment. The knowledge and information derived in this process
must then be disseminated to other GIA staff members.
Methods
The methods of experienced GIAs differ in many ways. This applies not only to the specific process of
assisting at procedures but also to the general activities within the endoscopy unit. There are different
systems for cleaning and storage of instruments, for preparation of patients for procedures, for moving
patients to and from the procedure room, for handling emergency situations, and for a great variety of
other functions. It is uncertain whether these differences are important and whether the approach of
one individual to a given procedure is more satisfactory than that of another. Most likely, there are
unique aspects of each experienced GIA's methods that will be useful to others.
The Society of Gastroenterology Nurses and Associates has made progress toward standardized
techniques for GIAs. The purpose in this is not to define recommended or approved methods, either
arbitrarily or by consensus, nor to impose uniformity. Rather, the value of a more standardized
approach lies in the establishment of a frame of reference and guiding principles for future GIAs.
Considering the many methods and techniques used by the GIA, it is appropriate to emphasize only
the most universal aspects.
Procedure Room Organization

The arrangement of the contents of the procedure room follows naturally from the general manner in
which endoscopic procedures are performed. A simple but useful plan of organization for the roomas
discussed in Chapter 4: The Endoscopy Unitis one that provides specific work areas for the GIA and
the endoscopist, with the examination table placed centrally. Ideally, the GIA's area should have a
large countertop work surface as well as adequate storage areas, all within easy reach during a
procedure.
Some features in the layout of a procedure room that pertain to the GIA's functions can be illustrated
by reference to a procedure room (Figure 52). During a procedure, the GIA works in the section of the
room between the examination table and the power column. The power column keeps necessary
equipment conveniently organized in a place that the GIA can reach without leaving the patient (Figure
53). Monitoring equipment, oxygen, suction, biopsy bottles, suction tubing, nasal cannula, irrigation
basin, and other items are found on the power column. A countertop work surface is behind the GIA,
attached to the power column, and another work surface that is the top of a cart is to the GIA's side.
During the procedure, this arrangement provides a natural separation for clean and contaminated
accessories (Figure 53). The cart located at the head of the table also holds various items of
equipment (e.g., electrosurgical generator). The position of this cart allows the GIA to perform certain
tasksworking with specimens, for examplewithout turning her or his back to the patient or the
endos-copist. Cleaning and storage facilities are provided in one area within the room. Instrument
storage cupboards are equipped with an inventory log board that lists the endoscopes assigned to that
cupboard and accounts for instruments that are away for repairs (Figure 54). Behind the power
column and to the left of the endoscope storage cupboard is a double sink with a rack for hanging
instruments while they are being cleaned (Figure 52).
(180)Figure 52. Typical endoscopy procedure room (Cleveland Clinic). An imaginary

diagonal line separates the gastrointestinal assistant's (GIA's) work area from the physician's.
The GIA's equipment is located on the left side of the power column. The endoscope washing
area behind the power column is complete with a double sink, a rack for holding endoscopes,
and an automatic endoscope washer (not shown). Ample cupboard space is provided.
(181)Figure 53. Close-up view of the GIA's work area. Frequently used equipment is stored
or housed on the power column. The GIA rarely has to leave the patient's side to gather
equipment.
(182)Figure 54. Instrument storage cupboard. Endoscopes are stored so that their distal ends
hang freely. Accessories are also stored in this cupboard.
Some Special Equipment

Carts
Regardless of how well an endoscopy unit and a hospital are equipped, certain endoscopic procedures
must be performed away from the unit. This means that virtually every item of equipment needed for
the procedure must be transported. The best method of transferring the functions of the procedure to
another site is by using a cart. The type and amount of equipment needed will vary according to the
nature of the procedure. The procedures include emergency endoscopy for gastrointestinal bleeding
with endoscopic methods of hemostasis, ERCP with ES, laser endoscopy, and bedside
sigmoidoscopy. It is difficult to set up a general purpose cart that would be suitable for all of these
procedures.
A single cart may be equipped and maintained for procedures that are frequently performed away from
the unit. The usual example is the emergency cart for EGD for acute upper gastrointestinal bleeding
(Figure 55). The equipment on the emergency cart used routinely in the Cleveland Clinic unit is listed
in Table 53. A relatively simple cart with open shelves and large wheels is preferable. This allows for
the interchange of assorted items of equipment, so that the cart can be used for various purposes.
Certain items commonly used in the endoscopy unit can also be placed on small carts with wheels, so
that they can be moved from one procedure room to another. It is often useful to place an
electrosurgical generator on such a cart.
TABLE 53
Emergency Cart Equipment
Basin for water

Plastic cups
Mouthguard
Gauze sponges (4" 4")
Gloves
Lubricant
Cytology brush
Cleaning brush
Biopsy forceps
Light source
Fiberoptic panendoscope
Lecturescope ("teaching attachment")
Heater probe (with large and small probes)
Suction machines (2)
Forms
Viscous lidocaine (Xylocaine) spray
50-ml syringe with blunt needle
Gastric lavage kit
Sclerosing needles (2)
Introducer (for accessory channel valve)
Additional items:
TABLE 53
Emergency Cart Equipment
3% sodium tetradecyl sulfate (Sotradecol) (one 2-ml

ampule)
Sterile 95% alcohol
12-ml syringes (6)
Normal saline solution (two 10-ml vials for injection)
(183)Figure 55. Emergency cart. Items maintained on this cart are listed in Table 53.
(Courtesy of Olympus Optical Company, Ltd.; and Allied Healthcare Products.)
Boxes
Certain small and frequently disposable items of equipment are required for specific parts of the
procedure unit. Such items are often used at the same time. An example would be the needles,
tourniquets, bandages, alcohol preparation sponges, and syringes for premedicating patients. One
method of organizing items of equipment that are used together is to put them in small boxes or
baskets.
Chapter 6 Medical Informatics

(184)
(185)
FARID NAFFAH, M.D., M.S.
Hospitals and large clinics have long recognized the utility of computers. Nowadays, it is difficult to
conceive of handling the vast amount of information generated in the delivery of health care without
computers. Technologic advances, the proliferation of hardware and software companies, market
competition, and the resulting decline in cost have all led to greater interest in computer systems.
Furthermore, the initial fear of a novel science and the reluctance to confront unfamiliar methods and
complicated machines are being replaced by the acquisition of personal skills, the confidence of a
growing experience, and the ready availability of qualified professionals.
The advantages of computers are indisputable (Table 61). 1 However, their ubiquity in medical
centers is not a measure of optimum utilization. The diversity of existing equipment (hardware,
operating systems, networking) gives rise to numerous options for the configuration of a system, which
can be confusing. The many criteria include processing power, memory allocation, data distribution,
accessibility, and transfer, as well as cost.2(186) When a platform is found that best serves the
data-processing requirements of an individual enterprise, the choice of software becomes the issue.
Usually, the data to be processed are readily evident or else easily defined. Unfortunately, this is not
true of medicine.
TABLE 61
Advantages of a Database
Compactness (minimizes storage problems)

Speed (optimizes retrieval time)
Currency (immediate availability of new data)
Efficiency (ease of maintenance)
Computerization in Medicine
Administrative
The introduction of computerization to medicine was inspired by its use in the corporate world. Initially,
computers dealt with only administrative aspects of medical care. Within a few years, a large
assortment of software programs became available for the electronic filing of patient demographics,
scheduling, and the tracking of charges and collections. The volume of patients at any particular
institution and the complexity of billing have virtually mandated the use of computers for these
purposes. Even relatively small office practices have come to rely on computers for routine
administrative tasks. To date, these activities still make up the bulk of computer use in medical
practice.
Clinical
The use of computers in medicine, however, has quickly expanded to encompass many other activities
related to clinical practice and research.3(187) These range from simple computations (e.g., drug
dosages) to elaborate statistical packages; from warnings and reminders (e.g., drug interactions) to the
interpretation of clinical tests (e.g., electrocardiography); from the transcription of chart notes ("the
computerized medical record")4(188) to analysis (e.g., quality assurance), discovery of clinical
associations, and the inference of new knowledge (e.g., cancer chemotherapy protocols); from
reference libraries (e.g., Medline) to diagnostic tools and medical decision making (e.g., expert
systems).5,6(189) This set of developments constitutes the world of medical informatics.
The discipline of medical informatics grew from the awareness that the application of computer science
to medicine requires unique constructs and special techniques. In science, knowledge is usually
focused and structured, its elements easily defined, and its interrelationships readily formulated.
Hypotheses can be tested, reasoning is logical, and decision making is methodical. Conclusions are
definite or carry a calculable uncertainty. Conversely, medical knowledge is vast and fragmentary,
disparate and unorganized. Much is unknown, cogent observations may be irrelevant, and
redundancies abound. Complex relationships between elements are incompletely understood and
change as scientific knowledge accrues. Hypotheses derive from personal experience rather than
established facts. Thus, reasoning is tainted by subjectivity, and decision making is largely
probabilistic. Much rests on convention, and conclusions are typically arguable.
In summary, the elements and nuances that constitute medical thinking are diverse and often elusive.
Nevertheless, they must be explicitly stated in order to be incorporated into a computer program. For
these reasons, many software programs that address a single issue or a limited purview have been
very effective. Conversely, the difficulty in formalizing medical thinking has resulted in a comparative
lack of success with respect to broader, more extensive systems.
Gastrointestinal Endoscopy
Many different aspects of endoscopy can benefit from the use of informatics;7(190) few have been
exploited.
The success of an endoscopic practice can be measured by various factors besides competence in
the performance of procedures. Many of these can be enhanced by computers, including
administrative and clinical factors.8(191)
Management of an endoscopy suite begins with the purchase of endoscopic equipment, its distribution
among procedure rooms, and its maintenance and replacement. It involves allocation of room time to
various endoscopists, which itself entails the assignment of nursing or other personnel to procedure
rooms and waiting and recovery areas. Patient scheduling is another complex task, as is billing for
procedures.
In an era when advancing endoscopic technology and ever-increasing regulation have combined to
increase expenses while the health care dollar inexorably shrinks, overhead costs have all but
mandated optimum use of time in the endoscopy unit. Given the many variables in the management of
a large and busy endoscopy suite, this is best achieved by a well-designed computer system. When a
suite consists of only one or two rooms or is sparsely used, optimum management is rarely a concern.
Computerization then becomes unnecessary and is likely burdensome.
At the very least, the clinical aspects of endoscopic practice include preparation of an endoscopic
report and retention of clinical data for patient follow-up. The endoscopic report is frequently intended
as the vehicle for communication with the referring physician. Often it also takes the place of the chart
note. The clinical data constitute the contents of the report. Other important uses for reports and
clinical data, depending on the size and interests of the particular institution, include quality
improvement and clinical research.
For a number of years, the perception in the endoscopic community has been that reporting methods
and data management could be greatly facilitated by the use of informatics. This has led to much
interest in the development of software intended for these purposes. Several programs now exist that
purport to perform many of the intended tasks, but none has gained wide acceptance and none is used
with any consistency. The luster of a new software product typically derives from some novel and
attractive feature such as the use of color and graphics or voice-recognition technology. Initial
enthusiasm quickly fades, however, and the product is abandoned as soon as closer scrutiny reveals
its weaknesses and limitations.
Why Have Software Programs Failed?

The first parameter for success of a computer system is the willingness of its intended clients to use it.
The second, equally important parameter is performance: Are the proposed goals being met, and at
what price?
Physician Resistance
Physician resistance was seen for many years as the principal obstacle to the success of computer
systems in medical practice.9(192) The emergence of a complex technology in a largely uneducated
medical community created a predictable aversion on the part of physicians. That the intricacies of
medical thinking could be processed by a machine was viewed by many as preposterous.10(193)
Perceptions have changed, however, and the period since the mid-1980s has witnessed a proliferation
of computer systems in medical institutions and private offices. Many physicians have become adept at
using them effectively and are eager to exploit their possibilities. The demand for more powerful
systems constantly increases, and expectations outdistance technologic feasibility. Yet no system for
gastrointestinal endoscopy has to date been embraced and applied broadly despite concerted efforts,
by researchers and industry, to develop a suitable product.
Cost
The costs of acquisition, maintenance, and operation have been touted as obstacles to the widespread
acceptance of computer systems.11(194) High price may be a cause for concern (current systems
including hardware and software cost between $20,000 and $50,000) but must be considered in
relation to potential benefits.12(195) Most systems provide, at a minimum, a report-generating
function, that is, the ability to print an endoscopic report based on data entered by the user. This
replaces the conventional report dictated by the endoscopist and typed by a secretary. Unlike the
conventional report, it is available immediately after the procedure. In a busy endoscopic practice,
automated reporting frees secretarial time and may result in savings large enough to offset the entire
cost of the system within the first year or two after purchase. Computer-assisted scheduling, billing,
and management of the endoscopy suite, when available, also contribute to efficiency and save clerical
time, resulting in additional savings.
Several auxiliary benefits provided by most commercial systems can enhance the functioning of an
endoscopic practice, although they do not lend themselves to comparative calculations of cost
effectiveness. For example, medical information is compactly stored and can be retrieved rapidly. This
saves time and space and facilitates patient follow-up. Reports of billing accounts, quality
improvement, and other desired information can be produced periodically, obviating manual
computations. Computer storage of digital endoscopic images preserves their integrity, provides a
safeguard against misplacement, and allows simultaneous cataloging by multiple attributesfor
example, type of lesion, anatomic location, therapeutic maneuvers, date, and patient demographics.
Computer retention also provides the physician with an extensive library of endoscopic images for
research, publication, and education.13(196)
Perception of Utility
To date, the utility of clinical systems remains in doubt. A clinical system is one that handles
information strictly relevant to patient care. In endoscopy, this information consists primarily of
endoscopic findings but also includes all the other facts deemed necessary to the complete
endoscopic report: indication(s) for the procedure, relevant medical history, description of the
procedure (e.g., endoscopist, assistants, type of instrument, accessories, medication, time and length
of the procedure), and complications, if any.
Two fundamental considerations enter into the decision to use a computer system for endoscopy. First,
the system must provide an adequate format for expression of endoscopic observations, that is,
without loss of content and with maximum observance of nuance and appropriate detail. Second, data
entry must be rapid. The time required to enter the substance of a procedure in the system should be
comparable to that spent in dictating a report. Despite the many advantages of informatics, dictation
will prevail as long as it is faster than entering data into a computer. Only after the challenges of
expressivity and speed of data entry have been resolved (and physicians are motivated to use a
computer system) will other benefits of a computer system be probed and appreciated.
Expressivity
For physicians to entrust their narration of clinical events to computers, they must be able to articulate
their thoughts with factual completeness and little deviation from their customary organization and style
of expression. They may find it useful to incorporate descriptive details, not strictly scientific, that impart
a general flavor to the report or convey an impression otherwise difficult to communicate. Insofar as
endoscopy is essentially observation, it entails interobserver variation and personal bias.
Free Text: a Viable Option?

To capture the subtlety of observation and preserve the endoscopist's organization of thoughts and
style of expression, free text would seem to be ideal. This is the strategy adopted by the COSTAR
system used within the Harvard Health Plan.14(197) Text dictated by the physician is transcribed by a
typist directly into the global database. When needed, a "hard copy" can be produced. This process is
identical to that for a standard dictated report, although the medium is electronic rather than paper. The
advantages of a free text system are limited to compactness of storage, accessibility, and availability of
data at remote locations. Any cost savings deriving from automated report generation are forfeited
because full-length text must be transcribed at a keyboard. Furthermore, free text does not lend itself
to systematic computer searches. Searching for the random appearance of certain words or groups of
words is exceedingly timeconsuming and fraught with errors owing to omissions, redundancies, varying
styles of expression, and the widespread use of synonyms in medical parlance. The value and
efficiency of a computer search depend directly on the structure of the stored elements.
Structure: Help or Hindrance?

In computer applications, structure exists at different levels: physical, logical, and user
interface.15(198) Structure at the physical level denotes the actual location of data in computer
memory. It is generally unknown to the user, although it is important from the standpoint of software
development. It determines the speed with which related data can be accessed and therefore plays a
significant role in the efficiency of a program.
Structure at the logical level reflects the cognitive basis of the field, in this case our understanding of
the art and science of endoscopy. Logical structure defines the elements of a body of knowledge and
their interrelationships. It might appear that this is fixed and immutable. This would be true if
endoscopic knowledge were stationary. In actual fact, the logical structure represents only our
conceptual interpretation of the field, this being subject to change as the field evolves.16(199) Although
impervious to the average system user, logical structure is critical for search and analysis of
accumulated data.
Finally, and most obviously, structure at the level of the user interface establishes the format of data
acquisition and display. It is conceived on the basis of various factors: user perception of the relative
importance of data and their utility, the usual sequence in which data are collected, a suitable order for
data viewing, synoptic value, and screen appeal, to mention but a few. For purposes of the present
discussion, the notion of structure is strictly confined to that of the user interface. It is the proposed
framework that the endoscopist must observe and that dictates the manner in which thoughts must be
expressed.
If imposition of structure limits expressivity, it also offers distinct advantages over free text. It reduces
error, suppresses duplication, minimizes oversight, and virtually eliminates verbiage. It also provides
the necessary skeleton for automation of the endoscopic report. The latter systematically molds itself
around data collected according to the existing structure. As discussed, when automated report
generation is satisfactory, cost savings can be considerable. Furthermore, structure at the user
interface ensures standardization of reporting and consistency in reported data. The counterpart of
such uniformity at the logical level is data analysis.17(200)
Unsatisfactory Design
Structure that is too restrictive for the effective expression of ideas is likely to be rejected by users,
whereas lack of structure in the organization of data offsets many of the benefits of
computerization.18(201) The challenge that has faced developers since the first efforts to computerize
endoscopic data and reporting is to maximize expressivity while minimizing reporting time and retaining
the advantages of structure.
Fixed structure imposes an order of data acquisition that is not necessarily that which the endoscopist
might choose or regard as most appropriate for a particular procedure.19(202) For example, some
systems divide an esophagogastroduodenoscopy (EGD) report into three headings, corresponding to
the component organs of the upper gastrointestinal tract (esophagus, stomach, duodenum), and force
the user to follow this sequence.20(203) This offers the advantage of clarity in simple cases but
detracts from important findings in others. The difficulties with this structure are illustrated by several
examples.
If a bleeding duodenal ulcer is found and coagulated, the system requires that findings pertaining to the
esophagus and stomach, however inconsequential, be mentioned first. This is not only clumsy but the
emphasis is also wrongly placed. The following scenario illustrates a different pitfall. Consider an EGD
performed to search for a bleeding source in the upper gastrointestinal tract. Although the examination
findings are negative, the uncomfortable patient retches sufficiently to suffer a Mallory-Weiss tear.
Reporting this under the heading "esophagus" suggests that the tear is the bleeding source rather than
a complication of the procedure.
Localizing a lesion only to an organ is often too vague. Generally, it is important to specify its site with
greater precision. For example, the significance of an esophageal diverticulum depends on its level
within the esophagus. An antral ulcer and one in the body of the stomach have different clinical
implications.
It seems that the problem of specifying location could be resolved by partitioning organs. Unfortunately,
suitable partitions are not readily evident. One set might be convenient for characterizing certain
endoscopic findings but not others. For example, the stomach can be divided histologically into body
and antrum or anatomically into body, fundus, cardia, antrum, and pylorus. The second option is better
for the description of ulcers and masses (e.g., to distinguish a pyloric channel ulcer from one in the
midantrum). However, this option is clumsy for the characterization of gastritis. Type A gastritis would
be listed in three distinct partitions (body, fundus, and cardia) instead of one (body). Similarly, the
esophagus could be arbitrarily divided into thirds (upper, middle, lower). Although adequate for
diverticula, such a scheme is cumbersome for describing esophagitis and localizing complex
esophageal strictures.
Despite the degree of detail achieved, partitioning may still be imprecise. For instance, it may be
important to note that a gastric ulcer is on the lesser or greater curvature. Unfortunately, further
partitioning of organs creates new problems. For one, it is onerous and exacting; it forces a degree of
precision that is only occasionally needed. Moreover, it is likely to introduce errors when lesions span
more than one partition.
Another issue relating to the anatomic partitioning of organs is that patients do not exhibit a uniform
anatomy.21(204) In certain cases, albeit rare, aberrant anatomy is congenital (e.g., pyloric duplication)
or merely an anatomic variant (e.g., pancreas divisum). In the majority, however, anatomic variability is
due to prior surgery. This creates a world of new possibilities. Not only are organs or portions of organs
removed or rearranged (e.g., gastric stapling, right hemicolectomy with ileotransverse anastomosis)
but also organs are interposed that do not ordinarily lend themselves to examination (e.g., jejunum in
Billroth II or Roux-en-Y anastomosis). Furthermore, a new set of pathologic findings becomes germane
(e.g., afferent loop syndrome, pouchitis in colectomy with ileoanal anastomosis). Structured
endoscopic reporting must accommodate all possible cases.
Descriptive reporting of lesions according to a preestablished sequence implies a static picture. Such a
format deals effectively with the majority of endoscopic procedures that are exclusively diagnostic. It
fails, however, to capture the dynamic aspect of endoscopy, which is most evident with therapeutic
maneuvers. Failure to recount the steps of a procedure in order of occurrence may result in loss of
valuable clinical information. It may also produce substantial inaccuracies in case studies intended for
quality assurance and for medical-legal concerns. Consider, for example, post-endoscopic retrograde
cholangiopancreatography (ERCP) pancreatitis. Although pancreatitis is a recognized complication of
ERCP, contributing events remain a matter of some controversy. A critical factor may be the repetitive
cannulation of the pancreatic duct that results from frustrated attempts to enter the biliary system. To
confirm or disprove this theory by retrospective review requires, at the very least, that the order of
ductal cannulation be explicitly stated. Tracking each step becomes quite elaborate, and other
variables must be taken into account (e.g., therapeutic maneuvers).
Voice-Recognition Technology: a Solution?

Despite the appeal of generating an endoscopy report with a computer program, the many drawbacks
of current systems have kept endoscopists from abandoning the dictated report. Fortuitously, the
emergence of voice-recognition technology seems promising in bridging the gap between dictation and
a computer-generated report.22(205)
Voice-recognition technology dispenses with the typist. A voice-recognition board in the computer
possesses a vocabulary from which words spoken by the user are recognized and stored in memory.
Sentences are formed by the system around the words chosen by the user to produce the report.
Unfortunately, systems using early voice-recognition technology have been disappointing. The training
period required of each user is lengthy, and mistakes in word recognition tend to occur frequently.
Recent advances have produced boards that are more accurate, eliminate user training, and possess
a much larger vocabulary.23(206)
Technology for the automatic transcription of freely enunciated sentences belongs to the future.
Voice-recognition systems cannot emulate the stylistic ease of free text dictation. However, they
provide ample latitude for text editing such that the adherence to a strict model for data entry becomes
less critical. Although this strategy may avoid some of the problems inherent in a rigid, structured
format for data collection, it does not address the issue of organization of the stored data. As
mentioned, the structure of the stored data is what determines the value of a computerized database
as a research tool.
Vocabulary
If computerization eliminates verbiage in medical reports, by the same token it requires that the
operating vocabulary be comprehensive and unambiguous. Several problems exist with respect to
verbal expression in medicine and are not unique to gastroenterology. Synonyms must be eliminated to
ensure that searches produce accurate results. For example, Crohn's disease and regional enteritis or
diverticulosis and diverticular disease cannot be used interchangeably.
Identification of lesions is also problematic. Those recognized by their appearance (e.g., ulcers,
diverticula) are usually reported as diagnostic entities. When a diagnosis cannot be ascertained from
the visual aspect of a lesion, the lesion is named by a descriptive term (e.g., polyp). Any elevation
above the mucosal surface may be technically identified as a polyp, even though the lesion may be
hyperplastic, inflammatory, adenomatous, or lipomatous. Some elevations are preferentially termed
nodules; other larger lesions may be designated masses. Nonetheless, the distinction between polyps,
nodules, and masses is blurred. The endoscopist may be unable to decide if a mucosal elevation is a
polyp or a prominent fold and may wish to express this uncertainty in the report. It seems appropriate,
in such cases, that the computerized report have the capacity to reflect such doubt by indicating
different possibilities. An alternative, more objective, approach would be a brief qualitative description
of the abnormality (e.g., erythematous, smooth mucosal elevation). This issue is even more evident
with diffuse lesions that have a similar appearance. It may be impossible, for example, to distinguish
chronic hemorrhagic gastritis from portal hypertensive gastropathy in a cirrhotic patient. Moreover, a
depiction of findings is often necessary, even when the nature of the lesion is evident, in order to
convey its type, extent, and severity. It is important to note not only the location of an ulcer but also its
size, its depth, and the presence of a visible vessel, adherent clot, or active bleeding. Varices may be
barely visible within the distal few centimeters of the esophagus, or they may be large, engorged, and
exhibiting red color signs or ongoing hemorrhage.
A qualitative description must use terminology that is meaningful and standardized. To achieve this
objective, it is necessary not only to classify each and every endoscopic finding according to a uniform
nomenclature but also to formulate precisely the list of attributes required for characterization of each
finding, each with its corresponding set of variables. Only then is it possible to draw conclusions by
comparison and analysis of data with the aid of a computer.24(207)
Endoscopic Images: a Solution?

Electronic endoscopy makes it possible to acquire and store high-resolution images. Photographs
saved on disk can be retrieved for illustration, case comparisons, patient follow-up, education, and
documentation. A suitably designed computer system can also relate an image to a specific finding
(e.g., an ulcer with a visible vessel) or a therapeutic maneuver (e.g., coagulation of that vessel).
The widespread availability of endoscopic photography would seem to obviate precise, albeit tedious
and imperfect, verbal descriptions of lesions. A few photographs obtained during a procedure would
circumvent the need for a comprehensive dictionary of findings and an explicit vocabulary of related
attributes. Indeed, this solution is adequate for day-to-day clinical practice. However, images without
data fall short in providing the basic structure needed for analysis and are not suitable alone for a wide
variety of purposes, including research. Unfortunately, despite advances of image understandinga

branch of artificial intelligenceit is not yet possible to analyze photographs without verbal definitions.
Some preliminary attempts have been made to explore the possibility of image processing in
endoscopy. Although they are encouraging, their scope remains narrow (e.g., the estimation of ulcer
size).25(208) In order to compare findings and conduct analyses to discover clinical associations and
successful therapies, objective cognitive features must be diligently elucidated, defined, itemized, and
evaluated. Any study conducted on a large number of elements possessing multiple attributes and
variables is best done with the assistance of a properly conceived computerized database.
A Computerized Database for Gastrointestinal Endoscopy

A database is a record-keeping system.26(209) It stores and catalogs information and facilitates
retrieval. As in any record-keeping system, the data must be organized so as to best serve the needs
of its intended users. Data organization may be thought of as a problem of optimum indexing, for it is
the indexing of data that permits their rapid retrieval and maximizes their utility. For example, if clinical
information regarding patients is always accessed by the patient's name, a single index, in this case a
name index, suffices. If information must also be accessed by date, patient age, type of procedure, or
diagnosis, then additional indexing is needed. If information is to be accessed by a combination of such
parameters, a data structure that supports complex requests becomes essential.27(210)
The term database is used loosely to refer to three distinct entities. First, a database is the aggregate
of collected data pertaining to some field of interest. Second, the term database may apply to the
conceptual framework devised expressly to accommodate that set of data. The framework defines
data items and their interrelationships and organizes the data according to a theoretical model, known
as the data model. Third, the term database may designate the database management system
(DBMS), a software program that facilitates the collection, storage, and manipulation of the data. A
DBMS is specific to a data model.
Data Models
Three principal data models exist, each with special properties that make it suitable to certain tasks:
network, hierarchical, and relational. When creation of a database is contemplated, it is most important
to make its purpose explicit. This not only indicates which data to collect but also makes evident the
best data model.28(211)
The network model is procedural. It is ideal for repetitious tasks. It is the most efficient of the three
models because items operated on are connected physically such that their simultaneous retrieval is
extremely rapid (Figure 61). For example, it is the model used by airline companies for reservations.
However, it lacks flexibility. Tasks that are not anticipated and planned cannot be executed (Figure
62).
(212)Figure 61. A network database. EGDesophagogastroduodenoscopy.
(213)Figure 62. A network model.

The hierarchical model has been the most popular model for medical databases.29(214) Conceptual
simplicity and efficiency in data retrieval make it attractive. Unfortunately, lack of flexibility limits its
usefulness.
The conceptual simplicity of the hierarchical model is reflected by its simple graphic representation: an
inverted tree. The trunk of the tree, which constitutes the key element, gives rise to branches, each
representing some feature related to the key element. Every branch may generate its own branches,
meant to identify selected properties. This process continues until all relevant data are included (Figure
63). The multilevel structure thus achieved seems particularly well suited for the representation of
medical information in that clinical occurrences (e.g., an endoscopic procedure) typically bring about
observations (e.g., organs examined, lesions), which invite finer characterization (e.g., description of
lesions, photography, biopsies), in turn necessitating specialized interpretation (e.g., histology, image
processing).
(215)Figure 63. A hierarchical database. EGDesophagogastroduodenoscopy.

From an operational standpoint, the main advantage of the hierarchical model lies in efficiency of data
retrieval. All data belonging to a particular record (i.e., an inverted tree) are physically linked in
computer memory, thereby minimizing access time. This requires, however, that data be accessed
according to the prescribed path, that is, by starting at the trunk of the tree. This arrangement is
satisfactory for the clinical follow-up of patients. The hierarchical model is used in the COSTAR system
(see earlier in this chapter) as well as in other databases for gastrointestinal endoscopy.
The advantage with the hierarchical model of rapid data retrieval becomes a drawback, however, if the
database is used as a research tool. Data that reside at deeper levels (farthest branches) of the
hierarchy cannot, as a rule, be accessed directly. Certain data items, judged to be of particular
importance when the database is designed, may be flagged and pointers drawn between them to
facilitate their independent recovery. This strategy is akin to having separate indexes for selected data.
Nonetheless, random queries on the vast majority of items in the hierarchical database remain
problematic. It is very impractical, if not impossible, to compare homologous data belonging to different
records, that is, corresponding branches of different trees. Moreover, the system collapses with
searches that use as parameters multiple attributes deeply embedded in the tree structure. Relatively
straightforward queries are unmanageable, as for example the following: "What is the percentage of
patients under 50 years of age without familial polyposis coli who, having been found to have one or
more adenomatous colonic polyps at a first colonoscopy, are found to have at least one other
adenomatous polyp within the ensuing 3 years?" Here the parameters include patient age, diagnosis,
procedure type, finding, and histology, each occupying a distinct level in the hierarchy. Searches in
hierarchical databases are therefore limited to simple queries. This severely limits their usefulness as
research tools (Figure 64). Attempts to extract new knowledge in gastrointestinal endoscopy from
existing hierarchical databases30(216) have been sparse31,32(217) and their results often
flawed.33(218)
(219)Figure 64. A hierarchical model.

Another serious disadvantage of hierarchical databases is predictable obsolescence. The design of a
hierarchical database is fixed. Once data items are defined and their interrelationships established,
they cannot be changed.34(220) Moreover, new data items cannot be easily incorporated, particularly if
they disrupt the existing structure. Suppose, for example, that location and maneuver are considered
qualifiers of a lesion (a maneuver is any action taken in relation to a lesion such as a biopsy, brushing,
or dilation). A gastric ulcer of which biopsies are obtained is easily expressed in such a structure. Now
suppose that new evidence emerges to suggest that duodenal ulcers associated with the presence of
Helicobacter pylori in the gastric antrum respond preferentially to some unconventional therapy.
Identification of such ulcers would require that biopsies be systematically obtained from the stomach.
To indicate that the biopsies are from the stomach when the ulcer is in the duodenum, the structure
would have to be amended. Indeed, the existing structure would suggest that the biopsies are from the
ulcer itself. Because a hierarchical database is static, it quickly becomes obsolete as knowledge
evolves.
The third and most robust model is the relational model.35(221) Its principal strength lies in the
absence of fixed relationships among data items, a property known as data independence.36(222)
Broadly speaking, each data item is represented by a table (Figure 65). A table is composed of fields,
one of which is the primary key. The primary key is a unique identifier of the data item itself, the
remaining fields being attributes directly related to it and characterizing it. For example, a patient table
could have a unique number ("patient #") as its primary key and first name, last name, sex, race, and
date of birth as its associated fields (Figure 65). An endoscopist table could have endoscopist number
as its primary key and first name, last name, pager number, and phone number as its associated
fields. Tables are independent of each other but can be joined by common fields known as foreign
keys. For example, it would be useful to have a field in the table called "procedure" (Figure 65) to
identify the corresponding endoscopist. Such a field, "endoscopist #", can be added to the procedure
table. It then functions as a foreign key to link that table to the table called "endoscopist" and to obtain
from the latter the desired information. The patient-procedure table serves to itemize the various
procedures undergone by each patient. Similarly, the indications table identifies the indications relevant
to the performance of each procedure.
(223)Figure 65. A relational database.

In the relational model, retrieval of data does not have to follow a preset path, as in the hierarchical
model. Rather, the access path depends on the particular query, and various techniques are used to
optimize that path.37(224) Complex queries can be accommodated without difficulty regardless of the
diversity and number of parameters involved (Figure 66).
(225)Figure 66. A relational model.

Tables and fields can be added to the relational model as medical knowledge advances. The previous
example of H. pylori-related duodenal ulcers serves to illustrate this point. In a properly designed
relational database, the data items "lesion" and "maneuver" are represented by separate tables. It is
then easy to add a field that indicates the location of the maneuver to the maneuver table. The location
of the maneuver is therefore distinguished from that of the lesion. In this case, it would indicate that
biopsies are obtained from the stomach and not the duodenum.
In summary, relational databases are the most flexible and powerful for research, and they offer the
potential for growth. They are also the most challenging from the standpoint of database design in that
they require in-depth understanding of the sphere of medical knowledge that is to be transformed into a
database.
A Global System
It is possible to conceive of several separate systems for gastrointestinal endoscopy, each assuming a
segment of the total work to be accomplished. However, a global system is far more efficient and
economical. Such a system ensures a unique vocabulary, minimizes redundancies and unnecessary
duplication of data, and eliminates errors associated with data transfer, particularly in the case of
updates. It greatly simplifies technical maintenance and facilitates the maturation of the software. It is
imperative, however, as the global system is conceived, that each and every task be considered
separately and guaranteed optimum efficiency.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
Blum B. Clinical Information Systems. New York: Springer-Verlag. 1986.

Bush IE. Trouble with medical computers. Perspect Biol Med 1979;22:60020.
Mishelevich DJ, Hudson BG, Van Slyke D, et al. Implementation of the IBM health care
support/patient care system. In Blum B (ed). Information Systems for Patient Care. New York:
Springer-Verlag, 1984;6282.
Lenhard R, Blum B. An information system for oncology. In Blum B (ed). Information Systems
for Patient Care. New York: Springer-Verlag, 1984;385403.
DeDombal FT, Leaper DJ, Staniland JR, et al. Computer-aided diagnosis of abdominal pain.
BMJ 1972;2:913.
Leaper DJ, Horrocks JC, Staniland JR, et al. Computer-assisted diagnosis of abdominal pain
using "estimates" provided by clinicians. BMJ 1972;4:3504.
Rozen P. Computer assistance in gastroenterology: A review. J Clin Gastroenterol
1982;4:40313.
Kuhn IM, Wiederhold G, Rodnick JE, et al. Automated ambulatory medical record systems in
the US. In Blum B (ed). Information Systems for Patient Care. New York: Springer-Verlag,
1984;199217.
Young DW. What makes doctors use computers? Discussion paper. J R Soc Med
1984;77:6637.
Dowling AF Jr. Do hospital staff interfere with computer system implementation?. HCM Rev
1980;5(4):2332.
Bradley E, Campbell JG. Information system selection: Methods for comparing service
benefits. In Proceedings of the Fifth Annual Symposium on Computer Applications in Medical
Care (CAMC), Washington, DC, November 14, 1981. New York: IEEE, 1981;9036.
Collen MF, Van Brunt EE, Davis LS. Problems of computerization of large computer medical
record systems. Med Inf (Lond) 1976;1:4753.
Drazen EL. Methods for evaluating costs of automated hospital information systems. In
Proceedings of the Fourth Annual Symposium on Computer Applications in Medical Care
(CAMC), Washington, DC, November 25, 1980. Vol I. New York: IEEE, 1980;6738.
Barnett, GO, Justice NS, Somand ME, et al. COSTAR systemA computer-based medical
information system for ambulatory care. Proc IEEE 1979;67:122637.
Blackburn JP. Microchips in medicine. J R Soc Med 1981;74:6445.
Weed LL, and Staff of Promis Laboratory. Representation of medical knowledge and Promis.
In Proceedings of the Second Annual Symposium on Computer Applications in Medical Care
(CAMC), Washington, DC, November 59, 1978. New York: IEEE, 1978;368400.
Williams BT, Mabry F, Alsberg P. Some database management considerations. Fed Proc
1981;40:21537.
Lincoln TL. Medical information science: A joint endeavor. JAMA 1983;249:6102.
Barnett GO, Greenes RA. Interface aspects of a hospital information system. Ann NY Acad
Sci 1969;161:75668.
Landau JA, Norwich KH, Evans SJ. Automatic speech recognitionCan it improve the
man-machine interface in medical expert systems?. Int J Biomed Comput 1989;24:1117.
Naffah F. Computers in the endoscopy unit. Gastroenterol Clin North Am 1992;2:28797.
Paramarajah S. A computer program for oesophago-gastroduodenoscopy records. Singapore
Med J 1986;27(2):1149.
Baker J. Dragon dictate. In Proceedings of Speech Technology. New York: Media
Dimensions, 1989;205.
Jeanty C. The computerized medical record in gastroenterology. Part 2. Morphological
(descriptive) data. Med Inf (Lond) 1978;3(4):2839.
Gaskill HV, Levine BA. Assessment of gastric mucosal ulceration by computerized image
processing. J Surg Res 1987;42:27783.
Wiederhold G. Database technology in health care. J Med Syst 1981;5(3):17596.
27.
28.
Date CJ. An Introduction to Database Systems. Reading, MA: Addison Wesley. 1986.
Barnett GO, Souder DE, Bowie J, et al. MUMPS: A support for medical information systems.
Med Inf (Lond) 1976;1:1839.
29.
Teorey TJ, Fry JP. Design of Database Structures. Englewood Cliffs, NJ: Prentice-Hall. 1982.
30.
Kruss DM. The ASGE database: Computers in the endoscopy unit. Endosc Rev
1987;4:6470.
31. Arrowsmith JB, Gertsman BB, Fleischer DE, et al. Results from the American Society for
Gastrointestinal Endoscopy/U.S. Food and Drug Administration collaborative study on
complication rates and drug use during gastrointestinal endoscopy. Gastrointest Endosc
1991;37:4217.
32. Zimmerman J, Shapiro E, Porath Y, et al. Computer application in a gastrointestinal
endoscopy unit. Endoscopy 1985;17:2079.
33. Naffah F. Databases in gastrointestinal endoscopy [Editorial]. Gastrointest Endosc
1991;37:48890.
34.
Ullman JD. Database and Knowledge-base Systems. Rockville, MD: Computer Science
Press. 1988.
35.
Codd EF. A relational model for large shared data banks. Communications of the ACM
1970;13:37787.
36.
Stonebraker MR. A functional view of data independence. In Rustin R (ed). Proceedings of
the 1974 ACM SIGMOD Workshop on Data Description, Access and Control, May 1974. New
York: ACM Press, 1974;6382.
37.
Selinger PG, Astrahan MM, Chamberlain DD, et al. Access path selection in a relational
database management system. In Bernstein PA (ed). Proceedings of the ACM SIGMOD
Conference, June 1979. New York: ACM Press 1979;2334.
Chapter 7 Electronic Image Management

(226)
(227)
MASAYUKI A. FUJINO, M.D., PH.D.
MASAHIRO IKEDA, M.D., PH.D.
The history of gastroenterologic endoscopy begins with attempts to peer into the gastrointestinal tract,
that is, the first rudimentary efforts to obtain images from within the lumen of the digestive tract.
Because endoscopic images have always been an invaluable asset to the gastroenterologic
endoscopist, the history of endoscopy can be viewed as a history of image acquisition and
documentation.
In the beginning, endoscopic images were documented as drawings made by an artist at the side of
the endoscopist. Attempts at photographic documentation date to the earliest phases of development
of cameras and photographic film, but endoscopic photographs of acceptable quality were not possible
until the 1950s. The development of the gastrocamera by Uji1(228) symbolizes the opening of the era
of gastroenterologic endoscopic photography (see Chapter 1: History of Endoscopy).
The accuracy of image documentation markedly improved with the introduction of photography, but
photographic materials were not easy to manage. Review of images on film strips through a projector,
as with gastrocamera systems, often resulted in damage to the film. Discoloration inevitably occurred
over long periods of time. In general, image degradation was a significant problem. Retrieval of a
desired image on a film strip or a 35 mm slide from among large numbers of images in storage was
not always easy, and the loss of images, particularly of important cases, was most annoying.2,3(229)
Supported by the advances in computers and related fields, electronic management of images became
a promising technology in the 1980s.110(230) In particular, the advent of the electronic endoscope by
Welch-Allyn Inc. (Skaneateles Falls, NY) in 1983 made feasible the electronic management of
endoscopic images.11(231) With the incorporation of the charge-coupled device (CCD) into an
endoscope, images of the digestive tract can be converted instantly to electric signals that are readily
managed by computers (see Chapter 3: Flexible Endoscope Technology: The Video Image
Endoscope). Advances in the development of computer memory media are occurring with amazing
speed.3(232) Image transmission technology is redefining the environment of the endoscopy unit, and
electronic image management is one of the newest developments. It is reasonable to believe that
these revolutionary changes will continue into the foreseeable future.
Technology of Image Management

Image management consists of acquisition and storage of images, together with related information,
and the organization of these data for specific purposes. Images are acquired by endoscopic
instrument systems, the ideal technology for image management. Once captured, images are archived
and organized according to requirements for patient care, research, or education.
The Digital Image

Before we analyze the detailed aspects of image management, it is essential to note the difference
between analog and digital images. This relates to the fact that computers are usually designed to
function by the use of binary coding systems.
An endoscopic image in its original state is a color picture that can be conceptualized as a set of
two-dimensionally arranged picture elements that are continuous in space and continuous in
brightness (i.e., amplitude) of color components. In reality, such an image cannot be represented with
absolute accuracy by numerical elements; rather, they comprise a continuous series of transitions from
point to point. A mathematical representation of this concept is a graph of a line with multiple curves
rather than a series of numbers. An image constituted by these continuous transitions in space and
amplitude is classified as an analog image.
When an endoscopic image is captured electronically by a CCD, the CCD generates, at each pixel, an
electric charge in proportion to the intensity of light that the pixel receives within a specified period of
time. The original image is sampled spatially according to the distribution of pixels on the CCD. The
image is transformed by the CCD into a set of electric signals, discrete in space but continuous in
amplitude (space-discrete signals). These signals are forwarded to the video processor (essentially a
computer), where the amplitude of the electric charge from each pixel is measured and converted to a
number, which is in turn represented by binary digits (i.e., 0 and 1). In this way, the original analog
image (a series of continuous transitions in space and amplitude) is converted to a digital or binary
image (i.e., one described by binary code). The waveform (analog) signal has, in effect, been reduced
to a series of numbers. This analog-to-digital process can never be so detailed that a digital image
becomes an exact duplicate of an analog image (although the digital representation becomes more
precise as the number of pixels increases). Nevertheless, the image that results from the
analog-to-digital conversion is suitable for most practical applications.
Because a digital image is essentially binary code, it has characteristics common to all digital data. For
example, it can be altered by specific software programs and transferred within a computer network
(see later in this chapter) or over telephone lines. It is helpful to conceptualize the digital image as
existing in a computer file comparable to the document file of a word processing computer program
(software). Whereas the latter file typically "contains" several "pages" of text, a digital image file
contains a single image. The term image can therefore denote a single digital image file. Although that
file may be a composite of several different pictures, it is nevertheless a single image file.
The digital image can be conceptualized as a matrix in much the same way as a CCD, that is, a
three-dimensional Cartesian coordinate system specified by three intersecting, perpendicular lines. In
the flat (two-dimensional) plane (defined by x- and y-axes), the CCD and digital images are both
composed of pixels. The location of each pixel in a digital image is therefore specified by two
coordinates (referred to as x and y coordinates). The third or vertical coordinate in the CCD image
matrix is the value of the electric charges that develop in the pixels. In the digital image, however, the
vertical axis is constituted by a binary code that specifies either color values or shades of gray (i.e., a
gray scale image).
The meaning of the term pixel changes depending on the context in which the word is used. Pixel
indicates an element on a CCD as well as a component of a digital image. The former is an actual
physical structure, whereas the latter is an abstract mathematical concept. Pixel can also refer to a
group of phosphor dots on the screen of a color television monitor. The potential for confusion is
compounded by the interrelationships among these several frames of reference. In effect, the number
and planar arrangement of pixels on the CCD determine the number and arrangement in the CCD
image. Because the television monitor must reproduce the image, the number and arrangement of
pixels it contains must conform to the digital image.
The smallest data element that a computer can distinguish is known as a bit. The number of colors (or
shades of gray) a pixel can display is specified by the number of bits it can contain (a factor termed bit
depth). As the number of bits increases, more colors or shades of gray can be represented. Bit depth
is a function of computing power. If the bit depth is one, only two colors or gray shades are possible,
that is, black and white. If the bit depth is increased to four, 16 colors or shades of gray are possible
(i.e., 2 to the fourth exponential power).
A CCD measures light intensity but not wavelength; that is, it detects brightness but not color. Several
ingenious methods have been developed to obtain color images with a CCD. These are discussed in
Chapter 3: Flexible Endoscope Technology: The Video Image Endoscope. A basic knowledge of the
trichromatic theory of color reproduction using the primary colorsred, blue, and greenis essential to
the following discussion of digital color images (see Chapter 3: Flexible Endoscope Technology: The
Video Image Endoscope).
Classification
Digital images are classified according to bit depth as follows: bilevel (1-bit, black and white), gray
scale (usually 8-bit), pseudocolor (usually 4-bit or 8-bit), and true color (24-bit). In a sense, bit depth
specifies the resolution of color or shades of gray in an image, that is, the number of different colors or
shades that can be distinguished.
Bilevel Images
Bilevel images are not germane to endoscopic image management systems because digital x-ray
images are composed of shades of gray, whereas endoscopic images must be rendered in color.
Bilevel images are used in printing, and clever methods have been developed to represent an image
using only the colors black and white.
Gray Scale Images
Gray scale images are composed of shades of gray ranging from absolute white to completely black.
Usually these are 8-bit images, which means that each pixel can display up to 256 shades of gray. The
"brightness" of a pixel is sometimes referred to as its gray value. A gray value of 0 is black; that for
pure white is 256. A gray scale image is analogous to a black-and-white photograph.
Pseudocolor Images
A 4-bit pseudocolor image can represent 16 colors; an 8-bit image of this class offers 256 colors. The
term pseudocolor is used to indicate that the color displayed in a pixel is not a direct function of its
value. Rather, a color is assigned to the pixel value from an index table of colors (referred to as a
palette) in the computer program. Usually, the palette table is stored with the image. Most
image-processing software supports 8-bit pseudocolor images. Commercially available electronic
endoscopic systems are based on 8-bit pseudocolor images.
True Color Images
Images of this class have 24 bits per pixel, that is, 8 bits for each of the three primary colors (see
Chapter 3: Flexible Endoscope Technology: The Video Image Endoscope). Therefore, as many as
16.7 million colors are possible with a true color image. This is well in excess of the number of
separate and distinct colors that can be distinguished by the human eye.
It is possible, within certain limits, to convert a digital image of one class to another of lesser bit depth.
An image of lesser bit depth can be converted to a class with greater depth, but this does not increase
the number of data available in the image; that is, the color and resolution remain the same.
Conversion of images from one class to another can sometimes be advantageous. For example, a
standard, commercially available (color) television camera can be used to capture digital images of
radiographic films. Because radiographs are black-and-white images, conversion of the pseudocolor
8-bit image to a gray scale image enhances detail and contrast.
Spatial Resolution
Spatial resolution is a second attribute of a digital image in addition to bit depth (brightness resolution).
This is specified by the number of pixels and expressed according to the two-dimensional coordinate
system, that is, the number of pixels on the x-axis multiplied by the number on the y-axis. For example,
an image with a resolution of 640 480 pixels contains 307,200 pixels; an image with a resolution of
1024 768 has 786,432 pixels.
Bit depth and spatial resolution are the main determinants of the quality of a digital image. Resolution
defines the amount of detail that can be seen in an image. The higher the resolution (number of
pixels), the better the definition and sharpness of the image.
Digital Versus Analog Images

Most image management systems use both digital and analog images. The digital image is
reconverted to an analog signal by the video processor so that it can be displayed on a standard
television monitor. Because television technology for consumers was originally developed for analog
signals, analog equipment is generally available at a lower price than digital equipment. Because of
financial considerations, analog devices have been used more widely for storage of endoscopic
images as, for example, by means of videotape or analog filing devices.12(233)
Some degradation in image quality usually occurs as a result of loss of information when analog
images are repeatedly stored and recovered for viewing. It is preferable, therefore, that images be
stored digitally.13(234) However, a permanently stored analog image is not degraded while in storage.
Some loss of data occurs when the image is initially stored and some additional loss occurs each time
it is recovered from storage, but the stored image itself remains essentially unchanged.
The difference between digital and analog signals can be illustrated using the analogy of a compact
audio disk (CD) compared with a long-playing phonograph record (LP). Sound is recorded on the LP
as a wave along a circular track; the amplitude of the wave (width of the lateral movement of the
pickup) specifies the intensity of the sound, and pitch is given by the frequency of the wave. On the
CD, amplitude of the sound wave is sampled at regular intervals, and these values are expressed as
binary code (i.e., a representation of any number by means of series of two digits, 0 and 1). Binary
code is represented on the CD itself by the presence or absence of a pit; the pickup of the CD player
detects the latter by means of a laser beam. On the CD, sound is thus recorded as digital signals.
Original sounds are easily reconstructed from binary code. Because the signals are represented by
discrete numbers, the shape of the analog signal wave is of no consequence when the "sound" (binary
code) is reproduced; distortion of the replayed sound (corresponding to the degradation of a
reproduced image) is unlikely to occur. Furthermore, copying the binary code (e.g., from one CD to
another) or long-distance transmission is unlikely to be associated with distortion. Little interference is
produced by noise (stray electric signals inherent to any electronic device); lost data can be recovered;
and methods are available for error correction. Although a sound signal is a function of time and
amplitude, the preceding discussion applies also to an image signal in which space replaces time.
Digitally recorded images are qualitatively superior to analog images. A digital image is also more
suitable for image processing by means of computer programs. Compared with an analog image,
however, a digital image (if uncompressed) requires more time for recording, for transmission, and for
recovery and redisplay from a memory storage device. When a memory medium of a given capacity is
used for recording, the number of uncompressed digital images that can be stored is much less than
the number of analog images; that is, a larger segment of the available memory of the device is
required. Technical problems also occur with storage of multiple digital images in series in order to
display motion.
Because the technology for storage of digital images is more involved, storing an image in digital as
opposed to analog format is more expensive. The cost is expected to be less problematic in the near
future as the use of digital technology becomes more widespread. The price will undoubtedly decrease
with increasing demand; at the current level of technology, however, digital storage of images is more
expensive.
An additional problem with the choice of the digital format for image management is speed of
transmission. For integrated image management, communication is vital, but the speed of digital image
transmissionas, for example, through a standard local area network (LAN)is limited.
The characteristics of analog and digital images are summarized and compared in Table 71.
TABLE 71
Analog Versus Digital Images
SPECIFICATIONS
Image recording
Recording capacity
Image search
Image transmission
Quality of recorded image
Transmitted image
Copied image
ANALOG
Quick (instantaneous with
buffer)
Large (108,000 images/300
mm optical disk)*
Quick (<1 sec/image*
Quick
Fair
Associated with image
degradation
Associated with image
degradation
Good
Fair
Moving image recording

Application for image
processing
MOmagneto optical disk.
* Data from the authors' system.
DIGITAL
Slow (instantaneous with
buffer)
Small (5000 images/130 mm
MO)*
Slow (<3 sec/image)*
Slow
Good
No image degradation
No image degradation
Fair
Good
Image Acquisition
Various types of images are appropriate and suitable for image management. Images acquired with
electronic or video endoscopes are obvious choices. However, images obtained through a fiberscope
can also be converted to a digital system by means of a video converter. However, resolution (the
ability to distinguish fine detail) in images obtained in this fashion is inferior to that of electronic
endoscope images and even that of photographs made by attaching a camera to a fiberscope.
In addition to endoscopic images, relevant images to be archived include ultrasound images obtained
by extracorporeal scanning and endoscopic ultrasonography, conventional radiographs, including plain
radiographs and contrast studies such as ERCP (endoscopic retrograde cholangiopancreatography)
radiographs as well as computed tomography (CT) and magnetic resonance imaging (MRI) scans.
Images in the form of photographs and radiographs can be "digitized" and incorporated into a digital
format storage system by means of a high-resolution image scanner. A complete system might include
images obtained by light or even electron microscopy (of biopsy specimens, for instance). Images of
resected specimens can also be important; macroscopic images can be acquired through a standard
video camera.
Data Formats
Data format is problematic with regard to all aspects of image management. This is somewhat
analogous to the situation that exists with text files generated by word processing software programs.
Text is stored in a computer memory (e.g., a hard drive or floppy disk) in a particular format that is
unique to each word processing program. This means that a text file stored in a format specific to one
program cannot be manipulated without difficulty by another word processing program. By means of a
uniform standard for text characters (i.e., American Standard Code for Information Interchange
[ASCII]), a text file written to a memory device can be read and manipulated by all word processing
programs. However, much of the usefulness that the word processing program adds to a text file is lost
when the digital data are written to memory using the ASCII format.
Formats for writing a digital image (file) to a memory device depend on the process by which the
manufacturer of the software acquired the digital image. In general, each format is unique to the
software and is proprietary in nature. Unlike the ASCII format for text files, no universal common
standard format exists for image files. Converting the format of a text file from that of one word
processing program to another (or to ASCII format) is relatively uncomplicated. Image files can also be
converted from one format to another, but conversion of this type is much more complex than that of
text files. To begin with, the number of data (binary code) in a text file, even a very large one, is
relatively small compared with that of a file containing a color image. Furthermore, exact conversions
between image formats are often technically difficult or impossible because images may become
distorted by the conversion process. The data format problem is compounded by the fact that all
images (e.g., endoscopic, radiographic, ultrasound) to be incorporated into a software-hardware image
management system must be in the same data format.
The type of file format is indicated by a file extension (i.e., a period followed by three letters after the
name of the file; e.g., FILENAME.TIF or FILENAME.TGA). Some commonly used file formats (and
corresponding software) are given in Table 72.
TABLE 72
Examples of File Formats Used for Digital
Images
TIFF
TARGA
BMP
CUT
GIF
IMG
MSP
Tagged information file format: A widely used general-purpose

format developed by Aldus Corporation and Microsoft
Corporation. Specifically designed for image processing (file
extension is .TIF).
Developed for image processing with TIPS software by
TrueVision Corporation. Works with TARGA series of frame
grabbers (extension is .TGA).
Used with Windows and OS/2 (extension is .BMP).
Developed by Media Cybernetics. Works with a program known
as Halo Paint Package (extension is .CUT).
Graphics interchange format: Developed by CompuServe, Inc.
(extension is .GIF).
Developed by Digital Research. Used in certain desktop
publishing programs (extension is .IMG).
Developed by Microsoft for a program called Microsoft Paint
(extension is .MSP).
TABLE 72
Examples of File Formats Used for Digital
Images
PCX
Developed by ZSoft Corporation (extension is .bmp).
The conventional (quick and easy) solution to the data format problem has been to use the format of
video signals as the interface; these consumer product television technology standards are NTSC
(National Television System Committee) for North America and Japan and PAL (phase-alternation line)
for Europe. Although this approach is economical, these standards are for analog signals. For storage
and retrieval, therefore, a digital image must undergo repeated digital-to-analog and analog-to-digital
conversions. Theoretically, this process results in subtle image degradation. Although one such
transfer (as occurs each time the image is recalled from memory) may be of no consequence in terms
of image quality, many such conversions result in an appreciable decrease in quality.
Efforts are under way to develop device-independent standards for diagnostic images; digital imaging
and communications in medicine (DICOM) standards for endoscopic images are being developed in
the United States of America by a joint working group.14(235)
Image Compression
A large amount of computer memory is needed to store an image with a high spatial resolution and bit
depth. For example, an 8-bit pseudocolor image with a spatial resolution of 480 640 requires about
300 kilobytes (K) of memory; a 24-bit true color image with the same spatial resolution requires almost
1 megabyte (MB) of storage capacity. Image "size" in terms of bits needed to represent the image also
influences the speed at which it can be transmitted, as in a FAX message. To deal with these
problems, standard methods of image encoding, also known as compression, have been developed.
Even with compression, however, a color digital image still requires a relatively large memory for
storage.
Image compression is fundamentally a mathematical technique that identifies redundant image data
and replaces them with codes that consist of smaller numbers of data bits. With elimination of
irrelevant and repetitious data, less memory is needed to store the image and speed of transmission
increases.
Several steps are required to encode an image. The first is known as mapping. In this step, the input
data from the pixels are transformed to reduce redundancy. Mapping may be of two types: destructive
(which eliminates varying amounts of data) and nondestructive (in which all data in an image are
preserved). Methods of compression that preserve all data are called lossless; those that do not are
referred to as lossy.
Discrete cosine transform (DCT), a commonly used compression for digital image storage and retrieval
in the Joint Photographic Experts Group (JPEG) format, is an example of destructive mapping. This
method divides an image into blocks where the spatial frequency components are extracted and
compressed by coarsely quantizing the high-frequency components (that slightly affect the image
quality). All methods of destructive (lossy) compression are nonreversible and degrade an image with
each compression.
Run length encoding (RLE) is a commonly used type of nondestructive mapping. This method is based
on the assumption that adjacent pixels frequently have the same brightness or value. RLE assigns a
number pair to each sequence or "run" of pixels that are alike. Typically, the first number of the pair
signifies the brightness of the first pixel in the run (which is equal to that of all of the other pixels in the
run). The second number specifies how many pixels the run contains. RLE is especially suitable for
images with large areas in which a color is constant.
Image Storage
Recording or storage of images is the essential element of image management. Two important
aspects of image storage are the selection of a memory medium and the security of the stored data.
Memory Media
Currently available memory media relevant to image storage are of three types: magnetic, optical, and
electric.
Magnetic memory media include the hard disk, floppy disk, and digital audio tape recorder (DAT). All of
these are rewritable; that is, it is possible to erase existing data. A hard disk consists of a substrate
coated with a thin metallic plating of a magnetic alloy. Characteristics of a hard disk include relatively
large memory capacity (40 to 4280 MB/disk) and high-speed access (less than 10 ms) to data. A floppy
disk is a flexible, lightweight disk made of ferric oxide-coated polyester, 3.5 or 5.25 inches in diameter.
The usual storage capacity is limited (approximately 1.4 MB per formatted disk), but the low price and
convenience of the floppy disk in an office environment are attractive features. DAT can be used for
archival storage.
Currently, three types of optical memory media are available: rewritable (i.e., erasable), write-once (i.e.,
non-rewritable), and read-many-times, or read-only. A magneto optical disk (Figures 71 and 72) is
rewritable and portable (90 to 130 mm in diameter) and has a large memory capacity (128 to 1000
MB/disk, double-sided). The write-once type of optical disk (occasionally abbreviated WORM for
write-once, read-many-times) (Figure 72) is non-rewritable and usually large (300 mm in diameter),
with a large capacity of memory (up to 7 gigabytes/disk, double-sided). The WORM disk is suitable for
archival storage of large numbers of data such as color images. It will replace the magnetic tape. A
compact disk with read-only memory (CD-ROM) is a portable optical disk with a memory capacity of
540 MB; the CD-ROM disk is especially suitable for atlases of endoscopic images.
(236)Figure 71. A magneto optical disk and its drive unit. A 130-mm, rewritable, magneto
optical disk (left) holds 646 megabytes. A disk stores more than 5000 endoscopic images in
one-tenth compression (irreversible).
(237)Figure 72. A 300-mm write-once type optical disk and a 130-mm magneto optical disk.
In the authors' system, the 300-mm write-once optical disk is used for backup storage. It
records 108,000 analog images per disk. (Courtesy of Olympus Optical Company, Ltd.)
IC memory cards are lightweight, credit card-shaped devices measuring 54 86 3.3 mm that
incorporate a microcomputer. Characteristics of the IC memory card relevant to image storage include
a large-capacity static random-access memory and quick access time (counted in nanoseconds). An
IC memory card may be used for temporary image storage or as a personal data file for an individual
patient. It can be used as a cache, that is, a type of quick access memory inserted between the central
processing unit and the main memory when the same data are read repeatedly. Data (e.g., an image)
retrieved for the first time from memory are "written" or transferred to the quick access memory, that is,
a cache file. Thereafter, these data are retrieved not from the main memory but from the quick access
memory or cache. This provides a solution to the problem of slowness in displaying digital images.
Security of Stored Data
Endoscopic images are as much a part of a patient's record as laboratory tests and chart notations. As
with all such records, patients have a legitimate expectation that this information will be kept
confidential. Endoscopic images and related information (e.g., procedure reports) must therefore be
protected against loss, theft, and unauthorized manipulation or modification. Data in any memory
medium must be physically protected from accidental erasure or overwriting. In this regard, the
write-once types of media are ideal. Nevertheless, it is mandatory that backup capability be
incorporated into the design of every system for image storage. To avoid violation of patient privacy,
access to data must be restricted to authorized personnel only. Generally, this can be accomplished by
use of passwords.
Image Retrieval and Display

Stored images are usually indexed and identified by patient identification (ID) code and demographic
information, date and sequence number of the procedure, and type of procedure. In most systems,
images captured successively during a procedure are assigned sequential frame numbers. Other
diagnostic information, such as biopsy results, must also be entered into the system. This is usually
accomplished by keyboard entry into a corresponding database.
Retrieval of endoscopic images is done for three basic purposes: patient care, education, and
research. It is frequently useful to compare endoscopic findings from a previous examination when
procedures are performed in follow-up. The ability to retrieve endoscopic images greatly enhances
endoscopy conferences, including case presentations. In addition, endoscopic images have numerous
uses with regard to endoscopic and other types of research.
Required images are retrieved by reference to a denominator (usually items of demographic
information), such as diagnosis, patient's date of birth, date of the procedure, and the patient's ID
number. A unique number is usually assigned to each patient by a hospital or health care system;
sometimes the patient's social security number is used. In general, use of the patient's name as a
denominator is unsatisfactory because two or more patients sometimes have the same name.
Because ID numbers may be entered incorrectly into a computer system, it is best to use a second
denominator to verify that the file(s) and images of the correct patient have been selected. The
patient's date of birth is usually suitable for this purpose.
Image Transmission
An endoscopy unit usually consists of a number of procedure rooms as well as other support rooms
that serve various purposes. It becomes necessary, therefore, to link these rooms and their equipment
into a common system for image management. This can be accomplished in a number of ways. In
most endoscopy units, the various components of an image management system are interconnected
by wire or fiberoptic cable. In such a design, software programs are needed to direct and regulate the
flow of images and other data. This type of system is generally referred to as a local area network
(LAN). Most commercially available systems provide software and other components needed to
establish a LAN.
It is not possible to discuss all aspects of LAN technology within the context of this chapter. However,
reference to a commonly used technology for LAN, called Ethernet, serves to illustrate certain essential
points. Ethernet refers to standards developed for a LAN mainly for connection of devices to the net
and transmission of data within a LAN. It was developed jointly by the Digital Equipment Corporation,
Xerox, and Intel.
The Ethernet LAN system can transmit data at the rate of 10 megabits/sec over coaxial cable with a
data link control protocol of carrier sense multiple access/collision detection (CSMA/CD). The
CSMA/CD is the software that prevents collisions of data files (images) when simultaneous requests
are made for transmission of data over the LAN, such as to or from a memory device. The CSMA/CD
protocol allows the various stations (nodes) on the system (e.g., several procedure rooms) to detect
the presence of transmitted messages on the network by sensing the carrier of the transmission. In
effect, CSMA/CD makes each station wait its turn to transmit until no other station is transmitting. As
the number of stations or nodes on a system increases, the chance of a collision also increases,
thereby making the waiting time longer. In general, digital image transmission is slower when many
stations are present.

Every procedure room in an endoscopy unit, as well as a central control room or area, must have a
station for transmission and display of images if the advantages of an image management system are
to be realized. This is the essential minimum. Beyond this point, many possible enhancements of the
system can increase its utility and improve the quality and efficiency of the endoscopy unit. Display
stations can be placed in consultation rooms to show patients their own endoscopic images. Stations in
physician work areas are useful. In particular, conference rooms and lecture theaters should have
display stations. Each operating room in the surgical theater should have access to stored endoscopic
images. All of these rooms (nodes, stations) must be connected by a LAN.
A System for Electronic Image Management

The system for electronic image management at the authors' hospital illustrates the process of design,
essential points that must be considered in building a system, and some possibilities for choices of
hardware and software.
The Environment
As a first step in designing and planning an electronic image management system, it is necessary to
consider the nature of the environment as well as the objectives of the system.
The authors' institution has a hospital with 600 beds distributed as 13 wards. Approximately 700
outpatient clinic visits occur each day; approximately 5000 digestive endoscopy procedures are
performed each year.15(238) The number of images acquired during a typical
esophagogastroduodenoscopy (EGD) is about 40.16(239) Most of the endoscopy procedures are
performed in the endoscopy unit, which has five examination tables, but ERCP and other procedures
that require fluoroscopy are performed in the radiology unit. Laparoscopy is performed in the operating
theater. The radiology unit and operating theater are both on different floors from the endoscopy unit.
Design of the System

A block diagram of the endoscopy image management system is shown in Figure 73.2,3,1725(240)
In this system, the endoscopy procedure tables in the endoscopy unit are connected on-line with the
main computer in a central control room, where images can be stored in both analog and digital
formats. The 130-mm magneto optical disks recorded with digital images are carried to distant
subsystems for display and review of images. All disks are kept in storage in the central control room.
(241)Figure 73. Block diagram of the authors' image management system. Within the
endoscopy unit, there is a central control room that houses equipment that stores images in
both analog and digital formats. All endoscopy tables are connected on-line with the central
analog image storage device through the multi-input unit. At the same time, all the endoscopy
tables have, both within and distant from the endoscopy unit, an individual digital image
recording device. Lecture theater and conference rooms have digital image display stations.
GIgastrointestinal; NTSCNational Television System Committee.
The flow of endoscopic images from the endoscopy procedure tables is diagrammed in Figure 74.
The images are recorded simultaneously as digital images at the endoscopy table and as analog
images centrally through the multi-input unit. An endoscopy station with procedure table is shown in
Figure 75, and the central control room is shown in Figure 76.
(242)Figure 74. Block diagram of endoscopy station at each endoscopy table. Endoscopic
images acquired with the electronic endoscope are sent to the video processor. The images are
further sent to both the digital image storage device that records images on the 130-mm
magneto optical disk and to the multi-input unit in the central control room. VTRvideo tape
recorder.
(243)Figure 75. View of an endoscopy station. A cable running from behind the television
monitor toward the ceiling connects the video processor via the multi-input unit with the
analog image storage device. In the bottom two sections of the rack, the digital image
recording devices are seen. These are an image compressor/decompressor (above) and a
magneto optical disk drive with its controlling computer (below). (Courtesy of Olympus
Optical Company, Ltd.)
(244)Figure 76. View of central control room. Equipment on the desk to the left is for digital
image display; that on the desk to the right is for data control and analog image recording and
display. The cables in the bundle seen in the right upper corner run to individual endoscopy
tables. In the front of the whitish box is the multi-input unit; below the right desk, in the lower
right corner, is the computer that controls the 300-mm optical disk recorder at the bottom. The
large television monitor at the extreme left of the left desk is a high-resolution computer
graphic monitor used for multiple image display of digital images. To its right is another
monitor showing the indices of the multiple display images; using these indices as a guide, it is
possible to choose an image for enlargement on the monitor above. Between the two lower
monitors (left desk) are two magneto optical disk drive units.
Multiple Image Display

An endoscopy conference in our department is shown in Figure 77. At the display station, an enlarged
endoscopic image (seen on the monitor on the left) is being discussed.21(245) The magneto optical
disks on which the required images are recorded have been brought to the conference room.
(246)Figure 77. Endoscopy conference using a display station. An enlarged endoscopic

image displayed on the left monitor is being discussed. The monitor on the right is a
high-resolution computer graphic monitor for displaying multiple images simultaneously; it is
possible to display up to 42 images simultaneously. The box above the left monitor is a
compression/decompression unit for digital images. Above the right monitor are four magneto
optical disk drive units that can search four disks in parallel for images.
Necessary images are retrieved through keyboard entry of a denominator, such as the patient's ID
number. If searched by patient ID number, all of the procedures performed on that particular patient
are listed first in chronologic order; a certain procedure performed on a given date is selected. Then all
of the images stored during that procedure are displayed, together with their indices, on the monitor
designated for multiple image display. If necessary, any particular image may be selected for
enlargement and detailed study. Figures 78, 79 and 710 are examples of images being displayed
on the monitors, including simultaneous display of multiple images on a computer graphic monitor
(Figure 78), selection of an image (image 26) for enlargement, display of this enlarged image on the
NTSC monitor (Figure 79), and a combination of different types of images (colonoscopic,
endosonographic, and microscopic) from a single patient, which are being edited for presentation
(Figure 710).
(247)Figure 78. Multiple image display on a computer graphic color monitor. Image 26 is
chosen for enlarged display. This is a case of acute gastric ulceration developing after
transcatheter arterial embolization therapy for hepatocellular carcinoma.
(248)Figure 79. Enlarged image display on the (NTSC) color television monitor. Image 26
(Figure 78) is shown enlarged for detailed analysis.
(249)Figure 710. Display of different kinds of color images. For presentation of a case of
colonic carcinoma, a colonoscopic image, two endosonographic images on the left, and a
histologic image of a resected specimen are displayed together. The histologic image was
input by attaching a television camera to a microscope.
Analog-To-Digital and On-Line/Off-Line Dual Hybrid System

A personal computer-based endoscopic image management system was developed by the authors to
minimize the implementation costs (Figure 711). In studying the picture archiving and communication
systems (PACS) in other university hospitals, we discovered that the delay in digital image
transmission was so marked that images to be reviewed during a conference had to be transferred to
the display unit in a conference room throughout the night preceding the meeting. We considered this
a serious problem because the average number of images captured during a procedure in Japan is
40,16(250) many more than are obtained in a conventional radiologic examination. Furthermore, the
number of endoscopic procedures far exceeds that of radiologic examinations. To circumvent this
problem, we developed an analog-to-digital and on-line/off-line dual hybrid system rather than the
standard on-line digital image management system. This system is integrated with the information
system of the hospital. When the shortcomings of the on-line digital image management system are
overcome, our dual hybrid system has the flexibility to be altered without difficulty to an on-line system.
(251)Figure 711. Analog/digital hybrid system for image management. The concept of
analog/digital and on-line/off-line dual hybrid system is used in the current system. The solid
lines show on-line signal flow, the broken line shows off-line signal flow.
Importance and Future Development

An electronic image management system organizes and integrates endoscopic images and makes
them readily available for a variety of purposes. Such a system offers a substantial degree of security
and safety and avoids problems of image degradation that are common to older methods of archiving
images. An electronic system facilitates image processing and analysis and increases the efficiency of
the endoscopy unit, especially those units in which photodocumentation is considered
essential.26(252) A tremendous library of images covering a wide variety of endoscopic findings and
diagnoses can be built within a relatively short period of time, and a properly designed system brings
together the many different types of images that pertain to each patient. All of this has a very favorable
influence on training programs and serves to elevate the level of endoscopic knowledge throughout the
endoscopy unit. With the further standardization of endoscopic terminology and the assimilation of
these terms into computer programs for the generation of endoscopy reports and retention of
endoscopic data, the endoscopy unit will truly enter the modern age of computers.27(253) Among the
remarkable changes that may be foreseen are the development of methods of image communication
between institutions. In this way, endoscopic images can become common assets within health care
systems. Computer networks will make it possible to transfer images from country to country for
consultation, research, and teaching.
REFERENCES
1.
Uji T. Studies on photography of gastric mucosa and its application. Tokyo Igaku Zasshi
1953;61:13541.
2. Fujino MA, Ikeda M, Yamamoto Y, et al. Development of an integrated filing system for
endoscopic images. Endoscopy 1991;23:115.
3.
Fujino MA, Morozumi A, Kawai T, et al. Management of endoscopic images and advances in
optical disk technology. Gastrointest Endosc Clin North Am 1992;2:31334.
4. Dwyer SJ III, Templeton AW, Martin NL, et al. The cost of managing digital diagnostic images.
Radiology 1982;144:3138.
5. James AE Jr, Erickson JJ, Carrol FE, et al. Medical image management: Practical, legal and
ethical considerations. Comput Biol Med 1986;16:24757.
6.
Cox GG, Templeton AW, Dwyer SJ III. Digital image management: Networking, display, and
archiving. Radiol Clin North Am 1986;24:20619.
7. Seshadri SB, Arenson RL, van der Voorde F, et al. Design of a medical image management
system: A practical cost-effective approach. Comput Methods Programs Biomed
1987;25:18592.
8. Rozen P. Computer assistance in gastroenterology: An update. Am J Gastroenterol
1988;83:13237.
9. Iber FL, Kruss DM. Computer storage of gastrointestinal procedures. Am J Gastroenterol
1988;83:663.
10. Publig W, Zandl Ch, Czitober H. EDV-gestutzte Dokumentation in der gastroenterologischen
Endoskopie. Simultane Archivierung und Befundausgabe. Wien Klin Wochenschr
1988;100:867.
11. Sivak MV Jr, Fleischer DE. Colonoscopy with a videoendoscope: Preliminary experience.
12.
Yamagata S, Oida M, Imaizumi H, et al. Management of endoscopic pictures using analog
filing system. Endoscopia Digestiva 1993;5:7318.
13. Classen M, Wagner F, Swobodnik W. Electronic data base in gastroenterological endoscopy.
Endoscopy 1991;23:2931.
14.
Endoscopic Image Exchange Ad Hoc Committee. Digital imaging and communications in
medicine (DICOM). Personal communication. .
15.
Fujino MA, Ikeda M, Yamamoto Y, et al. Introduction of endoscopy units (endoscopy centres).
A series: No. 45. Gastroenterol Endosc 1990;32:22834.
16.
Takasu Y, Kitamura A. Practice of panendoscopy. Stomach Intestine 1984;19:4754.
17.
Fujino MA, Ikeda M, Yamamoto Y, et al. Filing network of electronic endoscopic images. With
reference to a newly developed system. Image Technol Inf Disp Med 1990;22:117781.
18. Fujino MA, Ikeda M, Yamamoto Y, et al. Digital/analog hybrid system for filing of endoscopic
images. Comput Methods Programs Biomed 1992;37:2918.
19.
Fujino MA. Image management in video endoscopy. Organisation Mondiale d'Endoscopie
Digestive Newsletter 1994;1:7.
20.
Morozumi A, Fujino MA. Personal computer-based endoscopic picture archiving and
communication system (PACS). Image Technol Inf Disp Med 1992;24:68993.
21.
Otaka M, Fujino MA, Ikeda M, et al. Development of image display stations for endoscopic
conferences. Endoscopia Digestiva 1992;4:152933.
22.
Fujino MA, Morozumi A, Nakamura T, et al. Analysis on an off-line system for endoscopic
image management. Image Technol Inf Disp Med 1993;25:599602.
23.
Ikeda M, Fujino MA, Morozumi A, et al. Management of endoscopic images using digital
images. Off-line system/hybrid system. Endoscopia Digestiva 1993;5:73945.
24.
Fujino MA, Morozumi A, Nakamura T, et al. Future outlook for digestive endoscopy in the
context of computerisation. Endoscopia Digestiva 1993;5:7915.
25. Fujino MA, Ikeda M, Morozumi A, et al. Electronic endoscopy in perspective. J Gastroenterol
1994;29(suppl 7):8590.
26. Fujino MA, Kawai T, Morozumi A, et al. Endoscopic image manipulation: State of the art.
Endoscopy 1992;24(suppl):455550.
27. Crespi M, Delvaux M, Schapiro M, et al. Minimal standards for computerized endoscopic
database. Am J Gastroenterol 1994;89(suppl):S14453.
Chapter 8 Disinfection of Endoscopes and Accessories

(254)
(255)
ANTHONY T. R. AXON, M.D.
Since the mid-1980s, there has been a quickening of interest in the disinfection of endoscopes and
accessories. The recognition that serious infection may be transmitted endoscopically and that patients
have died as a result of inadequately disinfected equipment has caused endoscopists to review their
working practices. Interest has been further stimulated by the acquired immunodeficiency syndrome
(AIDS) epidemic, not only because of the risk that the human immunodeficiency virus (HIV) might be
transmitted to normal individuals but also because immunocompromised patients may harbor unusual
and potentially pathogenic organisms that may be passed to others.
The enormous increase in the number of patients undergoing endoscopy has meant that nurses and
endoscopy assistants are under pressure to "turn around" equipment more rapidly. They have
recognized that this can lead to unacceptable compromises in cleaning and disinfection, and they have
been attracted to mechanical aids to alleviate the pressure and tedium of these activities. The fact that
only aldehyde-based disinfectants appear at present to be totally reliable for safe disinfection has
meant that increasing numbers of endoscopists and staff have become sensitive to these toxic
chemicals. This has given rise to concern about safety at work and has led to litigation by affected
individuals against their employers. On the one hand, strong pressure is placed on endoscopists to
ensure that all equipment used is completely safe for all patients, whatever their immune status and
whoever underwent endoscopy before them. On the other hand, concern exists about the time taken to
disinfect equipment and the expense required to provide the necessary staff as well as added
resources to ensure that they are working in a safe environment.
These considerations stimulated a number of national endoscopy societies to produce guidelines for
the disinfection of equipment,15(256) and the World Congresses of Gastroenterology have made
similar recommendations on disinfection practice.6(257)
Potential Risks of Inadequately Disinfected Equipment

Infection Carried by Patients
A significant minority of patients referred for endoscopy are infected with pathogenic bacteria. This
applies particularly to individuals who undergo emergency endoscopy. A proportion have acute
infection, such as salmonellosis, that may have given rise to the upper or lower gastrointestinal
symptoms that prompted investigation; others are carriers of infectious disease. Infection with hepatitis
A and B is more common among individuals coming to endoscopy than in the population at large, and
a particularly high percentage of patients are infected with Helicobacter pylori. Because this latter
organism is now considered to be a pathogen (see Chapter 46: Peptic Diseases of the Stomach and
Duodenum), the majority of patients referred for endoscopy in most communities can be considered to
be infectious.
Patients who are immunocompromised frequently develop viral, bacterial, and fungal infections that
may lead to esophagitis, gastric symptoms, or diarrhea and require examination. Some of these
individuals are infected with the HIV virus; others, such as those undergoing chemo-therapy, although
not infected with a primary pathogen, nevertheless harbor other potential pathogens, especially atypical
mycobacteria, spore-forming organisms, and fungi. Although these may not be dangerous to normal
individuals, they may be devastating to other patients with an immune deficiency.711(258)
One of the major protections against enteric infection is gastric acid. The hypochlorhydric stomach
rapidly becomes colonized with a variety of enteric organisms. Many patients referred for endoscopy
are taking H2 receptor antagonists or proton pump inhibitors; others have achlorhydria as a result of
longstanding gastric pathology. Not only may these organisms be transmitted to others, but
hypochlorhydric patients are in greater danger of contracting infection from organisms transmitted to
them at endoscopy.
Infection Generated by Equipment

Endoscopic equipment is contaminated by use in patients, but it also becomes colonized by
saprophytic organisms during storage.1216(259) Opportunistic organisms present in the general
environment are able to colonize stagnant water, detergent, and even certain disinfectants. The warm,
moist channels of an endoscope left in the humid environment of an endoscopy cupboard in a centrally
heated room are rapidly colonized. Similar contamination occurs in ancillary endoscopic equipment
such as catheters and in endoscope washing machines. From these reservoirs of infection, the air
suction channel, the water channel, and the channel that contains the elevating mechanism for the
biopsy forceps may all become contaminated, leading to further spread within the endoscopy room.
The major risk of this contamination occurs with endoscopic retrograde cholangiopancreatography
(ERCP). However, in addition, these organisms, which are often antibiotic resistant, may be introduced
into a succession of patients, who may become colonized and transmit the infection after leaving the
endoscopy room.17,18(260) If they are introduced into patients who are immunologically
compromised, more serious problems may arise.
Salmonellosis
The best-documented cases of endoscopic transmission involve salmonella.1927(261) A number of
epidemics have been reported in which infection with this organism has been transmitted from one
patient to a succession of others. In each reported epidemic, the disinfection techniques used by the
units affected were below the standard now generally recommended. The size of this problem is
difficult to estimate because undoubtedly only a minority of epidemics are recorded in the scientific
literature. Some go unrecognized, others unreported. Most infection has arisen from
esophagogastroduodenoscopy (EGD), but one well-documented series of infections arose in patients
who had undergone colonoscopy with a variety of instruments. The source of infection was eventually
traced to the spiral spring of one pair of biopsy forceps.28(262) It follows that infection may be
transmitted not only by the endoscope but also by ancillary equipment used with the endoscope. It is
not necessary for the organism to be ingested by mouth; effective colonic seeding is also possible. The
fact that the colon is a con-taminated organ does not imply that less rigorous disinfection techniques
should be practiced with the colonoscope or flexible sigmoidoscope.
Helicobacter pylori
The most common pathogenic organism encountered at endoscopy is H. pylori, which infects 95% of
individuals with duodenal ulcer and more than 50% of patients with nonulcer dyspepsia.29(263) When
introduced into a normal stomach, the organism gives rise to an acute gastritic illness;3032(264) in
the majority of cases, that illness progresses to chronic active gastritis, which often persists throughout
life (see Chapter 46: Peptic Diseases of the Stomach and Duodenum). Infection is associated with an
increased risk of duodenal ulcer, gastric ulcer, and gastric cancer.
Two epidemics of helicobacter infection have been reported in volunteers undergoing experimental
intubation for gastric secretion studies.33,34(265) The infection was attributed to inadequately
sterilized pH probes, and although the cause of hypochlorhydria was not recognized at the time, it is
now believed that both were due to H. pylori. Well-documented cases have been reported of
helicobacter transmission by endoscopy in which individual patients became infected with the same
strain of organism as a patient who had undergone prior endoscopy.35(266) The disinfection technique
being practiced at that time did not meet the standards now generally recommended.
More recently, a study from Japan described a postendoscopy syndrome in which patients developed
upper gastrointestinal symptoms following endoscopy. This condition appears to be acute
helicobacter-associated gastritis as a direct result of the endoscopic transmission of the
organism.36(267) Unless equipment is properly disinfected, transmission rates may be as high as 1%
of all endoscopic examinations.
Hepatitis
The fear that hepatitis B might be transmitted by endoscopy has been a constant concern because
asymptomatic carriage of hepatitis B virus (HBV) is well recognized, and inevitably, many affected
patients undergo endoscopy without the clinician being aware of the hazard. It is surprising, therefore,
that only one case of HBV transmission has been reported.37(268) A number of longitudinal studies
have been done on individuals who inadvertently underwent endoscopy immediately following a carrier
who had undergone the same procedure.3846(269) None of these has demonstrated cross-infection,
and transmission of HBV by endoscopy appears to be an exceedingly rare occurrence.
Human Immunodeficiency Virus

To date, no cases of transmission of HIV by endoscopy have been reported. Patients with AIDS often
undergo endoscopy because they are prone to gastrointestinal complications. Fortunately, HIV is
labile4750(270) and less infective than HBV.5155(271) The fact that the latter organism has not
caused serious problems suggests that the danger of HIV transmission by means of endoscopy is very
low. If units follow appropriate disinfection procedures, the risk of transmission can be virtually
eliminated.56,57(272)
Opportunistic Organisms
Opportunistic organisms, particularly pseudomonas, are present throughout the environment and
readily colonize endoscopic equipment and ancillary apparatus. In 1974, Pseudomonas aeruginosa,
Pseudomonas fluorescens, atypical P. aeruginosa, and Pseudomonas maltophilia were isolated from
endoscopes following storage.12(273) At that time, it was recommended that equipment be disinfected
at the beginning of the endoscopy session as well as between each case. In the same year, fatal
pseudomonas septicemia was described following endoscopy in two patients with leukemia, and it was
confirmed that the infecting organism had come from the endoscope.7(274) In 1975, polymicrobial
sepsis was reported following ERCP, and the endoscopy equipment was again thought to be the
source.58(275) A further four patients with sepsis following ERCP were also described, one of whom
had a pseudocyst that was infected with pseudomonas.59(276) In 1976, two cases of fatal septicemia
following ERCP were described in which pseudomonas was the responsible organism.
In 1979, Parker et al.60(277) prospectively analyzed fever and bacteremia following ERCP. Ten of 50
patients became febrile. Organisms were grown from the blood in 7, and the source of infection was
the endoscope in at least 4. In 1980, Noy et al.16(278) reported three patients who became colonized
after EGD with a strain of P. aeruginosa identified as having come from the endoscope. The authors
believed that pseudomonas sepsis had contributed materially to the deaths of these patients and may
have been a major factor in determining their unsatisfactory response to surgery (they had been
admitted to the intensive therapy unit with gastrointestinal hemorrhage). These authors recommended
disinfection before each session. A prospective assessment of bacterial complications following ERCP
was performed in 101 endoscopies.61(279) Aspirates from the pancreaticobiliary duct after injection of
contrast material yielded growth of P. aeruginosa in 14. These had all come from the endoscope.
In 1982, Gerding et al.62(280) drew attention to the importance of 2% glutaraldehyde disinfection
followed by forced-air drying prior to storage. Using this technique, they were able to reduce overnight
contamination, but even in their hands and in a careful study, they still found contamination after
overnight storage in 4 of 63 instruments (6%). In the same year, another case of pseudomonas
septicemia was reported following ERCP,63(281) the organisms again coming from the endoscope. In
1987, pseudomonas infection was described in 12 patients undergoing ERCP.64,65(282) Davion et
al.66(283) reported multiple liver abscesses due to pseudomonas, and Allen et al.64(284) reported 10
patients in whom pseudomonas was cultured from bile following ERCP. Classen et al.,67(285) in 1988,
observed a single case of P. aeruginosa bacteremia following ERCP and then retrospectively reviewed
the cases of the previous year and identified six other infections. In all seven, the infection had
developed within 5 days of ERCP. In each case, the pseudomonas isolate was from equipment, and
each of the patients had received the first scheduled ERCP of the day. The mean duration between
cleaning of the ERCP endoscope and its subsequent use was significantly longer in these cases than
in matched controls. The authors emphasized that all gastrointestinal endoscopes were disinfected
according to American Society of Gastroenterologists and Centers for Disease Control guidelines for
high-level disinfection. At the end of the day endoscopes were washed, disinfected, dried, and
suspended, but "endoscopes were not washed or disinfected again before subsequent use."
In the same year, Godiwala et al.68(286) reported two cases of Serratia marcescens infection following
ERCP (death occurred in one case). The source of the infection was the endoscope. This organism is
usually regarded as a saprophyte and does not normally inhabit the gut. The two cases occurred 3
days apart. The authors of this paper draw attention to three other papers that describe S. marcescens
infection following ERCP.59,69,70(287)
Pseudomonas not only colonizes endoscopic equipment but may also infect tubes and water
bottles,12,13,15,71(288) disinfectant solution, and even full-strength disinfectant.18,72(289) Ancillary
equipment and reservoirs of fluids used in the endoscopy unit should be changed regularly and
cultured periodically. Once pseudomonas has colonized equipment, it may be difficult to eradicate,
particularly if it contaminates washing machines and cleaning apparatus, leading to contamination of
equipment by the machine during the rinsing program (see later in this chapter).
Other Organisms
Theoretical dangers exist that organisms such as fungi, mycobacteria, other bacteria, and viruses may
contribute to cross-infection. However, no reports have been published of transmitted disease at
digestive endoscopy, although fungal contamination of equipment has been described.73(290) In the
developing parts of the world,5(291) other organisms such as Giardia lamblia, amoebae, helminths,
and those associated with cholera represent further hazards, but no confirmed reports of their
transmission have been found in the readily available literature.
Immunocompromised Patients
With the increasing number of patients with AIDS, the use of immunosuppression for the treatment of
malignancy, the ability of physicians to manage patients with advanced liver and renal disease who
would previously have died, and an increase in the number of elderly patients, many patients referred
for endoscopy fall into a particularly vulnerable group with respect to infection. Not only are these
individuals at greater risk of developing disease from pathogenic organisms, but they may become
desperately ill when infected with organisms that would present no major difficulty for healthier people.
Some national societies have recommended that special precautions be taken to protect these patients
at endoscopy. The working party of the World Congresses of Gastroenterology, having taken account
of the published literature on this subject, took the view that if equipment is properly cleaned
mechanically and then totally submersed for 10 minutes in a disinfectant of equivalent power to 2%
activated glutaraldehyde, it should subsequently be safe for use in any individual.6(292)
Special Precautions to Be Taken in Patients Known to Be Infectious

Certain authorities have suggested that patients known to be infected with HIV, HBV, or other
infectious disease should be treated differently from other patients on the grounds that they could be a
risk to endoscopy staff or other patients. Suggestions include reserving a single endoscope for use
only on infectious patients, leaving "dangerous" patients to the end of the schedule of procedures, and
leaving endoscopic equipment in disinfectant fluid for a prolonged period after examination of such
patients. These special precautions are illogical and even dangerous. The implication of taking special
precautions is that standard disinfection practice is inadequate and those patients who carry HIV or
HBV, unknown to themselves and their medical attendants, represent a continual risk to staff and other
patients. Every patient submitted for endoscopy must be considered to be potentially infectious. Staff
must take universal precautions against needle-stick injury and must always employ a safe disinfection
protocol. Furthermore, in order to ensure that everyone is examined with properly disinfected
equipment, the principle of equality of safety must be employed. Whether a patient is the first on the
list, in the middle, or last, no benefit or detriment to safety should apply; that is, no patient must be
placed at greater risk than another, and this implies that no patient should be put at greater safety than
another.

Infective complications are especially important in ERCP because mortality is high when they occur.
Cholangitis and pancreatic sepsis may occur after ERCP because organisms already within the
pancreaticobiliary system are disturbed or may be disseminated by the procedure. Sanctuary areas in
the pancreas or in biliary calculi may be opened by the injection of contrast or by mechanical
manipulation. Alternatively, the injection pressure during retrograde cholangiography may cause a
backwash of infected bile into the bloodstream. Sometimes, following the procedure, a partial
obstruction may become complete owing to trauma, bleeding, or edema, in which case infection may
occur. The most common cause of sepsis, however, is none of these; rather, it is the injection of
contaminated radiographic contrast material into an obstructed duct.16,58,59,61,6368,71,7477(293)
Numerous examples have been reported of pseudomonas septicemia occurring after ERCP in which
the organism, grown from the blood, was the same as that identified in the endoscope. Sepsis is the
most common cause of death following ERCP,78(294) and, in the majority, the condition is iatrogenic.
Post-ERCP sepsis has become more of a problem since the introduction of stenting for malignant
disease. A recent study has shown that patients who underwent ERCP in another unit, before referral,
had a greater chance of developing pseudomonas septicemia than those who underwent the
procedure at the specialist unit.79(295) The explanation advanced for this finding is that the less
experienced units had not been as meticulous with their disinfection technique.
The patient first on the ERCP list is the one most likely to develop sepsis. The reason for this is that
many endoscopy units disinfect their equipment at the end of the day, leave it in storage overnight, but
do not disinfect it before use the following morning. This practice is dangerous. Pseudomonas colonize
equipment overnight, and although colonization does not occur if the equipment is thoroughly dried, it
is very difficult to be sure that all water has been eliminated from every channel. Even in those units
where great care was taken to follow these recommendations, colonization occurred in some of the
instruments. It is essential to observe the universal standards advocated previously. Endoscopic
equipment must be disinfected immediately before use in each patient. Failure to do this leads to
deaths, particularly in patients undergoing ERCP.
Principles of Disinfection and Sterilization

The principles of disinfection and sterilization have been reviewed by Bond.80(296) The process of
sterilization destroys all microbial life, including viruses, vegetative bacterial organisms, mycobacteria,
and spores. All equipment used in general surgery in which an epithelium is breached should be
sterilized. The same applies to needles used for parenteral injection or therapy. Some medical
equipment, however, having been sterilized or disinfected, may be left on an examination table for long
periods before use; such items include tongue depressors, cotton-tipped swabs, otoscopes, and nasal
specula. This equipment does not breach an epithelium, and the environmental organisms that
contaminate them are not pathogenic or are present in small, noninfective numbers. Furthermore,
these instruments are used in areas already contaminated by microorganisms. It is illogical to insist on
sterilization for a tongue depressor, for example, when people regularly ingest unsterile food and suck
the tips of pencils, cigarettes, knives, forks, digits, and other unsterile items. It is perhaps for this
reason that when upper digestive endoscopy was first introduced, relatively little attention was paid to
disinfection and sterilization.
Unlike the instruments used in general surgery, endoscopes are used in nonsterile areas. As soon as
the instrument is introduced into the mouth or anus, it becomes contaminated with microorganisms. It
is therefore illogical to demand "sterility" for endoscopes. Nevertheless, endoscopes and their related
equipment are increasingly used for therapeutic procedures that do breach the gastrointestinal
mucosa, and bleeding may even occur occasionally during diagnostic endoscopy. Unlike a tongue
depressor, however, the equipment is not disposable and is used in a succession of patients, often
with a very short turnaround time. Many patients are potentially infectious, and it is not necessary to
breach the epithelium to transmit their infectious organisms. The need to "disinfect" endoscopic
equipment is therefore apparent, but it is not necessary to "sterilize" it.
The process of disinfection is arbitrarily divided into high-, intermediate-, and low-level disinfection.
High-level disinfection implies that all vegetative organisms are destroyed, together with the majority of
spores. Intermediate-level disinfection is the destruction of all vegetative organisms, including
Mycobacterium tuberculosis var. bovis, but not spores. In low-level disinfection, mycobacteria may not
all be destroyed. Disinfection should be distinguished from antisepsis, a process in which gross
contamination is eradicated but not all vegetative organisms are destroyed. Examples of antisepsis are
the use of iodine on the skin and the wiping of a table top with a disinfectant. The distinction between
high-level disinfection and sterilization is essentially one of confidence. Certain processes inevitably
destroy all microbial life, such as autoclaving and some forms of radiation. Many of the chemicals used
for disinfection, for example, 2% activated glutaraldehyde, are capable of sterilization if left in contact
for a sufficient length of time. Any chemical that is used for high-level disinfection must, by definition,
be capable of sterilization.
The length of time needed for a disinfectant to produce high-level disinfection or sterilization depends
on a number of factors. Of great importance is the nature of the surface to be disinfected. Porous,
absorbent, or convoluted surfaces where disinfectant has greater difficulty in penetration require a
longer period of time than does a smooth, flat, nonporous surface. The penetration of certain
disinfectants may be enhanced by the use of surfactants, which increase the "wettability" of the
surface. Of particular importance is the "cleanliness" of the surface to be disinfected. If equipment is
adequately cleaned before the disinfection processthat is, all organic matter such as blood, mucus,
and other debris is removed, the length of time needed to disinfect that surface is less than if it
remained contaminated.
The cleanliness of the surface depends to some extent on its nature. A plastic lining, for example, may
be more difficult to clean than a smooth metallic surface. Some surfaces lend themselves to the
development of a biofilm, which may render disinfection particularly difficult, especially if the surface
itself is irregular or has discontinuities. Equipment comprising convoluted tubing is a particular
problem, not only because of the difficulty in mechanical cleaning but also because airlocks may
develop that prevent disinfectant from reaching the whole surface.

Access of the disinfectant to the surface is not the only factor that determines the rate at which
organisms are destroyed. In addition to the concentration of the disinfectant and length of time in
contact, the nature of the organism to be destroyed is important. Disinfectants vary in their
effectiveness toward different organisms. In general terms, certain organisms are recognized to be
more resistant to chemical disinfection than others. Spore-forming organisms are particularly resistant.
Certain fungi and mycobacteria are relatively resistant, and a league table of resistance can be drawn
up (Table 81).6(297)
Descending Order of Resistance to

Germicidal Chemicals*
TABLE 81
BACTERIAL SPORES
Baccilus subtilis
Clostridium sporogenes
MYCOBACTERIA
Mycobacterium tuberculosis var. bovis
NONLIPID OR SMALL VIRUSES
Poliovirus
Coxsackievirus
Rhinovirus
FUNGI
Trichophyton spp.
Cryptococcus spp.
Candida spp.
VEGETATIVE BACTERIA
Pseudomonas aeruginosa
Staphylococcus aureus
Salmonella choleraesuis
LIPID OR MEDIUM-SIZED VIRUSES
Herpes simplex virus
Cytomegalovirus
Respiratory syncytial virus
Hepatitis B virus (HBV)
Human immunodeficiency virus (HIV)
From Axon ATR, Bond WW, Bottrill PW, et al. Disinfection and endoscopy.
Working Party Report of the World Congress of Gastroenterology,
Sydney, 1990. J Gastroenterol Hepatol 1991; 6:2347.
* With the exception of HBV, the microorganisms listed under each heading
may be used in standardized laboratory tests to determine the germicidal
activity spectrum of a selected disinfectant chemical.
Fortunately, the two viruses that particularly worry endoscopists are among those most easily
destroyed. However, other factors also influence the effectiveness of disinfection. The presence of
organic matter, or "soil," within a piece of equipment materially reduces the effectiveness of a
disinfectant. Not only does organic matter inhibit or neutralize a disinfectant and prevent access of the
disinfectant to the surface, but organic material may become fixed by the disinfectant itself and form a
shell around other organic particles, which then remain in a sanctuary zone that the disinfectant is
unable to penetrate.
Endoscopic equipment presents particularly unsatisfactory surfaces for disinfection. It is fragile and
heat labile. It is made of glass, which precludes the use of irradiation. The channels within the
endoscope are made of a plastic material and are subject to the trauma of biopsy forceps and
therapeutic equipment, all of which may cause damage and give rise to surface irregularity. The
channels have bends and blind alleys. The air/water channel is too narrow to take a cleaning brush.
Finally, the equipment is used in circumstances in which patients undergo endoscopy one after
another, leaving little time for disinfection procedures to be carried out.
The shapes and compositions of the various ancillary equipment used in endoscopy are also
unsatisfactory for disinfection. Balloons and some catheters cannot be autoclaved. Biopsy forceps are
constructed from coiled wire, which has numerous interstices in which organic material and organisms
may reside and is difficult to clean. Dormia baskets and polypectomy snares comprise metallic wires
within a plastic sheath. The valves used in endoscopes are often heat labile and have a complicated
internal structure. Plastic water bottles, which represent a reservoir of potential infection, are linked to
the endoscope by plastic tubing. In short, the designs of endoscopes and ancillary equipment make
disinfection very difficult.
Disinfectants
A wide variety of commercial preparations is available for disinfection, but claims made by the
manufacturers are not always reliable. Assessing a disinfectant preparation adequately is an expensive
undertaking. Not only should the chemical be tested against a variety of organisms, including viruses
(the infectivity of which may be difficult to assess), but the compound should be tested "in the field"
before being accepted as a reliable preparation for use in endoscopy. Any disinfectant used in
endoscopy should be capable of "sterilization," should not be affected by the presence of organic "soil,"
must have good penetration, and should be rapidly effective. At the present time, the only generally
recommended disinfectants are aldehyde-based. The accepted standard is 2% activated
glutaraldehyde. This disinfectant destroys most vegetative organisms and viruses within 2 min.
However, in order to destroy mycobacteria and spores, the exposure time should be longer. An
international working party has recommended that if an endoscope has been properly cleaned, full
immersion of the endoscope and all channels in 2% activated glutaraldehyde for 10 min should be
sufficient to render the equipment safe for use in any patient.6(298)
Other disinfectant preparations may be equally efficacious, but proof of their efficacy should be sought
before they are recommended for use. Other available disinfectants include hydrogen peroxide, a
difficult compound to work with, and organic iodide-based preparations, to which some organisms may
be resistant81,82(299) and which may discolor the equipment.83(300) Other compounds have been
advocated, but they show a low level of activity against pseudomonas7,72,84,85(301) and
salmonella,2126(302) they lack tuberculocidal86(303) and virucidal activity, or they are rapidly
inactivated by organic material.83, 87(304) The alcohols have been suggested as an alternative to the
aldehydes. However, several studies have shown that they do not effectively eradicate
pseudomonas.12,14(305) In addition, owing to their flammability, they represent an environmental
hazard both within the endoscopy unit and with their disposal. Furthermore, they damage endoscopic
equipment after prolonged immersion.
For the reasons presented previously, only aldehyde-based disinfectants are recommended for use in
endoscopy at present. However, their major disadvantage lies in the environmental hazard they pose
to endoscopists and endoscopy assistants. Sensitivity to glutaraldehyde is a major problem in many
endoscopy units. The chemical causes conjunctivitis, rhinitis, asthma, and dermatitis;8890(306) some
individuals complain of general malaise when they are in contact with it. The disinfectant is volatile;
once sensitized, staff may develop symptoms when only trace amounts are present in the
environment. For these reasons, it is essential to take special precautions when it is used. Ideally, the
disinfection process should be carried out in a room that is completely isolated from both the
endoscopy area and the internal environment. Within that area, staff should be protected by masks
and special clothing. If this is not possible, a closed system should be used, with special attention given
to ventilation. Open troughs should never be used except in a specially constructed fume cupboard
with ventilation extraction to the outside. Special garments and gloves should always be used to avoid
skin contact. Ordinary surgical and disposable gloves are often not satisfactory, and the penetration of
glutaraldehyde through glove material is being assessed.9193(307)
Disinfection Technique
The most important part of disinfection is mechanical cleaning. Unless the endoscope and its
accessories are properly cleaned, the disinfection process cannot be effective. The presence of
organic debris within the endoscope channels or on the surface of the instrument causes problems in a
number of ways. Organic material exposed directly to most disinfectants denatures and becomes fixed
and hard within the endoscope channel. This leads to an irregularity that interferes with subsequent
cleaning and leads to a gradual build-up of an irregular film inside the endoscope channel. This porous,
organic matrix forms a nidus for colonization and provides sanctuary zones for organisms within it.
Particular problems arise at Y junctions, valve housings, biopsy ports, and exit ports, to which
particular attention must be directed.
Cleaning
Immediately following removal from the patient, the endoscope channels should be flushed by a
suction of water through the suction channel and ejection of clean water through the air/water channel.
This procedure should be performed by the endoscopist before handing the equipment to the
endoscopy assistant. The endoscope is then usually removed to a cleaning area, where it should be
totally immersed in warm water and detergent. The outside should be washed thoroughly with
disposable swabs. The distal end is brushed with a soft toothbrush, particular attention being paid to
the air/water outlet nozzle and the bridge elevator where appropriate. All of the valves should be
removed, disassembled, cleaned, and disinfected separately. The biopsy channel opening and suction
ports should be cleaned with a cotton-tipped swab.
Next, attention should be directed to the suction/ biopsy channel, which should be brushed through
using a cleaning brush of appropriate size for the specific instrument channel. This should be done
several times from top to bottom, and the cleaning brush should itself be washed in detergent using a
soft toothbrush each time it emerges from the distal end. Initially, the brush should be introduced
through the biopsy port; then it should be passed through the suction valve port, down through the
insertion tube, and lastly from the valve suction port proximally through the umbilical cord. Each
procedure should be carried out at least three times to ensure adequate cleansing of the channels.
If an automatic cleaning/disinfection machine is available, the equipment can then be transferred to it.
If such equipment is not available, however, manual cleaning should be continued. Each internal
channel should be flushed with detergent fluid. It is important to remember that some instruments have
more channels than others. This is particularly important in equipment with a raiser bridge (e.g.,
side-viewing duodenoscopes). After this cleaning has been done, all channels and the outside of the
instrument should be rinsed with tap water. Then the outside of the instrument should be dried and the
internal channels each flushed with air in order to expel as much water as possible prior to the
disinfection procedure.
Disinfection
The cleaning process reduces the number of contaminating organisms to a very low level. The
purpose of disinfection is to further reduce the number of organisms. The endoscope and all internal
channels should be soaked in 2% glutaraldehyde or disinfectant of similar potency for at least 10 min.
This prepares an instrument for use in any patient about to undergo endoscopy, no matter who has
been previously examined or whether the next patient on the list is immunocompromised.
Some authorities recommend 5 min of disinfection, which destroys viruses and vegetative bacterial
organisms. A number of spore-forming organisms and mycobacteria may remain; although these are
not a danger to most patients, they may represent a hazard to those who are immunocompromised.
For this reason the International Working Party6(308) recommends a 10-min soak.
Following disinfection, the equipment must be rinsed thoroughly, both internally and externally, with
drinking-quality water in order to remove all traces of the disinfectant. The outside of the endoscope is
then dried and air is flushed through the channels. The instrument is then ready for use.
Overnight Storage
In order to reduce overgrowth of commensal organisms in stored equipment, it is essential to follow
several procedures. At the end of the endoscopy session, the equipment needs to be cleaned and
disinfected in the usual manner. Following this, 70% alcohol should be introduced into each channel,
which should then be thoroughly dried using compressed air. After these procedures, the instrument
must be hung in a well-ventilated cupboard, not coiled in a closed case.
Washing Machines
Manual cleaning and disinfection of endoscopes are tedious, repetitive, and time-consuming activities.
Endoscopy assistants are under pressure from endoscopists to reduce the turnaround time to allow
more patients to undergo endoscopy per unit time, and staff may be tempted to take short cuts with
both the completeness of mechanical cleaning and the length of time equipment is in contact with
disinfectant. Most units use aldehyde-based disinfectants in endoscopy. These compounds are toxic to
staff, and small quantities of vapor in the environment may trigger unpleasant symptoms in sensitive
individuals. For all of these reasons, many endoscopy units have turned to automatic
washing/disinfection equipment to alleviate the workload. In addition, it is hoped that confining the
disinfectant to a closed system will decrease the likehood of its giving rise to sensitivity in staff
members.
Most endoscope washing machines are designed along the lines of a dishwasher, but without the
heating element. The apparatus does not "clean" the endoscope mechanically; it perfuses the internal
channels but does not introduce a brush, which is essential to mechanically dislodge organic material.
It is essential, therefore, that all endoscopic equipment be thoroughly cleaned mechanically, as
described previously, before being placed in the machine. Washing machines usually run on a
computerized cycle. Initially, detergent and water are flushed through the channels and the outside of
the equipment is sprayed. Then it is drained and rinsed. Following this, the endoscope is immersed in
disinfectant, the channels are filled, and the equipment is left in contact with the disinfectant for a
preprogrammed period of time. The disinfectant is then returned to a reservoir, and a rinsing cycle with
water completes the process. At the end of the cycle, the equipment is usually wet and needs to be
manually dried internally and externally.
A number of endoscope washing machines are available commercially, but unfortunately, most are
flawed in design; some actually represent serious danger to both patients and staff. Most of the
equipment marketed is bulky and immobile. Some are noisy and cannot be comfortably located within
the endoscopy unit. Most are very expensive, being priced at the level of a superior family automobile.
Some are designed to take only one endoscope at a time; others are limited to a single manufacturer's
endoscopic equipment.
Perhaps the most important flaw is that the majority have no satisfactory way to disinfect the machine
itself. Even a cursory examination reveals areas of stagnant water, which provides the potential for
commensals to multiply. Unless the autodisinfection mechanism is designed to perfuse every area of
tubing, including the detergent reservoir, the risk of serial endoscope contamination remains. The
particular area most commonly neglected is the piping between the main water supply and the inlet
valve. In many machines, the disinfectant follows a different pathway and exits through different
nozzles from the water used to rinse the equipment.
Several machines have a spray mechanism that, in the initial washing phase, splashes contaminated
water onto the inside of a perspex cover. This cover does not come in contact with the disinfectant in
the second phase. Thus, in the rinse cycle, the contaminated droplets are sprayed back onto the "clean
equipment" to be used for the next procedure. One machine was initially designed with an "oral" and an
"anal" program, presumably to ensure that disinfectant from colonoscopes would not become mixed
with disinfectant from upper gastrointestinal endoscopes. The thought process behind this
arrangement reveals a lack of understanding of the nature of "disinfection."
An effective autodisinfection cycle is essential, but in nearly all machines, it is either ineffective or
incomplete or requires such labor-intensive activity as to negate the advantage of having a machine in
the first place. Furthermore, attempting to autodisinfect some machines creates considerable risk of
spillage of disinfectant and atmospheric pollution by the vapor. Indeed, although endoscope washing
machines are said to be "closed," aldehyde vapor can be readily detected in the immediate
environment of most machines.
At present, no commercially available washing machine can be regarded as satisfactory. At least two
models, which have been shown to be unsafe, have been removed from the market in the United
States, although they continue to be marketed in other parts of the world.
Dilution of Disinfectant
With usage, the disinfectant becomes diluted and therefore inactive. It is very important for staff to be
aware that disinfectant should be regularly changed. The frequency depends more on the number of
times it has been used than on the length of time it has been in service. The effectiveness of the
disinfectant should be regularly monitored.
Ancillary Equipment
Ancillary equipment provides a potential source of cross-infection in endoscopy, so wherever possible
it is desirable to use disposable or sterilizable equipment. A particular problem relates to coiled springs
and cannulae, from which it may be difficult to dislodge organic matter. Endoscope valves should be
dismantled for cleaning purposes, and an ultrasonic cleaning device should be employed with this
equipment to ensure that mechanical cleansing is as complete as possible prior to sterilization.
Endoscopy Room Design

Design of the endoscopy room must take into consideration the cleaning, disinfection, and
maintenance of the equipment (see Chapter 4: The Endoscopy Unit). A purpose-designed cleaning
area separate from the rest of the unit should be provided so that toxic fumes are isolated. The
disinfection room should be well ventilated and contain a fume cupboard, so that if open troughs of
glutaraldehyde are required they can be isolated from the staff. Good ventilation is essential.
Depending on the policies of the unit, a bench-top autoclave or ultrasonic cleaning machine may be
incorporated along with the usual sinks and disinfection apparatus. In a large endoscopy unit, two-way
cupboards may be helpful so that dirty equipment can be delivered at one entry and clean equipment
collected from another without having to enter the disinfection room itself. Within the disinfection area,
staff safety is paramount. Apart from good ventilation, there should be enclosed systems, protective
clothing, and masks if necessary.
Training
Endoscopists and endoscopy assistants should receive instruction about the risk of infection to
themselves and to patients, and they must learn how to clean and disinfect endoscopic equipment and
ancillaries. Suitable guidelines should be laid down for each unit, and a designated individual should be
responsible for ensuring that endoscopic cleaning protocols are followed and kept up to date. He or
she should ensure that other employees within the department are properly trained and understand the
disinfection policies that have been laid down. Disinfection practice must be regularly audited.
All endoscopy staff must be made aware of the dangers of contact with disinfectants and should take
care to avoid such contact. It is most important that precautions do not become relaxed in this regard.
Staffing
Staffing numbers should be commensurate with the needs of the endoscopy facility, taking account of
the time and expertise needed to ensure that equipment is properly cleaned and maintained. The
scheduling of endoscopy sessions should correlate with the number of staff available and the
turnaround time required for disinfection; if necessary, extra endoscopic equipment or staff, or both,
must be provided so that each instrument is given the attention it requires. Using the "procedure
unit/weighted scale" system of analysis (discussed in Chapter 4: The Endoscopy Unit), an objective
assessment to determine the specific needs of a given facility can be made.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Weller IVD, Williams CB, Jeffries DJ, et al. Cleaning and disinfection of equipment for
gastrointestinal flexible endoscopy: Interim recommendations of a Working Party of the British
Society of Gastroenterology. Gut 1988;29:113451.
Cowen AE. Infection and Endoscopy. Sydney: Gastroenterological Society of Australia. 1987.
American Society of Gastrointestinal Endoscopy. Publication No. 1018. Infection control
during gastrointestinal endoscopy. Guidelines for clinical application. Gastrointest Endosc
1988;34(suppl):37S-40S.
Recommended Practices: Care of instruments, scopes and powered surgical instruments.
AORN J. 47:55669; 1988.
Tandon RK. Endoscopic disinfection: Practices and recommendations. J Gastroenterol
Hepatol 1991;6:379.
Axon ATR, Bond WW, Bottrill PM, et al. Disinfection and endoscopy. Working Party Report of
the World Congress of Gastroenterology, Sydney, 1990. J Gastroenterol Hepatol
1991;6:2347.
Greene WH, Moody M, Hartley R, et al. Esophagoscopy as a source of Pseudomonas
aeruginosa sepsis in patients with acute leukaemia: The need for sterilization of endoscopes.
Cohen FL, Koerner RS, Taub SJ. Solitary brain abscess following endoscopic injection for
sclerosis of esophageal varices. Gastrointest Endosc 1985;31:3313.
Barthel JS, Sprouse RF, Dix JD, et al. Fatal Candida esophageal abscess and sepsis
complicating endoscopic variceal sclerosis. Gastrointest Endosc 1987;33:10710.
Lai KH, Tsai YT, Lee SD. Spontaneous bacterial peritonitis after endoscopic variceal
sclerotherapy. Gastrointest Endosc 1986;32:303.
McGrew W, Goodin J, Stuck W. Fatal complication of endoscopic sclerotherapy: Serratia
marcescens bacteremia with delayed esophageal perforation. Gastrointest Endosc
1985;31:32931.
Axon ATR, Phillips I, Cotton PB, et al. Disinfection of gastrointestinal fibre-endoscopes.
Lancet 1974;1:6568.
Martin TR, Silvis SE, Vennes JA. The reduction of septic complications following ERCP in
obstructed patients. Clin Res 1976;24:566A.
Tolon M, Thofern E, Miederer SE. Disinfection procedures of fibrescopes in endoscopy
departments. Endoscopy 1976;8:249.
Lindstaedt H, Krizek L, Miederer SE, et al. Experiences and problems in the disinfection of
fibre-endoscopes. Endoscopy 1978;10:805.
Noy MF, Harrison L, Holmes GKT, et al. The significance of bacterial contamination of
fibreoptic endoscopes. J Hosp Infect 1980;1:5361.
Anskink-Schipper MC, Van Der Horst-Bruijn JGG, Zanen HC. The hidden source: Hospital
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
infections caused by endoscopes contaminated with Pseudomonas aeruginosa. Ned Tijdschr

Geneeskd 1981;125:9568.
Burdon DW, Whitby JL. Contamination of hospital disinfectants with Pseudomonas species.
BMJ 1967;2:1535.
Burkhardt F. Salmonellenubertragung durch Gastroscopie. Hosp Hyg 1976;68:3379.
Chmel H, Armstrong D. Salmonella oslo. A focal outbreak in a hospital. Am J Med
1976;60:2038.
Tuffnell PG. Salmonella infections transmitted by a gastroscope. Can J Public Health
1976;67:1412.
Centers for Disease Control. Salmonella gastroenteritis acquired from gastroduodenoscopy.
MMWR 1977;26:266.
Dean AG. Transmission of Salmonella typhi by fibreoptic endoscopy. Lancet 1977;2:134.
Beecham HJ, Cohen ML, Parkin WE. Salmonella typhimurium. Transmission by fibreoptic
upper gastrointestinal endoscopy. JAMA 1979;241:10135.
Schliessler KH, Rozendal B, Taal C, et al. Outbreak of Salmonella agona infection after upper
intestinal fibreoptic endoscopy. Lancet 1980;2:1246.
Hawkey PM, Davies AJ, Viant AC, et al. Contamination of endoscopes by Salmonella species.
J Hosp Infect 1981;2:3736.
O'Connor BH, Bennett JR, Alexander JG, et al. Salmonellosis infection transmitted by
fibreoptic endoscopes. Lancet 1982;2:8646.
Dwyer DM, Klein G, Istre GR, et al. Salmonella newport infections transmitted by fibreoptic
colonoscopy. Gastrointest Endosc 1987;33:847.
Dwyer B, Kaldor J, Tee W, et al. The prevalence of Campylobacter pylori in human
populations. In Rathbone BJ, Heatley RV (eds). Campylobacter pylori and Gastroduodenal
Disease. Oxford: Blackwell Scientific Publications, 1989;1906.
Sobala GM, Crabtree JE, Dixon MF, et al. Acute Helicobacter pylori infection: Clinical
features, local and systemic immune response, gastric mucosal histology, and gastric juice
ascorbic acid concentrations. Gut 1991;32:14158.
Marshall B, Armstrong J, McGechie D, et al. Attempt to fulfill Koch's postulates for pyloric
Campylobacter. Med J Aust 1985;142:4369.
Morris A, Nicholson G. Ingestion of Campylobacter pyloridis causes gastritis and raised
fasting gastric pH. Am J Gastroenterol 1987;82:1929.
Ramsey EJ, Carey KV, Peterson WL, et al. Epidemic gastritis with hypochlorhydria.
Gledhill T, Leicester RJ, Addis B, et al. Epidemic hypochlorhydria. Br Med J (Clin Res Ed)
1985;289:13836.
Langenberg W, Rauws EAJ, Oudbier JH, et al. Patient-to-patient transmission of
Campylobacter pylori infection by fibreoptic gastroduodenoscopy and biopsy. J Infect Dis
1990;161:50711.
Sugiyama T, Naka H, Yabana T, et al. Is Helicobacter pylori infection responsible for
post-endoscopic acute gastric mucosal lesions?. Eur J Gastroenterol Hepatol 1992;4(suppl
1):S936.
Birnie GG, Quigley EM, Clements GB, et al. Endoscopic transmission of hepatitis B virus. Gut
1983;24:1714.
Morris IM, Cattle DS, Smits BM. Endoscopy and transmission of hepatitis B. Lancet
1975;2:1152.
McDonald GB, Silverstein FE. Can gastrointestinal endoscopy trans-mit hepatitis B to
patients?. Gastrointest Endosc 1976;22:16870.
Centers for Digestive Disease Control. Minimizing the risk of hepatitis B transmission by
flexible fibreoptic endoscopes: Recommendations for use, routine cleaning and disinfecting.
Centers for Disease Control, Hepatitis surveillance report no. 41. Atlanta: Public Health
Services, US Department of Health, Education, and Welfare, 1977;223.
McClelland DBL, Burrell CJ, Tonkin RW, et al. Hepatitis B: Absence of transmission by
gastrointestinal endoscopy. BMJ 1978;1:234.

42. Moncada RE, Denes AE, Berquist KR, et al. Inadvertent exposure of endoscopy patients to
viral hepatitis B. Gastrointest Endosc 1978;24:2312.
43. Morgan AC, McAdam WAF, Walker BE. Hepatitis B and endoscopy. Lancet 1978;1:369.
44. Koretz RL, Camacho R. Failure of endoscopic transmission of hepatitis B. Dig Dis Sci
1979;24:214.
45. Ayoola EA. The risk of type B hepatitis infection in flexible fibreoptic endoscopy. Gastrointest
Endosc 1981;27:602.
46.
Kok ASF, Lai C-L, Hui W-M, et al. Absence of transmission of hepatitis B by fibreoptic upper
gastrointestinal endoscopy. J Gastroenterol Hepatol 1987;2:17580.
47. Resnik L, Veren K, Salahuddin SZ, et al. Stability and inactivation of HTLV III/LAV under
clinical and laboratory environments. JAMA 1986;255:188791.
48. Martin LS, McDougal JS, Loskoski SC. Disinfection and inactivation of the human T
lymphotropic virus type III/lymphadenopathy-associated virus. J Infect Dis 1985;152:4003.
49. Hanson PJ, Gor D, Jeffries DJ, et al. Chemical inactivation of HIV on surfaces. BMJ
1989;298:8624.
50.
Hanson PJ, Gor D, Chadwick MV, et al. Endoscopy and AIDS: An evaluation of the risk of
cross infection. Gut 1989;30:A742.
51.
Centers for Disease Control Update: Human immunodeficiency virus infections in health-care
workers exposed to blood of infected patients. MMWR. 36:28591987.
52.
McEvoy M, Porter K, Mortimer P, et al. Prospective study of clinical, laboratory and ancillary
staff with accidental exposures to blood or body fluids from patients infected with HIV. BMJ
(Clin Res Ed) 1987;294:15957.
53. Henderson DK, Saah AJ, Zak BJ, et al. Risk of nosocomial infection with human T-cell
lymphotropic virus type III/lymphadenopathy-associated virus in a large cohort of intensively
exposed health care workers. Ann Intern Med 1986;104:6447.
54. Hirsch MS, Wormser GP, Schooley RT, et al. Risk of nosocomial infection with human T-cell
lymphotropic virus III (HTLV-III). N Engl J Med 1985;312:14.
55.
Gerberding JL, Moss AR, Bryant CE, et al. Risk of acquired immune deficiency syndrome
(AIDS) virus transmission to health care workers (HCW) [Abstract]. Program and Abstract of
the Twenty-fifth Interscience Conference on Antimicrobial Agents and Chemotherapy.
Washington, DC: American Society of Microbiology, 1985;131.
56. Hanson PJV, Jeffries DJ, Collins JV. Viral transmission and fibreoptic endoscopy. J Hosp
Infect 1991;18(suppl A):13640.
57. Hanson PJV. AIDS: Practicing safe endoscopy. Baillieres Clin Gastroenterol 1990;4:47794.
58. Elson CO, Hattouri K, Blackstone MO. Polymicrobial sepsis following endoscopic retrograde
cholangiopancreatography. Gastroenterology 1975;69:50710.
59. Davis JL, Milligan FD, Cameron JL. Septic complications following endoscopic retrograde
cholangiopancreatography. Surg Gynecol Obstet 1975;140:3657.
60. Parker HW, Geenen JE, Bjork JT, et al. A prospective analysis of fever and bacteremia
following ERCP. Gastrointest Endosc 1979;25:1023.
61. Low DE, Allan MD, Micflikier B, et al. Infectious complications of endoscopic retrograde
cholangiopancreatography. Arch Intern Med 1980;140:10767.
62. Gerding DN, Peterson LR, Vennes JA. Cleaning and disinfection of fibreoptic endoscopes:
Evaluation of glutaraldehyde exposure time and forced-air drying. Gastroenterology
1982;83:6138.
63. Doherty DE, Falko JM, Lefkovitz N, et al. Pseudomonas aeruginosa sepsis following
retrograde cholangiopancreatography (ERCP). Dig Dis Sci 1982;27:16970.
64. Allen JI, O'Connor AM, Olson MM, et al. Pseudomonas infection of the biliary system resulting
from use of a contaminated endoscope. Gastroenterology 1987;92:75963.
65. Siegman-Igra Y, Isakov A, Inbar G, et al. Pseudomonas aeruginosa septicemia following
endoscopic retrograde cholangiopancreatography with a contaminated endoscope. Scand J
Infect Dis 1987;19:52730.
66. Davion T, Braillon A, Delamarre J, et al. Pseudomonas aeruginosa liver abscesses following
endoscopic retrograde cholangiography. Report of a case without biliary tract disease. Dig Dis
Sci 1987;32:10446.
67. Classen DC, Jacobson JA, Burke JP, et al. Serious Pseudomonas infections associated with
endoscopic retrograde cholangiopancreatography. Am J Med 1988;84:5906.
68. Godiwala T, Andry M, Agrawal N, et al. Consecutive Serratia marcescens infections following
endoscopic retrograde cholangiopancreatography. Gastrointest Endosc 1988;34:457.
69. Zimmon DS, Falkenstein DB, Riccobon C, et al. Complications of endoscopic retrograde
cholangiopancreatography. Analysis of 300 consecutive cases. Gastroenterology
1975;69:3039.
70.
Helm EB, Stille W. Infective complications. In Classen M, Geenen J, Kawai J (eds).
Nonsurgical Biliary Drainage. Berlin: Springer-Verlag, 1984;1139.
71. Bass DH, Oliver S, Bornman PC. Pseudomonas septicaemia after endoscopic retrograde
cholangiopancreatographyAn unresolved problem. S Afr Med J 1990;77:50911.
72. Sanford JP. Disinfectants that don't. Ann Intern Med 1970;72:2823.
73.
Singh S, Singh N, Kochhar R, et al. Contamination of an endoscope due to Trichosporon
beigelli. J Hosp Infect 1989;1:4953.
74. Thurnherr N, Bruhlmann WF, Krejs GI, et al. Fulminant cholangitis and septicemia after
endoscopic retrograde cholangiography (ERC) in two patients with obstructive jaundice. Am J
Dig Dis 1976;21:47781.
75. Schousboe M, Carter A, Sheppard PS. Endoscopic retrograde cholangiopancreatography:
Related nosocomial infections. N Z Med J 1980;92:2757.
76. Vennes JA, Jacobson JR, Silvis SE. Endoscopic cholangiography for biliary system diagnosis.
Ann Intern Med 1974;80:614.
77. Deviere J, Motte S, Dumonceau JM, et al. Septicemia after endoscopic retrograde
cholangiopancreatography. Endoscopy 1990;22:725.
78. Bilbao MK, Dotter CT, Lee TG, et al. Complications of endoscopic retrograde
cholangiopancreatography (ERCP). Gastroenterology 1976;70:31420.
79. Motte S, Deviere J, Dumonceau JM, et al. Risk factors for septicaemia following endoscopic
biliary stenting. Gastroenterology 1991;101:137481.
80. Bond WW. Disinfection and endoscopy: Microbial considerations. J Gastroenterol Hepatol
1991;6:316.
81.
Centers for Disease Control. Contaminated povidone-iodine solutionNortheastern United
States. MMWR 1980;29:5535.
82. Craven DE, Moody B, Connolly MG, et al. Pseudobacteraemia caused by povidone-iodine
solution contaminated with Pseudomonas cepacia. N Engl J Med 1981;305:6213.
83.
Spaulding EH, Cundy KR, Turner FJ. Chemical disinfection of medical and surgical materials.
In Block SS (ed). Disinfection, Sterilization and Preservation. 2nd ed. Philadelphia: Lea &
Febiger 1977;65484.
84.
Malizia WF, Gangarosa EJ, Goley AF. Benzalkonium chloride as a source of infection. N Engl
J Med 1960;263:8002.
85.
Lee JC, Fialkow PJ. Benzalkonium chlorideSource of hospital infection with gram-negative
bacteria. JAMA 1961;177:1446.
86. Kaslow RA, Mackel DC, Mallison GF. Nosocomial pseudo-bacteraemia. Positive blood
cultures due to contaminated benzalkonium chloride. JAMA 1976;236:24079.
87.
Slotnick IJ. Microbiology and sterilization of endoscopes. In Berci G (ed). Endoscopy. New
York: Appleton-Century-Crofts, 1976;15570.
88. Axon ATR, Banks J, Cockel R, et al. Disinfection in upper-digestive-tract endoscopy in Britain.
Lancet 1981;1:1093.
89. Corrado OJ, Osman J, Davies RJ. Asthma and rhinitis after exposure to glutaraldehyde in
endoscopy units. Hum Toxicol 1986;5:3257.
90. Burge PS. Occupation risks of glutaraldehyde. BMJ 1989;299:342.
91. Lehman PA, Franz TJ, Guin JD. Penetration of glutaraldehyde through glove material:
Tactylon versus natural rubber latex. Contact Dermatitis 1994;30:1767.

92. Mellstrom GA, Lindberg M, Boman A. Permeation and destructive effects of disinfectants on
protective gloves. Contact Dermatitis 1992;26:16370.
93.
Mellstrom G. Protective gloves of polymeric materials. Experimental permeation testing and
clinical study of side effects. Acta Derm Venereol Suppl 1992;163:154.
Chapter 9 Principles of Electrosurgery

(309)
ROBERT D. TUCKER, PH.D., M.D.
True electrocautery uses direct current from a battery to heat a wire, which in turn coagulates bleeding
tissue; however, electrocautery cannot cut tissue. In contrast, electrosurgery employs radio frequency
(RF) alternating currents that flow through the tissue to produce heat and cause both cutting and
coagulating effects. Electrosurgery is used in more than 18 million surgical procedures each year in the
United States and is discussed in this chapter.
History
Many scientists experimented with the effects of alternating currents on tissue, but the initial
development of electrosurgery is generally credited to d'Arsonval, who demonstrated in 1891 that
high-frequency alternating currents could be passed through living tissue, causing desiccation without
producing muscle and nerve stimulation.1(310) In the early 1900s various researchers such as
Riviere,2(311) Clark,3(312) and Nagelschmidt4(313) reported the use of electrosurgery in treating a
wide variety of clinical problems, such as the coagulation of intractable ulcers and the removal of
benign and malignant tumors of the head, neck, breast, and cervix. At the same time, Edwin Beer
initiated his pioneering effort to employ electrosurgery through a cystoscope to remove bladder
tumors.5(314) It was not until the late 1920s, however, that electrosurgery was accepted in surgical
practice.
Although several inventors produced electrosurgical machines for clinical use,6(315) only those
developed by William Bovie gained prominence. Bovie, a Harvard physicist, developed two separate
electrosurgical generators, one for cutting and another for coagulation. The acceptance of these
generators by surgeons was actively promoted by neurosurgeon Harvey Cushing. Cushing, Director of
Surgery at Boston's Peter Bent Brigham Hospital, performed various intracranial surgeries such as
removing meningiomas, vascular myelomas, and astrocytomas on patients who were previously
considered inoperable.7(316) By 1928, the Liebel-Flarsheim Company was producing the "Bovie" unit
commercially. Solid-state electrosurgical units, developed in the early 1970s, now offer reduced
generator size, added options, and enhanced electrosurgical safety for both patient and operator.
However, the principles of high-frequency alternating current effects developed by the work of early
pioneers have remained unchanged.
Electrosurgical Variables
The surgical action of electrosurgery is produced by passing alternating currents through tissue.
Because the tissue resists current flow, heat is produced. If heating is sufficiently high and rapid,
intracellular and extracellular fluids are vaporized and a cutting action is produced. If the amount of
heat produced is less and heating is slower, the cells are desiccated and the action produced is
coagulation. The quantity and time course of tissue heating are controlled in part by the generator
voltage and current output, waveform of the generator output, size of the active electrode, tissue
resistance, and application time.
The operating frequency of the electrosurgical generator is also important but not critical in relation to
tissue heating. Muscles and nerves are easily stimulated by alternating currents with frequencies of 2
KHz (2 kilohertz, i.e., 2000 cycles/sec) or less. Although muscles and nerves can be stimulated by
alternating currents as high as 1 MHz (1 megahertz, i.e., 1 million cycles/sec), the currents required are
extremely large and are not used in electrosurgery.8,9(317) Based on such neuromuscular stimulation
data, electrosurgical generators typically use frequencies between 400 KHz and 3 MHz. Generators
cause virtually no direct stimulation at these frequencies because they do not supply the necessary
high-current amplitude; the effect, therefore, is only thermal. The electrosurgery operating frequency
range is in the radio band, and thus the currents are referred to as radio frequency currents.
Voltage
Voltage (measured in volts, V) is the force required to push current (electrons or charged ions) through
a resistance. The voltage produced by an electrosurgical generator alternates between positive and
negative values, similar to household voltage except in the RF band. The number of volts from the
maximum to minimum value is termed the peak-to-peak voltage (Vp-p). The higher the voltage, the
more force is available to push current into tissue. Therefore, higher voltages cause deeper thermal
destruction (Figure 91).
(318)Figure 91. Tissue effect of increasing peak-to-peak voltage (Vp-p). As the voltage
increases, the arc intensity increases and thermal damage (shaded area) becomes deeper. At
high voltage, charring is seen on the tissue.
Tissue Resistance
As electric current (measured in amperes, A) flows through tissue, the tissue resists the flow; the
amount of tissue resistance is measured in ohms. The electrolyte content of tissue is directly
proportional to its resistance. Therefore, tissue such as liver with a high content of blood has a low
resistance, whereas fat and bone have high resistances.
Some typical tissue resistance values are blood, 30 to 50 ohms; skeletal muscle, 200 to 400 ohms;
bowel wall, 150 to 250 ohms; fat, 800 to 1000 ohms; bone, 1000 to 2000 ohms; and skin, 10K to 50K
ohms. (To be precise, these values should be in terms of resistivity, which is characteristic of a
material rather than of a particular specimen of a material; the unit of resistivity is ohm-cm). During
electrosurgery, the resistance of tissue varies dramatically: Prior to RF current flow, the tissue
resistance may be only 200 to 400 ohms; after a few seconds of desiccation or coagulation, the value
may rise to 1000 to 3000 ohms.
The term resistance, to be exact, applies only to direct current flow. When referring to alternating
currents, the term impedance is technically correct. Impedance consists of two components:
magnitude and a phase angle, because tissue has membrane capacitances as well as resistance. At
electrosurgical frequencies, however, the phase angles are small, and simply using resistance
measurements introduces only small errors.
Current Density and Power Density

The most important variable affecting clinical outcome that is directly under the control of the physician
during electrosurgery is current density. This is simply the RF current (I) that flows through a specific
cross-sectional area (a) and is expressed as I/a. The current density can be controlled by changing the
surface contact area between tissue and an active electrode, such as a snare for polypectomy,
specifically the diameter of the snare wire.
The amount of heat produced in the tissue by RF current flow is directly proportional to the power
dissipated in the tissue. Power density (P) is related to the current density (I/a) and the tissue
resistance (R) by:
The effect of current density on power delivered to the tissue is illustrated in Figure 92. Near the
snare, the cross-sectional area of the polyp tissue is 0.25 cm2. If 0.25 A flows from the snare to the
tissue, the current density is 1.0 A/cm2 (0.25 A/0.25 cm2). At the bowel wall, the area is 1.0 cm2.
Because the same current flows through this area, the current density is four times less than that at the
snare, or 0.25 A/cm2 (0.25 A/1.0 cm2). At the patient return electrode, the 0.25 A of RF current is
collected on an area of 100 cm2. Thus, the current density is 400 times less than that at the snare
(0.25 A/100 cm2, or 0.0025 A/cm2). Because the power delivered to the tissue is proportional to the
square of the current density, the power at the bowel wall and that at the return electrode are reduced
by factors of 16 (42) and 160,000 (4002), respectively, compared with the power at the snare. The
small power level at the patient return electrode causes virtually no rise in temperature; the power level
at the bowel wall (1/16 the level at the snare) causes a temperature rise. At high generator settings,
this could cause clinically significant thermal damage to the tissue.
(319)Figure 92. Effect of area on current density during polypectomy. The small
cross-sectional area of the polyp stalk near the snare (A) creates a high current density and
high temperature. At the bowel wall (B), the area through which the current flows (arrows) is
larger and the current density less. The large surface area of the return electrode (C) results in
a small current density and little rise in temperature.
The situation depicted in Figure 92 is idealized. Tissue resistance is not uniform; the polyp and bowel
wall consist of low-resistance tissues such as blood in vessels surrounded by higher-resistance
muscle. Current does not flow uniformly through actual tissue but through multiple paths of various
resistances. For example, current leaves the snare wire on at least two parallel paths: a blood vessel
with 50 ohms resistance and the surrounding muscle with 500 ohms resistance. The majority of the
current flows through the path of least resistance, approximately 90% through the low-resistance
vessel and the remaining 10% through the muscle.
Time and Energy

Electrosurgical effects are further complicated by changes that occur over time; conditions vary from
microsecond to microsecond. As RF current heats tissue, resistance increases, thus changing the path
of least resistance and the current flow. These changes directly affect the heat produced in a given
area. As the tissue resistance changes, generator power output also varies. The situation becomes
more complex as heat is removed from high-temperature areas by blood flow and tissue conduction.
Although the detailed time course of many electrosurgical conditions cannot be controlled by the
physician, methods are available by which the time course can be manipulated.
It is possible to control the duration of RF current application. The power level (P) multiplied by the time
(t) current flows through tissue is defined as the energy (E) dissipated in the tissue:
E=Pt
where E is measured in joules (J).
Although the total energy delivered to the tissue has a direct relation to thermal necrosis,10(320) the
time course of power delivery must also be considered. For example, 1 watt delivered over 20 sec (20
J) causes less destruction than 20 watts delivered over 1 sec (also 20 J) because blood flow removes
heat that is delivered slowly. Unfortunately, specific guidelines for relating the time course of power
delivered to tissue versus necrosis are complex because other variables such as blood flow have
significant effects. However, even at low levels of a few watts, power delivered through a small area
over a long time can cause tissue destruction.
Electrosurgical Waveforms
Another way to control the time course and clinical effect of electrosurgery is by choice of waveform.
Generator waveforms differ in their voltage and current characteristics over time, and these in turn
affect the physical electrosurgical effect. The electrosurgical action of the cutting waveform is produced
by heating tissue to a high temperature, that is, above 100 C, to vaporize water and explode cells.
Coagulation is produced by heating above 70 C but below 100 C in order to dry the tissue.
Coagulation is accomplished by desiccation or fulguration. In desiccation, the active electrode is in firm
contact with the tissue, whereas in fulguration, it does not touch the tissue and sparks jump from the
electrode to the tissue. The generator waveforms are optimized to accomplish these desired clinical
effects.
To some extent, the terms cut waveform and coagulate waveform are misleading. Low power settings
(low heat) in the cut mode (waveform) produce excellent contact desiccation, and high power settings
(high heat) in the coagulation mode produce a cutting action.
Cutting
Pure Cutting
Cutting is typically accomplished with a sine waveform (Figure 93A). The generator voltage must be
sufficiently high to cause ionic breakdown of air with arcing between the active electrode and the
tissue. These arcs produce high current density, causing rapid heating and tissue vaporization. If the
active electrode is placed firmly in contact with the tissue, cutting is delayed and is not initiated until the
tissue is desiccated and shrinks slightly from contact with the active electrode; at this point, arcing can
occur and the cut action begins. The cut progresses as arcing from active electrode to tissue allows
the electrode to move through the tissue.
(321)Figure 93. Typical electrosurgical waveforms over time (t): A, pure cut; B, blended cut;
C, coagulation; and D, fulguration.
Voltages of approximately 500 Vp-p are necessary to ionize air and establish arcing. Arcs occur twice
each cycle at approximately the maximum and minimum voltage of the sine wave; this produces one
arc each microsecond for a generator operating at 500 kHz. At this high arcing rate, ions produced in
the air gap between electrode and tissue and from the vaporized tissue have little opportunity to diffuse
away from the site. Therefore, the arcing occurs as a tight packet and produces a clean cut.11(322) As
the voltage is increased, the intensity of the arcing increases with greater current density; this produces
higher temperatures and greater thermal effects on the tissue adjacent to the cut. The depth of thermal
damage increases with the increasing voltage (see Figure 91), and at high voltages and current
densities, the tissue at the side walls of the cut becomes charred and the volume of necrotic tissue
becomes larger.
Blended Cutting
One method of obtaining adjacent sidewall coagulation and minimizing charring during cutting is to use
a blended cut waveform. These waveforms use higher voltages than pure cut waveforms, but the
generator does not supply current continuously as in cut mode (Figure 93B). Blended waveforms are
intermittent or interrupted; that is, the generator supplies current only part of the time during which it is
activated. Some generators have multiple blended settings. Typically, in the blended mode, the
generator supplies current during only 50 to 80% of the activation period. The term for this is duty
cycle. At 100% duty cycle, the generator supplies current in a continuous waveform, as in the pure cut
mode; at 50% duty cycle the generator supplies current during only 50% of the activation period. If
power is kept constant, lower duty cycles use higher voltages. The lower the duty cycle and the higher
the voltage, the greater the hemostatic effect. The higher voltages force current deep into the tissue,
causing increased sidewall coagulation; the lower duty cycle allows heat to be removed by blood flow
and tissue so that cutting occurs without unnecessary charring.
Figure 94 shows a comparison of waveform, voltage, and tissue effect of a pure cut and two blended
cut waveforms. All three have the same average voltage (Vavg) over time, and the average power is
therefore similar. However, because the blended cut waveforms are not supplying continuous current
yet employ higher voltages, the coagulation is deeper without charring than is the high-voltage pure cut
waveform. The ratio of peak voltage (Vp) to the Vavg (root mean square) is termed the crest factor.
(323)Figure 94. Tissue effect of various waveforms with the same average voltage (Vavg)
and increasing peak voltage (Vp): 1, pure cut; 2, blended one; and 3, blended two. Increasing
Vp increases the depth of thermal damage (d).
Coagulation
Coagulation waveforms have lower duty cycles than blended cut waveforms. Coagulation is
accomplished by two methods: (1) contact coagulation or desiccation, in which the active electrode is
touching the tissue, and (2) fulguration or spray coagulation, in which the active electrode is not
touching the tissue but rather sparks are generated between electrode and tissue. Typical coagulation
waveforms have a duty cycle less than 50% and as low as 5%.
Contact Coagulation
If an electrode is in contact with a tissue to be desiccated, a cut waveform (100% duty cycle) can be
used for coagulation provided that the peak-to-peak voltage (Vp-p) is less than 500 V (low generator
setting), so that coagulation is produced without sparking, cutting, or accompanying tissue
carbonization. In this case, the pure cut waveform produces excellent tissue desiccation; deep tissue
desiccation is produced with longer application times. This effect points out the misleading nature of
the names cut and coagulation waveforms.
The typical coagulation waveform (see Figure 93C) is also used for contact coagulation. The higher
voltage coagulation waveforms cause relatively deep tissue desiccation. As the surface tissue
desiccates, tissue resistance rapidly increases. If low settings are used, sparking does not occur. If
high voltage is used, however, sparking occurs from the electrode to the tissue, increasing the depth of
necrosis.
The sparking that occurs with the coagulation waveform differs in character from that that occurs in the
cut mode. Specifically, coagulation sparks are not confined to a compact area, but spray from the
active electrode may strike the tissue at many different points.11(324) The low duty cycle or
intermittence of the waveform is believed to be responsible for this phenomenon, as it allows the ions
to disperse between sparks; this requires the establishment of new ions and new spark paths, in
contrast to the cut mode, in which sparks follow the same ion cloud.12(325) If the voltage is sufficiently
high, carbonization of the tissue also occurs. Usually voltages vary from 500 Vp-p to several thousand
volts.
Fulguration or Spray Coagulation
Fulguration is used to treat large bleeding areas, typically multiple small vessels, such as occur in liver
resection or when retracted vessels become hidden in fissures and are not directly accessible. The
very large Vp-ps, up to 10,000 V (10 KV), create long sparks between the electrode and the tissue.
Although the voltage is extremely high (greater than 10 KV), the depth of tissue destruction is typically
shallow owing to the low duty cycle, about 5% (see Figure 93D), and lack of contact between the
electrode and the tissue. Clinically, fulguration is not as effective as contact coagulation for control of
bleeding from large vessels because tamponade or vessel coaptation cannot be achieved. The
sparking that occurs with fulguration is highly random, primarily because of the very low duty cycle. As
a spark strikes tissue, a high resistance area is created; therefore, the next spark is more likely to
strike a low-resistance area on the tissue.
Electrosurgical Generators and their Ramifications

Monopolar Systems
In monopolar electrosurgery, RF currents flow from the generator to an active electrode and then to the
surgical site. When tissue is entered, the current disperses, flows through the patient to a return
electrode with a large surface area, and then returns to the generator to complete the circuit.
Initially, all generators were ground referenced; that is, the return electrode was connected to earth
ground. As current seeks to return to the generator to complete the circuit, any break in the connection
between the patient return electrode and the generator causes current to flow through any ground path
back to the generator. These early grounded systems caused alternate-site burns if the RF
electrosurgical current returned to ground via unintentional paths (e.g., electrodes for
electrocardiography, conductive components of the operating table; Figure 95). Modern
ground-referenced generators have circuitry that monitors the integrity of the return electrode
connections. For example, a generator such as the Valleylab Force 4 compares the currents in the
active and return electrode leads and disables the generator if a substantial difference occurs.
(326)Figure 95. A ground-referenced electrosurgical generator can create alternate paths to

ground through other equipment, e.g., electrocardiography (ECG) electrodes.
Late-model electrosurgical generators are typically isolated; that is, the patient return electrode is not
connected to earth ground (Figure 96). If a break occurs in the return electrode connections, the
generator output is reduced virtually to zero and RF current does not flow through the patient to ground
by alternate paths. Alternate-site burns through ground paths are therefore eliminated. If care is not
taken to avoid inadvertently grounding the return electrode of an isolated generator, the system acts as
if it were grounded and all advantages of isolation are lost. Isolating the generator from earth ground
does increase the problem of capacitive coupling of stray RF currents in other nearby conductors (see
later in this chapter).
(327)Figure 96. An isolated electrosurgical generator eliminates alternate current paths to

ground. ECGelectrocardiography.
To avoid potential problems and complications of electrosurgery, the endoscopist should always
become familiar with the generator being used and its associated safety features.
Bipolar Systems
The patient return electrode is eliminated in bipolar electrosurgery by placing a second electrode a
small distance (several millimeters) from the active electrode. Current flows from the first electrode to
the tissue at the surgical site, through the tissue to the second electrode, and then returns to the
generator to complete the circuit. Therefore, bipolar electrosurgery confines the RF current flow to a
small volume of tissue between the two electrodes.

Classically, bipolar electrodes have been used only for coagulation and are well accepted by
endoscopists for this purpose. New bipolar electrodes have been developed for cutting. These have
met with varying degrees of clinical acceptance, however. Animal studies have demonstrated that
bipolar electrodes cut with reduced power and destroy less tissue than their monopolar counterparts
when electrodes of similar area are compared. Bipolar electrodes for clinical fulguration have not yet
been produced.
Generator Outputs
Power output and other characteristics of electrosurgical generators vary from model to model.
Endoscopists must therefore be fully aware of the output characteristics of the generator used for a
particular procedure. Many generators have arbitrary front panel output power settings, that is, a dial
marked from 0 to 10. The output at the maximum setting of 10 for monopolar cutting may be as low as
50 watts for one generator and as high as 400 watts for another. Although newer generators allow
selection of a specific power level, such as 50 watts, it must be remembered that even this is not
constant over time during cutting or coagulating.
Figure 97 shows the actual power output versus resistance for a typical multipurpose generator
(Valleylab SSE2L) in monopolar cut, blended, and coagulate modes with the generator set at 5, the
midpoint of the range of possible dial settings. Peak powers are produced in the typical tissue
impedance range, 300 to 500 ohms. However, the power output decreases for all modes of operation
as tissue impedance rises to several thousand ohms. In order to achieve cutting, cut and blended
modes produce higher powers at higher resistances than does the coagulation mode. In cut mode, the
RF current path resistances include tissue resistance at the surgical site, tissue resistance between the
surgical site and the return electrode, contact resistance of the return electrode and the patient's skin,
and all electric connections. Also included is the resistance of the arc from active electrode to tissue,
this being approximately 1100 ohms.13(328) Although spray coagulation also includes this arc
impedance, cutting action requires the delivery of higher power at high resistance in order to heat the
tissue to vaporization temperature. In contact coagulation without sparks, the power can decrease after
the tissue has been sufficiently desiccated (a tissue resistance of approximately 2000 ohms). In fact,
this effect is desirable to avoid excessive coagulation and deep necrosis.
(329)Figure 97. Generator power output (Valleylab SSE2L at a setting of 5) versus

resistance for cut, blended, and coagulation (Coag) modes.
The bipolar mode of a typical multipurpose generator has a very different power versus resistance
characteristic than does the monopolar mode. The peak power is generated at lower resistances, for
example, 100 to 150 ohms, these values being typical for the resistance of bleeding tissue between the
tines of a bipolar forceps. The power output of such a bipolar mode decreases much faster with
increasing resistance (Figure 98) than that of the monopolar mode, thus avoiding high heating and
complete vaporization of the tissue between the tines.
(330)Figure 98. Power output versus resistance for a typical multipurpose generator for
monopolar and bipolar applications. The curves were obtained at a setting of 3 in blend for the
monopolar output and a setting of 5 in the bipolar.
Special-purpose bipolar generators specific for endoscopic coagulation are available. The outputs of
these generators (e.g., CIRCON/ACMI, Stamford, CT) are optimized to coagulate bleeding tissue
between electrodes separated by several millimeters. These generators produce a maximum power
output of approximately 50 watts at 50 to 100 ohms and very little power at resistances greater than
1000 ohms (Figure 99).
(331)Figure 99. The power output versus resistance for the ACMI BICAP generator
designed for a bipolar coagulating electrode.
The bipolar output of either a multipurpose generator or a bipolar coagulation generator is
inappropriate for true bipolar cutting in endoscopy. Because the arc resistance is high and cutting
requires significant power output, effective bipolar cutting requires a power versus resistance curve
similar to that of a monopolar multipurpose generator. Several new bipolar generators are available
that are optimum for bipolar cutting (e.g., Everest Medical, Minneapolis, MN; Meditron, Hackensack,
NJ). These generators have bipolar power versus resistance characteristics analogous to those of
monopolar generators (Figure 910).
(332)Figure 910. The power output versus resistance for a bipolar generator (Everest
Medical) designed for bipolar cutting and coagulating electrodes.
Clinical Utilization
Although it is possible to give general guidelines for the clinical use of RF current to cut and coagulate
tissue, these techniques are largely a part of the art of medicine. They require subjective assessments
such as the appropriate amount of coagulation and the speed of cutting. Although understanding the
biophysics of electrosurgery aids in avoiding complications and achieving the desired outcome, the
technical art must be learned from a skilled endoscopist. It is prudent for the novice to utilize the same
equipment, settings, and techniques employed by their instructors until they develop the necessary
judgment and experience in a variety of clinical settings. The following sections are intended to
emphasize the role of electrosurgery in endoscopy and not to be a complete description of clinical
techniques.
Polypectomy
Electrosurgical polypectomy is accomplished by using one of the many commercially available
colonoscopes and a wire loop that is placed around the polyp and tightened. Three techniques are
commonly used to remove polyps with RF current:14,15(333) (1) desiccation followed by mechanical
cutting with the loop, (2) desiccation followed by electrosurgical cutting, and (3) a single-maneuver
electrosurgical cut. Some endoscopists open and close the polypectomy snare themselves, whereas
others have an assistant do this for them. The latter approach requires a substantial degree of
cooperation between endoscopist and assistant (see Chapter 5: The Gastrointestinal Assistant).
Probably the most common method of polypectomy is desiccation, at a relatively low generator setting,
of the entrapped tissue with a coagulation or blended cut waveform, followed by simple mechanical
transection with the wire loop. The necessary tissue desiccation is visually quantified by the amount of
mucosal blanching around the wire. Too little desiccation leads to bleeding during mechanical
transection; overdesiccation leads to dehydrated tissue that can be difficult to transect mechanically
and may cause snare entrapment.
The second polypectomy method is similar to the first except that the mechanical cut is replaced by
electrosurgical cutting. The initial step of coagulation is the same, and again, care must be taken to
avoid overdesiccation, which limits the ability to cut the tissue electrosurgically and may cause snare
entrapment. Endoscopists use pure cut, blended cut, and coagulation waveforms to accomplish
electrosurgical cutting. Because electrosurgical cutting requires arcing from electrode to tissue, an
excessively tight grasp of the polyp with the snare actually prevents the initiation of cutting. Therefore,
loosening the snare slightly enhances the cutting action. Although this is not intuitively obvious,
loosening the snare decreases the electrode-tissue contact area and thus increases the current
density.
Polypectomies are also performed in a single maneuver by an electrosurgical cut using a blended or
coagulation waveform without prior desiccation. With this technique, cutting speed strongly influences
the outcome; fast cutting yields less hemostatic effect than slow cutting. As discussed earlier, optimum
cutting is produced when the snare is in loose contact with the polyp to allow arcing.
All polypectomy techniques are influenced by the physical diameter of the wire used in making the
snare. A direct relationship exists between wire diameter and current density; that is, the current
density varies with the surface area of the contact between electrode and tissue. Popular commercial
snare wire diameters vary from 0.012 to 0.040 inch (0.3 to 1.0 mm). At a given power setting and a
given polyp size, a small-diameter wire results in a significantly higher current density and yields faster
cutting and coagulating than a large-diameter wire. If the diameter of the snare wire is changed, the
clinical resultintentional or unintentionalcan be different from that expected. Therefore, a change in
wire diameter may require a change in generator settings, depending on the desired result.
No typical generator power settings were given in the preceding discussion. Guidelines are general at
best, given the multiple variables involved in polypectomy. In fact, detailed power settings appropriate
to every situation cannot be given. At a specific power setting, voltage and type of waveform influence
clinical outcome. The correct power setting also depends on the amount of tissue trapped within the
snare; large polyps require higher power settings to achieve similar current densities than do small
polyps. When an equipment change is made (snare or generator), the endoscopist should seek advice
from manufacturers and other experienced endoscopists on ranges of power settings suitable to
particular clinical situations.
The polypectomy techniques discussed thus far are monopolar; current flowing from the snare to the
polyp continues flowing to the distant return electrode to complete the circuit. If the power setting is
high, the current density at the polyp base may be sufficient to cause full-thickness desiccation of the
bowel wall. Depending on the extent of necrosis, this injury can lead to the so-called postpolypectomy
syndrome or perforation. Another type of injury can also result from monopolar current flowing to the
return electrode. If the head of the polyp is in contact with the wall of the colon, the current effectively
has two paths to the return electrodean intended path through the snared tissue and an unintended
one through the head of the polyp to the opposite bowel wall (Figure 911). The current flow through
the unintended path increases as the tissue within the snare, typically the polyp stalk, is desiccated and
increases in resistance. This problem can be minimized by moving the snared polyp back and forth
during activation of the generator so that either contact with the opposite wall does not occur or the
duration of contact at any given point is not prolonged.
(334)Figure 911. Diagram of unintended current flow through the head of a polyp to the
opposite bowel wall. If the contact area is sufficiently small, a burn (darkened area) can
result.
Bipolar snares have been developed in an effort to eliminate or reduce the complications that occur in
monopolar polypectomy (e.g., contralateral wall burns and perforations). These are commercially
available from several sources (Bard, Tewksbury, MA; Everest Medical; and Meditron). Although
designs differ, all bipolar snares are composed of two wire electrodes separated by an insulator. RF
current passes from one electrode (segment or part of the snare loop) through the polyp to the second
electrode to complete the circuit. This design completely eliminates burns to the contralateral bowel
wall. Because bipolar snares cut with less power and the extent of tissue necrosis is significantly less
than with monopolar snares, the incidence of postpolypectomy syndrome and perforation may be
decreased.16(335) However, bipolar snares have not been available long enough to allow completion
of clinical comparison studies.
Sphincterotomy
Endoscopic retrograde sphincterotomy has become widely accepted since the mid-1970s. Cutting of
the sphincter of Oddi is performed by endoscopic cannulation of the papilla of Vater and passing of RF
current from a wire electrode on the cannulating catheter to a distant return electrode. This monopolar
electrosurgical procedure has become the treatment of choice for the removal of retained common
duct stones following cholecystectomy or with the gallbladder in situ (see Chapter 60: Endoscopic
Papillotomy, and Chapter 62: Calculus Disease of the Bile Ducts).
The cannulating catheter, known as a papillotome or sphincterotome, may be single or double lumen.
The single-lumen catheter carries only the electrosurgical wire electrode. A wire electrode occupies
one lumen of the double-lumen catheter and the second may be used for injection of a contrast agent
or passage of a guidewire to aid in difficult cannulations. In both catheter types, the distal end of the
catheter is bent into a bow when tension is applied to the cutting electrode wire so that the externally
located portion of the wire becomes the string of the bow (see Chapter 60: Endoscopic Papillotomy).
Sphincterotome cutting wires are available in a wide variety of designs. Wires may be stranded or
monofilament; diameters vary from 0.008 to 0.012 inch. The physical characteristics of the wire affect
its cutting characteristics. Monofilament wires have less surface area than stranded wires and for a
given diameter produce greater current density and faster cutting. Similarly, smaller-diameter wires
also produce greater current density and cut faster than those of larger diameter.
Cutting is typically performed in the blended cut mode. Because the risks of rapid cutting are bleeding
and perforation, use of more than 70 watts of power is uncommon. Many endoscopists cut with short
applications of electrosurgical current and only slight bending of the sphincterotome to achieve added
control. If cutting is slow, retraction of the wire electrode to shorten the length of wire in contact with the
papilla increases the current density and the speed of cutting. Slow cutting may also be encountered if
the papilla is fibrotic or a ductal calculus (having high electric resistance) is touching the active wire.
Repositioning the electrode in these cases usually enables cutting. Repeated application of ineffective
cutting current should be avoided, as this desiccates the tissue and raises the tissue resistance to high
levels, further slowing the cutting action. The practicing endoscopist should be familiar with both the
electric output of the generator and the cutting characteristics of the sphincterotome being used.
Gastrointestinal Bleeding
Endoscopic monopolar electrosurgical electrodes have been used for several decades to stop bleeding
from various sources in the upper gastrointestinal tract (see Chapter 28: Thermal Contact Methods for
Endoscopic Hemostasis). Although early attempts at electrosurgical control of bleeding were
associated with complications, including perforation and death,17(336) techniques and equipment
have been improved. Various distal tip configurations of monopolar coagulating electrodes are
available, but all of these are applied in a similar fashion. The electrode is used to tamponade the
bleeding vessel and coapt its walls; coagulation current then "welds" the vessel closed, producing
hemostasis. Excessive desiccation of the bowel wall can cause full-thickness burns, leading to
immediate or delayed perforation.
Bipolar electrodes were developed to reduce the tissue necrosis that occurs during monopolar
coagulation and to decrease the likelihood of a perforation. A bipolar coagulator typically consists of a
small (approximately 5 mm in length) ceramic shaft with a rounded end. Most have a central irrigation
channel. The electrodes at the distal end of the shaft have various configurations, the two most
common being (1) four or six metallic strips (about 1 mm wide) arranged longitudinally and equally
spaced around the end of the shaft with opposing strips electrically connected to form two sets of
electrodes and (2) two metallic strip electrodes equally spaced in a spiral pattern that winds around the
ceramic shaft. These endoscopic bipolar electrodes confine current flow to a much smaller volume of
tissue than do monopolar electrodes. Generators for use with bipolar electrodes have power versus
resistance curves that are optimized for the low resistance of blood; that is, maximum power output
occurs at approximately 50 ohms. The waveforms are pure sine wave (cut) with low voltage to
minimize the development of sparks and thus to provide contact desiccation. Bipolar systems have
been widely accepted by endoscopists.
Other thermal probes (e.g., "heat probe") are available for treatment of gastrointestinal bleeding
lesions. However, these are not actually electrosurgical devices because current does not flow in any
manner from the probe to the tissue, although the device is heated electrically (see Chapter 28:
Thermal Contact Methods for Endoscopic Hemostasis).
Complications of Electrosurgery
Alternate-Site Burns and Return Electrode Burns
The likelihood of producing a serious electrosurgical burn is minimized during endoscopy because
patients are awake and can experience pain. However, such burns can occur in patients who are
heavily sedated or unconscious or have spinal cord injuries.
As stated earlier, ground-referenced generators have the return electrode connected to earth ground.
Any discontinuity from the return electrode to the generator causes the electrosurgical current to flow
through any alternate path to ground in order to complete the circuit. Virtually any conductive material
can become incorporated into an alternate path, including metal tables, electrocardiographic
electrodes, and other monitoring equipment. Most newer grounded generators provide circuits that
monitor current in the return electrode and sound an alarm or disable the generator if any fault occurs
in the return circuit. Isolated machines were developed to eliminate alternate-site burns that result from
unintended ground paths. Because the return circuit is isolated from ground by a transformer, any
break in this circuit that prevents completion of the normal current path stops the generator output.
Ground-referenced generators and many isolated generators do not monitor the patient return
electrode contact. The surface area of the return electrode in contact with the patient must be relatively
large if current density is to remain low in the adjacent tissue. If the area of contact is small, as may
occur if the return electrode is placed improperly, all of the electrosurgical current flows through a
confined area, markedly increasing the probability of an RF burn. Some newer-model isolated
generators monitor the contact area between the return electrode and the patient by means of circuits
called return electrode monitors, or REMs. An REM electrode is actually two electrodes; that is, the pad
is split. The electrode-to-patient contact resistance is compared between the two separate return
electrode sections of the pad; if a significant difference exists, the generator is disabled. Thus, if the
return electrode is only partially in contact with the patient, a situation in which current density could be
high enough to cause a burn, the REM circuit eliminates all power output. REM generators and REM
return electrodes virtually eliminate return electrode burns. However, care must be taken not to use a
standard return electrode with an REM generator, as this defeats the monitor circuit and increases the
potential for burns.
Capacitive Coupled or Stray Currents

An RF current flowing through a conductor can induce an RF current on nearby conductors, even with
intact insulation and no direct electric connection between the two conductors. This phenomenon is
termed capacitive coupling. A capacitor is composed of two conductors separated by an insulator.
Alternating voltages across such a structure causes currents to flow through it. In endoscopy, several
capacitors have the potential to create deleterious effects for the patient and the endoscopist.
One capacitor is always present during endoscopic electrosurgery. This comprises the active wire lead
(e.g., polypectomy snare or sphincterotome) and the inner metal frame of the endoscope.18(337) As
current flows through the active lead within the endoscope, RF currents are induced in the metal frame
of the instrument. Because these appear to leak from the active wire, they are often called leakage or
stray currents. The induced currents may then flow from any exposed metal on the endoscope to the
endoscopist, who is unknowingly part of the path to the return electrode. Stray currents can result in
burns to the endoscopist, most likely from the eye piece of older fiberoptic endoscopes. The stray
currents can exceed 0.1 A. Given this high value, it is surprising that such burns do not occur more
often. However, another capacitive path couples most of the endoscope frame current to the patient.
This capacitor comprises the endoscope frame, the covering insulation of the instrument, and the
conductive patient tissue. Therefore, the majority of the stray current is coupled to the patient. Most
endoscopists have at times received a small burn or "shock" from these stray currents. (Although RF
currents cannot "shock," small burns are often described as a "shocklike" sensation.)
Patients are also at risk for a burn from stray currents if any part of the endoscope metal frame is
exposed by a small crack in the covering insulation. To reduce the potential for burns to endoscopists
and patient, a safety cord is used; this connects the endoscope to the patient return electrode and
removes all the stray currents induced on the endoscope and returns them safely to the generator.
However, if the return electrode is not properly in place on the patient, the endoscope becomes the
return electrode and current flows from patient to endoscope and returns to the generator via the safety
cord. Any point of contact between the patient and the metal endoscope frame is likely to result in a
serious burn. Therefore, the use of safety cords is no longer common and is not recommended.
Another capacitor that has the potential to create RF burns is found during guidewire-assisted
sphincterotomy.19(338) Many endoscopists prefer to cannulate the papilla using a guidewire,
especially in technically difficult cases. A double-lumen sphincterotome (one lumen for the cutting wire
and the other for the guidewire) can be advanced into the bile duct over the guidewire. The capacitor is
composed of the sphincterotome cutting wire, the insulating septum between the two lumens of the
catheter, and the metal guidewire. As current flows in the cutting wire, current is induced in the
guidewire by capacitive coupling; the resistance (impedance) for this capacitive coupling is as low as
2000 ohms. Many endoscopists leave the guidewire in the common bile duct during sphincterotomy in
order to maintain access. Only an insulated guidewire intended for use during electrosurgery is suitable
for this purpose; other types of guidewires must not be used during electrosurgery.
Any flaw in the insulation of a guidewire that exposes the metal wire to tissue during sphincterotomy
allows induced currents to flow to the patient. At typical operating settings in the blended mode, only a
few watts are induced from the guidewire. However, if the cutting wire is not in contact with tissue when
the generator is activated (open-circuit activation), up to 15 watts may be induced in the guidewire.
Although 1 or 2 watts can probably be tolerated, larger values over the small area of the flaw in the
guidewire insulation can yield power densities that may perforate the common bile duct. Therefore, if
the guidewire is to be left in the common bile duct during electrosurgery, high power settings,
prolonged generator activation, and activation with the active wire not in contact with the tissue must be
avoided.
Coupled currents between the active wire and the guidewire are reduced by thicker guidewire
insulation. Thus, sheathed guidewires offer an increased margin of safety if a guidewire is to be left in
the common bile duct during sphincterotomy. Some guidewires are simply spray coated with Teflon.
Although Teflon is an insulator, the coating has many imperfections that expose the underlying
metal.20(339) All guidewires to be left in place during cutting should be thoroughly examined for
defects in insulation.
A more serious potential clinical problem is the direct electric connection (shorting) between active wire
and guidewire; this may place full generator output power onto the guidewire. This cannot occur
without insulation failure on the guidewire, and sheathed guidewires are significantly less susceptible to
shorting. A guidewire should never be placed in the same channel of the sphincterotome as the cutting
wire because this invites electric shorting between the two wires. Unfortunately, even double-channel
catheters do not provide total isolation between the wires. In fact, new double-channel catheters have
been shown to have a significant incidence of septum defects that allow electric connection between
active wires and guidewires. Therefore, it is prudent not to leave a Teflon spray-coated guidewire or a
guidewire with any exposed metal in the common duct during electrosurgical cutting.
Demodulation Current
Although typical electrosurgical currents do not cause muscle or nerve stimulation, all surgeons are
aware that striated muscle fasciculations occur during electrosurgery. Endoscopists also report that
sedated patients occasionally complain of uncomfortable sensations during electrosurgical cutting even
though the bowel does not possess the necessary pain receptors. Both phenomena are explained by
the demodulation of RF currents, that is, the changing of high-frequency current into low-frequency or
direct current, both of which can stimulate muscle and nerve.21(340)
Demodulation occurs when RF current flows through nonlinear circuits. The most important of these
during electrosurgery is sparking. (Although tissue is also nonlinear, its contribution to demodulated
currents is minor). Therefore, occasional muscle and nerve stimulation should be expected during
electrosurgical cutting and fulguration. Unfortunately, such stimulation is inherent to RF currents and
cannot be prevented. However, no long-term consequences occur as a result of these demodulated
low-frequency and direct currents.
Electric Interference
Electrosurgical generators are essentially radio transmitters and therefore radiate small amounts of
radio waves into the atmosphere. These radiated waves as well as capacitively coupled currents may
interfere with other electronic devices. Electronic devices used in the same area with electrosurgical
generators usually have circuits designed to filter any RF interference. Interference may occasionally
occur with some video systems, but it is usually minor. However, in three clinical situations, RF
interference has potentially deleterious effects. These involve patients with pacemakers, implanted
defibrillators, and intracardiac catheters. Although patients undergoing open heart surgery are at
greatest risk for problems related to RF interference, endoscopists should be aware of these potential
complications.
Cardiac pacemakers are of two types: demand and fixed rate. Demand pacemakers sense the cardiac
cycle and deliver a stimulating pulse in the absence of an R wave. Because electrosurgical currents
may be interpreted as R waves, the pacemaker may thereby be inhibited. Demand pacemakers placed
in asynchronous mode are less susceptible to electrosurgical interference, although failures can occur
even in this mode.22(341) Fixed-rate pacemakers deliver stimulus at a predetermined rate, but their
internal circuitry can also be susceptible to RF interference generated during electrosurgery.
Although modern pacemakers have protective circuitry, programming to control RF interference, or
both, any pacemaker may have its intended output modified during electrosurgery. However,
endoscopic procedures can be performed safely on patients with pacemakers provided that several
simple precautions are observed. The return electrode should be placed away from the heart, that is, in
such a way that the current flow from the surgical site to the return electrode does not include the
heart. Endoscopists are encouraged to use low-power and low-voltage waveforms on patients with
pacemakers to reduce possible interference. Bipolar electrosurgery also decreases the potential for
pacemaker-associated problems. Electrocardiographic monitoring of patients with pacemakers is also
strongly recommended. Manufacturers can also provide specific details on the susceptibility of their
products to electrosurgical interference.
Potential RF interference complications also exist for patients with defibrillator implants. Electrosurgical
interference can cause unnecessary shocks or defibrillator output failure. Usually, inactivation of the
defibrillator and close patient monitoring are recommended. As with pacemakers, endoscopists are
encouraged to seek advice from the specific manufacturer and consultation with the patient's
cardiologist before performing an electrosurgical procedure.
Cases of ventricular fibrillation during electrosurgery have been reported in patients undergoing
thoracic surgery with its associated multiple monitoring systems, including central venous lines.23(342)
It has been demonstrated that ventricular fibrillation and a myocardial burn can occur via a
pressure-sensing catheter acting as the RF return for a grounded generator after a break in the patient
return circuit. Although such a situation is uncommon during endoscopy, it could occur when
electrosurgery is performed in patients with severe bleeding who are connected to various central
catheters and monitoring devices.
Bowel Gas Explosion

Although uncommon, injuries from bowel gas explosion during polypectomy have been reported since
this procedure was first introduced (see Chapter 80: Indications, Contraindications, and Complications
of Colonoscopy).24,25(343) Bowel gas is composed of carbon dioxide, nitrogen, oxygen, hydrogen,
and methane. The proportions of these gases are quite variable among individuals and depend in part
on diet and disease state. The combustible constituents of bowel gas result from swallowed air
(oxygen) and bacterial metabolism (hydrogen and methane).26(344) Nearly half of all patients with
unprepared colons have an explosive mixture of bowel gases.27(345) Therefore, preventive measures
to decrease concentrations of combustible gases must be taken prior to electrosurgical sparking.
Carbon dioxide insufflation during colonoscopy decreases the risk of explosion by decreasing the
concentration of combustible gases. However, meticulous bowel preparation prior to electrosurgery
results in extremely low concentrations of combustible gases. Several methods are available (see
Chapter 81: Technique of Colonoscopy). Only the use of carbohydrate alcohols (e.g., mannitol,
sorbitol) is contraindicated for bowel preparation because chemical breakdown by colonic bacteria
produces large amounts of hydrogen gas.28(346)
REFERENCES
1.
d'Arsonval MA. Action physiologique des courants alternatits. Comp Rend Soc Biol Paris
1891;43:2836.
2.
Riviere AJ. Action des courants de haute frequence et des efflures du resonateur Oudin sur
certaines tumeurs malignes. J Med Interne 1900;4:7767.
3.
Clark WL. Oscillatory desiccation in the treatment of accessible malignant growths and minor
surgical conditions. J Adv Therapy 1911;29:16983.
4.
Nagelschmidt CF. Lehrbuch der Diathermie fr rzte und Studierende. Berlin: J Springer.
1913.
5.
Beer E. Removal of neoplasms of the urinary bladder. A new method employing high
frequency (Oudin) currents through a catheterizing cystoscope. JAMA 1910;54:176874.
6.
Wyeth GA. The endotherm. Am J Electrother Radiol 1924;42:1867.
7.
Cushing H, Bovie WT. Electrosurgery as an aid to the removal of intracranial tumors. Surg
Gynecol Obstet 1928;47:75184.
8.
Hill AV, Foulerton RS, Katz B, et al. Nerve excitation by alternating current. Proc Roy Soc
Lond [Biol] 1936;121:74133.
9. LaCourse JR, Miller WT, Vogt M, et al. Effects of high frequency current on nerve and muscle
tissue. IEEE Trans Biomed Eng 1985;32:826.
10. Tucker RD, Kramolowski EV, Bedell E, et al. A comparison of urologic application of bipolar
versus monopolar five French electrosurgical probes. J Urol 1989;141:6625.
11.
McLean AJ. The Bovie electrosurgical current generator: Some underlying principles and
results. Arch Surg 1929;18:186373.
12.
Kelly HA, Ward GE. Electrosurgery. Philadelphia, London: WB Saunders, 1932;3944.
13.
Peares JA. Electrosurgery. Medical Instrumentation and Clinical Engineering Series. London:
Chapman & Hall 1986;87.
14. Cohen LB, Waye JD. Treatment of colonic polypsPractical considerations. Clin
Gastroenterol 1986;15:35976.
15.
Tedesco FJ. Colonic polypectomy. In Silvis SE (ed). Therapeutic Gastrointestinal Endoscopy.
New York: Igaku Shoin, 1985;26988.
16. Tucker RD, Platz CE, Sievert CE, et al. In vivo evaluation of monopolar versus bipolar
electrosurgical polypectomy snares. Am J Gastroenterol 1990;85:138690.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Koch H, Pesch HJ, Bauerle H, et al. Experimentelle Untersuchugen und klinische Erfahrunge
zur electrodogulation blutende Laisonan im oberen Gastrointestinaltrakt. Fortschr Endosk
1972;10:6771.
Barlow DE. Endoscopic application of electrosurgery: A review of basic principles.
Tucker RD, Voyles CR, Silvis SE. Capacitive coupled stray currents during laparoscopic and
endoscopic electrosurgical procedures. Biomed Instrum Technol 1992;26:30311.
Johlin FC, Tucker RD, Ferguson SD. The effect of guide wires during electrosurgical
sphincterotomy. Gastrointest Endosc 1992;38:53640.
Tucker RD, Schmitt OH, Sievert CE, et al. Demodulated low frequency currents from
electrosurgical procedures. Surg Gynecol Obstet 1984;159:3943.
Mangar D, Atlas GM, Kane PB. Electrocautery induced pacemaker malfunction during
surgery. Can J Anaesth 1991;38:6168.
Hungerbuhler RF, Swope JP. Ventricular fibrillation associated with use of electrocautery.
JAMA 1974;230:4325.
Geddes LA, Tacker WA, Cabler P. A new electrical hazard associated with electrocautery.
Med Instrum 1975;9:1123.
Carter HG. Explosion in the colon during electrodesiccation of polyps. Am J Surg
1952;84:5147.
Bigard MA, Gaucher P, Lassalle C. Fatal colonic explosion during colonoscopic polypectomy.
Levitt MD, Lasser RB, Schwartz MA, et al. Studies on a flatulent patient. N Engl J Med
1976;295:2602.
Ragins H, Shinya H, Wolf WI. The explosive potential of colonic gas during colonoscopic
electrosurgical polypectomy. Surg Gynecol Obstet 1974;138:5546.
Chapter 10 Laser Physics and Laser-Tissue Interactions

(347)
(348)
JOSEPH A. IZATT, PH.D.
The first brilliant flash of deep red laser light emitted from a perfect crystal of ruby at Hughes Research
Laboratories in 1960 marked the beginning of a renaissance in the study of light and its applications in
science, technology, and medicine.
Emerging from its origins in atomic physics, the laser has come to pervade disciplines ranging from
biophysics to satellite communications. Lasers are now used routinely in medical practice to cut,
shape, remove, and examine biologic tissue. Commercialization of lasers and delivery systems
specifically designed for particular applications has kept pace with the transformation from research
tool to clinical necessity. Increasing interest in lasers and optical techniques in medicine is due to a
synergism among the unique capabilities of lasers, rapid technologic advances in optoelectronic
technologies, and trends in clinical medicine.
The matchless qualities of lasers that make the light they generate especially attractive for medical
applications include high directionality (spatial coherence), wavelength selectivity and tunability, and
the generation of high peak and average optical powers. Indeed, specific laser sources can exhibit
each of these characteristics individually as well as simultaneously to a higher degree than any other
known sources of radiation.1(349)
The directionality of laser light is important for beam handling, shaping, focusing, and efficient coupling
into optical fibers. This characteristic makes possible the direct linkage of lasers to the rapidly
advancing optical fiber technology. That technology is critically important to the development of modern
clinical methods that are less invasive and more economical such as endoscopy and laparoscopy. The
tunability of laser light makes possible the selection of optimum wavelengths for specific applications,
in effect tuning the absorption and scattering properties of the tissue to match the problem at hand.
This characteristic complements recent advances in optical detector technology such as
charge-coupled devices and intensified detector arrays and enables real-time spectroscopic imaging of
tissue surfaces. The development of high-speed computers and related information technologies
permits the use of elaborate algorithms for therapy, imaging, and diagnostics in real time, thereby
amplifying the impact of advances in optoelectronic technology and matching new developments in
medical instrumentation and informatics.
In many medical applications, the use of laser light provides unique benefits over conventional
techniques for diagnosis and therapy. Therapeutic applications include ablation (vaporization) (see
Chapter 43: Palliation of Malignant Obstruction: Lasers and Tumor Probes), coagulation and
cauterization (see Chapter 29: Laser Therapy), and laser-induced photochemistry (see Chapter 19:
Photodynamic Therapy). Although it is a capital-intensive technology, use of lasers nevertheless offers
compelling advantages under the proper circumstances. Because no contact is required between the
light delivery system and the target tissue, the risk of infection is greatly reduced. For ablation and
coagulation, lasers generate significantly less radio frequency interference than alternative techniques
and are thus compatible with other medical devices such as pacemakers. For cutting applications,
laser ablation exerts only minimum pressure on tissue, in contrast to the scalpel blade, which cuts by
exerting pressure in excess of the local tissue tensile strength. The activation with laser light of
phototoxic exogenous chromophores, although still in the early phases of research and development,
is a promising technique for the treatment of cancer and some immune disorders.
Novel applications of lasers are also being developed for diagnosis. Laser scattering and fluorescence
techniques are the basis for sensors that monitor blood flow, local tissue temperatures, and blood gas
concentrations. Laser scanning confocal microscopy permits high-resolution "optical sectioning" of
microorganism preparations and tissue, enabling detailed subcellular examination without tissue
sectioning. It has been demonstrated that laser-induced fluorescence spectroscopy can be used in
tissue to indicate compositional and architectural changes associated with the onset and development
of atherosclerotic plaques and dysplastic tissue in the urinary bladder and gastrointestinal tract (see
Chapter 17: Biomedical Tissue Spectroscopy, and Chapter 18: Spectroscopic Diagnosis and
Treatment of Gastric Cancer). This most recent category of medical laser applications is in many ways
the most promising. When merged with other sophisticated diagnostic technologies, it has the potential
to elucidate the origins of certain diseases and to pinpoint their earliest manifestations.
Historical Background
Albert Einstein first suggested the theoretical concept of coherent "negative absorption" or amplification
of light in atomic systems in 1917.2(350) Because this idea arose before full development of the
quantum theory, which is necessary for accurately describing atomic systems, it remained only a
tantalizing speculation for many years. During the 1950s, groups in the United States and the Soviet
Union demonstrated weak microwave amplification in gases.3(351) Then, in 1958, Arthur Schawlow
and Charles Townes4(352) published an influential paper suggesting the possibility of amplification of
optical radiation. This initiated a flurry of research to demonstrate and expand this important concept
that continues to this day.
The first operational lasers were demonstrated in 1960 by Theodore Maiman et al.5(353) at Hughes
Research Laboratories (using ruby exposed to powerful flashlamps to produce red laser light) and Ali
Javan et al.6(354) at the Bell Telephone Laboratories (using an electric discharge in a gaseous mixture
of helium and neon to produce infrared, and later red, light). In the subsequent years, thousands of
different laser systems have been developed, generating perhaps more than 1 million discrete optical
wavelengths.
Attempts to use lasers for medical purposes occurred almost immediately after development of the first
operational devices. In 1961, Zaret et al.7(355) performed iris and retinal photocoagulation using the
ruby laser in rabbit eyes, and by 1963, several groups had developed ruby laser delivery systems for
retinal photocoagulation.8,9(356) Following development of the argon-ion laser, it was recognized to be
a superior source for achieving controlled photocoagulation, owing to its continuous wave nature and
strong absorption by melanin and hemoglobin (Figure 101).10,11(357) Clinicians quickly realized that
high-power lasers could be used for tissue removal or ablation, and by 1965, several groups had
begun using ruby and neodymium:yttrium-aluminum-garnet (Nd:YAG) lasers to vaporize melanomas,
sarcomas, and carcinomas in animal models.1218(358)
(359)Figure 101. Absorption spectrum of several important tissue chromophores, together

with the wavelength positions of the most widely used medical lasers. ArFargon-fluoride;
KrFkrypton-fluoride;
Nd:YAGneodymium:yttrium-aluminum-garnet;
XeClxenon-chloride;
He-Nehelium-neon;
Al2O3aluminum-oxide;
GaAsgallium-arsenic; Er:YAGerbium:yttrium-aluminum-garnet; CO2carbon dioxide;
HbO2oxyhemoglobin; HPDhematoporphyrin derivative. (Data from Boulnois J-L.
Photophysical processes in recent medical laser developments: A review. Lasers Med Sci
1986; 1:4766; data for the ultraviolet portion of the H2O spectrum is from Hale GM, Querry
MR. Optical constants of water in the 200nm to 2000nm wavelength region. Appl Opt 1973;
12:55565.)
The first attempt to use a photocoagulative effect to achieve hemostasis in the gastrointestinal tract
was by Goodale et al.19(360) in 1970, who reported success at delivering CO2 laser light at 10.6 mm
through a rigid gastroscope for control of bleeding from gastric erosions. Similar results were achieved
shortly afterward by the groups of Dwyer,20,21(361) Kiefhaber,22(362) and Frhmorgen23(363) using
argon-ion and Nd:YAG lasers coupled to optical fibers that were passed through flexible endoscopes.
Many controlled studies have since been published detailing efficacy of photocoagulative laser therapy
in the treatment of actively bleeding lesions2427(364) and angiodysplasia28,29(365) (see Chapter 29:
Laser Therapy).
The second prominent application of lasers in gastroenterologic practice is endoscopic laser ablation
(photoablation, vaporization) of neoplastic tissue. Although related conceptually to animal experiments
conducted 20 years previously, photoablation of malignant tumors in humans was not substantially
achieved until the technology for convenient delivery of high-power Nd:YAG laser light through flexible
endoscopes became commercially available in the early 1980s. Mostly palliative in intent, laser
vaporization has been well characterized for treatment of neoplasms of the esophagus,30(366)
stomach,31,32(367) duodenum, ampulla of Vater, colon, and rectum33(368) (see Chapter 43:
Palliation of Malignant Obstruction: Lasers and Tumor Probes.25(369)).
Two additional applications of lasers with relevance to gastrointestinal disease deserve mention.
Although the concept of photodynamic therapy was decades old, several groups began in the early
1970s to use lasers as directed light sources to activate toxicity in tumor-localized drugs, primarily
porphyrin derivatives.3438(370) Interest in this field has been heightened by the development of new
phototoxic drugs with improved selectivity for dysplastic tissue and reduced side effects such as skin
photosensitivity (see Chapter 19: Photodynamic Therapy). In addition, scattered and tentative reports
have described the use of very small doses of laser light for selective destruction of malignancy in
tissue cultures and animal models, a process termed biostimulation.3942(371) The mechanisms
underlying these effects remain poorly understood.
Laser Principles
Light Amplification by Stimulated Emission of Radiation
A laser is a device that generates and amplifies light by use of energy level transitions of quantum
mechanical systems enclosed within an electromagnetic resonant cavity.1,43,44(372) The word laser
is an acronym for light amplification by stimulated emission of radiation. Light in this context is defined
as the portion of the electromagnetic spectrum that the human eye perceives as colors, plus
neighboring frequencies in the ultraviolet and infrared ranges. Visible light frequencies range from
about 4.3 to 7 1014 cycles/sec, with corresponding wavelengths of about 400 to 700 nm (1 nm =
10-9 meter; see Figure 101). The essential components of a laser are illustrated in Figure 102.
These are (1) a lasing or gain medium, (2) a process for pumping or exciting the gain medium, and (3)
an electromagnetic oscillator or cavity enclosing the ends of the gain medium that intercepts and
recirculates light through the medium.
(373)Figure 102. Schematic diagram of a laser. Light amplified through the process of
stimulated emission is recirculated through the laser cavity (i.e., the gain or lasing medium) by
the end mirrors. Light partially transmitted through one end mirror constitutes the output laser
beam.
The gain (lasing) medium is the heart of the laser. The active elements of the gain medium are
quantum mechanical particles that exhibit discrete energy levels and absorb or emit radiation when
changing energy levels according to Planck's Law:
E=hv
where E is the energy shift of the transition (in Joules [J]), h is Planck's constant (h = 6.626 10-34
J-sec), and is the frequency of the radiation (in cycles/sec). Wavelength () is related to frequency
through the relation
where c is the speed of light in the medium (c = 3 108 m/sec in air). Lasing action has been
demonstrated with many different types of particles, including free electrons, atoms, molecules, and
electron-hole pairs in semiconductors. Although the details of a particle's energy level structure depend
on the type of particle and its surroundings, the underlying principles of laser action do not. Therefore,
for the purposes of this discussion, atom is used as a generic term for whatever species exhibits laser
gain.
Microscopic processes occurring within the lasing medium are illustrated schematically in Figure 102.
A pumping process such as an electric discharge is used to excite some of the atoms from their
ground state into higher energy states. Each excited state has a characteristic lifetime, which can be
thought of as the average time until the atom decays to a lower energy state or back to the ground
state. Several avenues are available for decay. The atom can simply release the excess energy in the
form of heat, this being termed nonradiative decay. The atom can also decay by emitting a photon, a
unit of radiation with energy equal to the transition energy and thus a frequency determined by Planck's
Law. This emission process, termed spontaneous emission, is responsible for the glow from
fluorescent lights, neon signs, and cathode-ray tubes. Light produced in this process is emitted in all
directions and can either escape the laser medium, scatter off other atoms or particles, be absorbed by
unexcited atoms, or if it encounters another excited atom, induce that atom to emit a second photon
through a process termed stimulated emission. This last process is unique in that the resulting pair of
photons is perfectly matched in frequency and phase and propagate together in precisely the same
direction. They can be conceptualized as an electromagnetic wave with the same frequency and phase
as the incident photon but with twice the amplitude.
The basic concept of laser action, as suggested by Einstein in 1917, is as follows: If a collection of
atoms could be specially prepared such that the population division between two of their energy states
is "inverted," that is, with the majority in the upper state, then through the process of stimulated
emission, the medium could produce and amplify a light beam of remarkable directionality, spectral
purity, and intensity. Physically, this means that the collection of atoms must be coaxed into a state in
which the probability of an incident photon of a given energy inducing stimulated emission is higher
than the sum of the probabilities of that same photon being absorbed, scattered, or otherwise lost. This
is a very unlikely situation in nature (in fact, it can be described as a condition of negative temperature),
and the difficulty of developing a means to trick a collection of atoms into this state was in large part
responsible for the long delay between Einstein's conceptualization of stimulated emission gain and the
first operational lasers in 1960.
One of the simplest and most practical methods for establishing the population inversion necessary for
laser action is the so-called four-level pumping scheme, as illustrated in Figure 103. Atoms in their
ground energy state are excited by a pump source into a highly excited state or group of states with
very short lifetimes for nonradiative transition to the upper lasing state, E2. The state E2 has a very
long lifetime, so that atoms tend to collect there and to be available for stimulated emission. Initial
photons emitted through spontaneous emission that encounter particles in this E2 state induce
stimulated emission, thereby reducing the particle's energy to the lower excited state E1 and producing
amplified light with a frequency of
(374)Figure 103. A four-level laser energy level diagram. Atoms are induced into a
condition of population inversion between two discrete energy levels by taking advantage of
differences between the lifetimes of states E1 and E2.
The state E1 has a short lifetime for deexcitation to the ground state, so that atoms do not collect in
this state and reabsorb the precious laser light. There must also be no other energy levels available at
an energy E2 - E1 above the ground state, or else absorption from the ground state would overwhelm
laser emission.
Laser action occurs when a few photons initially produced by spontaneous emission are amplified
many times by stimulated emission. If the condition of population inversion could be established in a
large enough collection of atoms, lasing would occur spontaneously. This has in fact been observed in
huge interstellar gas clouds and planetary atmospheres.45(375) A more practical laser, and one that
better preserves and enhances the directional and spectral purity of stimulated emission, is made by
enclosing a relatively small volume of particles within an optically resonant cavity. In a typical laser
oscillator (see Figure 102), mirrors placed at the ends of a lasing medium redirect only light emitted
within a limited solid angle back through the lasing medium to induce formation of a highly directional
laser beam. Laser light is extracted from the laser cavity either by allowing one of the mirrors to
transmit a slight fraction so that some light escapes with every round trip or by some other means such
as the use of very fast optical switches that periodically open the laser cavity to permit escape of short
pulses of light.
Unique Properties of Laser Light

Several qualities of laser light distinguish it from light from any other source and make it particularly
useful in medical applications. These qualities, illustrated in Figure 104, all result from the directional
and phase-maintaining characteristics of the stimulated emission process, preserved and enhanced by
the surrounding optical resonant cavity.
(376)Figure 104. Unique properties of laser light.

Spatial Characteristics
The spatial characteristics of laser beams are largely determined by the laser optical cavity. Typical
simple lasers emit relatively narrow beams (a few millimeters in diameter) with an intensity profile that
decays radially as a Gaussian function (see Figure 104). The regenerative feedback mechanism
inherent in laser gain lends laser beams the property of being highly collimated; that is, the beam
diverges very slowly with distance. A typical collimated laser beam has a divergence on the order of 1
milliradian, although lasers can be constructed with microradian divergence. The directionality of laser
light makes it easy to direct laser beams with mirrors and to manipulate them with inexpensive,
low-numerical-aperture optics. The collimation of laser light is also directly related to the minimum
attainable focused spot size; a well-collimated beam can be focused with an appropriate lens to a spot
with a diameter comparable to a single wavelength. This degree of spatial coherence allows for laser
beams to be coupled into optical fibers and to be used for very high resolution microscopy.
Frequency Characteristics
Lasers emit light over a very narrow wavelength range, owing both to the limited gain bandwidth
intrinsic to the stimulated emission process and to the restricted number of discrete electromagnetic
modes that can oscillate within the resonant optical cavity. Although the visible spectrum extends over
approximately 300 nm, the spectral width of a gas laser line can be narrower than 0.01 nm. The
spectral purity or monochromaticity of laser light, defined as the ratio of the frequency spread of the
emitted light to the frequency of the emitted light itself, can thus range as high as 1 part in 106. In
optical systems, this property eliminates chromatic aberration as a design consideration and aids in
focusing the beam to the smallest spot sizes. In medical applications, this feature can be used to excite
or target very specific species (ranging down to specific molecules) in biologic tissue in order to identify
them or alter their function.
Radiometric Characteristics
For many medical applications, the most important property of a laser beam is the power or energy it
carries. Radiometry is the quantitative measurement of electromagnetic radiation; definitions of
radiometric terminology for medical lasers are summarized in Figure 104. Pulsed lasers are rated in
terms of their energy per pulse, termed radiant energy (in Joules). Continuous-wave lasers are rated in
terms of radiant power, which is energy per unit time (1 watt = 1 J/sec). The radiant power in a typical
continuous-wave laser beam available in an operating room, although not exceptional in itself (a few
watts are comparable to the optical output of an incandescent light bulb), is truly remarkable in that all
of this power resides within a highly collimated beam and a very narrow spectral line width. Typical
laser radiant powers range from less than 1 milliwatt for compact disk player or laser printer
semiconductor lasers, to kilowatts of power in industrial CO2 lasers used for materials processing. The
highest reported laser powers have occurred only instantaneously during extremely short laser pulses.
Research lasers have generated in excess of 1015 watts of instantaneous power,46(377) a factor
more than 100 times the total installed electric generating capacity of the United Statesalthough only
for a few tenths of femtoseconds, equal to 1 millionth of 1 billionth sec.
The extent of laser effects on materials usually depends on the area over which the laser beam has
been focused. Thus, irradiance is defined as the laser radiant power divided by the beam area.
Specific Laser Systems

During the past several decades of laser development, the total number of discrete laser wavelengths
demonstrated probably exceeds 1 million. In terms of sheer numbers of wavelengths available,
molecular systems represent the largest category by far of laser gain media owing to their large and
complicated energy level structures. However, the number of laser systems that are commercially
viable for specific applications is much smaller, and a relative handful of systems are the workhorses
for medical and industrial applications.47,48(378) A selection of laser systems relevant to medical
applications and their key characteristics are provided in Table 101. Although relatively few types of
lasers are in current use, it is important to note that appropriate laser designs can often be
implemented very rapidly as new applications are developed that require novel laser wavelengths and
other special characteristics.
TABLE 101
LASER
(TYPE)
Excimer
(pulsed)
Argon ion (cw)
Dye (pulsed or
cw)
Semiconductor
(pulsed or cw)
Nd:YAG (pulsed
or cw)
Laser Systems Used in Medical Applications

LASING
PARTICLES
(MEDIUM)
WAVELENGTH
(NM)
Excited dimers
(gas)
Argon ions
(gas)
193, 249,
308, 351
488, 514
Dye molecules
(liquid)
Electron-hole
pairs (solid)
Neodymium
ions (solid)
APPROXIMATE
PENETRATION
DEPTH IN
TISSUE
TISSUE EFFECTS
1 mm
Ablation with little

thermal damage
Coagulation
4001000
110 mm
Vaporization
Fluorescence
excitation
Tunable absorption
7501550
110 mm
1064
5 mm
1100 m
CO2 (pulsed or
CO2 molecules 10,600
10 m
cw)
(gas)
cwcontinuous wave; Nd:YAGneodymium:yttrium-aluminum-garnet.
Sensitizer triggering
Deep penetration
Volume heating
Vaporization
Plasma and shock
Precise cutting
TYPICAL
APPLICATION
Corneal surgery
Laser angioplasty
Panretinal photocoagulation
Hemostasis
Cancer palliation
Fluorescence
spectroscopy
Pigmented lesions
Photodynamic therapy
Photodynamic t
Optical tomography
Hemostasis
Cancer palliation
Photodisruption
Laser microsurg
Tissue debulking
Extensions of Laser Technology

Lasers, because of their remarkable qualities, have since their discovery stimulated a renaissance in
the field of optics. Many phenomena would not have been discovered without the use of laser sources;
some of these phenomena form the basis of useful extensions of laser technology itself. Examples
include developments in electro-optics, acousto-optics, nonlinear optical phenomena, and fiberoptics.
The first two of these provide techniques now commonly used to perform optical switching within lasers
to create short pulses as well as techniques to direct, shift, and examine the frequency of laser
beams.1,49(379) The third category, nonlinear optics, encompasses a large range of very useful
phenomena that have substantial commercial applications.50,51(380)
Nonlinear optics describes the unusual phenomena that occur when very intense radiation propagates
through materials, including the generation of new frequencies from one laser beam and frequency
mixing of more than one laser beam. These important effects make available a wider range of
wavelengths from established laser systems and play a large role in medical devices.
The basic principle of nonlinear optics is illustrated by the example of second-harmonic generation
(Figure 105). When an electromagnetic wave propagates through a clear medium such as a crystal,
electrons within the crystal are induced to oscillate along with the wave. The induced oscillations are
termed the optical polarization of the material, and this polarization in turn induces reradiation, as is
generated by any oscillating charge. When the incident intensity is small enough that the induced
motion of the electrons is small and their motion is unencumbered by the surrounding crystal lattice,
then the electron motions are proportional to the incident field, following the sinusoidal oscillations of
the incident wave. In this case, the reradiated field is indistinguishable from the incident field, and the
light propagates unchanged through the crystal. When the incident intensity becomes high enough,
however, electrons are driven beyond their linear range of motion, and the induced polarization is no
longer sinusoidal and acquires frequency components other than the incident field frequency.
(381)Figure 105. Schematic representation of second-harmonic generation, an example of

nonlinear optics. Incident electromagnetic (EM) radiation generates asymmetric oscillation of
electrons in crystal lattices that have the appropriate symmetry properties. The oscillating
electrons reradiate at both the incident light frequency and its second harmonic.
Second-harmonic generation occurs in crystals that lack so-called inversion symmetry. In these
crystals, the structure in which the electrons are located is asymmetric, with the result that the
electrons are freer to oscillate with larger amplitudes in one direction than in the opposite. As illustrated
in Figure 105, the excursion of the electrons is different for the positive-going wave than for the
negative-going wave. In this case, Fourier analysis demonstrates that the induced polarization can be
constructed from the sum of two sinusoidal waves, one at the incident oscillating frequency and a
second at twice that frequency, plus a small direct current offset. As a result, the oscillating electrons
reradiate at both the incident wave frequency and its second harmonic. When the incident radiation is
an intense laser beam and the crystal is oriented at just the right direction so that all of the radiating
electrons are in phase for both the incident frequency and the second harmonic, a second beam at
twice the frequency of the incident light (i.e., half the wavelength) can be generated, with a significant
proportion of the power converted from one frequency to the other. Similar effects can be used for
generating higher harmonics of a laser frequency or for adding or subtracting frequencies of more than
one laser beam.
Medical Laser Delivery Systems

Developments in fiberoptic technology have greatly accelerated the use of lasers for medical
procedures. Early laser delivery systems required awkward articulated arms containing folding mirrors.
Optical fibers capable of transmitting substantial amounts of energy are now available for wavelengths
from about 300 nm to 2.5 m.48(382) Unfortunately, this wavelength range excludes CO2 radiation at
10.6 m as well as 2.9 m, where water and hence tissue absorption are even stronger (see Figure
101). Some optical fibers and hollow flexible wave guides are available for these infrared
wavelengths, but current versions tend to be unacceptably lossy and inflexible. There is an intense
effort to develop fibers suitable for this important wavelength range.
Optical fibers modify the properties of laser light that they conduct depending on their construction and
size. Most commercially available optical fibers trap light by means of total internal reflection at the
interface between cylindrical glass layers with dissimilar indices of refraction (Figure 106). The inner
cylinder in such "stepindex" fibers is termed the core, and the surrounding glass is termed the cladding
(see also Chapter 2: Flexible Endoscope Technology: The Fiberoptic Endoscope). Total internal
reflection occurs over a range of angles of incidence at the core-cladding interface if the index of
refraction of the core is greater than that of the cladding. Two major classes of fibers are available.
Single-mode fibers have very small core diameters (<10 m) and fully preserve the spatial coherence
of fully coherent laser beams; that is, the light exiting the fiber can be recollimated by a lens to a beam
with diffraction-limited divergence and Gaussian cross section. Unfortunately, single-mode fibers are
delicate and can conduct only limited optical powers. Also, the laser beam coupling site is very
sensitive to shocks and vibrations. For these reasons, the majority of medical laser delivery systems
use multimode fibers with core diameters of 50, 100, or 200 m or larger. Laser light propagating in a
multimode fiber loses a substantial portion of its spatial coherence and, if the fiber is long enough,
emerges with the optical power evenly spread over the surface of the output core face and at the
maximum angle allowed by the fiber. A plot of the distribution of laser energy emerging from a bare
multimode fiber is provided in Figure 107. Various focusing and diffusing tips for medical use are
available, the choice depending on the application.
(383)Figure 106. Schematic of a step-index optical fiber in cross (a) and longitudinal (b)
section. Light incident on the core-cladding interface at angles less than the critical angle () is
totally internally reflected throughout the length of the fiber (see Chapter 2: Flexible
Endoscope Technology: The Fiberoptic Endoscope). the index of glass used in the core
and cladding.
(384)Figure 107. Plots of the spot size and light irradiance as a function of distance from a
fiber tip. The calculations are performed for a 600-m core, 0.2 numerical aperture multimode
fiber carrying a total radiant power of 1 watt.
Laser-Tissue Interactions
Laser-Tissue Interaction Regimes
The interaction of laser light with biologic tissue is a complex process that depends on many factors:
the wavelength, power, and illuminating geometry of the incident laser beam; the optical, thermal, and
mechanical properties of the tissue; and the biologic host response.48,52(385) Laser-tissue interaction
processes may be grossly divided into qualitative regimes that are determined primarily by the intensity
of the laser beam and its interaction time with the tissue (Figure 108).48,53(386) Within each of these
regimes, variations in the resultant effects can be observed, depending mostly on the laser wavelength
and the optical properties of the specific tissue type.
(387)Figure 108. Laser-tissue interactions map including the primary laser-tissue interaction
regimes. CWcontinuous wave. (Data from Boulnois J-L. Photophysical processes in recent
medical laser developments: A review. Lasers Med Sci 1986; 1:4766.)
Very high laser intensities combined with short laser pulses (<10-6 sec) result in ionization of tissue
(and the air or water above it) and the formation of a plasma. This process, termed photodisruption in
ophthalmology, is useful for cutting of transparent membranes within transparent media in intraocular
surgery.47(388) It is also the mechanism responsible for the initiation of photoacoustic fracturing of
urinary calculi with pulsed dye lasers.54(389) However, this regime is not optimum for soft tissue
vaporization because it has poor efficiency, the light can damage optical fibers, and dosage is difficult
to control because the initiation mechanism is very sensitive to tissue and medium impurities.
The physical basis of many surgical applications of lasers is tissue ablation. The use of moderately
high laser beam intensities with exposure times of milliseconds to seconds results in rapid deposition
of heat and subsequent tissue vaporization or decomposition, processes that characterize thermal and
thermoacoustic ablation. The ablation process depends substantially on the thermodynamic properties
of water, the primary constituent of tissue. For short laser pulses, the thermodynamics of water
specifies that under conditions of inertial confinement, when not enough time is available for material to
move, the energy densities typically deposited during ablation can lead to the generation of
tremendous pressure. Ablation occurs when this pressure exceeds a threshold that depends on the
structural properties of the tissue.55(390) For longer pulses or quasicontinuous-wave exposures,
inertial confinement is not achieved, but ablation can be modeled as a rapidly moving vaporization front
at relatively constant pressure.56(391)
Low-intensity or defocused laser beams used for periods of a few seconds have a more gentle effect
of simply heating without tissue removal. This regime is used for tissue coagulation and cauterization.
Careful modeling of heat deposition and diffusion has led to the prediction and observation of specific
tissue transformations that occur at particular temperatures (Table 102).
TABLE 102
TEMPERATURE
( C)
45
60
80
100
210+
Relationship Between Temperatures and Tissue Event

HISTOLOGIC EVENT
Cell death, edema, endothelial
damage, vasodilation
Protein coagulates
Denatured collagen contracts,
blood vessels constrict
Tissue water boils
Dehydrated tissue burns
ENDOSCOPIC MANIFESTATION
Erythema, edema cuff
Tissue turns gray-brown, blood turns black
Tissue "puckers"
Vaporization causes divot

Blackened tissue disappears, leaves glowing
embers
From Fleischer D. Lasers in gastroenterology. Am J Gastroenterol 1984; 79(5):40615.
Finally, nonthermal interactions are produced at low laser intensity levels delivered over minutes or
hours. These interactions are strongly coupled to the tissue history and host response and include
various forms of photochemically mediated reactions such as photodynamic therapy. This regime is
also used in laser-tissue diagnostics, including tissue spectroscopy and imaging.
Laser-Tissue Interaction Processes

The laser-tissue interaction regimes relevant to gastroenterologic practice and research are those
occurring at moderate to low laser powers and exposure times (see Figure 108), that is, vaporization,
coagulation, photochemistry, and tissue diagnostics. Each of the regimes depend on many component
processes, including tissue optics, heat diffusion, ablation, and ionization; in general, the higher the
intensity, the greater the complexity of the processes and correspondingly less knowledge of the
mechanisms involved. The topics of tissue optics and heat diffusion are indispensable to a basic
understanding of any laser-tissue interaction and are discussed in the following sections. The
remaining topics are beyond the scope of this chapter; interested readers are directed to the
references concerning these very active areas of research.
Tissue Optics
All of the laser-tissue interaction processes depicted in Figure 108 are initiated by the deposition of
light within tissue. Many tissue components absorb laser light, the absorbing tissue species being
termed chromophores. The absorption spectra of some important tissue chromophores are included in
Figure 101. Light is absorbed in tissue through the action of a variety of chromophores, including
water, various proteins, the hemoglobin in blood, and tissue pigments such as melanin.48(392)
Because absorption of almost all tissue constituents is lowin the 600- to 1300-nm region of the
spectrumlaser light within this wavelength range penetrates more deeply into tissue. This range of
wavelengths is therefore referred to as the therapeutic window. Chromophore absorption spectra are
typically a strong function of the laser wavelength and have relatively sharp features because they
depend on the details of the atomic or molecular structure of the absorbing species (see Figure 101).
The strength of light absorption in tissue, which is proportional to the sum of the concentration of all of
the chromophores at a given wavelength, can be characterized by an absorption coefficient, a, which
is related to the probability of absorption per unit length in the tissue. The inverse of a is the
mean-free-path for absorption, la, which can be interpreted as the average distance an incident photon
propagates in the tissue before being absorbed.
Just as in the laser medium itself, absorption of an incident photon leads to excitation of the absorbing
species, and several avenues are available for deexcitation. Excited chromophores can release their
energy as heat, this being the basis for tissue heating and ablation. Alternatively, excited
chromophores can release part of their energy as photons of a lower frequency than the laser beam.
This is the process of fluorescence. Monitoring of emitted fluorescence can be used to detect and
measure the concentrations of chromophores specific to particular disease states.
Laser light is also scattered in biologic tissue.57(393) In scattering events, incident laser photons
essentially bounce off tissue components, losing in this process little or none of their energy. The
amplitude of scattering in tissue varies less strongly with wavelength than does absorption by tissue
chromophores, and it does not exhibit such sharp features. The direction in which photons are
scattered, however, depends very strongly on the size of the scattering particle and its orientation with
respect to the incident photon. In bulk tissue, the average scattering direction is effectively randomized
and can be characterized statistically only as the average deviation from the forward direction. The
strength of scattering in a tissue type can also be characterized by a scattering coefficient (s) and its
inverse (ls), the mean-free-path for scattering. This latter quantity represents the average distance a
photon travels in tissue between scattering events.
A significant characteristic of tissue optics is the fact that in almost all tissues except the clear part of
the eye, la is much longer than ls; that is, scattering dominates absorption. The ratio of scattering
length to absorption length in tissue is typically in the range of 0.01 to 0.1, indicating that photons
average 10 to 100 scattering events before being absorbed. The dominance of scattering over
absorption indicates that incident laser light is drastically redistributed within tissue, typically taking on a
roughly exponential shape, as illustrated in Figure 109. The exponential constant or penetration depth
of the laser light within the tissue is a complicated function of the absorption and scattering
mean-free-paths as well as the mean scattering deviation angle and is very difficult to predict
theoretically as a function of chromophore and scattered concentrations.58,59(394) For this reason,
the best estimations of light distribution profiles are obtained using Monte Carlo calculations on
computers (the profiles in Figure 109 were generated in this way).60(395)
(396)Figure 109. Contour plot of light irradiance levels in tissue resulting from illumination
with a flat-profile laser beam of diameter 5 mm with 10 watts radiant power. The calculation
was performed using a Monte Carlo computer program with absorption and scattering
coefficients of a = 1 cm-1 and s = 100 cm-1 as parameters. The data summarize the
simulated trajectories of 10,000 photons.
Typical measured values for tissue optical properties as well as theoretically predicted values for the
penetration depth and other tissue physical properties are listed in Table 103. Measured optical
penetration depths in most tissues of clinical interest range from fractions of a micron in the deep
ultraviolet and mid-infrared regions, up to several millimeters in the near-infrared. An excellent
tabulation of all known published tissue optical properties as of 1990 may be found in Cheong et
al.61(397)
TABLE 103
Summary of Tissue Physical Properties
PARAMETER
Absorption coefficient
Absorption mean-free-path
Scattering coefficient
Scattering mean-free-path
Total attenuation coefficient
Albedo
Penetration depth
SYMBOL
a
Ia
PROPERTIES
Optical Properties
Typical values in therapeutic window (
~ 6001300 nm)
0.01 mm-1
1 - 100 mm
s
Is
1 - 100 mm-1
0.01 - 1 mm
t=a+s
s / (a + s)
1 - 100 mm-1
0.5 - 0.99
Specific heat per unit volume
Heat diffusivity
Speed of sound
Maximum elongation
Maximum stress
c
s
Thermal Properties
3.96 x 10-3 (J/ C-mm3)
0.106 mm2/sec
Mechanical Properties
Soft Tissue
Calcified Tissue
1500 m/s
2800 m/s
25100%
2.5%
600 bars
5002500 bars
Tissue Thermal Processes

Absorption of visible and near-ultraviolet light in tissue occurs by exciting electronic transitions in tissue
molecules, primarily proteins and lipids. Near-infrared and mid-infrared light is absorbed via
vibrational-rotational excitations in biomolecules and in water. The lifetimes of both the electronic and
the vibrational-rotational states are much less than laser pulse durations of hundreds of nanoseconds
or longer. As soon as the initial excited state distributions thermalize, the energy deposited by the laser
pulse may be considered to be heat, and its dissipation is described by classic thermal diffusion. Thus,
once an estimate of a laser light deposition profile has been secured, centuries-old analytical
techniques developed for thermal diffusion may be used to describe the subsequent heat flow.62(398)
A very useful prediction of heat diffusion theory provides an approximate prediction of the characteristic
time for heat diffusion in tissue. This quantity , indicates how short a laser pulse duration must be to
ablate without causing thermal damage in peripheral tissue; its inverse determines the maximum
repetition rate for avoiding a cumulative rise in local temperature with successive laser exposures. The
expression for is calculated under the assumption that the diameter of the tissue spot under
irradiation is larger than the penetration depth of the laser light in the tissue.48(399) It is
where is the heat diffusivity of tissue, provided in Table 103. A plot of this characteristic time as a
function of penetration depth and the heat diffusion regimes it defines is provided in Figure 1010.
(400)Figure 1010. Illustration of laser pulse duration thermal regimes. For pulses longer
than the characteristic thermal diffusion time (solid line), thermal diffusion occurs during the
pulse and can lead to significant heat damage peripheral to the illuminated spot.
Summary
Laser technology offers distinct advantages in many medical applications, both diagnostic and
therapeutic. Commercialization of lasers and delivery systems specifically designed for particular
applications plays an important role in acceptance of new techniques as well as in the laser
marketplace itself. Although technology for laser applications in many medical disciplines is maturing,
an abundance of exciting prospects for significant advances remain. The rapid development of
semiconductor lasers, which have already redefined much of the optoelectronics industry, will be soon
be felt in medical therapy as new low-cost devices begin to supplant older technology. New
developments in tissue imaging and spectroscopy coincide with rapid integration of optics and lasers
into the sensor industry in general. These have the potential to deliver significant new diagnostic
capabilities in a cost-effective manner.
REFERENCES
1.
2.
3.
Siegman AE. Lasers. Mill Valley, CA: University Science Books. 1986.
Einstein A. On the quantum theory of radiation. Physikal Z 1917;18:1218.
Gordon JP, Zeiger HJ, Townes CH. The maserNew type of microwave amplifier, frequency
standard, and spectrometer. Phys Rev 1955;18:126474.
4.
Schawlow AL, Townes CH. Infrared and optical masers. Phys Rev 1958;112:19409.
5.
Maiman TH. Stimulated optical radiation in ruby. Nature 1960;187:4934.
6.
Javan A, Bennett WR Jr, Herriott DR. Population inversion and continuous optical maser
oscillation in a gas discharge containing a He-Ne mixture. Phys Rev Lett 1961;6:10610.
7.
Zaret MM, Breinin GM, Schmidt H, et al. Ocular lesions produced by an optical maser (laser).
Science 1961;134:15258.
8.
Campbell CJ, Rittler MC, Koester CJ. The optical maser as retinal photocoagulator: An
evaluation. Trans Am Acad Ophthalmol Otolaryngol 1963;67:5867.
9.
Kapany NS, Peppers NA, Zweng HC, Flocks M. Retinal coagulation by lasers. Nature
1963;199:1469.
10. L'Esperance FA Jr. An ophthalmic argon laser photocoagulation system: Design,
construction, and laboratory investigations. Trans Am Ophthalmol Soc 1968;66:827904.
11. Little HL, Zweng HC, Peabody RR. Argon laser slit lamp retinal photocoagulation. Trans Am
Acad Ophthalmol Otolaryngol 1970;74:8596.
12.
McGuff PE, Deterling RA Jr, Bushnell D, et al. Laser radiation of malignancies. Ann N Y Acad
Sci 1965;122:74757.
13.
Minton J, Ketcham A, Dearman JR. Tumoricidal factor in laser radiation. Surg Forum
1964;15:33556.
14.
Minton JP, Ketcham AS. The laser, a unique oncolytic entity. Am J Surg 1964;108:8458.
15.
Minton JP, Ketcham AS, Dearman JR, et al. The application of pulsed, high-energy laser
radiation to multiple intra-abdominal tumor implants in experimental animals. Surgery
1965;58:1221.
16.
Minton JP, Ketcham AS, Dearman JR, et al. The effect of neodymium laser radiation on two
experimental malignant tumor systems. Surg Gynecol Obstet 1965;120:4817.

17. Mullins F, Hoye R, Ketcham AS, et al. Studies in laser destruction of chemically induced
primate hepatomas. Am Surg 1967;33:298303.
18.
McGuff PE, Bushnell D, Soroff HS, et al. Studies of the surgical applications of laser (light
amplification by stimulated emission of radiation). Surg Forum 1983;14:1435.
19. Goodale RL, Okada A, Gonzales R, et al. Rapid endoscopic control of bleeding gastric
erosions by laser radiation. Arch Surg 1970;101:2114.
20. Dwyer RM, Yellin AE, Craig J, et al. Gastric hemostasis by laser phototherapy in man. A
preliminary report. JAMA 1976;236:13834.
21. Dwyer RM, Haverback BJ, Bass M, Cherlow J. Laser-induced hemostasis in the canine
stomach. JAMA 1975;231:4869.
22.
Nath G, Gorish W, Kreitmair A, Keifhaber P. Transmission of a powerful argon laser beam
through a fiberoptic flexible gastroscope for operative gastroscopy. Endoscopy 1973;5:213.
23.
Frhmorgen P, Reidenbach J, Bodem R, et al. Experimental examinations on laser
endoscopy. Endoscopy 1974;6:116.
24. Silverstein FE, Protell RL, Gilbert DA, et al. Argon vs. neodymium YAG laser photocoagulation
of experimental canine gastric ulcers. Gastroenterology 1979;77:4916.
25. Rutgeerts P, Van Trappen G, Broeckaert L, et al. Controlled trial of YAG laser treatment of
upper digestive hemorrhage. Gastroenterology 1982;83:4106.
26. Vallon AG, Cotton PB, Laurence BH, et al. Randomised trial of argon laser photocoagulation
in bleeding peptic ulcers. Gut 1981;22:22833.
27. Swain CP, Storey DW, Northfield TC, et al. Controlled trial of argon laser photocoagulation in
bleeding peptic ulcers. Lancet 1981;2:131316.
28.
Jensen DM, Machicado GA, Tapia JI, et al. Endoscopic treatment of hemangioma with argon
laser in patients with gastrointestinal bleeding. In Atsumi K, Nimsakul N (eds). LaserTokyo
'81. Tokyo: Inter Group Corp, 1981;S205.
29.
Etienne J, Raimbert P, Dorme N. Successful Nd:YAG laser photocoagulation in Rendu
Osler's and Willebrand's diseases. In Atsumi K, Nimsakul N (eds). LaserTokyo '81. Tokyo:
Inter Group Corp, 1981;S267.
30. Fleischer D, Kessler F. Endoscopic Nd:YAG laser therapy for carcinoma of the esophagus: A
new form of palliative treatment. Gastroenterology 1983;85:6006.
31. Swain CP, Bown SG, Edwards D, Salmon PR. Neoplastic gastric outflow tract obstruction
relieved by argon laser at endoscopy. Gastrointest Endosc 1984;30:312.
32. Richey GD, Dixon JA. Ablation of atypical gastric mucosa and recurrent polyps by endoscopic
application of laser. Gastrointest Endosc 1981;27:2247.
33.
Bowers J. Laser therapy of colonic neoplasms. In Fleischer D, Jensen D, Bright-Asare P
(eds). Therapeutic Laser Endoscopy in Gastrointestinal Disease. Boston: Martinus Nijhoff,
1983;13950.
34.
Figge FH, Weiland GS, Manganiello LO. Cancer detection and therapy: Affinity of neoplastic,
embryonic, and traumatized tissues for porphyrins. Proc Soc Exp Biol Med 1948;68:6401.
35.
Lipson RL, Blades EJ. The photodynamic properties of a particular hematoporphyrin
derivative. Arch Dermatol 1960;82:50816.
36. Diamond I, McDonagh AF, Wilson CB. Photodynamic therapy of malignant tumors. Lancet
1972;2:11757.
37. Dougherty JJ. Activated dyes as anti-tumor agents. J Natl Cancer Inst 1974;52:13336.
38. Tomson SH, Emmett EA, Fox SM. Photodestruction of mouse epithelial tumors after oral
acridine orange and argon laser. Cancer Res 1974;34:31247.
39.
Fu-Shou Y. The effect of laser irradiation on human liver carcinoma cells (BEL-7404 line). In
Atsumi K, Nimsakul N (eds). LaserTokyo '81. Tokyo: Inter Group Corp, 1981;378.
40.
Nakata M, Ohnishi T, Kamikawa K. Anti-tumor effects of near ultraviolet irradiation. In Atsumi
K, Numsakul N (eds). LaserTokyo '81. Tokyo: Inter Group Corp, 1981;224.
41.
Karu T. Photobiological fundamentals of low-power laser therapy. IEEE J Quant Elect
1987;QE-23:170317.
42.
Itzkan I, Bourgelais DBC. Laser wound healing can be explained by the photodissociation of
oxyhemoglobin. Lasers Lif Sci 1988;2:24955.
43.
Marcuse D. Principles of Quantum Electronics. New York: Academic Press. 1980.
44.
Yariv A. Quantum Electronics. 3rd ed. New York: John Wiley & Sons. 1989.
45.
Elitzur M. Physical characteristics of astronomical masers. Rev Mod Phys 1982;54:122560.
46.
Ditmire T, Perry MD. Terawatt Cr:LiSrAlF6 laser system. Opt Lett 1993;18:4268.
47.
Steinert RF, Puliafito CA. The Nd:YAG Laser in Ophthalmology. Philadelphia: WB Saunders.
1985.
48.
Katzir A. Lasers and Optical Fibers in Medicine. New York: Academic Press. 1993.
49.
Yariv A. Optical Electronics. 2nd ed. New York: Holt, Reinhart, and Winston. 1985.
50.
Bloembergen N. Nonlinear Optics. New York: Benjamin. 1965.
51.
Shen YR. Principles of Nonlinear Optics. New York: John Wiley & Sons. 1984.
52.
Muller G (ed). Medical Optical Tomography: Functional Imaging and Monitoring. Bellingham,
WA: Society of Photo-Instrumentation Engineers. 1993.
53.
Boulnois J-L. Photophysical processes in recent medical laser developments: A review.
Lasers Med Sci 1986;1:4766.
54.
Teng P, Nishioka NS, Anderson RR, Deutsch TF. Optical studies of pulsed laser
fragmentation of biliary calculi. Appl Phys B 1987;42:738.
55.
Albagli D, Dark M, Perelman LT, et al. Photomechanical basis of laser ablation of biological
tissue. Opt Lett 1994;19:16846.
56. Partortovi F, Izatt JA, Cothren RM, et al. A model for laser ablation of biological tissue using
laser radiation. Lasers Surg Med 1987;7:14154.
57.
Ishimaru A. Wave Propagation and Scattering in Random Media. New York: Academic Press.
1978.
58. Jacques SL, Prahl S. Modeling optical and thermal distributions in tissue during laser
irradiation. Lasers Surg Med 1987;6:494503.
59. Welch AJ, Yoon G, van Gemert MJC. Practical models for light distribution in laser-irradiated
tissue. Lasers Surg Med 1987;6:48893.
60.
Wang L, Jacques SL. Monte Carlo Modeling of Light Transport in Multi-Layered Tissues in
Standard C. Houston: M D Anderson Cancer Center. 1992.
61.
Cheong WF, Prahl SA, Welch AJ. A review of the optical properties of biological tissues. IEEE
J Quant Elect 1990;QE-26:216685.
62.
Carslaw HS, Jaeger JC. Conduction of Heat in Solids. Oxford: Clarendon Press. 1947.
Chapter 11 Principles of Endoscopic Ultrasonography

(401)
DAVID M. PAUSHTER, M.D.
Endoscopic ultrasonography is a technologic development of extraordinary application for the
endoscopist. It requires, however, a considerable degree of endoscopic expertise, along with additional
knowledge of the anatomy and ultrasonographic representations of the chest and abdomen. To be
able to integrate this information and perform the procedure competently, the endoscopist must begin
with a practical knowledge of ultrasonography.1(402)
Properties of Ultrasound Waves

Ultrasound waves within the human body are mechanical pressure waves that are propagated in
longitudinal fashion. They are characterized by their frequency (the distance between successive
waves) and their wavelength (the span of a wave in space). The speed of ultrasound waves in a
medium is called the acoustic velocity and is the product of the wavelength times the frequency.
Therefore, of necessity, ultrasound waves of high frequency must have short acoustic wavelengths.
The acoustic velocity of ultrasound waves in the human body is relatively constant, measuring 1540
m/sec. As shall be seen later, variations in acoustic velocity have great import for ultrasound imaging.
The acoustic impedance of a tissue is related to the tissue's mechanical properties and is defined as
its density times the acoustic velocity. Acoustic impedance is a primary predictor of echo reflection
between two different tissues.
Ultrasound imaging is accomplished when pulsed ultrasound is produced by a transducer and is
reflected by tissue interfaces in the human body. Reflected sound is received by the transducer.
Because acoustic velocity within soft tissues is assumed to be constant at 1540 m/sec and the time
between sound pulse transmission and reception is known, the distance between the transducer and
the reflecting object or surface may be calculated. In this sense, time is equated to distance in
ultrasound imaging.
In addition to the positional information provided by the time delay between sound transmission and
reception, the strength or amplitude of the received echoes is displayed on a monitor. Typically, a black
background is used, with a gray scale shading toward white representing increasing echo amplitude.
Interaction of Ultrasound with Matter

Displayed echoes on an ultrasound screen are created by an acoustic impedance mismatch between
adjacent tissues. Large, smooth surfaces, such as the liver capsule, are specular reflectors that
transmit a portion of the ultrasound beam into deeper soft tissues and reflect the rest. Specular
reflectors are visualized only when the incident ultrasound beam is perpendicular to the reflector
surface (Figure 111). Smaller reflectors result in more diffuse echoes that are scattered in all
directions, with a small proportion returning to the transducer receiver (Figure 112). These diffuse
reflectors form the basis for imaging organs and organ boundaries within the human body.24(403)
(404)Figure 111. Specular reflector. A, When sound leaving the transducer (T) strikes a
specular reflectorsuch as that which occurs between two tissue interfaces (T1Tr)in a
perpendicular fashion, a portion of beam will be propagated distally and a portion will be
reflected back to the transducer for display (dashed arrow). B, If the incident beam is not
perpendicular to the specular reflector, the reflected portion of the beam will not be returned
to the transducer (dashed arrow), and the angle of incidence will equal the angle of reflection.
(405)Figure 112. Smaller reflectors. Scattering from smaller tissue reflectors occurs in all
directions. A small portion of the beam will be reflected back to the transducer (T) for display.
When sound enters the human body, it can be absorbed, refracted, or reflected. Attenuation of the
ultrasound beam is due primarily to tissue absorption, with energy conversion to a small amount of
heat.24(406) Refraction is the change in direction of an ultrasound beam that occurs as it crosses
different tissues with different acoustic velocities. Refraction, which may occur at a soft tissue-fat
interface within the abdomen, results in a misregistration of the reflector position on the ultrasound
monitor (Figure 113).
(407)Figure 113. Refraction. Sound waves (arrows) traveling through two tissues (T1, T2)
with differing acoustic velocities may misregister the position of a reflector on the television
monitor owing to refraction. Light graytissue interface; black ovalactual reflector
position; dark gray ovalprojected position on monitor.
Attenuation of the ultrasound beam is directly related to transducer frequency; the reduction in intensity
with increasing frequency is logarithmic. Therefore, evaluation of deeper structures within the human
body requires lower-frequency transducers despite their limited resolving powers. Ultrasound
equipment compensates for ultrasound beam attenuation by boosting echoes returning to the
transducer from deeper tissues. This represents the time-gain compensation (TGC) curve, which can
be manipulated by the operator. The TGC curve prevents echoes arising in deeper structures from
appearing as decreased in amplitude owing to beam attenuation. In addition, the amplitude of returning
echoes is often compressed because of ultrasound monitor limitations. The degree of compression
can be manipulated in a postprocessing fashion to enhance subtle echo amplitude differences
between tissues.
Instrumentation
Originally, ultrasound information was displayed in amplitude mode (A-mode), with a vertical axis
representing the amplitude of returning echoes and the horizontal axis distance. Current ultrasound
equipment uses a brightness (B-mode) display on the cathode ray or television tube, with brightness
representing received echo amplitude. Representation of echo amplitudes in shades of gray is called
gray scale imaging. Returning echoes are stored and processed by a scan converter, which then
displays the spatial origin of echoes and gray scale amplitude on the monitor screen.4(408)
The original B-mode scanners stored information on spatial location of echoes and echo amplitude
formed during a prolonged scan time. The image was displayed in a static fashion. All modern
ultrasound equipment is capable of acquiring data at a rate fast enough to produce a real-time image
with true temporal resolution. Because all echoes from a pulse of ultrasound must be returned to the
transducer before the next pulse is generated, a limitation is imposed on the information that can be
displayed. As the frame rate increases, temporal resolution is improved but spatial resolution
decreases. Therefore, real-time ultrasound imaging represents a compromise between spatial and
temporal information.2,3(409)
Transducer Technology
Ultrasound transducers contain piezoelectric crystals that are deformed when a voltage is applied.
Therefore, ultrasound waves are produced with a frequency corresponding to the alternating voltage
source. Of equal importance are the returning echoes from the human body. These exert mechanical
pressure on the piezoelectric crystals, which in turn induces a voltage. This voltage, when amplified, is
the basis for echo display on a television monitor. Most piezoelectric crystals are manufactured and
can be shaped for different transducer specifications.24(410)
Many types of ultrasound units and transducers are available for real-time imaging. The simplest is a
mechanical sector that usually employs a single oscillating element and a traditional triangular field of
view. The element motion may be expanded to 360 degrees in certain transducers for specific
purposes, such as endoscopic ultrasonography.
A linear array transducer is composed of multiple elements that are fired sequentially along a relatively
broad surface. Focusing of the ultrasound beam is better than with mechanical sector transducers, and
visualization of the tissues adjacent to the transducer is improved. The curved linear array is similar but
provides a more accessible scan head than does linear array equipment.
Phased array transducers use time delays between excitation of adjacent elements and reception of
returning echoes to electronically steer the ultrasound beam. Phased array transducers produce a
sector-shaped field of view and have the advantages of a relatively small size and electronic focusing.
Annular array transducers employ concentric ring elements with mechanical scanning. This
arrangement also allows electronic focusing, which is not possible with single-crystal transducers.
Determinants of Image Quality

Spatial resolution of an ultrasound image is determined by the number of lines per displayed frame. As
previously noted, the lines per frame and therefore spatial resolution must decrease as frame rate
increases. Imaging blurring occurs as a result of echo averaging of smaller reflectors and reception of
echoes from strong receptors at multiple points. Small sample volumes diminish blur inherent in
ultrasound images and improve spatial resolution.
Resolution has both axial and lateral components. Axial resolution refers to the ability to distinguish
echo reflectors in the same direction of the ultrasound beam. Axial resolution is improved by
shortening or damping the ultrasound pulse (Figure 114). High-frequency transducers produce brief
pulses and thus improve axial resolution. Lateral resolution refers to the ability to separate echo
reflectors that are perpendicular to the ultrasound beam. Lateral resolution is intimately related to
transducer beam width. As the beam width decreases, the sample volume and lateral resolution
improve. Narrower beam patterns are achieved with higher-frequency transducers, resulting in an
increase in lateral resolution.
(411)Figure 114. Axial resolution. A, A short ultrasound pulse is able to distinguish two
echo reflectors oriented in the same direction as the ultrasound beam. B, With a longer
ultrasound pulse, axial resolution is degraded, and the two echo reflectors are displayed as a
single, elongated echo complex on the monitor. Ttransducer.
The resolution of an ultrasound beam is not constant through its field of view. Although axial resolution
does not vary substantially throughout the image, the width of the ultrasound beam changes and
results in perceptible changes in lateral resolution (Figure 115). The transducer focal zone represents
the portion of the ultrasound beam that produces the best spatial resolution (Figure 116). The focal
length is the distance from the transducer face to the narrowest portion of the focal zone. The near
field represents the distance from the transducer face to the start of the focal zone (Figure 117). In
the near field, returning echo information is distorted for unfocused transducers and does not represent
the actual pattern of reflectors in this region.24(412)
(413)Figure 115. Lateral resolution. Two reflectors perpendicular to the ultrasound beam
axis are best discriminated when beam width is minimized. In the widest portion of the beam,
two echo reflectors may be displayed on the monitor as a single echo complex. Ttransducer.
(414)Figure 116. Focused transducer. The focal zone represents the narrower portion of the
ultrasound beam with optimum spatial resolution. The focal length is the distance from the
transducer (T) to the narrowest portion of the focal zone.
(415)Figure 117. Unfocused transducer. The ultrasound beam is composed of a near field
and a far field. Reflector information obtained in the near field may be distorted.
Ttransducer.
Although the focal length and focal zone can be manipulated, the length of the focal zone increases
with increasing focal length. As the transducer aperture increases, the transducer focus improves with
a narrower beam, but focal length decreases. These principles have important implications for
transducer design and selection for clinical applications. A transducer designed with a short focal
length to evaluate structures close to the transducer has a correspondingly short focal zone.
Single-crystal transducers are limited by a fixed focal zone; therefore, good spatial resolution is
achieved in only a very narrow range of depths. Single-element transducers are focused by means of
an acoustic lens or curved piezoelectric crystal design. Multielement transducers allow the operator to
select the focus zone most appropriate to the structures being interrogated. More complex focusing
techniques provide optimum focusing characteristics throughout all depths of the ultrasound image.
Ultrasound Artifacts
Artifacts occur commonly in ultrasound imaging and must be recognized by the operator in order to
produce accurate, high-quality images and avoid interpretive errors. In addition, artifacts often provide
useful information that helps to characterize structures visualized during the course of the examination.
Shadowing occurs when the ultrasound beam is completely reflected or absorbed at a tissue interface.
Typically, this is seen with gallstones and is represented by a lack of sound or a black streak deep to
the reflecting structure (Figure 118). Gas also may produce acoustic shadowing that is characterized
as "dirty" owing to reverberation and ring-down artifact. Reverberation occurs when a pulse of
ultrasound is reflected multiple times, therefore delaying its return to the transducer. This is displayed
as artifactual echoes in the region of shadowing. A ring-down or "comet-tail" artifact occurs when a
reflector prolongs the train of echoes returning to the transducer. This type of artifact is typically seen
with air or metal and has also been described with cholesterol deposits and adenomyomatosis
involving the gallbladder.
(416)Figure 118. Shadowing. When the ultrasound beam is completely reflected or absorbed
at a tissue interface, such as calcification (Ca), absence of sound distally will result in acoustic
shadowing compared with surrounding tissues (long arrows).
Fluid-filled structures produce enhanced transmission of sound. In essence, fluid attenuates the sound
beam less than adjacent tissues; therefore, an increase in transmitted sound deep to the fluid structure
is seen (Figure 119). This results in increased amplitude of returning echoes, with a width
corresponding to the width of the fluid collection. Fluid structures also cause refraction of the
ultrasound beam with bending inward of the sound. This also produces brighter echoes deep to a fluid
structure. At the edges of a cyst or other well-defined fluid collection, refractive shadowing occurs
owing to bending of the edge of the sound beam away from the center of the beam and a resultant
decrease in echo intensity.
(417)Figure 119. Increased sound transmission. Sound originating from the transducer
(black arrows) is not significantly attenuated by fluid as is sound in surrounding tissues (gray
arrows). This results in apparent increased echo amplitude beyond a fluid collection.
A mirror-image artifact occurs at highly reflective interfaces. A portion of the returning ultrasound beam
may be delayed in returning to the transducer and produce a mirror image of the reflecting surface
displayed at a different depth. If the ultrasound beam is wider than an interrogated fluid collection,
external reflectors may be projected into the fluid. This is commonly seen as artifactual sludge within
the gallbladder. Although the speed of ultrasound is assumed to be constant in soft tissues, acoustic
velocity does vary and can produce misregistration of reflector positions. Fat within the human body
has an acoustic velocity approximately 6% less than that of other soft tissues. The location of the fatty
reflectors is misrepresented in the ultrasound image, and reflectors deep to the fat are also
misregistered owing to prolonged ultrasound beam transit time.
Finally, owing to inhomogeneity in the ultrasound beam, off-axis artifacts may occur where reflectors
outside the beam return echoes to the transducer. This results in an artifactual display of these
reflectors within the ultrasound image.
Principles of Doppler/Duplex Sonography

When ultrasound waves strike moving reflectors, such as red blood cells, the returning frequency of
the reflected wave is shifted. Such Doppler-shifted frequencies are measured in the kilohertz range, as
opposed to frequencies in the megahertz range in diagnostic ultrasound imaging. Doppler equipment
can measure these frequency shifts and, by application of the Doppler equation, produce an accurate
estimate of blood velocity.
Continuous-wave Doppler ultrasound produces a continuous signal, and a separate receiving
transducer receives this signal for Doppler-shifted frequencies. These Doppler-shifted frequencies are
displayed graphically as frequency versus time. Continuous-wave Doppler ultrasound provides no
positional information and may interrogate multiple blood vessels in its course simultaneously.
Pulsed Doppler ultrasound works on the same principle as ultrasound imaging: A limited pulse of
ultrasound is transmitted and received by a single transducer, and the time delay of the returning pulse
provides positional information. This gives the operator the ability to select a particular sample volume
depth but cannot provide exact positional information unless it is coupled with ultrasound imaging.
Duplex sonography is a combination of pulsed Doppler ultrasound and real-time imaging. This allows a
visual depiction of sample volume location, and the size of the sample is adjustable as well. By directly
visualizing sampled vasculature, the angle between the transmitted ultrasound beam and the blood
flow axis, termed the Doppler angle, can be accurately characterized, thus producing highly accurate
velocity tracings. Duplex sonography systems rapidly oscillate between imaging and Doppler
capabilities to provide apparent real-time characteristics, but competition between these two
components causes loss of detail.5,6(418)
Color Doppler Imaging

Color Doppler imaging provides a visual, color-encoded representation of flow superimposed on a
two-dimensional ultrasound image. A process of autocorrelation is usually used in color Doppler
imaging to detect relative motion over time. This approximates pulsed Doppler sampling, in which the
entire field of view is summated by multiple gates to provide a large sample volume.
Color Doppler imaging provides visual information on relative direction of blood flow and mean blood
velocity (see Chapter 25: Endoscopic Blood Flow Analysis). Color encoding is used to depict the
direction of blood flow, with flow toward the transducer conventionally portrayed as red and flow away
from the transducer as blue. In addition, Doppler-shifted frequency information is used to display mean
velocity of flow as color shading or hue; this may vary, for example, from deep red for slower-moving
blood to white for higher-velocity systems.7(419)
Color Doppler imaging has several advantages in the evaluation of vascular structures. It provides a
rapid assessment of vascular patency and flow directionality. For small vessels or diseased vessels
with off-axis flow streams, color Doppler imaging may improve estimation of the Doppler angle for
pulsed Doppler sampling and provide more accurate velocity information. Color Doppler imaging allows
visualization of small vessels not seen with gray scale imaging, which is particularly important in the
evaluation of smaller branch vasculature. The addition of color Doppler imaging often decreases
examination time and increases operator confidence by visually depicting normal and abnormal flow
patterns.7,8(420)
The information provided by color Doppler imaging is, however, limited. In many clinical circumstances,
detailed information on the characteristics of blood flow is needed, such as absolute blood velocity and
resistance in the vascular system. In these instances, information obtained by pulsed Doppler imaging
is critical, and color Doppler imaging provides guidance for pulsed Doppler sampling.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
Kremkau FW. Diagnostic Ultrasound: Principles, Instruments and Exercises. 3rd ed.
Philadelphia: WB Saunders. 1989.
Goldstein A. Physics of ultrasound. In Rumack CM, Wilson SR, Charboneau JW (eds).
Diagnostic Ultrasound. St. Louis: Mosby-Year Book, 1991;218.
Sprawls P. Physical Principles of Medical Imaging. Gaithersburg, MD: Aspen Publishers.
1993.
Curry ST III, Dowdey JE, Murry RC. Christensen's Physics of Diagnostic Radiology.
Philadelphia: Lea & Febiger. 1990.
Nelson TR, Pretorius DH. The Doppler signal: Where does it come from and what does it
mean?. AJR 1988;151:43947.
Taylor KJW, Holland S. Doppler US: Part I. Basic principles, instrumentations, and pitfalls.
Radiology 1990;174:297307.
Mitchell DG. Color Doppler imaging: Principles, limitations, and artifacts. Radiology
1990;177:110.
Foley WD, Erickson SJ. Color Doppler flow imaging. AJR Am J Roentgenol 1991;156:313.
Chapter 12 Chromoscopy
(421)
(422)
KAZUNORI IDA, M.D.
MASAHIRO TADA, M.D.
Esophagus, Stomach, and Duodenum

Chromoscopy, a unique method of accurate diagnosis of many gastrointestinal disorders, provides a
variety of new findings and information that cannot be obtained by standard endoscopy.1,2(423)
Premedication
No special premedication is necessary for chromoscopy except for examination of the stomach.
Gastric mucosa is covered with large quantities of mucus that inhibits observation. Therefore,
procedures for the elimination of mucus are always necessary for chromoscopy of the stomach.
An anticholinergic agent is given intramuscularly to suppress evacuation of the mucus-clearing solution
from the stomach. The gastric mucus-clearing solution consists of dimethylpolysiloxan, Pronase (a
proteinase enzyme) (Kaken Pharmaceutical Co., Ltd.), and sodium bicarbonate (Figure 121). The
temperature of the solution should be about 40 C, as this promotes the activity of the enzyme. Sodium
bicarbonate maintains the Pronase solution at an optimum pH value. The mucus-clearing solution
must wash over all gastric mucosal surfaces. This may be accomplished by asking the recumbent
patient to roll over once every minute for about 10 or 15 min.
(424)Figure 121. Procedure for elimination of gastric mucus using a proteinase enzyme.
Methods for Spraying the Dye Solution

Two spraying methods are used: direct and indirect. With the direct method, which can be used
throughout the gastrointestinal tract, a dye solution is simply sprayed over the mucosa during
endoscopic observation by means of a syringe and a catheter inserted through the accessory channel
of the endoscope. The indirect method is used only in the stomach. After premedication and the gastric
mucus-clearing procedure have been completed, the patient takes the dye orally and once again
repeatedly changes position by rolling over, so that the dye reaches all gastric mucosal surfaces. At the
conclusion of the examination, any remaining pools of dye should be aspirated.
Technique and Application

Three basic types of chromoscopy are used in gastrointestinal endoscopy at present: contrast,
staining, and reaction methods (Table 121).
TABLE 121
METHOD
Contrast
Staining
Reaction
Chromoscopy
MECHANISM
Dye collects in
mucosal
depressions
Absorption or
permeation of
dye
APPLICATION
DYE
More precise
diagnosis
Indigo carmine
Intestinal
metaplasia;
intestinal and
colonic diseases
Methylene blue
Esophagitis;
esophageal
carcinoma
Defines
acid-secreting
mucosa
Lugol's solution
Reaction with cell

constituent or
secretion
Lugol
Congo red
Congo red
Contrast Method
The contrast method highlights irregularities in the mucosal architecture by pooling of a blue dye
solution in mucosal grooves and other interstices. This improves the precision of endoscopic diagnosis
by defining minute and inconspicuous lesions and structures that might otherwise be overlooked with
conventional endoscopic methods. Some of the applications of this method are:
1. Observation of the minute structure of the gastrointestinal tract, such as areae gastricae,
duodenal villi, and colonic areas.
2. Confirmation of the existence of a very small lesion, such as a gastric erosion or a small colonic
polyp.
3. Differentiation of certain lesions as either benign or malignant.
4. Determination of the extent of infiltration of a malignant lesion.
5. Certain applications in conjunction with magnifying endoscopy.
Dye spraying may be either direct or indirect with the contrast method (Figure 122). In the direct
method, a 0.2% indigo carmine solution is sprayed after the standard endoscopic evaluation has been
completed. If the stomach is to be examined, the mucus-clearing procedure must be completed before
endoscopy. In the indirect method, which is applied in the stomach and the duodenal bulb, 10 ml of a
3% indigo carmine solution is given orally at the same time that the patient takes the mucus-clearing
solution. We prefer the simple indirect method.
(425)Figure 122. Contrast method using indigo carmine solution.

Staining Method
The staining method is based on absorption of dye by epithelial cells or permeation of the dye into
necrotic tissue.3(426)
Staining can be used to diagnose certain diseases, pathologic conditions, and states of the mucosa
that are difficult to recognize with certainty by conventional endoscopic observation. Some examples
are gastric intestinal metaplasia and the status of the colonic mucosa in ulcerative colitis, that is,
whether the inflammatory process is active or healed.
Direct and indirect methods of dye application may be used in the stomach as outlined in Table 122,
but the indirect method is preferable because directly sprayed dye promptly flows away and is not fully
absorbed. In the duodenum, small intestine, and colon, however, a 0.1 to 0.5% dye solution is sprayed
directly over the mucosa. After 1 to 2 min, the mucosa must be washed with water to observe the
stained surface.
Staining Method Using

Methylene Blue Dye
TABLE 122
DIRECT METHOD
1. Procedure for elimination of gastric mucus
2. Conventional endoscopic evaluation
3. Spray 0.1 to 0.5% methylene blue solution on mucosa
4. Wash mucosa with water 1 to 2 min later
5. Endoscopic chromoscopic observation
INDIRECT METHOD
1. Procedure for elimination of gastric mucus
2. Oral administration of 20 ml of 0.5 to 0.7% methylene
blue solution in conjunction with the gastric
mucus-eliminating solution
3. Patient rolls over once every minute for 10 to 15 min
4. Endoscopic chromoscopic observation
Reaction Method
In the reaction method, a dye applied to a mucosal surface reacts with a constituent of the epithelial
cell or with some mucosal secretion. There are two types: the Lugol method and the Congo red
method.
Lugol's Method
Nonkeratinized squamous epithelial cells contain abundant glycogen that reacts with Lugol's
solution.4(427) This reaction has been used in the diagnosis of esophageal diseases such as
esophagitis and carcinoma.5(428) After standard endoscopic evaluation, 1.5 to 3.0% Lugol's solution is
sprayed directly over the entire esophageal mucosal surface. Normally, the mucosa turns brown within
1 to 2 sec.
Congo Red Method
The Congo red method is based on the reaction that occurs between this dye and secreted
hydrochloric acid.6(429) It is therefore useful in defining the extent of the acid-secreting fundic mucosa.
The fundic mucosa changes from red to dark blue; the antral mucosal surface does not change color.
The original technique for the Congo red method is as follows: After conventional endoscopic
observation, a mixture of 0.3% Congo red and 5% sodium bicarbonate solution is sprayed over the
gastric mucosa. Tetragastrin, 5 g/kg, is then injected intramuscularly. The mucosa is then observed
for 15 to 30 min until no further extension of discolored areas occurs. This method is not suitable for
morphologic observation.
Endoscopic Appearance
Esophagus
Reaction Method
After direct spraying with Lugol's solution, the normal esophageal mucosa stains brown or dark brown
and has a fine mucosal pattern suggesting a wrinkled texture. Areas of leukoplakia stain a striking dark
brown. Esophagitis with or without erosions and esophageal cancer are not stained by Lugol's solution.
This method is useful for detecting early cancer and for defining the boundaries of invasive cancer. A
superficial esophageal carcinoma, shown in Figure 123A, is well delineated as a whitish lesion after
spraying with Lugol's solution (Figure 123B).
(430)Figure 123. Superficial cancer of the esophagus. A, Conventional endoscopy. B, After

being sprayed with Lugol's solution, the cancer is more clearly delineated.
Stomach and Duodenum

Contrast Method
With the contrast method, the normalthat is, acid-secretingfundic mucosa of the proximal stomach
is thick and reddish, whereas the pyloric (antral) mucosa and non-acid-secreting fundic mucosa that
are affected by fundal gastritis are thin and yellowish (Table 123).7(431) Nonsecreting mucosa
observed in the proximal stomach corresponds to atrophic gastritis.
Chromoscopic Features of Fundic

and Pyloric (Antral) Mucosa
TABLE 123
FEATURE
Mucosa
Thickness
Color
Luster
Areae Gastricae
Size
Shape
Contour
Arrangement
FUNDIC
MUCOSA
PYLORIC (ANTRAL)
MUCOSA
Thick
Reddish
Glossy
Thin
Yellow
Dull
Regular
Uniform
Well defined
Close, regular
Regular-irregular
Uniform-multiform
Poorly defined
Regular-irregular
The term areae gastrica, it will be recalled, refers to a small area superimposed on the gastric mucosa
by a system of furrows. These grooves subdivide the mucosa into areas (areae gastricae) of 1 to 5 mm
in diameter. The areae gastricae of the fundic mucosa are usually regular in size and arrangement;
those of the other non-acid-secreting fundic mucosa and pyloric mucosa are generally irregular. The
former are called Type F and the latter Type P. Each type is divided, based on appearance, into four
subtypes (Figures 124, 125 and 126). At the higher end of the scale, that is, F3 or P3, the mucosa
is atrophic. Areae gastricae express histopathologic findings that can be diagnostic for atrophic
gastritis.
(432)Figure 124. Classification of areae gastricae.
(433)Figure 125. Examples of classification of areae gastricae in the fundic mucosa. A, F0;
B, F2.
(434)Figure 126. Examples of classification of areae gastricae in the pyloric (antral) mucosa.
A, P1; B, P2.
The distal border of the acid-secreting fundic mucosa in the stomach can be easily recognized
because of the differences in mucosal characteristics and the pattern of the areae gastricae. The
border is not uniform in position in the stomach from individual to individual, especially along the lesser
curvature. Two general types are recognized. In the open type, the lesser curvature of the gastric body
is covered with non-acid-secreting fundic mucosa; in the closed type, the lesser curvature mucosa is
acid-secreting fundus (Figure 127). The open and closed types are subdivided into three subtypes
according to the extent of the pyloric type of mucosa; that is, in the order of Type C0 to O3 the extent
of the normal fundic mucosa progressively decreases, with a corresponding decrease in acid
secretion. An example of a patient with a Type C2 pattern is shown in Figure 128.
(435)Figure 127. Schematic diagram illustrating closed and open types of (fundic-pyloric)
mucosal border patterns in the stomach as well as the subclassification of each type. The top
row of figures represents the entire stomach opened as a gross specimen along the greater
curvature. The bottom row of figures illustrates endoscopic views of the stomach
corresponding to the figures in the top row.
(436)Figure 128. Fundic-pyloric mucosal border. A, Region of the stomach proximal to the
angulus in a patient with the closed type of distribution. B, Mucosal border on the greater
curvature of the antrum.
The contrast method is very effective in detecting early gastric cancer, especially minute and Type IIb
carcinomas (Japanese classification of early gastric cancer), and in recognizing the extent of
cancerous infiltration in the mucosa.8(437) An example of early gastric cancer in the body of the
stomach (Type IIc, similar to IIb) is shown in Figure 129.
(438)Figure 129. Early gastric cancer (Type IIc, similar to Type IIb). A, Conventional
endoscopy. B, Direct contrast method demonstrates irregular erosion, greater extent of
involvement, spread into the surrounding mucosa, abnormal red granulation, and tapered
folds. All findings indicate malignancy.
The contrast method is useful in differentiating benign from malignant ulcers. A small erosion that
produces a very shallow depression without elevation of the surrounding mucosa is very difficult to
detect endoscopically. However, with the contrast method, recognizing such lesions and defining their
size, shape, number, and distribution are not difficult (Figure 1210).
(439)Figure 1210. Erosive gastritis, antrum. A, Conventional endoscopy showing a few red
spots. B, Multiple small erosions demonstrated by contrast chromoscopy with indigo carmine
solution.
Precise endoscopic evaluation of peptic ulcer healing is extremely important because the approach to
treatment may be significantly altered by endoscopic findings. Chromoscopy using the contrast method
provides a very accurate assessment. Even when conventional endoscopy evaluation indicates that
the ulcer has healed, in the majority of cases, the contrast method reveals a small depression
consisting of a defect in the regenerated epithelium (Figure 1211). In the strict sense of the term, the
ulcer has not yet healed and may recur. Therefore, evaluation of ulcer healing should be performed
using the contrast method and detailed observation.
(440)Figure 1211. Evaluation of gastric ulcer healing. A, Conventional endoscopy suggests a

healed ulcer. B, Chromoscopy (indirect method) performed immediately after conventional
endoscopy reveals a depression in the center that indicates incomplete healing.
Chronic active gastritis can accompany gastric erosions, which are thought by some investigators to
produce a variety of symptoms, including epigastric discomfort. Chronic active gastritis may be
manifest as redness of the mucosa. This may have a comblike patternthat is, multiple, long,
frequently narrow red lines in the fundic mucosa that are arranged in parallel fashion (Figure 1212).
The use of contrast chromoscopy reveals associated erosions in many patients with comblike redness
and unexplained epigastric discomfort or pain.
(441)Figure 1212. Comblike redness and linear erosions, gastric body. A, Comblike redness
by conventional endoscopy. B, Linear erosion on a reddened fold are revealed by contrast
method.
The villous pattern of the duodenum and individual villi are clearly defined by the contrast method. In
the example shown (Figure 1213), two microerosions are present in a duodenal bulb where the
mucosal pattern has an otherwise normal villous appearance. Such erosions are found frequently. The
contrast method is also helpful in determining whether a duodenal ulcer has healed completely.
(442)Figure 1213. Normal villi and two microerosions in the duodenal bulb (contrast
method).
Staining Method
Two groups of stomach lesions are defined by in vivo staining chromoscopy. Group 1 comprises the
lesions that have absorptive cells and regular papillae; like villi, they are stainable by spraying
methylene blue on their surface. The papillae thus stained can be seen when the tip of the endoscope
is close to the mucosal surface or when high-magnification endoscopy is used (Figure 1214). The
lesions in Group 2 do not have stainable, regular papillae. The staining method in this group is based
on permeation of methylene blue into tumor tissue. The lesions in Group 1 include intestinal
metaplasia, adenoma, and polyps consisting of intestinalized mucosal tissue. Group 2 lesions
comprise protruded gastric cancers that are covered by necrotic tissue and debris.
(443)Figure 1214. Close-up view of intestinal metaplasia stained with methylene blue.
Reaction MethodCongo Red
The Congo red method demonstrates the extent of acid-secreting mucosa (Figure 1215). Once
discoloration takes place, the junction of the acid-secreting and non-acid-secreting mucosal areas is
sharply defined.
(444)Figure 1215. Endoscopic appearance after Congo red method. Very dark purple color
develops in acid-secreting mucosa.
Mucosal Border and Gastroduodenal Disease
As noted previously, the distribution pattern for acid-secreting fundic mucosa and the
non-acid-secreting mucosa in the stomach and hence the border between these two areas differ from
individual to individual. In general, the demarcation between the two types of mucosa moves in an oral
direction with increasing age; that is, in a group of patients, the predominant classification changes
from C0 toward O3 as age increases (Figure 1216). Gastric cancer and polyp formation, within this
classification, are found with highest frequency in patients with the greatest proximal extent of the
pyloric (antral) type of mucosa, that is, those in Type O3. Furthermore, the frequency of these
disorders gradually increases as the classification changes from C0 toward O3. Conversely, the
frequency of duodenal ulcer, gastric erosions, and superficial gastritis decreases as the border moves
in a distal direction; that is, the highest incidence is found with the C0 classification.
(445)Figure 1216. Types of mucosal border by age group.

The more proximal the location of a peptic ulcer, the more proximal the demarcation between the two
types of mucosa. In other words, an ulcer in a proximal position is found in fundic mucosa, which
occupies a relatively small percentage of the total gastric mucosal surface area, and generally the acid
output of the stomach is low. More than 90% of patients with active duodenal ulcers have a closed-type
mucosal pattern, and open types O2 and O3 are not found. The comblike redness characteristic of
chronic active fundal gastritis, as expected, is more often found when the mucosal pattern is the closed
type.
Small Intestine
The mucosal surface of the small intestine is covered by numerous minute villi of various shapes. The
villi of the duodenal bulb and ileocecal valve have a ridgelike form, but those of the distal parts of the
duodenum, jejunum, and ileum have a regularly arranged finger shape.
Application of chromoscopy in the small intestine focuses mainly on observation of the villi;
high-magnification fiberscopes are frequently used in conjunction with chromoscopy (see Chapter 13:
Magnification Endoscopy). Two of the various magnifying endoscopes suitable for close observation of
the villi are the SIF-M and the SIF-HM (both Olympus Optical Company, Ltd.) (Table 124). Use of the
SIF-M and SIF-HM for close observation permits photography at magnifications of 10 to 35.9(446)
TABLE 124
Optical system
Angle of view field
Range of observation
Magnifying rate
Insertion tube
Outer diameter (mm)
Angle mechanism
Up, down
Technical Specifications of Magnifying Endoscopes

SIF-M
SIF-HM
CF-HM
CF-UHM
CF-200HM
65
2.54.5
x10
70
2.3100
x35
70
2.3100
x35
100
3100
x170
(0.0950.105)*
76
2090
x50
(2.83.1)*
10
12.5
14.4
14
13.7
150,
120
90
180, 90
180
120
180
100
160
110
160
2.8
1420
1645
2.8
1420
1670
2.8
1035
1270
2.8
1330
1620
Right, left
Biopsy channel
Inner diameter (mm)
2.8
Working length (mm)
1860
Whole length (mm)
2015
* Magnifying optical system.
The magnifying enteroscope is introduced into the small intestine via the anus or mouth. With the
instrument in place, and after the standard inspection, 10 ml of 0.1% methylene blue solution is
sprayed on the mucosa. Staining of the villi occurs rapidly.
The villi have any of several different shapes or forms.10(447) In almost all parts of the small intestine
they are finger-like. A few villi with a wider form (tonguelike) may be found among the finger-shaped
villi, but these are thought to be a variation of normal.
The villi of normal duodenal mucosa are well stained with methylene blue. This method reveals
complete or incomplete healing of a duodenal ulcer and demonstrates complete versus incomplete
reepithelialization. An ulcer scar takes the same stain as surrounding mucosa, and this indicates
complete reepithelialization and therefore complete healing, with the likelihood that a particular ulcer
will not recur.
The villi are irregular in shape and arrangement in inflammatory bowel disease and are sometimes lost
completely. In an area of tuberculous scarring of the intestine, the villi are finger-shaped or leaf-shaped
but vary in size, and their arrangement is irregular and sporadic (Figure 1217); also, the absorption of
methylene blue in the area of the scar is less than that of normal villi. In Crohn's disease, the villous
pattern is irregular with a leaf or ridge convolution form (Figure 1218). It is interesting that in Crohn's
disease the villi that are distant from an obvious inflammatory focus and seem to be intact by
endoscopic observation actually have an atrophic appearance when studied by magnification
chromoscopy.
(448)Figure 1217. Atrophic villi of intestinal tuberculosis.
(449)Figure 1218. Irregular villi of the ileum in Crohn's disease.

When a villus becomes abnormal, a striking reduction occurs in the total number of constituent
epithelial cells. Thus, a leaf-shaped villus has only about one fourth the number of epithelial cells of a
normal finger-shaped villus; those with a convolution form have only one eighth the normal number.
Such morphologic abnormalities can be found, for example, in the intestinal villi of postgastrectomy
patients. These and similar abnormalities are thought to correspond to a malabsorptive condition in the
mucosa and to reflect the pathophysiologic state of the small intestine.
Colon
The surface of the colonic mucosa has a granular appearance and is demarcated by innominate
grooves; demarcated regions are called the colon area on the basis of radiologic and endoscopic
findings. The innominate grooves circumscribe areas on the colonic mucosa that contain numerous
pits, representing the glands or crypts of Lieberkhn. These minute pits are arranged in a regular
pattern and are round or somewhat oval. The diameter of each pit is 40 to 50 m. About 40 to 60 pits
are present in one colon area.
When a blue dye such as indigo carmine or methylene blue is sprayed on the surface of the colonic
mucosa, the colon areas as defined by grooves are observable without difficulty.11(450) Several
normal variations in the shape of the colonic mucosal areas can be seen: spindle, oval, and so on.
The minute structures of the colonic mucosa can be observed as a fine network pattern with the aid of
chromoscopy, but smaller pits cannot be seen. Observation of such minute pits requires the use of a
magnifying colonoscope. Several instrument types have now been devised (Table 124).11,12(451)
Magnifying colonoscopes have almost the same characteristics as conventional colonoscopes except
that their optical systems can provide mucosal images that are magnified from about 35 to 170.
With the use of chromoscopy and a magnifying colonoscope, numerous minute pits, which represent
the glands of Lieberkhn, are seen in the colonic mucosal areas circumscribed by the innominate
grooves (Figure 1219). These minute pits are arranged regularly and are round and oval.
(452)Figure 1219. Magnified view of the normal colon mucosa using CF-200HM (50).
Minute round pits are arranged regularly.
Polypoid Lesions
The surface appearance of a polyp or the margin of a flat or diminutive polyp is more clearly observed
with chromoscopy than with conventional colonoscopy alone. The minute structure of polypoid lesions
can also be observed with a magnifying colonoscope.
Polyps are classified into four categories according to the shape or arrangement of pits (Table 125):
circular, tubular, sulcus, and irregular (Figure 1220). Metaplastic polyps belong to the circular type;
adenomatous polyps may be circular, tubular, or sulcus types and are quite regular in shape and
arrangement. All cases of advanced cancer fall into the irregular type. Almost all cases of early cancer
also belong to the irregular type. However, some have a tubular or sulcus pattern like that of the benign
adenoma. The diagnosis of early or diminutive cancer with the magnifying colonoscope remains
problematic, and further improvements in both magnifying capabilities of endoscopes and
chromoscopy are required.
adenoma. The diagnosis of early or diminutive cancer with the magnifying colonoscope remains
problematic, and further improvements in both magnifying capabilities of endoscopes and
chromoscopy are required.
Classification of Minute Structure

of Polypoid Lesions of the Colon
TABLE 125
Round type
Tubular type
Sulcus type
Irregular type
Regular round or narrowly deformed pits; a little

larger than normal mucosal pits
Various patterns from circular to oval pits
No pits; formation of cerebral gyrus-like sinuses
Surface shows cauliflower-like, rough, irregular or
nonstructured appearance
(453)Figure 1220. Various types of minute structures of the colon polyp. A, Round type;
round or very nearly round pits that are slightly larger than normal mucosal pits. B, Tubular
type; various patterns from circular to oval. C, Sulcus type; no pits but sinus formation like
cerebral gyrus. D, Irregular type of early cancer; surface shows a cauliflower-like, rough,
irregular pattern.
Ultra-Magnifying Endoscopy
The ultra-high-magnification colonoscope was introduced in 1982 to enhance the recognition of small
colonic cancers.12(454) This instrument, the CF-UHM, which has undergone improvements, has a
magnification capability of 170.
Ultra-high magnification with the aid of chromoscopic techniques offers endoscopic views of the minute
structure of the pits in flat, normal mucosa (Figure 1221). The pits are round or somewhat oval.
Nuclei of the cells are arranged peripherally around the pits, each nucleus being round and regular in
shape. Nuclei of the stroma around the intestinal gland are also visualized clearly.
(455)Figure 1221. Ultra-magnified view of the normal colon mucosa, using CF-UHM
(170).
The pits of adenomatous polyps are enlarged and elongated. Nuclei of the gland cells are arranged
peripherally as in flat, normal mucosa. Each nucleus is regular in its arrangement and size. The stroma
is displaced by the glands and reduced in size.
The minute surface structure of early cancer differs from that of both benign polypoid lesions and
normal, flat mucosa (Figure 1222). Pits are displaced by irregular sulcus formation. Nuclei are also
irregular in shape and can be easily differentiated from those of benign colon polyps. Thus, colon
cancer can be accurately diagnosed in its early or diminutive stage with the ultra-magnifying
colonoscope. It is expected that more diminutive cancers will be diagnosed at the stage in which
cancer shows only cellular atypism and not structural atypism.
(456)Figure 1222. Ultra-magnified view of early cancer.
REFERENCES
1.
Yamakawa K, Naito S, Kanai J, et al. Superficial staining of gastric lesions by fiberscopy. In

Proceedings of the First Congress of the International Society of Endoscopy. Tokyo,
1966;58690.
2.
Ida K, Misaki J, Kohli Y, Kawai K. Fundamental studies on the dye scattering method for
endoscopy. Jpn J Gastroenterol Endosc 1972;14:2616.
3.
Ida K, Hashimoto Y, Kawai K. In vivo staining of gastric mucosa. Its application to endoscopic
diagnosis of intestinal metaplasia. Endoscopy 1975;7:1824.
4.
Schiller W. Early diagnosis of carcinoma of the cervix. Surg Gynecol Obstet 1933;56:21022.
5. Nothmann BJ, Wright JR, Schuster MM. In vivo vital staining as an aid to identification of
esophagogastric mucosal junction in man. Am J Dig Dis 1972;17:91924.
6.
Okuda S, Saegusa T, Ito T, et al. An endoscopic method to investigate the gastric acid
secretion. In Proceedings of the First Congress of the International Society of Endoscopy.
Tokyo, 1966;2216.
7.
Ida K, Kohli Y, Shimamoto K, et al. Endoscopical findings of fundic and pyloric gland area
using dye scattering method. Endoscopy 1973;5:216.
8. Ida K, Hashimoto Y, Takeda S, et al. Endoscopic diagnosis of gastric cancer with dye
scattering. Am J Gastroenterol 1975;63:31620.
9.
Tada M, Kawai K. Small-bowel endoscopy. Scand J Gastroenterol 1984;19(suppl 102):3952.
10. Tada M, Misaki F, Kawai K. Endoscopic observation of villi with magnifying
enterocolonoscopes. Gastrointest Endosc 1982;28:179.
11. Tada M, Misaki F, Kawai K. A new approach to the observation of minute changes of the
colonic mucosa by means of magnifying colonoscope, type CF-MB-M (Olympus). Gastrointest
Endosc 1978;24:1467.
12.
Tada M, Nishimura S, Kawai K. A new method for the ultra-magnifying observation of the
colon mucosa. J Kyoto Pref Univ Med 1982;91:34954.
Chapter 13 Magnification Endoscopy

(457)
(458)
TADAYOSHI TAKEMOTO, M.D.
NOBUHIRO SAKAKI, M.D.
Magnification endoscopy refers to a diagnostic endoscopic system for observing changes that occur in
the pits (openings of glands) and the villi of the digestive tract. Diagnosis by observation of morphologic
changes must be based on an established frame of reference. Standard endoscopic diagnosis is
equivalent to observation of findings visible to the naked eye. The frame of reference in magnification
endoscopy is the fine structure formed by punctiform or linearly arranged recesses in the mucosa of
the digestive tract of about 0.1 mm in size. As such, magnification endoscopy is intermediate between
macroscopic and microscopic observation. The structures and findings observable by magnification
endoscopy correspond to those that may be seen by means of a dissecting (stereoscopic) microscope.
Although all endoscopes provide some magnification, the power of standard instruments is insufficient
for observation of fine detail such as pit patterns. Special magnifying endoscopes have been
developed for this purpose. Use of these instruments also requires new systems of endoscopic
diagnosis, such as that which we developed for observation of the gastric mucosa.1(459) This system
of gastroscopy and its related instrumentation were introduced in the first edition of this book.
Excellent resolving power as well as computer processing of digital images (e.g., edge enhancement)
makes detailed observations possible with modern electronic endoscopes (see Chapter 3: Flexible
Endoscope Technology: The Video Image Endoscope). However, the development in Japan of
endoscopic mucosectomy (strip biopsy) for mucosal cancers and its associated requirement for
detailed diagnosis have redirected attention to the need for magnifying instruments (see Chapter 24:
Early Colorectal Cancer and Endoscopic Resection). The fundamentals of magnification endoscopy
and its application to endoscopic diagnosis using magnifying electronic endoscopes are described in
this chapter.
History of Magnification Endoscopy

Early Development
The origin of magnification endoscopy, that is, the first attempt to observe gastric pits, dates to the era
of rigid gastroscopes. Gutzeit and Teitige2(460) observed foveolae in the body of the stomach and
published their findings in 1954. Subsequently, a diagnostic system based on the shapes of the gastric
pits was developed within the field of pathology. This made use of the dissecting microscope. Salem
and Truelove3(461) in 1964 stressed the value of this system for examination of gastric biopsies,
especially in the diagnosis of gastritis. In Japan, Matsumoto4(462) reported that normal gastric pit
patterns are destroyed and replaced by pseudo-pit patterns in cases of gastric ulcer and gastric
cancer, each having its own characteristic pattern. Yoshii5(463) described normal fine mucosal
patterns (foveolate, foveolate-sulciform, and sulciform) as well as the changes that occur in these
patterns in certain disorders. These types of investigation provide the basis for the diagnostic system of
magnification endoscopy.
The Magnifying Fiberscope

Special fiberscopes for magnification endoscopy have been produced in Japan since l967. Although
many prototypes were produced (Table 131), detailed study1,6,7(464) of the fine mucosal patterns of
the gastrointestinal tract awaited development of the most powerful magnifying fiberscopes such as the
ML series offered by Machida Endoscope Company in 1977 and the HM series offered by Olympus
Optical Company, Ltd. in 1980. High-magnification gastroscopy was performed with the FGS-ML II and
GIF-HM instruments (Figures 131 and 132); the FCS-ML II and CF-HM instruments were used for
magnification colonoscopy. In general, these instruments were forward-viewing fiberscopes with
magnification capabilities ranging from 30 to 35. They provided excellent images of pit patterns, but
they were not popular for clinical applications owing to certain limitations in their handling
characteristics. Production was discontinued in the mid-1980s.
TABLE 131
AREA
Stomach
Development of Magnifying Fiberscopes

INSTRUMENT
OF USE
FGS-ML type 1
MGF-I
FGS-ML, type 2
GIF-M
FGS-ML II
GIF-HM
FGS-SML
Colon
CF-MB-F
FCS-ML II
CF-HM
Cf-UHM
* Magmagnification capability.
MAG*
x3
x15
x15
x15
x30
x35
x170
x10
x30
x35
x170
DEVELOPED BY
Okuyama
Suzaki and Miyake
Maruyama and Takemoto
Ida and Kawai
Sakaki and Takemoto
Ooida and Okabe
Sakaki
Tada and Kawai
Nishizawa and Kobayashi
Tada
Tada
(465)Figure 131. Magnified view of the gastric mucosa at the greater curvature of gastric
angle, using the FGS-ML II model endoscope offered by Machida Endoscope Company.
(466)Figure 132. Magnified view of the gastric mucosa at the greater curvature of the gastric
angle, using the GIF-HM endoscope offered by Olympus Optical Company, Ltd.
An ultra-high-magnification instrument, CF-UHM (Olympus Optical Company, Ltd.), with a magnifying
capacity of 170, was developed by Tada et al.8(467) in 1982. Another system that we9(468)
developed was the FGS-SML (Machida Endoscope Company) with 170 magnification. Cellular
features and the nucleus are considered the frame of reference for these ultra-high-magnification
fiberscopes (Figure 133).
(469)Figure 133. Magnified view (170) of a colonic cancer, using an

ultra-high-magnification endoscope (CF-UHM, Olympus Optical Company, Ltd.). (Courtesy
of Dr. M. Tada.)
Magnifying Endoscopy with Electronic Endoscopes

When electronic endoscopes were initially developed, their excellent resolution, based on high
numbers of picture elements (pixels), was used to advantage as in the TGS-50D introduced by
Toshiba Corporation-Machida Endoscope Company in 1986 (Figure 134). Similar high-resolution
instrument models were offered by the Fuji Photo Optical Company (HR2 series) in 1988 (Figure 135)
in response to interest in observing mucosal pit patterns. Further improvements in resolution were
achieved with the introduction by the Toshiba Corporation in 1992 of instruments (TRE-3000 series)
with a "color" charge-coupled device (CCD) having a high pixel density (Figure 136). All of these are
practical endoscopes that offer not only magnification but also a wide angle of view (more than 100
degrees), features that usually represent conflicting design specifications in a single instrument.
Although the degree of magnification with these instruments is low, it is nevertheless satisfactory for
the observation of pit patterns.
(470)Figure 134. Close-up view of a minute Type IIc early gastric cancer, using the
TGS-50D model endoscope (Toshiba Corporation-Machida Endoscope Company) after
spraying a solution of indigo carmine dye.
(471)Figure 135. View of Type IIa + IIc early gastric cancer, using an EG7-HR2 model
endoscope (Fuji Photo Optical Company).
(472)Figure 136. View of a gastric ulcer scar, using the TGI-3000D endoscope (Toshiba
Corporation). (Courtesy of Dr. Y. Hoshihara.)
Another approach to magnification endoscopy has been to place a movable lens with variable
magnification between the objective lens and the CCD of an electronic endoscope, the first such
instrument being the GIF-V10Z offered by the Olympus Optical Company, Ltd. This company
subsequently introduced the model GIF-200Z instrument in 1991. After observation in the standard
mode, this instrument is placed close to the mucosa for magnified observation. Once the objective lens
is close to the area of interest, a magnified view is produced by twisting a ring that is incorporated into
the control section of the instrument. Maximum magnification is 35; the magnified image is displayed
as a 1-inch image on a television monitor. Pit patterns can be observed clearly with these endoscopes
(Figure 137). However, the range of focus at magnified views is very narrow; that is, the image is out
of focus unless the objective lens remains within a narrow range of distance from the mucosal surface.
For this reason, time and effort are required to learn the technique of magnification endoscopy with this
type of instrument.
(473)Figure 137. Magnified view of fundic gastric mucosa, using the GIF-200Z model
magnifying endoscope (Olympus Optical Company, Ltd.).
The fine gastric mucosal pattern consists of gastric pits; the diameter of the elements in this pattern is
about 0.1 mm. High resolving power, that is, the ability to distinguish structures as small as 0.01 mm,
such as that offered by FGS-ML and GIF-HM instruments, is required to observe the detailed
morphology of pit patterns. In our experience, a resolving power of less than 0.5 mm, such as that
available with recently introduced electronic endoscopes, is sufficient only for low-magnification
endoscopy, although this allows for recognition of characteristics and differences in pit patterns. The
resolving power of standard endoscopes is about 0.1 cm by our measurements with a metal test plate
of the type used in roentgenography (Micro-chart, Micromedical Company).
Fundamentals of High-Magnification Gastroscopy

The gastric mucosal surface, when observed with magnifying fiberscopes, has a pattern composed of
variously shaped gastric pits. Areas that are slightly reddish are elevated parts of the mucosal surface;
whitish yellow dots and lines are recesses, the former being the gastric pits. We named this
appearance the fine gastric mucosal pattern.
The fine gastric mucosal pattern is divided into two main patterns of depression: discontinuous and
continuous. We have classified these two main patterns as follows: The discontinuous pattern may be
Type A, which has "dotted" depressions (Figure 138), and Type B, which has short linear depressions
(Figure 139). The continuous pattern is divided into Type C, in which the elevated areas are
surrounded by depressions that are characterized as striped (Figure 1310), and Type D, which has
circular depressions (Figure 1311). Three intermediate patterns are also recognized, so that the
combination of Types A and B is named Type AB, the combination of Types B and C is Type BC, and
the combination of Types C and D is Type CD. Therefore, seven pattern types are possible (Figure
1312).
(474)Figure 138. Type A fine gastric mucosal pattern (dotted depressions) under
magnification view, using the FGS-ML II model instrument (Machida Endoscope Company).
(475)Figure 139. Type B fine gastric mucosal pattern (short linear depressions).
(476)Figure 1310. Type C fine gastric mucosa pattern (striped continuous depressions).
(477)Figure 1311. Type D fine gastric mucosal pattern (circular continuous depressions).
(478)Figure 1312. Classification schema of the fine gastric mucosal patterns for
magnification endoscopy.
The fine gastric mucosal pattern is altered in various disease states. The Type A pattern is
characteristic of normal fundic mucosa. This is transformed into the Types B and C patterns with
mucosal atrophy. The normal pyloric mucosa has a fine Type C pattern that transforms to a coarse
Type C pattern when the mucosa is atrophic. Well-differentiated (intestinal) cancer has a characteristic
Type C pattern (Figure 1313), but poorly differentiated (diffuse) cancer shows an uneven surface with
no pit pattern. The Type D pattern indicates previous mucosal destruction and regeneration and is
observed mainly with scarring from ulcers and erosions. The pattern of a hyperplastic polyp is the
same as that of the surrounding mucosa.
(479)Figure 1313. Characteristic irregular type C pattern of a well-differentiated tubular

adenocarcinoma (FGS-ML II, Machida Endoscope Company).
Small Gastric Cancers

Gastric carcinoma can be recognized by its characteristic pit pattern in magnified views. With
advanced carcinoma, however, it is difficult to observe the characteristic pit pattern owing to tissue
destruction and regeneration. Standard observation is more appropriate for the clinical diagnosis of
advanced carcinoma.
Endoscopic Diagnosis
The diagnosis of a small (diameter up to 1 cm) or minute (diameter up to 5 mm) gastric cancers is
difficult by standard methods of observation. Detailed observation using the dye-contrast method
(chromoscopy) is usually required (Figure 1314) (see Chapter 12: Chromoscopy). These small
lesions, in which the characteristic pit patterns are well preserved, are good subjects for magnification
endoscopy (Figure 1315). They have been treated in Japan by endoscopic mucosectomy using the
technique of strip biopsy devised by Tada et al. in 1984.10(480) Well-differentiated mucosal
carcinomas in the stomach with diameters less than 2 cm are considered in Japan to be good
indications for endoscopic mucosectomy, which also has applications in the colon and esophagus (see
Chapter 24: Early Colorectal Cancer and Endoscopic Resection). Flat or depressed cancers can be
resected safely, along with a portion of surrounding normal mucosa. The spread of carcinoma in an
endoscopically resected specimen is assessed by differences in pit patterns observed with a dissecting
microscope (Figure 1316) prior to standard histologic evaluation.
(481)Figure 1314. Close-up view with a conventional fiberscope (GIF-P20, Olympus

Optical Company, Ltd.) of a minute gastric cancer sprayed with indigo carmine dye.
(482)Figure 1315. Magnified view of the margin of the characteristic pit pattern of a minute
gastric cancer (same lesion as in Figure 1314). (GIF-200Z, Olympus Optical Company,
Ltd.).
(483)Figure 1316. Spread of a mucosal gastric cancer as assessed by difference in pit

patterns observed with a dissecting microscope in an endoscopically resected specimen.
(Courtesy of Dr. Y. Iwasaki.)
An early gastric cancer of the papillary type is recognized as a reddish lesion with a characteristic Type
C pattern that can be likened to a tree branch (Figure 1317). A fine striped pattern is characteristic in
well-differentiated tubular adenocarcinoma. In these cases, background atrophic gastric mucosa also
exhibits a Type C pattern. However, this has a coarse appearance and can be differentiated from the
labyrinth-like pattern with irregular alignment of the carcinomatous area (Figure 1318).
(484)Figure 1317. Papillary adenocarcinoma with characteristic tree branch-like type C

pattern (TGS-50D, Toshiba Corporation-Machida Endoscope Company).
(485)Figure 1318. Well-differentiated tubular adenocarcinoma with labyrinth-like Type C

pattern aligned irregularly (GIF-200Z, Olympus Optical Company, Ltd.).
Moderately differentiated tubular adenocarcinoma has the same fine mucosal pattern as background
atrophic gastric mucosa. Therefore, diagnosis of this type of cancer is difficult even with magnifying
endoscopic observation. A Type IIb (flat) early gastric cancer with a diameter of 1.5 cm is shown in
Figure 1319. Recognition of this lesion as an early gastric cancer was difficult because of the
unevenness in the mucosal surface and color tone as well as a lack of differences in pit patterns
between the cancerous and the benign mucosa.
(486)Figure 1319. Type IIb (flat) moderately differentiated tubular adenocarcinoma that
exhibits the same fine mucosal pattern as the surrounding atrophic mucosa (TGS-50D,
Toshiba Corporation-Machida Endoscope Company).
Correct diagnosis of signet-ring cell carcinoma or poorly differentiated tubular adenocarcinoma, which
do not form defined glandular ducts, is difficult by observation of the fine mucosal pattern alone. These
lesions are for the most part eroded so that the pit pattern is lost and replaced by a flat area (Figure
1320). Occasionally, the eroded area is surrounded by mucosa with irregularly shaped Type D
patterns of various sizes (Figure 1321). Although considered to be characteristic of diffuse (poorly
differentiated) carcinoma, this finding may not be present. When cancer cells are scattered in the
mucosa around the margin of the eroded area, the fine mucosal structure is the same as that of
surrounding normal mucosa and is difficult to recognize (Figure 1322).
(487)Figure 1320. Type IIc type signet-ring cell carcinoma that was observed as a
nonspecific erosion without a pit pattern (TGS-50D, Toshiba Corporation-Machida
Endoscope Company).
(488)Figure 1321. Type IIc poorly differentiated adenocarcinoma in which irregularly

shaped Type D patterns of various sizes surround the eroded area (EG7-FP3, Fuji Photo
Optical Company).
(489)Figure 1322. Marginal area of Type IIc + IIb signet-ring cell carcinoma in which
cancer cells are scattered. The pit pattern is the same as that of the surrounding mucosa
(GIF-300Z, Olympus Optical Company, Ltd.).
Differential Diagnosis
Elevated submucosal lesions have a fine mucosal pattern that is the same as that of surrounding
mucosa. An erosion with associated edema and fibrosis on the surface of a small elevated
submucosal lesion is seen as a defect or as mucosa with a regenerated Type D pattern. On rare
occasion, it is necessary to differentiate a hyperplastic polyp from a papillary carcinoma. This is not
difficult because the former lesion has a coarse Type B or C pattern (Figure 1323).
(490)Figure 1323. Hyperplastic polyp that exhibited a coarse Type BC pattern (GIF-200Z,
Olympus Optical Company, Ltd.).
The elevated Type IIa early neoplastic lesion has a well-demarcated characteristic pattern.
Well-differentiated tubu-lar adenocarcinoma is recognized by a labyrinth-like Type C pattern with a
sharp boundary; a coarse Type C pattern is noted in the surrounding atrophic mucosa. However, it is
difficult to differentiate this lesion from an adenoma.
Small red spots can be divided into surface redness and erosions by magnification endoscopy (Figure
1324). If a fine gastric mucosal pattern is preserved in a reddish area, it is referred to as a redness, a
descriptive term that includes angiodysplasia. An erosion, if acute, is seen as a shallow depression
without a pit pattern; if it has healed, a regenerated Type D pattern is observed. For differential
diagnosis of small depressed lesions, assessment of the fine mucosal pattern in the depressed area
and the adjacent mucosa is important. If the area exhibits a coarse Type D regenerating mucosal
pattern without variation in size, it is an ulcer scar or the remnant of an erosion, especially when the
boundary is not clearly defined. A depressed Type IIc early gastric cancer is easily recognized by the
characteristic pit pattern of the differentiated cell type as seen by magnification endoscopy.
(491)Figure 1324. Classification of gastric red spots using magnification endoscopy.
Gastric Ulcer Healing

Magnification endoscopy can be used to study the healing process of gastric ulcers, including the
morphologic changes that occur during regeneration of the mucosal surface. A magnified view of the
initial stages of healing of an ulcer in the body of the stomach 2 weeks after initiation of treatment is
shown in Figure 1325. A narrow reddish belt without mucosal structure surrounds a white coating,
and an enlarged pit pattern due to edema is present. At 4 weeks, regenerated mucosa in a Type C
pattern, with a radial alignment, was observed (Figure 1326). This ulcer became a scar that showed a
coarse granular regenerated mucosal pattern and transformed subsequently to a fine granular pattern,
similar to that of the surrounding mucosa (Figure 1327).
(492)Figure 1325. Gastric ulcer in the initial stage of healing (R0 stage) at 2 weeks after
treatment (GIF-D10, Olympus Optical Company, Ltd.).
(493)Figure 1326. Gastric ulcer in the initial stage of healing (R2 stage) at 4 weeks after
treatment (GIF-D10, Olympus Optical Company, Ltd.).
(494)Figure 1327. Gastric ulcer scar (Sc scar) at 36 weeks after treatment (GIF-D10,
A recurrent gastric ulcer exhibits a different morphology. A Type C regenerating mucosal pattern was
observed at 4 weeks; after 6 weeks, the pattern tended to be more granular (Figure 1328). A coarse
Type C regenerated mucosal pattern with a central depression was observed (Figure 1329) after
disappearance of the white coating. This transformed into a coarse granular regenerated mucosal
pattern with a central depression (Figure 1330).
(495)Figure 1328. Recurrent gastric ulcer (R2 stage) at 6 weeks after treatment (GIF-D10,
(496)Figure 1329. Recurrent gastric ulcer scar (Sa scar) at 8 weeks after treatment
(GIF-D10, Olympus Optical Company, Ltd.).
(497)Figure 1330. Recurrent gastric ulcer scar (Sa scar) at 36 weeks after treatment
(GIF-D10, Olympus Optical Company, Ltd.).
We proposed a classification of the phases of gastric ulcer healing in 1984 based on our experience
with magnification endoscopy and pathologic studies;11(498) this was modified for application to
standard endoscopy in 1988.12(499) In our classification, the healing process is divided into two main
phases: a regenerating phase with white coating and a scar phase (Figure 1331). The regenerating
phase is divided into three stages based on regenerated mucosal patterns: R0 (no regenerating
mucosal pattern), R1 (fine regenerating mucosal pattern), and R2 (coarse regenerating mucosal
pattern). The scar phase is divided into three stages: Sa (coarse regenerated mucosal pattern with
central depression), Sb (central coarse regenerating mucosal pattern), and Sc (fine pattern similar to
that of the surrounding mucosa).
(500)Figure 1331. Classification of the time phase of the healing process of gastric ulcers by
magnification endoscopy. (From Sakaki N, Takemoto T. Endoscopic study on stage of healing
process of duodenal ulcer [Japanese, English abstract]. Gastroenterol Endosc 1988;
30:19149.)
The scar pattern as seen at endoscopy correlates with the degree of histologic regeneration and
healing.13(501) Histologic regeneration is inadequate if the Sa and Sb patterns persist, even if these
remain present over long periods of time. Transformations in the scar pattern are also related to the
depth of the ulcer defect as noted at endoscopy.14(502) When the ulcer defect is no deeper than the
submucosal layer, only the Sc pattern is seen. When the defect penetrates into the muscularis propria,
regeneration occurs by transformation from the Sa or Sb pattern to the Sc pattern.
We investigated the relationship between our endoscopic classification of ulcer scars and ulcer relapse
in a prospective endoscopic follow-up study of 77 patients who received maintenance therapy with an
H2 blocker (at half the therapeutic dosage).14(503) Transformation of scar patterns and ulcer relapses
during the 2-year follow-up period are shown in Table 132. A high rate of relapse (84%) was observed
in patients with persistence of the Sa scar. The relapse rate for patients with an Sb scar pattern was
also high (33%) but less than with persistence of the Sa pattern. By contrast, relapses did not occur in
the group with the Sc pattern. We concluded that ulcer scar patterns as seen at endoscopy relate to
the depth of the ulcer defect and can be correlated with the likelihood of ulcer recurrence.
Transformation of Scar Patterns and Ulcer Relapses*
TABLE 132
SCAR
PATTERN
TRANSFORMATION
CASES WITH
ULCER RELAPSE
MEAN INTERVAL
TO RELAPSE
0/13
Sc Sc
0/17
Sa, Sb Sc
5/15 (33%)
23.2 months
Sb
Sa, Sb Sb
27/32 (84%)
11.6 months
Sa
Sa Sa
From Sakaki N, Takemoto T. The relationship between endoscopic finding of gastric ulcer scar
and ulcer relapse. J Clin Gastroenterol 1993; 17(suppl 1): S649.
* Prospective 2-year follow-up study in 77 patients with maintenance H2-blocker therapy at half
dosage.
P < .01: Sc vs. Sb, Sc vs. Sa, and Sb vs. Sa.
Sc
Other Digestive Tract Organs

Colon
The diagnostic unit (frame of reference) for magnification colonoscopy is the colonic pit. These
structures are actually the orifices of the crypts of Lieberkhn. Chromoscopy is useful for endoscopic
observation of the colonic pits, and the normal colonic mucosa stains deeply with methylene blue (see
Chapter 12: Chromoscopy, and Chapter 24: Early Colorectal Cancer and Endoscopic Resection). The
shapes of colonic pits change under the influence of certain disease states and show characteristic
patterns. However, these are very similar to those seen in the stomach and can be described by our
classification. Fundamentally, normal colonic pits are characterized by a Type A pattern. During the
healing process of inflammatory bowel disease, a Type B pattern is transiently observed. A Type C
pattern is characteristic of neoplastic lesions.
Small Intestine
The villi of the small intestinal mucosa range in size from 0.2 to 1.0 mm. They are easily observed with
low magnification and are often clearly seen with standard instruments. Normal villi show Types C and
D patterns in magnified views. A coarse Type D pattern is characteristic of regeneration.
Esophagus
Magnification endoscopy of esophageal mucosa can be performed using the technique of iodine
staining (see Chapter 12: Chromoscopy). However, the technique of magnification esophagoscopy is
not well developed, and information on the magnified view of esophageal epithelium is lacking.
REFERENCES
1.
Sakaki N, Iida Y, Okazaki Y, et al. Magnifying endoscopic observation of the gastric mucosa,
particularly in patients with atrophic gastritis. Endoscopy 1978;10:26974.
2.
Gutzeit H, Teitige H. Die Gastroskopie, Lehrbuch und Atlas. Munchen: Urban &
Schwarzenberg. 1954.
3.
Salem SN, Truelove SC. Dissection microscope appearance of the gastric mucosa. BMJ
1964;2:15034.
4.
Matsumoto M. Dissecting microscopic study of gastric ulcer and gastric cancer. Gastroenterol
Endosc 1973;15:63965.
5.
Yoshii T. Staining dissecting microscopic observation and its application for endoscopy. In
Takemoto T, Kawai K (eds). Gastrointestinal Endoscopy With Application of Dye. Tokyo:
Igaku Shoin, 1974;1120.
6. Nishizawa M, Kariya A, Kobayashi S, Shirakabe H. Clinical application of an improved
magnifying fiber-colonoscope (FCS-ML II) with special reference to the remission features of
ulcerative colitis. Endoscopy 1980;12:7680.
7. Tada M, Misaki F, Kawai K. Endoscopic observation of villi with magnifying
enterocolonoscopes. Gastrointest Endosc 1982;28:179.
8.
Tada M, Nishimura S, Watanabe Y, et al. A new method for the ultra-magnifying observation
of colonic mucosa. J Kyoto Pref Univ Med 1982;91:34954.
9.
Sakaki N, Takeuchi K, Saito M, et al. Study for development of ultra-high magnifying
fibergastroscope [Japanese, English abstract]. Gastroenterol Endosc 1985;27:659.
10.
Tada M, Shimada M, Murakami F, et al. Development of the strip-off biopsy [Japanese,
English abstract]. Gastroenterol Endosc 1984;26:8339.
11.
Sakaki N, Harada H, Takeuchi K, et al. Magnifying and endoscopic diagnosis of gastric ulcer
healing [Japanese, English abstract]. Stomach Intestine 1984;19:97986.
12.
Sakaki N, Takemoto T. Endoscopic study on stage of healing process of duodenal ulcer
[Japanese, English abstract]. Gastroenterol Endosc 1988;30:19149.
13. Takemoto T, Sakaki N, Tada M, et al. Evaluation of peptic ulcer healing with a highly
magnifying endoscope: Potential prognostic and therapeutic implications. J Clin Gastroenterol
1991;13(suppl 1):S1258.
14. Sakaki N, Takemoto T. The relationship between endoscopic finding of gastric ulcer scar and
ulcer relapse. J Clin Gastroenterol 1993;17(suppl 1):S649.
Chapter 14 The Use of Histoacryl Tissue Adhesive for the Treatment

of Variceal Bleeding
(504)
(505)
KENNETH F. BINMOELLER, M.D.
NIB SOEHENDRA, M.D.
Since the mid-1980s, endoscopic sclerotherapy has become a routine treatment for acute esophageal
variceal bleeding. A major advantage of sclerotherapy is that it can be performed at the same time as a
diagnostic endoscopy. Randomized controlled studies comparing sclerotherapy with balloon
tamponade and vasopressin infusion have shown sclerotherapy to be more efficacious in arresting
variceal hemorrhage and to have lower complication rates.14(506) However, sclerotherapy is not a
panacea for the treatment of bleeding varices. In particular, the treatment of gastric varices has been
problematic, and studies have shown poor results and high complication rates.5,6(507) Further, early
recurrence of bleeding after sclerotherapy of esophageal varices still occurs, approaching 30% or more
of patients in some studies.1,2,7(508) In several trials comparing sclerotherapy with balloon
tamponade and vasopressin infusion, no significant difference was found in early recurrence of
bleeding or in mortality rates among the treatment groups.1,2(509)
Endoscopic variceal obliteration with the polymer tissue adhesives isobutyl-2-cyanoacrylate (bucrylate,
not available commercially) and N-butyl-2-cyanoacrylate (Histoacryl, Braun, Melsungen, Germany) is
one answer to the shortcomings of sclerotherapy. The fundamental technique for tissue adhesive
injection is the same as that for sclerotherapy. This chapter describes the use of tissue adhesives for
the management of bleeding gastroesophageal varices and describes the role of these agents in
relation to conventional sclerotherapy.
Historical Perspective
Since their discovery in 1949, cyanoacrylate polymers have been widely studied and applied clinically
in various surgical subspecialties as tissue adhesives. Since the mid-1970s, interventional radiologists
have used the polymers for embolization of aneurysms, arteriovenous malformations, fistulas, and
vascular lacerations.8,9(510) Martin et al.10(511) used a tissue adhesive
(trifluoro-2-propyl-2-cyanoacrylate) at endoscopy in an effort to control hemorrhage in six patients with
a variety of bleeding lesions (esophagitis, malignant tumor, varix, duodenal ulcer). However, Protell et
al.11(512) found that use of this agent was not very effective in controlling hemorrhage from ulcers in
an experimental animal model. Vallieres et al.12(513) reported injection of N-butyl-2-cyanoacrylate
monomer into a bleeding duodenal ulcer. Although bleeding was controlled initially, pancreatoduodenal
necrosis developed within a matter of hours and the patient underwent pancreatoduodenectomy.
In 1974, Rsch et al.13(514) proposed the use of bucrylate for the management of gastroesophageal
varices following successful embolization of the gastric coronary vein in an animal model. The first
clinical obliteration of gastric and esophageal varices was reported by Lunderquist et al.14(515) (1978)
via a percutaneous transhepatic approach. Gotlib and Zimmermann15(516) (1984) described the first
use of bucrylate for the endoscopic obliteration of esophageal varices and Soehendra et al.16(517)
(1986) discussed its use for gastric varices. Whereas Gotlib and Zimmermann used bucrylate in lieu of
a sclerosant for the elective treatment of varices, Soehendra et al. (1987)17(518) proposed that
bucrylate treatment be restricted to acutely bleeding varices and the elective treatment of gastric
varices.
Properties of Cyanoacrylate Polymers

The cyanoacrylate polymers have a viscosity and appearance similar to those of water. Polymerization
occurs on contact with water to form a solid complex tightly bound with any underlying tissue.
Polymerization is almost instantaneous in blood. Experimental and clinical studies have shown that
cyanoacrylate polymers have both bacteriostatic and hemostatic activity.18,19(519) Half-life,
biodegradability, and tissue reactivity vary according to the alkyl chain length of the different homologs;
half-life increases and tissue toxicity decreases with lengthening of the alkyl chain. Longer alkyl chain
homologs have less bacteriostatic activity. Histoacryl, which has a long alkyl chain, is the least
histotoxic of the cyanoacrylate polymers currently available commercially.20(520) The use of Histoacryl
in the treatment of varices is described.
Technique and Indications for Variceal Obliteration

Preparation and Instruments
Patient preparation for Histoacryl injection is the same as that for conventional sclerotherapy. The use
of Histoacryl does not require any special laboratory tests or monitoring. All personnel working with
Histoacryl should wear eye protection as a precaution against inadvertent spraying of the tissue
adhesive during injection. The patient's eyes should also be protected during injection.
An endoscope with a large working channel (at least 3.7 mm in diameter) is recommended to allow for
suction after inserting the injection catheter. Several injection catheters should be on hand. We use an
injection needle measuring 7 mm in length and 0.7 mm in diameter. The needle should be checked
before use, ensuring that it can be fully retracted and extended from the covering sheath. Two suction
outlets should be available, one for oropharyngeal and the other for endoscopic suction.
The liquid tissue adhesive can instantaneously solidify at the tip of the endoscope or within the working
channel, thereby damaging the instrument. This can be avoided by lubricating the insertion tube
liberally with silicone oil and aspirating oil through the accessory channel. With this precautionary
measure, we have not experienced any damage to an endoscope.
To prevent premature solidification during injection, Histoacryl is diluted with the oily contrast agent
Lipiodol (Byk Gulden, Konstanz, Germany). When Histoacryl is mixed in a ratio of 0.5 ml (volume per
tube of Histoacryl) to 0.8 ml Lipiodol (total volume, 1.3 ml per syringe), hardening is delayed by
approximately 20 sec.21(521) The two components are drawn up together into a 2-ml plastic syringe
and then mixed by inverting the syringe several times. To help prevent Histoacryl from adhering to the
catheter wall, several milliliters of Lipiodol are injected into the catheter, followed by a syringe full of air.
This coats the inner wall of the catheter with Lipiodol.
Although Histoacryl has a viscosity similar to that of water, the addition of Lipiodol increases the
viscosity considerably. Rapid injection of the Histoacryl mixture may therefore require considerable
injection pressure. Use of a large-bore sclerotherapy needle (0.7 mm in diameter) makes the injection
easier. We also recommend 2-ml syringes for the injection of Histoacryl. Use of a syringe with a
Luer-Lok prevents inadvertent spraying of Histoacryl during injection.
Most of the Histoacryl is retained in the dead space of the injection catheter following injection. To
ensure that the total volume of Histoacryl injected is actually deposited in the varix, it is necessary to
follow the Histoacryl injection with a second injection of distilled water. The volume injected should be
equal to the dead space of the catheter, about 0.7 ml for standard catheters. Several 2-ml syringes
filled with distilled water should be prepared for this purpose.
Injection Technique
Once localized, the bleeding varix is punctured in the usual manner for intravariceal injection (Figures
141 and 142). Theoretically, the tissue adhesive should be injected distal to the bleeding point
because of the direction of venous blood flow in portal hypertension. However, we have found this to
be irrelevant as long as the injection is intravariceal and the puncture is close to the bleeding point.
Paravariceal injection may lead to severe ulceration. Intravariceal position can be verified by first
injecting a small volume of distilled water into the varix. This should not cause mucosal swelling.
Because Lipiodol is radiopaque, the injection can also be monitored fluoroscopically. This is a distinct
advantage if the equipment is available.
(522)Figure 141. Endoscopic photographs of Histoacryl injection of esophageal varices. A,

Actively bleeding esophageal varix. B, Intravariceal injection of Histoacryl. C, Cessation of
bleeding after injection of Histoacryl.
(523)Figure 142. Endoscopic photographs of Histoacryl injection of gastric (fundic) varices.

A, Conglomerate of polypoid varices in the gastric fundus as seen in retroflexion. B,
Intravariceal injection of Histoacryl. Histoacryl is seen emerging from the varix rupture site
(arrow). C, Magnified view showing plugging of the rupture site with solidified Histoacryl.
The Histoacryl-Lipiodol mixture is injected in small aliquots into esophageal (0.5 ml) and gastric (1 ml)
varices, followed immediately by an injection of distilled water to flush out any Histoacryl retained in the
catheter (see earlier in this chapter). After injecting the tissue adhesive, the injection catheter is
advanced several centimeters beyond the endoscope tip and flushed continuously with distilled water
to keep it patent for further injections. As a precautionary measure to prevent endoscope damage, one
should not aspirate through the accessory channel during and for about 20 sec after the injection of the
tissue adhesive.
Treatment of Nonbleeding Varices

Active bleeding from gastroesophageal varices can usually be arrested with one or two injections of
Histoacryl. After hemostasis is achieved, nonbleeding varices are treated at the same session to
decrease the risk of further variceal bleeding.
We use Histoacryl or a conventional sclerosant to treat nonbleeding esophageal varices. Although
Histoacryl can obliterate varices more rapidly than sclerotherapy, it may cause ulcerations when
inadvertently injected paravariceally. Therefore, we limit its use to patients judged to be at higher risk
for variceal bleeding and those who have a higher risk of death from a variceal bleed. These subsets of
patients are more likely to benefit from rapid varix obliteration. We use endoscopic criteria (size,
presence of red color signs) as well as the clinical severity of underlying liver disease to judge the risk
for variceal bleeding.22(524) Patients with varices at low risk for bleeding are treated with conventional
sclerotherapy.
We perform sclerotherapy with 1% polidocanol (Aethoxysclerol, Kreussler, Wiesbaden, Germany), but
no sclerosant has been shown to be superior to another. No consensus exists as to the optimum
injection technique. We have found a combined perivariceal and intravariceal technique to be the most
efficacious (Figure 143).21(525) A maximum of 5 ml of sclerosant is first injected immediately
adjacent to the varices to produce submucosal wheals that compress the varix. The injection needle is
then introduced intravariceally and, while injecting 1 to 2 ml of the sclerosant, the needle is shifted to
and fro in a deliberate attempt to damage the intima of the varix. Both injury to the intima and
compression of the varix by submucosal deposits of sclerosant are thought to aid in securing
thrombosis of the varix. Recommended volumes for injection apply only to 1% polidocanol; the
selection of a different sclerosant may dictate a different injection volume.
(526)Figure 143. Diagram demonstrating the sequence of combined intravariceal and

perivariceal sclerotherapy technique. A, Perivariceal injection on both sides of the varix
produces visible submucosal wheals. B, Intravariceal injection. While the sclerosant is
injected, the needle is shifted back and forth in an attempt to damage the intima of the varix.
C, Submucosal deposits of sclerosant produce varix compression on all sides.
We treat gastric varices exclusively with Histoacryl. This includes varices located immediately distal to
the gastroesophageal junction without extension into the gastric fundus, which we prefer to call
junctional varices. Junctional varices deserve special mention because they are easily overlooked.
Junctional varices may appear during or after the eradication of esophageal varices and may cause
recurrent bleeding. When bleeding occurs, these varices may be more difficult to treat than esophageal
or fundic varices owing to their location at the cardia.
We categorize fundal varices into three groups according to endoscopic appearance (Figure 144).
Type I is a single polypoid varix with feeding tributaries (Figure 145). The appearance may mimic that
of a large polyp. This type is easily obliterated with Histoacryl, usually requiring only a single injection.
Varix obliteration can be tested for by gentle probing with the tip of the injection catheter (needle
withdrawn); the structure should be firm and noncompressible. A radiograph may also be taken to
verify varix obliteration (Figure 146). Type II is a conglomerate of large, polypoid varices that tend to
be confluent. Several Histoacryl injections into different varices are required, and more than one
treatment session may be necessary. Type III varices are slender, serpiginous varices that tend to be
widely distributed in the gastric fundus and body. Varices with this morphology rarely bleed, and
therefore we do not treat them. Intravariceal injection with Histoacryl may also be difficult owing to the
slender caliber.
(527)Figure 144. Diagram representing the morphologic classification of fundal varices as

outlined in the text. A, Type I: a single polypoid varix. B, Type II: a conglomerate of large
polypoid varices. C, Type III: slender, serpiginous varices.
(528)Figure 145. Endoscopic photographs of fundal varices (Type I) after injection with
Histoacryl. A, Fundal varix with feeding tributaries. B, Same patient 3 months following
treatment showing obliteration of varices with some residual scarring.
(529)Figure 146. Roentgenographic views corresponding to Figure 145. A, Histoacryl

filling a Type I fundal varix and feeding tributaries. B, No residual Histoacryl is seen 3 months
after treatment.
Although controversial, our policy is to treat nonbleeding gastric varices following the treatment of
bleeding esophageal varices. Similarly, we treat any nonbleeding esophageal varices that may
accompany bleeding gastric varices.
Patient Follow-Up
Esophagogastroscopy is performed 4 days after the index episode of bleeding and at weekly intervals
thereafter until all varices are obliterated. Gastric varices are checked for obliteration in the manner
described previously. Incompletely obliterated varices are reinjected with Histoacryl. The number of
injection sessions using Histoacryl depends on the morphology of the gastric varices. Sclerotherapy for
esophageal varices is continued on a weekly basis until all varices are eradicated and the inner wall of
the esophagus is fibrosed (Figure 147).
(530)Figure 147. Large esophageal varices. A, Before sclerotherapy with 1% polidocanol. B,

After sclerotherapy with 1% polidocanol.
Results
Published reports of experience with the cyanoacrylate polymer adhesives in the management of
gastroesophageal varices stem primarily from centers in France and Germany.
Gotlib and Zimmermann15(531) reported the first results of esophageal varix obliteration with bucrylate
in 96 patients. Bucrylate was used during active bleeding in 21 patients, and hemostasis was achieved
in 20 (95% success rate). Bleeding recurred in 6 patients (30%) within 8 days of treatment. Of 75
patients treated electively with bucrylate within 8 to 20 days of a bleeding episode, 32 patients (43.2%)
experienced recurrent bleeding during a mean follow-up of 9 months.
Ramond et al.23(532) reported the results of a study of bucrylate in 49 patients with bleeding
gastroesophageal varices, 12 of whom had active bleeding at initial endoscopy. Bleeding was
controlled in 14 of 15 bleeding episodes. The cumulative rate of recurrent bleeding was 37% at 6
months and 42% at 1 year. In a second study by these investigators, 27 patients with large or actively
bleeding gastric varices were treated with either bucrylate or Histoacryl.24(533) The polymer adhesives
stopped bleeding in all 6 patients treated, but bleeding recurred within 6 hours in 2 patients. The
cumulative rate of recurrent bleeding was 37%. Feretis et al.25(534) performed Histoacryl injections in
23 patients with actively bleeding esophageal (19 patients) or gastric (4 patients) varices; hemostasis
was achieved in 22 patients after a single injection of Histoacryl. During a mean 2.4-month follow-up
period, patients with esophageal varices underwent conventional serial sclerotherapy, and only 1
episode of recurrent bleeding was encountered.
We began using Histoacryl in 1985 as part of an empirical treatment protocol for actively bleeding
gastroesophageal varices and nonbleeding gastric varices.21(535) Over a 6-year period (1985 to
1991), a total of 283 patients were treated with Histoacryl, 139 of whom had acute variceal bleeding at
endoscopy. Among the patients with acute bleeding, 61 had bleeding esophageal varices, 52 bleeding
"junctional" varices, and 26 bleeding fundic varices. Hemostasis was achieved with Histoacryl in all
patients. In the Histoacryl-treated group of patients, 17 (12.2%) experienced early recurrence of
bleeding (13 from esophageal varices and 4 from gastric varices). The in-hospital mortality rate was
7.2%. Causes of death were related primarily to hepatic failure, and no deaths were directly attributable
to bleeding. When the periods before and after the introduction of Histoacryl for acute variceal bleeding
are compared, the tissue adhesive substantially improved the results of treatment (Table 141).
Reflecting this improvement, balloon tamponade and decompressive shunt surgery have become
obsolete treatments for variceal bleeding in our unit.
Results of Sclerotherapy in
Patients With Acute Variceal Bleeding*
TABLE 141
19781985
(n = 168)
n (%)
19851992
(n = 119)
n (%)
Initial hemostasis
With Histoacryl
98 (82.4%)
With polidocanol
124 (73.8%)
21 (17.6%)
With polidocanol and
40 (23.8%)
balloon tamponade
Unsuccessful
4 (2.4%)
Early recurrence of bleeding

50 (30.5%)
17 (14.3%)
Hospital mortality
53 (31.5%)
13 (10.9%)
* Comparison of two treatment periods at the University
Hospital, Hamburg. Includes only patients undergoing initial
hemostasis.
Adverse Effects and Complications

Local inflammatory reactions and neural toxicity were reported in surgical series that evaluated some
of the early smaller-chain monomers for wound closure. Recent surgical studies have shown markedly
reduced tissue toxicity when Histoacryl is used for blepharoplasty and other surgical wounds.20(536)
Endoscopic obliteration of varices with bucrylate was found to cause acute ulcerations of the
esophageal wall in autopsy studies.26(537) However, we have not observed ulceration to occur when
Histoacryl injections are strictly intravariceal, in contrast to inadvertent paravariceal injection, which can
cause extensive ulceration.
A concern has been raised regarding potential carcinogenicity of the polymer cyanoacrylates. In
experimental studies, polymer cyanoacrylates have been shown to induce sarcomas when implanted
into rat peritoneum.27(538) However, the dose used was 100 times that used in humans, and the rat
species selected for this study is known to have a predisposition to sarcoma formation. These findings
have not been reproduced in other experimental models or in humans. In particular, extensive clinical
experience with intraperitoneal implantation of cyanoacrylates has not shown carcinogenicity; in more
than 2000 patients treated with intravascular instillation of bucrylate, none has been reported to have
associated or concomitant neoplasm.28(539)
In the application of Histoacryl for the obliteration of varices, the tissue adhesive is injected into the
varix, solidifies, and is gradually extruded en masse over a period of weeks into the lumen and passed
in the feces (Figures 148 and 149). As shown in Figure 146, the complete disappearance of
Histoacryl can be demonstrated roentgenographically. This minimizes a long-term foreign body
reaction.
(540)Figure 148. Endoscopic photograph showing extrusion of a Histoacryl cast 1 month

after varix obliteration.
(541)Figure 149. Photograph of an extruded Histoacryl venous cast.

Serious side effects of endoscopic injection of polymer cyanoacrylates have been limited to two cases
(one fatal) of cerebrovascular stroke resulting from acute embolization.29(542) It is thought that
anomalous right-to-left shunting predisposed to this rare complication. Other complications have been
minor and self-limited. In a review of 317 patients treated with bucrylate or Histoacryl over a 5-year
period, Gotlib30(543) reported no procedure-related mortality and only minor complications, including
dysphagia without stenosis, esophageal stenosis, bacteremia (2 cases), and pyrexia (1 case). In our
hands, the intravariceal injection of Histoacryl has not been associated with any serious complications.
These excellent results may be due to the technique and the doses used. In contrast to our findings,
the serious complications occurred with the use of larger-injection aliquots (1 to 2 ml) and two or three
injections per varix.
Approximately 1 week after intravariceal injection of Histoacryl, the mucosa overlying the obliterated
varix begins to slough. As mentioned earlier, the tissue adhesive is treated as a foreign body and is
gradually expelled into the lumen (see Figures 148 and 149). Several months may elapse before it is
completely eliminated from the wall. This usual sequence of events following Histoacryl injection is not
associated with increased bleeding or other adverse effects.
Summary
Endoscopic variceal injection is widely accepted as the treatment of choice for acutely bleeding
esophageal varices. The availability and simplicity of the method have contributed to its popularity.
However, high rates of recurrent bleeding and technical difficulty treating bleeding gastric varices have
been major shortcomings of the technique. Endoscopic variceal obliteration with polymer
cyanoacrylates addresses these shortcomings. In a limited number of centers, Histoacryl has been
shown to be safe and highly effective in arresting acute gastroesophageal variceal bleeding and
permanently obliterating varices. These encouraging results indicate a need for further multicenter
clinical trials. The polymer cyanoacrylates do not replace conventional sclerotherapy, which is well
suited for eradicating nonbleeding varices and has the final aim of fibrosis of the inner esophageal wall.
Techniques for the use of polymer cyanoacrylates are still in a state of evolution, and the ultimate role
of each remains to be determined.31(544)
Acknowledgments
The authors would like to thank Dr. J. McCarthy, for his assistance in the preparation of this
manuscript, as well as the support of the nursing staff during these studies.
REFERENCES
1.
Westaby D, Hayes PC, Gimson AES, et al. Controlled clinical trial of injection sclerotherapy
for active variceal bleeding. Hepatology 1989;9:2747.
2. Sderlund C, Ihre T. Endoscopic sclerotherapy v. conservative management of bleeding
oesophageal varices. Acta Chir Scand 1985;151:44956.
3. Larson AW, Cohen H, Zweiban B, et al. Acute esophageal variceal sclerotherapy: Results of a
prospective randomized controlled trial. JAMA 1986;255:497500.
4. Barsoum MS, Bolous FI, El-Rooby AA, et al. Tamponade and injection sclerotherapy in the
management of bleeding oesophageal varices. Br J Surg 1982;69:768.
5. Trudeau W, Prindiville T. Endoscopic injection sclerosis in bleeding gastric varices.
6.
Yassin ZM, Eita MS, Hussein A. Endoscopic sclerotherapy for bleeding gastric varices. Gut
1985;26:A1105.
7. Copenhagen Esophageal Varices Sclerotherapy Project. Sclerotherapy after first variceal
hemorrhage in cirrhosis. N Engl J Med 1984;311:15941600.
8. Zanetti PH, Sherman FE. Experimental evaluation of a tissue adhesive as an agent for the
treatment of aneurysms and arteriovenous anomalies. J Neurosurg 1972;36:729.
9.
Lehman RA, Hayes GJ. The toxicity of 2-cyanoacrylate tissue adhesives: Brain and blood
vessels. Surgery 1967;91522.
10. Martin TR, Onstad GR, Silvis SE. Endoscopic control of massive upper gastrointestinal
bleeding with a tissue adhesive (MBR4197). Gastrointest Endosc 1977;24:746.
11.
Protell RL, Silverstein FE, Gulacsik C, et al. Failure of cyanoacrylate tissue glue (Flucrylate,
MBR4197) to stop bleeding from experimental canine gastric ulcers. Dig Dis 1978;23:9038.
12. Vallieres E, Jamieson C, Haber GB, Mackenzie RL. Pancreatoduodenal necrosis after
endoscopic injection of cyanoacrylate to treat a bleeding duodenal ulcer: A case report.
Surgery 1989;106:9013.
13. Rsch J, Goldman ML, Dotter CT. Experimental catheter obstruction of the gastric coronary
vein. Possible technique for percutaneous intravascular tamponade of the gastroesophageal
varices. Invest Radiol 1975;10:20611.
14.
Lunderquist A, Borjesson B, Owman T, et al. Isobutyl 2-cyanoacrylate (Bucrylate) in
obliteration of gastric coronary vein and esophageal varices. AJR 1978;130:16.
15.
Gotlib JP, Zimmermann P. Une nouvelle technique de traitement endoscopique des varices
oesophagiennes: L'obliteration. Endoscopia Digestiva 1984;7:102.
16. Soehendra N, Nam V Ch, Grimm H, et al. Endoscopic obliteration of large esophagogastric
varices with Bucrylate. Endoscopy 1986;18:256.

17. Soehendra N, Grimm H, Nam V Ch, et al. N-Butyl-2-cyanoacrylate: A supplement to
endoscopic sclerotherapy. Endoscopy 1987;19:2214.
18. Eiferman RA, Snyder JW. Antibacterial effect of cyanoacrylate glue. Arch Ophthalmol
1983;101:95860.
19. Matsumoto T, Hardaway RM, Heisterkamp CA, et al. Cyanoacrylate adhesive and
hemostasis. Arch Surg 1967;94:85860.
20. Kamer FM, Joseph JH. Histoacryl. Its use in aesthetic facial plastic surgery. Arch Otolaryngol
Head Neck Surg 1989;115:1937.
21. Soehendra N, Grimm H, Maydeo A, et al. Endoscopic sclerotherapyPersonal experience.
Hepatogastroenterology 1991;38:2203.
22. The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices.
Prediction of the first variceal hemorrhage in patients with cirrhosis of the liver and
esophageal varices. A prospective multicenter study. N Engl J Med 1988;319:9839.
23. Ramond MJ, Valla D, Gotlib JP, et al. Obturation endoscopique des varices oeso-gastriques
par le Bucrylate. Itude clinique de 49 malades. Gastroenterol Clin Biol 1986;10:5759.
24. Ramond MJ, Valla D, Mosnier JF, et al. Successful endoscopic obturation of gastric varices
with butyl cyanoacrylate. Hepatology 1989;10:48893.
25. Feretis C, Tabakopoulos D, Benakis P, et al. Endoscopic hemostasis of esophageal and
gastric variceal bleeding with Histoacryl. Endoscopy 1990;22:2824.
26. Fabiani B, Degott C, Ramond MJ, et al. Obturation endoscopique des varices oeso-gastriques
par le Bucrylate. IItude morphologique a partir de 12 cas autopsies. Gastroenterol Clin
Biol 1986;10:5803.
27.
Reiter A. Sarkomerzeugende Wirkung des Gewebeklebers Histoacrylblau an der Ratte. Z Exp
Chir Transplant Knstliche Organe 1987;20:5560.
28. Samson D, Marshall D. Clinical versus experimental use of isobutyl-2-cyanoacrylate. J
Neurosurg 1987;67:3189.
29. See A, Florent C, Lamy P, et al. Accidents vascularies cerebraux apres obturation
endoscopique des varices oesophagiennes par l'isobutyl-2-cyanoacrylate chez deux malades.
Gastroenterol Clin Biol 1986;10:6047.
30.
Gotlib JP. Endoscopic obturation of esophageal and gastric varices with a cyanoacrylic tissue
adhesive. Can J Gastroenterol 1990;9:6378.
31. Binmoeller KF, Soehendra N. Endoscopic sclerotherapy in 1991: Still evolving [Editorial].
Endoscopy 1991;23:2868.
Chapter 15 Tissue Glue for Fistulas

(545)
(546)
MICHAEL JUNG, M.D.
Interventional endoscopy is frequently an alternative to surgery. Injection techniques for gastrointestinal

hemorrhage, endoscopically assisted dilation, and prosthesis insertion have replaced more complex
and hazardous methods of surgical intervention. In recent years, endoscopic techniques have been
developed for treating defects and fistulas of the gastrointestinal tract by application of a variety of
agents, including artificial glues and quick-hardening substances.
Cyanoacrylate (Histoacryl), a rapidly polymerizing substance, can induce the formation of granulation
tissue in a fistula by foreign body reaction.1,2(547) Initially, this agent functions like an external clamp
that only briefly seals the opening of the fistula and is then rejected. Cyanoacrylate was first used
successfully for gluing recurrent congenital tracheoesophageal fistulas (in esophageal atresia).3,4(548)
Frequent failures, complicated revisions, and equipment defects caused by the glue were
counterbalanced by successful treatment in some cases.5(549) If injected intramucosally or
submucosally, polidocanol (Aethoxysclerol, Kreussler, Wiesbaden, Germany) causes occlusive edema
and severe, partially destructive inflammation of the tissue. In individual cases, the local inflammatory
stimulus has been used to close defects.6,7(550) Electrocoagulation induces similar tissue changes
and has been used to induce local wound healing.8(551) Prolamin sulfate (Ethibloc), a curing amino
mixture, has been used mainly for blocking the pancreatic duct.9,10(552) Cases have also been
reported of successful precipitation of this agent in necrotic tumor cavities involving the esophagus and
rectum, in addition to gluing of gastrointestinal fistulas.1113(553) High viscosity and hardening times
of up to 30 min, however, have limited the use of this substance.14(554)
None of the agents or gluing techniques described previously has proved to be entirely satisfactory. In
recent years, interest has therefore focused on fibrin glues, biologic substances that meet diverse
prerequisites for gluing and healing defects (Tissucol/Tisseel, Immuno, Heidelberg/Vienna; Beriplast,
Behringwerke, Marburg).15(555)
Theoretical Aspects of Fibrin Gluing

Fibrin glues were used in surgery for the first time in 1943 by Michael and Abbott16(556) for skin
transplantation. A year later, Cronkite et al.17(557) used them for gluing nerves and for skin
transplantation. Today, surgical indications are myriad: establishing hemostasis, securing
anastomoses, gluing of ruptured parenchymatous organs, and treating fistulas.18,19(558)
Fibrin gluing imitates the last phase of plasmatic coagulation, whereby wound healing is stimulated in a
physiologic fashion.20(559) The gluing system incorporates two separate intermediary substances,
namely, highly concentrated human fibrinogen, fibronectin, and Factor XIII, as well as thrombin in
different concentrations dissolved in calcium chloride (4 IU/ml or 500 IU/ml; Tissucol, Immuno,
Heidelberg). The components are drawn up separately into two different syringes and attached to a
specially designed syringe holder. Mixing the substances together causes fibrinogen to convert to
fibrin. Activation by Factor XIII induces the cross-linking of fibrin with fibronectin. On the macroscopic
level, this combination results in a gray, ultimately white clot that develops, depending on the thrombin
component, within 30 to 120 sec (rapid or slow gluing) (Figure 151). Factor XIII is relevant for the
cross-linking of fibrin by means of covalent bonding to the fibrin molecule. This plays a crucial role in
the graft's resistance to tearing. Endogenous fibrinolysis is impeded by adding aprotinin (approximately
3000 KIU/ml) to the glue mixture.
(560)Figure 151. The forming fibrin clot. Note the color change from gray to white.
For use in practice, both agents required for gluing are supplied in a double-syringe system (Duploject
system- application catheter Duplocath, Immuno, Heidelberg) that separates them until application, at
which time they are combined via a Y-shaped end piece (Figure 152). In this way, the two
components are not united until they reach the end of the carrier channel where the fibrin plug forms.
(561)Figure 152. Double syringe (Duploject) with adapted catheter. The catheter for fibrin
glue instillation is passed through the channel of a flexible bronchoscope with an insertion
tube diameter of 6 mm (BF 1T10, Olympus Optical Company, Ltd.).
By interconnecting with local tissue, the glue initially forms a seal. Subsequently, granulation tissue
grows. Because the percentage of fibrin is 30 times higher than in a naturally formed blood clot, the
resulting fibrin clot has much greater resistance to tearing.20(562)
Comparative studies of gluing agents currently in use suggest that they have physicochemical and
electron microscopic differences.21(563) Solubility patterns and the osmotic pressures of the
fibrinogen and thrombin solutions differ within specific limits. Apparently, the Tisseel clot most closely
approximates physiologic fibrin cross-linking.
Because human fibrinogen is a component of the fibrin glue, the glue components must be obtained
from the pooled human plasma of controlled long-term donors. At every plasmapheresis, the donor
plasma is tested for viral genomes of human immunodeficiency virus (HIV), hepatitis B virus, and
hepatitis C virus, and any significant elevations in transaminases. The manufacturing procedure
requires six logarithmic levels, so that no transmission of HIV is possible.22(564) No infections induced
by fibrin glue have been reported.
Application Technique
Fibrin glue (Tissucol) is available as a deep-frozen, lyophilized substance in a double syringe. Before
use, the components must be thawed. Application must be performed at a temperature of 37 C, with
both components in a completely dissolved form. Application at temperatures other than the optimum
of 37 C results in instability of the clot and inadequate therapy. In interventional endoscopy, the
comparatively long pathways that the components must travel through the endoscope must be
considered. Double-lumen catheters can be used to guide the injection of the components separately
to the tissue surface. However, consecutive injection of the two components with single-lumen
catheters is also a reliable technique for gluing.15(565) A spray-type catheter is not required.
Technique
Successful gluing in the digestive tract requires contact with tissue that has the capability of local
regeneration. Because this is necessary for the development and adaptation of the fibrin clot, the
tissue surrounding a fistula must have the capacity for wound healing. The glue does not function
exclusively as a sealing plug but provides a substrate for the migration of fibroblasts. Ideally, the glued
surface is replaced by scar tissue and the fibrin glue completely decomposes.23(566) At least 4 weeks
are required to reach this stage.
A standardized, generally accepted endoscopic technique for gluing is outlined in Table 151. After
endoscopic visualization and cannulation of the defect, deepithelialization should first be performed by
electrocoagulation of the margins or by roughening the fistulous canal with a brush. Mild bleeding is
desirable. These measures improve adhesion of the fibrin clot. The actual gluing is performed by
completely filling the defect with glue and sealing the inner defect with a clot. After 3 to 5 min, the clot
attains about 70% of its maximum strength.24(567) Thereafter, mechanical stress to the glued surface
should be avoided. Patients with esophageal leakage and fistulas are fed enterally via a nasoduodenal
tube until the wound has completely healed. Administration of antibiotics is recommended. The fistula
is then assessed once per week endoscopically and radiologically (Gastrografin swallow), and gluing
sessions are repeated until the fistula has closed completely.
TABLE 151
Technique for Endoscopic Fibrin
Gluing
Drain fistula-connecting cavity (pleural, abdominal)
Localize fistula endoscopically and radiologically
Deepithelialize margins and canal of fistula (electrocoagulation,
cytology brush)
Instill fibrin glue into canal and around borders of fistula
Avoid mechanical stress to glued surface
Reevaluate endoscopic and radiologic control (weekly)
Further gluing sessions (up to five if necessary)
Indications
The general indications for endoscopic gluing are outlined in Table 152. Wound sealing is most
successful if the fistula is tubular and short. Endoscopic gluing is not indicated for broad, window-like
openings between the esophagus and the trachea. Therapy may fail for tissue that is chronically
infected or damaged by radiation therapy owing to the lack of regenerative capacity.25,26(568) The
clot rarely sticks to malignant fistulas.
General Indications for Endoscopic

Tissue Gluing in the Digestive Tract
TABLE 152
General
Absolute
Relative
Difficult
Not indicated
Small, tubular defects that can be well visualized

and are not connected to the abdominal cavity
Postoperative leaks, inflammatory fistulas,
instrumental perforations involving esophagus,
trachea, and bronchus
Defects induced by radiation therapy
Chronic, infectious fistulas
Tumor fistulas
Defects and leaks involving the esophagus that can be reached using a flexible endoscope represent
the primary indications for gluing. The best results are achieved for this indication with favorable
anatomic position of the esophagus in the posterior and middle mediastinum and its complete
separation from the pleura and abdomen. Fistula penetration in an extraesophageal cavity (e.g.,
pleural space, abdomen) requires that the cavity be drained to relieve pressure prior to gluing.
Likewise, fistulas due to breakdown of a bronchial stump and bronchopleural fistulas can be glued after
primary thoracic drainage.7,15,27(569)
Esophagus
The main indications for endoscopic gluing are recurrent congenital tracheoesophageal fistulas in the
case of esophageal atresia, anastomotic breakdown and leakage (esophagogastrostomy,
esophagojejunostomy), spontaneous inflammatory fistulas, and traumatic or instrumental perforations
of the esophagus.24,2831(570) The extent and the speed of wound healing depend primarily on the
degree of local tissue irritation. Fistulas involving malignant tumors and those induced by radiation
therapy are equally problematic; multiple glue applications are rarely successful. The therapy of choice
for tracheoesophageal fistulas that develop in association with malignant tumors is implantation of an
endoprosthesis (see Chapter 40: Benign and Malignant Tumors of the Esophagus). Occasionally, the
use of glue (fibrin or artificial glue) can be of assistance in cases in which prosthetic bridging is
incomplete.15,32(571)
The more comprehensive reports on successful gluing refer mostly to esophageal leakage and
fistulas.26,30,31(572) The success rates for treatment of esophageal wounds are stated to be 80 to
90% (Table 153). The conservative endoscopic treatment of esophageal perforations should
therefore include the therapeutic concept of fibrin gluing (Figures 153, 154 and 155).
TABLE 153
CAUSE OF FISTULA
Results of Endoscopic Fibrin Sealing in the Gastrointestinal (GI) Tract

UPPER GI TRACT
LOWER GI TRACT
FISTULA CLOSURE
UNSUCCESSFU
TABLE 153
CAUSE OF FISTULA
Results of Endoscopic Fibrin Sealing in the Gastrointestinal (GI) Tract

UPPER GI TRACT
LOWER GI TRACT
FISTULA CLOSURE
UNSUCCESSFU
Congenital
14
11
(2 x Histoacryl)
Postoperative
13
2
14
Radiation induced
3
2
Inflammatory
1
1
Trauma
1
1
Tumor
4
1
Total
36
2
30
Adapted from Jung M, Manegold BC, Brands W. Endoscopic therapy of gastrointestinal fistulae with fibrin tissue sealant.
Waclawiczek HW (ed). Progress in Fibrin Sealing. Heidelberg: Springer-Verlag, 1989; 4352
(573)Figure 153. Gastrografin swallow x-ray demonstrating a high-grade esophageal

stenosis caused by mediastinal lymph node metastasis in a 60-year-old woman with breast
carcinoma.
(574)Figure 154. A, Gastrografin x-ray demonstrating implantation of an endoprosthesis for

relief of dysphagia in the same patient as in Figure 153. B, Gastrografin x-ray demonstrating
a large perforation (arrows) into the mediastinum. C and D, Endoscopic views of the
perforation.
(575)Figure 155. A and B, Endoscopic views showing the technique of fibrin sealing of the
perforation (same patient as in Figures 153 and 154). C and D, Endoscopic views of the
fibrin seal (same patient as in Figures 153 and 154). E, Gastrografin x-ray demonstrating
complete healing of the perforation at 3 weeks after initiation of gluing therapy. The patient
had no further dysphagia.
Stomach
The treatment of fistulas involving the stomach and duodenum is difficult as long as a connection to the
abdominal cavity exists. However, success has been repeatedly achieved in gluing even the toughest
of putrid, fistulizing anastomotic failures (Figures 156 and 157).6,3336(576) The strategy for
management of such cases often starts with endoscopic visualization of the fistula (fistuloscopy) to
establish the degree of leakage and the nature of fistulous communications.37,38(577) For purulent
fistulas, abscess drainage, antibiotic therapy, and treatment of the fistula with a
streptokinase/streptodornase solution are necessary before gluing is attempted.
(578)Figure 156. A, Barium contrast x-ray demonstrating leakage at the

esophagojejunostomy site in a 37-year-old woman who underwent gastrectomy for early
gastric cancer in the cardia of the stomach. B, Endoscopic views of the anastomosis showing
suture material and small anastomotic fistula. C, Endoscopic fistula cannulation.
(579)Figure 157. A and B, Endoscopic views of fibrin clot occluding the inner opening of
fistula (same patient as in Figure 156). C, Barium contrast x-ray of the anastomotic region 10
days after endoscopic fibrin sealing demonstrating complete healing of the fistula.
Cases of perforated ulcers successfully treated by endoscopic plus intraabdominal (laparoscopic)
applications of fibrin glue have been reported.39,40(580)
Biliary Tract and Pancreas

It is possible to glue fistulas in the region of the biliary tract. After rinsing with an antibiotic solution and
administering somatostatin, the biliary and pancreatic secretions can be reduced and a clot
successfully attached to the epithelialized tissue.36,41,42(581)
Endoscopic pancreatic duct occlusion with prolamin solution has been used to control the pain
associated with chronic pancreatitis. Occlusion, however, can damage the exocrine cellular system
and lead to atrophy of the exocrine pancreas and subsequent pancreatic fibrosis. The results of this
form of treatment are not very satisfactory.9,10(582)
Fibrin glues remain only briefly in the pancreatic ductal system. Owing to the copious pancreatic
secretion with its high fibrinolytic activity, the clot dissolves within 12 to 24 hours.43(583) Short-term
blockage can be used, however, to help protect pancreatic anastomoses. When somatostatin is
administered, cutaneopancreatic fistulas can be satisfactorily sealed with fibrin glue, although
documented success is limited to individual case reports.
Colon/Rectum
Fibrin gluing of enterocutaneous and enteral fistulas in Crohn's enterocolitis is controversial. It is
performed after complicated pretreatment by peroral lavage, enteral tubal feeding, systemic
administration of antibiotics (metronidazole), medication with Factor XIII (1250 IU/day), and
7-S-immunoglobulin for 5 days with local antibiotic rinsing.44(584) The few reports of successful
outcomes are regarded with skepticism because convincing follow-up studies are lacking.
Furthermore, fibrin gluing for chronic inflammatory fistulas is known to be predominantly unsuccessful.
The healing of deep fistulas from colorectal anastomoses can be accelerated by endoscopic
application of gluing agents, provided that the defect is located below the peritoneal reflection.
Similarly, dehiscence and rupture following endoscopic bougienage of fibrous rectal strictures can be
treated successfully by gluing. As in the esophagus, improved wound healing with reduced scar
formation has been impressive.15,45(585)
Summary
Endoscopic gluing techniques play an increasing role in the definitive treatment of gastrointestinal
fistulas and leaks. The main targets for gluing are postoperative fistulas and iatrogenic defects
involving the esophagus. Compared with all other agents, fibrin glue is the most attractive because it
promotes natural wound healing. To date, the results of gluing techniques have been reported only as
case series, although the results have been spectacular in relatively large numbers of patients. Proof in
the form of randomized studies that fibrin gluing is superior to other conservative forms of treatment is
not available. Implementation of such studies poses considerable difficulties because the patient
populations that may benefit from gluing are mainly the severely ill. These issues are compounded by
the fact that the agents used for fibrin gluing are very expensive. Nevertheless, alternative therapies
may be poorly effective or nonexistent for many of the patients in whom gluing may be appropriate. For
these reasons, the further refinement of endoscopic gluing techniques and their investigation in
patients with suitable indications are clearly warranted.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Barthelemy C, Audigier JC, Fraisse H. A non-tumoral esophago-bronchial fistula managed by

isobutyl-2-cyanoacrylate. Endoscopy 1983;15:3578.
Marco C, Doncel F, Veloso E, et al. Non-surgical closure of a benign oesophagobronchial
fistula. Br J Surg 1987;74:415.
Gdanietz K, Wiesner B, Krause I, et al. Gewebekleber zum Verschluss von
sophagotrachealfisteln bei Kindern. Z Erkr Atmungsorgane 1974;141:4650.
Waag KL, Joppich I, Manegold BC. Endoscopic closure with Histoacryl of the recurrent
oesphago-tracheal fistula following oesophageal atresia. Z Kinderchir 1975;17:248.
Brands W, Lochbhler H, Raute-Kreinsen U, et al. Die Fibrinklebung angeborener
sophagusmissbildungen. Zbl Chir 1983;108:8037.
Meyer G, Lange V, Wenk H, Schildberg FW. Endoscopic sealing of gastrointestinal fistulae.
Surg Endosc 1988;2:116.
Waclawiczek HW, Chmelizek F, Koller I. Endoscopic sealing of infected bronchus stump
fistulae with fibrin following lung resections. Experimental and clinical experience. Surg
Endosc 1987;1:99102.
Rangecroft L, Bush GH, Lister J, Irving IM. Endoscopic diathermy obliteration of recurrent
tracheoesophageal fistulae. J Pediatr Surg 1984;19:413.
Rsch W, Phillip J, Gebhardt C. Endoscopic duct obstruction in chronic pancreatitis.
Endoscopy 1979;11:436.
Pap A, Flautner L, Tihanyi T, et al. Pancreatic function after endoscopic and surgical
occlusion of the pancreatic duct in patients with chronic pancreatitis. Endoscopy
1985;17:603.
Riemann JF, Nowak G. Tumorfistelokklusion im Rektum mit Ethibloc. Dtsch Med
Wochenschr 1985;110:15923.
Ell C, Riemann JF, Demling L. Endoscopic occlusion of a neoplastic esophagomediastinal
fistula by a fast hardening amino acid solution. Gastrointest Endosc 1986;32:2878.
Marone G, Santoro LM, Torre V. Successful endoscopic treatment of GI-tract fistulas with a
fast-hardening amino acid solution. Endoscopy 1989;21:479.
Lange V, Meyer G, Wenk H. Endoskopischer Verschluss gastrointestinaler Fisteln. In
Manegold BC, Jung M (Hrsg). Fibrinklebung in der Endoskopie. Berlin, New York:
Jung M, Schlicker H, Manegold BC. Therapeutische Endoskopie mit Fibrinkleber. Med Welt
1987;38:1416.
Michael G, Abbott W. The use of human fibrinogen in reconstructive surgery. JAMA
1943;123:279.
Cronkite EP, Lozner EL, Deaver JM. Use of thrombin and fibrinogen in skin grafting. JAMA
1944;124:9768.
Schlag G, Redl H (eds). Fibrin Sealant in Operative Medicine. Vols 16. Heidelberg:
Springer-Verlag. 1986.
Brands W, Mennicken C, Beck M. Preservation of the ruptured spleen by gluing with high
concentrated human-fibrinogen: Experimental and clinical results. World J Surg
1982;6:3668.
Redl H, Schlag G. Fibrin sealant and its modes of application. In Schlag G, Redl H (eds).
Fibrin Sealant in Operative Medicine. General Surgery and Abdominal Surgery. Vol 6. Berlin,
New York: Springer-Verlag, 1986;1326.
Sawada K, Yoshima H, Yamamoto H, et al. Comparative studies on the physiochemical
properties of Tisseel and a newly developed fibrin sealant. In Schlag G, Redl H (eds). Fibrin
Sealing in Surgical and Nonsurgical Fields. Vol 1. Wound Healing. New York: Springer-Verlag
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
1994;4551.
Kaeser A. Qualitts- und Sicherheitsanforderungen an Fibrinkleber. In Manegold BC, Jung M
(Hrsg). Fibrinklebung in der Endoskopie. Berlin, New York: Springer-Verlag, 1988;716.
Pesch HJ, Scheele J. Lokaler Fibrinkleberabbau im TierexperimentHistomorphologische
Untersuchungen. In Scheele J (ed). Fibrinklebung. Heidelberg: Springer-Verlag, 1984;3844.
Kaeser A, Dum N. Grundlagen der FibrinklebungWirkprinzip und Infektions sicherheit von
Tissucol. Z Herz-Thorax-Gefsschir 1987;1(suppl 1):510.
Jung M, Brands W, Manegold BC. The use of fibrin sealant for the endoscopic treatment of
fistulae. Schweiz Rundsch Med Prax 1988;77:35.
Jung M, Manegold BC, Brands W. Endoscopic therapy of gastrointestinal fistulae with fibrin
tissue sealant. In Waclawiczek HW (ed). Progress in Fibrin Sealing. Heidelberg:
York EL, Lewall DB, Hirji M, et al. Endoscopic diagnosis and treatment of postoperative
bronchopleural fistula. Chest 1990;97:13902.
Kohler B, Khler G, Riemann JF. Spontaneous esophagotracheal fistula resulting from ulcer
in heterotopic gastric mucosa. Gastroenterology 1988;95:82830.
Groitl H, Scheele J. Initial experience with endoscopic application of fibrin tissue adhesive in
the upper gastrointestinal tract. Surg Endosc 1987;1:937.
Groitl H, Horbach T, Stangl R, et al. Endoscopic application of fibrin glue in the treatment of
anastomotic defects, perforation, and fistulas in the gastrointestinal tract. In Schlag G,
Wayand W (eds). Fibrin Sealing in Surgical and Nonsurgical Fields. Vol 8. Endo-scopy. New
York: Springer-Verlag 1994;337.
Lang V, Maiwald G, Souvatzi T, et al. Endoscopic approaches for occlusion of fistulas. In
Schlag G, Wayand W (eds). Fibrin Sealing in Surgical and Nonsurgical Fields. Vol 8.
Endoscopy. New York: Springer-Verlag, 1994;5864.
Straumann A, Gyr K, Stalder GA. Tissue adhesive in the prevention of esophageal tube
migration: Report on a new method. Schweiz Rundsch Med Prax 1984;73:10867.
Cadoni S, Ottonello R, Maxia G, et al. Endoscopic treatment of a duodeno-cutaneous fistula
with fibrin tissue sealant (Tissucol). Endoscopy 1990;22:1945.
Eleftheriadis E, Tzartinoglou E, Kotzampassi K, Aletras H. Early endoscopic fibrin sealing of
high-output postoperative enterocutaneous fistulas. Acta Chir Scand 1990;156:6258.
Willems EG. Transgastric perforation of a PEG and successful endoscopic management of
the complication using the fibrin sealing technique. Endoscopy 1990;22:288.
Costantino V, Alfano D'Andrea A, Pedrazzoli S. Human fibrin sealants and postoperative
fistulas. In Schlag G, Wayand W (eds). Fibrin Sealing in Surgical and Nonsurgical Fields. Vol
8. Endoscopy. New York: Springer-Verlag, 1994;7881.
Lang V, Meyer G, Wenk H, et al. Fistuloscopyan adjuvant technique for sealing
gastrointestinal fistulae. Surg Endosc 1990;4:2126.
Nakagawa K, Momono S, Sasaki Y, et al. Endoscopic examination for fistula. Endoscopy
1990;22:20810.
Barthelm G. Erfolgreiche Fibrinklebung eines perforierten malignen Ulkus. In Manegold BC,
Jung M (Hrsg). Fibrinklebung in der Endoskopie. Berlin, New York: Springer-Verlag
1988;12430.
Moviel J, Katkhouda N, Iovine L. Laparoscopic surgical treatment of duodenal ulcer disease.
In Schlag G, Wayand W (eds). Fibrin Sealing in Surgical and Nonsurgical Fields. Vol 8.
Endoscopy. New York: Springer-Verlag, 1994;209.
Prinz RA, Pickleman J, Hoffman JP. Treatment of pancreatic cutaneous fistulas with a
somatostatin analog. Am J Surg 1988;155:3642.
Brckner M, Grimm H, Nam VC, Soehendra N. Endoscopic treatment of a pancreatic abscess
originating from biliary pancreatitis. Surg Endosc 1990;4:2279.
Doertenbach JG, Hanisch E, Kiseleczuk J, Ohliger H. Pancreatic duct occlusion with fibrin
sealant: An experimental study. In Waclawiczek HW (ed). Progress in Fibrin Sealing.
Heidelberg: Springer-Verlag, 1989;1119.
44.
45.
Eimiller A, Berg P, Born P, et al. Fibrin sealing of fistulas in Crohn's disease. In Waclawiczek
HW (ed). Progress in Fibrin Sealing. Heidelberg: Springer-Verlag, 1989;614.
Liedgens P, Joppich I, Brands W. Endoskopische Behandlung tiefer Rektum
anastomosenstenosen im Kindesalter. In Manegold BC, Jung M (Hrsg). Fibrinklebung in der
Endoskopie. Berlin, New York: Springer-Verlag 1988;1647.
Chapter 16 Endoscopic Sewing Machine

(586)
CHRISTOPHER PAUL SWAIN, M.D.
The development of endoscopically controlled sewing and stapling machines might enable
endoscopists and surgeons to perform a range of surgical procedures without making external
incisions. Clinical applications for such devices might include the following:
1. Endoscopic treatment of gastrointestinal hemorrhage by oversewing or stapling of bleeding ulcers
and varices. Such improved endoscopic technology is needed for treatment of gastrointestinal
hemorrhage because existing methods (lasers, diathermy, heater probe, injection) are less
effective for terminating bleeding from large arteries than mechanical methods such as ligation or
stapling.
2. Development of new, less invasive surgical methods such as creation of anastomoses, closure of
perforation, collection of full-thickness gastrointestinal biopsy samples using flexible endoscopes,
antireflux procedures, and operations for control of obesity. Endoscopic delivery of these
techniques might avoid the need for laparotomy, thus lessening discomfort to patients, shortening
hospital stay, and perhaps improving survival.
3. Attachment of devices to the gastrointestinal wall might offer new opportunities for diagnosis,
monitoring, and treatment. Placement of radiotelemetry capsules (RTCs) for measurements of pH
and pressure at the intestinal wall may allow detailed long-term monitoring of the gastrointestinal
environment. The ability to attach tubes to the stomach wall might make nasogastric feeding a
more reliable prospect. Drugs in membrane-bound form for topical treatment of lesions or in
slow-release form, if attachable to specific sites in the gastrointestinal tract, might offer the
prospect of new methods of treatment.
The design problems involved in developing such machines are not trivial. Access is available to only
one side of the tissue to be sewn and stapled. The suture or staple depth must be precisely controlled
to limit the risk of damage to other organs. The machines must be sufficiently small to be swallowed.
Power to and control of the machines occur via a small-diameter flexible channel either within or
parallel to the flexible endoscope.
Sewing Machines
Chain-Stitch Sewing Machines Using Thread
Sewing machines as developed in the 18th and perfected in the 19th century were designed to allow
stitches to be sewn faster than could be achieved by the human hand holding a threaded needle. The
aim of sewing in flexible endoscopy differs in that the technique should allow access to an inaccessible
site in the human body for placement of stitches without recourse to laparotomy. Conventional sewing
machines require access to both sides of the material to be sewn, a circumstance not available at
endoscopy.
One method of overcoming this difficulty is to pleat the tissue so that the needle can pierce a double
thickness.1(587) This allows the needle pushing and catch mechanisms to remain within the bowel
lumen without requiring access to the serosal surface. This principle was used in the design of the first
endoscopic sewing machine (Figure 161).
(588)Figure 161. Chain-stitch endoscopic sewing machine. A prototype endoscopic sewing

machine of transparent acrylic showing needle, thread, hooked catch mechanism, and
endoscope.
The pleat or double fold of tissue can be formed by using suction to bring a fold of tissue into a cavity in
the machine. A needle threaded through an eye close to the needle tip pierces the double layer of
tissue. A catch mechanism retains the loop of thread, and the machine is moved to the site of the next
stitch. The needle pierces tissue and then the first loop with the next loop, which in turn is retained by
the catch mechanism to form a chain of stitches (Figure 162). The machine is then withdrawn through
the mouth, pulling the loop of thread through the tissue until the threads passing through the tissue can
be detached at the mouth from the sewing machine. The thread may then be tied and a knot run down
to the sewn tissue.
(589)Figure 162. Mechanism of action of an endoscopic sewing machine. A, Sewing

machine and its relationship to the endoscope are shown, indicating tissue, needle, thread,
catch mechanism, and control wires. B, Fold of tissue is sucked into the cavity while the
threaded needle pierces a double layer of tissue. C, Catch mechanism retains the loop of
thread that has been passed through the tissue. D, A row of chain stitches is formed in the
tissue by the machine.
Rows of chain stitches were formed using this machine at endoscopy in the postmortem human
esophagus and stomach (Figure 163). A modular design of an endoscopic sewing machine (Figure
164), having interchangeable cavity sizes to provide adjustment of suture length and depth, has been
used at endoscopy to form single stitches and chains of stitches at the cardioesophageal junction and
stomach in survival studies in dogs with and without laparotomy.
(590)Figure 163. A chain of stitches in a postmortem human stomach formed by a prototype

endoscopic sewing machine.
(591)Figure 164. Modular sewing machine. A modular design of the sewing machine with
hand controls and a two-channel endoscope.
Acute ulcers of 1 cm diameter (n = 10) were formed with a suction biopsy device (Quinton ulcer maker)
in canine mucosa and then oversewn using the endoscopic sewing machine. The machine was
positioned over the ulcer, and the entire ulcer was sucked into the cavity. Suction pressure exerted by
the machine was usually sufficient to terminate bleeding. The threaded needle was then passed from
one side of the ulcer through the muscularis propria or serosa to appear at the other side of the ulcer.
The loop of thread was then caught and the ulcer oversewn; bleeding was terminated by compressing
the vessels with a single stitch or a chain of stitches that was subsequently tied at endoscopy.
This method can be simplified to deliver one loop of thread that can be tied by means of half-hitches or
other knots formed externally to the patient.
Single-Stitch Hollow-Needle Sewing Machine

A different approach to forming stitches in tissue is possible using a hollow needle that is passed
through tissue. Part of a preformed stitch is forced through the needle to appear on the far side of the
tissue while another part is released on the proximal side (Figures 165 and 166).
(592)Figure 165. Mechanism of action of a single-stitch sewing machine. A, Cross-sectional

view of the machine positioned above the target tissue. One end of a preformed nylon stitch is
loaded into the hollow needle. B, Full-thickness double layer of tissue has been sucked into the
machine's cavity and pierced by the needle. C, The tilting arm of the preformed nylon stitch
has been forced out of the needle beyond the tissue, and the radiotelemetry capsule (RTC) is
being ejected from its cavity. D, Needle and pushing rod have been returned to their starting
positions, and the suction has been turned off to allow the tissue to slip out of the machine's
cavity, leaving the RTC secured to the nylon stitch sewn through a full-thickness double layer
of gastric tissue.
(593)Figure 166. Close-up preformed nylon stitches in postmortem human tissue.

The single-stitch hollow-needle sewing machine was designed to deliver a preformed nylon H-shaped
"tilt" stitch through a full-thickness double layer of gastrointestinal tissue (Figure 167; see also Figure
165).2(594) It is possible to attach objects to this stitch, leaving them secured to the wall of the gut.
The preformed stitch, attached for example to an RTC (Figure 168), is preloaded into the machine,
passing one end into the hollow needle and securing the stitch under a retainer that can release the
stitch in due course. A double layer of tissue is pierced by the hollow needle by pulling on a wire within
a wire-wound flexible cable. The leading part of the stitch is then forced by another wire through the
hollow needle by a pushing rod to a point beyond the fold of tissue where it adopts a T shape that
prevents it from coming back through the tissue. The movement of the pushing rod component of the
machine simultaneously releases the following part of the stitch on the near side of the tissue fold. The
needle and pushing rod are returned to their starting position, and the suction vacuum is released to
allow the tissue with the nylon stitch through it to slide out of the cavity of the machine. The RTC is
ejected from the body of the machine by pulling on a third wire, which forces it out.
(595)Figure 167. Single-stitch endoscopic sewing machine.
(596)Figure 168. RTC attached to a stomach by a preformed stitch.

The sewing machine can be front-loaded with its control cables passing through the accessory
channels of a conventional endoscope, or it can be guided using an endoscope positioned alongside
the control cables. To facilitate insertion and multiple stitch placements, the machine was designed to
be passed optionally through an esophageal overtube with an 11.5-mm internal diameter. One
thousand stitches were delivered in 25 postmortem human stomachs, 50 in 5 esophagi, 20 in 2 small
bowel specimens, and 20 in 2 colon specimens. The nylon stitch had a 6.6-lb breaking strain. Stitches
through muscularis propria required a 4.5-lb force to pull them out; stitches extending only through
muscularis mucosa required 2.1 lb of force.
Radiotelemetry
The single-stitch sewing machine was used to implant pH-sensitive RTCs on 14 occasions for up to 87
days in the stomachs of unrestrained beagle dogs.3(597) Continuous recordings were made in 48-hr
periods using a belt aerial and solid-state recorder worn in a jacket. Studies were carried out during
normal feeding and with 20 and 40 mg (2 to 4 mg/kg) omeprazole (OM) once daily. There were no
complications (hemorrhage or perforation) during the period of implantation, and follow-up

endoscopies disclosed no evidence of ulceration or focal gastritis. The median (range) intragastric
control pH (C) was 3.4 (2.8 to 7.2). Median pH during acid inhibition after 5 days dosing with 20 mg OM
was 4.2 (1.2 to 7.6), and 5.2 (2.0 to 7.6) after a similar period on 40 mg OM (p = .01 C vs. OM 20 and
OM 40, p = ns OM 20 vs. OM 40). After the drug was stopped, the median pH returned to 2.8 (1.0 to
6.8) on days 1 to 3 but fell to 1.2 (1.0 to 3.2) by days 5 to 7 (p < .001 C vs. period 5 to 7 days). This
technique, for the first time, has allowed monitoring of gastric pH for long periods of time without the
use of wires or tubes.
The single-stitch endoscopic sewing machine has been used to achieve continuous noninvasive
monitoring of gastric pH in humans.4(598) Using an endoscopic sewing machine, an RTC was sewn
into the stomach of one of its inventors with some help from his friends. An 11.5-mm outer diameter
endoscopic sewing machine was used to sew an RTC to the wall of the body of the stomach at 56 cm
from the teeth without laparotomy using a single transmural nylon stitch (Figure 169).
(599)Figure 169. Stitch in a human stomach.

The RTC stayed in position for 159 days and transmitted a radiofrequency signal indicating gastric pH
for the whole time. A total of 1200 hr of ambulatory recordings of gastric pH, that is, 50 24-hr periods,
were made using a body-borne aerial belt and portable pH recorder. A consecutive series of
experiments was carried out comparing the effect of regular antacids (Asilone; Maalox; Gaviscon, 40
ml qid with food; OM 20 mg; OM 40 mg; ranitidine [RAN] 300 mg nocte; RAN 300 mg twice a day), with
a control period of 1 week between each dose regimen. These studies were carried out without loss of
time at work; indeed, the subject was at work on an endoscopy list less than 1 hr after the RTC was
sewn into place. No discomfort was experienced; weight did not alter. No ulceration or inflammation
was observed at the site of the stitch at endoscopy on day 0 or day 14. Acid suppression was
expressed as the percent pH greater than 4.
These experiments showed that antacids altered gastric pH only marginally compared with control data
(control < pH 4: 50%, Gaviscon 51%, Maalox 66%, Asilone 49%). OM 40 mg and RAN 600 mg caused
significantly more acid inhibition (OM 40 mg < pH 4: 94%; RAN 600 mg 100%) than OM 20 mg and
RAN 300 mg (OM 20 mg 79%, RAN 300 mg 73%). Stressful events were not associated with alteration
of baseline pH in this study. This is by far the longest period of continuous monitoring of gastric pH in
humans, and this technique is far more comfortable than measurement by means of a nasogastric pH
probe. Long-term monitoring of gastrointestinal function by means of biosensors implanted at flexible
endoscopy is possible in humans.
Stapling Machine
This instrument is the first flexible, remotely controlled device designed to place multiple transmural
staples in double rows through gastrointestinal tissue without requiring laparotomy.5(600)
The stapler is mounted in front of a conventional endoscope, with a control cable passing through the
accessory channel of the endoscope. The dimensions of the stapler were 16 mm outer diameter and 8
cm long. The staples were made of stainless steel or titanium wire (0.25 mm) and measured 3.5 3.5
mm.
The stapler works in the following sequence (Figure 1610): Gastrointestinal tissue to be stapled is
sucked into a cavity within the stapling device to form a double layer (Figure 1610A). The dimensions
of the cavity determine the depth of staple penetration, and the device was designed to deliver staples
that penetrate mucosa and muscularis propria to serosa. The mechanism of action requires three
movements of parts within the stapling device. The tissue must be compressed after suction into an
optimum thickness for staple closure (Figure 1610B). The second movement (Figure 1610C)
pushes the staples through their location slots by means of staple rammers, so that the staples
penetrate the double layer of tissue to engage shaped anvils that cause them to bend inward and
backward on themselves (into a B conformation) to exert compressive force on the tissue once they
have pierced it (producing an audible "clunk"). The considerable force of 12 to 15 lb required to close
these staples through tissue is exerted by means of strong steel wires inside a wire-wound flexible
(Bowden) cable similar to a bicycle brake cable. The third action (Figure 1610D) is to release the
tension transmitted to the staples through these wires to allow a release spring within the body of the
stapling device to open the cavity and permit the stapled tissue to slide out of the device.
(601)Figure 1610. Mechanism of action of the endoscopic stapling device. A, Fold of tissue
is sucked into the cavity of the stapler. B, The tissue is compressed to an optimum thickness
prior to stapling. C, The staple rammers force the staples through the tissue onto shaped anvils
that bend the staples into a B configuration. D, Spring opens the cavity to allow the stapled
tissue to slide out of the stapling device. Note: This diagram has been simplified to show the
closure of only two staples. In the device, the staples were arranged in an overlapping double
layer forming a semicircle 1 mm from the outer circumference of the stapler.
Two 1-cm-wide strips of postmortem human tissue were stapled together, and the force required to
tear the anastomosis apart was measured. Strips joined with five staples could support weights of up
to 2 lb.
The stapling machine (Figures 1611 and 1612) was used in the postmortem human stomach (35
stomachs, 120 stapling experiments), esophagus (20 esophagi, 40 stapling experiments), and colon
(15 colons, 30 stapling experiments). The force required to close rows of staples varied with the organ
being stapled. An average of 2 lb more force was needed to close rows of staples in the esophagus
than in the stomach. The device was also used to seal experimental 1-cm full-thickness perforating
wounds in postmortem human gastrointestinal tissue (10 experiments).
(602)Figure 1611. Endoscopic stapling machine in opened position.
(603)Figure 1612. Endoscopic stapling machine in closed position.

The stapler was used at endoscopy, with and without laparotomy, to deliver rows of staples in survival
studies in the canine esophagus (12 experiments) and porcine stomach (4 experiments). It was used
at laparotomy to terminate bleeding from standard experimental bleeding gastric ulcers in dogs (10
experiments) and to close 1-cm full-thickness perforating scalpel incisions (4 experiments) in
uncomplicated survival experiments. The stapler was used to oversew experimental bleeding ulcers
and to close small, full-thickness incisions in the canine stomach.
The stapling device is similar in some respects to the endoscopic sewing machine and also to existing
surgical anastomotic stapling guns.611(604) However, it is the first flexible, endoscopically controlled
device for remote placement of multiple staples through gastrointestinal tissue that does not require
laparotomy. Potential applications of such a device include the stapling of bleeding varices (the present
model has been specifically designed for this application and offers a prospect of improved
hemostasis), closure of internal wounds or fistulas, assistance in the removal of normal or abnormal
tissue, safe procurement of full-thickness biopsies of gastrointestinal tissue, and stapling across
bleeding or perforating ulcers. Modifications of the device could be used to assist in the formation of
anastomoses and for antireflux procedures and surgery for the control of obesity.
Other Mechanical Devices for Attaching Material to Tissue at Endoscopy

A number of related mechanical methods for attaching things to tissue have also been devised. An
elegant clip-fixing device that can be passed through the accessory channel of an endoscope was
developed in 1975 by Hayashi et al.12(605) This has been used to treat bleeding ulcers, to mark the
site of insertion of a colonoscope, and to secure pH electrodes in the stomach. The Sakura J clip,
designed in the mid-1970s by Hayashi et al.12,13(606) and modified by Hachisu,14(607) can now be
passed through the accessory channel (2.8 mm diameter) of a conventional endoscope. This clip is
opened and then closed (Figure 1613) tight on the bleeding point by pulling the proximal loop of a
compressible metal clip into a metal ring, so that the clip closes in the shape of the number 8 to
compress a bleeding vessel.13(608) This device was reported as effective in controlling bleeding from
22 of 24 gastric ulcers and 3 of 4 duodenal ulcers. Dislodgment was relatively frequent (28%), and
further bleeding was noted in 23%.
(609)Figure 1613. Clip-fixing device. The clips are shown open and closed on tissue.
The invention of a rubber-band ligation device (Figures 1614 and 1615) for use at flexible
endoscopy, similar to that for treatment of hemorrhoids, was first reported in 198515(610) and has
subsequently been applied by Stiegmann et al. in the treatment of esophageal variceal bleeding (see
Chapter 36: Endoscopic Elastic Band Ligation for Bleeding Esophageal Varices).16,17(611)
(612)Figure 1614. Rubber-band ligation device.
(613)Figure 1615. Rubber-band placed on tissue.

An ulcer clamp (Figure 1616), which sucks tissue into a cavity to which compression can be exerted
by means of a spring, was developed and reported in 1985.15(614)
(615)Figure 1616. Ulcer clamp.

An instrument that exerts sufficient rotational torque to screw a corkscrew-like device into tissue at high
speed (600,000 rpm) was developed and has been used to treat bleeding gastric ulcers in two clinical
cases.18(616)
A suturing device for use with rigid endoscopes has been reported by Da Silva.19(617) A long plastic
needle is restrained within a curved track that allows it to be passed through tissue at the end of the
rigid open tube and the thread retrieved once it has passed through the tissue.
Anastomosis at Flexible Endoscopy

In 1892, Murphy20(618) popularized anastomosis at laparotomy by a compression button method in
which circular gastrointestinal anastomoses were formed by ischemic compression of tissue between
two buttons held together by a spring. This technique has been modified to allow the formation of
anastomoses at flexible endoscopy.
The compression buttons used were circular with a raised rim designed to introduce a thin circle of
ischemia in the intervening tissue when the two buttons were pressed together. Both buttons had a
central locating hole through which a threaded needle could be passed, and one (the female part) had
a locking mechanism composed of a sprung wire that could catch a serrated hollow locating pin
passed through the other (the male part). When locked, the two buttons were pressed together by a
compression spring positioned over the locating pin (Figures 1617 and 1618).
(619)Figure 1617. Compression buttons open.
(620)Figure 1618. Compression buttons closed.

Gastrojejunostomies were formed at flexible endoscopy in six adult greyhounds in the following
fashion.21(621) Through a 2-cm minilaparotomy incision, a suitable loop of jejunum adjacent to the
stomach was identified. One flexible endoscope was passed by mouth into the stomach. Meanwhile, a
small incision was made in the jejunum to allow a second flexible endoscope to be passed a few
centimeters into the jejunum. By maneuvering the tip of this second endoscope, the jejunal wall could
be brought into close proximity to the stomach. Transillumination and indentation of the stomach wall
by the jejunal endoscope could be observed (Figure 1619). A needle with thread attached was
pushed through the jejunal and stomach walls and caught by a wire loop snare. Both endoscopes were
removed, leaving the thread passing through the abdominal incision into the jejunum, then through into
the stomach to emerge through the mouth. The male button was passed over the thread into the
jejunum. A knot was tied on the abdominal side, so that the force exerted on the thread at the mouth
would pull the jejunal male button in a cephalad direction. The second, female button was passed over
the thread and pushed into the stomach by the gastroscope. When the gastric button was in apposition
with the gastric wall, the thread was pulled so as to force the serrated pin of the male button through
the jejunal and gastric walls and into the female button. The two buttons were thus locked together
under compression and the intervening tissues squeezed between the raised circumferential rims. The
endoscope in the stomach was held steady, so that its tip could act as a stop to hold the female button
in place as the central serrated pin on the male button was pulled by means of the thread passing
through the accessory channel. Once the buttons were locked into position, a white rim indicating
ischemia could be seen around the edge of the buttons and bile-stained jejunal juice refluxed through
the hollow central pin (Figure 1619). The endoscopes were withdrawn, and the thread was removed
by pulling it out of the jejunal and abdominal incisions, which were then closed.
(622)Figure 1619. Compression buttons at endoscopy. A rim of ischemic necrosis can be

seen at the edge of the button. A drop of bile from the jejunum has run through the needle into
the stomach.
Using the preceding technique, eight gastrojejunostomies were formed in six animals without
complications (in two experiments adjacent double anastomoses were formed). The buttons left the
stomach and emerged in the stool at 48 to 96 hr after placement. Endoscopies at 7 days demonstrated
that the gastroenterostomies could be crossed by a 9-mm-diameter endoscope. The anastomoses
were still present and patent 9 months later (Figure 1620).
(623)Figure 1620. Formed gastrojejunal anastomosis at 9 months.

Summary
A need remains for further development and application of technology for sewing, stapling, and forming
anastomoses. Knot tying at rigid and flexible endoscopy remains cumbersome although
feasible.22(624) Despite these difficulties, further technical development, together with the use of the
techniques and instruments described above in combination, has made it possible to perform relatively
complex operations without resorting to laparotomy.23(625) Moreover, continued technical
development will lead to simplified devices with potential clinical utility.24(626) Considerable scope
exists for joint endoscopic and laparoscopic projects involving sewing and stapling, and it is a pity that
so little crossover of technology occurs between the domains of rigid and flexible endoscopy.
Surgery has three basic technical aspects: cutting, controlling bleeding, and joining living tissue with
materials or other living tissue. Striking advances have been made in the clinical application of
endoscopic surgical techniques for cutting tissue and controlling bleeding. If the problems of sewing
and joining tissue and material at flexible endoscopy can be overcome, it is possible to consider the
advent of a new gastrointestinal surgery in which laparotomy is rarely required and the transabdominal
and transthoracic insertion of rigid instruments is much reduced.
REFERENCES
1.
Swain CP, Mills TN. An endoscopic sewing machine. Gastrointest Endosc 1986;32:367.
Swain CP, Brown G, Van Someren N, et al. Design and testing of a single stitch endoscopic
sewing machine. Gut 1989;30:A15145.
3.
Evans DF, Brown G, Kalabakas A, et al. Acid breakthrough with high dose omeprazole
revealed by a new non-invasive method of continuous pH monitoring from implanted
telemetry capsules. Gastroenterology 1991;100:A62.
4.
Swain CP, Evans DF, Glynn M, et al. Endoscopic sewing machine used to achieve continuous
non-invasive monitoring of gastric pH for three months in man. Gastrointest Endosc
1992;38:278.
5. Swain CP, Brown GJ, Mills TN. An endoscopic stapling device: The development of a new
flexible endoscopically controlled device for placing multiple transmural staples in
gastrointestinal tissue. Gastrointest Endosc 1989;35:3389.
6. Rohscek F. The birth of the surgical stapler. Surg Gynecol Obstet 1980;150:57982.
7.
Fraser I. An historical perspective on mechanical aids in intestinal anastomosis. The
surgeon's library. Surg Gynecol Obstet 1982;155:55674.
8.
Ravisch MM, Steichen FM. Technics of staple suture in the gastrointestinal tract. Ann Surg
1972;1975:81537.
9. Goliger JC. Recent trends in the practice of sphincter-saving excision for rectal cancer. Ann R
Coll Surg Engl 1979;61:16976.
10.
Ravisch MM, Steichen FM. Contemporary stapling instruments and basic mechanical suture
techniques. Surgical stapling techniques. Surg Clin North Am 1984;634:42540.
11. Waxman BP. Large bowel anastomosis II. The surgical staplers. Br J Surg 1983;70:647.
12.
Hayashi T, Yonezawa M, Kuwabara T, et al. The study on staunch clips for treatment by
endoscopy. Gastroenterol Endosc 1975;17:92101.
13.
Hayashi J. The study of endoscopic staunching clips for bleeding of the exposed vessel end
of the ulcer. In Okabe M, Honda T, Oshiba S (eds). Endoscopic Surgery. New York: Elsevier,
1984;11724.
14. Hachisu T. Evaluation of endoscopic hemostasis using an improved clipping apparatus. Surg
Endosc 1988;2:137.
15.
Swain CP, Mills TN, Northfield TC. Experimental studies of new mechanical methods of
endoscopic haemostasis: Stitching, banding, clamping and ulcer removal. Gut
1985;26:A1151.
16. Stiegmann GV, Cambre T, Sun JH. A new endoscopic elastic band ligating device.
2.

17. Stiegman GV, Goff JS. Endoscopic esophageal varix ligation: Preliminary clinical experience.
18. Escourrou J, Delvaux M, Buscail L, et al. First clinical evaluation and experimental study of a
new mechanical suture device for endoscopic hemostasis. Gastrointest Endosc
1990;36:4947.
19. Da Silva EG. Suturoscope: A new device that allows endoscopic sutures to be performed with
traditional threads. Surg Endosc 1990;4:2203.
20.
Murphy JB. Cholecystointestinal, gastrointestinal, enterointestinal anastomoses and
approximation without suture (original research). Med Rec NY 1892;42:66576.
21. Swain CP, Mills TN. Anastomosis at flexible endoscopy: An experimental study of
compression button gastrojejunostomy. Gastrointest Endosc 1991;37:62831.
22. Swain CP, Kadirkamanathan SS, Gong F, et al. Knot tying at flexible endoscopy. Gastrointest
Endosc 1994;40:7229.
23. Kadirkamanathan SS, Evans DF, Gong F, et al. Antireflux operations at flexible endoscopy
using endoluminal stitching techniques: an experimental study. Gastrointest Endosc
1996;44:13343.
24. Swain CP, Brown GJ, Gong F, Mills TN. An endoscopically deliverable tissue-transfixing
device for securing biosensors in the gastrointestinal tract. Gastrointest Endosc
1994;40:7304.
Chapter 17 Biomedical Tissue Spectroscopy

(627)
(628)
JOSEPH A. IZATT, PH.D.
The visual sensation of colors of light and their variations have long played a fundamental role in the
interaction of humans with nature. Our most highly developed special sense, vision, is a sophisticated
and versatile tool for analyzing the spatial and spectral distributions of light in our environment. Optical
spectroscopy, the quantitative study of the colors of light, has stimulated many of the great discoveries
in the history of science. Isaac Newton first stipulated that colors of light correspond to frequencies of
vibration of waves, and he demonstrated color separation with prisms.1(629) Late in the 19th century,
when it was determined that light waves correspond to vibrations of coupled electric and magnetic
fields, it became apparent that light spectra could be used to probe the elementary structure of matter.
This is because electromagnetic waves interact with atoms and molecules at characteristic frequencies
(i.e., colors) that are highly specific to their internal composition and microenvironment. Early in the
20th century, observations of optical spectra of simple atoms and molecules demonstrated that the
frequencies of light that are absorbed and emitted are quantized; that is, they occur only at discrete
energy intervals. These observations led directly to the development of quantum theory, the theoretical
tool needed to understand and predict atomic and molecular spectra. Many other developments in
science and technology in the 20th century can be traced to the study or use of optical spectra. Optical
and radio spectroscopy of radiation from the heavens provides the only available direct experimental
evidence concerning the origins and history of the universe.
Various forms of optical spectroscopy have been applied in the identification and monitoring of
physical, chemical, and biologic processes. Qualitative optical spectroscopy has also played a
fundamental role in medicine. From earliest to present practitioners, visual assessment of color, hue,
and pallor has been an essential element of physical diagnosis. These indicators have only increased
in value with the development of modern, minimally invasive diagnostic technology, which allows for
visual assessment of tissue surfaces in previously inaccessible regions of the body. Despite the
availability of these imaging modalities of unparalleled sophistication, diagnosis is still based largely on
qualitative visual observations. Unfortunately, the highly sensitive and specific optical biochemical
assays developed in other disciplines using previously available technologies have not been directly
transferable to diagnostic applications in humans. In any case, these are insufficiently sophisticated in
relation to the complexity of clinical diagnosis.
Rapid advances in optical and photonic technologies over recent decades have led to the development
of new techniques that make possible the application of quantitative optical spectroscopy to the early
detection and staging of human disease. The fundamental advance that stimulated the renaissance of
optics in science, technology, industry, and now medicine was the invention of the laser (see Chapter
10: Laser Physics and Laser-Tissue Interactions). When used in conjunction with low-loss flexible
optical fibers, lasers make it possible to apply quantitative spectroscopic techniques within the human
body.
Two additional technologies developed in the last decade are of critical importance to making real-time
quantitative spectroscopic diagnosis a reality: (1) highly sensitive, large-area multipixel imaging
detectors (e.g., charge-coupled devices [CCDs]) that enable collection of quantitative spectra through
optical fibers in near real time; and (2) fast, inexpensive computers that can rapidly collect, process,
and display imaging and spectral information.
Current research also holds promise in the near future of capabilities even greater than those provided
by imaging detectors and computers. Ultrafast optical technologies, for example, can potentially
combine spectroscopic characterization of disease with high-resolution, three-dimensional imaging of
tissue microstructure.2,3(630) Advances in the understanding of human visual processing may lead to
minimally invasive technologies for visualization of and interaction with unique three-dimensional
environments.4(631)
Light-Tissue Interactions
The interaction of light with tissue is a special case of the interaction of light with matter. This
interaction is a particularly difficult system to describe in detail, as it involves a complex interplay
among many component processes. Some of these processes act to change the frequency distribution
of light, whereas others act to redistribute light spatially once it has entered the tissue.57(632) Two
other intensively studied physical systems that involve simultaneous spatial and spectral redistribution
of radiation are the scattering of light in planetary atmospheres8(633) and the scattering of neutrons in
nuclear reactors.9(634) Many of the mathematical foundations underlying current studies of light-tissue
interactions had their origins in the study of these phenomena.7(635)
Although the light-tissue interaction process as a whole is complex, its components are well
understood (Figure 171). At its basis, the interaction of light with matter depends on the transfer of
energy between particles of light (photons) and particles of matter. Particles of matter include
submicroscopic components such as electrons, atoms, and molecules, as well as larger (microscopic)
particles such as cells, granules, and organelles. Although the latter may scatter photons (described
later), a quantum description is necessary only to describe interactions that occur between photons
and submicroscopic particles.
(636)Figure 171. Spectroscopic light-tissue interactions of diagnostic significance.

Molecules are the submicroscopic particles that have the richest spectral content and thus are the
primary concern of this discussion. Both photons and molecules exhibit a strong quantum nature,
having characteristics of both particles and waves. Light photons carry energy proportional to their
frequency, according to Planck's Law (see Chapter 10: Laser Physics and Laser-Tissue Interactions).
The wavelength of light () is inversely proportional to its frequency () according to the relation = c/,
where c is the velocity of light.
According to quantum mechanics, only discrete amounts of energy may be stored in molecules. This
quantization applies to the electron orbitals, which may be occupied in the molecule's constituent
atoms, as well as to vibrations between constituent atoms, rotations of parts of the molecule with
respect to other parts, and rotations of the whole molecule. Vibrational and rotational energies are
typically small compared with electron energy levels (Figure 172). A photon can interact with a
molecule only if the photon energy corresponds to an available energy level difference in the molecule.
Similarly, if a molecule occupies an energy level above that of its ground state (having been excited, for
example, by a previous photon), it may relax to a lower state, accompanied by emission of a photon
with a frequency corresponding to the energy level difference. Molecules also interact with their
environment in other ways. They can be excited by other energy sources such as chemical reactions,
heat, or sound, and they may relax from excited states by emission of any of these or other forms of
energy.
(637)Figure 172. Schematic diagram of the energy level of a molecule. Solid horizontal
lines represent atomic energy levels that are surrounded by closely spaced
vibrational-rotational energy levels (horizontal dashed lines) with lower energy spacing. The
horizontal dotted line represents virtual energy levels, which only occur in the presence of
high-intensity light. Horizontal sinusoidal lines represent photons with energy hf, where f
represents the type of interaction. hPlanck's constant; frequency. Vertical solid lines
represent energy level transitions resulting in the absorption or emission of photons, and the
diagonal dashed line represents energy level transitions by other means, such as simple
relaxation.
Several light-tissue interaction processes have significance for optical diagnosis (Figure 173; see also
Figure 171). These can be described by several fundamental properties. Interactions can be elastic or
inelastic. In elastic scattering processes, the frequency of the incident radiation and the energy state of
the molecule remain largely unchanged, but the direction of light propagation may be altered. Two
important categories of elastic scattering include Rayleigh and Mie scattering.7,10(638) Rayleigh
scattering refers to scattering of light by particles much smaller than a wavelength, with the result that
(unpolarized) incident light is scattered essentially uniformly in all directions. Mie scattering theory
describes the scattering of light by particles of a size comparable to or larger than a wavelength; its
theoretical description is much more complicated. In Mie scattering, most photons are scattered into a
relatively small cone in the same direction as that of the incident photon; however, some photons are
scattered at high angles to that direction with a distribution that is highly sensitive to the exact size and
shape of the particle. Thus, small shifts in the wavelength of the incident light can lead to large
changes in the scattered light distribution for Mie-sized particles. Most (>90%) of the scattering of
visible wavelengths of light by tissue falls into the Mie scattering category.11(639)
(640)Figure 173. Schematic diagram of several possible light-tissue interactions processes.

In inelastic scattering processes, part or all of the energy carried by a photon may be absorbed by the
particle. The absorbed energy may be dissipated to the local surroundings in the form of heat, or
reemitted as another photon, generally with a lower frequency. Cases of both elastic and inelastic
scattering are frequency dependent; thus, by monitoring the frequencies that are preferentially
absorbed and scattered, as well as monitoring the emission of light at secondary frequencies,
information may be obtained about the size distribution and composition of tissue particles.
Figures 172 and 173 include energy level diagrams as a shorthand notation for keeping track of the
transfer of energy that characterizes light-matter interactions. These can be interpreted as progressing
from left to right; a photon approaches an atom from the left and interacts with it, and another photon
may be emitted and exit to the right. The time dependence of an interaction is typically
noninstantaneous only when an atom occupies an intermediary excited state for some period of time
before emitting the final photon. The quantum yield of a specific interaction is defined here as that
fraction of the photons involved in interactions that are products of that specific interaction, compared
with the total number of photons incident on the medium. Elastic scattering (see Figure 171) is by far
the most likely light-tissue interaction in the visible and near-infrared wavelength regions of the
spectrum. This property of tissue, which accounts for its translucence, implies that each photon is likely
to be scattered many times (hundreds or thousands) before and between inelastic collisions. This
property also makes it particularly challenging to identify the spatial origins of photons reemitted from
tissues on a microscopic scale. Biologic specificity is a qualitative measure of how distinctive a given
type of interaction is in relation to biologic activity.
Optical Spectroscopy in Biomedicine

The earlier technical introduction to the physics of tissue spectroscopy forms the basis for a discussion
of specific approaches to spectroscopic characterization of normal and diseased tissue. The following
represents the current state of research in this field.
Absorption Spectroscopy
The most common application by far of spectroscopy in biomedicine is the routine analytical
characterization of biochemical compounds by spectrophotometry. Absorption spectra of compounds
extracted from human tissues can serve as indicators of metabolic state or the presence of
disease.1214(641) Absorption spectroscopy is performed by monitoring the attenuation of light in
samples of known thickness in order to detect and quantify the concentrations of specific molecules
with known quantum transition energy levels. These transition energies vary with the state and
microenvironment of the molecule (e.g., pH or oxidation-reduction state). It is well established that
hemoglobin, for example, exhibits distinct shifts in absorption spectra between its oxygenated and its
deoxygenated states (Figure 174).15,16(642) Deoxyhemoglobin absorbs strongly in the
red-wavelength region, thus causing deoxygenated blood to appear blue, whereas oxyhemoglobin
absorbs very little in the red region and thus causes oxygenated blood to reflect red colors and appear
bright red.
(643)Figure 174. Optical absorption spectra of oxyhemoglobin (HbO2) and

deoxyhemoglobin (Hb) in the visible and near-infrared wavelength ranges. (From Szaflarski
NL, Cohen NH. Use of pulse oximetry in critically ill adults. Heart Lung 1989; 18:44455.)
Although absorption spectroscopy has a long, rich history as the mainstay of analytical and clinical
chemistry, it is very difficult to apply quantitatively in highly scattering biologic tissues, primarily for two
reasons: (1) Native tissue contains such an abundance of biochemical compounds, each having
distinct but overlapping visible absorption spectra, that it is virtually impossible to separate the
contributions of the individual component molecules of interest. This is especially true in the lower
wavelength regions of the visible spectrum, where most biomolecules absorb and fluoresce.17(644) (2)
Even when abundant molecules with characteristic spectra in the infrared region (e.g., hemoglobin,
porphyrin compounds) can be identified, the prevalence of scattering in the tissue makes it virtually
impossible to accurately determine the path length of a light source shining through the tissue. For this
reason, quantitative determinations of molecular concentration cannot be performed.
Although quantitative measurements of chromophore concentration are very difficult to achieve using
absorbance measurements, fractional concentrations may still be measured using ratio
measurements. An example of this technique is pulse oximetry, in which the oxygen saturation, which
is the relative concentration of oxyhemoglobin to the total hemoglobin concentration, is
measured.18,19(645) This is accomplished by measuring absorbance at two separate wavelengths in

the red and infrared regions of the spectrum (see Figure 174). Measurements are performed at
systole and diastole in rapid succession to separate out absorbance due to arterial blood only in the
finger tip or earlobe. Another use of absorption spectroscopy in diagnosis lies in the administration of
drugs with an increased affinity for diseased tissues, unique spectral signatures, and sufficient tissue
concentrations to overwhelm other tissue chromophores within their characteristic wavelength regions.
Porphyrin compounds, which preferentially concentrate in neoplastic tissues, are under investigation as
potential optical markers for disease.14,2027(646)
Despite its limitations for obtaining definitive diagnostic information, absorption does play an important
role in other forms of spectroscopy, including reflectance and fluorescence, by spectrally filtering
light.28,29(647) Because absorption is an established standard for biochemical measurements,
considerable effort has been devoted to the development of sophisticated techniques for extraction of
quantitative absorbance information from other optical tissue-probing techniques. Some of the
experimental techniques used to discriminate tissue optical properties, including absorption and
scattering coefficients, are attenuated total reflectance spectroscopy,30(648) coherent backscatter
spectroscopy,31(649) low-coherence interferometry,32(650) and various time-resolved3335(651) and
frequency-domain36(652) methods. Analytical and numerical theoretical techniques have also been
used in attempts to obtain tissue optical properties from reflectance spectra.11,37,38(653)
Reflectance Spectroscopy
Reflectance spectroscopy is the analysis of the intensity of total light reemitted from tissue surfaces as
a function of wavelength. Because the detected light contains contributions from all possible
light-tissue interactions shown in Figure 171, quantitative extraction of specific molecular
concentrations is problematic with reflectance spectroscopy. Nonetheless, exploratory studies of
reflectance spectroscopy in tissue have been performed. Because the laser light-tissue interactions
with the highest quantum yields are elastic scattering and absorption, reflectance spectroscopy reveals
information primarily about these interactions. Absorption within tissue leads to negative peaks, or
valleys, in the reflectance spectrum.3941(654) Scattering, on the other hand, reveals information
primarily about the distribution of scatters of a size comparable to the wavelength of light being
used.10(655) As noted above, the main elastic scattering interaction in biologic tissue is Mie scattering,
in which the distribution of scattered light is expected to vary rapidly with the size of the scattering
particles. These variations are most dramatic when the particles are approximately the same size as or
larger than the light wavelength being used. Thus, reflectance spectroscopy reveals potentially
important information concerning the microscopic architecture of tissues on the scale of individual cells
and cellular components.42(656)
Research using reflectance spectroscopy has included studies of bladder,28,43(657) gastrointestinal
(GI) tract,44(658) and skin cancers.4548(659) Differences in reflectance spectra from cancerous
versus normal tissues are described in these studies. These differences are attributed preliminarily to
the sensitivity of reflectance spectroscopy to microscopic tissue changes associated with the onset of
cancer, such as increased nuclear-cytoplasmic ratio. However, data reported to date are insufficient to
convincingly demonstrate efficacy.
Fluorescence Spectroscopy
Native Fluorescence Spectroscopy
Laser-induced fluorescence (LIF) spectroscopy refers to a sequence of physical processes that
represents one of the most promising forms of tissue spectroscopy under investigation.17(660) The
first step in LIF is the absorption by a target group of chromophores of a particular wavelength of light.
In order to stimulate a sufficient number of molecules so that the results can be easily detected, light is
usually delivered by means of a laser. Typically, the incident light selected is from the energetic
ultraviolet region of the spectrum. Thus, the chromophore molecules are excited to one of their higher
atomic energy levels. The excited molecule has both radiative and nonradiative mechanisms available
for deexcitation; the probability of the occurrence of each mechanism depends on the detailed
structure of the molecule and its microenvironment. The excited molecule typically starts to deexcite in
a cascade process, falling down an energy-level ladder until it reaches a state from which further
deexcitation by thermal means is less likely than radiative deexcitation. If conditions are favorable, the
remaining excitation energy may be lost through radiation of photons with a lower energy (higher
wavelength) than the exciting photons.
Unlike elastic scattering processes, the fluorescence process is not instantaneous because the
molecule may go through many intermediate steps between excitation and reemission. The process
typically requires approximately 10-10/10-6 sec. In some cases the energy can be transferred to
another portion of the molecule or even to a separate molecule, in which case the process is termed
phosphorescence, and deexcitation times can be as long as several seconds.49(661)
The fluorescence and phosphorescence processes are more specific to tissue biochemistry than is
reflectance spectroscopy in the following ways: (1) An excitation wavelength can be chosen to target
chromophores selectively. By means of an appropriate excitation wavelength, molecules that have
different fluorescence properties under the metabolic conditions being probed or different properties in
the presence of biochemical environments specific to a particular disease can be selectively
excited.50,51(662) (2) The amount and frequency distribution of incident photon energy reemitted as
fluorescence are intricately related to the state and environment of the probe molecule because these
factors determine the relative probabilities of radiative versus nonradiative deexcitation. (3) The
fluorescent light is easily separated from the incident and elastically scattered light because it has a
separate wavelength. Thus, fluorescence-sensing systems can be constructed with relatively simple
laser sources and optical means for analysis of the spectrum of the reemitted light. These systems can
be used to examine only that portion of the incident light that has interacted with the tissue in a highly
selective manner.
Cellular Basis of Tissue Fluorescence
Biologists have long used fluorescence to understand the structure and dynamics of large organic
molecules. Chromophores that fluoresce primarily in the ultraviolet and visible regions of the spectrum
have been observed among molecular systems, including proteins, lipids, nucleic acids, and free
nucleotide coenzymes (Table 171).5052(663) These fluorophores display specific spectral
alterations in response to changes in their local microenvironment, including pH, redox potential,
bonding sites, polarity, and ion concentrations.17(664)
TABLE 171
Assignment of Autofluorescent Tissue Chromophores
Potential
Flurophore
Tryptophan
Pyridoxol 5-phosphate (in
vitamin B6)
4-Pyridoxic acid (in vitamin
B6)
Collagen (Types I,III)
Excitation maximum
(NM)
Emission maximum
(NM)
Observed in Tissue Type

(Reference)
290
305
330
385
Colon [63, 74]

Colon [63]
315
430
Colon [63]
330
390
Colon [63, 74, 126]

Artery [73]
Cervix [136]
TABLE 171
Assignment of Autofluorescent Tissue Chromophores
Potential
Flurophore
Excitation maximum
(NM)
Emission maximum
(NM)
Observed in Tissue Type

(Reference)
Nicotinamide adenine
dinucleotide (NADH),
Nicotinamide adenine
dinucleotide phosphate
(NADPH)
Elastin
3409
460
Colon [63, 74, 126]

Cervix [136]
370
415
Lipapigments
Endogenous porphyrins
370
390
510
630, 680
Pyridoxol 5-phosphate (in

vitamin B6)
Flaven adenine dinucleotide
(FAD)
410
520
Artery [73]
Cervix [136]
Colon [74]
Colon [74, 126]
Esophagus [130]
Lung [80, 82]
Breast [80, 82]
Kidney (rat) [79]
Prostate (rat) [79]
Bladder (rat, mouse) [79]
Colon [63]
460
520
Colon [74, 126]

Lung [80, 82, 86]
Breast [80, 82]
Esophagus [130]
Kidney (rat) [79]
Prostate (rat) [79]
Bladder (rat, mouse) [79]
Cervix [136]
It was established in the 1950s and 1960s that the line shapes of the brightest fluorescence observed
in intact cells correspond closely to those of the aromatic amino acids tryptophan and tyrosine.52(665)
Pure tryptophan has an excitation peak near 290 nm and an emission peak near 330 nm when in
aqueous solution.53(666) This emission from tryptophan overwhelms all other sources of tissue
fluorescence when deep ultraviolet (wavelengths less than 325 nm) excitation is used. To observe
information-bearing fluorescence emission from other cellular molecules, longer excitation
wavelengths are usually required.
Early microfluorometric studies demonstrated that, excepting the amino acid contribution, the second
most intense cellular fluorescence arose within mitochondria, where fluorescent emission was
attributed to proteins bound to respiratory enzymes. Using microfluorometric techniques, fluorescence
changes were found to correspond to oxidation-reduction changes within living
mitochondria.5456(667) Potential molecular chromophores responsible for this fluorescence include
nicotinamide adenine dinucleotide (NADH) and nicotinamide adenine dinucleotide phosphate
(NADPH), both of which function as coenzymes in oxidation-reduction reactions and both of which
have broad emission maxima in the 460-nm region of the spectrum when excited in the 340-nm
region.57,58(668) Flavins, such as riboflavin, fluoresce in the 520-nm range in the green region of the
spectrum when excited near 460 nm, with the specific wavelength peak depending on the
oxidation-reduction state of the molecule's environment.5962(669) Riboflavin is part of the coenzyme
flavin adenine dinucleotide (FAD), which is responsible for oxidation-reduction reactions in the
mitochondria.
Other natural intracellular chromophores that fluoresce in the visible region have also been identified.
Several peaks in tissue excitation and emission spectra have been correlated with molecular
chromophores related to vitamin B6, including pyridoxal 5-phosphate.63,64(670) Porphyrin
compounds, which occur naturally in hemoglobin, urine, stool, and erythrocytes, fluoresce brightly in
the red region of the spectrum between 590 and 680 nm.50(671) The earliest studies that suggested
the use of fluorescence spectroscopy to localize disease, conducted in the first half of the 20th century,
correctly attributed the red fluorescence visually observed in human and animal tumor tissues to
natural and injected porphyrin-containing compounds.25,6567(672)
For the purpose of detection of early-stage cancer with fluorescence spectroscopy, it is encouraging to
note that several of the tentatively identified intracellular fluorescent compounds, as mentioned earlier,
have been associated with dysplasia and cancer in independent studies. Intracellular concentrations of
NAD+ and NADH have been observed to decrease dramatically following neoplastic conversion in
cultured fibroblasts.57(673) Decreased serum levels of pyridoxal 5-phosphate have been observed in
patients with cancer.68(674) Increased concentrations of endogenous porphyrins have been found in
numerous neoplasms of various organ systems.14,22,2427,66,67,69(675) This increase has been
attributed to either an altered metabolism of cancerous tissue70(676) or a microbial synthesis.26(677)
Extracellular Basis of Tissue Fluorescence
Many early studies of cellular fluorescence (see Table 171) for the purpose of understanding cellular
metabolism and biochemistry did not account for the potential contributions of extracellular substances.
Recent spectroscopic and microspectrofluorescent studies of bulk tissues have begun to elucidate the
important fluorescent role of extracellular material, particularly connective tissues. Studies of
fluorescence in bulk samples of coronary arteries have demonstrated that the observed line shapes
correlated well with fluorescence line shapes observed in purified extracts of the connective tissue
components collagen and elastin when the composite spectra are corrected for oxyhemoglobin
absorption artifacts.40,7173(678) Intense collagen and elastin fluorescence has also been observed
directly in microspectrofluorometric studies of both arteries and the GI wall.71,74,75(679) Bright
fluorescence in the green-yellow region from localized granular structures of subcellular dimensions
has also been described and attributed variously to eosinophils,75,76(680) lipid
microdroplets,74,77(681) and lipopigment-loaded lysosomes.74,78(682) Fluorescence in the green
visible region (525 to 540 nm) attributed to keratin and melanin has also been observed.7981(683)
In Vitro Studies of Tissue Fluorescence
The first efforts to monitor physiologic processes and diagnose disease by means of quantitative
measurements of autofluorescence spectra were undertaken by several pioneering groups in the
1980s. In these preliminary studies, changes in fluorescence line shapes were observed in vitro in
cancerous tissue from several human and animal organ systems, including skin, kidney, bladder,
prostate, breast, lung, and artery.79,80,8290(684) Some of these early in vitro studies did not account
for an artifact due to self-absorption of fluorescence by oxyhemoglobin released from lysed red blood
cells.29(685) This reabsorption contaminated samples and led to false conclusions concerning the
appearance of new spectral differences.
Ultraviolet LIF in human coronary artery samples was shown to be sufficient to distinguish normal from
atherosclerotic segments of human arteries.40,73,9193(686) The origin of fluorescence in artery
samples was attributed entirely to extracellular chromophores intrinsic to structural proteins contained
in collagen and elastin, with a small contribution from ceroid, an oxidation product of lipids.40,73(687)
The important contribution of oxyhemoglobin reabsorption in fluorescence spectra was recognized in
these studies, which included corrections for this factor. The fluorescence changes in atherosclerotic
tissue were primarily a decrease in overall fluorescence (attributed to a decreased concentration of the
more brightly fluorescing elastin in diseased arteries) and a slight shift in fluorescence line shape
toward an increase in fluorescence in the yellow region of the spectrum (attributed to the presence of
ceroid in atherosclerotic tissue).
In vitro studies of fluorescence spectroscopy in GI tissues have been performed using colonic
tissues.63,74,9497(688) To date, these have been limited in scope, concentrating primarily on
differentiation of hyperplastic and adenomatous polyps from normal mucosal tissue. Nonetheless, the
adenomatous polyp as a model for colonic dysplasia has become one of the most thoroughly
characterized of all autofluorescent human tissues, both in vitro and in vivo. In vitro studies have been
conducted on surgically excised tissue specimens so as to avoid the oxyhemoglobin reabsorption
artifact encountered in other early work. All practical excitation and emission wavelengths have been
examined thoroughly and plotted in an instructive format called the excitation-emission matrix
(EEM).63(689)
The EEM is the most general format for presentation of spectra from inelastic scattering
events.98(690) An EEM displays all possible excitation wavelengths and emission spectra over given
wavelength regions on a two-dimensional contour plot (Figure 175). Excitation wavelengths are
plotted on the ordinate (y-axis) and emission wavelengths on the abscissa (x-axis). The lines on the
plots are contours of constant fluorescent intensity; some values for the contour levels are labeled on
the plots. The more familiar single line shape plots of excitation and emission spectra correspond to
vertical and horizontal slices, respectively, through the three-dimensional EEM data sets. No data are
ever present in an EEM for emission wavelengths shorter than the corresponding excitation
wavelength. This relates to the principle of conservation of energy. Shorter wavelengths correspond to
higher-energy photons. Because energy is not created in scattering events, the wavelength of
scattered (i.e., fluorescent) light is always equal to or greater than the wavelength of the excitation light.
(691)Figure 175. Excitation-emission matrices (EEMs) obtained in human colon samples in

vitro. In all cases, the excitation wavelength is plotted on the ordinate and the emission
wavelength is plotted on the abscissa. Contour lines connect points of equal fluorescence
intensities. Although fluorescence intensities are given in arbitrary units, the same scale of
units is maintained between plots. A, Average EEM of four normal human colon samples.
Three sets of equally spaced contours are shown: 10 from 1 to 10 units; 10 from 10 to 100
units; and 2 from 150 to 200 units. B, Average EEM of 11 adenomatous colon polyp samples.
Four sets of linearly spaced contours are shown: 2 from 0.5 to 1 unit; 10 from 1 to 10 units;
and 2 from 150 to 200 units. C, Ratio of the average adenomatous EEM divided by the
average normal EEM. Twenty contours are spaced linearly from 0.5 to 1.0. Local maxima
represent regions where adenomatous tissue exhibits increased fluorescence compared with
normal tissue, and local minima represent regions where adenomatous tissue exhibits
decreased fluorescence compared with normal tissue. Optimum excitation regions for
differentiating the fluorescence spectra of normal and adenomatous tissues are indicated by
20-nm-wide horizontal bands centered at exc = 330, 370, and 430 nm. (A-C, From
Richards-Kortum R, Rava RP, Petras RE, et al. Spectroscopic diagnosis of colonic dysplasia.
Photochem Photobiol 1991; 53:77786.)
Elastically scattered light, the brightest component of reemitted light owing to its high quantum yield, is
for the most part removed from an EEM by means of optical filters. However, remnants appear as a
dark diagonal line at matching excitation and emission wavelengths (see Figure 175). This is because
elastic scattering involves no frequency change. If it had not been specifically filtered out in this
experiment, the data at the 45-degree line (see Figure 175) would represent the absorption spectrum
of the tissue, because excitation and emission wavelengths are matched on that line.
The EEM is a very useful device in spectroscopy because it allows for identification of the origins of
spectral features using both degrees of freedom that are available from both exciting and emitted light.
It is reasonable to conjecture that, except for absorption artifacts, each "mountain" or "valley" in the
contour plot of an EEM (see Figure 175) arises from some substance within the tissue, which can be
excited over a range of wavelengths and emits fluorescent light over a slightly longer range of
wavelengths. For example, the features on the EEM plots in Figure 175 correlate well with previously
observed chromophore peak absorption and emission wavelengths, as listed in Table 171. It is
notable that both intracellular and extracellular chromophores appear to be involved in the generation
of an EEM. Prominent absorption bands are expected to appear as valleys in the fluorescence spectra
along both excitation and emission axes. For example, absorption bands due to hemoglobin can be
observed along both axes at 420, 540, and 580 nm.
Averaged EEMs obtained by Richards-Kortum et al.63(692) in a study of normal colonic mucosa and
adenomatous polyps are plotted, respectively, in Figure 175A and 175B. Both tissues exhibited
peaks attributed to tryptophan and NADH/NADPH at (ex, em) = (290, 330 nm) and (ex, em) =
(345, 365 nm), respectively, although the latter peak was approximately twice as intense in normal as
in adenomatous tissue. A number of other subsidiary peaks appear in Figure 175A and 175B, some
of which are notably different in intensity and location between the two types of tissue.
A useful technique for comparing EEMs in different tissues is to plot their ratio, as illustrated in Figure
175C. Such a plot can be used to develop algorithms that rapidly differentiate tissues. From the
relative flatness (lack of peaks) of the ratio EEM over most of the visible region (see Figure 175C), it
is evident that both tissues exhibit fluorescence peaks in similar locations, with normal tissues having
approximately two to three times the fluorescence intensity of adenoma. A potentially important
exception to this trend is a local maximum in the ratio map in the region of (ex, em) = (290, 330 nm).
This could be indicative of an observable increase in the concentration of endogenous porphyrins in
adenoma. Richards-Kortum et al.63(693) suggest three 20-nm-wide excitation bands corresponding to
330 + 10 nm, 370 + 10 nm, and 430 + 10 nm as optimum excitation wavelengths for differentiating
normal from adenomatous tissue in rapid (i.e., in vivo) fluorescence-based diagnostic systems.
Although EEMs provide quantitative information regarding the spectral characteristics of fluorescent
chromophores in bulk tissue samples, the spatial distribution of fluorescence in tissues is lost using
this technique. However, it is possible to achieve greater precision in defining the chromophores that
are responsible for autofluorescence through the use of microspectrofluorometry
techniques.71,74,75(694)
Microspectrofluorometry is the study of fluorescence spectra of intact tissue samples through
microscopes, in which the spatial organization of the sample is preserved and the fluorescence
distribution analyzed. An example of the increased information content of the microscopic technique is
provided in Figure 176, which is a comparison of cross-sectional fluorescence images in normal and
low-grade dysplastic human colon samples. These samples were imaged as unstained fresh frozen
sections using a confocal laser scanning fluorescence microscope with ultraviolet (351 to 364 nm)
illumination and a 515-nm long-pass filter. Intense extracellular fluorescence is observed in connective
tissue components in the lamina propria and submucosa. Intense fluorescence is also noted as
intracellular granules, which have been attributed variously to eosinophils75,76(695) and
lipopigments.74(696) The extracellular fluorescence intensities are approximately equal for the two
samples, with the possible exception of an increased quantity of fluorescing granules in the dysplastic
tissue sample. However, a notable difference in fluorescence intensity is evident for cytoplasmic
fluorescence between normal and dysplastic epithelial crypt cells. Very little fluorescence is seen in
normal epithelial cells in this wavelength range, but it is readily evident that dysplastic cells
fluorescence. The source of this cytoplasmic fluorescence has not yet been conclusively elucidated.
Nevertheless, the presence of distinct fluorescing chromophores within dysplastic epithelial cells raises
the possibility that novel, perhaps disease-specific, chromophores may be present.75,97(697)
(698)Figure 176. Laser scanning confocal microscope image (left) of 10-m unstained fresh
frozen samples of normal human colonic tissue obtained using 351- to 364-nm argon-ion laser
illumination with a 515-nm barrier filter. Intense autofluorescence is observed in connective
tissue and intracellular granules. Minimal fluorescence is observed within columnar epithelial
cells. Corresponding image (right) of low-grade dysplastic crypts within a tubular adenoma
from a different patient. Increased fluorescence is observed within epithelial cells lining the
dysplastic crypts.
Phototoxic Dye-Enhanced Fluorescence Spectroscopy
One potential approach to enhance the detection of early-stage cancers by means of fluorescence
spectroscopy is the use of low-dose exogenous phototoxic dyes and/or drugs. Because these agents
have an affinity for neoplastic tissue, they can serve as characteristic fluorescence markers. It has long
been recognized that porphyrin compounds have an affinity for neoplastic
tissues14,22,2427,66,67,69(699) as well as characteristic fluorescence emission with near-ultraviolet
excitation.14,23,24,99(700) The use of such exogenous chromophores may supplement tissue
autofluorescence when the intensity of the latter is too weak to be reliably detected.100(701) The
development of a new generation of phototoxic drugs based on phthalocyanine compounds is of
particular interest with regard to enhanced tissue spectroscopy. These agents have greater
phototoxicity and fewer side effects than porphyrin compounds but still exhibit bright
fluorescence.101,102(702)
Time-Resolved Fluorescence Spectroscopy
A novel recent development in the field of fluorescence spectroscopy is the use of time-resolved and
frequencydomain methods to measure fluorescence lifetimes rather than fluorescence intensity
alone.2,23,3335,103109(703) As noted earlier, inelastic scattering events may be noninstantaneous
if they involve energy transitions to intermediate energy levels, which must decay before secondary
(fluorescence) light is emitted. This is the case with fluorescence emission, which typically involves
excited-state lifetimes ranging from 10-10 to 10-6 sec. Fluorescence lifetimes of specific molecular
transitions can be equally or more sensitive to changes in the molecule's microenvironment than to
changes in fluorescent intensities.
Fluorescence lifetime measurements also overcome many of the difficulties inherent to
fluorescence-intensity sensing such as photobleaching, light source instabilities, and irreproducibilities
in the sample illumination geometry. Several diagnostic algorithms based on fluorescence
spectroscopy that have been proposed for detection of dysplasia in the colon, for example, depend
primarily on changes in fluorescence intensity over a certain range of wavelengths (see later). To be
maintained, these algorithms thus depend on accurate calibration of fluorescence intensities. In
practice, this may be difficult to achieve in an unobtrusive manner within an organ system, such as the
gut, which has a tortuous lumen. Time-resolved measurements have the potential in such situations to
provide diagnostic algorithms free from calibration artifacts.
Developments in time-resolved fluorescence spectroscopy have been stimulated by rapid technologic
advances in ultrafast optical detectors and electronics, including the capability for two-dimensional
imaging of fluorescence lifetimes with sub-nanosecond time resolution.105(704) Frequency-domain
techniques, which rely on measurements of phase delay and modulation depth of reemitted
radiofrequency-modulated continuous light, constitute a particularly promising and cost-effective
technology for fluorescence lifetime sensing.13,103,105,108,110,111(705)
Most developments in fluorescence lifetime imaging to date relate to applications in analytical and
clinical chemistry, including measurements of pH, blood gases, and glucose as well as intracellular
calcium and other ions. These measurements in turn depend on the development of specific molecular
probes, or moieties, whose fluorescent lifetimes are altered in the presence of a specific ion or
molecule. This dependence on the synthesis and construction of specific probes is at present one of
the principal drawbacks of the technique with respect to imaging of bulk tissue. Probes that operate at
the near-infrared wavelengths necessary for deep tissue penetration and excitation by cost-effective
diode lasers are now becoming available.105(706) Thus, fluorescence lifetime sensing is as yet
untested as an approach to imaging in bulk tissue.
Raman Spectroscopy
The Raman effect was discovered by Sir C. V. Raman in 1928 in a remarkable feat of experimental
physics in which he used the sun as a light source and his eye as a detector. In the Raman effect,
photons experience a small shift in frequency when scattered from molecules, the frequency shift
being equal to a vibrational or rotational energy transition within the molecule. The photons can either
gain or lose energy from molecules. Because most molecules are in their ground vibrational states at
room temperature, the more likely scenario is for the molecule to be excited and the photon to be
shifted to a lower frequency.
Raman spectroscopy is performed by shining a light source on a collection of molecules and recording
the spectra of the scattered light. Raman spectroscopy is a much more specific probe of molecular
chemistry than fluorescence spectroscopy because vibrational/rotational energy levels are specific to
molecular bonds rather than atomic electron configurations as in fluorescence.112(707) The limitation
of Raman spectroscopy is that its quantum yield is several orders of magnitude less than that for
fluorescence (see Figure 171). Raman spectroscopy was initially discounted as impossible in biologic
tissue owing to this decrease in quantum yield; however, remarkable advances in laser and detector
technology have resulted in the 1990s in demonstration of Raman spectroscopy in tissue for the first
time.106,113123(708)
Near-Infrared Raman Spectroscopy
Vibrational (Raman) spectroscopy has the potential for increased specificity in the diagnosis of disease
owing to its intrinsically narrow line shapes and rich spectral content, which is reflective of molecular
structure. Raman spectroscopy is similar to infrared absorption spectroscopy in that frequencies of
molecular vibrations and rotations are identified. The use of the Raman technique, however, makes the
vibrational and rotational spectra accessible using commonly available optics and detectors designed
for use in the visible wavelengths, this information being otherwise obtainable only in the infrared
region. In particular, laser sources can be used to excite Raman spectra in the near-infrared
wavelength region, where tissue penetration is deep and high quantum efficiency multichannel
detectors can be used. The use of excitation light in the near-infrared region of the spectrum also
avoids excitation of endogenous fluorescence, which would overwhelm the weak Raman signal.
Raman spectroscopic studies using laser excitation wavelengths at 1064 nm and near 800 nm have
successfully identified the presence of cholesterols and phosphates in atheromatous and calcified
atherosclerotic plaques, respectively.114117(709)
Ultraviolet Resonance Raman Spectroscopy
An alternative for elimination of electron fluorescence artifact in Raman spectroscopy is the use of
deep ultraviolet light to excite high-lying electronic states; quenching and relaxation from these states
either eliminate fluorescence or cause it to occur at higher wavelengths. This technique, known as
ultraviolet resonance Raman (UVRR) spectroscopy, has been used to probe vibrational lines specific
to nucleic acids in adenine and guanine residues.124(710) In studies comparing UVRR spectroscopy
in normal human colonic mucosa versus adenocarcinoma, the nucleic acid bands were more
pronounced in the cancerous tissue, indicating increased cellular nuclear content (Figure
177).120(711) Although this technique is unlikely to find clinical application owing to the poor
penetration and potential mutagenicity of the ultraviolet laser light, it is an excellent example of the
potential of highly specific advanced spectroscopic techniques in the detection of disease in biologic
tissue.
(712)Figure 177. A, Ultraviolet resonance Raman spectroscopy (UVRR) of DNA (salmon

sperm) and tyrosine dissolved in water (1 mg/ml), obtained at an excitation wavelength of 250
nm. The intense peaks near 1580 cm-1, 1485 cm-1, and 1337 cm -1 have been identified as
originating in the purine bases guanine and adenine of DNA. Tyrosine, a typical protein
aromatic amino acid, exhibits bands of the benzene system at 1603 cm-1 and 1180 cm-1. B,
UVRR of normal human colonic mucosa and adenocarcinoma of the colon, also obtained at an
excitation wavelength of 250 nm. Comparison with the spectra in A indicates that the intense
bands are due to the nucleic and amino acid content in the colonic tissue. The adenomatous
samples has more pronounced nucleic acid bands than the normal sample, revealing increased
nuclear content in the neoplasm. (A and B, From Manoharan R, Wang Y, Dasari RR, et al.
Ultraviolet resonance Raman spectroscopy for detection of colon cancer. Lasers Life Sci
1995; 6:21727.)
Clinical Tissue Spectroscopy Studies

Gastrointestinal Tract
Several studies of colonic mucosal LIF in humans have been reported; all of these used the colonic
adenoma as a model for dysplasia.125128(713) Typical spectra reported in one such study by
Cothren et al.125(714) are reproduced in Figure 178; a summary of selected results from various
clinical LIF studies is provided in Table 172. Although the results reported to date of clinical studies
using the colon polyp model are for the most part consistent with those from previous in vitro studies
(see Figure 176), these data are interpreted differently by several investigators. Two major groups of
investigators have demonstrated that LIF has sensitivity and specificity in the differentiation of
adenoma from normal colonic mucosa that are comparable to or better than those for histopathologic
interpretation.125128(715) However, differences of interpretation have arisen with regard to the
origins of the observed fluorescent signals and their significance.
TABLE 172
FIRST
AUTHOR
Selection of In Vivo Fluorescence Spectroscopy Clinical Studies

EXCITATION
(NM)
SPECIMENS
DISCRIMINATION
GOAL
ALGORITHM
Adenoma from
normal and
hyperplastic
Intensity ratio of
two emission
wavelengths
SENSITIVITY
%
Colon
Cothren125
370
32 Normal
31
Adenomatous
polyp
4 Hyperplastic
polyp
100
TABLE 172
Selection of In Vivo Fluorescence Spectroscopy Clinical Studies
FIRST
AUTHOR
EXCITATION
(NM)
Schomacker126
337
SPECIMENS
86 Normal
DISCRIMINATION
GOAL
ALGORITHM
SENSITIVITY
%
Adenoma and
carcinoma from
normal and
hyperplastic
Multivariate
linear
regression of
12 emission
wavelengths
80
Normal from
adenocarcinoma
Differential
normalized
fluorescence
98
Moderate, severe,
CIS invasive from
normal
Proprietary
73
1. Abnormal from
normal
2. Preneoplastic,
neoplastic from
non-neoplastic,
abnormal
Ratio of slope
to peak
Ratio of two
slopes
92
49
Adenomatous
polyp
35 Hyperplastic
polyp
7 Carcinoma
Esophagus
Vo-Dinh130
410
45 Normal
29 Adenocarcinoma
Lung
Lam132
440
173 Normal
Dysplasia:
14 Mild
33 Moderate
15 Severe
29 CIS
64 Invasive Ca
Cervix
Ramunujam136
337
66 Normal
23 Cervical
intraepithelial
neoplasia
5 Inflammation
21 Human
papillomavirus
Cacancer; CIScarcinoma in situ.
(716)Figure 178. A, Typical spectra obtained from in vivo fluorescence measurements of

normal colonic mucosa, a hyperplastic polyp, and an adenomatous polyp using 370 nm
87
excitation. B, Ratio of the mean spectrum from 33 adenomatous polyps divided by the mean
spectrum from 43 normal colonic mucosa sites. Over most of the spectrum (400 to 600 nm),
adenomatous tissue autofluorescence is similar in lineshape to normal tissue autofluorescence
but weaker in fluorescence intensity by approximately a factor of 3. Lineshape differences
occur in the red region, where the adenoma fluoresces more brightly than normal tissue. (A,
and B, From Cothren RM, Sivak MV, Van Dam J, et al. Detection of dysplasia at colonoscopy
using laser-induced fluorescence: A blinded study. Gastrointest Endosc 1996; 44:16876.)
Studies conducted by Schomacker et al.126,127(717) using ultraviolet excitation at wavelengths of
less than 350 nm consistently found that the transition from normal colonic tissue to hyperplastic polyp
to adenomatous polyps corresponds to decreasing collagen fluorescence in the 380- to 400-nm region
and a relative increase in fluorescence, attributed to NADH and FAD, at wavelengths greater than 450
nm. These investigators compared observed fluorescence spectra from all three histologic types of
colon tissue with spectra obtained in purified solutions of Type I collagen, NADH, and FAD. They found
that the spectra obtained in vivo and in vitro were nearly the same except in the wavelength region
from 400 to 460 nm, where hemoglobin absorption artifact was suspected. Schomacker et
al.126,127(718) also observed peaks attributed to endogenous porphyrins in abnormal tissue only in
the 650-nm region. It is notable that with excitation wavelengths of less than 350 nm, the largest
observed fluorescence associated with the transformation to dysplastic tissue was a significant
decrease in the collagen fluorescence in the 380- to 400-nm region. Changes in the intensity of this
peak played a correspondingly large role in their diagnostic algorithm. Based primarily on these
observations, these investigators stipulated that the changes sensed by their algorithm were primarily
due to changes in polyp morphology rather than to changes in fluorescence that are specific for
adenoma. This is because most of the collagen, the source of the most intense peak in their data, is
assumed to be localized to connective tissue in the submucosa. Thus, the decrease in 380- to 400-nm
fluorescence as measured at the mucosal surface in dysplastic tissue is interpreted as being due
primarily to shielding of both the excitation light and the fluorescent light by a thickened dysplastic
mucosa.
The clinical LIF studies by Cothren et al.125,128(719) of the colon adenoma model for dysplasia used
an excitation wavelength of 370 nm, this choice of wavelength being based on data from previous in
vitro EEM studies.63(720) At the 370-nm wavelength, collagen fluorescence is not as strongly excited.
Thus, NAD(P)H, FAD, and porphyrins appear to play an increased role in the observed spectra. In this
case, it was reliably observed that the transformation from normal colonic mucosa to hyperplastic polyp
to adenomatous polyp corresponded to a strong decrease in fluorescence in the 400- to 600-nm range,
accompanied by an increase in fluorescence at wavelengths longer than 650 nm. The decrease in 400to 600-nm fluorescence could be explained in part by a thickened dysplastic mucosa. However, the
observed differences could also be explained by the presence of intrinsic fluorescence differences
between dysplastic and normal tissue arising from as yet unidentified molecular species. This point of
view is supported by the observation of unique fluorescence arising in the cytoplasm of dysplastic
epithelial cells (see Figure 175).
Overall, the results of clinical LIF studies of colon adenomas indicate that differentiation of normal from
adenomatous tissue by means of this technique is at least as reliable as histopathologic interpretation.
Differentiation of normal tissue from hyperplastic polyps by LIF is less reliable. The debate concerning
mechanisms of LIF has important implications for the future design of clinical instrument systems. If
differences in LIF spectra can be found that result from biochemical rather than morphologic
differences in tissue, the necessary technology needed to detect such line shape differences is more
straightforward to design and implement than that which would be required for detection of intensity
differences, as the latter systems would be sensitive to calibration errors, stray reflections, and system
noise.
A consensus view concerning the interpretation of LIF spectra appears to be emerging in more recent
publications.74,97,128(721) This holds that two characteristic features differentiate normal mucosa
from adenomatous tissue at an excitation wavelength of 370 nm. For adenomas, fluorescence intensity
is decreased in the 400- to 550-nm spectral range and increased in the red region at wavelengths
greater than 550 nm. The intensity of adenoma fluorescence is less than that of normal mucosa in the
400- to 550-nm region because (1) mucosal collagen fluorescence is reduced as a result of
enlargement of the crypts; (2) adenoma exhibits greater absorption in this range as a result of
increased hemoglobin content; and (3) fluorescence arising in the submucosa is decreased in
adenoma as a result of shielding by the thickened mucosa. The fluorescence intensity of adenoma is
increased in the red region (i.e., at wavelengths greater than 550 nm) owing to the increased intrinsic
fluorescence of dysplastic crypt cells.
Thus, LIF is much more informative when an excitation wavelength of 370 nm is used rather than
wavelengths less than 350 nm. With the latter frequency, collagen fluorescence in the 380- to 400-nm
region is totally dominant, and the difference between normal tissue and adenoma is primarily
attributable to mucosal shielding. The observed spectral differences with 370-nm excitation are due to
a combination of factors related to intrinsic and biochemical changes as well as morphologic
alterations.
A clinical study of LIF diagnosis of esophageal malignancy has recently been reported.129,130(722)
Fluorescence in normal and malignant esophageal tumors was excited at 410 nm and collected over
the visible range. The results were similar to those of studies of the colon adenoma model at excitation
wavelengths greater than 350 nm. A large reproducible decrease in fluorescence was observed in the
450- to 550-nm range in abnormal tissues compared with normal esophageal wall. An algorithm
sensitive to this intensity change was developed; the correlation between algorithm predictions and
histopathologic classification was excellent (see Table 172).
It is important to bear in mind that the results of clinical LIF studies, as tabulated in Table 172,
represent the performance of algorithms applied retrospectively to the data sets from which they were
derived. In a recent study of in vivo LIF diagnosis, however, Cothren et al.128(723) have applied an
algorithm in a blinded, prospective manner. The algorithm developed in this study is probability based
and presents results of LIF in terms of likelihoods of disease or normal states in contrast to the binary
decision algorithms used in previous studies. An example of the "decision space" produced by this
algorithm is shown in Figure 179. This type of algorithm is more likely representative of the true
nature of progressively developing GI cancer.
(724)Figure 179. Decision space resulting from a probabilistic algorithm differentiating

normal colonic mucosa from adenomas and hyperplastic polyps. The contours indicate the
probability that a spectrum represented by parameters at that point is from the indicated tissue
type; bold lines separate the three tissue classifications. The dotted ellipses encompass most of
the data from the retrospective data set from which the probabilities were calculated. (From
Cothren RM, Sivak MV, Van Dam J, et al. Detection of dysplasia at colonoscopy using
laser-induced fluorescence: A blinded study. Gastrointest Endosc 1996; 44:16876.)
Respiratory Tract
Currently, the most advanced system for clinical fluorescence-based cancer diagnosis has been used
in LIF studies of malignancy in the respiratory tract.86,87,131134(725) This system consists of a
real-time imaging device coupled to a standard fiberoptic bronchoscope. It makes use of band-pass
filters and intensified CCD cameras to perform ratio imaging of entire fluorescence images in the 500and 650-nm regions at an illumination wavelength of 442 nm.87,132134(726) Lam et al.132(727)
have demonstrated that this system provides identical specificity (94%) to but increased sensitivity
(72.5%) over standard bronchoscopy with a white-light source (48.4%) for the detection of dysplasia
and carcinoma in situ in more than 100 patients. Fluorescence intensity differences have been
attributed by these investigators primarily to differences in the oxidation-reduction state of flavins.
Urogenital Tract
The majority of in vivo studies of LIF to date have been conducted in the GI and respiratory tracts,
although these techniques have also been applied in studies of other organ systems.135137(728)
Following exploratory studies in vitro,135(729) Ramanujam et al.136(730) investigated spectroscopic
identification of cervical intraepithelial neoplasia (CIN) in vivo (see Table 171) in a two-part study. In
the first part, histologically abnormal tissues were differentiated from colposcopically normal samples
with both a sensitivity and a specificity greater than 90% (see Table 172). These data were then
reanalyzed to separate abnormal samples into neoplastic and nonneoplastic categories because of the
clinical necessity of differentiating CIN from infection by the human papillomavirus. This differentiation
was also made with a specificity superior to that of visual assessment by a skilled colposcopist. The
observed spectroscopic differences were consistent with observations described earlier for colon
polyps at excitation wavelengths of less than 350 nm. Specifically, the progression from normal to
abnormal tissues was accompanied by decreasing collagen and elastin138(731) fluorescence and
increasing NAD(P)H fluorescence. Porphyrin fluorescence was not observed in this study.
Summary
GI endoscopy has attained a high level of sophistication in the diagnosis of digestive diseases. For the
most part, however, these advances are related to the development of flexible endoscopes that extend
the range of visual observation within the gut and of accessories for the procurement of tissue samples
for histopathologic diagnosis. In fact, endoscopic diagnosis frequently depends on histologic
confirmation. This dependency imposes a serious limitationnamely, it is frequently difficult or
impossible to diagnose disease that cannot be recognized by visual observation. Despite remarkable
progress in the design and construction of endoscopic instruments, endoscopic diagnosis has
remained fundamentally unchanged for the last 200 years. It depends on the human eye and the
experience of the endoscopist.
Tissue spectroscopy opens the distinct possibility of new methods of endoscopic diagnosis that may
dramatically extend the range of visual observation. Within certain limits, it may be possible to arrive at
a tentative histopathologic diagnosis by means of novel imaging systems based on the principles of
spectroscopy. This has important implications with respect to a variety of diseases, but especially
those that carry a high risk of malignancy. As compelling as this prospect may be, much further work is
needed to define the value and limitations of tissue spectroscopy in endoscopic diagnosis.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Born M, Wolf E. Principles of Optics. New York: Pergamon Press. 1980.

Chance B, Kang K, Sevick E. Photon diffusion in breast and brain: Spectroscopy and imaging.
Opt Photon News 1993;4:913.
Izatt JA, Hee MR, Huang D, et al. Micron-resolution biomedical imaging with optical
coherence tomography. Opt Photon News 1993;4:149.
Griffin WP. Three-dimensional imaging in endoscopic surgery. Biomed Instrument Tech
1995;May/June:1839.
Katzir A. Lasers and Optical Fibers in Medicine. New York: Academic Press. 1993.
Chance B (ed). Photon Migration in Tissues. New York: Plenum Press. 1990.
Ishimaru A. Wave Propagation and Scattering in Random Media. New York: Academic Press.
1978.
Chandrasekhar S. Radiative Transfer. New York: Dover. 1960.
Henry AF. Nuclear-Reactor Analysis. Cambridge, MA: MIT Press. 1982.
Bohren CF, Huffman DR. Absorption and Scattering of Light by Small Particles. New York:
John Wiley & Sons. 1983.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
Cheong WF, Prahl SA, Welch AJ. A review of the optical properties of biological tissues. IEEE
J Quant Elect 1990;QE-26:216685.
Chance B. Comparison of time-resolved and unresolved measurements of deoxyhemoglobin
in brain. Proc Natl Acad Sci USA 1988;85:49715.
DeBlasi RA. Cerebral and muscle oxygen saturation measurement by frequency-domain
near-infrared spectrometer. Med Biol Eng Comput 1995;33:22830.
Cubeddu R, Canti G, Vanentini G. Absorption spectrum of hematoporphyrin derivative in vivo
in a murine tumor model. Photochem Photobiol 1994;60:5825.
Horecker BL. The absorption spectra of hemoglobin and its derivatives in the visible and near
infra regions. J Biol Chem 1943;148:17383.
Roughton FJW, Kendrew JC. Haemoglobin. New York: Interscience. 1949.
Lackowicz JR. Principles of Fluorescence Spectroscopy. New York: Plenum Press. 1985.
Yoshiya I, Shimada Y, Tanaka K. Spectrophotometric monitoring of arterial oxygen saturation
in the fingertip. Med Biol Eng Comput 1980;18:2732.
Bowes WAI, Corke BC, Hulka J. Pulse oximetry: A review of the theory, accuracy, and clinical
applications. Obstet Gynecol 1989;74:5416.
Bottiroli G, Ramponi R, Croce AC. Quantitative analysis of intracellular behaviour of
porphyrins. J Photochem Photobiol 1987;46:6637.
Bottiroli G, Ramponi R. Equilibrium among hematoporphyrin-derivative components: Influence
of the interaction with cellular structures. J Photochem Photobiol 1988;47:20914.
Chung-Ho S, Duzman E, Mellott J, et al. Spectroscopic, morphologic, and cytotoxic studies on
major fractions of hematoporphyrin derivative and photofrin II. Lasers Surg Med
1987;7:1719.
Cubeddu R, Ramponi R, Taroni P. Time-gated fluorescence spectroscopy of porphyrin
derivatives incorporated into cells. J Photochem Photobiol 1990;6:3948.
Dal Fante M, Bottiroli G, Spinelli P. Behaviour of haematoporphyrin derivative in adenomas
and adenocarcinomas of the colon: A microfluorometric study. Lasers Med Sci
1988;3:16571.
Figge FH, Weiland GS, Manganiello LO. Cancer detection and therapy: Affinity of neoplastic,
embryonic, and traumatized tissues for porphyrins. Proc Soc Exp Biol Med 1948;68:6401.
Harris DM, Weekhaven J. Endogenous porphyrin fluorescence in tumors. Lasers Surg Med
1987;7:46772.
Kessel D. Tumor localization and photosensitization by derivatives of hematoporphyrin: A
review. IEEE J Quant Electron 1987;QE-23:1718.
Mourant JR, Bigio IJ, Boyer J, et al. Spectroscopic diagnosis of bladder cancer with elastic
light scattering. Lasers Surg Med 1995;17:3507.
Liu CH, Tang GC, Pradhan A, et al. Effects of self-absorption by hemoglobins on the
fluorescence spectra from normal and cancerous tissues. Lasers Life Sci 1990;3:16776.
Baraga JJ, Feld MS, Rave RP. Infrared attenuated total reflectance spectroscopy of human
artery: A new modality for diagnosing atherosclerosis. Lasers Life Sci 1992;5:13.
Eddowes MH, Mills TN, Delpy DT. Monte Carlo simulations of coherent backscatter for
identification of the optical coefficients of biological tissues in vivo. Appl Opt 1995;34:226167.
Schmitt JM, Knuttel A, Bonner RF. Measurement of optical properties of biological tissues by
low-coherence reflectometry. Appl Opt 1993;32:603242.
Patterson MS, Chance B, Wilson BC. Time resolved reflectance and transmittance for the
noninvasive measurement of tissue optical properties. Appl Opt 1989;28:23316.
Yamada Y, Proskurin SG, Takahashi Y. Absorption coefficient measurements of highly
scattering media in slabs and cylindrical phantoms by means of time-resolved optical
spectroscopy. In Chance B, Alfano RR (eds). Optical Tomography, Photon Migration, and
Spectroscopy of Tissue and Model Media: Theory, Human Studies, and Instrumentation.
Society of Photo-Instrumentation Engineers 1995;2389:15766.
Oda I, Eda H, Takada M, Tsunazawa Y. New algorithm to determine absorption and
scattering coefficient from time-resolved measurement. In Chance B, Alfano RR (eds). Optical
Tomography, Photon Migration, and Spectroscopy of Tissue and Model Media: Theory,
Human Studies, and Instrumentation. Society of Photo-Instrumentation Engineers
1995;2389:7586.
36.
Gratton E, Fantini S, Franchesini MA, Fishkin JB. Near-infrared optical spectroscopy of
tissues using an LED frequency domain spectrometer. In Alfano RR (ed). Advances in Optical
Imaging and Photon Migration. Orlando, FL: Optical Society of America, 1995;27881.
37.
Liu H, Boas DA, Beauvoit B, et al. Absorption and scattering properties in a highly absorbing
turbid medium. In Alfano RR (ed). Advances in Optical Imaging and Photon Migration.
Orlando, FL: Optical Society of America, 1995;2727.
38.
Wang L, Jacques SL. Use of a laser beam with an oblique angle of incidence to measure the
reduced scattering coefficient of a turbid medium. Appl Opt 1995;34:23626.
39.
Wu J, Feld MS, Rava RP. Analytical model for extracting intrinsic fluorescence in turbid
media. Appl Opt 1993;32:358595.
40.
Richards-Kortum R, Rava RP, Cothren R, et al. A model for extraction of diagnostic
information from laser induced fluorescence spectra of human artery wall. Spectrochim Acta
1989;45A:8793.
41.
Keijzer M, Richards-Kortum RR, Jacques SL, Feld MS. Fluorescence spectroscopy of turbid
media: Autofluorescence of human aorta. Appl Opt 1989;28:428692.
42.
Boyer J, Mourant JR, Bigio IJ. Theoretical and experimental investigations of elastic scattering
spectroscopy as a potential diagnostic for tissue pathologies. In Alfano RR (ed). Advances in
Optical Imaging and Photon Migration. Orlando, FL: Optical Society of America, 1994;2658.
43.
Bigio IJ, Mourant JR, Boyer JD, et al. Noninvasive identification of bladder cancer with
subsurface backscattered light. In Alfano RR (ed). Advances in Laser and Light Spectroscopy
to Diagnose Cancer and Other Diseases. Vol I. Society of Photo-Instrumentation Engineers
1994;2135:2635.
44.
Mourant JR, Boyer JD, Johnson TM, et al. Detection of gastrointestinal cancer by elastic
scattering and absorption spectroscopies with the Los Alamos optical biopsy system. In
Alfano RR (ed). Advances in Laser and Light Spectroscopy to Diagnose Cancer and Other
Diseases. Vol II. Society of Photo-Instrumentation Engineers 1995;2387:2107.
45. Marchesini R, Brambilla M, Clemente C, et al. In vivo spectrophotometric evaluation of
neoplastic and non-neoplastic skin pigmented lesions. I: Reflectance measurements.
Photochem Photobiol 1991;53:7784.
46.
Zeng H, MacAulay C, Palcic G, McLean DI. A computerized autofluorescence and diffuse
reflectance spectroanalyser system for in vivo skin studies. Phys Med Biol 1993;8:23140.
47.
Hajizadeh-Saffar M, Feather JW, Dawson JB. An investigation of factors affecting the
accuracy of in vivo measurements of skin pigments by reflectance spectrophotometry. Phys
Med Biol 1990;9:130115.
48.
Tomatis S, Marchesini R, Bartoli C, et al. Reflectance imaging spectroscopy: Possible aid in
the clinical diagnosis of melanoma. In Alfano RR (ed). Advances in Laser and Light
Spectroscopy to Diagnose Cancer and Other Diseases. Vol II. Society of
Photo-Instrumentation Engineers 1995;2387:95101.
49.
Sevick-Muraca EM, Burch CL. Origin of phosphorescence signals remitted from tissues. Opt
Lett 1994;19:192830.
50.
Udenfriend S. Fluorescence Assay in Biology and Medicine. Vol I. New York: Academic
Press. 1962.
51.
Udenfriend S. Fluorescence Assay in Biology and Medicine. Vol II. New York: Academic
Press. 1969.
52.
Barenboim GM, Domanskii AN, Turoverov KK. Luminescence of Biopolymers and Cells. New
York: Plenum Press. 1969.
53. Bent DV, Hayon E. Excited state chemistry of aromatic amino acids and related peptides. III.
Tryptophan. J Am Chem Soc 1975;97:26129.
54.
Chance B, Baltschefsky H. Respiratory enzymes in oxidative phosphorylation. J Biol Chem
1958;233:736.
55.
Chance B, Cohen P, Jobsis F, Schoener B. Intracellular oxidation-reduction states in vivo.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
Science 1962;137:499508.
Duysens LNM, Sweep G. Fluorescence spectrophotometry of pyridine nucleotide in
photosynthesizing cell. Biochem Biophys Acta 1957;25:136.
Schwartz JP, Possoneau JV, Johnson GS, Pastan I. The effect of growth conditions on
NAD(MathematicalPi-One )1 and NADH concentration and the NAD(MathematicalPi-One
)1:NADH ratio in normal and transformed fibroblasts. J Biol Chem 1974;249:4138.
Mayevsky A. Brain NADH redox state monitored in vivo by fiber optic surface fluorometry.
Brain Res Rev 1984;7:4968.
Chance B, Schoener B. Fluorometric studies of flavin compounds of the respiratory chain. In
Slater EC (ed). Flavins and Flavoprotein. Amsterdam: Elsevier. 1966.
Fasman GD. Handbook of Biochemistry and Molecular Biology. 3rd ed. Cleveland: CRC
Press. 1975.
Bessey OA, Lowry OH, Love RH. The fluorometric measurement of the nucleotides of
riboflavin and their concentration in tissue. J Biol Chem 1949;180:75569.
Kunz WS, Kunz W. Contributions of different enzymes to flavoprotein fluorescence of isolated
rat liver mitochondria. Biochim Biophys Acta 1985;841:23746.
Richards-Kortum R, Rava RP, Petras RE, et al. Spectroscopic diagnosis of colonic dysplasia.
Photochem Photobiol 1991;53:77786.
Wolfbeiss OS, Leiner MJ. Mapping of the total fluorescence of human blood serum as a new
method for its characterization. Anal Chim Acta 1985;167:20315.
Policard A. tudes sur les aspects offerts par des tumeurs experimentales examines la
lumiere de Wood. Compt Rend Soc Biol 1924;91:14234.
Auler H, Banzer G. Untersuchungen ber die Rolle der Porphyrine bei Geschwulstkranken
menschen end tieren. Z Krebsforsch 1942;53:658.
Lipson RL, Baldes EJ, Olsen AM. The use of a derivative of hematoporphyrin in tumor
detection. J Natl Cancer Inst 1961;26:18.
Merril AH, Henderson JM. Diseases associated with defects in vitamin B6 metabolism. Add
Rev Nutr 1987;7:13756.
Yuanlong Y, Yanming Y, Fuming L, et al. Characteristic autofluorescence for cancer diagnosis
and its origin. Lasers Surg Med 1987;7:52832.
Leibovici L, Schoenfeld N, Yehoshua HA, et al. Activity of porphobilinogen deaminase in
peripheral blood mononuclear cells of patients with metastatic cancer. Cancer
1988;62:22972300.
Baraga JJ, Rava RP, Fitzmaurice M, et al. Characterization of the fluorescent morphological
structures in human arterial wall using ultraviolet-excited microspectrofluorimetry.
Atherosclerosis 1991;88:114.
Neuman RE, Logan MA. The determination of collagen and elastin in tissues. J Biol Chem
1950;186:54956.
Richards-Kortum R, Rava RP, Fitzmaurice M, et al. A one-layer model of laser-induced
fluorescence for diagnosis of disease in human tissue: Applications to atherosclerosis. IEEE
Trans Biomed Eng 1989;36:12232.
Bottiroli G, Croce AC, Locatelli D, et al. Natural fluorescence of normal and neoplastic human
colon: A comprehensive "ex vivo" study. Lasers Surg Med 1995;16:4860.
Romer TJ, Fitzmaurice M, Cothren RM, et al. Laser-induced fluorescence microscopy of
normal colon and dysplasia in colonic adenomas: Implications for spectroscopic diagnosis.
Am J Gastroenterol 1995;90:817.
Barnes D, Aggarwal S, Thomsen S, et al. A characterization of the fluorescent properties of
circulating human eosinophils. Photochem Photobiol 1993;58:297303.
Wolman M. Lipid pigments (chromolipids): Their origin, nature and significance. Pathobiol
Annu 1980;10:25367.
Koenig H. The autofluorescence of lysosomes. Its value for the identification of lysosomal
constituents. J Histochem Cytochem 1963;11:5567.
Alfano RR, Tata DB, Cordero J, et al. Laser induced fluorescence spectroscopy from native
cancerous and normal tissue. IEEE J Quant Electron 1984;QE-20:150711.

Alfano RR, Tang GC, Pradhan A, et al. Fluorescence spectra for cancerous and normal
human breast and lung tissues. IEEE J Quant Electron 1987;QE-23:180611.
81.
Kozikowski SD, Wolfram LJ, Alfano RR. Fluorescence spectroscopy of eumelanins. IEEE J
Quant Electron 1984;QE-20:137982.
82.
Alfano RR, Asima P, Tang GC, Wahl SJ. Optical spectroscopic diagnosis of cancer and
normal breast tissues. Opt Soc Am B 1989;6:101523.
83.
Anderson PS, Kjellin E, Montan E, et al. Autofluorescence of various rodent tissues and
human skin tumor samples. Lasers Med Sci 1987;2:419.
84.
Anderson PS, Montan S, Svanberg S. Multispectral system for medical fluorescence imaging.
IEEE J Quant Electron 1987;QE-23:17981805.
85.
Andersson-Engels S, Johansson J, Svanberg K, Svanberg S. Laser-induced fluorescence in
medical diagnosis. Society of Photo-Instrumentation Engineers 1990;1203:7696.
86. Hung J, Lam S, LeRiche JC, Palcic B. Autofluorescence of normal and malignant bronchial
tissue. Lasers Surg Med 1991;11:99105.
87.
Lam S, LeRiche JC, MacAulay C, Palcic B. Detection of early lung cancer and preneoplastic
lesions. Lung Cancer 1994;11(suppl 2):1612.
88.
Svanberg K, Clemente LP, Clemente MP, et al. Pharmacokinetic studies of
(MathematicalPi-One )d-amino levulinic acid-induced protoporphyrin IX build-up in some
malignant tumors. In Alfano RR (ed). Advances in Laser and Light Spectroscopy to Diagnose
Cancer and Other Diseases. Vol II. Society of Photo-Instrumentation Engineers
1995;2387:3042.
89.
Tang GC, Pradhan A, Sha W, et al. Pulsed and CW laser fluorescence spectra from
cancerous, normal, and chemically treated normal human breast and lung tissues. Appl Opt
1989;28:233742.
90. Tang GC, Pradhan A, Alfano RR. Spectroscopic differences between human cancer and
normal lung and breast tissues. Lasers Surg Med 1989;9:2905.
91. Richards-Kortum R, Mehta A, Hayes G, et al. Spectral diagnosis of atherosclerosis using an
optical fiber laser catheter. Am Heart J 1989;118:381.
92. Garrand TJ, Stetz ML, O'Brien KM, et al. Design and evaluation of a fiberoptic fluorescence
guided laser recanalization system. Lasers Surg Med 1991;11:10616.
93.
Feld MS, Kramer JR, Albagli D, et al. LAS II: An integrated system for spectral diagnosis,
guidance, and ablation in laser angiosurgery. In Vogel JHK, King SB (eds). The Practice of
Interventional Cardiology. 2nd ed. St. Louis: Mosby-Year Book, 1993;189202.
94.
Yashke PN, Bonner RF, Cohen P, et al. Laser induced fluorescence spectroscopy may
distinguish colon cancer from normal human colon. Gastrointest Endosc 1989;35:184.
95. Kapadia CR, Kutruzzola FW, O'Brien K, et al. Laser-induced fluorescence spectroscopy of
human colonic mucosa. Detection of adenomatous transformation. Gastroenterology
1990;99:1507.
96. Marchesini R, Brambilla M, Pignoli E. Laser-induced fluorescence of adenomas,
adenocarcinomas and non-neoplastic mucosa in human colon. I. In vitro measurements. J
Photochem Photobiol B 1992;14:21930.
97. Zonios GI, Cothren RM, Arendt JT, et al. Morphological model of human colon tissue
fluorescence. IEEE Trans Biomed Eng 1996;43:11322.
98.
Christian GD, Callis JN, Davidson ER. Array detectors and excitation-emission matrices in
multicomponent analysis. In Wenry EL (ed). Modern Fluorescence Spectroscopy. New York:
Plenum Press 1981;Chapter 4.
99.
Wang I, Andersson-Engels S, Idvall I, et al. In vitro laser-induced fluorescence studies of
breast tumors following low-dose injection of Photofrin. In Alfano RR (ed). Advances in Laser
and Light Spectroscopy to Diagnose Cancer and Other Diseases. Vol I. Society of
100. Baert L, Berg R, Van Damme N, et al. Clinical fluorescence diagnosis of human bladder
carcinoma following low-dose Photofrin injection. Urology 1993;41:32230.
80.
101.
Oleinick NL, Antunez AR, Clay ME, et al. New phthalocyanine photosensitizers for
photodynamic therapy. J Photochem Photobiol 1993;57:2427.
102. Panjehpour M, Sneed RE, Frazier DL, et al. Quantification of chloroaluminum phthalocyanine
tetrasulfonate concentration in rat tissue using laser-induced fluorescence spectroscopy.
Lasers Surg Med 1993;13:2330.
103.
Burch CL, Lakowicz JR, Sevick-Muraca EM. Biochemical sensing in tissues: Determination of
fluorescent lifetimes in multiply scattering media using frequency-domain spectroscopy. In
Alfano RR (ed). Advances in Laser and Light Spectroscopy to Diagnose Cancer and Other
Diseases. Vol I. Society of Photo-Instrumentation Engineers 1994;2135:28699.
104. Jacques SL. Time resolved reflectance spectroscopy in turbid tissues. IEEE Trans Biomed
Eng 1989;36:115561.
105.
Lakowicz JR, Szmacinski H, Thompson RB. Fluorescence-lifetime-based sensing:
Applications to clinical chemistry and cellular imaging. In Cohn GE (ed). Ultrasensitive
Laboratory Diagnostics. Los Angeles, CA: Society of Photo-Instrumentation Engineers,
1993;1895:217.
106. Liu CH, Das BB, Sha Glassman WL, et al. Raman, fluorescence, and time-resolved light
scattering as optical diagnostic techniques to separate diseased and normal biomedical
media. J Photochem Photobiol B 1992;16:187209.
107.
Chance B, Miwa M, Ueda Y. Development of time-resolved spectroscopy system for
quantitative noninvasive tissue measurements. In Chance B, Alfano RR (eds). Optical
Tomography, Photon Migration, and Spectroscopy of Tissue and Model Media: Theory,
Human Studies, and Instrumentation. Society of Photo-Instrumentation Engineers
1995;2389:1429.
108.
Wilson BC, Sevick EM, Patterson MS, Chance B. Time-dependent optical spectroscopy and
imaging for biomedical applications. Proc IEEE 1992;80:91830.
109.
Wu J, Wang Y, Feld MS. Time-resolved multichannel imaging of fluorescent object
embedded in turbid media. Opt Lett 1995;20:48990.
110.
Wilson BC, Jacques SL. Optical reflectance and transmittance of tissues: Principles and
applications. IEEE J Quant Electron 1990;QE-26:218699.
111.
Sevick-Muraca EM, Hutchinson CL, Paithankar DY. Optical tissue biodiagnostics using
fluorescence lifetime. Opt Photon News 1996;7:2545.
112.
Szymanski HA. Raman Spectroscopy. New York: Plenum Press. 1967.
113.
Alfano RR, Liu CH, Sha WL, et al. Human breast tissues studied by IR Fourier transform
Raman spectroscopy. Lasers Life Sci 1991;4:16.
114.
Rava RP, Baraga JJ, Feld MS. Near infrared Fourier transform Raman spectroscopy of
human artery. Spectrochim Acta 1991;47A:50912.
115.
Baraga JJ, Feld MS, Rava RP. Rapid near-infrared Raman spectroscopy of human tissue with
a spectrograph and CCD detector. Appl Spectrosc 1992;46:18790.
116. Baraga JJ, Feld MS, Rava RP. In situ optical histochemistry of human artery using near
infrared Fourier transform Raman spectroscopy. Proc Natl Acad Sci USA 1992;89:34737.
117. Manoharan R, Baraga JJ, Feld MS, Rava RP. Quantitative histochemical analysis of human
artery using Raman spectroscopy. J Photochem Photobiol B 1992;16:21133.
118.
Redd DCB, Feng ZC, Yue KT, Gansler TS. Raman spectroscopic characterization of human
breast tissues: Implications for breast cancer diagnosis. Appl Spectrosc 1993;47:78791.
119. Manoharan R, Baraga JJ, Rava RP, et al. Biochemical analysis and mapping of
atherosclerotic human artery using FT IR microspectroscopy. Atherosclerosis
1993;103:18193.
120.
Manoharan R, Wang Y, Dasari RR, et al. Ultraviolet resonance Raman spectroscopy for
detection of colon cancer. Lasers Life Sci 1995;6:21727.
121.
Nahadevan-Jansen A, Richards-Kortum R. Raman spectroscopy for the detection of cancers
and precancers. Biomed Opt 1995;1:3170.
122.
McCreery RL, Frank CJ, Redd DC. Raman spectroscopy of human biopsy specimens. In
Lakowicz JR (ed). Advances in Fluorescence Sensing Technology II. Society of

Ramanujam A, Ramanujam N, Mitchell MF, et al. Optical techniques for diagnosis of cervical
precancer: A comparison of Raman and fluorescence microscopy. In Lakowicz JR (ed).
Advances in Fluorescence Sensing Technology II. Society of Photo-Instrumentation
Engineers 1995;2388:11020.
124.
Grygon CA, Spiro TG. UV resonance Raman spectroscopy of nucleic acid duplexes
containing A-U and A-T base pairs. Biopolymers 1990;29:70715.
125. Cothren RM, Richards-Kortum R, Sivak MV, et al. Gastrointestinal tissue diagnosis by
laser-induced fluorescence spectroscopy at endoscopy. Gastrointest Endosc
1990;36:10511.
126. Schomacker KT, Frisoli JK, Compton C, et al. Ultraviolet laser-induced fluorescence of
colonic tissue: Basic biology and diagnostic potential. Lasers Surg Med 1992;12:6378.
127. Schomacker KT, Frisoli JK, Compton CC, et al. Ultraviolet laser-induced fluorescence of
colonic polyps. Gastroenterology 1992;102:115560.
128. Cothren RM, Sivak MV, Van Dam J, et al. Detection of dysplasia at colonoscopy using
laser-induced fluorescence: A blinded study. Gastrointest Endosc 1996;44:16876.
129. Panjehpour M, Overholt BF, Schmidhammer JL, et al. Spectroscopic diagnosis of esophageal
cancer: New classification model, improved measurement system. Gastrointest Endosc
1995;41:57781.
130. Vo-Dinh T, Panjehpour M, Overholt BF, et al. In vivo cancer diagnosis of the esophagus using
differential normalized fluorescence (DNF) indices. Lasers Surg Med 1995;16:417.
131.
Qu J, MacAulay C, Lam S, et al. Fluorescence in normal and carcinoma bronchial tissue. In
Jacques S (ed). Laser-Tissue Interaction. Vol V. Society of Photo-Instrumentation Engineers,
1994;2134A:23646.
132. Lam S, MacAulay C, Hung J, et al. Detection of dysplasia and carcinoma in situ with a lung
imaging fluorescence endoscope device. J Thorac Cardiovasc Surg 1993;105:103540.
133. Palcic B, Lam S, Hung J, MacAulay C. Detection and localization of early lung cancer by
imaging techniques. Chest 1991;99:7423.
134. Lam S, MacAulay C, Palcic B. Detection and localization of early lung cancer by imaging
techniques. Chest 1993;103(suppl):12S-14S.
135. Mahadevan A, Mitchell MF, Silva E, et al. Study of the fluorescence properties of normal and
neoplastic human cervical tissue. Lasers Surg Med 1993;13:64755.
136. Ramanujam N, Mitchell MF, Mahadevan A, et al. In vivo diagnosis of cervical intraepithelial
neoplasia using 337-nm-excited laser-induced fluorescence. Proc Natl Acad Sci USA
1994;91:101937.
137.
Glassman WS, Liu CH, Tang GC, et al. Ultraviolet excited fluorescence spectra from
non-malignant and malignant tissues of the gynecological tract. Lasers Life Sci 1992;5:4958.
138. Deyl Z, Macek K, Adam M, Vancikov A. Studies of the chemical nature of elastin
fluorescence. Biochim Biophys Acta 1980;625:24854.
123.
Chapter 18 Spectroscopic Diagnosis and Treatment of Gastric Cancer

(732)
SHU-DONG XIAO, M.D.
Historical Background
Spectroscopic diagnosis and treatment of gastric cancer are based on differences in spectral
properties between normal and abnormal tissues that retain endogenous and exogenous porphyrin
compounds.
Spectroscopic diagnosis was first reported by Policard in 1924.1(733) He noted that some human and
animal tumors fluoresced under illumination with a Wood lamp and claimed that the red fluorescence
emitted from the tumors was attributable to endogenous porphyrins arising from secondary infection by
hemolytic bacteria. Auler and Banzer2(734) in 1942 observed red fluorescence in rat tumors, but not in
normal tissue, following systemic injections of hematoporphyrin, a nonmetallic porphyrin derived from
hemoglobin. The general nature of this apparent tumor localization of hematoporphyrin was
demonstrated in 1948 by Figge et al.3(735) By examining various types of induced and transplanted
tumors in mice, these authors found porphyrin fluorescence in several types of sarcomas and
carcinomas, as well as in lymph nodes, traumatized tissues, and embryonic tissues. Red fluorescence
in human tumors after systemic injection of hematoporphyrin derivative (HpD) was first demonstrated
by Lipson et al.4(736) in 1961. In addition to fluorescence of frankly malignant tumors, reports also
indicate that injected HpD may cause premalignant lesions to fluoresce (e.g., carcinoma in situ of the
uterine cervix and urinary bladder).5(737)
Complete eradication of experimental tumors with spectroscopic treatment or photodynamic therapy
(PDT) was first performed in 1975.6(738) The first case of bladder carcinoma treated by transmitting
white light transurethrally via a quartz rod following intravenous HpD was reported in 1976.7(739) Since
then, a large number of studies concerning PDT have been performed throughout the world using HpD
as the photosensitizer for a wide variety of malignant tumors, including gastric cancer (see also
Chapter 19: Photodynamic Therapy).
Fukutomi et al.8(740) studied specimens of stomach cancer exposed to argon laser light at a
wavelength of 514.5 nm and observed yellow autofluorescence in the cancerous tissue. These
researchers suggested that autofluorescence might be applied in the diagnosis of gastric cancer. We
began our studies with laser-induced fluorescence (LIF) for detection of gastric cancer in
1985.9,10(741) Prior to our studies, interest was growing in LIF for detection of malignant tumors,
usually with the administration of various fluorescent porphyrin compounds such as HpD.11,12(742)
Basic Principles of LIF (Autofluorescence)

Low-power laser illumination (excitation) can induce tissue fluorescence. The spectral characteristics
of LIF with and without prior administration of photosensitizing compounds have been found to be
different for normal and cancerous tissue, although the reason(s) for these differences is a matter of
considerable conjecture and debate.1315(743) The term autofluorescence is used here to denote the
fluorescence emitted by the tissue when it is excited by laser light without prior administration of HpD
or other photosensitizing compounds (Figure 181A). Thus, LIF in its broad sense also includes
autofluorescence (see Chapter 17: Biomedical Tissue Spectros-copy).
(744)Figure 181. Typical laser-induced fluorescence (LIF) spectra. A, LIF spectra of normal
tissue. B, Hematoporphyrin derivative (HpD) standard. C, LIF spectra of cancerous tissue.
If cancerous tissue is excited at a wavelength of 365 nm (supplied by a pulsed xenon ion laser) or
353.3 nm (pulsed third harmonic wave of a neodymium:yttrium-aluminum-garnet [Nd:YAG] laser) at 48
to 72 hr after intravenous injection of HpD, the LIF spectra from the tissue usually have characteristic
peaks near 630 nm and 690 nm (Figure 181B and 181C). No such peaks appear in the spectra of
corresponding normal mucosa or benign lesions (Figure 181A). Even without prior administration of
HpD, the autofluorescence spectrum from cancerous tissue appears to be the same spectroscopic
pattern as that seen after HpD (Figure 182). Moreover, as after HpD administration, no such
characteristic autofluorescence peaks are ever found in the spectrum of the corresponding normal
tissue. It has been hypothesized that the characteristic autofluorescence peaks from cancerous tissue
may originate from endogenous porphyrin compounds formed in the human body.16,17(745)
(746)Figure 182. Comparison of autofluorescence spectra at endoscopy of gastric cancer (C)

and adjacent normal tissue (A) with fluorescence spectra of HpD (B).
Endoscopic Spectroscopy
Instruments
Either a xenon ion laser or an Nd:YAG laser (third harmonic wave) can be used as a source of
excitation light (Figure 183).18(747) The pulsed xenon ion laser operating at wavelength 365 nm has
an output energy of 0.5 mJ per pulse, with a duration of 0.6 s (FWHM, i.e., full-width-half-maximum)
and a repetition rate of 10 Hz. The third harmonic wave of the Nd:YAG laser emitting at wavelength
353.3 nm has an output energy of 0.8 mJ per pulse, with a duration of 15 ns (FWHM) and a repetition
rate of 5 Hz.
(748)Figure 183. An endoscopy-compatible optical fiber spectroscopic system that utilizes

third harmonic wave of the neodymium:yttrium-aluminum-garnet (Nd:YAG) laser (at left).
A coaxial optical probe is used to transmit laser light and to collect emitted fluorescent light. This has a
coaxial design with a single central quartz fiber for transmission of the laser light. Arranged around the
central fiber is a bundle of glass fibers for collection of the fluorescent light. The optical probe is 2.2
mm in diameter and can be inserted through the accessory channel of an endoscope and into the
stomach cavity.
Normal tissue emits strong fluorescence in the blue or green region of the spectrum. To demonstrate
clearly the fluorescence that is characteristic of the cancer, it is necessary to use a cut-off filter to block
fluorescent emission below a wavelength of 550 nm. An optical multichannel analyzer is used for signal
detection. The monochromator is fixed without scanning and with the exit slit removed (Figure 184).
Fluorescent light is collected by the optical system and directed to the optical multichannel analyzer
system. The entire fluorescent light distribution for every laser pulse is captured. The spectra are
stored in a computer and can be displayed on a television monitor or printed out on paper. The firing of
the laser, synchronized sampling, and data storage are all controlled by a computer program.
(749)Figure 184. Block diagram of a system used to collect fluorescence spectra of gastric
mucosa or cancer in vivo during gastroscopy.
Technique
Pharmacologic preparation for endoscopic LIF or autofluorescence spectroscopy is the same as that
for routine endoscopy. Before endoscopic spectroscopy is performed, the spectrum and intensity of an
HpD standard are measured. The patient lies in the left decubitus position during insertion of the
endoscope. However, the patient's position may be changed as required to facilitate targeting. During
spectroscopic investigation, the distal end of an optical probe should be aimed perpendicularly at the
target at a distance of 2 to 3 mm. Therefore, endoscopic spectroscopy is usually most easily performed
when the target lesion is located in the antrum rather than in the fundus or cardia of the stomach.
Spectroscopic Diagnosis
Autofluorescence Spectroscopy
Fluorescence intensity peaks in the spectra of gastric cancers are often found in the regions near 630
nm and 690 nm. These peaks are not seen in surrounding normal tissue. These characteristic
autofluorescence peaks have been observed in 75% of patients with gastric cancer.19(750) However,
the detection rate for the characteristic peaks also correlates with the location of the cancer. When the
tumor is in the antrum, it is relatively easy to obtain the desired perpendicular orientation at 2 to 3 mm
distance of the optical probe from the target. The approach to lesions in the body or cardia of the
stomach is more difficult. Therefore the detection rate of the characteristic peaks is likely to be higher
when the cancer is in the antrum.
Autofluorescence spectra can be recorded within 10 sec during endoscopy. The procedure is
considered to be a fast, safe, and simple means for detecting gastric cancer, particularly as a guide to
biopsy sites when the lesion is suspected to be malignant.
LIF Spectroscopy Following HpD Administration

HpD remains the most common photosensitizer used in the diagnosis of malignant tumors. After
intravenous injection, HpD is taken up selectively by malignant tumor and retained with a 10-fold
greater concentration than in normal tissue. A small dose of HpD (0.5 to 1 mg/kg body weight) is
usually given intravenously 48 to 72 hr before endoscopic LIF spectroscopy is performed. HpD has
optimum tumor-localizing properties and very few undesirable side effects when a small dose is given
intravenously. As with autofluorescence, excitation light from a xenon ion laser or Nd:YAG laser (third
harmonic wave) can be used to induce fluorescence in normal or diseased tissue. The advantage of
LIF following administration of a small dose of HpD is that the intensity of fluorescence spectra of the
cancerous tissue at 630 nm and 690 nm is stronger and more conspicuous.
Spectroscopic Treatment of Gastric Cancer

Spectroscopic treatment or PDT, in which the photosensitizer HpD localized in malignant tissue is
activated in vivo by red light, has been shown to be effective against a number of malignant tumors
such as cancers of the skin, urinary bladder, and gastrointestinal tract.2022(751) The basic theory
and principles of PDT are discussed in Chapter 19: Photodynamic Therapy.
The most commonly used photosensitizers are HpD and dihematoporphyrin ether (DHE, Photofrin II).
The dosages used for treatment are considerably higher than those used for diagnosis. The dosage for
HpD is 3 to 5 mg/kg body weight and that for DHE is 2 mg/kg body weight. After intravenous injection
of HpD, the porphyrin compound is taken up by all tissues to an equal extent at first. During the
following 48 to 72 hr, the porphyrin compound is retained in the cancerous tissue while it is cleared
from normal tissue. When the tumor containing the photosensitizer HpD is exposed to red light from an
argon dye laser (wavelength 630 nm) or gold vapor laser (wavelength 628 nm), rapid necrosis of
cancer cells ensues. Red light is used for PDT because it penetrates tissue better and more deeply
than other wavelengths. Tumor destruction occurs very rapidly, but the effects of PDT persist for a
relatively long period of time. Therefore, the overall result cannot be assessed until 4 weeks after
therapy.
Endoscopic PDT is performed at 48 to 72 hr after intravenous injection of HpD or DHE. Premedication
includes diazepam and meperidine. We use the argon dye laser (wavelength 630 nm) as the excitation
source. The laser beam is focused on a quartz fiber of 400 m in diameter; the output power at the
distal end of the fiber is 500 to 800 mW. The quartz fiber is inserted into the gastric lumen through the
accessory channel of the endoscope. The distance between the distal end of the quartz fiber and the
target is maintained at 1 to 2 cm, and each part of the cancerous lesion is exposed to the red light for
15 to 20 min. Cylinder quartz diffuser fibers, 1.5 cm or 2.0 cm in length, that provide a 360-degree field
of irradiation are used to treat tumors of the esophagus and gastric cardia.
As irradiation with red light proceeds, the cancerous tissue rapidly becomes edematous and hyperemic
(Figure 185). The edema is transitory and subsides after 48 hr. Necrosis and slough can be seen
subsequently. Depending on the response of the tumor, the need for further PDT can be considered
after about 48 to 72 hr. The disadvantage of PDT is skin photosensitization that makes it necessary for
the patient to avoid exposure to sunlight for 4 to 8 weeks after treatment. PDT can be repeated after 3
to 4 weeks, but the dosage of HpD or DHE should be reduced.
(752)Figure 185. A and B, Endoscopic views of edematous and hyperemic mucosal tissue
reaction during photodynamic therapy (PDT).
Clinical Studies of PDT

PDT has been used to treat patients with obstruction of the esophagus or gastric cardia owing to
advanced, inoperable cancer (see Chapter 19: Photodynamic Therapy, for further discussion of clinical
applications). Improvement in the ability to swallow is achieved in most patients. In a study of 52
patients by Jin et al.,23(753) 77% of upper gastrointestinal tract tumors responded favorably to PDT.
We have had similar good results with PDT of obstructing cancer of the cardia of the stomach.24(754)
Hayata et al.25(755) reported the results of PDT in 16 patients with early gastric cancer. Surgical
resection was performed after PDT in 12 patients. No evidence of residual tumor was found in 5
resected specimens, but residual cancer was discovered by serial section in the other 7. Surgery was
not performed for various reasons in 4 patients. Two of these patients died of causes unrelated to
cancer, 1 at 13 months and 1 at 5 months after PDT. Although initial remission had been achieved in
these 2 patients, evidence of recurrence was seen in both (1 by endoscopic biopsy, 1 at autopsy). In a
third patient followed for 31 months, a recurrence was discovered at 27 months after PDT. A fourth
patient was apparently disease free at 30 months after PDT.
Tajiri et al.26(756) treated 13 patients with early gastric cancer by PDT. Eight of the patients had Type
IIc lesions, 4 had Type IIa, and 1 had Type I. The cancers were 2 cm or less in diameter in 10 patients.
Five patients were initially treated by photoablation of the tumor using the Nd:YAG laser. Surgical or
autopsy specimens were available from 4 patients, and residual tumor was found in 2, both of these
being greater than 2 cm in diameter at initial diagnosis. Ten patients who did not undergo surgery were
followed by serial endoscopy procedures and biopsies for periods of 4 to 28 months. No evidence of
recurrent cancer was found in 9 patients who remained alive.
We have performed PDT in 21 patients with gastric cancer, 13 of which were located in either the
antrum or the body of the stomach. In 8 patients, the cancer arose in the cardia. Dysphagia was
relieved by PDT in all cases. Early gastric cancer was found in 3 patients. One patient with a Type IIc
early cancer complained of abdominal pain and passed tarry stools for 2 days immediately after PDT.
An active gastric ulcer was found at endoscopy on the lesser curvature of the stomach 1 week later
(Figure 186). This patient subsequently underwent surgical resection, and histologic examination of
the specimen showed cancerous cells in tissue surrounding the ulcer.
(757)Figure 186. A, Endoscopic view of an active ulcer after PDT treatment of an early
gastric cancer (Type IIc). B, Macroscopic appearance of the lesion.
Song et al.22(758) reported on 23 patients with gastric cancer treated by PDT. A partial response was
achieved in 10 and no response was seen in 7 patients. Hu et al.27(759) treated 42 patients with
carcinoma of the gastric cardia; they found that adenocarcinoma and mucinous adenocarcinoma
responded well to PDT, but papillary adenocarcinoma and squamous cell carcinoma were refractory.
The treatment was effective in patients with limited and superficial cancerous involvement in the gastric
wall but unsatisfactory in those with wider and deeper infiltration.
Although PDT has proved to be effective in the treatment of many types of solid malignant tumors,
including cancer of the esophagus and gastric cardia, data on its use remain relatively limited.28(760)
Although PDT appears to have a place in the palliative management of patients with obstructing
tumors, its role in the treatment of early-stage lesions remains uncertain, and further research in this
area is needed. For the present, therefore, PDT remains an investigative procedure for this indication.
Acknowledgments
The author is grateful for the cooperation of Prof. De-Zhong Zhang, Prof. Qing Zhu, Dr. Su-Ming Wu,
and Ms. Hong-Yu Luo; and to Mr. Yen-Ming Ye, Mr. Yuan-Long Yang, and Mr. Jing-Fang Xia from the
Department of Physics of Shanghai Fu-Dan University for technical assistance.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Policard A. tudes sur les aspects offerts par des tumeurs experimentales examinees la
Auler H, Banzer G. Untersuchungen ber die Rolle der Porphyrine bei Geschwulstkranken
menschen und tieren. Z Krebsforsch 1942;53:658.
Figge FHJ, Weiland GS, Manganiello LOJ. Cancer detection and therapy. Affinity of
neoplastic embryonic and traumatized tissues for porphyrins and metalloporphyrins. Proc Soc
Exp Biol Med 1948;68:6401.
Lipson RL, Baldes EJ, Olsen AM. The use of a derivative of hematoporphyrin in tumor
detection. J Natl Cancer Inst 1961;26:18.
Kyriazis GA, Balin H, Lipson RL. Hematoporphyrin derivative fluorescence test colposcopy
and colpophotography in the diagnosis of atypical metaplasia, dysplasia and carcinoma in situ
of the cervix uteri. Am J Obstet Gynecol 1973;117:37580.
Dougherty TJ, Grindey GB, Fiel R, et al. Photoradiation therapy II. Cure of animal tumors with
hematoporphyrin and light. J Natl Cancer Inst 1975;55:1159.
Kelly JF, Snell ME. Hematoporphyrin derivative: A possible aid in the diagnosis and therapy of
carcinoma of the bladder. J Urol 1976;115:1501.
Fukutomi H, Kawakita I, Sakita T. New method of gastric cancer detection by argon laser. In
Joffe SN (ed). Neodymium-YAG Laser in Medicine and Surgery. New York: Elsevier Science,
1983;1748.
Luo HY, Xiao SD, Li RR, et al. Porphyrin concentrations in implanted S-180 sarcoma in mice
and its various organs by laser-induced fluorescence spectroscopy. Chin J Optoelec Med
1993;2:1720.
Xiao SD, Zhang DZ, Li RR, et al. A study of fluorescent substance in human gastric cancer
tissue. Chin J Dig 1993;13:1413.
Doiron DR, Profio E, Vincent RC, et al. Fluorescence bronchoscopy for detection of lung
cancer. Chest 1979;76:27.
Cortese DA, Kinsey JH, Woolnen LB, et al. Clinical application of a new endoscopic technique
for detection of in situ bronchial carcinoma. Mayo Clin Proc 1979;54:63542.
Kapadia CR, Cutruzolla FW, O'Brien KM, et al. Laser-induced fluorescence spectroscopy of
human colonic mucosa. Detection of adenomatous transformation. Gastroenterology
1990;99:1507.
Schomacker KT, Frisoli JK, Compton CC, et al. Ultraviolet laser-induced fluorescence of
colonic tissue: Basic biology and diagnostic potential. Lasers Surg Med 1992;12:6378.
Cothren RM, Richards-Kortum R, Sivak MV Jr, et al. Gastrointestinal tissue diagnosis by
1990;36:10511.
Ye YM, Yang YL, Xia JF, et al. Single pulse cancer diagnosis by laser excited
autofluorescence. Chin J Lasers 1987;14:2413.
Yang YL, Ye YM, Li FM, et al. Characteristic autofluorescence for cancer diagnosis and its
origin. Lasers Surg Med 1987;7:52832.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Wu SM, Xiao SD, Luo HY, et al. Autofluorescence induced by third harmonic wave of Nd:YAG
laser in the diagnosis of gastric cancer. Lasers Med 1993;1:79.
Xiao SD, Wu SM, Luo HY, et al. Autofluorescence excited by endoscopic xenon ion laser in
the diagnosis of gastric cancer. J Shanghai 2nd Med Univ 1990;4:726.
Dougherty TJ, Lawrence G, Kaufman JH, et al. Photoradiation in the treatment of recurrent
breast carcinoma. J Natl Cancer Inst 1979;62:2317.
McCaughan JR, James S. Progress in photoradiation treatment of cancer following
administration of HpD. Laser Focus 1983;19:4856.
Song SZ, Li JH, Zou J, et al. Hematoporphyrin derivative and laser photodynamical reaction in
the diagnosis and treatment of malignant tumors. Lasers Surg Med 1985;5:616.
Jin ML, Yang BQ, Li R, et al. Analysis of hematoporphyrin derivative and laser photodynamic
therapy of upper gastrointestinal tumors in 52 cases. Lasers Med Sci 1986;2:514.
Zhang DZ, Xiao SD, Zhu Q, et al. Assessment of photodynamic therapy in the treatment of
gastric cancer and gastric precancerous lesions. Acta Shanghai 2nd Med Univ 1988;8:2246.
Hayata Y, Kato H, Okitsu H, et al. Photodynamic therapy with hematoporphyrin derivative in
cancer of the upper gastrointestinal tract. Semin Surg Oncol 1985;1:111.
Tajiri H, Daikuzono N, Joffe SN, et al. Photoradiation therapy in early gastrointestinal cancer.
Hu C, Li JY, Jin ML, et al. Pathologic study on treatment effect of hematoporphyrin derivative
in carcinoma of gastric cardia. Chin J Tumor 1988;10:1979.
Hayata Y, Kato H, Konaka C, et al. Photoradiation therapy in early stage cancer cases of the
lung, esophagus, and stomach. In Andreoni A, Cubeddu R (eds). Porphyrins in Tumor
Phototherapy. New York: Plenum Press 1984;40512.
Chapter 19 Photodynamic Therapy

(761)
(762)
STEPHEN K. HEIER, M.D., F.A.C.P., F.A.C.G.
Photodynamic therapy (PDT) is effective in treating a variety of tumors. PDT combines a relatively
non-toxic chemotherapy with new laser delivery systems and endoscopy (see Chapter 18:
Spectroscopic Diagnosis and Treatment of Gastric Cancer). Tumors pharmacologically sensitized to
specific wavelengths of light undergo necrosis as a result of photochemical reaction. After years of
development, an understanding of the mechanism of action of PDT is emerging, and an awareness of
the utility of this form of therapy is growing.
History
Nearly a century has passed since fluorescent and cytotoxic properties of drugs were first correlated
with acridine dye1(763) and eosin.2(764) Policard3(765) identified porphyrins as possible
photosensitizers in 1924. Hematoporphyrin, a breakdown product of hemoglobin that is devoid of iron
atoms, was injected into rats with implanted tumors by Auler and Banzer in 1942.4(766) The tumors
demonstrated the salmon red fluorescence characteristic of porphyrin sensitization. In 1960, Lipson
and Baldes5(767) prepared hematoporphyrin derivative (HpD), a chemical mixture resulting from
acetylation and alkalization of hematoporphyrin. Tumor localization was improved with the use of HpD,
which produced fluorescent activity that could be visualized in human tumors, including those viewed
endoscopically.68(768) HpD was used with some success in 1966 by Lipson et al.9(769) to treat a
large, fungating, and ulcerated breast tumor.
Dougherty et al.10(770) refined HpD by isolating its potent fraction in 1983. The major components of
this refined mixture, named dihematoporphyrin ethers (DHE, porfimer sodium, Photofrin), are esters
and ethers of hematoporphyrin.11(771) The other porphyrin components of HpD, abundantly present
but producing minimum toxicity, do not bind to tumor cells and do not confer tumor sensitization when
injected into animals.10,12(772) Research before clinical trials used HpD for the most part, but clinical
experience is largely based on the use of DHE.
Pharmacokinetics
The selective nature of PDT (i.e., tumor selectivity) results from longer retention of the photosensitizing
drug in tumor compared with normal tissue. A delay between intravenous drug injection and laser light
application optimizes differential tissue effects. Transplanted tumors in mice maintain high drug
concentrations for at least 72 to 96 hr, in contrast to muscle tissue, which clears HpD by 24 hr.13(773)
Differential tissue effects have also been demonstrated throughout the gastrointestinal tract in several
animal models.14,15(774) Significant drug retention occurs in the reticuloendothelial cells of the liver,
spleen, and kidney and in skin.13,16(775) Only skin retention imparts toxicity (i.e., skin
photosensitivity); nevertheless, a differential response exists between skin tumors and adjacent normal
skin.17,18(776)
In aqueous solutions, including serum, DHE forms aggregates that are not photoreactive until
disaggregation.19,20(777) It has been postulated that the leaky neovasculature of tumors allows influx
of large aggregates of DHE16(778) and that disaggregation occurs on tight cellular binding to tumor
cells.21(779) DHE has an affinity for low-density lipoprotein (LDL) in serum.22(780) Tumor cells
possess elevated LDL receptor activity,2325(781) which results in differential tumor endocytosis of
the LDL-DHE complex.26,27(782)
Laser Activation
Absorption of light by HpD and DHE is maximum in the violet range of the spectrum at a wavelength of
about 405 nm.28,29(783) This wavelength stimulates maximum fluorescence from the surface of a
tumor. The wavelength used for therapy is 630 nm (red), which is a minor absorption peak of HpD.
This wavelength was selected as a compromise between that of maximum drug absorption and the
greater tissue penetrance of longer visible wavelengths.
Mechanism of Action
Photons of laser light absorbed by DHE can result in fluorescent photon emission or, more importantly,
energy transfer to ground state triplet oxygen with conversion to reactive singlet oxygen.30,31(784) In
animal models, decreasing the availability of the oxygen substrate for this pivotal photochemical step,
as by producing ischemia in targeted tissue, reduces or abolishes photodynamic
destruction.3236(785) Tissue destruction results from singlet oxygen-mediated oxidation of cellular
constituents, including enzymes3641(786) and structural proteins.40,42,43(787) Damage is
particularly severe to the plasma membrane,40,42,43(788) with marked cell blebbing, increased
permeability, and inhibition of membrane transport processes, and to the
mitochondria,3639,41,4447(789) with inhibition of cytochrome c oxidase and reduction of cellular
ATP levels. Whether hypoxic areas of human tumors are resistant to these effects remains to be
elucidated.
Photodynamic damage to neovasculature, with secondary tumor necrosis, is at least contributory and
perhaps central to the mechanism of action. Henderson et al.48(790) demonstrated that rapidly
progressive cell death after PDT can be prevented by immediately excising animal tumors and
explanting tumor suspensions to a cell culture system. Although it has been reported that tumor cells
grown in culture take up greater amounts of HpD than normal cells,49(791) this finding is the
exception.50,51(792) Using autoradiography with 3H-labeled HpD in mice, Bugelski et al.16(793)

determined that most HpD localizes in the neovascular stroma instead of the tumor parenchymal cells.
Electron microscopy revealed early bleb changes in perivascular tumor cells, followed by endothelial
damage, red blood cell extravasation, and generalized tumor destruction.52(794) Vascular
compromise47,53,54(795) leads to tumor hypoxia within an hour after PDT.55,56(796)
Although hypothesized,57(797) confirmatory support for an immunopotentiating mechanism of action is
scant. Transient increases in the production of tumor necrosis factor after in vitro PDT have been
demonstrated in murine macrophages obtained by peritoneal lavage.58(798) Increased macrophage
activation occurs in animals after low-dose photoactivation.59(799) Chronic inflammatory infiltrates
have been associated with prolonged responses after PDT in transitional cell carcinoma of the human
urinary bladder,60(800) and transient increases in various urinary cytokines (e.g., interleukin-1,
interleukin-2, and tumor necrosis factor) have been measured in three patients.61(801) The occasional
confluent inflammatory stricture after esophageal PDT and the longer duration of response compared
with neodymium:yttrium-aluminum-garnet (Nd:YAG) laser therapy observed by our group may be
compatible with an immunopotentiation mechanism. Conversely, animal contact hypersensitivity is
suppressed after intraperitoneal62(802) and cutaneous PDT,63,64(803) and transient decreases in
splenic natural killer cell activity and spleen size occur after hindleg PDT.65(804)
Equipment
Filtered light has been used for certain PDT applications.18,66(805) However, despite the relatively low
energy levels needed for drug activation, coupling energy losses associated with transendoscopic fiber
placement necessitate the use of lasers. The argon dye laser is an organic dye laser consisting of an
argon gas laser that optically pumps a liquid dye jet stream; the resonating output wavelength is tuned
by the operator by turning a birefringent filter (Figure 191). The spectrum of wavelengths obtained
depends on the dye selected for loading into the laser. Generally, the output wavelength is confirmed
by a spectrometer (Figure 192) and the output power by an integrating sphere. The
potassium-titanyl-phosphate (KTP) dye laser is similar to the argon dye laser, except that pulsed
Nd:YAG laser output, frequency doubled after passing through a KTP crystal, is used to pump the dye
laser. An alternative laser system occasionally used for PDT is the gold vapor laser, although biologic
equivalence is controversial.67,68(806)
(807)Figure 191. Blue-green light from an argon laser is seen bouncing between mirrors in a
dye module, with the path perpendicular to a jet stream of dye (center arrow). The resonating
wavelength (red) is selected by turning a diffraction grating (right arrow) placed along the
path of the uppermost beam of light.
(808)Figure 192. Confirmation of the output wavelength of an argon dye laser is

accomplished by shining the fiber output into a spectrometer and looking through its eye piece
at an interior calibration scale.
The types of delivery fibers (Figure 193) available for endoscopic transmission of light include a
microlens tip variety that provides a uniform spot circle and the more commonly used genre with a
cylinder diffusing tip that provides circumferential light delivery over tip lengths ranging from 1.0 to 5.0
cm. The cylinder diffusing fiber can also be implanted into a tumor mass to provide interstitial light
delivery.
(809)Figure 193. Fibers for endoscopic photodynamic therapy (PDT). A, Microlens fiber for
forward surface illumination provides a uniform beam spot. B, Cylinder diffusing fiber for
diffuse illumination of a cylindrical cavity.
Light Dosimetry
We use the following nomenclature to describe light dosimetry. Tissue dose is the total amount of
energy received per unit surface area of tissue treated (reported in units of Joules per centimeter
square [J/cm2]). Energy density is the amount of energy per second received per unit surface area of
tissue (mW/cm2). Light dose is the total amount of energy delivered from the fiber tip (measured as
J/cm fiber in the case of a cylinder diffusing fiber). Light dose delivery rate is the amount of energy per
second delivered from the fiber tip (milliwatts [mW], measured as mW/cm fiber in the case of a cylinder
diffusing fiber). Power setting is the laser unit output per second (mW).
An example can illustrate the use of these terms in dosimetric calculations. To deliver a light dose of
300 J/cm from a 2.5-cm-long cylinder diffusing fiber at a light dose delivery rate of 400 mW/cm, a
power setting of 1000 mW is selected (400 mW/cm 2.5 cm) and maintained for 750 seconds (300
J/cm 400 mW/cm). A 5-cm-long esophageal tumor could then be treated in two sequential segments
with the diffusing fiber.
Low light dose delivery rates are chosen to avoid thermal effects. For esophageal tumors, my
colleagues and I analyzed the depth of tumor necrosis for individual tumor segments after 77 light
applications delivered during 44 separate courses of therapy.69,70(810) Light doses were varied from
200 to 600 J/cm fiber, resulting in a range of tissue doses that depended on luminal diameter. A
significant correlation was found between the tissue dose and the resultant depth of necrosis (r = .67, p
< .001). Based on this experimentation, a standardized set of treatment parameters was selected: 300
J/cm light dose, 400 mW/cm delivery rate (treatment time 750 sec). The resultant range of tissue
doses provides significant necrosis in tightly stenotic tumor segments but ensures minimum necrosis in
segments with more patent lumens, minimizing the risk of fistula formation. The higher light doses
previously used by others probably accounted for an unfavorable frequency of complications, including
fistulas and pleural effusions.71,72(811)
Results
Curative Therapy
Clinical studies throughout the world have focused on a potential role for PDT in the management of
early gastrointestinal cancer. In a report from China, 12 of 13 early esophageal squamous cell cancers
(ESCC) detected by balloon cytology screening with endoscopic confirmation were undetectable after
PDT over a mean follow-up of 24.7 months (range 21 to 32 months).73(812) The 1 patient in whom
residual cancer cells were detected by balloon cytology at 3 days after therapy subsequently underwent
endoscopy and surgical resection, and no residual cancer was found. Complete responses were
achieved in 12 of 14 patients with ESCC in a multi-institutional study from Japan with a 2-year mean
follow-up (range 7 to 48 months); 2 patients in whom tumor recurred responded to repeated
therapy.74(813) Complete responses were achieved in 12 of 15 patients with ESCC in Switzerland
(mean follow-up 20.1 months; range 0.5 to 5 years) who were identified by screening patients with
malignancies of the head and neck region for second primary cancers.75(814) In another study from
France, a complete response was initially achieved in 32 of 37 patients with ESCC; this was sustained
in 28 patients over periods of 2 to 40 months, with tumor recurrences in 4 patients.76(815) However,
many of these patients received adjunctive irradiation and chemotherapy. A complete response was
achieved in seven patients with esophageal adenocarcinoma who underwent PDT in one of four
centers in the United States. Of these seven patients, four were treated as part of larger series in which
the primary intent was palliation; one had residual cancer later detected at autopsy at 30
months,77,78(816) one had Barrett's esophagus with early adenocarcinoma,79(817) and two had
Barrett's esophagus with carcinoma in situ.80(818) In the latter three patients, partial79(819) or
complete80(820) ablation of Barrett's mucosa occurred simultaneously.
As a result of these reports, attention has been directed toward PDT of Barrett's and associated
dysplasia. Wang et al. used low-dose PDT with HpD (at a drug dose of 1.51.75 mg/kg instead of 3
mg/kg; light dose of 150200 J/cm instead of 300 J/cm) to treat 24 patients with Barrett's, and
compared them with 11 control patients.80a(821) Both groups were maintained on omeprazole. After a
mean follow-up of 25 weeks, only the PDT group demonstrated regression, with a decrease in the
linear extent from 7 cm to 5 cm. In addition, 29% of the PDT patients (7 patients) had a 4-cm or more
decrease in the length of their Barrett segment. Mild sunburn was the most common complication,
occuring in 7 of the 24 patients.
Overholt and Panjehpour used PDT with either a bare cylinder diffusing fiber or a centering balloon, to
treat Barrett's patients with either T2 cancer (2), Tis-T1 cancer (12), high-grade dysplasia (23), or
low-grade dysplasia (7).80b(822) Twenty-one dysplasia patients were treated once, 7 required 2
sessions, and 2 required 3 sessions. Patients were maintained on omeprazole 20 mg twice a day for 3
months, and every day for 9 months. With a follow-up of 660 months, cancer was eliminated in all 14
patients, high-grade dysplasia in 19 of 23 (4 patients reduced to low-grade dysplasia), and low-grade
dysplasia in 6 of 7 (one progressed to high-grade dysplasia). Of import, approximately 75 to 80% of
treated areas healed with squamous mucosa (i.e., with regression of Barrett's). Seventeen patients
developed strictures that responded to dilatation.
Our group is currently analyzing new treatment time parameters for PDT of Barrett's esophagus. These
parameters are aimed at accomplishing same-day injection and endoscopic therapy, instead of the 2day waiting period used during the palliative esophageal cancer trial (see later). The earlier treatment
time should allow use of a lower drug dose, because of enhanced tissue effects soon after injection
when tissue and serum drug levels are higher. Skin sensitivity will thereby be minimized because of the
reduced drug dose administered. Treatment times selected for analysis were derived from a canine
study of Photofrin levels, and initial clinical results have been promising (unpublished data). However,
achieving the optimum degree of tissue damage may require fine-tuning of drug and light doses.
Most reports of PDT for early gastric cancer have originated in Japan and China (see Chapter 18:
Spectroscopic Diagnosis and Treatment of Gastric Cancer). In several published series, complete
responses have been achieved as follows: 28 of 30 patients (follow-up >1 year for all patients),81(823)
8 of 14 patients (mean follow-up 19 months, longest follow-up 35 months),82(824) 11 of 13 patients
(follow-up 4 to 28 months),83(825) 9 of 16 patients (mean follow-up 17 months, range 5 to 31
months),84(826) and 4 of 5 patients (mean follow-up 4 months).15(827) As stressed by Xiao in the
preceding chapter, these outcomes and those of his group should be viewed with caution pending
studies with longer follow-up.
Nonmetastatic rectosigmoid cancers less than 4 cm in diameter were treated using PDT in a French
study. Complete responses were achieved in 10 of 21 patients (mean follow-up 21.7 months for those
who responded, 32.6 months for 6 patients not dying of other causes).85(828) In a series from
England of 10 patients with colon cancer less than 5 cm long (4 patients with metastatic disease)
treated by PDT, a complete response was achieved in 2 patients.86(829) One of these 2 patients had
an anastomotic recurrence (follow-up 28 months), and the other had a tumor documented by
endoscopic ultrasound to be transmural despite prior Nd:YAG laser therapy (follow-up 20 months).
There is also a report of eradication by PDT of sessile colorectal villous adenomas in three of five
patients.87(830)
Palliative Therapy
PDT using light at 630 nm offers certain advantages and disadvantages because of the limited tissue
penetration of this wavelength. Although this restricts the use of PDT for control of bulky tumors in
many areas of the body, it may be ideal for defining selective necrosis within the boundaries of the
gastrointestinal tract.
At New York Medical College, 44 courses of PDT were delivered in 32 patients with malignant
esophageal obstruction.70(831) Before therapy, 5 patients were unable to swallow their saliva, 16 had
dysphagia with liquids, 18 experienced dysphagia with soft foods, and 5 patients had difficulty only with
solid food. Baseline minimum lumen diameters were 3.5 2.6 mm. Light applications were delivered 2
days after intravenous DHE and were repeated 2 to 4 days later if necessary. At 1 week, mean
dysphagia grades increased almost two full levels, and lumen diameter increased by 9.9 2.6 mm. At
1 month, the Karnofsky performance status and esophageal grades had improved significantly, and the
weights of most patients remained stable or decreased modestly. By multiple parameters, my
colleagues and I were unable to detect a difference in specific activity of PDT against adenocarcinoma
(n = 15) or squamous cell carcinoma (n = 17).70,88(832) However, the more distal location of the
adenocarcinomas may have accounted for a somewhat longer duration of response (108 vs. 75 days),
although this difference was not statistically significant. In another study of 40 similar patients,77(833)
some of whom received concurrent Nd:YAG laser therapy initially,71,89(834) similar improvements in
dietary tolerances and esophageal grades occurred.
Of 10 patients with colonic adenocarcinomas treated in the previously described study from England, 6
patients experienced an improvement in symptoms, with cessation of bleeding and abdominal pain and
regularization of bowel habits.86(835)
Forty-two patients were entered into our randomized trial comparing palliative PDT with Nd:YAG laser
therapy for treating obstructing esophageal cancer.70(836) PDT patients had a greater increase in
their Karnofsky performance status and esophageal grades at 1 month (p < .001 and p = .002,
respectively) and a longer duration of response (p = .008). Complications of therapy (PDT vs. YAG)
included fistulas (1 vs. 2), strictures (0 vs. 2), and skin photoreactions (4 vs. 0).
Lightdale et al. reported results of the multicenter randomized trial comparing palliative PDT with
Nd:YAG laser therapy for the treatment of 236 patients with obstructing esophageal cancer.89a(837)
The light dose used (300 J/cm fiber) was the standardized dose indicated by our light dosimetry trials.
The authors concluded that although temporary photosensitivity is a limitation, PDT is equally effective,
easier to administer, and associated with fewer acute perforations than Nd:YAG laser therapy.
Complications
Treatment-related problems and toxicity encountered in our first 32 patients undergoing PDT (Table
191) included luminal plugging by necrotic tumor tissue that necessitated additional endoscopic
management in 9 patients; transient low-grade fever in 8 patients; 10 instances of limited first-degree
sunburns in 6 patients (mean 21 days after DHE, range 6 to 72 days); 2 delayed strictures
characterized by confluent inflammatory exudates, occurring after multiple treatment courses (three
DHE injections each); and 1 fistula.
Complications of Photodynamic Therapy for

32 Patients With Malignant Obstruction Treated at Westchester
County Medical Center, Valhalla, New York
TABLE 191
COMPLICATION
NO. OF PATIENTS (%)
Luminal plugging
9 (28.1)
Complications of Photodynamic Therapy for

32 Patients With Malignant Obstruction Treated at Westchester
County Medical Center, Valhalla, New York
TABLE 191
COMPLICATION
Fever
Sunburn/erythema
Stricture
Fistula
NO. OF PATIENTS (%)

8 (25.0)
6 (18.8)
2 (6.3)
1 (3.1)
The most frequent toxic effect reported by investigators is sunburn. Photobleaching, a process by
which the drug itself becomes oxidized after light exposure through drug-induced singlet oxygen
generation, may paradoxically shorten the duration of photosensitivity in patients least compliant with
sun exposure restrictions.90(838)
Higher light doses than my colleagues and I have been using increase the frequency of significant
complications. In one study of 14 patients with malignant esophageal obstruction who received a mean
light dose of more than 500 J/cm, 3 patients suffered mediastinitis, 2 developed fistulas, 2 patients
died, 9 had febrile responses, and 1 developed a delayed stricture.72(839)
Limitations
Photodynamic-mediated cytotoxicity requires drug interaction with light and oxygen. Impediments to
light transmission, such as gastric rugae, peristalsis, and tumor ledges, may prevent adequate
photoactivation; deeply seated tumor cells and hypoxic regions may be inaccessible to photoactivation.
It is certain that the irregular geometry of tumors and secretions or blood covering the tumor surface
render light dosimetry calculations imprecise. Because the region of photodynamic destruction is
defined by local light penetration, the benefits and risks of adjunctive systemic therapies require
assessment.
Skin photosensitivity, although requiring temporary precautions against direct sunlight for about 1
month, has not been a major deterrent for patients seeking potentially curative therapy and has not
resulted in major lifestyle changes for the elderly, the debilitated, or even most of the patients who are
actively employed. It is occasionally necessary to exclude a patient with an outdoor occupation or
leisure activity from consideration for PDT. Drugs producing less skin photosensitivity are being
evaluated.
The dye laser, once an anathema to the clinician without a team of engineers, has undergone
continuous refinement. Technical difficulties do occur, and all units require regular maintenance (e.g.,
changing the dye after it has undergone degradation), but these problems are not insurmountable. The
price of the improved dye lasers is somewhat greater than that of the Nd:YAG laser, although
incremental cost is limited because these lasers also provide argon, KTP, or Nd:YAG wavelength
capabilities.
Advantages
The advantages of PDT derive from the selective and nonthermal nature of its photoreaction with
tumors. In combination, these two properties make diffuse illumination a safe and effective method of
delivering activating light. These properties may enhance the curative potential of PDT in early cancer,
because microscopic clusters of malignant cells can be selectively destroyed in areas of tissue that
appear macroscopically to be normal. The preservation of submucosal collagen maintains intestinal
integrity even after full-thickness necrosis, as documented by histologic analysis and bursting
pressures in experimental studies.91(840) These responses to PDT contrast sharply with the
indiscriminate effects of thermally induced tissue damage.92,93(841)
For the treatment of advanced cancer, diffuse illumination simplifies light application and obviates the
need for dozens of precisely aimed pulses of laser energy with Nd:YAG laser therapy. Because normal
tissue adjacent to tumor contains lower drug concentrations, injury is negligible compared with damage
that may occur because of operator error during thermal destruction. Within the tumor lumen,
excavated areas are exposed to lower tissue doses of light than constricted segments, and this
augments the margin of safety for PDT.
Unlike Nd:YAG laser therapy, the nonthermal light delivery used in PDT does not result in the high
temperatures that can cause discomfort for patients and damage endoscopes. Because smoke is not
generated during PDT, endoscopic visualization is not obscured, as occurs frequently with Nd:YAG
laser therapy. Personnel are not exposed to noxious smoke, and the laser light itself, both visible and
of lower intensity, represents less of a safety hazard.
PDT or Nd:YAG laser therapy can be repeated when tumor regrowth produces recurrent luminal
obstruction. The duration of response, however, appears to be longer after PDT.
PDT is currently approved for the treatment of esophageal cancer that is not amenable to or is difficult
to treat with Nd:YAG laser therapy.70,89a,93a(842) These circumstances may include completely
obstructing tumors, cervical tumors, tumor ingrowth/overgrowth of metal stents, tumors at the
gastroesophageal junction, long tumors, tumor recurrence after radiation therapy, and submucosal
tumors.
Summary
PDT is unique in that it is laser therapy without associated thermal effects, and it provides selective
chemotherapy without associated systemic toxicity (other than transient photosensitivity). It is safe and
effective as an alternative for the palliative treatment of cancerous luminal obstruction. PDT can
achieve complete responses in patients with early cancer, although the frequency and duration of
these responses remain uncertain. Studies with periods of longer follow-up than those of existing
investigations are required to accurately define the curative potential of PDT. Preferably, these should
include patients who are carefully selected according to strict imaging criteria for staging, including
endoscopic ultrasonography. Controlled trials should assess the curative role of PDT in addition to its
palliative function.
New tumor-sensitizing agents and light delivery devices are being developed. A second generation of
sensitizing agents undergoing evaluation includes substituted porphyrins, phthalocyanines,
napthalocyanines, chlorins, and aminolevulinic acid.94(843) These tissue sensitizers are characterized
by less photosensitivity, activation at wavelengths that penetrate deeper into tumors, and shorter
intervals between drug injection and the delivery of therapy. Delivery devices under development
include the fiber-containing balloon.95(844) This device has two advantages. It centers the diffusing tip
of the fiber in the transverse plane of the lumen to distribute light uniformly, and it anchors the diffusing
tip in position along the longitudinal axis of the lumen to ensure accurate tumor irradiation despite
peristalsis and other motion. These balloons can be partially covered with opaque shielding to minimize
injury to normal tissue during treatment of noncircumferential tumors. Advances in tissue sensitizers
and delivery technology should further enhance the efficacy and safety of PDT and simplify its
application.
REFERENCES
1.
2.
Raab O. ber die Wirkung Fluorescirender Stoffe auf Infusorien. Z Biol 1900;39:5246.
Von Tappeiner H, Jesionek A. Therapeutische Versuche mit fluoreszierenden Stoffen. Munch
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Med Wochenschr 1903;47:20424.

Policard A. tudes sur les aspects offerts par des tumeurs experimentales examinees a la
Auler H, Banzer G. Untersuchungen ber die Rolle der Porphirine bei Geschwulstkranken
Menschen und Tieren. Z Krebsforsch 1942;53:658.
Lipson RL, Baldes EJ. The photodynamic properties of a particular hematoporphyrin
derivative. Arch Dermatol 1960;82:50816.
Lipson RL, Baldes EJ, Olsen AM. Hematoporphyrin derivative: A new aid of endoscopic
detection of malignant disease. J Thorac Cardiovasc Surg 1961;42:6239.
Lipson RL, Baldes EJ, Olsen AM. A further evaluation of the use of hematoporphyrin
derivative as a new aid for the endoscopic detection of malignant disease. Dis Chest
1964;46:6769.
Lipson RL, Pratt JH, Baldes EJ, et al. Hematoporphyrin derivative for the detection of cervical
cancer. Obstet Gynecol 1964;24:7884.
Lipson RL, Gray MJ, Baldes EJ. Hematoporphyrin derivative for detection and management
of cancer. Proceedings of the Ninth International Cancer Congress, Tokyo, 1966;393.
Dougherty TJ, Boyle DG, Weishaupt KR, et al. Photoradiation therapyClinical and drug
advances. Adv Exp Med Biol 1983;160:313.
Dougherty TJ. Studies on the structure of porphyrins contained in Photofrin II. Photochem
Photobiol 1987;46:56973.
Dougherty TJ. Characterization of intra-tumoral porphyrin following injection of
hematoporphyrin derivative or its purified component. Photochem Photobiol 1987;46:6770.
Gomer CJ, Dougherty TJ. Determination of [3H]- and [14C]hematoporphyrin derivative
distribution in malignant and normal tissue. Cancer Res 1979;39:14651.
Douglass Jr HO, Nava HR, Weishaupt KR, et al. Intra-abdominal applications of
hematoporphyrin photoradiation therapy. Adv Exp Med Biol 1983;160:1521.
Okuda S, Mimura S, Otani T, et al. Experimental and clinical studies on HPD-photoradiation
therapy for upper gastrointestinal cancer. In Andreoni A, Cubeddu R (eds). Porphyrins in
Tumor Phototherapy. New York: Plenum Press, 1984;41321.
Bugelski PJ, Porter CW, Dougherty TJ. Autoradiographic distribution of hematoporphyrin
derivative in normal and tumor tissue of the mouse. Cancer Res 1981;41:460612.
Gilson D, Ash D, Driver I, et al. Therapeutic ratio of photodynamic therapy in the treatment of
superficial tumors of skin and subcutaneous tissues of man. Br J Cancer 1988;58:6657.
Dougherty T, Kaufman J, Goldfarb A, et al. Photoradiation therapy for the treatment of
malignant tumors. Cancer Res 1978;38:262835.
Redmond RW, Land EJ, Truscott TG. Aggregation effects on the photophysical properties of
porphyrins in relation to mechanisms involved in photodynamic therapy. Adv Exp Med Biol
1985;193:293302.
Kessel D, Thompson P, Musselman B, et al. Chemistry of hematoporphyrin-derived
photosensitizers. Photochem Photobiol 1987;46:5638.
Dougherty T. Photosensitization of malignant tumors. Semin Surg Oncol 1986;2:2437.
Kessel D. Porphyrin-lipoprotein association as a factor in porphyrin localization. Cancer Lett
1986;33:1838.
Norata G, Canti G, Ricci L, et al. In vivo assimilation of low density lipoproteins by a
fibrosarcoma tumour line in mice. Cancer Lett 1984;25:2038.
Ho Y, Smith R, Brown M, et al. Low-density lipoprotein (LDL) receptor activity in human acute
myelogenous leukemia cell. Blood 1978;52:1099114.
Gal D, MacDonald P, Porter J, et al. Cholesterol metabolism in cancer cells in monolayer
culture. III. Low-density lipoprotein metabolism. Int J Cancer 1981;28:3159.
Goldstein JL, Anderson RGW, Brown MS. Coated pits, coated vesicles and
receptor-mediated endocytosis. Nature 1979;279:67985.
Maziere J, Morliere P, Santus R. The role of low density lipoprotein receptor pathway in the
delivery of lipophilic photosensitizers in the photodynamic therapy of tumors. J Photochem
Photobiol B 1991;8:35160.
28. Moan J, Sommer S. Action spectra for hematoporphyrin derivative and Photofrin II with
respect to sensitization of human cells in vitro to photoinactivation. Photochem Photobiol
1984;40:6314.
29.
Moan J, Christensen T, Jacobson P. Porphyrin-sensitized photoinactivation of cells in vitro. In
Doiron DR, Gomer CJ (eds). Porphyrin Localization and Treatment of Tumors. New York:
Alan R Liss, 1984;41942.
30. Weishaupt KR, Gomer CJ, Dougherty TJ. Identification of singlet oxygen as the cytotoxic
agent in photoinactivation of two murine tumors. Cancer Res 1976;36:23269.
31. Keene J, Kessel D, Land E, et al. Direct detection of singlet oxygen sensitized by
hematoporphyrin and related compounds. Photochem Photobiol 1986;43:11720.
32. Gomer CJ, Razum NJ. Acute skin response in albino mice following porphyrin
photosensitization under oxic and anoxic conditions. Photochem Photobiol 1984;40:4359.
33. Lee See K, Forbes IJ, Betts WH. Oxygen dependency of phototoxicity with hematoporphyrin
derivative. Photochem Photobiol 1984;39:6314.
34. Mitchell JB, McPherson S, DeGraff W, et al. Oxygen dependence of hematoporphyrin
derivative-induced photoinactivation of Chinese hamster cells. Cancer Res 1985;45:200811.
35. Moan J, Sommer S. Oxygen dependence of the photosensitizing effect of hematoporphyrin
derivative in NHIK 3025 cells. Cancer Res 1985;45:160810.
36. Gibson SL, Hilf R. Interdependence of fluence, drug dose and oxygen on hematoporphyrin
derivative-induced photosensitization of tumor mitochondria. Photochem Photobiol
1985;42:36773.
37. Hilf R, Smail DB, Murant RS, et al. Hematoporphyrin derivative-induced photosensitivity of
mitochondrial succinate dehydrogenase and selected cytosolic enzymes of R3230AC
mammary adenocarcinomas of rats. Cancer Res 1984;44:14838.
38. Perlin DS, Murant RS, Gibson SL, et al. Effects of photosensitization by hematoporphyrin
derivative on mitochondrial adenosine triphosphate-mediated proton transport and membrane
integrity of R3230AC mammary adenocarcinoma. Cancer Res 1985;45:6538.
39. Atlante A, Moreno G, Passarella S, et al. Hematoporphyrin derivative (Photofrin II)
photosensitization of isolated mitochondria: Impairment of anion translocation. Biochem
Biophys Res Commun 1986;141:58490.
40. Dubbleman TMAR, van Steveninck J. Photodynamic effect of hematoporphyrin derivative on
transmembrane transport systems of murine L929 fibroblasts. Biochim Biophys Acta
1984;771:2017.
41. Gibson SL, Hilf R. Photosensitization of mitochondrial cytochrome c oxidase by
hematoporphyrin derivative and related porphyrins in vitro and in vivo. Cancer Res
1983;43:41917.
42. Bellnier DA, Dougherty TJ. Membrane lysis in Chinese hamster ovary cells treated with
hematoporphyrin derivative plus light. Photochem Photobiol 1982;36:437.
43.
Volden G, Christensen T, Moan J. Photodynamic membrane damage of hematoporphyrin
derivative-treated NHIK 3025 cells in vitro. Photobiochem Photobiophys 1981;3:10511.
44. Hilf R, Murant RS, Narayanan U, et al. Relationship of mitochondrial function and cellular
adenosine triphosphate levels to hematoporphyrin derivative-induced photosensitization in
R3230AC mammary tumors. Cancer Res 1986;46:2117.
45. Ceckler T, Gibson S, Kennedy S, et al. Heterogeneous tumour response to photodynamic
therapy assessed by in vivo localized 31P NMR spectroscopy. Br J Cancer 1991;63:9169.
46. Hilf R, Gibson SL, Penney DP, et al. Early biochemical response to photodynamic therapy
monitored by NMR spectroscopy. Photochem Photobiol 1987;46:80917.
47.
Franken NAP, Vrenson GFJM, van Delft JL, et al. Early morphological changes induced by
photodynamic therapy in amelanotic Greene melanoma implanted in the anterior eye
chamber of rabbits. Lasers Med Sci 1988;3:2734.
48. Henderson B, Waldow S, Mang T, et al. Tumor destruction and kinetics of tumor cell death in
two experimental mouse tumors following photodynamic therapy. Cancer Res 1985;45:5726.
49. Christensen T, Sandquist T, Feren K, et al. Retention and photodynamic effects of

haematoporphyrin derivative in cells after prolonged cultivation in the presence of porphyrin.
Br J Cancer 1983;48:3543.
50. Henderson BW, Bellnier DA, Ziring B, et al. Aspects of the cellular uptake and retention of
hematoporphyrin derivative and their correlation with the biological response to PRT in vitro.
Adv Exp Med Biol 1983;160:12938.
51.
Cheng T, Chang C, Dougherty TJ. Photoradiation therapy: Kinetics and thermodynamics of
porphyrin uptake and loss in normal and malignant cells in culture. Radiat Res 1978;74:498.
52. Nelson J, Lih-Huei L, Orenstein A, et al. Mechanisms of tumor destruction following
photodynamic therapy with hematoporphyrin derivative, chlorin, and phthalocyanine. J Natl
Cancer Inst 1988;80:1599605.
53. Star WM, Marijnissen HPA, van den Berg-Blok AE, et al. Destruction of rat mammary tumor
and normal tissue microcirculation by hematoporphyrin derivative photosensitization observed
in vitro in sandwich observation chambers. Cancer Res 1986;46:253240.
54.
Delaney T, Bonner R, Smith P, et al. Photodynamic therapy for surface malignancies: Clinical
results and correlation with blood flow effects secondary to photoirradiation. Int J Radiat Oncol
Biol Phys 1987;13(suppl):1867.
55. Hetzel FW, Farmer H. Dose effect relationships in a mouse mammary tumor. Prog Clin Biol
Res 1984;170:58390.
56. Reed M, Mullins A, Anderson G, et al. The effect of photodynamic therapy on tumor
oxygenation. Surgery 1989;106:949.
57.
Logan PM, Newton J, Richter A, et al. Immunological effects of photodynamic therapy.
Society of Photo-Optical Instrumentation Engineers 1990;1203:1538.
58. Evans S, Matthews W, Perry R, et al. Effect of photodynamic therapy on tumor necrosis factor
production by murine macrophages. J Natl Cancer Inst 1990;82:349.
59. Yamamoto N, Homma S, Sery T, et al. Photodynamic immunopotentiation: In vitro activation
of macrophages by treatment of mouse peritoneal cells with haematoporphyrin derivative and
light. Eur J Cancer 1991;27:46771.
60. Shumaker BP, Hetzel FW. Clinical laser photodynamic therapy in the treatment of bladder
carcinoma. Photochem Photobiol 1987;46:899901.
61.
Nseyo UO, Whalen RK, Duncan MR, et al. Immune response following photodynamic therapy
for bladder cancer. Society of Photo-Optical Instrumentation Engineers 1989;1065:6672.
62. Jolles CJ, Ott MJ, Straight RC, et al. Systemic immunosuppression induced by peritoneal
photodynamic therapy. Am J Obstet Gynecol 1988;158:144651.
63. Elmets CA, Bowen KD. Immunological suppression in mice treated with hematoporphyrin
derivative photoradiation. Cancer Res 1986;46:160811.
64. Musser DA, Fiel RJ. Cutaneous photosensitizing and immunosuppressive effects of a series
of tumor localizing porphyrins. Photochem Photobiol 1991;53:11923.
65. Gomer CJ, Ferrario A, Hayashi N, et al. Molecular, cellular, and tissue responses following
photodynamic therapy. Lasers Surg Med 1988;8:45063.
66. Tse T, Kersten R, Anderson R. Hematoporphyrin derivative photoradiation therapy in
managing nevoid basal-cell carcinoma syndrome. Arch Ophthalmol 1984;102:9904.
67.
McKenzie AL, Carruth JAS. A comparison of gold-vapour and dye lasers for photodynamic
therapy. Lasers Med Sci 1986;1:11720.
68. Mortsyn G, Kaye AH, Thomas R, et al. High dose photoirradiation therapy of human tumors
using the gold metal vapor laser. Photochem Photobiol 1987;46:9458.
69.
Heier SK, Rothman K, Heier LM, et al. Endoscopic laser therapies for obstructing esophageal
tumors. Proceedings of the Twelfth Annual International Conference of the IEEE Engineering
in Medicine and Biology Society 1990;12:11445.
70. Heier SK, Rothman KA, Heier LM, Rosenthal WS. Photodynamic therapy for obstructing
esophageal cancer: Light dosimetry and randomized comparison with Nd:YAG laser therapy.
71. McCaughan JS, Williams TE, Bethel BH. Palliation of esophageal malignancy with
photodynamic therapy. Ann Thorac Surg 1985;40:1139.

72. Thomas RJ, Abbott M, Bhathal PS, et al. High-dose photoirradiation of esophageal cancer.
Ann Surg 1987;206:193199.
73. Tian M, Qui S, Qing J. Preliminary results of hematoporphyrin derivative-laser treatment for
13 cases of early esophageal carcinoma. Adv Exp Med Biol 1985;193:215.
74. Fujimaki M, Nakayama K. Endoscopic laser treatment of superficial esophageal cancer.
Semin Surg Oncol 1986;2:24856.
75.
Wagnieres G, Monnier P, Savary M, et al. Photodynamic therapy of early cancer in the upper
aerodigestive tract and bronchi: Instrumentation and clinical results. Society of Photo-Optical
Instrumentation Engineers 1990;6(suppl 1):24971.
76.
Lambert R, Sabben G, Souquet JC, et al. Clinical experiences. In Riemann JF, Ell C (eds).
Lasers in Gastroenterology. New York: Thieme Medical, 1989;939.
77. McCaughan JS, Nims TA, Guy JT, et al. Photodynamic therapy for esophageal tumors. Arch
Surg 1989;124:7480.
78.
Heier SK, Rothman K, Heier LM, et al. Complete responses of esophageal adenocarcinomas
to photodynamic therapy. Gastrointest Endosc 1992;38:279.
79. Overholt B, Panjehpour M, Teftellar E, Rose M. Photodynamic therapy for treatment of early
adenocarcinoma in Barrett's esophagus. Gastrointest Endosc 1993;39:736.
80.
Nava HR, Zera R, Douglass HO, et al. Photodynamic therapy in carcinoma in situ in Barrett's
esophagus. Proceedings of the Third Biennial Meeting of the International Photodynamic
Association 1990;32.
80a. Wang KK, Gutta K, Laukka MA. A prospective randomized trial of low dose photodynamic
therapy in the treatment of Barrett's esophagus [Abstract]. Gastroenterology 1994;106:A208.
80b.
Overholt B, Panjehpour M. Photodynamic therapy destroys dysplasia and early cancer in
Barrett's esophagus. Gastrointest Endosc 1996;43:342.
81. Suzuki H, Miho O, Watanabe Y, et al. Endoscopic laser therapy in the curative and palliative
treatment of upper gastrointestinal cancer. World J Surg 1989;13:15864.
82. Ito Y, Kameya A, Kano T, et al. Indications and limitations of laser treatment for early gastric
cancer and palliative treatments for malignant obstruction of the esophagus and stomach. Jpn
J Cancer Chemother 1988;15:14359.
83. Tajiri H, Daikuzono N, Joffe SN, et al. Photoradiation therapy in early gastrointestinal cancer.
84. Hayata Y, Kato H, Okitsu H, et al. Photodynamic therapy with hematoporphyrin derivative in
cancer of the upper gastrointestinal tract. Semin Surg Oncol 1985;1:111.
85.
Patrice T, Foultier MT, Yactayo S, et al. Endoscopic photodynamic therapy with
haematoporphyrin derivative in gastroenterology. J Photochem Photobiol B 1990;6:15765.
86.
Barr H, Krasner N, Boulos PB, et al. Photodynamic therapy for colorectal cancer: A
quantitative pilot study. Br J Surg 1990;77:9396.
87.
Krasner N, Chatlani PT, Barr H. Photodynamic therapy in gastroenterologyA review. Lasers
Med Sci 1990;5:2338.
88. Heier SK, Rothman K, Heier LM, et al. Adenocarcinoma vs. squamous cell carcinoma of the
esophagus: The response to palliative photodynamic therapy. Gastrointest Endosc
1992;38:279.
89. Chemaly P. Phase III trials of photodynamic therapy now under way. J Natl Cancer Inst
1989;13:9756.
89a. Lightdale CJ, Heier SK, Marcon NE, et al. Photodynamic therapy with porfimer sodium versus
thermal ablation therapy with Nd:YAG laser for palliation of esophageal cancer: A multicenter
randomized trial. Gastrointest Endosc 1995;42:50712.
90. Boyle DG, Potter WR. Photobleaching of Photofrin II as a means of eliminating skin
photosensitivity. Photochem Photobiol 1987;46:9971001.
91. Barr H, MacRobert AJ, Tralau CJ, et al. The significance of the nature of the photosensitiser
for photodynamic therapy: Quantitative and biological studies in the colon. Br J Cancer
1990;62:7305.
92. Barr H, Tralau CJ, Boulos PB, et al. The contrasting mechanisms of colonic collagen damage
between photodynamic therapy and thermal injury. Photochem Photobiol 1987;46:795800.
93.
Barr H, Bown SG, Krasner N, et al. Photodynamic therapy for colorectal disease. Int J
Colorectal Dis 1989;4:159.
93a.
Likier HM, Levine JG, Lightdale CJ. Photodynamic therapy for completely obstructing
esophageal carcinoma. Gastrointest Endosc 1991;37:758.
94.
Heier SK, Heier LM. Tissue sensitizers. Gastrointest Endosc Clin North Am 1994;4:32752.
95.
Panjehpour M, Overholt BF, DeNovo RC, et al. Centering balloon to improve esophageal
photodynamic therapy. Lasers Surg Med 1992;12:6318.
Chapter 20 Stereoscopic Endoscopy

(845)
(846)
MARC F. CATALANO, M.D.
Attempts to determine the size of pathologic lesions of the gastrointestinal tract have been made since
the beginning of gastrointestinal endoscopy. Endoscopists have tried to define the size of a variety of
abnormalities, including benign and malignant tumors, ulcers, polyps, varices, and strictures. Along
with other endoscopic findings, the dimensions of a lesion contribute to the accuracy of diagnosis (e.g.,
potential for carcinoma within a polyp). Change in size may be the only indicator of the effectiveness of
treatment (e.g., the healing rate of peptic ulcers in response to H2 receptor antagonists), and size may
also indicate the need for therapy (e.g., diameter of esophageal varices).
The most common method of endoscopic measurement of lesions is by reference to a marker with
known dimensions, such as an open biopsy forceps. Because of the obvious limitations of this simple
but crude approach, several innovative techniques to increase accuracy have been proposed since the
mid-1980s. For the most part, these are also inaccurate and cumbersome. In response to this
problem, a novel endoscope capable of obtaining precise, stereoscopic three-dimensional
measurements has been developed for clinical research. This instrument has undergone "bench"
testing to determine its accuracy and potential benefit, and this technical evaluation is the subject of
this chapter.
Since the first report by Sivak and Fleischer1(847) describing its utility and efficacy, the video
endoscope has become the standard of practice. This instrument produces an image by means of an
integrated circuit image sensor or charged-coupled device (CCD) (see Chapter 3: Flexible Endoscope
Technology: The Video Image Endoscope). A prototype electronic endoscope that incorporates two
CCDs has been developed (Olympus Optical Company Ltd., Tokyo, Japan). With this revolutionary
stereoscopic endoscope (i.e., stereoscope), an endoscopist can measure matching objective points in
right and left images, compute three-dimensional coordinates (X, Y, and Z) by triangulation (i.e.,
parallax), and determine the distance (d) of the object point from the endoscope tip. This instrument
provides precise, three-dimensional measurements of the length, width, and height or depth of a lesion
seen at endoscopy.
Endoscopic Methods for Estimating Dimensions

Past and current techniques for determining the size of a lesion at endoscopy vary from the
unsophisticated open biopsy forceps method24(848) to complex graphic processing systems that use
lasers. All have inherent limitations. A certain degree of optical distortion exists in all commercially
available fiberoptic and video endoscopes, especially those that offer wide fields of view. Errors in
measurement are therefore unavoidable.
Sonnenberg et al.5(849) studied the accuracy of experienced endoscopists in estimating ulcer size, in
vitro and in vivo, using an open biopsy forceps or a graduated measuring probe. The true size of ulcers
was invariably underestimated by 29 40% (mean SD). This study demonstrated that estimation of
ulcer size at endoscopy is grossly inaccurate.

Several groups independently studied the accuracy of placing reference markers of known dimensions
adjacent to peptic ulcers as a method of estimating the size of the ulcers. Dancygier et al.6,7(850)
described the use of a system that consists of a graphic measuring tablet, computer, and television
monitor. After transmitting the image onto the monitor, the ulcerated area was measured directly with
an electronic overlay marker. The computer then calculated the circumscribed ulcerated area by
relating this to the known dimensions of the endoscopically introduced reference area marker. Based
on multiple measurements obtained using different reference areas, this method proved to be reliable,
with a margin of error of only 4.2 0.5%. However, the optical distortion introduced by the endoscope
lens was not taken into account. Employing the same technique, Okada et al.8(851) obtained results
with a substantially greater aberration ratio (i.e., percentage deviation between the measured and the
true volumes). This was accounted for by the optical distortion introduced by the endoscopic
instruments used in the study. Later, Mitsuhashi et al.9(852) used this same endoscopic marker
method in evaluating 127 patients with peptic ulcers who were treated with a variety of modalities. The
investigators found an exponential, rather than linear, rate of ulcer healing and concluded that the
marker method was reliable for following the effects of antiulcer medications.
Okabe et al.10(853) described a modification of the reference disk technique that corrected for the
optical distortion that was not accounted for in prior studies. With an endoscope, they photographed a
mesh pattern of squares (2.5 2.5 mm) at lens-object distances (LOD) from 1 to 5 cm. These pictures
were copied onto transparent glass plates. Next, an endoscopic photograph of an ulcer with a
5-mm-diameter rubber disk in the center was placed under the glass plate with the square pattern of its
corresponding LOD. By counting the number of squares for which at least half of the area of the
square was contained within the area of the ulcer, a more precise determination of the size of the
lesion could be made. The average aberration ratio was 5.6 1.34%.
An innovative endoscopy system for three-dimensional measurement of ulcers was described by
Yamaguchi et al.1113(854) The components of this system include an argon laser and a side-viewing
endoscope that contains a diffraction double-layer glass fiber coupled to a graphic processing unit. The
grating defracts the laser beam and causes a two-dimensional array to be projected on the mucosa.
The depth, diameter, area, and height of a mucosal lesion are determined by graphic processing that
involves detection of the deviation of the laser spot array on the mucosa. When tested in vitro, this
system had an aberration ratio of 7.5 4.8%.
The use of fiberscopes with measuring scales built into the lens system has been described
independently by Sakita and Oguro,14(855) Yamaguchi et al.,15(856) and Maruyama et al.16(857)
With this type of system, aberration ratios as high as 11.2 6.3% have been found. As described by
Yamaguchi et al.,15(858) these instruments consist of a flexible fiberscope with a scaled objective lens
attached to a digital display indicator. A sensor guide, operated by an air pressure system through the
accessory channel, is advanced beyond the distal end of the instrument to the center of the lesion to
be measured. The system then provides a measurement of this distance. By means of a simple
formula that depends on the LOD, the lesion can be measured from the scaled optical lens.
Stereoscope Specifications
The stereoscope is nearly indistinguishable from standard upper gastrointestinal endoscopes,
especially the Olympus model GIF-100 (Olympus Company, Ltd., Tokyo, Japan). The length of the
insertion tube is 100 cm; the field of view is 120 degrees. The insertion tube diameter is 12.5 mm,
slightly larger than current upper gastrointestinal endoscopes. There is a small-diameter (2.0 mm)
working accessory channel (Figures 201 and 202). The system also includes two video monitors,
two video processors with light sources (CV-1) (Figure 203), two Frame Memory Banks (VP-1125),
and a central processing computer (PC-9801 RA) complete with hard disk, keyboard, and mouse
(Figure 204).
(859)Figure 201. Distal end of a stereoscopic endoscope with two charge-coupled devices
(CCDs), two light guide bundles, and an accessory channel. (From Catalano MF, Van Dam J,
Bedford R, et al. Preliminary evaluation of the prototype stereoscopic endoscope: Precise
three-dimensional measurement system. Gastrointest Endosc 1993; 39:238.)
(860)Figure 202. Stereoscopic endoscope with dual light source adaptors (length 100 cm,
diameter 12.5 mm). (From Catalano MF, Van Dam J, Bedford R, et al. Preliminary evaluation
of the prototype stereoscopic endoscope: Precise three-dimensional measurement system.
Gastrointest Endosc 1993; 39:238.)
(861)Figure 203. Stereoscopic endoscopy system with two vertically mounted CV-1 video
processor systems. (Courtesy of Olympus Optical Company Ltd., Tokyo, Japan.)
(862)Figure 204. Three-dimensional stereoscopic endoscopy system consisting of two frame

memory banks (VP-1125), central processing computer (PC-9898-9801 RA5), keyboards,
and mouse. (Courtesy of Olympus Optical Company Ltd., Tokyo, Japan; and NEC, Boxboro,
MA.)
Essentially, the stereoscope consists of two optical imaging systems. Images can be captured
simultaneously with each optical system (Figure 205) and stored in the computer memory (up to 15
images). After endoscopic examination, three-dimensional measurements can be calculated from the
stored images, so that the length of the endoscopic portion of the procedure is the same as that for a
conventional examination.
(863)Figure 205. Central processing computer showing a menu of stored, dual images.
Clinical Applications
Most attempts to develop more precise methods for determining size endoscopically are indirect, and
although many methods have been proposed, none has yet been found that is exact, uncomplicated,
and reliable. Commonly employed techniques using a biopsy forceps or a graduated probe, for
example, are inaccurate and cumbersome. The stereo endoscope was developed specifically to
address the problem of obtaining accurate measurements at endoscopy.
The importance of precise, three-dimensional measurements within the gastrointestinal tract is
considered in relation to three common lesions: peptic ulcer, esophageal and gastric varices, and
gastrointestinal tumors.
Peptic Ulcer
The efficacy of antiulcer drugs has been assessed in controlled clinical trials by measuring the size of
ulcers before, during, and after treatment.17(864) However, the accuracy and reproducibility of
endoscopic measurements have often been poor. As a result of barrel distortion (i.e., the peripheral
distortion of an image inherent to all wide-field optical systems), errors in measurement are likely to
occur.
Sonnenberg et al.5(865) assessed the reliability of measuring ulcer size by endoscopy using the open
biopsy forceps method. Initially, surface areas of 23 artificial ulcers in a mannequin were estimated,
and then gastric ulcers were measured in 35 consecutive patients by six experienced endoscopists. Of
the 138 estimations made in the mannequin, 80% underestimated the true size of the ulcer. The ratio
of the largest and the smallest estimate (R) of the same ulcer by different endoscopists varied on
average by a factor of 4.5 3.8%. For patients, the scatter of the estimates was even larger (R = 7.8
6.3%). Sonnenberg et al.5(866) concluded that estimates of ulcer size by endoscopy are highly
inaccurate and that changes in size as measured by the open biopsy forceps technique should not be
used as an index of the effect of antiulcer therapy.
Ulcer healing has been investigated in several therapeutic trials.2,3,17(867) The methods used to
assess healing include time to complete healing, the percentage of ulcers healed over a given period
(e.g., 6 to 8 weeks), and the percent reduction in ulcer size per unit time. The major problem with these
methods, as pointed out by Sheurer et al.,4(868) is the need for frequent endoscopic surveillance
procedures to assess healing of the ulcer. This becomes expensive, impractical, and time consuming.
In the study of Sheurer et al.,4(869) 38 patients with peptic ulcer disease completed the protocol; 24
had gastric ulcers, and 14 had duodenal ulcers. Ulcer size was measured with a calibrated measuring
device on days 1 and 21 of the study; some ulcers were measured at 42 days in patients who received
a placebo. Healing could be described as an exponential function from which an ulcer half-life could be
calculated (1.7 weeks for gastric and 1.9 weeks for duodenal ulcers). Sheurer et al.4(870) concluded
that calculated ulcer half-life is a sensitive indicator of ulcer healing that can be used to decrease the
need for repeated surveillance endoscopy procedures in trials of ulcer therapy.
Gastrointestinal Malignancy
Accurate three-dimensional measurements of tumor size are critically important for preoperative
assessment and for evaluating response to chemotherapy and other treatment modalities.1820(871)
The initial size estimate and subsequent response to therapy are determined by a combination of
endoscopic and radiographic findings, along with the subjective response of the patient to treatment
(e.g., improvement in dysphagia score). No established method for the three-dimensional assessment
of intraluminal lesions is objective and accurate.
Esophageal Varices
In patients with portal hypertension and esophageal varices, the magnitude of portal pressure is
suggested by variceal radius. In general, the portal pressure is higher in patients with large varices
than in those with small esophageal varices.2123(872) Precise determination of the size of
esophageal varices is therefore a critical factor in predicting which patients are at increased risk of
hemorrhage. Larger varices are associated with a higher rate of bleeding.2426(873) However, there
is considerable interobserver variation with regard to the grading of esophageal varices.27,28(874) An
accurate method for estimating variceal size would be of critical importance in the evaluation of
patients for therapy and for assessing the results of therapeutic trials (e.g., prophylactic
sclerotherapy).29(875)
Laboratory Tests
The prototype stereoscope described in this chapter underwent extensive bench testing in a series of
experiments using two-dimensional planar target grids. These studies were designed to determine
ease of use (i.e., function), reproducibility of measurements (i.e., to determine consistency with which a
result is obtained), barrel distortion, and accuracy. The results of these tests indicated that
1. There was no learning curve; special training and practice are unnecessary prior to using this
system.
2. Measurements were most reproducible at short distances from the target (<30 mm) when the
target was near the center of the image field.
3. The barrel distortion associated with all wide-field optical systems was most prominent when
measuring targets in the periphery of the field and at distances of more than 5 cm from the
endoscope.
4. All three spatial coordinates were overestimated by a proportionality factor of 0.83.
Ex vivo studies were done to determine the accuracy of the stereoscopic endoscope in measuring
objects of known size and to compare the results with measurements obtained using established
endoscopic methods.29(876) Fifteen objects, including 10 rectangles, 2 cubes, and 3 marbles, were
placed within an artificial representation of the sigmoid colon. After placing the stereoscope about 2 cm
from each object, an image was acquired and retained in the computer memory. Measurements were
then obtained of three matching (objective) points (X, Y, Z) in the right and left images (Figures 206,
207 and 208). Measurements estimated with the open biopsy forceps technique by an endoscopist
without prior knowledge of the shape or size of the objects were compared with measurements
obtained with the stereoscope. The measurements obtained by the two methods were then compared
with the actual measurements of the objects using the Dunnett multiple comparison procedure. No
significant difference was found between the actual measurements and those obtained with the
stereoscopic system. However, measurements based on the open biopsy forceps method consistently
underestimated actual measurements.
(877)Figure 206. Right and left monitor images showing cursors measuring identical
(objective) points and triangulation (parallax) computing three-dimensional coordinates.
(878)Figure 207. Central processing computer showing tabulation of X, Y, and Z

coordinates. (From Catalano MF, Van Dam J, Bedford R, et al. Preliminary evaluation of the
prototype stereoscopic endoscope: Precise three-dimensional measurement system.
Gastrointest Endosc 1993; 39:238.)
(879)Figure 208. Depth or height calculation of an object.

Differences in means between volume as determined by the open biopsy forceps technique (Vbx), that
obtained by the stereoscopic system (Vs), and the actual volume of the objects (Vt) are given in Table
201. The 95% simultaneous confidence intervals demonstrate the magnitude of the differences
between the volume provided by each technique and the actual volume. The percent deviation
between the true and the measured volumes (i.e., aberration ratios) for the 15 objects using the
stereoscope and the open biopsy forceps technique are presented in Table 202. With respect to this
parameter, the stereoscope is clearly superior to the forceps method.
True Volume Versus Volume as Determined

From Stereo-scopic Images and the Open Biopsy Forceps
Method
TABLE 201
COMPARISON
MEAN DIFFERENCE
95% SIMULTANEOUS C.I.
True Volume Versus Volume as Determined

From Stereo-scopic Images and the Open Biopsy Forceps
Method
TABLE 201
COMPARISON
MEAN DIFFERENCE
95% SIMULTANEOUS C.I.
Vs Vt
343.6
(-436.9, 1124.1)
Vbx Vt
-1157.8
(-1938.3, -377.3)
From Catalano MF, Van Dam J, Bedford R, et al. Preliminary evaluation of the prototype
stereoscopic endoscope: Precise three-dimensional measurement system. Gastrointest
Endosc 1993; 39:238.
C.I.confidence interval; Vttrue volume; Vsvolume as determined from stereoscope
images; Vbxvolume as determined with the open biopsy forceps
Actual Volumes and Measured Volumes of 15 Objects Using the Open

Biopsy Forceps and Stereoscopic Endoscope
TABLE 202
OPEN BIOPSY FORCEPS

OBJECT
ACTUAL
VOLUME (MM3)
Measured
Volume (mm3)
Aberration
Ratio (%)
STEREOSCOPIC ENDOSCOPE
Measured Volume
(mm3)
Aberration Ratio
(%)
1
1498.5
381.5
1579.0
74.5
2
2350.9
1436.0
2099.7
38.9
3
7234.6
3052.1
9563.7
57.8
4
970.3
512.0
1080.1
47.2
5
3723.9
1728.0
3774.6
53.6
6
599.1
768.0
28.2
591.0
7
1198.3
864.0
1340.5
27.9
8
1228.8
1152.0
1359.4
6.3
9
1812.6
1512.0
2116.2
16.6
10
2396.5
1296.0
2613.3
45.9
11
2419.3
1944.0
2589.5
19.7
12
2442.2
1944.0
2607.7
20.4
13
3625.2
3200.0
3905.9
11.7
14
4793.1
2016.0
5641.8
57.9
15
4823.4
1944.0
5407.9
59.7
Mean
2741.1 1807.7
1583.3 821.1
34.0 26.8
3084.7 2332.8
SD
From Catalano MF, Van Dam J, Bedford R, et al. Preliminary evaluation of the prototype stereoscopic
endoscope: Precise three-dimensional measurement system. Gastrointest Endosc 1993; 39:238.
5.4
10.7
32.2
11.3
1.4
1.4
11.9
10.6
16.8
9.1
7.0
6.8
7.8
17.7
12.1
9.2 9.5
The results of the ex vivo bench testing of the stereoscope are demonstrated graphically in Figure
209. Along the 45-degree line, estimated volume equals actual volume. The data points obtained with
the stereoscope (indicated by squares) are in general near the 45-degree line. Data points obtained by
means of the open biopsy forceps method (indicated by circles) are consistently below the 45-degree
line, and this method consistently underestimates volume.
(880)Figure 209. Graphic results, comparing the open biopsy technique (circles) with the
stereoscopic results (squares). The 45-degree line is the line for which the estimate of the
volume equals the actual volume. (From Catalano MF, Van Dam J, Bedford R, et al.
Preliminary evaluation of the prototype stereoscopic endoscope: Precise three-dimensional
measurement system. Gastrointest Endosc 1993; 39:238.)
Summary
Although exact three-dimensional measurement of lesions at endoscopy is usually unnecessary,
precise measurements are desirable in certain circumstances. The methods employed at endoscopy
to obtain such measurements have intrinsic limitations; most are complex and cumbersome, and all
are relatively inaccurate. The stereoscopic endoscopy system represents a new approach to the
problem of obtaining accurate measurements of lesions encountered at endoscopy. It is relatively
simple to use and has proved to be highly accurate when tested under standardized laboratory
conditions. Trials including animal models and human subjects are necessary to demonstrate the
clinical usefulness and accuracy of the stereoscope.
REFERENCES
1.
Sivak MV, Fleischer DE. Colonoscopy with a videoendoscope: Preliminary experience.

2. Dyck WP, Belsito A, Fleshler B, et al. Cimetidine and placebo in the treatment of benign
gastric ulcer: A multicenter double blind study. Gastroenterology 1978;74:4105.
3. Englert E, Freston JW, Graham DY, et al. Cimetidine, antacid and hospitalization in the
treatment of benign gastric ulcer. A multicenter double blind study. Gastroenterology
1978;74:41625.
4. Scheurer U, Witzel L, Halter F, et al. Gastric and duodenal ulcer healing under placebo
treatment. Gastroenterology 1977;72:83841.
5.
Sonnenberg A, Giger M, Kern L, et al. How reliable is determination of ulcer size by
endoscopy?. BMJ 1979;2:13224.
6.
Dancygier H, Wurbs D, Classen M. New method for endoscopic determination of ulcer size.
Gut 1980;21:A895.
7. Dancygier H, Wurbs D, Classen M. A new method for the endoscopic determination of
gastrointestinal ulcer area. Endoscopy 1981;13:2146.
8.
Okada N, Katsumata T, Saigenjii K, et al. A basic study on endoscopic measurement of
gastric ulcerated areas using a marker disk method. Gastroenterol Endosc 1983;25:24753.
9.
Mitsuhashi T, Okada N, Ohida M. Clinical study on gastric ulcer therapy using endoscopic
marker method. Gastroenterol Endosc 1984;26:102840.
10. Okabe H, Ohida M, Okada N, et al. A new disk method for the endoscopic determination of
gastric ulcer area. Gastrointest Endosc 1986;32:204.
11.
Yamaguchi M, Iida Y, Ohtani T, et al. A new method for the endoscopic measurement of
gastrointestinal lesion by the diffracted laser beam and graphic processingThe first report.
Gastroenterol Endosc 1983;25:86874.
12.
Yamaguchi M, Okazaki Y, Matsuda K, et al. A new method for the endoscopic measurement
of gastrointestinal lesion by the diffracted laser beam and graphic processingThe second
report. Gastroenterol Endosc 1985;27:19519.
13. Yamaguchi M, Okazaki Y, Yanai H, et al. Three-dimensional determination of gastric ulcer
size with laser endoscopy. Endoscopy 1988;20:2636.
14.
Sakita T, Oguro Y. Measurements of the lesion size under direct vision with optical
fiberscope. Gastroenterol Endosc 1973;15:45664.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Yamaguchi M, Iida Y, Sataka M, et al. Clinical trial of a newly developed measuring gastric
fiberscope. Gastroenterol Endosc 1983;25:3945.
Maruyama M, Ichioka S, Takemoto T, et al. A clinical trial of a gastrofiberscope with scale.
Resbeut M, Prise-Fleury EL, Ben-Hassel M, et al. Squamous cell carcinoma of the
esophagus: Treatment by combined vincristine-methotrexate plus folinic acid rescue and
cisplatin before radiotherapy. Cancer 1985;56:124650.
Seitz JF, Giovannini M, Padaut-Cesana J, et al. Inoperable nonmetastatic squamous cell
carcinoma of the esophagus managed by concomitant chemotherapy (5-fluorouracil and
cisplatin) and radiation therapy. Cancer 1990;66:2149.
Wolfe WG, Burton GV, Seigler HF, et al. Early results with combined modality therapy for
carcinoma of the esophagus. Ann Surg 1987;205:56371.
Garcia-Tsao G, Groszmann RJ, Fisher RL, et al. Portal pressure, presence of
gastroesophageal varices and variceal bleeding. Hepatology 1985;5:41924.
Lebrec D, DeFleury P, Rueff B, et al. Portal hypertension, size of esophageal varices and risk
of gastrointestinal bleeding in alcoholic cirrhosis. Gastroenterology 1980;79:113944.
Snady H, Feinman L. Prediction of variceal hemorrhage: A prospective study. Am J
Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by esophageal
endoscopy. Gastrointest Endosc 1981;27:2138.
Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology
1981;80:8009.
North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices. Prediction
of the first variceal hemorrhage in patients with cirrhosis of the liver and esophageal varices.
A prospective multicenter study. N Engl J Med 1988;319:9839.
Bendstein F, Skovgaard LT, Sorensen TIA, et al. Agreement among multiple observers on
endoscopic diagnosis of esophageal varices before bleeding. Hepatology 1990;11:3417.
Catalano MF, Van Dam J, Sivak MV. Are endoscopic features diagnostic of portal
hypertension (PH)? Interobserver variation. Gastroenterology 1992;102:A789.
Groszmann RJ. Drug therapy of portal hypertension. Am J Gastroenterol 1987;82:10713.
Catalano MF, Van Dam J, Bedford R, et al. Preliminary evaluation of the prototype
stereoscopic endoscope: Precise three-dimensional measurement system. Gastrointest
Endosc 1993;39:238.
Chapter 21 Endoscopic Ultrasonography with Linear Array

Instruments: Anatomy and EUS-Guided Fine-Needle Aspiration
(881)
(882)
KENNETH J. CHANG, M.D.
RICHARD A. ERICKSON, M.D.
Endoscopic ultrasonography (EUS)-guided fine-needle aspiration (FNA) has been shown to be an

effective modality for establishing a histologic diagnosis of primary malignant lesions within and
adjacent to the gastrointestinal tract and for documenting the spread of malignancy to lymph nodes,
evaluating fluid collections, and assessing liver metastases.16(883) Novel techniques such as
EUS-guided celiac blockade and EUS-guided cholangiography are being developed as corollaries of
EUS-guided FNA.
EUS-guided FNA is performed using a curved linear array echoendoscope (Pentax Precision
Instruments, Orangeburg, NY) (Figure 211). Color-flow mapping and Doppler ultrasonography, which
are useful in the assessment of vascular invasion by tumor, can be performed only with a linear array
echoendoscope. Although the Pentax FG-32 UA model is the only instrument commercially available,
prototype curved linear array instruments are being developed by other companies (Olympus America,
Inc., Melville, NY) (Figure 212).
(884)Figure 211. Pentax curved linear array echoendoscope (FG-32 UA) with
GIP/Medi-globe needle.
(885)Figure 212. Olympus prototype linear array echoendoscopes, XGF-UC4 (top) and
XGF-UC3 (bottom).
The first part of this chapter presents endosonographic anatomy as demonstrated by linear array
instruments and includes a detailed description of the way in which various structures are imaged. The
second part focuses on the new modality of EUS-guided FNA, including technical considerations and a
review of published results.
Linear Array Endosonographic Anatomy

Training in endosonography has focused mainly on the use of radial scanning echoendoscopes.
However, radial scanning and curved phased array instruments display endosonographic anatomy
quite differently. It is therefore important to grasp the basic concepts of linear array endosonographic
anatomy before embarking on EUS-guided FNA.
Radial scanning EUS employs a 360-degree ultrasonographic field oriented perpendicular to the long
axis of the echoendoscope. With the probe in the esophagus, the images obtained are similar in
orientation to those obtained by computed tomography (CT) or magnetic resonance imaging of the
chest. Linear array EUS provides a wedge-shaped ultrasonographic field of 100 to 180 degrees that is
oriented parallel to the long axis of the echoendoscope. For any given probe position within the chest,
images obtained with a linear array transducer are rotated 90 degrees from those obtained with the
radial scanning echoendoscope.
Gastrointestinal Wall
The five echolayer images of the wall of the esophagus, stomach, duodenum, and rectum appear
similar with linear array and radial scanning transducers. However, the linear array transducer offers
frequencies of 7.5 and 5.0 MHz but not the 12 MHz available with the radial scanners. The five
echolayers of the gastrointestinal wall, especially the thin-walled esophagus, are less well visualized
with the curved phased array echoendoscope. Techniques for acoustic coupling to enhance images of
the wall layers include filling the gastrointestinal lumen with water (i.e., water-fill method) or injecting
water while imaging and then promptly freezing the display (i.e., dynamic water-fill method).
Mediastinum
Knowledge of the linear array anatomy of the mediastinum is prerequisite for FNA of lymph nodes that
is performed to stage esophageal and non-small cell carcinomas of the lung. Most endosonographers
employ radial endosonography for identification of pathologic abnormalities in the mediastinum. After a
lesion warranting FNA has been recognized, the ability to relocate it using the curved linear array
echoendoscope becomes essential; this depends on a thorough understanding of mediastinal
endosonographic anatomy.
As in assessing the abdomen, vascular anatomy is the key to mastering linear echoendoscopic
anatomy of the mediastinum. Most of the structures visible by EUS are anterior to the esophagus, and
it is therefore necessary to understand mediastinal anatomy from a posterior perspective, as shown in
Figure 213. The major structures can be identified by following a standard pattern during the
procedure.
(886)Figure 213. Posterior view of the paraesophageal mediastinal anatomy. The numbers
refer to reference points (i.e., stations) for the endosonographic views that follow.
Initially, the linear echoendoscope is inserted into the lower third of the esophagus, which is 30 to 35
cm from the incisors (see Station 1, Figure 213), at which point the transducer is usually oriented
toward the patient's left side. By rotating the insertion tube of the echoendoscope a little to the right
(i.e., clockwise) or left (i.e., counterclockwise), the descending aorta can be easily recognized.
Endosonographically, the descending aorta is seen as a large, echolucent, longitudinal structure with a
deeper wall that is very bright (i.e., echodense) owing to the air interface with the left lung (Figure
214A).
(887)Figure 214. Linear array endosonographic views of the mediastinum using

Pentax/Hitachi FG-32 UA system at 7.5 MHz. All station numbers refer to Figure 213. A,
Descending aorta (Station 1). B, View through the left (L) atrium and mitral valve (MV) into
the left (L) ventricle (Station 1). C, View of the subcarinal region (Station 2), with the
cephalad portion of the left (L) atrium, pulmonary artery (PA), aortic valve (AV), and
ascending aorta. D, Simultaneous cross-sectional views of the pulmonary artery (PA) and
aortic arch.
The descending aorta is the reference structure seen throughout most of the mediastinum. From the
descending aorta, rotating the echoendoscope to the right moves the transducer from posterior to
anterior along the patient's left esophageal wall. This sequentially brings into view the left lung and left
atrium. On continued rotation past the left atrium, the transducer moves from anterior to posterior
along the patient's right, with visualization of the right lung, azygous vein, and spine.
Rotating to the right from the descending aorta in the distal esophagus brings into view the left atrium
as a contracting, thin-walled echolucent chamber (see Figure 214B). The mitral and aortic valves can
be visualized through the left atrium. Starting from the descending aorta, when rotating the
echoendoscope insertion tube to the right, the mitral valve is usually seen before the aortic valve (see
Figure 214B). Further rightward rotation and withdrawal a slight distance brings into view the aortic
outflow tract of the left ventricle. Slight further withdrawal brings the aortic valve and the root of the
ascending aorta into full view deep to the left atrium.
By withdrawing a few centimeters more, the echoendoscope is in the subcarinal region (see Station 2,
Figure 213). Visible here is the large, pulsatile, cephalad portion of the left atrium or upper pulmonary
veins just distal to a round cross-sectional view of the pulmonary artery (see Figure 214C). Depending
on the exact orientation and individual patient anatomy, the ascending aorta may be seen deep to
these structures. This is an important view for localization of subcarinal lymph nodes for FNA. By the
modified American Thoracic Society classification given in Table 211 of mediastinal lymph
nodes,7(888) nodes seen in the area from the midpoint of the pulmonary artery distal for 2 to 3 cm
represent Nodal Station 7, the subcarinal nodes. Depending on the patient's anatomy and the exact
orientation of the echoendoscope, the superior vena cava may be seen in cross section interposed
between the ascending aorta and the pulmonary artery.
American Thoracic Society Definitions of

Regional Nodal Stations
TABLE 211
X
Supraclavicular nodes
American Thoracic Society Definitions of

Regional Nodal Stations
TABLE 211
2R
Right upper paratracheal nodes: Nodes to the right of the midline of

the trachea, between the intersection of the caudal margin of the
innominate artery with the trachea and the apex of the lung
2L
Left upper paratracheal nodes: Nodes to the left of the midline of the
trachea, between the top of the aortic arch and the apex of the lung
4R
Right lower paratracheal nodes: Nodes to the right of the midline of
the trachea, between the cephalic border of the azygos vein and the
intersection of the caudal margin of the brachiocephalic artery with the
right side of the trachea
4L
Left lower paratracheal nodes: Nodes to the left of the midline of the
trachea, between the top of the aortic arch and the level of the carina,
medial to the ligamentum arteriosum
5
Aortopulmonary nodes: Subaortic and paraaortic nodes lateral to the
ligamentum arteriosum or the aorta or left pulmonary artery, proximal
to the first branch of the left pulmonary artery
6
Anterior mediastinal nodes: Nodes anterior to the ascending aorta or
the innominate artery
7
Subcarinal nodes: Nodes arising caudal to the carina of the trachea
but not associated with the lower lobe bronchi or arteries within the
lung
8
Paraesophageal nodes: Nodes dorsal to the posterior wall of the
trachea and to the right or left of the midline of the esophagus
9
Right or left pulmonary ligament nodes: Nodes within the right or left
pulmonary ligament
10R
Right tracheobronchial nodes: Nodes to the right of the midline of the
trachea, from the level of the cephalic border of the azygos vein to the
origin of the right upper lobe bronchus
10L
Left tracheobronchial nodes: Nodes to the left of the midline of the
trachea, between the carina and the left upper lobe bronchus, medial
to the ligamentum arteriosum
11
Intrapulmonary nodes: Nodes removed in the right or left lung
specimen, plus those distal to the mainstem bronchi or secondary
carina
From Glazer GM, Gross BH, Quint LE, et al. Normal mediastinal lymph nodes:
Number and size according to American Thoracic Society mapping. Am J
Roentgenol 1985; 144:2615; adapted from American Thoracic Society. Clinical
staging of primary lung cancer. American Thoracic Society node mapping
scheme. Am Rev Respir Dis 1983; 127:65964.
Further withdrawal of the echoendoscope presents a blind area to the echoendoscope as it passes
over the left mainstem bronchus. The arch of the aorta then comes into view as a large circular
structure adjacent to the esophagus (see Station 3, Figure 213). By rotating the echoendoscope
slightly right or left, the origins of the left subclavian or the left common carotid arteries can sometimes
be seen. Occasionally visible deep to the arch is the left innominate (brachiocephalic) vein. By rotating
the instrument right or left, the pulmonary artery and aortic arch can usually be brought simultaneously
into cross-sectional view (see Figure 214D).
Between the pulmonary artery and the aortic arch is the aortopulmonary window, another important
area for FNA of mediastinal lymph nodes. Nodes visualized between the cephalad border of the aortic
arch and the midpoint of the pulmonary artery represent Nodal Station 4L, the left lower paratracheal
nodes, by the modified American Thoracic Society classification. Lymph nodes visualized in the area
deep to the aortic arch and pulmonary artery are in Nodal Station 5. These aortopulmonary lymph
nodes are not accessible to FNA through the esophageal wall.
Rotating the echoendoscope further to the right at the level of the aortic arch usually visualizes only the
bright interface with the posterior wall of the trachea, the right lung, and the spine. Lymph nodes seen
at this level on the patient's right between the midline of the trachea and the spine are classified as
Nodal Station 4R and are called the right lower paratracheal nodes.
On withdrawing the echoendoscope toward the neck, few structures can be visualized because the
esophagus is wedged among the echo-impenetrable trachea anteriorly, the spine posteriorly, and the
apices of the lungs. However, longitudinal views of the left and occasionally right common carotid
arteries and the deeper internal jugular veins can be obtained when the echoendoscope is oriented to
the patient's left or right in the proximal esophagus.
Any mediastinal lymph nodes visualized above the level of the aortic arch are classified as Nodal
Stations 2R or 2L and are called the upper paratracheal nodes. The midline of the trachea is the
dividing line between these nodal stations.
In the middle to distal esophagus, rotation to the left from the descending aorta promptly brings the
azygous vein into view as a thin, longitudinal echolucency close to the wall of the esophagus. Rotating
the echoendoscope to the right or left of the descending aorta at the level of the gastroesophageal
junction usually brings into view the hepatic veins draining into the inferior vena cava. The
paraesophageal lymph nodes visualized below the subcarinal region are classified as Nodal Station 8.
Right or left parabronchial nodes (Nodal Stations 10R and 10L) are generally not visible by EUS
because of air in the bronchus and surrounding lungs.
In anticipation of attempts to locate a lesion as seen by radial EUS using a linear array
echoendoscope, it is advisable to first identify the general position of the radial echoendoscope so that
the linear array instrument can be placed in approximately the same region. This can be done in the
esophagus by reference to the distance of the transducer from the incisor teeth or mandibular ridge.
The next step is to delineate the relationship of the lesion in question to surrounding major vascular
structures. When the linear array echoendoscope is placed in the same region, the vessels used as
reference points at radial scanning can be used to localize the lesion. If there is any difficulty during
linear array echoendoscopy in differentiating an artery, systemic vessel, or splanchnic vein, Doppler
and color-flow ultrasonography can be used. Once the reference vessels are located, it is only
necessary to rotate or advance the linear array echoendoscope in the appropriate direction to visualize
the lesion in question.
Abdomen
The stomach and duodenum serve as the windows for imaging abdominal organs and vascular
structures (Figure 215). Unlike echoimaging through the wall of the esophagus, in which anatomic
relationships tend to be fixed, endosonography through the gastric wall is more difficult because of the
larger capacity and greater anatomic variation of the stomach, and the differences in shape and size
change the relationships to surrounding organs.
(889)Figure 215. Anterior view of the paragastroduodenal anatomy. The numbers refer to
reference points (stations) for the endosonographic views that follow.
The major reference structures for imaging through the stomach are the abdominal aorta and the
mesenteric vessels. The examination begins at the gastroesophageal junction. With the transducer
positioned in the cardia, the echoendoscope is rotated until the abdominal aorta is imaged
longitudinally, moving downward from the thorax (see Station 1, Figure 215).
By advancing the transducer a few centimeters along the aorta, the celiac trunk comes into view as a
longitudinal structure branching off the aorta at an angle of about 45 degrees (Figure 216). Color-flow
and Doppler ultrasonography can be used for verification. The celiac artery has a typical mesenteric
arterial waveform, with a sharp upstroke followed by a substantial diastolic flow owing to the low
resistance of the mesenteric vascular bed. It is important to visualize the celiac artery as a reference
structure in the identification of celiac lymph nodes for EUS-guided FNA8(890) and for the
performance of celiac nerve blocks.9(891)
(892)Figure 216. View of the celiac axis (CA) and superior mesenteric artery (SMA)
coming off the abdominal aorta at Station 1 of Figure 215. Color mapping shows laminar
flow.
After the celiac trunk has been identified, further advancement of the transducer 1 to 2 cm images the
superior mesenteric artery (SMA) coming off the aorta at an angle of about 30 degrees (see Figure
216). Between the celiac trunk and the SMA (see Station 2, Figure 215), the body of the pancreas
can be imaged in cross section, with the pancreatic duct appearing as an echolucent circle (Figure
217A). To view the distal pancreas, left kidney, and spleen (see Stations 3 and 4, Figure 215), the
endosonographer rotates his or her shoulders to the right, adjusts the left or right control, and
withdraws the echoendoscope slightly. The tail of the pancreas can be visualized along with the splenic
vein coursing by the left kidney toward the splenic hilum (see Figure 217B). Generally, the splenic
artery is serpiginous, cephalad to the pancreas, and tends to be more circular than the splenic vein,
which is more oval and has a larger diameter (Figure 218). Deep to these vessels, the renal vein and
artery can be imaged; the vein is larger and closer to the transducer (see Station 4, Figure 215).
Further rotation and tip deflection allows imaging of the left kidney and spleen (see Figure 217B). The
left adrenal gland, readily identified on radial scanning,10(893) can also be imaged by linear array
scanning. It is located medial to the left kidney (see Station 5, Figure 215 and Figure 217C).
(894)Figure 217. Linear array endosonographic views of the abdomen using Pentax/Hitachi
FG-32 UA system at 7.5 MHz. All station numbers refer to Figure 215. A, Pancreatic
(PANC) body with pancreatic duct (PD) in cross section (Station 2). SAsplenic artery;
SVsplenic vein. B, Tail of pancreas (P) overlying the left kidney (K) (Stations 3 and 4). C,
Left adrenal gland showing typical elliptical shape (Station 5). (D1 is a distance marker and
has no significance.) D, Common bile duct (CBD) and pancreatic duct (PD) at the head of the
pancreas (Station 6).
(895)Figure 218. Color-flow mapping shows the splenic vein (SV) seen in blue and the
splenic artery (SA) in red at the level of the pancreatic body and tail at Station 3 of Figure
215.
After imaging the structures of the upper abdomen (see Stations 1 through 5, Figure 215), the
echoendoscope is advanced into the duodenum to imaging Stations 6 through 8. The pancreatic
uncinate process, head, and neck are imaged by withdrawing the echoendoscope in a continuous
sweep from the duodenum to the antrum and body of the stomach.
Station 6 in Figure 215 is imaged by identifying the ampulla of Vater endoscopically and then placing
the transducer against the mucosa with the water balloon filled. The uncinate process is identified just
distal to the ampulla. The ventral pancreas is often hypoechoic, and the appearance of the uncinate
process can mimic a tumor (i.e., pseudotumor).11(896) This should be kept in mind when performing
EUS-guided FNA. At the level of the ampulla, the head of pancreas is imaged with the common bile
duct (CBD) and pancreatic duct seen as parallel structures (see Figure 217D). The CBD is always
closer to the transducer. The extrapancreatic CBD is usually imaged adjacent to the distal duodenal
bulb (Figure 219A). By following the CBD from the level of the ampulla during gradual withdrawal of
the echoendoscope toward the duodenal bulb, the CBD can be evaluated for the presence of stones
(see Figure 219A). Preliminary studies using radial scanning echoendoscopes suggest that EUS may
be a less invasive and more cost-effective alternative to endoscopic retrograde
cholangiopancreatography (ERCP) for evaluating patients for choledocholithiasis before laparoscopic
cholecystectomy.12,13(897) The linear array instrument can also be used to diagnose
choledocholithiasis (see Figure 219A). The CBD at the level of the duodenal bulb is a good location to
perform EUS-guided cholangiography (see Figure 219B). The distal duodenal bulb is also the best
location for evaluating the head of pancreas for tumor and performing EUS-guided FNA.
(898)Figure 219. Linear array endosonographic views of abdomen using Pentax/Hitachi

FG-32 UA system at 7.5 MHz. All station numbers refer to Figure 215. A, Stone in the
common bile duct (CBD) seen parallel with the larger portal vein (PV) (Station 7). B,
Endoscopic ultrasonography (EUS)-guided cholangiogram. Echoendoscope (Pentax) is seen
fluoroscopically with a fine-needle aspiration (FNA) needle injecting contrast into the
common bile duct through the duodenal wall (Station 7). C, Portal vein (PV) with
characteristic venous hum on Doppler pulsewave (Station 7). CBDcommon bile duct;
PANChead/neck of pancreas. D, Neck of pancreas above portal vein (PV), with the superior
mesenteric artery (SMA) below (Stations 7 and 8).
The confluence of the portal and splenic veins is imaged from the duodenum (usually proximal second
portion or bulb) and is deep to the pancreatic head and neck. Portal vein and splenic vein Doppler
waveforms are similar and are characterized by a continuous venous hum that extends throughout
systole and diastole (see Figure 219C). Color-flow imaging at this level may enhance the ability to
assess portal vein invasion in patients with cancer arising in the pancreatic head (Figure 2110). The
neck of the pancreas is imaged from the duodenal bulb or the antrum. Decreasing the amount of water
in the balloon facilitates the gradual transition from the bulb to the antrum and prevents sudden
"popping" through the pylorus. At the level of the pancreatic neck, the superior mesenteric vein and
splenic vein may be seen merging into the portal vein (see Figure 219D). The SMA is usually
visualized as a longitudinal hypoechoic structure running parallel and below the superior mesenteric
vein and portal vein (see Figure 219D). The echoendoscope is then further withdrawn into the body of
the stomach, from where the pancreatic body is imaged. Occasionally, it is possible to orient the
transducer to image the pancreas and splenic vein transversely, in which case the SMA is seen in
cross section. This orientation is similar to that of radial scanning.
(899)Figure 2110. Color Doppler endosonogram shows turbulent flow in the portal vein
(PV) with invasion of the vessel by a malignant tumor (T) in the head of the pancreas. The
vessel wall adjacent to the tumor is irregular with the presence of pseudopods.
The other major vascular structures imaged through the duodenum are the inferior vena cava and
abdominal aorta. The inferior vena cava has a typical pulsatile, "sawtooth" waveform, with some
reversal of flow due to right atrial contraction. The gallbladder is best imaged from the duodenum or
antrum (Figure 2111). When conventional imaging studies have been unrevealing in patients with
unexplained biliary type pain, EUS imaging of the gallbladder may be useful to look for gallbladder wall
thickening or microlithiasis.14(900)
(901)Figure 2111. Gallbladder imaged through the duodenal bulb (Station 7) with a small
stone casting a shadow.
Linear array endosonography, which can perform EUS-guided FNA, EUS-guided injections, and color
mapping and Doppler ultrasonography, is becoming a valuable procedure. Familiarity with linear array
anatomy is the critical foundation on which these diagnostic and therapeutic procedures can be
performed safely and effectively.
Endoscopic Ultrasonography-Guided Fine-Needle Aspiration

EUS as an imaging modality is highly useful in the detection, characterization, and staging of
gastrointestinal and pancreatic malignancies.1520(902) However, EUS is not specific in differentiating
benign inflammation from malignant lesions. For example, in staging lymph nodes, EUS cannot
differentiate benign from malignant nodes with a high degree of certainty.2122(903) Similarly, EUS is
limited in its ability to distinguish between focal pancreatitis (pseudotumor) and pancreatic
cancer.23,24(904) In the evaluation of submucosal tumors, EUS can characterize the lesion, but it
does not give a definitive diagnosis in most cases, with the exception of submucosal varices. Overall,
the clinical utility of EUS alone is less than optimum because it lacks the specificity of a histologic
diagnosis.
The initial attempts to obtain a tissue diagnosis by EUS used a technique now referred to as
EUS-assisted FNA.25,26(905) Radial scanning EUS is first performed to determine the exact size,
location, vascularity, and consistency of a lesion. FNA is then attempted with a standard endoscope
and aspiration needle. Although shown to be effective, EUS-assisted FNA is limited to lesions that are
visible endoscopically; aspiration of small submucosal lesions and extrinsic structures such as lymph
nodes and the pancreas are precluded. This procedure also does not permit visualization of the needle
as it penetrates the lesion.
Echoendoscopes with a sector scan oriented along the long axis of the insertion tube (e.g., FG-32 UA,
Pentax/Precision Instruments) (see Figure 211) offer real-time visualization of the entire aspiration
needle and allow for performance of EUS-guided FNA. Unlike the EUS-assisted technique,
EUS-guided FNA can be used to sample lesions outside the gastrointestinal tract that do not create an
intraluminal bulge.
EUS-guided FNA was initially described in the aspiration of cystic pancreatic lesions in two case
reports (i.e., a large pancreatic pseudocyst and a mucinous cystadenoma).27,28(906) We reported
EUS-guided FNA in the diagnosis of a pancreatic carcinoma undetected by abdominal CT (Figures
2112 and 2113).29(907) EUS-guided FNA of mediastinal lymph nodes in a patient with lung cancer
was also described.30(908) These case descriptions were followed by reports of series of EUS-guided
FNA of various lesions within and adjacent to the gastrointestinal tract.15(909) Preliminary data
suggest that EUS-guided FNA enhances the diagnostic and staging capabilities of EUS in patients with
gastrointestinal and pulmonary malignancies.
(910)Figure 2112. Diagram of EUS-guided FNA of a lesion in the head of the pancreas.
Transducer is in the duodenal bulb, with the needle going through the duodenal wall into the
tumor.
(911)Figure 2113. Endosonographic view of EUS-guided FNA of a pancreatic tumor (T),

with a biliary stent seen deep to the tumor. Arrows show the FNA needle. Cytologic
examination showed adenocarcinoma.
All studies of EUS-guided FNA have been performed with the curved linear array echoendoscope
(FG-32 UA, Pentax Precision Instruments). Radial scanning echoendoscopes visualize an aspiration
needle only in cross section as a dot on the ultrasonographic field. Some ultrasonographic guidance is
possible with this type of echoendoscope, but the tip of the needle cannot be advanced under real-time
ultrasonographic visualization with this system.
Technique
The technique of EUS-guided FNA begins with positioning the transducer so that the target lesion is at
the 6 o'clock position. A sequence of steps is used for performing EUS-guided FNA with the
GIP/Medi-globe needle (Medi-globe, Tempe, AZ):
1. Obtain color map and Doppler flow studies of the lesion and surrounding structures to avoid
puncturing vascular structures.
2. Insert the catheter with needle and stylet into the accessory channel of the echoendoscope.
3. Anchor the handle to the inlet of the accessory channel.
4. Advance the needle with the stylet about 1 cm, until the stylet is seen in the ultrasound field, and
continue to advance until the stylet abuts the duodenal or gastric wall.
5. Withdraw the stylet 5 mm, allowing the sharp needle tip to be exposed.
6. Advance the needle with the retracted stylet in 1-cm increments through the gastrointestinal wall
into the lesion.
7. With the needle in the lesion, advance the stylet to its original position to "unplug" the needle tip,
and then withdraw the entire stylet.
8. Use a 10-ml Luer-Lok syringe with 5 ml of suction applied.
9. Move needle to and fro within lesion about five times with constant suction.
10. Release the suction, and gently allow the vacuum to equilibrate for a few seconds with the needle
still in the lesion.
11. Withdraw the needle into the catheter.
12. Unscrew the handle, and remove the catheter from the echoendoscope.
13. Use a syringe to expel one large drop of specimen onto each of two glass slides, which are
smeared and processed for Papanicolaou stain (reinsert the stylet if the specimen is plugged in
the needle).
14. The remaining specimen (including occasional core specimen) can be placed in formalin for a cell
block.
The needle can be advanced up to 10 cm beyond the catheter. The GIP/Medi-globe needle is similar to
the needle used by Vilmann et al.4(912) in a series of EUS-guided FNA of lesions in the upper
gastrointestinal tract. Prophylactic antibiotics are not routinely given, except for cystic lesions or lesions
along the lower gastrointestinal tract. The technique for EUS-guided FNA using the GIP/Medi-globe
needle is described in greater detail elsewhere.31(913)
Lesions in the head of the pancreas are best accessed from the duodenal bulb (see Figures 2112
and 2113), and lesions in the body or tail are best sampled through the posterior wall of the gastric
body and fundus. Partial distention of the transducer balloon with water can help to "anchor" the
echoendoscope tip in the duodenum during FNA of the pancreatic head. Care must be taken not to
fully distend the balloon with water, or it may be punctured by the needle. Partial distention of the
balloon yields a larger contact surface for acoustic coupling during FNA regardless of the anatomic
location.
For EUS-guided FNA of mediastinal lymph nodes, the transducer is placed in the esophagus, and the
lymph node is localized by withdrawing the echoendoscope while rotating the transducer in a circular
fashion. Celiac, hepatic, and splenic lymph nodes are approached through the posterior wall of the
gastric body and fundus (Figure 2114). Lymph nodes adjacent to the inferior vena cava and
midabdominal aorta may be targeted from the duodenum.
(914)Figure 2114. Endosonographic view of EUS-guided FNA of an enlarged celiac lymph

node.
Lesions and lymph nodes adjacent to the lower gastrointestinal tract can also be sampled by
EUS-guided FNA.1(915) Antibiotics should be given prophylactically before EUS-guided FNA from
rectum or colon.
Safety
Wiersema et al.32(916) performed a multicenter assessment of the complication rate for EUS-guided
FNA of 554 lesions in various anatomic locations in 435 patients. The overall complication rate was 1%
(5 of 435 patients) including 1 esophageal perforation due to dilation, 1 duodenal perforation, 1 cyst
hemorrhage, and 2 febrile episodes after EUS-guided FNA of cystic lesions. Four of these patients
required surgery, but there were no deaths. The complication rate was significantly higher for cystic
lesions. In another study specifically designed to assess the safety of EUS-guided FNA of the
pancreas, complications occurred in 4 (2%) of 164 patients.33(917)
The complication rate for EUS-guided FNA is similar to those of abdominal percutaneous FNA34(918)
or transrectal FNA of the prostate,35(919) each of which is approximately 1%. In EUS-guided FNA,
complications directly related to the needle aspiration may be less frequent because of more precise
positioning of the tip of the needle, which is visible ultrasonographically at all times; the shorter
insertion length of the needle; and the use of color Doppler imaging to avoid puncturing vessels.
However, the added risk of endoscopy with conscious sedation and the possible need for esophageal
dilation may offset this advantage.
With early prototype needles, numerous instances of echoendoscope damage resulting from puncture
of the accessory channel of the instrument were reported. Our current needle (GIP/Medi-globe) has a
metal coil catheter that extends just beyond the accessory channel at the distal end of the
echoendoscope and is attached to a handle that locks onto the proximal inlet of the accessory channel
(designed by Vilmann). This minimizes the possibility of inadvertent damage to the echoendoscope.
Malignant seeding along the needle track during percutaneous FNA, particularly in pancreatic
carcinoma, is a small albeit significant concern.36(920) However, because most malignant seeding
occurs cutaneously, this may be less of an issue with EUS-guided FNA. For example, lesions in the
head of the pancreas are sampled through the duodenum, and in patients who have resectable
tumors, the duodenum is removed during surgery. Nevertheless, the potential problem of malignant
seeding requires further study.
Adequacy of Cytologic Specimens

A specimen is defined as adequate for cytologic diagnosis only if sufficient numbers of cells from the
target site are detected in the aspirate. Based on our recent series37(921) and on multicenter
experience, adequate specimens are obtained from more than 90% of patients. In most instances of
inadequate specimens, epithelial cells from the mucosal surface of the bowel were obtained. The site
of FNA did not appear to have a bearing on specimen adequacy. However, the average number of
needle passes required to obtain adequate tissue from pancreatic lesions is higher compared with that
for lymph node sampling.33(922)

Certain tissue characteristics of lesions, the type of needle used, and the presence of a cytologist
during the procedure are factors that influence the adequacy of the specimen obtained by EUS-guided
FNA. EUS-guided FNA is more difficult when a lesion has associated fibrosis, necrosis, or surrounding
inflammation. With fibrotic or necrotic lesions, placing the needle at the periphery of the lesion may
sometimes provide a better yield than targeting the center of the tumor.
The force generated at the handle of the needle must be transmitted to the needle tip to pierce fibrotic
areas. Some earlier prototype needles used polyethylene tubing to connect the needle with the handle.
This did not allow adequate responsiveness at the working end of the needle. The GIP/Medi-globe
needle consists of a solid stainless steel, 22-gauge needle that is secured onto the handle; this
optimizes control of the needle tip.
The presence of a cytology technician or cytopathologist during EUS-guided FNA is important in
obtaining an adequate specimen. In our initial experience, 18 patients underwent procedures without a
cytologist present, and 6 required repeat procedures because of inadequate specimens. In contrast,
specimens adequate for cytologic diagnosis were obtained in all 20 patients who underwent the
procedure with a cytologist in attendence.1(923) After each FNA pass, the cytologist assesses the
cellularity of the specimen, determines the types of cells present, and assists the endosonographer in
deciding whether additional passes are needed. In our experience, as many as 12 passes have been
necessary before the cytologic diagnosis of cancer could be made. Although a cytologic diagnosis of
malignancy terminates the procedure, the number of passes required for diagnosis of a benign lesion
with EUS-guided FNA is less certain and depends on the clinical circumstances and the index of
suspicion of malignancy.
Diagnostic Accuracy
The diagnostic accuracy was 87% for all lesions targeted by EUS-guided FNA in our initial series of 38
patients with a sensitivity and specificity for malignant lesions of 91% and 100%, respectively.1(924)
Wiersema et al.2(925) found an overall diagnostic accuracy of 90% for 18 of 20 patients for whom
surgical confirmation or clinical follow-up was available.
Data concerning the diagnostic accuracy for specific target lesions have become available as the
number of studies of EUS-guided FNA has grown. The pancreas is a target organ of particular interest
for EUS-guided FNA. Rsch et al.23(926) found that EUS alone did not fully differentiate malignant
from inflammatory pancreatic masses. In this study of 132 patients, the overall accuracy was 76% for
malignancy and 46% for focal inflammation. In our collaborative study of EUS-guided FNA of the
pancreas in 164 consecutive patients from four centers, the sensitivity, specificity, diagnostic accuracy,
negative predictive value, and positive predictive value were 83%, 90%, 85%, 80%, and 100%,
respectively. These results were significantly superior to CT, for which the corresponding values were
56%, 37%, 50%, 28%, and 65% respectively.33(927)
EUS-guided FNA of lymph nodes is also highly accurate. In our experience with 29 periesophageal and
celiac lymph nodes, the specimen was adequate for cytologic examination of 93%, and the overall
diagnostic accuracy was 90% (18 of 20) (see Figure 2114).37(928) The diagnostic accuracy for
EUS-guided FNA of periesophageal nodes was 78%, and that for enlarged celiac nodes was 100%.
Seven (54%) of 13 celiac nodes were not detected by CT. In the series of Wiersema et al.,2(929)
seven patients underwent EUS-guided FNA of mediastinal lymph nodes, and two of seven had no
evidence of mediastinal lymphadenopathy on CT. The specimen was adequate in all seven cases, and
the diagnostic yield was 100% (4 of 4). Mediastinal lymph node sampling is not limited to
gastrointestinal malignancies; EUS-guided FNA may also be useful in staging patients with lung
cancer.38(930)
When sampling lymph nodes associated with a primary gastrointestinal malignancy, a potential risk
exists for contaminating the specimen with the cancer cells from the primary tumor as the needle
traverses the bowel wall. To avoid such contamination, only those lymph nodes adjacent to uninvolved
segments of the bowel wall should be sampled.
We performed EUS-guided FNA in 17 patients with submucosal lesions in our initial series.37(931)
Endoscopy with procurement of biopsy specimens was attempted in these cases before EUS-guided
FNA, except for 1 patient with a pancreatic pseudocyst. An adequate specimen was obtained in all 17.
The types of lesions sampled were gastric ulcer or inflammation (5), esophageal cancer (2), gastric
cancer (2), gastrinoma (2), lipoma (1), gastric lymphoma (1), ampullary cancer (1), rectal inflammation
(1), and unknown (2). The overall diagnostic accuracy of EUS-guided FNA in our initial experience was
significantly lower for submucosal (64%) than for extrinsic lesions (93%). This result has been
confirmed in a multicenter study in which the overall sensitivity for the diagnosis of malignant disease
was 60% for intraluminal and 90% for extraluminal lesions (p < .001).32(932)
Because EUS-guided FNA is less efficacious in the diagnosis of submucosal lesions, other means of
obtaining a tissue diagnosis may be needed. Despite the lower diagnostic yield compared with extrinsic
lesions, EUS-guided FNA may still be worthwhile because a cytology specimen positive for malignancy
has great clinical significance and the alternatives for obtaining a histologic diagnosis may be more
invasive. These include endoscopic biopsy techniques such as large particle, Tru-Cut, and snare
biopsy. Lesions with normal overlying mucosa may still be inaccessible by these techniques, and
surgery then becomes the only alternative.
EUS-guided FNA has been used to sample pleural and ascitic fluid collections detected by EUS.6(933)
In many instances, the fluid detected at EUS is not appreciated by more established imaging studies
such as transabdominal ultrasonography and CT. EUS-guided FNA can also be used to obtain
histologic confirmation of liver metastasis (Figure 2115).
(934)Figure 2115. Endosonographic view of EUS-guided FNA of a 5-mm lesion in the left
lobe of the liver in a patient with esophageal cancer. The needle tip is seen within the lesion.
Cytologic examination revealed metastatic squamous cell carcinoma.
Although the positive predictive value of EUS-guided FNA should approach 100%, the negative
predictive value is influenced by the experience of the endosonographer and cytologist, the type of
equipment available, and the nature of clinical conditions for which EUS-guided FNA is being
performed. In our multicenter trial of EUS-guided FNA of the pancreas, the negative predictive value
was 79%. A negative FNA result is not completely reassuring with regard to patient management, and
the value of this finding is related to the clinical setting and to the skill and experience of the
endosonographer and cytologist.
Clinical Utility
EUS-guided FNA may play an important role in the diagnosis of primary lesions in situations for which
conventional modalities are unsuccessful or technically difficult. These include small pancreatic tumors
not detected by CT or transabdominal ultrasonography, periluminal lymph nodes, and submucosal
gastrointestinal lesions. For example, a CT scan may reveal only nonspecific enlargement of the
pancreas and clinical decision making based on such a finding is often difficult. EUS then becomes an
attractive alternative to ERCP, with its possible complication of pancreatitis and its lower diagnostic and
cytologic yield.
Among patients undergoing EUS-guided FNA of the pancreas at our institution, surgery was precluded
for 38% because of the additional information provided by this procedure.39(935) These initial data
suggest significant clinical and economical impacts of EUS-guided FNA on the diagnosis and staging
of pancreatic carcinoma.40(936)
EUS-guided FNA of celiac or periluminal lymph nodes can provide histologic confirmation of
involvement by malignancy.8(937) With esophageal, gastric, and pancreatic carcinomas, such
involvement carries a poor prognosis and generally excludes patients from curative surgery. In our
series of 14 patients with enlarged celiac nodes, malignancy was detected by EUS-guided FNA in
11.37(938) A diagnosis of liver metastasis can be obtained by EUS-guided FNA; aspiration of pleural
and peritoneal fluid may provide a positive cytologic diagnosis of the spread of malignancy.6(939) In
the preoperative staging of gastrointestinal and pulmonary malignancies, the addition of EUS-guided
FNA of lymph nodes, pleural and peritoneal fluid, and hepatic lesions increases the clinical value of
EUS compared with EUS without FNA.
The nonsurgical alternative to EUS-guided FNA of lymph nodes and pancreatic lesions is CT or
transabdominal ultrasonography-guided FNA. EUS has a higher sensitivity than either of these
modalities. For FNA of pancreatic masses under CT or transabdominal ultrasonography guidance, the
needle is usually directed at the center of the mass to ensure that it is within the lesion because of the
limited resolution of these imaging techniques. We have found that sampling the periphery of the tumor
may at times give a better yield; approaching the tumor from different angles may also be necessary.
These techniques of sampling are more difficult under CT or transabdominal ultrasonography
guidance. In the detection of large pancreatic carcinomas, ERCP is comparable to EUS; however,
ERCP has a lower sensitivity for small pancreatic tumors and offers little information regarding
vascular invasion and lymph node status.19,24,41(940) The various methods for obtaining cytologic
specimens during ERCP generally have lower yields than CT-guided FNA for pancreatic
carcinoma.42,43(941) Many patients require EUS for local staging when CT demonstrates no obvious
metastasis, particularly in esophageal, gastric, and pancreatic malignancies. Rather than having the
patient undergo a second CT for FNA or an ERCP, FNA can be performed at the time of tumor staging
by EUS.
Future Modalities
EUS-guided cholangiography can be performed when cannulation of the bile duct is difficult or
unsuccessful. The CBD is identified adjacent to the duodenal wall by EUS, and the fine needle is
guided directly into the duct, with subsequent injection of a radiographic contrast medium (see Figure
219B).
Celiac nerve block or neurolysis can be performed under EUS guidance (Figure 2116). After the
celiac trunk has been visualized, a needle is inserted under EUS guidance, followed by injection of
bupivacaine and absolute alcohol into the region. Wiersema et al.9(942) performed this technique in 20
patients with intractable abdominal pain owing to intraabdominal malignancy. In this preliminary study,
a lessening of pain occurred in 18 (90%) of 20 patients. For 86%, this benefit was found to have
persisted at 8 weeks' follow-up. Four patients experienced transient diarrhea, but no other side effects
or complications occurred.
(943)Figure 2116. Diagram of EUS-guided celiac nerve block.
REFERENCES
1.
2.
3.
Chang KJ, Katz KD, Durbin TE, et al. Endoscopic ultrasound-guided fine-needle aspiration.
Wiersema MJ, Kochman ML, Cramer HM, et al. Endosonography guided real-time fine-needle
aspiration biopsy. Gastrointest Endosc 1994;40:7007.
Wegener M, Adamek RJ, Wedmann B, Pfaffenbach B. Endosonographically guided fine
needle aspiration puncture of paraesophagogastric mass lesions: Preliminary results.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
Endoscopy 1994;26:58691.
Vilmann P, Hancke S, Henriksen FW, Jacobsen GK. Endoscopic ultrasonography guided
fine-needle aspiration biopsy of lesions in the upper gastrointestinal tract. Gastrointest Endosc
1995;41:2305.
Giovannini M, Seitz JF, Monges G, et al. Fine-needle aspiration cytology guided by
endoscopic ultrasonography: Results in 141 patients. Endoscopy 1995;27:1717.
Chang KJ, Albers CG, Nguyen P. Endoscopic ultrasound-guided fine needle aspiration of
pleural and ascitic fluid. Am J Gastroenterol 1995;90:14850.
Glazer GM, Gross BH, Quint LE, et al. Normal mediastinal lymph nodes: Number and size
according to American Thoracic Society mapping. AJR 1985;144:2615.
Chang K, Durbin T, Katz K, et al. Endoscopic ultrasound (EUS) guided fine needle aspiration
(FNA) of upper gastrointestinal peri-luminal and celiac lymph nodes (LN) [Abstract 190].
Gastrointest Endosc 1994;40:P62.
Wiersema M, Sandusky D, Carr R, et al. Endosonography guided celiac plexus neurolysis
(EUS CPN) in patients with pain due to intra-abdominal (IA) malignancy [Abstract].
Chang KJ, Erickson RA, Nguyen P. Endoscopic ultrasound (EUS) and EUS-guided fine
needle aspiration of the left adrenal gland. Gastrointest Endosc 1996;44:56872.
Savides TJ, Gress FG, Zaidi SA, et al. EUS detects the pancreatic ventral anlage [Abstract
66]. Gastrointest Endosc 1995;41:312.
Canto M, Chak A, Sivak MV. Endoscopic ultrasonography (EUS) versus cholangiography for
diagnosing extrahepatic biliary stones: A prospective, blinded study in pre- and
post-cholecystectomy patients [Abstract 383]. Gastrointest Endosc 1995;41:391.
Erickson RA, Chang KJ. ERCP, EUS (MathematicalPi-One )1 ERCP, or MRCP
(MathematicalPi-One )1 ERCP prior to laparoscopic cholecystectomy (LC): A cost-benefit
analysis [Abstract]. Gastrointest Endosc 1996;43:311.
Dill J. Comparative accuracy of endoscopic ultrasound, transabdominal ultrasound and biliary
drainage in the diagnosis of cholecystitis. Gastroenterology 1995;108:A414.
Rice TW, Boyce GA, Sivak MV, et al. Esophageal carcinoma: Esophageal ultrasound
assessment of preoperative chemotherapy. Ann Thorac Surg 1992;53:9727.
Akahoshi K, Misawa T, Fujishima H, et al. Preoperative evaluation of gastric cancer by
endoscopic ultrasound. Gut 1991;32:47982.
Boyce GA, Sivak MV Jr, Lavery IC, et al. Endoscopic ultrasound in the pre-operative staging
of rectal carcinoma. Gastrointest Endosc 1992;38:46871.
Rsch T, Lightdale CJ, Botet JF, et al. Localization of pancreatic endocrine tumors by
endoscopic ultrasonography. N Engl J Med 1992;326:17216.
Rsch T, Lorenz R, Braig C, et al. Endoscopic ultrasound in pancreatic tumor diagnosis.
Rsch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by
endoscopic ultrasonography. Comparison with conventional sonography, computed
tomography, and angiography. Gastroenterology 1992;102:18899.
Heintz A, Mildenberger P, Georg M, et al. Endoscopic ultrasonography in the diagnosis of
regional lymph nodes in esophageal and gastric cancerResults of studies in vitro.
Endoscopy 1993;25:2315.
Grimm H, Hamper K, Binmoeller KF, Soehendra N. Enlarged lymph nodes: Malignant or not?.
Endoscopy 1992;24(suppl 1):3203.
Rsch T, Lorenz R, Braig C, et al. Endoscopic ultrasonography in diagnosis and staging of
pancreatic and biliary tumors. Endoscopy 1992;24(suppl 1):3048.
Yasuda K, Mukai H, Fujimoto S, et al. The diagnosis of pancreatic cancer by endoscopic
ultrasonography. Gastrointest Endosc 1988;34:18.
Wiersema MJ, Hawes RH, Tao LC, et al. Endoscopic ultrasonography as an adjunct to fine
needle aspiration cytology of the upper and lower gastrointestinal tract. Gastointest Endosc
1992;38:359.
26. Caletti GC, Brocchi E, Ferrari A, et al. Guillotine needle biopsy as a supplement to
endosonography in the diagnosis of gastric submucosal tumors. Endoscopy 1991;23:2514.
27. Vilmann P, Jacobsen GK, Henriksen FW, et al. Endoscopic ultrasonography with guided fine
needle aspiration in pancreatic disease. Gastrointest Endosc 1992;38:1723.
28. Grimm H, Binmoeller KF, Soehendra N. Endosonography-guided drainage of a pancreatic
pseudocyst. Gastrointest Endosc 1992;38:1701.
29. Chang KJ, Albers CG, Erickson RA, et al. Endoscopic ultrasound-guided fine needle
aspiration of pancreatic carcinoma. Am J Gastroenterol 1994;89:2636.
30. Wiersema MJ, Kochman ML, Chak A, et al. Real-time endoscopic ultrasound guided
fine-needle aspiration of a mediastinal lymph node. Gastrointest Endosc 1993;39:42931.
31. Chang KJ. Endoscopic ultrasound-guided fine needle aspiration in the diagnosis and staging
of pancreatic tumors. Gastrointest Endosc Clin North Am 1995;5:72334.
32.
Wiersema M, Vilmann P, Giovannini M, Chang K. Prospective multicenter evaluation of EUS
guided fine needle aspiration biopsy (FNA): Diagnostic accuracy and complication (CX)
assessment (Abstract). Gastrointest Endosc 1996;43:432.
33.
Chang KJ, Wiersema M, Giovannini M, et al. Multi-center collaborative study on endoscopic
ultrasound (EUS) guided fine needle aspiration (FNA) of the pancreas (abstract). Gastrointest
Endosc 1996;43:417.
34. Edoute Y, Ben-Haim SA, Malberger E. Value of direct fine needle aspirative cytology in
diagnosing palpable abdominal masses. Am J Med 1991;91:37782.
35. Mondal A, Ghosh E, Ghose A. The role of transrectal fine needle aspiration cytology in the
diagnosis of prostatic nodules suspicious of malignancyA study of 126 cases. Indian J
Pathol Microbiol 1990;33:239.
36. Fornari F, Civardi G, Cavanna L, et al. Complication of ultrasonically guided fine-needle
abdominal biopsy. Results of a multicenter Italian study and review of the literature. The
Cooperative Italian Study Group. Scand J Gastroenterol 1989;24:94955.
37.
Chang KJ, Durbin TE, Katz KD, et al. The clinical value of endoscopic ultrasound (EUS)
guided fine needle aspiration (FNA) in gastrointestinal and pulmonary malignancies [Abstract
355]. Am J Gastroenterol 1994;89:1704.
38.
Gress F, Savides T, Ikenberry S, et al. A prospective comparison study of endoscopic
ultrasound (EUS), computed tomography (CT) and EUS directed fine needle aspiration biopsy
(EUS/FNA) of the mediastinum in the preoperative staging of non-small cell lung cancer
(NSCLCA) [Abstract 34]. Gastrointest Endosc 1995;41:304.
39.
Chang KJ, Nguyen P, Durbin TE, et al. The utility of endoscopic ultrasound (EUS) guided fine
needle (FNA) in the diagnosis and staging of pancreatic carcinoma. World Congress
Gastroenterol Abstr 1994;I:75EC.
40.
Chang KJ, Nguyen P, Erickson RA, et al. The clinical and economic impact of endoscopic
(EUS) guided fine needle aspiration [Abstract]. Gastroenterology 1995;108(suppl):A455.
41. Niederau C, Grendell JH. Diagnosis of pancreatic carcinomaImaging techniques and tumor
markers. Pancreas 1992;7:6686.
42. Ryan ME. Cytologic brushings of ductal lesions during ERCP. Gastrointest Endosc
1991;37:13942.
43. Scudera PL, Koizumi J, Jacobson IM. Brush cytology evaluation of lesions encountered during
ERCP. Gastrointest Endosc 1990;36:2814.
Chapter 22 Flexible Endoscope Technology: Three-Dimensional,

Nonradiographic Imaging
(944)
CHRISTOPHER B. WILLIAMS, B.M., F.R.C.P.
BRIAN P. SAUNDERS, M.B., M.R.C.P.
G. DUNCAN BELL, M.D., F.R.C.P.

DUNCAN F. GILLIES, PH.D.
CHRISTOPHER N. GUY, PH.D.
JOHN S. BLADEN, PH.D.
Background: The Problem

The flexible endoscopes that evolved in the 1960s were intended principally for use in the stomach and
for reaching previously hidden areas such as the duodenum or the gastric fundus. The predictable
anatomy and fixations of the stomach have made flexible upper gastrointestinal endoscopy relatively
simple from a technical standpoint. The same cannot be said of the colon, where variations in length
and retroperitoneal fixation allow extraordinarily variable and sometimes painful looping to occur as the
colonoscope is inserted.
Skilled endoscopists perform colonoscopy empirically through a combination of maneuvers, including
retracting or twisting the insertion tube of the instrument, changing the patient's position, or
compressing the abdominal wall, to conjure the straightened instrument through the whole length of the
colon in more than 95% of patients. However, difficulty and even trauma is encountered in perhaps 10
to 20% of cases. Less-skilled endoscopists may achieve total colonoscopy in 75% or fewer patients, a
success rate diminished by the need for higher doses of sedative drugs and the potential for
complications. Even experienced endos-copists must sometimes guess the instrument's configuration
and the location of its tip. During insertion of the instrument, in as many as one third of patients the
endoscopist's sense of location may be grossly disorientated, and she or he can mistake the
sigmoid-descending colon junction for the splenic flexure, or the splenic flexure for the hepatic flexure.
In good faith, the endoscopist may believe she or he has performed colonoscopy to the cecum,
although the colonoscope has reached no farther than the mid-transverse colon or hepatic flexure.
Clues about location can be misleading; the triangular configuration associated with the transverse
colon can be seen in the descending colon; the blue extracolonic coloration typical of the hepatic
flexure can be seen at the splenic flexure or in the sigmoid colon. The only fixed and definite landmark
is the ileocecal valve, and until this is identified, the endoscopist can be uncertain about the location of
the instrument tip in the colonand more seriously, mistaken on some occasions when she or he is
certain. Even when tip localization is judged correctly, there is no way of knowing what loops may have
formed or of predicting how best to straighten them.
Errors of localization may result in inappropriate maneuvers and technical difficulty during insertion of
the colonoscope. Information given to a surgeon may be misleading, a possibility that has led some
surgeons, especially laparoscopic surgeons, to insist on a preoperative barium enema to accurately
localize a tumor that has been identified endoscopically.
Although the anatomic uncertainties of colonoscopy often make the procedure difficult or traumatic,
teaching another physician to perform colonoscopy can be a nightmare. The "expert" standing near the
pupil can make only a best guess about the situation, knowing that corrective advice may be wrong. As
a result, skilled colonoscopy has remained an art rather than developing into a predictable science.
Since the mid-1970s, endoscopists have debated the merits and demerits of using fluoroscopy during
colonoscopy. The few endoscopists who have x-ray equipment in the endoscopy suite know the
confidence in localization and logic of maneuver that imaging brings in some examinations, not to
mention its considerable value during the learning phase of colonoscopy. Most endoscopists do not
have fluoroscopy equipment, and they correctly point out the advantages of developing simplified,
dexterous endoscopic technique, unencumbered by lead aprons, darkened rooms, and the potential
hazard of radiation.
Even enthusiasts for fluoroscopy admit its limitations. The approximately 9-inch field size visible on the
x-ray monitor, covering less than a quarter of the abdomen, can give an entirely misleading partial view
of convoluted loops, especially if the patient is not placed supine. The two-dimensional fluoroscopic
image provides less than ideal information when an endoscopist seeks to control loops of bowel that
are three dimensional (3-D). It is difficult to apply hand pressure to a loop and view it simultaneously
without exposing the assistant to radiation.
Development of Nonradiographic Imaging

We were frustrated by recurrent difficulties in performing and teaching colonoscopy, despite
fluoroscopy being available to both of our groups. We challenged physicist colleagues in 1989 and
1990 with the need for a nonradiographic solution for imaging the colonoscope within the abdomen.
The St. Mark's Hospital and Imperial College group, already working on microcomputer teaching
simulators, "intelligent endoscopes," and other possibilities, started to modify the electronics and
software of an existing single-coil electromagnetic device under development for other purposes. The
Ipswich Hospital and the Sheffield University group, with assistance from British Telecom, initially
investigated the feasibility of a fiberoptic solution before also coming to the conclusion that
electromagnetic methodology was preferable. Laboratory experimentation by both groups ended in
limited clinical trials of prototype systems in late 1992 (Figure 221), followed by simultaneous
publication of the results of these trials.1,2(945) Evaluation of fully functional systems began in early
1994.
(946)Figure 221. Schematic diagram of electronic imager system.

Our rival systems are so similar in final format that we hereafter describe their general principles as
one. Preliminary evaluation has been a joint venture.
Principles of the Imaging Systems

The underlying basis of both imager systems is that of Faraday's demonstration that a changing
magnetic field induces electric current to flow in a coil of wire exposed to it. Large-sized coils mounted
below the patient examination table generate low-frequency, low-power magnetic fields that are
detected by miniature sensor coils within the endoscope. Because the strength and direction of the
fields vary with position, the information from each miniature sensor coil enables its position and
orientation to be computed. The positional information from each of about 15 sensor coils within the
instrument, together with knowledge of the mechanical characteristics of the instrument tip and
insertion tube, allow the 3-D configuration of the whole endoscope to be determined using a point
location algorithm.1,3(947) The prototype systems produce effectively real-time images at about 5
frames/sec, but it is expected that a 30 frames/sec frequency will be achieved in development.
Because the system uses low-power and low-frequency currents, there is no electromagnetic hazard.
The multiple-point 3-D image is rendered by the computer graphics unit to resemble the insertion tube
of the instrument and is displayed on the monitor in simulated 3-D by using different intensities of gray
scale, ranging from white (nearest to the viewer) to black (farthest from the viewer) (Figure 222). For
a more detailed understanding of particular loops or anatomic variants, the positional information
displayed on the monitor can be "rotated" to present a lateral or oblique view. Alternately,
anteroposterior and lateral real-time images can be viewed synchronously.
(948)Figure 222. Ipswich/Sheffield electronic imager "three-dimensional graphic," with

simultaneous x-ray image for comparison. (From Bladen JS, Anderson AP, Bell GD, et al.
Non-radiological technique for three-dimensional imaging of endoscopes. Lancet 1993;
341:71922. by The Lancet Ltd., 1993.)
One or more body-marker coils can be used in addition to the endoscope sensor array. With these
coils, the endoscopist registers on-screen positions of anatomic landmarks, such as the costal
margins, the lumbar spine, or the anus. A marker coil on the assistant's hand can be used to show the
efficacy of various degrees of pressure on particular loops. Alternatively, markers can be used to
maintain the screen view of the endoscope in anatomic anteroposterior configuration for easier
interpretation, even though the patient may be placed in different positions.
The sensor coils for prototype evaluation are mounted within a catheter inserted into the accessory
channel of the endoscope, ideally a "large-channel" or two-channel instrument that can maintain
suction. The coils could be incorporated into the insertion tube of the endoscope during manufacture or
repair. The prototype imagers have used a separate monitor screen, but for clinical purposes, it may
be more convenient, using a video card and mixer, to place the 3-D graphic adjacent to the endoscopic
image on the video monitor.
The image data from each examination can be stored and thereafter retrieved to review technique and
measure anatomic relationships and instrument movements. This could lead to a better understanding
of the principles of flexible endoscopy. The complete data record of several examinations can be
transferred onto floppy disk for transport and analysis. Loops and straightening maneuvers involved in
a procedure can be reviewed retrospectively on the monitor screen from any view, including
anteroposterior, lateral, oblique, and axial orientations. Precise anatomic measurements to within a few
millimeters can be made of the distance between any sensor coil and any body marker, illustrating the
degree to which the configuration of the intestine with its attachments can be modified by the action of
the colonoscope. It is also easy to reproduce 3-D graphics of selected endoscope configurations for
teaching purposes (Figure 223).
(949)Figure 223. Three-dimensional graphics from the electronic imager used for a teaching
presentation.
Effect on Performance and Teaching of Colonoscopy

Initial results show that electronic imaging has considerable influence on the performance of
colonoscopy.4(950) The on-screen image gives the endoscopist exact, real-time information on the
location of the endoscope tip and the configuration of the insertion tube. In "difficult" examinations of
patients with redundant colons, the confidence of seeing odd loops forming enhances the examiner's
ability to make appropriate maneuvers and to target hand pressure, which make the procedure
substantially easier. Derotation and straightening of complex loops becomes logical, whether by
clockwise or counterclockwise twisting as suggested by the 3-D display, rather than remaining
empirical as in the trial-and-error approach. Some repeated examinations regarded as exceedingly
difficult based on prior experience have been made quick and easy by use of the imager. For particular
bends or loops, the manipulative dexterity and close-up visual judgment of the expert endoscopist are
still essential, but the "extra dimension" given by imager information is enormously helpful and
reassuring.
Real-time 3-D imaging has its greatest impact on the learning process. For example, the inexperienced
endoscopist sees the folly of pushing in farther when an N loop is forming in the sigmoid colon and the
patient has discomfort. The relevance of torque or twisting movements and their simultaneous
maintenance during insertion or withdrawal become logical, rather than a mystical trick of the "expert."
The endoscopist need no longer guess the location of the instrument tip. Sometimes, this means that
an examination can be completed more rapidly or that complete insertion can be ensured by combining
imager and endoscopic views in cases for which the anatomy is atypical.
Formal evaluation of the 3-D nonradiographic imagers is in progress. However, it is evident on using
the system or seeing others using it that it represents a significant step forward in the development of
easier and more accurate colonoscopy.
Other Possible Endoscopic Applications

Current catheter-mounted sensor coil arrays have the obvious disadvantage that, except with
endoscopes with very large or twin accessory channels, they prevent effective suction during use.
Further miniaturization of the coils or installation of the coil array within the insertion tube of the
endoscope could make the electronic imager more widely applicable, including uses other than
colonoscopy. In the operated stomach, for instance, the imager could show which loop of a Billroth type
II anastomosis is being examined. Even during conventional gastroscopy, a "bird's eye view" of the
instrument looping in the greater curve or of the tip angling around into the duodenum should enable
less-experienced endoscopists to become more skillful. For enteroscopy, especially of the "push"
variety, seeing the configuration of the insertion tube and the position of the tip may increase the speed
and efficacy of the procedure. Precise demonstration of the axis of the ultrasound endoscope should
simplify the interpretation of endosonography images.
Miniaturized sensor coils have potential for imaging accessories (e.g., balloons, guidewires, overtubes)
when these must be passed out of the endoscopist's view. Because the coils and their wire
connections are simple, cheap, and small, there is no reason they cannot be incorporated into
disposable devices or fixed onto existing ones.
REFERENCES
1.
2.
3.
4.
Williams CB, Guy C, Gillies D, et al. Electronic three-dimensional imaging of intestinal

endoscopy. Lancet 1993;341:7245.
Bladen JS, Anderson AP, Bell GD, et al. Non-radiological technique for three-dimensional
imaging of endoscopes. Lancet 1993;341:71922.
Bladen JS, Anderson AP, Bell GD, Heatley DJ. A non-radiological technique for the real-time
imaging of endoscopes in three dimensions. Conference Record of the IEEE Nuclear Science
Symposium and Medical Imaging Conference, San Francisco, 1993;18914.
Saunders BP, Bell GD, Williams CB, et al. First clinical results with a real-time electronic
imager as an aid to colonoscopy. Gut 1995;36:9137.
Chapter 23 Needle Biopsies of Submucosal Lesions

(951)
GIANCARLO CALETTI, M.D.
PAOLO BOCUS, M.D.
ENRICO RODA, M.D.
Most lesions of the gastrointestinal tract can be recognized by their endoscopic appearances. A
clinician's suspicion about the nature of a lesion usually can be confirmed by obtaining endoscopic
biopsies and brush cytology specimens. The use of these two sampling techniques, alone or together,
results in correct identification of approximately 95% of malignant lesions.13(952) A small proportion
of tumors remain undiagnosed by these conventional techniques, because only surface cells or tissues
are sampled. Lesions that are primarily infiltrative, such as linitis plastica, recurrent tumors, and less
common histologic types such as lymphomas and sarcomas, are difficult to diagnose from standard
biopsy or brush cytology specimens.24(953) When used in combination, these techniques give
false-negative results for 5% of lesions,510(954) but this percentage increases to as high as 50% for
infiltrative or stenotic lesions,4,7,10(955) malignancies with necrotic surfaces,7(956) and submucosal
tumors such as lymphomas and sarcomas.4,11,12(957)
An endoscopically identified "submucosal tumor" may be a benign or malignant pathologic growth
within the gastrointestinal wall itself or may be the result of compression by an adjacent normal or
pathologic organ. Radiologic and endoscopic examinations evaluate only the luminal appearance of a
nonepithelial neoplasm. Multiple endoscopic biopsy specimens are often inadequate for a histologic
diagnosis because of insufficient tissue depth and the small size of these samples.
Techniques for obtaining biopsies of submucosal lesions are difficult. Despite an intended effort to
obtain submucosal tissue at endoscopy, Hedenbro et al.13(958) made a definitive diagnosis based on
biopsies of only 35% of such lesions. One explanation may be that use of a standard biopsy forceps for
this purpose provides specimens of inadequate size. However, Kaneko et al.14(959) and Siegel et
al.15(960) used much larger biopsy forceps with no greater success. More aggressive techniques for
obtaining biopsies have been proposed, but the safety of these methods is uncertain, and they are of
no value in cases of extrinsic compression.16(961)
Because of the limitations of standard methods of tissue sampling in certain disorders, several devices
have been developed to obtain biopsy and cytology specimens. These methods include aspiration
needle biopsy and guillotine needle biopsy.
Aspiration Needle Cytology

Aspiration needle biopsy is a form of aspiration cytology that is increasingly used to obtain tissue during
bronchoscopy. The so-called skinny needle (22 gauge) is used routinely by radiologists to obtain tissue
specimens from sites previously regarded as inaccessible.1722(962) The availability of flexible
needles for injecting esophageal varices raised the possibility that needle aspiration cytology could be
applied to gastrointestinal lesions visualized endoscopically (Figure 231). This technique makes it
possible to obtain samples from lesions lying deep to necrotic debris or normal mucosa that are difficult
to diagnose by conventional sampling methods.
(963)Figure 231. Flexible needle for injecting esophageal varices; used for aspiration
cytology (Olympus NM-12L, 21 G, 8 mm).
Technique
Aspiration samples are taken using a transbronchial aspiration needle (21 gauge, 13 mm long) or a
sclerotherapy needle (23 gauge, 4 mm long), both of which are retractable at their distal ends. A 10-ml
disposable syringe is attached at the proximal end of the needle.
After visualizing the target lesion, the needle with the tip retracted is passed through the accessory
channel of the endoscope. The needle is then protruded beyond the protective sheath and inserted into
the lesion. Suction is applied with the syringe while the needle is plunged in and out of the lesion. After
easing the suction, the needle is retracted, and the whole unit is withdrawn from the endoscope. The
aspirated material is ejected onto clean glass slides, air dried, stained with May-Grnwald-Giemsa
stain, and examined by an experienced cytologist. If concomitant brush cytology specimens or forceps
biopsies must be obtained, the sampling techniques should be employed in the following order: needle
aspiration, brush cytology, and forceps biopsy.
Results
A number of investigators have advocated endoscopic aspiration cytology as a way of increasing
diagnostic yield when endoscopic biopsies and brush cytology specimens are obtained in the upper
gastrointestinal tract.
Using aspiration needle cytology, Iishi et al.17(964) obtained a positive diagnosis of malignancy in 11
cases of diffusely infiltrative carcinoma of the stomach for which no correct diagnosis could be made
by prior histologic or cytologic examination. Similarly, endoscopic needle aspiration cytology was the
only sampling method that provided the correct diagnosis for 2 of 10 adenocarcinomas of the upper
gastrointestinal tract reported by Ingoldby et al.23(965) Kochhar et al.24(966) also successfully applied
this technique to the diagnosis of esophageal and gastric carcinomas in 10 patients and found it
especially useful in diagnosing a case of necrotic esophageal carcinoma for which conventional biopsy
and brush specimens had been inconclusive. The diagnostic yield with forceps biopsy, brush cytology,
and needle aspiration cytology, as reported by Kochhar et al.,25(967) for 46 cases of
gastroesophageal malignancies was 88.8%, 80.4%, and 89.1%, respectively. The combination of
forceps biopsy and brush cytology gave an accuracy of 93.5%, and the addition of fine-needle
aspiration cytology increased the yield to 100%.
Graham et al.26(968) compared the diagnostic yields of forceps biopsy, two types of needles (i.e.,
plastic and metal) for aspiration cytology, and "salvage cytology" in diagnosing gastrointestinal cancer.
Salvage cytology involved placement of a specimen trap at the suction port of the umbilical connector
of the endoscope. Suction is applied to the accessory channel of the endoscope after removal of the
biopsy forceps (with a specimen) and during the intervals when tissue sampling is not being performed.
Any cellular material that might otherwise be lost within the accessory channel is salvaged for cytologic
examination. Forceps biopsies produced positive diagnostic results in 78%, metal needle biopsies in
84%, plastic needle biopsies in 77%, and salvage cytology in 91% of cases. Overall accuracy was
increased by combining techniques. In the series of 50 cases of upper gastrointestinal malignancies
reported by Bhasin et al.,27(969) the diagnostic yield of salvage cytology was 96%, which was equal to
that for standard biopsy specimens but greater than that for cytology specimens obtained by brushing.
In two cases of ulcerated and stenotic tumors, the results of endoscopic aspiration cytology were
positive when those of standard biopsy specimens were nondiagnostic.
Zargar et al.28(970) reported a series of 318 lesions sampled by needle aspiration cytology in which 14
submucosal malignancies and 5 benign submucosal lesions were identified. The accuracy in
diagnosing submucosal, infiltrative, and ulceronecrotic malignancies for needle aspiration was
significantly higher than that for biopsies obtained with a forceps or with cytologic specimens obtained
by brushing (92.9% vs. 7.1% and 14.3%, respectively; p < .001). The higher diagnostic yield of needle
aspiration was attributed to its ability to obtain adequate cytologic samples from deeper layers.
Kochhar et al.29(971) reported results of transproctoscopic fine-needle aspiration cytology for 51
consecutive patients, each referred with a presumptive diagnosis of a rectal mass. Fine-needle
aspiration cytology, brush cytology, and biopsy samples were taken during proctoscopic examination.
Of 30 patients with rectal malignancy in whom all three sampling techniques were applied, biopsy
results were positive for 27 (90%), brush cytology for 25 (83.3%), and fine-needle aspiration cytology
for 29 (96.6%). A combination of fine-needle aspiration cytology and brush cytology gave a positive
yield of 96.6%; fine-needle aspiration cytology with biopsy gave a yield of 100%. Fine-needle aspiration
cytology was most helpful in diagnosing infiltrative tumors. All 10 patients with tumor spread to the
pouch of Douglas or the rectovesical pouch and a patient with submucosal rectal carcinoma were
correctly diagnosed by fine-needle aspiration cytology. There were no false-positive results and no
complications with this technique.
Limitations
Fine-needle aspiration cytology has certain limitations. Its role in assessing benign lesions is primarily
that of excluding malignancy, but a negative aspirate result does not necessarily eliminate the
possibility of malignancy. Fine-needle aspiration cytology cannot be used for typing benign lesions.
To improve the diagnostic accuracy for submucosal lesions, Wiersema et al.30(972) combined
endoscopic ultrasonography (EUS) with fine-needle aspiration cytology. A diagnostic result was
achieved for 17 of 20 consecutive patients with lesions that aroused a suspicion of malignancy or that
were not amenable to standard biopsy technique. EUS appeared to be helpful in determining the
feasibility of fine-needle aspiration and the direction of approach for aspiration. EUS allowed definition
of the regional anatomy surrounding the mass and assessment of malignant features.
Guillotine Needle Biopsy

EUS has been used to characterize submucosal tumors.3136(973) Unfortunately, attempts to
differentiate benign from malignant submucosal lesions on the basis of size, shape, and sonographic
appearance have been disappointing.3336(974)
Histologic determination of the tissue of origin or the presence of malignancy is essential in making
appropriate decisions for therapy of submucosal tumors. This determination is less important for
symptomatic, large, or diffuse lesions, because they are already considered to be candidates for
surgery. However, laparotomy and excision of small (>4 cm diameter), benign lesions in patients who
are fit for surgery also seems excessive. A bioptic (i.e., guillotine) needle (Flexi-Temno, Bauer
Company, Pieve di Cento, Bologna, Italy) can be used to obtain a tissue sample adequate for
histologic confirmation from solid submucosal lesions that are demonstrated by EUS.37,38(975)
Technique
The bioptic needle is a unique flexible needle designed for use with endoscopes that have accessory
channels of standard diameters. Its tip is a guillotine device that provides submucosal samples 8 mm
long and 1.1 to 2.1 mm in diameter. The guillotine is controlled by means of a handle similar to that of
a biopsy forceps. The device comes in several lengths for various applications, and it is protected by
an outer Teflon sheath (Figure 232).
(976)Figure 232. The guillotine needle shown closed (top) and open (bottom).
During endoscopy, the guillotine needle is employed in a step-by-step manner. When facing the
submucosal tumor, the Teflon sheath is retracted, the guillotine device is opened, and the needle is
thrust completely into the tumor. At this point, the guillotine device is closed, and the needle is gently
withdrawn.
Results
Guillotine needle biopsy was performed in 21 patients. One to four punctures (mean of three) were
necessary to obtain one adequate sample. Samples were not obtained from two lesions that were later
confirmed as leiomyomas at surgery. One of these, a lesion that was smaller than 2 cm in diameter,
was located on the prepyloric lesser curvature of the stomach; the other, which was larger than 4 cm in
diameter, was on the greater curvature of the body of the stomach.
No major complication was encountered during this study. Fifteen patients had minimal, self-limited
bleeding, comparable to that commonly observed after obtaining a standard forceps biopsy.
The following diagnoses were made by histologic appraisal of the guillotine biopsy specimens: 1
leiomyoblastoma (Figure 233), 2 lipomas (Figures 234 and 235), and 14 leiomyomas. Two
previously unsuspected leiomyosarcomas were diagnosed; by EUS, both lesions were thought to be
benign leiomyomas. One was less than 4 cm in diameter and the other larger than 4 cm. Surgery was
performed and the diagnosis confirmed for 14 patients: 1 lipoma, 1 leiomyoblastoma, 2
leiomyosarcoma, and 10 leiomyomas. EUS and guillotine biopsy diagnoses were in agreement for 1
lipoma and 4 leiomyomas, all in asymptomatic patients and each less than 4 cm in diameter; these
lesions are being followed at 6-month intervals with EUS and guillotine needle biopsy.
(977)Figure 233. Photomicrograph of a guillotine needle biopsy sample of a

leiomyoblastoma. A, From left to right, mucosa and deeper layers. B, Enlargement (100). In
the muscularis propria layer, leiomyoblastoma is demonstrated.
(978)Figure 234. Endoscopic ultrasonography (EUS) (10 MHz) of a gastric lipoma showing
a large solid hyperechoic mass originating from the third layer. The first two layers are
normal. The fourth layer (arrows) is clearly depicted and is normal.
(979)Figure 235. EUS (7.5 MHz) of an esophageal leiomyoma (L). A well-demarcated

hypoechoic mass contiguous with the muscularis propria (arrowheads) is displayed. Aaorta;
Bwater-filled balloon.
Use of the guillotine needle provides submucosal samples in 90% of cases. In two benign myogenic
tumors, the technician failed to obtain adequate samples because of a marked fibrotic component that
gave the mass a hard consistency; the guillotine needle was more easily inserted into all three
malignant tumors.
Guillotine needle biopsy has proved in our experience to be a simple and safe technique. In this
chapter, its use has been described only in gastric submucosal lesions. The technique can be used
safely and easily in other organs provided EUS has demonstrated a solid submucosal tumor without
evidence of vascular structures. We have performed guillotine needle biopsy in the esophagus (eight
leiomyo-mas; Figures 236 and 237), stomach (infiltrative gastric disorders, including three
lymphomas, two linitis plastica lesions, two cases of Menetrier's disease; Figures 238 and 239),
duodenum (one leiomyoblastoma, one ampullary cancer), and rectum (one submucosal cancer
recurrence).
(980)Figure 236. Photomicrograph of the guillotine needle biopsy of the lesion

demonstrated in Figure 234. Histology demonstrated that the lesion was a lipoma (100).
(981)Figure 237. A, Photomicrograph of a guillotine needle biopsy of the esophageal

leiomyoma demonstrated in Figure 236. B, Enlargement (100). Histology demonstrated a
leiomyoma.
(982)Figure 238. EUS (10 MHz) of a gastric lymphoma. In the anterior wall of the gastric
body, a hypoechoic thickening is displayed (large arrows). Thickening is localized and
limited to the second and third sonographic layers. Normal gastric wall is clearly depicted
(small arrows).
(983)Figure 239. Photomicrograph of a guillotine needle biopsy of lesion demonstrated in

Figure 238. Histology confirmed the presence of lymphoma (100).

Although EUS can demonstrate the presence of a solid submucosal tumor, histologic diagnosis by
ultrasonographic findings alone is often problematic and unreliable. In such cases, a guillotine needle
biopsy can provide a definitive histologic diagnosis. Asymptomatic patients with solid submucosal
tumors can be regularly followed with EUS and guillotine needle biopsy and operated on when
malignancy is revealed.
Conclusions
It is often difficult at endoscopy to differentiate submucosal tumors and tumescent lesions caused by
extraluminal compression. Because the false-positive rate of diagnosis is about 27%,13(984) additional
diagnostic procedures, such as EUS for visualization of the deeper layers of the gut wall and its
immediate surroundings, are needed. When EUS clearly shows that an elevation in the wall is caused
by extrinsic compression by a normal anatomic structure, no further tests are required. If EUS
demonstrates extraluminal compression due to a pathologic process, aspiration needle cytology is
indicated. If the presence of a submucosal tumor is established by EUS, guillotine needle biopsy can
be used to exclude malignancy, to determine histologic type of the tumor, and to keep the lesion under
surveillance during periodic follow-up examinations.
REFERENCES
1.
Benvenuti GA, Hattori K, Levin B, et al. Endoscopic sampling for tissue diagnosis in
gastrointestinal malignancy. Gastrointest Endosc 1975;21:15961.
2.
Kobayashi S, Yoshii Y, Kasugai T. Biopsy and cytology in the diagnosis of early gastric
cancer. Endoscopy 1976;8:538.
3.
Smithies A, Lovell D, Hishon S, et al. Value of brush cytology in diagnosis of gastric cancer.
BMJ 1975;4:326.
4. Winawer SJ, Posner G, Lightdale CJ, et al. Endoscopic diagnosis of advanced gastric
carcinoma. Factors influencing yield. Gastroenterology 1975;69:11837.
5. Witzel L, Holter F, Gretillat PA, et al. Evaluation of specific value of endoscopic biopsies and
brush cytology for malignancies of the oesophagus and stomach. Gut 1976;17:3757.
6. Winawer SJ, Leidner SD, Hagdu SI, et al. Colonoscopic biopsy and cytology in the diagnosis
of colon. Cancer 1978;42:284953.
7. Qizilbash A, Castelli M, Kowalski M, et al. Endoscopic brush cytology and biopsy in the
diagnosis of cancer of the upper gastrointestinal tract. Acta Cytol 1980;24:3138.
8. Chambers LA, Clark WE. The endoscopic diagnosis of gastroesophageal malignancy. Acta
Cytol 1986;30:70114.
9. Chen YL. The diagnosis of colorectal cancer with cytologic brushing under direct vision at
fiberoptic colonoscopy. A report of 59 cases. Dis Colon Rectum 1987;30:3424.
10. Jan GM, Dewani K, Koul V, et al. Role of brush cytology in GIT and biliary tract lesions. Indian
J Cancer 1988;25:228.
11. Kobayashi S, Prolla JC, Winans CS, et al. Improved endoscopic diagnosis of
gastroesophageal malignancy: Combined use of direct vision brushing cytology and biopsy.
JAMA 1970;212:20869.
12. Young JA, Hughues HE. Three year trial of endoscopic cytology of the stomach and
duodenum. Gut 1980;21:2416.
13. Hedenbro JL, Ekelund M, Wetterberg P. Endoscopic diagnosis of submucosal gastric lesions.
14. Kaneko E, Kumagai J, Honda N, et al. Evaluation of the new giant-biopsy forceps in the
diagnosis of mucosal and submucosal gastric lesions. Endoscopy 1983;15:3226.
15. Siegel M, Barkin J, Rogers SA, et al. Gastric biopsy: A comparison of biopsy forceps.
16.
Bjork JT. Non-epithelial neoplasms of the stomach. Gastrointest Endosc 1981;23:1623.
17. Iishi H, Yammamoto R, Tatsuta M, et al. Evaluation of fine-needle aspiration biopsy under
direct vision gastrofiberscopy in diagnosis of diffusely infiltrative carcinoma of the stomach.
Cancer 1986;57:13659.
18. Lundgreen R. A flexible thin needle for transbronchial aspiration biopsy through the flexible
fiberoptic bronchoscope. Endoscopy 1980;12:1802.
19. Oho K, Kato H, Ogawa I, et al. A new needle for transfiberoptic bronchoscopic use. Chest
1979;76:492.
20. Nguyen GK. Fine-needle aspiration biopsy cytology of hepatic tumors in adults. Pathol Annu
1986;21:32149.
21. Nguyen CK. Percutaneous fine-needle aspiration cytology of the pancreas. Pathol Annu
1985;20:22138.
22. Dickey JE, Haaga JR, Stellato TA, et al. Evaluation of computed tomography guided
percutaneous biopsy of the pancreas. Surg Gynecol Obstet 1986;163:497503.
23. Ingoldby CH, Mason MK, Hall RI. Endoscopic needle aspiration cytology: A new method for
the diagnosis of upper gastrointestinal cancer. Gut 1987;28:11424.
24. Kochhar R, Gupta SK, Malik AK, et al. Endoscopic fine needle aspiration biopsy. Acta Cytol
1987;31:4814.
25. Kochhar R, Rajwanshi A, Malik AK, et al. Endoscopic fine needle aspiration biopsy of
gastroesophageal malignancies. Gastrointest Endosc 1988;34:3213.
26. Graham DY, Tabibian N, Michaletz PA, et al. Endoscopic needle biopsy: A comparative study
of forceps biopsy, two different types of needles, and salvage cytology in gastrointestinal
cancer. Gastrointest Endosc 1989;35:2079.
27. Bhasin DK, Kochhar R, Rajwanshi A, et al. Endoscopic suction cytology in upper
gastrointestinal tract malignancy. Acta Cytol 1988;32:4524.
28. Zargar SA, Khuroo MS, Mahajan R, et al. Endoscopic fine needle aspiration cytology in the
diagnosis of gastro-oesophageal and colorectal malignancies. Gut 1991;32:7458.
29. Kochhar R, Rajwanshi A, Wig JD, et al. Fine needle aspiration cytology of rectal masses. Gut
1990;31:3346.
30. Wiersema MJ, Hawes RH, Tao LC, et al. Endoscopic ultrasonography as an adjunct to fine
needle aspiration cytology of the upper and lower gastrointestinal tract. Gastrointest Endosc
1992;38:359.
31. Yasuda K, Nakajima M, Yoshida S, et al. The diagnosis of submucosal tumors of the stomach
by endoscopic ultrasonography. Gastrointest Endosc 1989;35:105.
32. Caletti GC, Zani L, Bolondi L, et al. Endoscopic ultrasonography in the diagnosis of gastric
submucosal tumor. Gastrointest Endosc 1989;35:4138.
33. Nakazawa S, Yoshino J, Nakamura T, et al. Endoscopic ultrasonography of gastric myogenic
tumor. A comparative study between histology and ultrasonography. J Ultrasound Med
1989;8:3539.
34. Tio TL, Tytgat GNJ, Den Hartog Jager FCA. Endoscopic ultrasonography for the evaluation of
smooth muscle tumors in the upper gastrointestinal tract: An experience with 42 cases.
35. Yasuda K, Cho E, Nakajima M, et al. Diagnosis of submucosal lesions of the upper
gastrointestinal tract by endoscopic ultrasonography. Gastrointest Endosc 1990;36:S1720.
36.
Aibe T, Takemoto T. Benign lesions of gastrointestinal tract. In Kawai K (ed). Endoscopic
Ultrasonography in Gastroenterology. Tokyo: Igaku-Shoin 1988;4455.
37.
Caletti GC, Gullini S, Brocchi E, et al. Endoscopic ultrasonography (EUS) with the guillotine
needle biopsy in the diagnosis of gastric submucosal tumors (GMST). Gastrointest Endosc
1990;36:222.
38. Caletti GC, Brocchi E, Ferrari A, et al. Guillotine needle biopsy as a supplement to
Chapter 24 Early Colorectal Cancer and Endoscopic Resection

(985)
(986)
SHIN-EI KUDO, M.D., PH.D.
Early colorectal carcinoma, defined as malignancy limited to the mucosa and submucosa, may
manifest in various forms. A macroscopic classification system similar to that for early gastric cancer
has been proposed (Figure 241). This classification is based on the macroscopic morphology of early
colorectal cancers and includes shapes that range from protruded to flat to depressed.14(987)
Various intermediate shapes and combinations are possible in this classification.
(988)Figure 241. Macroscopic classification of early colorectal cancer (heavy black line
indicates cancer). LSTlaterally spreading tumor. (From Kudo S. Endoscopic mucosal
resection of flat and depressed types of early colorectal cancer. Endoscopy 1993; 25:45561.)
Among the several types of early colorectal cancers, the depressed lesions are especially difficult to
recognize with endoscopy. Although often discrete, they are not distinctly elevated lesions and are
therefore difficult to recognize by conventional endoscopic methods.
Tumor progression, manifested by transmural growth, may be rapid for early cancers of the depressed
type. This is supported by the fact that size-matched early colorectal cancers that are classified as
protruded types have a lower submucosal invasion rate than any other type, but the depressed type
has the highest rate (Table 241). In our experience over a 10-year period, 12.4% of 908 cases of
early colorectal cancer were of the depressed type. Therefore, diagnosis of small depressed cancers is
extremely important despite the difficulties that this presents at endoscopy.
Rate of Submucosal Invasion in Colorectal Neoplasm
TABLE 241
SIZE (MM)
lp, lsp, ls
lla, llb
lla+llc
llc, llc+lla
610
1115
1620
21
TOTAL
0/2887
0%
2/2989
0.1%
1/144
0.7%
11/131
25/1884
1.3%
2/399
0.5%
7/48
14.6%
11/30
43/467
9.2%
1/25
4.0%
16/24
66.7%
4/5
26/137
19.0%
0/2
0%
9/10
90.0%
0
25/80
31.3%
0/1
0%
6/12
50.0%
0
119/5455
2.2%
5/3416
0.1%
39/238
16.4%
24/166
8.4%
36.7%
80.0%
15.7%
0/139
5/42
19/77
24/258
LST
0%
11.9%
24.7%
9.3%
14/6151
45/2361
64/660
40/191
50/170
213/9533
Total
0.2%
1.9%
9.7%
20.9%
29.4%
2.2%
Data from Akita Red Cross Hospital, Akita-City, Akita, Japan, 1985 through February 1995.
ppedunculated; spsubpedunculated; ssessile (see Fig. 241); LSTlaterally
spreading tumor.
Macroscopic Lesion Definition

Type IIc Lesions
Type IIc lesions are slightly depressed tumors without an elevated margin (see Figure 241). Most
Type IIc lesions are small submucosal (sm) cancers, and strict attention must be given to the clinical
management of these tumors.
Type IIc+IIa Lesions

Type IIc+IIa lesions are depressed with elevated margins (see Figure 241). The elevated margins are
composed of normal or hyperplastic mucosa with a nonneoplastic surface. Endoscopically, Type IIc+IIa
resembles the IIa+IIc type; however, the former can be discriminated from the latter by dye-spraying
endoscopy (i.e., chromoscopy), magnifying endoscopy, or both methods. Either technique reveals the
nonneoplastic structure of the elevated margin of the Type IIc+IIa lesion. In principle, the actual level of
the central "depressed" region of the IIc+IIa tumor should be equal to or lower than the surface of the
surrounding normal mucosa.
Type IIa+IIc Lesions

Type IIa+IIc lesions may be described as elevated tumors with a central depression (see Figure 241).
Most such lesions are larger than 1 cm in diameter and submuco-sally invasive. Both the elevated and
the depressed portions consist of neoplastic tissue. Endoscopically, Type IIa+IIc lesions resemble
Type IIc+IIa lesions, but they can be discriminated by chromoscopy, magnifying colonoscopy, or the
combination of both methods. The elevated margin of Type IIc+IIa tumors consists of normal or
hyperplastic mucosa.
Type IIa+IIb Lesions

So-called flat adenomas are absolutely flush (Type IIb) or slightly elevated (Type IIa),5(989) the latter
being more frequent. These lesions are not usually malignant until they are quite large (see Table
241).
Patterns of Invasion
The histopathologic patterns of invasion into the submucosal layer by depressed types of cancer are
classified into three types: penetrating, laterally expanding, and superficially spreading (Figure 242).
(990)Figure 242. Infiltrative patterns of flat and depressed types of colorectal cancer (black
areas indicate cancer). (From Kudo S. Endoscopic mucosal resection of flat and depressed
types of early colorectal cancer. Endoscopy 1993; 25:45561.)
Penetrating Type
A penetrating tumor invades the submucosal layer through the perivascular space while the tumor is
relatively small. The cancer can be invasive when the lesion is as small as 5 mm in diameter. The
cancer cells proliferate in the submucosal layer, forming a mass, which results in a slight elevation of
the lesion as a whole.
Laterally Expanding Type

There is preferential lateral extension with the laterally expanding type of cancer (see Figure 242).
The lesion may extend over a horizontal distance of nearly 10 mm before there is invasion of the
submucosal layer. The surrounding normal mucosa is elevated owing to the compression by the tumor,
but the surface of the lifted margin is composed of normal glands or reactive hyperplastic tissue.
Superficially Spreading Type

The superficially spreading type develops by extension onto the mucosal surface layer (see Figure
242). Because this type is seldom accompanied by any color changes that may be visible during
endoscopy, it is especially difficult to detect. Even if the dye-spraying technique is used, the borders
are not always clearly demarcated. When the cancer invades the submucosal layer, a nodular
protrusion is formed within the lesion.
Detection
The detection of the depressed types of early colorectal cancer during routine colonoscopy requires
knowledge of the appearance of the various types and an awareness that these lesions have no
elevation above the mucosal surface. The depressed types of cancer are easily overlooked at
colonoscopy. Clues that aid in their endoscopic detection include the following: color changes, bleeding
spots, friability, mural deformity, and disappearance of the vascular network pattern.
Color Changes
Redness
The depressed types of early colorectal carcinoma are most frequently identified by a change to a light
red color. The typical histologic characteristic that accounts for this appearance is hyperplasia of
capillaries in the mucosal lamina propria. Deeper reddening is usually associated with nonneoplastic
lesions, including benign erosions, inflammatory changes, and angiodysplasia.
Discoloration or Paleness
Small and flat, pale lesions frequently found in the large bowel, especially the rectum, are usually
hyperplastic polyps. Rarely, early carcinoma may have a whitish color. The discoloration may be
attributable to decreased interstitial constituents and increased glandular density. Slightly reddish
lesions may look pale when melanosis coli is present.
Bleeding Spots
Spontaneous bleeding from a lesion may suggest that the lesion is malignant. The colonoscopist
should be alert for the presence of hemorrhagic spots that may indicate an early form of colon cancer.
Friability
When inspecting red or pale tissues, bleeding may occur during observation, particularly during
deformation of the wall of the colon induced by the insufflation of air. This manifestation strongly
suggests malignancy provided trauma or other external factors are clearly not responsible for the
bleeding.
Mural Deformity
On endoscopic observation, slightly uneven mucosa (compared with the smooth surface of
surrounding normal mucosa) or mural deformity may indicate the presence of cancerous lesion. These
changes may be observed with relative ease when the mucosal surface is seen in profile. By using the
air insufflation and suction controls of the endoscope, it is usually possible to bring the mucosal surface
into this viewing plane, which is the most ideal approach for detection of surface irregularities. After
such a lesion is detected, the degree of mucosal deformity can also be assessed by varying the
volume of air within the colon.
Disappearance of Vascular Network Pattern

Disappearance of the distinct vascular network pattern of the colonic mucosa often suggests the
existence of a slightly elevated or depressed lesion. This may be observed by contrasting the
appearance of such a lesion with that of the surrounding mucosa. However, this finding occurs
frequently within normal mucosa, especially of the transverse colon, and close attention to detail is
needed to discriminate the cancer-derived pattern from the normal mucosal pattern.
Qualitative Diagnosis: Differential between Early Carcinoma and Benign

Adenoma
At endoscopy, depressed types of early colon cancers appear chiefly as light red or pale areas with
irregular margins. After detecting a lesion, further observation is needed to determine whether it is
benign or malignant. In addition to simple observation, two other endoscopic maneuvers are required
for a qualitative colonoscopic diagnosis: air-induced deformation and study of the innominate grooves.
Air-Induced Deformation
Varying the volume and pressure of air within the colon deforms small lesions such as early-stage
depressed cancers. After a flat or depressed type of tumor is detected, the first step in the diagnosis is
to observe morphologic changes in the tumor itself and the surrounding mucosa using this technique.
Decreased air volume elevates the normal and hyperplastic mucosa surrounding the tumor and brings
the tumor depression into relative prominence (Figure 243). Increased air volume makes the
depression less prominent, and the tumor appears to be flatter than it actually is. With a benign lesion,
both the depression and the surrounding rim are elevated when suction is applied, and no significant
change in shape is observable.
(991)Figure 243. Transformation of the shape of early colorectal cancer by varying the
levels of air insufflation during colonoscopy. With maximum wall distention, the lesion
becomes flat; with less insufflation, the elevated margins and central excavation become more
evident. (From Kudo S. Endoscopic mucosal resection of flat and depressed types of early
colorectal cancer. Endoscopy 1993; 25:45561.)
Interruption of the Innominate Grooves

Innumerable fine grooves, called the innominate grooves, demarcate the normal colonic mucosa into
areas that contain the crypts of Lieberkhn. The innominate grooves are arranged circumferentially
with reference to the axis of the lumen. They are clearly observed when a dye such as indigo carmine
is sprayed over the mucosal surface.
The innominate grooves disappear in neoplastic lesions, but they remain visible in nonneoplastic
lesions and in normal colonic mucosa (Figure 244). I have found this observation to be highly useful
for differentiating depressed-type carcinoma from nonneoplastic changes. Disappearance of the
innominate grooves is characteristic for depressed-type tumors. However, in elevated lesions, whether
nonneoplastic or neoplastic, innominate grooves are absent. The presence or absence of the pattern
of innominate grooves is more clearly defined by magnifying endoscopy (see Chapter 13: Magnification
Endoscopy).
(992)Figure 244. Disappearance of innominate grooves after dye-spraying. With neoplastic

lesions such as cancers and adenomas, the innominate grooves disappear. These grooves
remain on nonneoplastic lesions and in normal mucosa. (From Kudo S. Endoscopic mucosal
resection of flat and depressed types of early colorectal cancer. Endoscopy 1993; 25:45561.)
Quantitative Diagnosis: Depth of the Lesion

Qualitative endoscopic diagnosis establishes the presence of early cancer, but quantitative diagnosis
assesses involvement of the submucosa. Submucosal extension of early colorectal cancer is classified
histologically according to the vertical depth of invasion as sm1 (infiltration confined to the upper third),
sm2 (invasion to middle third), and sm3 (infiltration to lower third). Sm1 lesions are further divided into
subgroups a, b, and c according to the degree of horizontal extension (Figure 245). Most early-stage,
depressed-type colorectal cancers are small but invade the submucosa. It is therefore important to
identify submucosal invasion at endoscopy. The essential endoscopic findings include fold
convergence, air-induced deformation, absence of structure within the depressed portion of the lesion,
and pit pattern.
(993)Figure 245. Classification of submucosal invasion (black areas indicate cancer). (From
Kudo S. Endoscopic mucosal resection of flat and depressed types of early colorectal cancer.
Endoscopy 1993; 25:45561.)
Fold Convergence Pattern

The possibility of submucosal infiltration must be considered when mucosal folds converge toward a
depressed type lesion. In mucosal cancers of Type IIc, fold convergence is rare. However, reduction of
the volume of air gives rise to apparent fold convergence even in mucosal cancers, although this
appearance is not static or fixed and disappears with increased air pressure. When a Type IIc lesion
exhibits fixed fold convergence, there is always extensive submucosal invasion (i.e., sm3
classification), or the lesion is an advanced cancer.
Air-Induced Deformation
Air-induced deformation is a typical feature of flat and of depressed types of tumors. The depression
becomes more pronounced as the air pressure within the colon is decreased. The degree to which this
occurs may be slight or more severe. If the depression becomes especially prominent (i.e., severe),
submucosal cancer may be suspected; the lesion may be rather superficial (i.e., sm1a to sm1c
classification). Severe air-induced deformation is thought to result from fixation to the muscularis
mucosa. When there is massive submucosal or deeper invasion (i.e., sm3 cancer), the lesion is fixed
as a mass, and its depressed portion does not become more prominent with a reduction in air
pressure. Accordingly, fixed depressed lesions are diagnosed as having massive involvement of the
submucosa or deeper layers.
Nonstructural Base of the Depressed Portion

If standard endoscopic observation demonstrates the base of a depressed lesion to be monotonous
and without structure, involvement of the submucosa is almost always present, and the lesion may be
diagnosed as a submucosal cancer. The depressed bases of these lesions are infrequently associated
with cancerous ulcerations. Although appearing to be without structure, the base usually consists
histologically of a dense glandular assembly.
Pit Pattern
The innominate grooves circumscribe colonic mucosal areas that contain between 40 and 60 pits (i.e.,
glands or crypts of Lieberkhn). These are round or slightly oval, have a diameter of around 70 m,
and are arranged in a regular pattern. The colonic pits can only be observed with a magnifying
endoscope or by stereo-microscopy (see Chapter 13: Magnification Endoscopy). The normal pit
pattern is altered by certain disease states (Figure 246). Among the several kinds of pit patterns,
small pits (Type IIIs) or very irregular, nonstructural pits (Type V) are characteristic of depressed-type
early co-lorectal cancer. If a Type V pit pattern is found within the depressed portion of a lesion,
invasion of the submucosal is likely (i.e., an sm cancer).
(994)Figure 246. Classification of the pit pattern of the colorectum. (From Kudo S.
Endoscopic mucosal resection of flat and depressed types of early colorectal cancer.
Endoscopy 1993; 25:45561.)
Stereomicroscopy and Pit Pattern

The minute surface structure of resected specimens (e.g., those obtained by strip biopsy) should be
examined after staining using a stereomicroscope with a magnification of 64. The colonic pits, which
are the orifices of crypts, are the structural features seen on stereomicroscopy. Based on the study of
these specimens, five pit patterns are possible for colorectal mucosa under normal and pathologic
conditions (see Figure 246).
Tubular pits (Type III) and branched pits (Type IV) are characteristic of neoplastic tumors; large tubular
(Type IIIL) and branched pits (Type IV) are characteristic of protruded or flat adenomas; and small
tubular pits (Type IIIs) are typical of depressed-type lesions. A depressed-type early carcinoma may be
associated with an elevated rim (Type IIc+IIa), but the pits of the elevated rim are usually normal (Type
I) or star-shaped, hyperplastic pits (Type II), and not one of the neoplastic types. Therefore, the
elevation of the rim is presumably related to compression of normal tissue by the carcinoma.
Magnifying Endoscopy
It is possible to observe colonic pits and pit patterns in vivo by means of magnifying endoscopes with a
magnification of 100 (see Chapter 13: Magnification Endoscopy). Certain of these instruments make
it possible to "zoom in" on lesions in stepwise fashion during conventional colonoscopy and magnify
the images up to 100 times.
Observation of pit patterns by magnifying endoscopy in conjunction with dye-spraying is a reliable
method for differentiation of tumors from nonneoplastic lesions. Because adherent mucus is present in
the depressed portion of depressed lesions, thorough washing is required. After washing, the lesion is
stained with cresyl violet.
Examples of the depressed type of early colorectal cancer are shown in Figures 247, 248 and 249.
(995)Figure 247. A, Endoscopic image of a 4-mm diameter lesion in the sigmoid colon with
superficial depression and redness (Type IIc+IIa). B, Lesion after dye-spraying. The central
depression and slight marginal elevation are more clearly defined. C, Histologic cross section
showing a depressed adenocarcinoma with peripheral elevation and massive invasion of the
submucosa. (A-C, From Kudo S. Endoscopic mucosal resection of flat and depressed types of
early colorectal cancer. Endoscopy 1993; 25:45561.)
(996)Figure 248. A, Endoscopic image of a superficial depressed lesion with redness and an
irregular margin in the sigmoid colon (Type IIc+IIa) after dye-spraying. B, Stereomicroscopic
image. Lesion measures 5 4 mm with a Type IIIs pit pattern in the depressed area that is
strongly suggestive of cancer. C, Histologic cross section, low-power view, showing
adenocarcinoma invading the superficial submucosa. D, Histologic cross section, high-power
view, showing the tumor invading the superficial submucosa with an increase in collagen.
(A-D, From Kudo S. Endoscopic mucosal resection of flat and depressed types of early
colorectal cancer. Endoscopy 1993; 25:45561.)
(997)Figure 249. A, Endoscopic image of a superficial depressed lesion with redness in the
sigmoid colon (Type IIc). B, After dye-spraying, the central depression and margin are clearly
shown. C, Histologic cross section, low-power view, showing the center of the depression with
a low-grade malignancy that is limited to the mucosa. There is an increase of glands and
marked atypical structure. (A-C, From Kudo S. Endoscopic mucosal resection of flat and
depressed types of early colorectal cancer. Endoscopy 1993; 25:45561.)
Treatment
Treatment of depressed types of early colorectal cancer is achieved by endoscopic strip biopsy or
surgery. Strip biopsy is the first step in treatment of small cancers. Additional treatment depends on the
results of histopathologic analysis, with particular attention to findings such as the presence of tumor
remnants at resection margins and depth of invasion.
The technique of endoscopic mucosal resection has been used in the diagnosis and treatment of a
wide variety of early-stage cancerous lesions of the digestive tract, including early gastric
cancer,58(998) gastric lymphoma,9(999) superficial esophageal carcinoma,1012(1000) duodenal
carcinoma,13(1001) and carcinoid tumor of the rectum.14(1002) Various modifications of the
technique have been described.15,16(1003) In contrast to laser photoablation, endoscopic resection
provides a tissue specimen for accurate histologic diagnosis. However, a combination of endoscoptic
resection and laser therapy has been found to be potentially useful in the diagnosis and treatment of
superficial gastric cancer, depending on technical problems related to the location of the
lesion.7,8(1004) High-frequency ultrasonographic "miniprobes" are also thought to be useful in
assessing whether small gastric tumors are suitable for strip biopsy.6,17(1005)
Endoscopic Resection Technique

The technique of endoscopic resection is shown in Figure 2410. A flat or depressed lesion is elevated
above the mucosal surface by injecting saline beneath the lesion, which is then removed in a manner
that resembles standard polypectomy, albeit with some important differences.
(1006)Figure 2410. Endoscopic resection technique (heavy black line indicates cancer).
(From Kudo S. Endoscopic mucosal resection of flat and depressed types of early colorectal
cancer. Endoscopy 1993; 25:45561.)
Optimum endoscopic visualization of the target lesion should be achieved before undertaking
treatment. Submucosal injection underneath the cancer should be performed in such a way that the
lesion itself is exactly at the top of the artificial swelling that is created. The elevated portion of the
mucosa is then captured with a snare and resected using only cutting current. The resection specimen
is carefully grasped with a grasping forceps, and the instrument is then withdrawn. Damage to the
specimen must be avoided, and it is therefore advisable not to attempt to recover the specimen by
pulling it through the accessory channel of the endoscope.
Lesions with 3-cm or smaller diameters can be removed by the endoscopic resection technique. Most
depressed (i.e., Types IIc and IIc+IIa) and Type IIa+IIc lesions 10-mm or larger in diameter are
invasive cancers. The endoscopic resection technique is indicated primarily for minute Types IIc and
IIc+IIa tumors. It can be considered as definitive treatment for intramucosal cancer. However, like the
penetrating type of cancer, the depressed type may invade the submucosa even if it is minute (;5 mm
in diameter). Infiltration should be assessed and classified as described previously.
Endoscopic Resection and Submucosal Invasion

In my experience, about 10% of submucosally invasive early cancers metastasized to lymph nodes.
Malignant lesions with only slight invasion of the submucosa (sm1a to sm1b) without vessel infiltration
showed no nodal metastasis, so that they may be treated by endoscopic resection alone. However,
when the degree of submucosal invasion is more extensive (sm1c) or deeper (sm2, 3), the risk of
residual carcinoma and the possibility of lymph node metastasis increase. These lesions should
therefore be treated surgically.
Significance of Depressed-Type Early Colorectal Carcinomas

The importance of minute, flat or depressed adenomas of the colon and rectum is
controversial.18(1007) Riddell18(1008) credits Muto et al.19(1009) with the first description of the flat
adenoma. These authors described 33 of these lesions in 1985 and 128 such adenomas in
1991.20(1010) A number of investigators have subsequently called attention to the discovery of small,
early-stage, depressed or flat cancers2124(1011) or to flat adenomas with severe dysplasia.25(1012)
Although most reports are from Japan, these lesions have been recognized elsewhere in the
world.26,27(1013) However, minute and depressed colorectal adenomas and small depressed
colorectal carcinomas have not been observed in large prospective colonoscopic studies of Western
patients with adenomatous polyps.28,29(1014)
Iishi et al.21(1015) discovered 21 superficial, early-stage cancers over a period of about 3.5 years.
These accounted for 8% of a group of 256 early-stage (Dukes A and B1) cancers diagnosed over the
same period. Most superficial, early-stage cancers were well differentiated, and five (24%) invaded the
submucosa. Most (86%) of the lesions were not associated with adenomas. In a prospective study,
Karita et al.25(1016) found 31 flat adenomas in 29 patients. Histologically, all lesions were flat
adenomas confined to the mucosa and had a maximum diameter of 2 cm. Thirteen lesions exhibited
severe dysplasia; they were depressed lesions with elevation of the surrounding mucosa or flat lesions
with no changes evident in the surrounding mucosa.
Matsumoto et al.30(1017) described the macroscopic and histologic appearance of 34 lesions with
diameters less than 5 mm that were detected by colonoscopy. Eight lesions were flat, and 26 were flat
but slightly elevated at colonoscopy. Macroscopic study revealed the center of the seemingly flat
lesions to be elevated slightly; the center of the apparently elevated flat ones, however, had a central
depression. In these latter lesions, the entire thickness of the mucosa was replaced by adenomatous
glands.
There is some histologic evidence that flat adenomas and depressed adenomas differ histologically
and morphologically from their polypoid and elevated counterparts. Moreover, high-grade dysplasia is
more common in the adenomas. Katakura et al.31(1018) compared 25 superficial, depressed tubular
adenomas with diameters of less than 5 mm with 71 elevated lesions of approximately the same
diameter, focusing particularly on nuclear stratification, loss of nuclear polarity, glandular crowding, and
nuclear pleomorphism. Glandular crowding and nuclear pleomorphism were significantly more
common in the depressed lesions.
Using immunohistochemical staining, Yao et al.32(1019) compared 28 depressed adenomas, 40 flat
adenomas without central depressions, and 29 polypoid adenomas. Pericryptal fibroblasts were most
evident in polypoid adenomas and least evident in depressed lesions. Pericryptal fibroblasts were
poorly developed in the depressed adenomas. These investigators thought that the depressed growth
of certain adenomas might be explained in part by the depletion of fibroblasts observed in these
lesions.
The existence of small, flat or depressed cancers and adenomas necessitates certain modifications in
the well-established theory of the pathogenesis of colorectal cancer. In particular, the dictum that
adenoma size correlates with cancer risk is called into question. The pathogenesis of colorectal cancer
has been explained by the concept of the polyp-to-cancer sequence, and ample evidence
demonstrates that this explanation is reasonably correct. The current approach to the prevention of
colorectal cancer is based on the polyp-to-cancer hypothesis. The existence of small, depressed
cancers raises the possibility that alternative mechanisms for the pathogenesis of colorectal cancer
may exist. It also is not difficult to speculate that colorectal cancer may arise de novo without the
intervening benign developmental phase known as the adenoma. These possibilities have significant
implications for colonoscopy and for the early diagnosis and prevention of colorectal cancer by
colonoscopic means.
Aside from the arguments surrounding the endoscopic recognition and treatment of minute adenomas
and carcinomas, there are questions concerning the histologic evaluation of these lesions. For
example, Uno et al.33(1020) found a significant discrepancy between the histologic interpretations of
seven flat or depressed lesions by two experienced pathologists. One pathologist diagnosed one of the
seven lesions as cancer, and the other pathologist found five of the lesions to be cancers. The
explanation for such differences in interpretation, whether a matter of subjective opinion or a lack of
objective criteria, is uncertain.
Despite problems of histologic interpretation, the issue of the minute, flat or depressed adenomas and
cancers cannot be dismissed as merely a problem of histologic interpretation. There is too much
evidence that such lesions do exist and that they can be systematically studied with new endoscopic
methods.3437(1021) How common they are remains to be seen, and their frequency may vary
considerably in different populations. Increased recognition can be attributed to an increased
awareness on the part of some endoscopists, and as the risk of cancer from the typical adenoma is
reduced by colonoscopic polypectomy, it is possible that other mechanisms for the development of
colorectal cancer will become more evident. Alternatively, a change in the biologic behavior of
colorectal cancer and polyps may be occurring. There are many questions concerning minute, flat or
depressed adenomas and carcinomas that remain to be answered.
REFERENCES
1.
Kudo S. Endoscopic diagnosis and treatment of flat and depressed type of early colorectal
cancer. Stomach and Intestine 1989;24:31729.
2.
Kudo S. Diagnosis of minute colorectal cancer. Stomach Intestine 1990;25:80112.
3.
Kudo S. Histological diagnosis of flat and depressed types of early colorectal cancerStrip
biopsy and dealing of material. Gastroenterol Endosc 1989;31:28456.
4. Kudo S. Endoscopic mucosal resection of flat and depressed types of early colorectal cancer.
Endoscopy 1993;25:45561.
5. Mizumoto S, Misumi A, Harada K, et al. Evaluation of endoscopic mucosal resection (EMR)
as a curative therapy against early gastric cancer. Nippon Geka Gakkai Zasshi
1992;93:10714.
6. Ohashi S, Okamura S, Mitake M, et al. The utility of endoscopic ultrasonography (EUS) in
endoscopic mucosal resection of early gastric cancer. Nippon Shokakibyo Gakkai Zasshi
1993;90:286672.
7. Watanabe M, Sugino Y, Imai Y, et al. Reevaluation of endoscopic laser therapy for treatment
of early cancer in the stomach. Keio J Med 1993;42:2068.
8. Yasuda K, Mizuma Y, Nakajima M, Kawai K. Endoscopic laser treatment for early gastric
9. Suekane H, Iida M, Kuwano Y, et al. Diagnosis of primary early gastric lymphoma. Usefulness
of endoscopic mucosal resection for histologic evaluation. Cancer 1993;71:120713.
10. Endo M, Yoshino K, Takeshita K, et al. Diagnosis and surgical treatment of early esophageal
cancer. Kyobu Geka 1989;42(8 suppl):6904.

11. Makuuchi H, Machimura T, Mizutani K, et al. Controversy in the treatment of superficial
esophageal carcinomaIndications and problems of the procedures. Nippon Geka Gakkai
Zasshi 1992;93:105962.
12. Endo M. Endoscopic resection as local treatment of mucosal cancer of the esophagus.
Endoscopy 1993;25:6724.
13. Obata S, Suenaga M, Araki K, et al. Use of strip biopsy in a case of early duodenal cancer.
Endoscopy 1992;24:2324.
14. Fujimura Y, Mizuno M, Takeda M, et al. A carcinoid tumor of the rectum removed by strip
biopsy. Endoscopy 1993;25:42830.
15. Takechi K, Mihara M, Saito Y, et al. A modified technique for endoscopic mucosal resection of
small early gastric carcinomas. Endoscopy 1992;24:2157.
16. Karita M, Tada M, Okita K. The successive strip biopsy partial resection technique for large
early gastric and colon cancers. Gastrointest Endosc 1992;38:1748.
17. Takemoto T, Yanai H, Tada M, et al. Application of ultrasonic probes prior to endoscopic
resection of early gastric cancer. Endoscopy 1992;24(suppl 1):32933.
18. Riddell RH. Flat adenomas and carcinomas: Seeking the invisible?. Gastrointest Endosc
1992;38:7213.
19. Muto T, Kamiya J, Sawada T, et al. Small "flat adenoma" of the large bowel with special
reference to its clinicopathological features. Dis Colon Rectum 1985;28:84751.
20. Adachi M, Muto T, Okinaga K, Morioka Y. Clinicopathological features of the flat adenoma.
Dis Colon Rectum 1991;34:9816.
21. Iishi H, Kitamura S, Nakaizumi A, et al. Clinicopathological features and endoscopic diagnosis
of superficial early adenocarcinomas of the large intestine. Dig Dis Sci 1993;38:13337.
22. Kuramoto S, Oohara T. Flat early cancers of the large intestine. Cancer 1989;64:9505.
23. Karita M, Tada M, Okita K, Kodama T. Endoscopic therapy for early colon cancer: The strip
biopsy resection technique. Gastrointest Endosc 1991;37:12832.
24. Mion F, Desseigne F, Napoleon B, et al. Failure of endoscopic detection of a de novo
carcinoma of the colon in a patient with adenomatous polyps. Gastrointest Endosc
1992;38:7036.
25. Karita M, Cantero D, Okita K. Endoscopic diagnosis and resection treatment for flat adenoma
with severe dysplasia. Am J Gastroenterol 1993;88:14213.
26. Hunt DR, Cherian M. Endoscopic diagnosis of small flat carcinoma of the colon. Report of
three cases. Dis Colon Rectum 1990;33:1437.
27. Wolber RA, Owen DA. Flat adenomas of the colon. Hum Pathol 1991;22:704.
28. Winawer SJ, Zauber AG, Ho MN, et al. for the National Polyp Study Workgroup. Prevention of
colorectal cancer by colonoscopic polypectomy. N Engl J Med 1993;329:197781.
29. Winawer SJ, Zauber AG, O'Brien MJ, et al. for the National Polyp Study Workgroup.
Randomized comparison of surveillance intervals after colonoscopic removal of newly
diagnosed adenomatous polyps. N Engl J Med 1993;328:9016.
30. Matsumoto T, Iida M, Kuwano Y, et al. Minute non-polypoid adenoma of the colon detected by
colonoscopy: Correlation between endoscopic and histologic findings. Gastrointest Endosc
1992;38:64550.
31.
Katakura S, Satake Y, Aksoz K, et al. Endoscopic and histopathological study about 25 cases
of minute superficial depressed neoplastic lesions in the large intestine. Dig Endosc
1993;5:312.
32. Yao T, Tada S, Tsuneyoshi M. Colorectal counterpart of gastric depressed adenoma. A
comparison with flat and polypoid adenomas with special reference to the development of
pericryptal fibroblasts. Am J Surg Pathol 1994;18:55968.
33. Uno Y, Munakata A, Tanaka M. The discrepancy of histologic diagnosis between flat early
colon cancers and flat adenomas. Gastrointest Endosc 1994;40:16.
34. Kudo S, Tamure S, Nakajima T, et al. Depressed type of colorectal cancer. Endoscopy
1995;27:547.
35. Kudo S, Hirota S, Nakajima T, et al. Colorectal tumours and pit pattern. J Clin Pathol
1994;47:8805.
36. Kudo S, Tamura S, Nakajima T, et al. Depressed type of colorectal cancer. Endoscopy
1995;27:547.
37.
Kudo S. Early colorectal cancer. New York, Igaku-Shoin. 1996.
Chapter 25 Endoscopic Blood Flow Analysis

(1022)
NOBUHIRO SATO, M.D., PH.D.
SHINGO TSUJI, M.D., PH.D.
SUNAO KAWANO, M.D., PH.D.
This chapter focuses on the measurement of gastrointestinal (GI) hemodynamics under endoscopy, for
which there are three principal techniques: the hydrogen gas clearance method, laser Doppler
velocimetry, and reflectance spectrophotometry. Recent developments in electronic endoscopy and
image processing have also made it possible to assess blood hemoglobin distribution in GI mucosa
directly from an endoscopic image.
Microcirculation in GI tissue has been the subject of physiologic studies in recent decades.
Improvements in techniques for measuring mucosal blood flow have significantly increased our
understanding of the physiopathology of the GI circulation. Various techniques have permitted
assessment of microcirculation, and endoscopy is one of the best of these. The increase in number of
endoscopic studies of human GI hemodynamics is therefore not surprising. This chapter presents the
current status of these studies in two ways: a description of endoscopic methods for assessment of
tissue hemodynamics including basic theory and a summary of the results of studies of the human GI
circulation.
Methods for Endoscopic Assessment of Tissue Circulation

Hydrogen Gas Clearance
Conventional Method
The hydrogen gas clearance method has been used successfully by a number of
investigators.15(1023) The results from this work have been accepted as a standard for the
estimation of tissue blood flow and have been used to determine electromagnetic flowmetry in
physiologic experiments.
This method is based on the dissociation of molecular hydrogen at a platinum surface into hydrogen
ions and electrons. A platinum electrode positioned on the mucosal tissue and a reference calomel,
mercurous chloride, or silver-silver chloride electrode placed on the skin are connected to a
polarographic unit and chart recorder. As hydrogen gas is inhaled, it accumulates in the tissue and
causes current in the circuitry. This current is the result of hydrogen dissociation on platinum (H2
2H+ + 2e); reportedly, this reaction is independent of the hydrogen ion concentration or pH in the
vicinity of the electrode.1(1024) On discontinuation of the external supply of hydrogen gas, the current
gradually drops, depending on the decrease of hydrogen gas concentration in the tissue. Theoretically,
the change in hydrogen gas concentration can be expressed as:
-dC(t)/dt = C(t)Q,
where C(t) is the concentration of hydrogen gas and Q is blood flow. This equation means that the
change in concentration of a marker substance in the tissue depends on the concentration and tissue
blood flow during a unit of time. The equation is expressed as:
C(t) = C0e-Qt,
where C0 is an initial concentration of the marker. Therefore,
Q = ln2/T
= 0.693/T (ml/min/g tissue)
= 69.3/T (ml/min/g tissue)
where T is a time (in minutes) at which concentration of the marker decreases to one half of the
initial concentration. A platinum electrode transforms the hydrogen gas concentration into the current.
Tissue blood flow (Q) can therefore be calculated from the exponential change in the current (Figure
251). The blood flow calculation correlates well with results obtained by the microsphere method,
venous outflow, and electromagnetic flowmetry.57(1025)
(1026)Figure 251. Change in current measured by the hydrogen gas clearance method. This
is a schematic presentation of the current between a platinum electrode inserted in the gastric
mucosa and a reference silver-silver chloride (Ag-AgCl) electrode placed on the skin. In order
to determine the T1/2 the current is plotted on a semilog scale as shown in the window of the
figure.
The hydrogen gas clearance method has several variations in its instrumentation. Conventional
hydrogen gas electrodes are inserted into the tissue, which is therefore invasive. It is also difficult to
know into which layer of the GI wall the platinum face of the needle electrode is inserted. Several
investigators have developed a contact-type electrode for this reason.2,3,8(1027) However, it is difficult
to hold a contact probe in the same location with the same pressure. This is problematic because
mechanical compression of the tissue perturbs the tissue microcirculation. Endoscopists have applied
soft rubber suction cups to the tip of an endoscope or a flow probe to control movement.912(1028)
However, it is also difficult to apply a large suction cup at endoscopy. Consequently, a group of
endoscopists developed a probe that has two platinum electrodes for simultaneous measurement of
gastric mucosal and submucosal blood flow.13(1029)
Inhalation of explosive hydrogen gas is not physiologic for humans. Some investigators have
emphasized the usefulness of a nonexplosive mixture of air and a low concentration of hydrogen
gas.11(1030) Others have utilized hydrogen gas generated in the tissue by electrolysis. This
electrolytic method is more convenient than the conventional hydrogen gas clearance method, but
endoscopic application presents theoretical and practical difficulties.
Electrolytic Hydrogen Gas Clearance Method
In this method, electrolysis of H3O+ generates hydrogen gas locally within the tissue. This hydrogen
gas dissolves in tissue and produces an electric potential at a platinum electrode that serves to
estimate tissue blood flow.1416(1031) In this electrolytic hydrogen gas clearance method, however,
diffusion of the generated gas cannot be ignored. To discuss the diffusion problem in greater detail, it
is necessary to use a generalized equation rather than equation -dC(t)/dt = C(t)Q, given previously.
Fortunately, this diffusion effect can be practically estimated by measuring a T of hydrogen current
with zero blood flow, that is, after the animal is sacrificed. Blood flow (Q) may be estimated by the
equation:
Q = (69.3/T) - (69.3/T0 )
where T0 is the T in the dead animal. In human studies, T0 is assumed to be the same value as
that of a surgically resected specimen or a dead animal.

Disadvantages
The hydrogen gas clearance method does have a few disadvantages, as do other techniques for
estimating tissue circulation. Theoretically, it estimates blood flow in a homogeneous tissue from a
washout curve of the gas within the tissue. GI mucosa, however, consists of several layers. Blood flow
differs in each of these layers, and the hydrogen gas may diffuse from one layer to another. For the
conventional as well as the electrolytic methods for estimation of GI mucosal blood flow, gas diffusion
between layers cannot be eliminated and represents inherent uncertainty in the calculations.
Laser Doppler Velocimetry Method

This method employs light from low-power sources such as a diode laser or a helium-neon (He-Ne)
laser.1722(1032) A block diagram of a laser Doppler velocimeter is shown in Figure 252. Laser light
at a frequency of 0 is conducted through a fiberoptic bundle and illuminates the tissue. After repeated
reflections by nonmoving components in the tissue, the light strikes blood vessels. Moving particles in
the flowing blood, such as erythrocytes, cause Doppler shifts in the wavelength of the light. These
frequencies of the shifted wavelengths of light are:
1, 2, . . . , x.
(1033)Figure 252. Block diagram of a laser Doppler velocimeter. Low-power laser light
(helium-neon [He-Ne] diode laser, etc.) is conducted by a fiberoptic bundle and illuminates
the tissue. The laser light is partially modulated by a Doppler effect owing to moving particles
such as erythrocytes within tissue capillaries. Both modulated and unmodulated light are
collected by additional fiberoptic bundles and produce an electric signal in a photodiode. This
signal contains a multifrequency beat that is essentially due to the difference in frequencies
between the Doppler-shifted light and the unmodulated light. A signal processor analyzes the
frequencies and intensities of the beat, then estimates mucosal blood flow velocity, blood
flow, or mucosal blood volume.
Doppler shifting of wavelengths depends on the velocities and directions of the moving particles as well
as the directions from which photons strike the particles. However, in the laser Doppler method for
assessment of blood flow, photons strike these moving particles from all directions at random. As a
result, the directional factors for particle movement (blood flow) and photons will be canceled. The
shifted and deflected light returns to the mucosal surface after additional scattering and is collected by
a fiberoptic bundle connected to the apparatus. Light reflected by nonmoving tissue components is not
Doppler-shifted, but it also returns to the mucosal surface for collection by the fiberoptic bundle. This
reflected, nonshifted light serves as a reference source whose frequency is 0. When these two types
of light interfere with each other, the shifted wavelengths produce light whose intensity is modulated at
the frequencies of:
0 1, 0 2, . . . , 0 x
A photodetector changes these wavelengths to a current modulated by the sum of the frequencies of:
0 1, 0 2, . . . , 0 x
A Fourier transformation analysis of the current estimates distribution of these Doppler shifts. Data
obtained reflect flow velocity, flow, or blood volume of the tissue, depending on the algorithm used.
One may suspect that angulation of the probe against the tissue would be a critical factor in this
technique because laser light radiated at certain angles might not return to the optical probe and may
not produce Doppler shifts. However, this angulation effect can be canceled if laser light is reflected by
the nonmoving component a number of times. Ahn et al.23(1034) reported that angulation within 60 to
120 degrees does not affect the flowmeter signal, so that angulation of the probe is not such a critical
factor in laser Doppler velocimetry.
One advantage of the laser Doppler method is that it is minimally invasive with respect to tissue; the
light source is low power, and the light can be conducted without difficulty through a thin fiberoptic
bundle. Furthermore, the measuring time is much shorter than that for the hydrogen gas clearance
method. However, the accuracy of the laser Doppler method is readily influenced by the movement of
the subject or the probe. Motion artifact is therefore a major drawback.
Another point to be considered is spatial resolution or the depth sensitivity of each laser Doppler
velocimeter. A phantom study demonstrated that a laser Doppler velocimeter was most sensitive in an
area 0.6 mm distant from the tip of the probe.24(1035) However, this spatial resolution changes
according to the design of the probe, the wavelength and intensity of the light, and the optical
characteristics of the subject. Shepherd et al.17,19(1036) reported that the electric signal obtained with
a laser Doppler velocimeter reflected mucosal blood flow in the canine intestine. On the other hand,
Ahn et al.18(1037) reported that data obtained with a He-Ne laser Doppler flowmeter reflected the
blood flow within an entire layer of cat intestinal wall. Ahn et al. also reported that gastric blood flow
estimates using the two different systems are different.23(1038) The endoscopist who wishes to use a
laser Doppler velocimeter should therefore consult the manufacturer of the instrument concerning the
issue of spatial resolution.
Organ Reflectance Spectrophotometry
This method estimates blood hemoglobin content and blood hemoglobin oxygen saturation within a
region of tissue a few millimeters in depth and with a diameter determined by the size of the probe. The
principle of reflectance spectrophotometry is based on photon diffusion theory.25(1039) A white light is
conducted through a fiberoptic bundle into the tissue. However, the light is not conducted straight
through the tissue; instead, photons diffuse into and reflect back from the tissue by scattering (Figure
253) and are absorbed by pigments such as oxyhemoglobin and deoxyhemoglobin within the tissue.
The resulting light is collected by the fiberoptic bundle and is conducted to a spectrophotometer that
consists of a grating (or a prism), a photodiode array, and a microprocessor. This spectrum is
subtracted from that of a white scatter (e.g., magnesium oxide) that has been "memorized" during
calibration of the instrument. The calculated difference reflects the spectra of oxyhemoglobin,
deoxyhemoglobin, and other pigments. Hemoglobins are the predominant pigments in the GI mucosa
when calculated from the mucosal concentration of hemoglobin26(1040) and cytochromes and their
extinction coefficients at wavelengths between 550 and 650 nm.27,28(1041) The following equation
can be used to estimate the concentration of hemoglobin using the difference in absorption (and
scattering) of two wavelengths:
where I0 and Ir are the reflected light intensities from a standard white scatter and from tissue,
respectively. F is the extinction and scattering from tissue except hemoglobin pigment. When 1 and
2 are close to each other, then the F1 and F2 should be similar. Thus log(Fl/F2) is neglected. If
l, a light path in a tissue, is constant, then c, the hemoglobin concentration, can be calculated from the
preceding equation.
(1042)Figure 253. Organ reflectance spectrophotometry. The incident light from the light
source is directed to the gastric mucosa through fiberoptic bundles. This light diffuses into the
mucosa and is attenuated by oxyhemoglobin and deoxyhemoglobin in erythrocytes within the
mucosal capillaries. The light reflected from the mucosa is then guided to the photodiode
array through the grating (or prism) of the spectrophotometer (not shown in this figure).
M.M.muscularis mucosae. (From Sato N, Kamada T, Shichiri M, et al. Measurement of
hemoperfusion and oxygen sufficiency in gastric mucosa in vivo: Evidence of mucosal
hypoxia as the cause of hemorrhagic shock-induced gastric mucosal lesion in rats.
Gastroenterology 1979; 76:8149.)
A series of organ reflectance spectrophotometers have been developed in our laboratory and used for
assessment of tissue microcirculation and oxygen metabolism. The TS-200, a commercially available
version of the reflectance spectrophotometer, was developed in collaboration with Sumitomo Electric
Co., Osaka, Japan. A microprocessor in the TS-200 instrument automatically calculates the indices of
tissue hemoglobin content (IHb) and tissue hemoglobin oxygen saturation (ISO2) according to the
following equations:29(1043)
where constant K (= 0.673) is calculated from the spectra of a purified hemoglobin solution using the
extinction coefficients presented by van Assendelft.30(1044)
The TS-200 instrument has a contact sensor that minimizes the pressure exerted by the probe.
Following early endoscopic studies of the esophagus, gastroduodenal mucosa, colon mucosa, and
liver in humans as well as experimental animals,28,3138(1045) growing numbers of endoscopists are
now using this instrument in their investigative work.3945(1046)
One advantage of organ reflectance spectrophotometry is a short measuring time. Hundreds of
seconds are required for a conventional spectrophotometer equipped with a photomultiplier to measure
a few hundred nanometers of spectra. With a linear photodiode array that simultaneously detects the
same spectra, the TS-200 instrument estimates mucosal blood hemoglobin content and mucosal
hemoglobin oxygen saturation in 30 to 180 ms. Another advantage is data stability.
The original developers of the reflectance spectrophotometry method reported on the hemoglobin
content and hemoglobin oxygen saturation in the stomach37,46,47(1047) and duodenum48(1048)
(Figure 254) using computer-assisted color displays. In this image analysis, 20 points or 24 points of
mucosal blood hemoglobin distribution and blood hemoglobin oxygen saturation in a stomach and a
duodenum were measured endoscopically in a point-by-point manner. The data were then processed
with a computer and displayed as a map on a color monitor. This technique is suitable for the
macroscopic analysis of mucosal blood distribution and mucosal blood oxygen saturation in the whole
stomach, although more than 10 minutes is required to obtain a color map in one patient.
(1049)Figure 254. Computer-assisted color display of gastric mucosal hemoglobin

concentration. The stomach is shown as an extended shape as if it was cut along the greater
curvature. On the right is the corpus, and on the left is the antrum. The color key in the left
lower part of the figure expresses the different levels of mucosal blood hemoglobin content.
Seven different colors (dark blue, green, light blue, yellow, pink, red, and white) are used.
Electronic Endoscopic Image Analysis of Hemoglobin Distribution

It is feasible to use an electronic endoscope and computer-assisted image processing for analysis of
mucosal hemoglobin distribution.4953(1050) Tsuji et al.53(1051) adapted computer-aided image
analysis to an electronic endoscopic system that uses the color wheel method of illumination (see
Chapter 3: Flexible Endoscope Technology: The Video Image Endoscope) to perform quantitative
assessment of the mucosal hemoglobin content. The GI mucosa is illuminated sequentially with red
(R), green (G), and blue (B) light. A charge-coupled device (CCD) detects light reflected from the
mucosa as corresponding electric signals (Vr for red light, Vg for green light, and Vb for blue light,
respectively) from which a color image is reconstructed by the videoprocessor in the endoscopic light
source. However, the Vr and Vg signals are also transferred to the computer memory of the image
analyzer, which produces a computerized image for hemoglobin distribution using the algorithm:
32{log2(Vr/Vg)}.
Because the 32{log2(Vr/Vg)} value at a localized area of mucosa relates linearly to mucosal blood
hemoglobin content as estimated by conventional organ reflectance spectrophotometry, and to tissue
blood flow measured by a hydrogen gas clearance technique in the respective area,53(1052)
distribution in 32{log2(Vr/Vg)} reflects the distribution of mucosal blood volume as well as mucosal
blood flow. Studies have shown that mucosal hemoglobin content at the margin of an ulcer is high
during healing stages and low when the ulcer is in an active stage (Figure 255).51(1053) These
results agree with earlier reports concerning mucosal blood volume at the margin of gastric
ulcers,32(1054) including those describing tissue blood flow at the ulcer margin using a hydrogen gas
clearance technique.3(1055)
(1056)Figure 255. Direct image analysis of mucosal hemoglobin (Hgb) distribution around
peptic ulcers at the gastric angle using an electronic endoscope. A, Top, electronic endoscopic
image of an active ulcer in a 52-year-old man. Bottom, color display shows mucosal Hgb
distribution around the ulcer. B, Top, electronic endoscopic image of an ulcer at the healing
stage. Bottom, corresponding color display of mucosal Hgb distribution. C, Top, electronic
image of an ulcer at the scarring stage. Bottom, corresponding color display of mucosal Hgb
distribution. The index of mucosal Hgb concentration is displayed according to an increasing
order, using the following eight colors: black, dark blue, light blue, green, pink, red, yellow,
and white. (A-C, From Tsuji S, Kawano S, Hayashi N, et al. Analysis of mucosal blood
hemoglobin distribution in gastric ulcers by computerized color display on electronic
endoscopy. Endoscopy 1991; 23:3214. Georg Thieme Verlag, Stuttgart, New York.)
Direct measurement of mucosal blood hemoglobin content distribution from a CCD image has various
advantages over other methods for measurement of tissue hemodynamics. Although this approach
estimates mucosal blood hemoglobin content rather than flow or flow velocity, the area measured is
smaller than 1 mm2, that is, theoretically, the resolution of the endoscope or that of the image
analyzer, whichever is lower. Furthermore, through image analysis of an electronic endoscopic image
as many as 10,000 points of tissue blood hemoglobin content can be measured and the distribution
can be displayed on a television monitor simultaneously (within 1 min); the procedure also avoids
problems inherent to contact methods that require that a probe be placed endoscopically on the
mucosa.
Other Methods
Valuable, albeit less popular, approaches to the assessment of GI hemodynamics have been reported
since the mid-1970s.
Acoustic Doppler Velocimetry

Unlike the laser Doppler method, which is used to measure tissue blood velocity, the acoustic or
ultrasound Doppler method measures blood velocity in a single vessel because ultrasound penetrates
more deeply into tissue than light does. Intraesophageal ultrasound probes have been used to assess
cardiac contractility and blood flow in large vessels. Identifying the feeding artery in a bleeding peptic
ulcer is important in the treatment of patients with massive hemorrhage, and a Doppler ultrasound
probe has also been developed for this purpose.54(1057) Matre et al.55(1058) reported on the use of
prototype ultrasound Doppler transducers (frequencies 5 and 10 MHz) for recording flow velocity in
vessels and varices within the wall of the GI tract.
Heat Clearance
Heat can also serve as a means to assess tissue blood flow using Fick's principle.9,10,12,56,57(1059)
A source of heat may be an electric heater, laser light,57(1060) or even water. A thermocoupler is the
most commonly used detector. The thermographic approach has been used in animal
experiments57(1061) but not in endoscopic studies.
Dye Appearance
Another approach to the assessment of GI blood flow is to observe the mucosa for the appearance of
a dye that has been administered in a blood vessel. Franke et al.58(1062) described intragastric
photography using indocyanine green (ICG) and a gastrocamera with infrared film. They injected 2
mg/kg of ICG into the cubital vein. The spread of the dye in the mucosa can be recorded with a series
of near-infrared photographs, although the appearance of the dye cannot be recorded quantitatively
with the naked eye. The density of ICG in a unit area probably reflects mucosal capillary density.
Delayed coloration of a specific mucosal area may indicate an insufficient blood supply to that region.
It is also possible to use ICG with an infrared electronic endoscope and a video tape recorder. When
used in conjunction with computer-assisted image analysis, it is possible to evaluate tissue blood
capillary density more quantitatively than with infrared photographs. Sodium fluorescein has also been
employed as a dye.59(1063)
Assessment of mucosal blood flow using dyes appears to be less quantitative than conventional
methods such as hydrogen gas clearance and reflectance spectrophotometry. However, the dye
method remains useful for assessment of the distribution of mucosal blood flow.
Microscopic Observation
In vivo microscopy is one of the standard methods to observe and measure tissue microcirculation.
Ultramagnifying endoscopes, with magnification factors of 170-fold, have been used successfully to
observe gastric60(1064) and colon mucosa [unpublished data] in humans. Velocities of erythrocytes,
leukocytes, and plasma are measured directly from a sequence of microscopic pictures (flying-spot
method) or indirectly by a comparison of video signals between points upstream and downstream on
the same vessel (dual-slit method).
Aminopyrine Clearance
Until recently, aminopyrine clearance was a standard method for determining gastric mucosal blood
flow only in animals. However, this method may be employed at endoscopy.6163(1065) It is based on
the principle that diffusion of weak bases (a pKa of 5 to 10) through lipid membranes depends on the
degree of ionization of the compound. At the pH of blood, the nonionized base diffuses freely through
the membrane. When it comes in contact with the considerably lower pH of the gastric juice, the base
dissociates and cannot diffuse back into the blood. Both plasma and gastric juice specimens are
obtained at intervals and analyzed for the weak base. The clearance, which represents pouch mucosal
plasma flow, is calculated by using the equation:
C = GV/P
where C is the clearance (ml/min), G and P are gastric juice and plasma concentration of the base,
respectively, and V is the gastric secretory rate. Aminopyrine was the most commonly used base, but
aniline, 99mTc,64(1066) and neutral red65(1067) have also been utilized. The method is currently less
popular because aminopyrine secretion is also related to the acid-secreting activity of gastric mucosa.
Xenon Gas Clearance
Omata66(1068) has used 133Xe gas for measurement of gastric blood flow. Spatial resolution in the
isotopic xenon gas clearance technique is low because of the size of the detector.
Oxygen Electrode
Tissue oxygen tension, which can be determined by polarography, is not a direct measure but an
indirect estimate of blood flow. A typical oxygen electrode consists of a silver electrode, a platinum
electrode, and a Teflon membrane permeable to oxygen. Although contact-type and needle-type
oxygen electrodes have been developed successfully by several investigators, the commercially
available electrodes do not appear stable or durable enough for clinical use.
Laser Speckle
This method, currently under development, utilizes low-power laser illumination of tissue. Light
reflected from the subject has noise or so-called speckles, the number of which is thought to depend
on tissue blood flow. By analyzing the speckle pattern, blood flow can be estimated. An advantage of
this method is that an image of blood flow distribution can be obtained. Because of the unstable nature
of the speckles, however, the range of reliable measurement appears to be limited.
Intravascular-Pressure Measurement
Variceal vessels have been punctured for the purpose of measuring intravascular pressure. This is
usually followed by endoscopic injection sclerotherapy. A pressure-sensitive gauge has also been used
for noninvasive measurement of esophageal variceal pressure at endoscopy.67(1069)
Histologic Analyses of Biopsy Specimens
Although it is neither quantitative nor objective, histologic observation may be of assistance in the
assessment of GI blood flow. Staining specifically for vascular endothelium (e.g.,
immunohistochemistry against Factor VIII) is preferable. Careful observation of erythrocytes within
mucosal and submucosal capillaries may suggest the congestive state of the GI microcirculation.
Human GI Tract and Other Digestive Organs

This section focuses on physiologic and pathologic aspects of tissue circulation in the human GI tract.
Data based on work with experimental animals or results obtained by nonendoscopic procedures such
as angiography are beyond the scope of this review. For full comprehension of this topic, the reader is
encouraged to consult other reviews that include GI blood flow in experimental models.68,69(1070)
Esophagus
Various arteries supply the esophagus: inferior thyroidal arteries supply the cervical esophagus; proper
esophageal arteries supply the thoracic esophagus; and left gastric artery and infraphrenic artery
supply the abdominal esophagus. It is not surprising, therefore, that the distribution of mucosal blood
flow is heterogeneous.70(1071)
Kawano et al.36(1072) reported that blood volume and blood hemoglobin oxygenation in esophageal
tissue are not affected by age. There are no differences in mucosal circulation between men and
women. Esophageal mucosal blood volume is significantly higher in patients with liver cirrhosis than in
subjects without liver disorders.36(1073) The estimated oxygen saturation is lower in patients with liver
cirrhosis than in normal subjects. Changes in these parameters are more remarkable in patients with
esophageal varices than in those without varices. With regard to the risk of variceal bleeding, Rigau et
al.67(1074) reported that variceal pressure was significantly higher in patients who had sustained
variceal bleeding than in those who had not bled (15.7 2.8 vs. 12.1 2.6 mm Hg) (see Chapter 31:
Variceal Bleeding).
Stomach
Intragastric Blood Flow Distribution and the Effect of Aging
By all methods of analysis (hydrogen gas clearance, laser Doppler flowmetry, reflectance
spectrophotometry, electronic endoscopic image analysis), mucosal blood flow is greater in the corpus
of the stomach than in the antrum and at the angulus, and it is less at the lesser curvature than at the
greater curvature. Murakami et al.3,71(1075) reported that gastric mucosal blood flow is larger in the
proximal corpus (approximately 80 ml/min/100 g tissue) and middle corpus, and smaller in the angulus,
antrum, and distal corpus. Intragastric distribution of mucosal hemoglobin content was investigated
more intensively by Kamada et al.32(1076) and Kawano et al.47(1077) Mucosal hemoglobin content in
healthy subjects is higher at the greater curvature in the proximal and middle corpus and lower at the
lesser curvature in the antrum and angulus (see Figure 254). It is reported that gastric mucosal blood
flow and mucosal hemoglobin content decrease with age.71,72(1078)
Neural and Psychologic Effects
Although the first observations of neural and psychologic influences on gastric mucosal blood flow
were made in a patient with a gastric fistula (Alexis St. Martin) by Beaumont73(1079) in 1833,
information on this topic remains very limited with respect to human subjects. In patients with severe
head injury, however, the gastric mucosal hemoglobin content and the mucosal hemoglobin oxygen
saturation decrease (see later in this chapter).
Prostaglandins and Nonsteroidal Anti-Inflammatory Drugs
Prostaglandins have been reported to play important roles in the regulation of mucosal blood flow, acid
secretion, and so-called cytoprotection against various necrotizing agents.74,75(1080) A double-blind
study using reflectance spectrophotometry demonstrated that a synthetic prostaglandin increases
gastric mucosal hemoglobin content.37(1081) Some nonsteroidal anti-inflammatory drugs such as
indomethacin decrease gastric mucosal blood flow in humans, whereas others do not.76(1082)
Smoking
Smoking is one of the most important factors in delayed healing and recurrence of peptic ulcer
disorders. It takes only 1 to 3 puffs of cigarette smoke to decrease gastric mucosal hemoglobin content
and mucosal oxygen saturation for a period of 20 min.77(1083) These ischemic changes are greater in
nonhabitual smokers than in habitual smokers.78(1084) Factors other than nicotine are involved in
decreased gastric mucosal blood flow after smoking. Both ordinary tobacco and nicotine-free
cigarettes decrease mucosal blood hemoglobin content and mucosal oxygenation [unpublished data].
Physical Stress
In patients with severe burns or head injuries, gastric mucosal blood volume and mucosal blood
oxygen saturation decrease before the development of stress ulceration.31(1085) Trench ulcer is a
large and shallow ulceration developed at the gastric corpus and is often found in patients who suffer
severe somatic stress. An aminopyrine clearance method study suggests that gastric mucosal blood
flow may also decrease in patients with a trench ulcer.63(1086) Surgical intervention such as proximal
gastric vagotomy lowers gastric perfusion.79(1087)
Peptic Ulcer
Mucosal blood volume (i.e., mucosal blood hemoglobin content) decreases at the edge of an early
stage (active stage) gastric ulcer (see Figure 255A);32,47(1088) this indicates the presence of gastric
mucosal ischemia at this stage. Gastric mucosal blood flow and mucosal blood hemoglobin content at
the ulcer margin reportedly increase during the healing stages of tractable gastric ulcers (see Figure
255B and 255C).3,20,32,80(1089) In an intractable ulcer, however, mucosal blood hemoglobin fails
to increase.32(1090)
Liver Cirrhosis
Gastric bleeding in addition to esophageal bleeding is often encountered at endoscopy in patients with
liver cirrhosis or portal hypertension. Although circulation in the GI wall is reportedly in a hyperdynamic
state, circulation in gastric mucosa appears congestive. In a study of 23 cirrhotic patients, the gastric
mucosal hemoglobin content increased, and the mucosal hemoglobin oxygen saturation decreased, as
the portal pressure rose.44(1091) Decreased gastric mucosal blood flow and mucosal hypoxemia were
found in another study;81(1092) a hydrogen gas clearance study demonstrated decreased gastric
mucosal blood flow and increased submucosal blood flow in patients with portal hypertension
compared with control subjects.82(1093) Using the laser Doppler method, however, Chung et
al.21(1094) reported increased gastric blood flow in subjects with portal hypertension. These conflicting
results may be due to differences in instrumentation or to differences between groups of subjects.
Although a number of investigators disagree, endoscopic sclerotherapy appears to alter gastric
mucosal hemodynamics.21(1095) When patients with esophageal varices are subjected to esophageal
transection, gastric mucosal blood flow is reduced by approximately 30% during surgery and by 20% at
4 weeks after operation. During and after left gastric vena cava shunting, however, the mucosal blood
flow tends to be preserved with a rate of reduction of less than 10%.82(1096)
Transcatheter arterial embolization is used to treat patients with liver cirrhosis and hepatocellular
carcinoma. Gastric mucosa becomes ischemic at the lesser curvature of the antrum after transcatheter
embolization from the hepatic artery proper. This may be due to regurgitation of the emboli into the
right gastric artery.
Aortic Aneurysm
Patients with aortic aneurysm have an increased incidence of peptic ulcer disease. Gastric mucosal
blood flow at the antrum, angulus, and corpus has been reported as significantly lower in patients with
aortic aneurysm than in control subjects.83(1097)
Duodenum
Murakami et al.3(1098) reported that bulbar mucosal blood flow was 52 9 ml/100 g at the anterior
wall, 70 10 ml/min/100 g at the lesser curvature, and 62 13 ml/min/100 g at the posterior wall.
Blood hemodynamics in patients with duodenal ulcers has been studied by Fukuda and
Kamada48(1099) (Figure 256). Duodenal mucosal hemoglobin content at the ulcer edge significantly
decreases during the active stage, whereas that of the mucosa surrounding the edge increases.
During the healing stage, mucosal hemoglobin content increases at the ulcer edge and in the
surrounding mucosa; in this stage, mucosal hemoglobin content is higher at the ulcer edge than at the
surrounding mucosa. The index of mucosal hemoglobin content remains high at the early-stage scar,
the so-called red scar, whereas the hemoglobin content returns to normal at the late-stage scar, the
white scar. The index of mucosal hemoglobin oxygen saturation does not change significantly during
the healing process of the tractable ulcer. Using similar instrumentation, Leung et al.84(1100) reported
that mucosa at the ulcer margin is less oxygenated in an intractable duodenal ulcer than in a tractable
lesion.
(1101)Figure 256. Gastroduodenal mucosal blood hemoglobin distribution in a patient with

duodenal ulcer (D.U.). A, Gastroduodenal mucosal hemoglobin distribution in a patient with
an active duodenal ulcer. B, Gastroduodenal mucosal blood hemoglobin distribution in a
patient with a healing duodenal ulcer. The index of mucosal blood hemoglobin content is
expressed using seven different colors as shown in Figure 254. (A and B, From Fukuda M,
Kamada T. Duodenal mucosal hemodynamics in the healing process of duodenal ulcer.
Gastroenterol Endosc 1985; 27:20668.)
Small Intestine
Nojiri and Tanaka85(1102) applied reflectance spectrophotometry and the hydrogen gas clearance
method for measurement of small intestinal hemodynamics in humans and dogs in 1986. They found
that an amino acid solution increased intestinal blood flow and intestinal mucosal hemoglobin content
in dogs. Studies in humans demonstrated that intravenous injection of butyl scopolamine bromide
possibly decreases mucosal blood flow. However, these authors found it difficult to insert a platinum
electrode into the intestinal mucosa. Information on circulation in the human small intestine in vivo
remains very limited.
Colon
Blood Flow Heterogeneity
Miwa et al.86(1103) found marked differences in colonic blood flow depending on the segment of the
colon studied. Using a needle-type hydrogen gas clearance method, they reported colon blood flows of
65 18 ml/min/100 g tissue at the splenic flexure and 46 15 ml/min/100 g tissue at the sigmoid
colon. Differences such as these may be due to differences in the vasculature between segments of
the colon, although it remains unknown which layer of colon wall was actually measured.
Inflammatory Bowel Diseases
Although most animal studies have demonstrated mucosal blood flow stasis and congestion during
experimental inflammation in the colon, data obtained at endoscopy have been conflicting. Su et
al.87(1104) have reported that colon mucosal hemoglobin content and mucosal hemoglobin oxygen
saturation (IHb and ISO2, respectively) were significantly higher in patients with active inflammatory
bowel diseases than in healthy subjects. Using similar instrumentation and electronic endoscopic
image analysis, however, Kawano et al.88(1105) found a significant increase in mucosal hemoglobin
content and a decrease in mucosal hemoglobin oxygen saturation in patients with ulcerative colitis.
Liver, Pancreas, and Other Abdominal Organs

Hepatic hemodynamics have been studied at laparoscopy using reflectance
spectrophotometry34,35,8991(1106) and the hydrogen gas clearance method.92(1107) Glucagon is
reported to increase hepatic hemoglobin content. Lactate increases hepatic oxygen consumption and
hepatic hemoglobin content. Regional hepatic hemoglobin content decreases in patients with chronic
hepatitis and liver cirrhosis as these diseases develop. Pancreatic blood flow was measured by Ishida
et al.93(1108) during laparoscopy, which can also be used to determine blood flow in other abdominal
tissues.
REFERENCES
1.
2.
Aukland K, Bower BP, Berliner RW. Measurement of local blood flow with hydrogen gas. Circ
Res 1964;14:16487.
Murakami M, Moriga M, Miyake T, et al. Contact electrode method in hydrogen gas clearance
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
technique: A new method for determination of regional mucosal blood flow in animals and
humans. Gastroenterology 1982;82:45767.
Murakami M, Inada M, Miyake T, et al. Regional mucosal blood flow and ulcer healing. In Koo
A, Lam SK, Smaje LH (eds). Microcirculation of the Alimentary Tract. 1st ed. Vol 1.
Singapore: World Scientific, 1983;293302.
Miyamoto J, Takase Y, Takeshima T, et al. Gastric mucosal blood flow in patients with gastric
ulcer. Gastroenterol Endosc 1981;23:78191.
Ashley SW, Cheung LY. Measurement of gastric mucosal blood flow by hydrogen gas
clearance. Am J Physiol 1984;247:G33945.
Cheung LY, Sonnenschein LA. Measurement of regional gastric mucosal blood flow by
hydrogen gas clearance. Am J Surg 1984;147:327.
Cheung LY. Gastric mucosal blood flow: Its measurement and importance in the mucosal
defense mechanisms. J Surg Res 1984;36:2828.
Leung FW, Guth PH, Scremin OU, et al. Regional mucosal blood flow measurements by
hydrogen gas clearance in the anesthetized rat and rabbit. Gastroenterology 1984;87:2836.
Demling L, Ottenjann R, Wachsmann F. Method for measurement of blood flow and acidity in
the human stomach. Am J Dig Dis 1964;9:51724.
Bell PRF, Shelley T. Gastric mucosal blood flow and acid secretion in conscious animals
measured by heat clearance. Am J Dig Dis 1968;13:68596.
Leung FW, Washington J, Kauffman GLJ, et al. Endoscopic measurement of gastric corpus
mucosal blood flow in conscious dogs. Dig Dis Sci 1986;31:62530.
Shimakura S, Fukutomi H, Miyamoto J, et al. Endoscopic application of thermal diffusion
method for continuous measuring gastric mucosal blood flow in dogs. Gastroenterol Endosc
1987;29:88292.
Hirayama Y, Niwa H, Miki K, et al. Measurement of gastric mucosal blood flow through
endoscope with a bipolar electrode. Gastroenterol Endosc 1982;24:8737.
Stosseck K, Luebbers DW, Coffin N. Determination of local blood flow (microflow) by
electro-chemically generated hydrogen. Construction and application of the measuring probe.
Pflugers Arch 1974;348:22538.
Miyamoto J, Fubkutomi H, Shimakura H, et al. Experimental investigation measuring gastric
blood flow using hydrogen gas generated by electrolysis. Gastroenterol Endosc
1983;25:165865.
Koshu K, Kamiyama K, Oka N, et al. Measurement of regional blood flow using hydrogen gas
generated by electrolysis. Stroke 1982;13:4837.
Shepherd AP, Riedel GL. Continuous measurement of intestinal mucosal blood flow by
laser-Doppler velocimetry. Am J Physiol 1982;242:G66872.
Ahn H, Lindhagen J, Nilsson GE, et al. Evaluation of laser-Doppler flowmetry in the
assessment of intestinal blood flow in cat. Gastroenterology 1985;88:9517.
Shepherd AP, Riedel GL, Kiel JW, et al. Evaluation of an infrared laser-Doppler blood
flowmeter. Am J Physiol 1987;252:G8329.
Saita H, Murakami M, Tatsumi K, et al. Gastric mucosal blood flow during the healing process
of the initial and the recurrent gastric ulcers. Gastroenterol Endosc 1987;29:87381.
Chung RS, Bruch D, Dearlove J. Endoscopic measurement of gastric mucosal blood flow by
laser Doppler velocimetry: Effect of chronic esophageal variceal sclerosis. Am Surg
1988;54:11620.
Okano H, Tsuji H, Kodama T, et al. Effect of endoscopic sclerotherapy and propranolol on the
blood flow of esophageal varices. Am J Gastroenterol 1990;85:303.
Ahn H, Ivarsson LE, Johansson K, et al. Assessment of gastric blood flow with laser Doppler
flowmetry. Scand J Gastroenterol 1988;23:120310.
Nilsson GE, Tenland T, Oberg PA. Evaluation of a laser Doppler flowmeter for measurement
of tissue blood flow. IEEE Trans Biomed Eng 1980;27:597605.
Wilson BC, Jacques SL. Optical reflectance and transmittance of tissues: Principles and
applications. IEEE J Quantum Electron 1990;26:218699.
26. Sato N, Kamada T, Shichiri M, et al. Measurement of hemoperfusion and oxygen sufficiency
in gastric mucosa in vivo: Evidence of mucosal hypoxia as the cause of hemorrhagic
shock-induced gastric mucosal lesion in rats. Gastroenterology 1979;76:8149.
27.
Sato N, Shichiri M, Matsumura T, et al. Non-destructive measurement of tissue oxidative
functions: Measurement of concentration of respiratory enzymes and the rate of oxygen
consumption in living liver tissue by reflectance spectrophotometry. In Scarpa A, Dutton PL,
Leigh J (eds). Frontiers in Biological Energetics. Vol II. New York: Academic Press,
1978;150714.
28. Sato N, Kamada T, Kawano S, et al. Oxidative and phosphorylative activities of the gastric
mucosa of animals and humans in relation to the mechanism of stress ulcer. Biochim Biophys
Acta 1978;538:23643.
29. Sato N, Matsumura T, Shichiri M, et al. Hemoperfusion, rate of oxygen consumption and
redox levels of mitochondrial cytochrome c((MathematicalPi-One )1cl) in liver in situ of
anesthetized rat measured by reflectance spectrophotometry. Biochim Biophys Acta
1981;634:110.
30.
Van Assendelft OW. Spectrophotometry of Haemoglobin Derivatives. Springfield, IL: CC
Thomas. 1970.
31. Kamada T, Sato N, Kawano S, et al. Gastric mucosal hemodynamics after thermal or head
injury: A clinical application of reflectance spectrophotometry. Gastroenterology
1982;83:53540.
32. Kamada T, Kawano S, Sato N, et al. Gastric mucosal blood distribution and its changes in the
healing process of gastric ulcer. Gastroenterology 1983;84:15416.
33.
Sato N, Kamada T, Kawano S, et al. Microcirculation, mucosal oxygen utilization and gastric
ulcer. In Koo A, Lam SK, Smaje LH (eds). Microcirculation of the Alimentary Tract. 1st ed. Vol
1. Singapore: World Scientific, 1983;30312.
34. Sato N, Hayashi N, Kawano S, et al. Hepatic hemodynamics in patients with chronic hepatitis
or cirrhosis as assessed by organ-reflectance spectrophotometry. Gastroenterology
1983;84:6116.
35. Kamada T, Hayashi N, Sato N, et al. Estimated hepatic oxygen consumption in patients with
chronic liver diseases as assessed by organ reflectance spectrophotometry. Dig Dis Sci
1986;31:11924.
36. Kawano S, Fukuda M, Tsuji S, et al. The esophageal mucosal hemodynamics in patients with
liver cirrhosis. Nippon Shokakibyo Gakkai Zasshi 1986;83:2297301.
37.
Sato N, Kawano S, Fukuda M, et al. Misoprostol-induced changes in gastric mucosal
hemodynamics. A double-blind parallel study in human volunteers. Am J Med 1987;83(suppl
1):15S-21S.
38. Fukuda M, Kawano S, Sato N, et al. Gastric mucosal hemodynamics and mucosal
oxygenation in patients with liver cirrhosis. Nippon Shokakibyo Gakkai Zasshi 1986;83:2632.
39.
Shimazu R, Sato N, Kawano S, et al. Evaluation of hemodynamics in the gastric mucosa,
submucosa and muscle using organ-reflectance spectrophotometry. In Kitajima M (ed).
Cytoprotection and Cytobiology. Vol 2. Amsterdam: Excerpta Medica, 1985;1608.
40.
Shimazu R, Sato N, Kawano S, et al. Effects of ethanol on hemodynamics in gastric mucosa
and submucosa. In Tsuchiya M (ed). Cytoprotection and Cytobiology. Vol 3. Amsterdam:
Excerpta Medica, 1986;199205.
41.
Kamada T, Sato N. Endoscopy in future: Function imaging by optoelectronics. Gastroenterol
Endosc 1983;25:7712.
42.
Kawano S, Sato N, Tsuji S, et al. Gastric microcirculatory changes associated with
physicochemical and ionic mediators. In Manabe H, Zweifach BW, Messmer K (eds).
Microcirculation in Circulatory Disorders. 1st ed. Vol 1. Tokyo: Springer-Verlag, 1988;18994.
43.
Tominaga S, Ueno K, Saito H, et al. Evaluation of reflectance spectrophotometry (model
TS-200) on the measurement of gastric mucosal blood flow in dog and human. Gastroenterol
Endosc 1985;27:72637.
44.
Shimotsuura Y, Toyonaga A. Clinical application of reflectance spectrophotometry in portal
hypertension. Gastroenterol Endosc 1986;28:24325.

45. Leung FW, Slodownik E, Jensen DM, et al. Gastroduodenal mucosal hemodynamics by
endoscopic reflectance spectrophotometry. Gastrointest Endosc 1987;33:2848.
46.
Kawano S, Fukuda M, Sato N, et al. A three-dimensional display and a colour display of
gastric mucosal blood volume. Gastroenterol Endosc 1982;24:13748.
47. Kawano S, Sato N, Tsuji S, et al. Two-dimensional computer color graphics of gastric
mucosal blood distribution in normal subjects and ulcer patients. Endoscopy 1991;23:31720.
48.
Fukuda M, Kamada T. Duodenal mucosal hemodynamics in the healing process of duodenal
ulcer. Gastroenterol Endosc 1985;27:20668.
49.
Tsuji S, Sato N. Electronic endoscopy and fiberoptic endoscopy: An infrared endoscopy.
50.
Tsuji S, Sato N, Kawano S, et al. Prediction of gastric ulcer healing in the view of mucosal
hemodynamics using electronic endoscopy and reflectance spectrophotometry [Abstract].
Gastroenterology 1988;94:A466.
51. Tsuji S, Kawano S, Hayashi N, et al. Analysis of mucosal blood hemoglobin distribution in
gastric ulcers by computerized color display on electronic endoscopy. Endoscopy
1991;23:3214.
52. Sato N, Tsuji S, Kawano S, et al. Computer-assisted two-dimensional analysis of gastric
mucosal hemoglobin distribution using a newly modified electronic endoscope. Endoscopy
1992;24(suppl 2):5226.
53. Tsuji S, Sato N, Kawano S, et al. Functional imaging for the analysis of the mucosal blood
hemoglobin distribution using electronic endoscopy. Gastrointest Endosc 1988;34:3326.
54. Beckly DE, Casebow MP, Pottengell KE. The use of a Doppler ultrasound probe for localising
arterial blood flow using upper gastrointestinal endoscopy. Endoscopy 1982;14:1467.
55. Matre K, Odegaard S, Hausken T. Endoscopic ultrasound Doppler probes for velocity
measurement in vessels in the upper gastrointestinal tract using a multifrequency pulsed
Doppler meter. Endoscopy 1990;22:26870.
56. Hiramatsu A, Watanabe T, Okuhira M, et al. Endoscopical determination of gastric mucosal
blood flow by crossed thermocouple method. Gastroenterol Jpn 1984;19:21620.
57. Kamada T, Sato N, Kakubari N, et al. Noninvasive assessment of microcirculation of living
organs and tissues using laser. Lasers Surg Med 1983;2:27580.
58. Franke J, Lux G, Demling L. Intragastric infrared photography in conjunction with infrared
absorption angiography. Gastrointest Endosc 1985;31:879.
59. Perbeck L, Nyberg B, Sonnenfeld T, et al. Endoscopic fluorescein flowmetry in the evaluation
of gastric and duodenal mucosal blood flow. Scand J Gastroenterol 1987;22:43742.
60.
Sakura H, Sato N, Kawano S, et al. Endoscopic observation of gastric microstructure and
gastric microcirculation. In Tsuchiya M, Asano M, Mishima Y, et al. (eds).
MicrocirculationAn Update. 1st ed. Vol 1. Amsterdam: Elsevier Science, 1987;28990.
61. Jacobson ED, Linford RH, Grossman MI. Gastric secretion in relation to mucosal blood flow
studied by a clearance technique. J Clin Invest 1966;45:113.
62. Guth PH, Baumann H, Grossman MI. Measurement of gastric mucosal blood flow in man.
63.
Fusamoto H, Saito M, Masuzawa M, et al. Possible relationship between development of
Trench ulcer and gastric mucosal blood flow. Gastroenterol Endosc 1979;21:7007.
64. Taylor TV, Pullan BR, Elder JB, et al. Observations of gastric mucosal blood flow using
99mTc in rat and man. Br J Surg 1975;62:78891.
65.
Knight SE, McIsaac RL. Neutral red clearance as an estimate of gastric mucosal blood flow in
man. J Physiol (Lond) 1977;272:62P-3P.
66.
Omata S. Secretin. J Jap Soc Intern Med 1974;63:103741.
67. Rigau J, Bosch J, Bordas JM, et al. Endoscopic measurement of variceal pressure in
cirrhosis: Correlation with portal pressure and variceal hemorrhage. Gastroenterology
1989;96:87380.
68.
Guth PH, Leung FW. Physiology of the gastric circulation. In Johnson LR (ed). Physiology of
the Gastrointestinal Tract. 2nd ed. Vol 1. New York: Raven Press, 1987;103153.
69. Sato N, Kawano S, Tsuji S, et al. Microvascular basis of gastric mucosal protection. J Clin
Gastroenterol 1988;10(suppl 1):S138.
70.
Miyoshi H, Shikata J, Tokura Y. A clinical study on mucosal tissue blood flow of esophagus
using endoscopic laser Doppler flowmetry. Gastroenterol Endosc 1992;34:12527.
71. Murakami M, Miyake T, Moriga M, et al. Measurement of regional gastric mucosal blood flow
in chronic gastric ulcer patients and polypectomy-induced ulcer patients. Nippon Shokakibyo
Gakkai Zasshi 1983;80:234751.
72. Sato N, Nakagawa A, Kawano S, et al. Gastric mucosal hemodynamics and aging: Clinical
application of a non-invasive, consecutive, high-speed reflectance spectrophotometry. Nippon
Shokakibyo Gakkai Zasshi 1981;78:20748.
73.
Beaumont W. Experiments and Observations on the Gastric Juice and the Physiology of
Digestion. 1st ed. New York: Dover Publications. 1833.
74.
Miller TA. Protective effects of prostaglandins against gastric mucosal damage: Current
knowledge and proposed mechanism. Am J Physiol 1983;245:G6013.
75. Robert A. Cytoprotection by prostaglandins. Gastroenterology 1979;77:7617.
76. Konturek SJ, Kwiecien N, Obtulowicz W, et al. Effect of carprofen and indomethacin on
gastric function, mucosal integrity and generation of prostaglandin in men.
77. Kawano S, Sato N, Tsuji S, et al. Effect of cigarette smoking on the gastric mucosal blood
volume index and hemoglobin oxygenation in man. Gastroenterol Jpn 1989;24:17.
78. Kawano S, Sato N, Fukuda M, et al. Effect of cigarette smoking on the gastric hemodynamics:
Analysis by reflectance spectrophotometry. Nippon Shokakibyo Gakkai Zasshi
1982;79:18792.
79. Ahn H, Johansson K, Lindhagen J. Local blood flow changes in association with proximal
vagotomy. Scand J Gastroenterol 1986;21:9614.
80. Li ZS. Changes of mucosal blood flow in the healing process of peptic ulcer. Chung Hua I
Hsueh Tsa Chih 1989;69:2035.
81.
Adachi H. The study of gastric mucosal changes and its mechanism in patients with liver
cirrhosis. Gastroenterol Endosc 1987;29:45571.
82. Mitarai Y, Kobayashi M. Correlation between gastric microcirculation and mucosal injury after
surgical therapy of esophageal varices. Nippon Geka Gakkai Zasshi 1990;91:10113.
83. Konno H, Sakaguchi S, Aoki K, et al. The formation of gastric ulcers with a tendency to
hemorrhage in association with aortic aneurysm. Jpn J Surg 1991;21:26871.
84. Leung FW, Reedy TJ, Van Deventer GM, et al. Reduction in index of oxygen saturation at
margin of active duodenal ulcers may lead to slow healing. Dig Dis Sci 1989;34:41723.
85.
Nojiri H, Tanaka M. Endoscopic measurement of hemodynamics of the small intestinal
mucosa. Gastroenterol Endosc 1986;28:24302.
86.
Miwa Y, Maeda A, Akagami A, et al. Colonic mucosal blood flow measurement in men and
rats. Gastroenterol Endosc 1983;25:4217.
87. Su KC, Leung FW, Guth PH. Assessment of mucosal hemodynamics in normal human colon
and patients with inflammatory bowel disease. Gastrointest Endosc 1989;35:227.
88.
Kawano S, Sato N, Ogihara T, et al. Hemodynamics of colon mucosa in patients with
ulcerative colitis. A new evaluation method by electronic endoscopy. Gastroenterology
1990;98:A180.
89. Kasahara A, Hayashi N, Sasaki Y, et al. Hepatic circulation and hepatic oxygen consumption
in alcoholic and nonalcoholic fatty liver. Am J Gastroenterol 1988;83:8469.
90. Sato N, Kawano S, Matsumura T, et al. Characterization of hepatic hemodynamics in
cirrhotics and non-cirrhotics. Effect of glucagon infusion. Scand J Gastroenterol
1989;24:22330.
91. Lieber CS, Baraona E, Hernandez-Munoz R, et al. Impaired oxygen utilization. A new
mechanism for the hepatotoxicity of ethanol in sub-human primates. J Clin Invest
1989;83:168290.
92. Hayashi N, Kasahara A, Kurosawa K, et al. Oxygen supply to the liver in patients with
alcoholic liver disease assessed by organ reflectance spectrophotometry. Gastroenterology
1985;88:8816.
93. Ishida H, Makino T, Kobayashi M, et al. Laparoscopic measurement of pancreatic blood flow.
Endoscopy 1983;15:10710.
Chapter 26 Acute Upper Gastrointestinal Bleeding

(1109)
(1110)
DAVID A. GILBERT, M.D.
FRED E. SILVERSTEIN, M.D.
This chapter deals with the role of endoscopic diagnosis in upper gastrointestinal bleeding. Several
excellent reviews of this subject have been published.15(1111) Gastrointestinal bleeding in relation to
portal hypertension is covered in detail in subsequent chapters. The advantages of endoscopy for the
bleeding patient, specifically the information gained by endoscopy that may aid in management and
influence the ultimate outcome of the bleeding episode, are discussed. This topic was the subject of
two consensus workshops sponsored by the National Institutes of Health in the United
States.6,7(1112)
Upper gastrointestinal bleeding is commonly encountered in clinical medicine. It is estimated that
between 50 and 150 patients per 100,000 members of the general population are hospitalized annually
for bleeding.8,9(1113) In the United States, with a population of approximately 250 million, about
250,000 patients present with upper gastrointestinal bleeding yearly.10(1114) Despite major advances
in monitoring and intensive care, the mortality rate of approximately 10% for such patients has
changed little since 1940.11,12(1115) One possible explanation, reflecting advances in other aspects
of medical care, is that patients in later studies are older (>60 years) than the patients studied during
the period from the mid-1950s to the mid-1960s and that they often have other serious illnesses.
Bordley et al.13(1116) identified six early factors that predicted a good outcome for patients with acute
gastrointestinal bleeding: age younger than 75 years, absence of an unstable comorbid illness,
absence of ascites, a normal prothrombin time, systolic blood pressure of 100 mm Hg or higher, and
the absence of free or fresh blood in the nasogastric aspirate. The mortality rate for a group of 1017
patients with upper gastrointestinal bleeding in the study of Katschinski et al.14(1117) was 9%. All but
4 of the 91 patients who died were older than 60 years of age. Death followed recurrent bleeding in 13
patients (14%). In some cases, gastrointestinal hemorrhage had developed in patients with advanced
diseases of other organ systems. However, bleeding was not the immediate cause of death of most
patients; 81% died of a concomitant disease that had been exacerbated by gastrointestinal bleeding.
In a retrospective review of the records of 115 patients treated for upper gastrointestinal bleeding,
including patients with portal hypertension, Larson et al.15(1118) identified several factors that
predicted mortality: age older than 60 years, disease in three organ systems, transfusion of five or
more units of blood, surgery for control of bleeding, and pulmonary or liver disease (or both). A total of
26 (23%) patients had died, and the deaths of 19 were attributed to another underlying condition.
Predictably, mortality was highest for patients with bleeding varices (36%) and lowest for those with
duodenal (7.7%) and gastric (15.8%) ulcers.
Factors that predicted death of patients with upper gastrointestinal bleeding in the study of Provenzale
et al.16(1119) were severity of bleeding (i.e., hematochezia, absence of melena, short duration of
bleeding, and 5% decrease in hematocrit), liver disease (i.e., prolonged prothrombin time or
encephalopathy), and renal disease. The findings in the study of Supe et al.17(1120) were similar.
Advanced age, clinical evidence of shock, hemoglobin concentration less than 8.0 g/dl, and hepatic
dysfunction were associated with an increase in mortality.
Matthewson et al.18(1121) prospectively studied 139 consecutive patients with peptic ulcers who
presented with bleeding and 74 patients with ulcers who presented with pain without bleeding. Patients
with bleeding were more likely to have had a previous complication of peptic ulcer disease, to have
taken nonsteroidal anti-inflammatory drugs (NSAIDs) within the prior month, to be older (66 vs. 51
years), and to have ulcers with diameters larger than 2 cm. The mortality rate for those with upper
gastrointestinal bleeding in a survey of 44 centers in 21 countries sponsored by the World Organization
of Gastroenterology was 8.3%. However, the mortality rate for patients with recurrent bleeding,
including those with portal hypertension, was 30.7%.12(1122)
Approach to the Patient

When a patient presents with upper gastrointestinal bleeding, an initial brief assessment is appropriate.
Often, this requires the judicious and expeditious use of the entire range of the physician's skills. The
assessment may be carried out during or immediately after institution of resuscitative measures. It
should include inquiry concerning prior episodes of upper gastrointestinal bleeding, known peptic ulcer
disease or gastrointestinal malignancy, and ingestion of drugs that cause gastrointestinal inflammation
or ulceration. Because death related to gastrointestinal bleeding frequently results from exacerbation of
a concurrent serious condition involving another organ system, it is important to determine whether the
patient has any other disorders that may have a bearing on outcome.
NSAIDs have been recognized as an important etiologic factor in peptic ulcer disease, a common
cause of major gastrointestinal hemorrhage. The United States Food and Drug Administration has
estimated that, among the more than 10 million people in the United States who consume these
agents on a daily basis, 2 to 4% who do so for 1 year develop a symptomatic ulcer, an episode of
bleeding from an ulcer, or an ulcer perforation.19(1123) Among patients who bleed from ulcers,
approximately half may give a history of recent NSAID ingestion.18,20(1124) In a hospital-based
case-control study, Levy et al.21(1125) found that the risk of major gastrointestinal bleeding was
substantially increased by the regular (i.e., 4 days per week) and by the occasional use of aspirin. A
history of ingestion of NSAIDs is of interest for any patient with gastrointestinal hemorrhage.22(1126)
All NSAIDs appear to be capable of causing ulceration, and the risk is reported to increase with intake
of multiple NSAIDs and with high doses of these agents.20(1127) Increased ulcer risk also appears to
be linked with concomitant use of corticosteroids and NSAIDs.23(1128)
Gastrointestinal bleeding is sometimes problematic in patients who receive anticoagulant therapy.
Wilcox and Truss24(1129) described 15 such patients. A diagnosis could be established for 81% of
cases of upper gastrointestinal bleeding, but the cause of lower gastrointestinal bleeding frequently
remained occult despite appropriate evaluations. The most common cause of upper gastrointestinal
bleeding was peptic ulcer disease.
Symptoms of recent bleeding, such as melena, passage of red blood per rectum, and hematemesis,
should be elucidated. Historical or physical evidence of other intercurrent illnesses must not be
overlooked, because certain conditions such as diabetes, various forms of heart disease, and
hypertension have a bearing on management. During the brief but thorough physical examination, the
acuteness and severity of the bleeding must be estimated. Accurate determinations of the heart rate
and blood pressure, including orthostatic changes in blood pressure, are essential. A careful inspection
for signs of liver disease, including spider telangiectasia, jaundice, and ascites, is also essential.
Emesis and stool specimens should be examined if possible.
As the assessment proceeds, resuscitative measures should be instituted, starting with placement of
one or more large-bore intravenous catheters. According to the estimated severity of blood loss,
appropriate fluids should be given for volume replacement and hemodynamic stabilization. An
appropriate number of units of whole blood or packed cells should be typed and crossmatched based
on the estimate of the volume of blood loss and the severity of bleeding. In some situations, it may be
necessary to use rapid methods for crossmatching, in consultation with blood bank personnel,
especially when there is a danger of exsanguination.
Appropriate consultations should be sought as the patient is being admitted rather than as urgent
measures when the patient's condition suddenly deteriorates. These consultations are based on the
overall assessment of the patient's status; however, as a general rule it is best that a gastroenterologist
and a surgeon be involved in most cases. The level of such consultation can range from active
participation by both members of this team in the management of gravely ill patients to a request that
consultants remain available for assistance should the need arise.
Laboratory tests appropriate at initial presentation include hematocrit, hemoglobin, and white blood cell
count. Coagulation studies, such as a bleeding time to assess platelet function, platelet count,
prothrombin time, and partial thromboplastin time, may also be of importance. Serum electrolyte values
and simple tests of renal function should be obtained. If there is any question of liver disease, basic
liver function tests should be included, because underlying hepatic disease is a significant risk factor in
the outcome of a bleeding episode.25,26(1130)
Diagnostic and Management Considerations After Initial Assessment and

Resuscitation
The diagnostic and management issues that arise after initial assessment and resuscitation include
determination of the level of the bleeding within the gastrointestinal tract, assessment of the severity of
bleeding, diagnosis of the cause of the bleeding, prognosis based on endoscopic findings, timing of
endoscopy, the effect of diagnostic and therapeutic endoscopy on outcome, and a rational approach to
endoscopy in the patient with upper gastrointestinal bleeding.
Determination of the Level of Bleeding

The level of bleeding within the gastrointestinal tract should be determined as soon as possible after
the patient's condition is stabilized. Although in many cases it is obvious that bleeding is upper
gastrointestinal in origin, the level of bleeding is uncertain in others. Immediate management may not
be altered by this information, but knowledge of the level of bleeding within the gut does have a
profound influence on the sequence and extent of the patient's subsequent evaluation. If the patient
reports hematemesis and can be considered a reliable observer, it may be assumed that the source of
bleeding is proximal to the ligament of Treitz. However, the patient may be obtunded, reliability may be
questionable, or in some cases, hematemesis may not have occurred. In such instances, it is essential
that a tube be passed into the stomach to sample the gastric contents for blood. A sump tube (18
French) is recommended.
In a retrospective series of 353 patients with gastrointestinal bleeding, Luk et al.27(1131) reported a
less than 1% chance of bleeding proximal to the ligament of Treitz if the gastric aspirate did not contain
blood. However, in the prospective national survey by the American Society for Gastrointestinal
Endoscopy (ASGE), which included data on 2225 patients who were evaluated for upper
gastrointestinal bleeding, an active source of bleeding was found by endoscopy in 15.9% of patients
whose gastric aspirates were clear of blood.28(1132) The different conclusions of these two studies
may be explained by the retrospective nature of the report of Luk et al.27(1133) and by the fact that
only 209 of 353 patients had endoscopy in this study.
The endoscopic diagnoses in the patients with clear nasogastric aspirates in the ASGE survey are
listed in Table 261. Duodenal ulcer was most common, although gastric and esophageal lesions are
represented. It can be concluded that a gastric aspirate positive for blood confirms that bleeding is
from the upper gastrointestinal tract but that a negative aspirate result does not eliminate this
possibility.
American Society for Gastrointestinal

Endoscopy Bleeding Survey: Endoscopic Findings in 214
Patients With Clear Nasogastric Aspirates
TABLE 261
FINDING
NUMBER OF PATIENTS
INCIDENCE (%)
Duodenaal ulcer
64
29.8
Gastric erosions
57
26.5
Gastric ulcer
47
21.9
Esophagitis
23
10.7
Duodenitis
21
9.8
Varices
11
5.1
Mallory-Weiss tear
10
4.7
Neoplasm
8
3.7
Stomal ulcer
7
3.3
Esophageal ulcer
2
0.9
Telangiectasia
0
Other
18
8.4
None
24
11.2
From Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE
survey on upper gastrointestinal bleeding. III. Endoscopy in upper
gastrointestinal bleeding. Gastrointest Endosc 1981; 27:94102.
Cuellar et al.29(1134) evaluated the sensitivity and specificity of the appearance of the gastric aspirate
as predictors of bleeding in 62 patients. They found that among 36 patients thought by clinicians to be
actively bleeding on the basis of the physical appearance of the aspirate, only 19 were actively
bleeding at endoscopy. Five of 26 patients thought not to have acute upper gastrointestinal bleeding
were found to have ongoing bleeding at endoscopy. Although there is some doubt concerning the
reliability of the information obtained from the gastric aspirate, its high negative predictive value for the
absence of active bleeding can provide valuable information in many cases.
There are several reasons for a clear gastric aspirate despite upper gastrointestinal bleeding. The
nasogastric tube may curl in the distal esophagus, preventing sampling of the stomach contents; the
stomach may have emptied itself of blood, and bleeding may have stopped; or the bleeding lesion may
be distal to the pylorus with no reflux of blood into the stomach, which may occur with lesions such as
an ulcer in the descending duodenum or an aortoduodenal fistula. On occasion, especially when the
clinical circumstances are highly problematic, endoscopy may be necessary to determine if bleeding is
emanating from an upper gastrointestinal source.
Assessment of the Severity of Bleeding

Patients with upper gastrointestinal bleeding are a heterogeneous group. They vary in general medical
status from relatively healthy to very ill with other concomitant diseases and systemic illnesses.
Similarly, there is great variation in the severity of blood loss. Patients may present in shock or in a
stable condition with a history of black stools several days earlier. The initial assessment of the activity
of bleeding is important for determining prognosis and immediate management. In the ASGE survey,
bleeding activity in patients with upper gastrointestinal hemorrhage, judged by color of gastric aspirate
(Table 262) or color of stool (Table 263), significantly correlated with outcome in terms of mortality,
the need for transfusion of more than five units of blood, the need for surgery, and the occurrence of a
complication during the bleeding episode.25(1135) Additional insight into prognosis can be gained by
combining information on the color of the nasogastric aspirate and stool (Table 264). Patients with
brown stools and clear nasogastric aspirates had a relatively low mortality rate, but those with red
nasogastric aspirates and red stools had a mortality rate approaching 30%.25(1136)
American Society for Gastrointestinal Endoscopy Bleeding Survey:

Nasogastric Aspirate Color and Patient Outcome
TABLE 262
NASOGASTRIC
ASPIRATE
COLOR
NUMBER OF
PATIENTS
MORTALITY
(%)
COMPLICATIONS
(%)
>5 UNITS
OF BLOOD
(%)
SURGERY
(%)
Clear
234
6.0
8.6
12.0
9.8
Coffee grounds
650
9.2
10.6
26.0
12.9
Red (blood)
734
17.9
18.9
41.3
22.9
From Silverstein FE, Gilbert DA, Tedesco FJ, et al. The national ASGE survey on upper gastrointestinal bleeding.
II. Clinical prognostic factors. Gastrointest Endosc 1981; 27:8093.
American Society for Gastrointestinal Endoscopy Bleeding

Survey: Stool Color and Patient Outcome
TABLE 263
STOOL
COLOR
NUMBER OF
PATIENTS
MORTALITY
(%)
COMPLICATIONS
(%)
>5 UNITS
OF BLOOD
(%)
SURGERY
(%)
Brown
465
11.0
9.9
16.3
12.3
Black
1312
8.1
10.8
24.8
13.5
Red
373
20.1
16.4
45.3
26.3
From Silverstein FE, Gilbert DA, Tedesco FJ, et al. The national ASGE survey on upper
gastrointestinal bleeding. II. Clinical prognostic factors. Gastrointest Endosc 1981; 27:8093.

Endoscopy Bleeding Survey: Nasogastric Aspirate and Stool
Color Versus Mortality
TABLE 264
NASOGASTRIC
ASPIRATE
COLOR
Clear
Coffee grounds
Red (blood)
STOOL
COLOR
Brown
Black
Red
Brown
Black
Red
Brown
Black
NUMBER OF
PATIENTS
MORTALITY
(%)
38
149
41
128
412
84
160
382
7.9
4.7
7.3
7.8
8.3
19.1
19.4
12.3

Endoscopy Bleeding Survey: Nasogastric Aspirate and Stool
Color Versus Mortality
TABLE 264
NASOGASTRIC
ASPIRATE
COLOR
STOOL
COLOR
NUMBER OF
PATIENTS
MORTALITY
(%)
Red
171
28.7
From Silverstein FE, Gilbert DA, Tedesco FJ, et al. The national ASGE survey
on upper gastrointestinal bleeding. II. Clinical prognostic factors. Gastrointest
Endosc 1981; 27:8093.
Gastric lavage may provide an additional assessment of the severity of bleeding. Before diagnostic
endoscopy, a large-bore (32 to 36 French) soft rubber tube is passed through the mouth to the
stomach to allow gentle irrigation to determine whether bleeding is active (and if so, to estimate rate of
bleeding), to clear the stomach of clots and fluid to facilitate endoscopy, and to reduce the possibility of
aspiration during endoscopy. The lavage tube should have extra holes cut near its tip to ensure
adequate drainage of the irrigation fluid. The patient is placed in a left decubitus position, with the head
in a slightly dependent, downward position. About 150 to 200 ml of irrigation fluid may be introduced at
a time. Whenever possible, fluid should be allowed to flow out of the stomach without applying suction
to the tube. This decreases the risk of traumatic injury to the gastric wall, with resultant submucosal
lesions that may be confused at endoscopy with gastric erosions or mucosal hemorrhages. The tube
should be allowed to drain passively by gravity into a container at the bedside. If the tube does not
drain, squeezing it with a milking motion often dislodges a clot from the intragastric tip and allows it to
drain properly. Changing the position of the tube by moving it in and out slightly frequently restores
proper drainage.
The purpose of lavage is not to stop bleeding. No controlled studies support the notion that lavage with
cold or warm saline or water, with or without vasoconstrictors, can stop acute gastrointestinal
hemorrhage. One study in an animal model of iced lavage solutions with and without vasopressin failed
to demonstrate efficacy for the control of acute upper gastrointestinal hemorrhage.30(1137) Aside from
the lack of therapeutic efficacy, many practitioners omit gastric lavage in patients requiring urgent
endoscopy with the possibility of therapeutic intervention in an effort to avoid creation of mucosal
artifacts before the endoscopic procedure. The efficacy for removing large clots, even with large-bore
tubes, has been questioned.5(1138)
Endoscopy can effectively establish the level or degree of gastrointestinal bleeding. If seen
endoscopically early in the course of the bleeding episode, bright red blood emanating from a lesion
establishes the presence of active, ongoing bleeding. This may be an essential parameter in selecting
patients for endoscopic hemostatic therapy.31(1139) In some situations, although bleeding appears by
clinical assessment to have stopped, an actively bleeding lesion, including a spurting artery, may be
found at endoscopy. This should alert the medical or surgical team to the serious nature of the patient's
condition. Endoscopy can also identify risk factors associated with an increased likelihood that bleeding
will recur. Techniques such as the use of technetium-labeled red blood cells with gamma camera
scanning to determine activity and location of bleeding32(1140) and visceral angiography33,34(1141)
are needed only rarely to assess the degree of upper gastrointestinal bleeding.
Diagnosis of the Cause of Bleeding

Several approaches can be used in determining the cause of gastrointestinal bleeding. Using clinical
acumen and the history of a patient's prior bleeding episodes, the clinician can venture a guess about
the cause of current bleeding. However, several studies suggest that the accuracy of clinical,
history-based diagnosis is as low as 40%.35,36(1142) In the ASGE survey, an ulcer was found to be
the cause of gastrointestinal bleeding in fewer than 50% of patients with a history of gastric or
duodenal ulcer; a similar poor correlation with a history of reflux esophagitis has been
observed.7(1143) In patients with portal hypertension and known varices who presented with upper
gastrointestinal bleeding, it has been reported that 30 to 50% were bleeding from sources other than
varices.3739(1144) The actual bleeding sources in patients with known varices that were not
bleeding, as reported in the ASGE survey, are listed in Table 265.

Endoscopy Bleeding Survey: Final Diagnoses for 109
Patients With Varices at Endoscopy That Were Not the
Cause of Bleeding
TABLE 265
DIAGNOSIS
NUMBER OF PATIENTS
INCIDENCE (%)
Gastric erosions
62
56.8
Mallory-Weiss tear
16
14.6
Gastric ulcer
15
13.8
Duodenal ulcer
15
13.8
Esophagitis
12
11.0
Duodenitis
7
6.4
Esophageal ulcer
3
2.7
Neoplasm
3
2.7
Stomal ulcer
1
0.9
Telangiectasia
1
0.9
Other
3
2.7
From Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE
survey on upper gastrointestinal bleeding. III. Endoscopy in upper
gastrointestinal bleeding. Gastrointest Endosc 1981; 27:94102.
Upper gastrointestinal barium x-ray studies have been useful in determining the cause of upper
gastrointestinal bleeding. Single contrast x-ray films have demonstrated lesions such as varices,
carcinoma of the stomach and esophagus, and deep gastric and duodenal ulcers.40(1145) The results
of some studies have even suggested that a vortex of blood can be seen spurting into a
gastrointestinal lumen that has been filled with barium, thereby making possible a diagnosis of active
bleeding.31(1146) This is, however, a rarely observed phenomenon. The results of other investigations
indicate that subtle mucosal lesions such as varioliform gastritis or erosive esophagitis can be
diagnosed with the enhanced resolution of air contrast x-ray films.41(1147)
Many studies have compared the diagnostic accuracy of endoscopy with that of upper gastrointestinal
barium contrast radiography in the diagnosis of gastrointestinal bleeding.4244(1148) Three
representative prospective studies are compared in Table 266. These demonstrate that accuracy with
respect to site of bleeding is approximately 25 to 50% with single contrast x-ray films compared with 70
to 95% for fiberoptic panendoscopy.
Endoscopy Versus Radiography in the

Diagnosis of Upper Gastrointestinal Bleeding
TABLE 266
diagnosed with the enhanced resolution of air contrast x-ray films.41(1147)

Many studies have compared the diagnostic accuracy of endoscopy with that of upper gastrointestinal
barium contrast radiography in the diagnosis of gastrointestinal bleeding.4244(1148) Three
representative prospective studies are compared in Table 266. These demonstrate that accuracy with
respect to site of bleeding is approximately 25 to 50% with single contrast x-ray films compared with 70
to 95% for fiberoptic panendoscopy.
Endoscopy Versus Radiography in the

Diagnosis of Upper Gastrointestinal Bleeding
TABLE 266
IDENTIFICATION OF SITE
STUDY
NUMBER OF
PATIENTS
By
Endoscopy
(%)
By
Radiography
(%)
Katon and Smith42

Morris et al. 43
90
93
30
54
69
22
McGinn et al. 44
134
88
53
Although the air contrast upper gastrointestinal technique is more sensitive than single contrast
studies, it is not as accurate as endoscopy.45(1149) Thoeni and Cello46(1150) compared the accuracy
of endoscopy and double-contrast radiography in a prospective study of 100 patients with significant
upper gastrointestinal bleeding. Endoscopy detected the primary site of bleeding in 93% of patients
while radiography detected the principle site in 80%. Endoscopy can detect superficial mucosal
abnormalities such as esophagitis, Dieulafoy's lesion, and Mallory-Weiss tears. Ulcers are identified
with greater accuracy by endoscopy than by x-ray studies.42(1151)
The use of upper gastrointestinal contrast radiography in the evaluation of the bleeding patient has
certain disadvantages. Barium obscures the gastrointestinal mucosa and makes subsequent
endoscopy impossible for at least 6 to 12 hr. Similarly, diagnostic and therapeutic visceral angiography
cannot be undertaken for at least several hours. It is generally accepted that endoscopy is superior to
roentgenography in determining the cause of bleeding. For the 5 to 10% of stable patients in whom
endoscopy fails to disclose the bleeding source, an upper gastrointestinal x-ray contrast study may be
helpful.
Visceral angiography is less available than endoscopy and does not provide a precise diagnosis of
bleeding lesions. Complications may occur from the arterial puncture and injection of a contrast agent.
Angiography is generally reserved for the patient with problematic recurrent bleeding in whom
angiographic therapy is a consideration. Technetium-labeled red blood cell studies are rarely helpful for
diagnosing the cause of upper gastrointestinal bleeding.47(1152)
The frequency of the various causes of upper gastrointestinal bleeding has been accurately
determined using endoscopy (Table 267).28(1153) Multiple lesions have been found in approximately
one third of patients.28,4850(1154) The 1978 through 1982 survey data on upper gastrointestinal
bleeding (44 centers in 21 countries) obtained by the Research Committee of the World Organization
of Gastroenterology showed peptic ulcer disease to be the most common cause of bleeding
determined by upper gastrointestinal endoscopy (36%) followed by esophagogastric varices (13%),
gastric erosions (6%), and multiple causes (22%), with the most common combination being peptic
ulcer and hiatus hernia.12(1155)
one third of patients.28,4850(1154) The 1978 through 1982 survey data on upper gastrointestinal
bleeding (44 centers in 21 countries) obtained by the Research Committee of the World Organization
of Gastroenterology showed peptic ulcer disease to be the most common cause of bleeding
determined by upper gastrointestinal endoscopy (36%) followed by esophagogastric varices (13%),
gastric erosions (6%), and multiple causes (22%), with the most common combination being peptic
ulcer and hiatus hernia.12(1155)
American Society of Gastrointestinal

Endoscopy Bleeding Survey: Endoscopic Diagnoses for Upper
Gastrointestinal Bleeding in 2097 Patients
TABLE 267
DIAGNOSIS
NUMBER OF PATIENTS
FREQUENCY (%)
Gastric erosions
620
29.6
Duodenal ulcer
477
22.8
Gastric ulcer
457
21.9
Varices
323
15.2
Esophagitis
269
12.8
Duodenitis
191
9.1
Mallory-Weiss tear
168
8.0
Neoplasm
78
3.7
Esophageal ulcer
46
2.2
Stomal ulcer
39
1.9
Telangiectasia
10
0.5
Other
152
7.3
From Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE survey
on upper gastrointestinal bleeding. III. Endoscopy in upper gastrointestinal
bleeding. Gastrointest Endosc 1981; 27:94102.
Gastric erosions represented the most frequent finding in patients with upper gastrointestinal bleeding
of various severities in the ASGE survey (see Table 267).28(1156) Later studies indicate that severe
or life-threatening upper gastrointestinal bleeding most often results from varices, gastric or duodenal
ulcer, or Dieulafoy's lesion.5,51(1157)
Before the widespread use of endoscopy for diagnosis of gastrointestinal bleeding, the Mallory-Weiss
syndrome was regarded as a rare occurrence, usually found in alcoholic patients who had protracted
or severe episodes of vomiting. Endoscopy has shown that this condition is relatively common and may
develop in a wide variety of clinical circumstances. A Mallory-Weiss tear was found by endoscopy in
7.0% of the patients in the ASGE survey (see Table 267).28(1158) Sugawa et al.52(1159) identified
this lesion as the cause of bleeding in 10.3% of patients (n = 2175) who underwent endoscopy for
upper gastrointestinal bleeding; Clain et al.53(1160) observed the lesion in 7.8% of such patients (n =
1667). Paquet et al.54(1161) determined that a Mallory-Weiss tear can account for bleeding in patients
with esophageal varices; in some, it is the first manifestation of bleeding. The Mallory-Weiss syndrome
also occurred in patients who had undergone prior sclerotherapy. The severity of bleeding can range
from mild to severe, although it is self-limited in most cases (Figure 261).52,53,55(1162) Additional
lesions may be found at endoscopy in some patients who have Mallory-Weiss tears.
(1163)Figure 261. A, Endoscopic view of a Mallory-Weiss tear with active bleeding (gastric
lumen is at top left). B, Endoscopic view of an organized clot adherent to a Mallory-Weiss tear
(gastric lumen is at bottom left ). (A and B, From the collection of Dr. M. V. Sivak, Jr.)
Aortoenteric fistula is an uncommon but exceedingly important cause of upper gastrointestinal bleeding
because the mortality associated with this condition is virtually 100% without treatment. There have
been numerous case reports in which endoscopy played a prominent role in the diagnosis.5659(1164)
Because of the mortality associated with this condition, upper gastrointestinal endoscopy should be
performed as soon as possible in any patient with upper gastrointestinal bleeding and a history of
abdominal aortic surgery. Partial visualization of the aortic graft through the eroded wall of the third part
of the duodenum is the classic endoscopic finding of an aortoenteric fistula (Figure 262). Other
endoscopic findings are possible; in some cases, the only finding may be a clot adherent to the
duodenal wall. A fistula may form to segments of the gastrointestinal tract other than the duodenum.
Endoscopy and angiography may fail to identify an aortoenteric fistula, in which case computed
tomography may be helpful.60(1165) Aortoenteric fistula is considered in greater detail in Chapter 49:
Diseases of the Duodenum.
(1166)Figure 262. Endoscopic view of abdominal aortic graft that has eroded into the third
part of the duodenum. (From the collection of Dr. M. V. Sivak, Jr.)
Dieulafoy's lesion (i.e., exulceratio simplex) is an uncommon but potentially dangerous cause of upper
gastrointestinal bleeding. Because the lesion is small, endoscopic diagnosis may be difficult, especially
if bleeding is severe. Dieulafoy's lesion is most often located in the proximal stomach and appears as a
small, red, round lesion that may be slightly elevated above the mucosal surface (Figure 263). If there
is a fibrin plug in the vessel, it may look white rather than red. In more than 80% of cases, the lesion is
within 6 cm of the gastroesophaeal junction, frequently on the lesser curvature of the stomach.
(1167)Figure 263. Endoscopic view of a Dieulafoy lesion on the lesser curvature of the
stomach. (From the collection of Dr. M. V. Sivak, Jr.)
The pathogenesis of Dieulafoy lesion is poorly understood, but the fundamental element is thought to
be the presence of an artery of unusually large diameter (up to 3 mm) in the submucosa of the
stomach.61(1168) Although the vessel appears histologically normal, it has an unusually large
diameter in relation to its location, and it usually runs a tortuous course through the submucosa.
Evidence of associated inflammation, vasculitis, or aneurysmal dilation is usually lacking. The factors
that lead to rupture of the vessel are largely unknown.
Hemorrhage from a Dieulafoy lesion is usually severe and frequently recurs within a few days so that
the clinical course is relatively short. Based on a review of 101 cases, Veldhuyzen Van Zanten et
al.62(1169) estimated the mortality rate associated with this lesion at 25%. Baettig et al.63(1170) found
that a Dieulafoy lesion accounted for 5.8% of episodes of gastrointestinal bleeding in 480 patients who
underwent endoscopy. Although numbers of reported cases are small, it appears that endoscopic
therapy may be effective.63,64(1171)
Overt upper gastrointestinal bleeding may occasionally be caused by angiodysplasia of the stomach or
duodenum (Figure 264). Quintero et al.65(1172) found that these lesions accounted for 2.1% of
hospital admissions for upper gastrointestinal hemorrhage and 3.7% of admissions for severe bleeding
(defined as transfusion of two or more units of blood). In the study of Clouse et al.,66(1173)
angiodysplasia of the stomach or duodenum was encountered in 30 (4%) of 676 patients who
underwent endoscopy for upper gastrointestinal bleeding; multiple lesions were found in 63%.
Hematemesis or melena occurred in 23 patients (77%). Unless angiodysplasia is suspected, the
diagnosis can be missed at endoscopy, especially if the mucosal surface is obscured by fresh or old
blood or clots. In the study of Quintero et al.,65(1174) endoscopy at the first episode of bleeding was
nondiagnostic in all patients who did not have hereditary hemorrhagic telangiectasia (i.e.,
Osler-Weber-Rendu disease). For the latter condition, endoscopy was diagnostic in all cases,
undoubtedly because of a higher index of suspicion by the clinician. Angiodysplasia accounted for a
large percentage of cases labeled as "bleeding of unknown etiology" after initial evaluation. At
endoscopy, the telangiectasias of hereditary hemorrhagic telangiectasia are indistinguishable from the
nonhereditary form. Gastric antral vascular ectasia, the so-called watermelon stomach, usually results
in chronic rather than acute blood loss, and patients typically present with iron deficiency anemia
(Figure 265).
(1175)Figure 264. Endoscopic view of a vascular ectasia (angiodysplasia) in the duodenum.

(From the collection of Dr. M. V. Sivak, Jr.)
(1176)Figure 265. Endoscopic view of the gastric antrum with watermelon stomach. The
pylorus is at top center. Note the linear distribution pattern of the vascular lesions arranged
radially around the pylorus. (From the collection of Dr. M. V. Sivak, Jr.)
Gastrointestinal bleeding that develops in patients with malignant tumors, including tumors of the
gastrointestinal tract, can be problematic in several respects. Coagulopathies including
thrombocytopenia may be caused by treatment with chemotherapeutic agents or may result from
metastatic disease. Although a diffuse mucosal ooze of a blood is sometimes suspected in patients
with gastrointestinal bleeding and thrombocytopenia, Chu et al.67(1177) found evidence of this at
endoscopy in only 1% of patients. A single source of bleeding was found in most patients; it was
usually a relatively common type such as peptic ulcer disease. In the study of Shivshanker et
al.,68(1178) most patients with cancer were found to be bleeding from one of the more common
benign lesions such as peptic ulcer disease; approximately one third of patients with gastrointestinal
cancers were bleeding from another source.
Patients with chronic renal failure, including those undergoing chronic hemodialysis, may have a
different pattern of upper gastrointestinal bleeding from those with normal renal function. Zuckerman et
al.69(1179) identified 59 patients with chronic renal failure and gastrointestinal bleeding in a group of
482 patients who underwent upper gastrointestinal endoscopy for bleeding over a 42-month period.
The most common cause of bleeding in these patients was angiodysplasia of the stomach or
duodenum. Erosive esophagitis was also significantly more common in patients with renal failure.
Bleeding was more likely to be recurrent in these patients (25% vs. 11% in patients with normal renal
function), and angiodysplasia was the most frequent cause.
Unusual causes of upper gastrointestinal bleeding include invasion of the duodenal wall by malignant
tumors of surrounding organs.70(1180) Most of these are pancreatic tumors,71(1181) but invasion by
malignant tumors from other organs may be responsible for gastrointestinal bleeding.72(1182) Rarely,
gastrointestinal bleeding may be the first mani-festation of pancreatic carcinoma.73(1183) Neurogenic
tumors, including those associated with cutaneous neurofibromatosis (i.e., von Recklinghausen's
disease),74,75(1184) can be the source of acute upper gastrointestinal bleeding. Acute hemorrhage
(i.e., hemosuccus pancreaticus) rarely may arise from within the pancreatic duct.7678(1185) Bleeding
from the papilla of Vater due to passage of a gallstone has been observed.79(1186) Although unusual,
there are several reports of bleeding from duodenal varices.8082(1187) One patient with amyloidosis
presented with upper gastrointestinal bleeding.83(1188)
Prognostic Value of Endoscopy

Certain endoscopic findings, usually called stigmata of recent hemorrhage, predict a poor clinical
outcome for patients with gastrointestinal bleeding. These findings may therefore identify patients who
can benefit from early and, in some cases, more aggressive therapeutic intervention. Endoscopic
findings that predict nonvariceal hemorrhage are considered; those that predict variceal bleeding are
discussed in Chapter 31: Variceal Bleeding.
The definition of stigmata of recent hemorrhage is inexact, but the term usually encompasses active
bleeding or oozing of blood from an ulcer, adherent fresh or altered clot within a crater, or a "visible
vessel" within an ulcer crater (Figure 266). The term visible vessel refers to an elevated, red, purplish,
yellow-white, or black spot within an ulcer crater (Figure 266A). A nonelevated, red or black spot
within an ulcer crater is not a visible vessel (Figure 266B and Figure 266C). In most cases, the
visible vessel is probably a clot protruding from an arterial vessel. Stigmata of hemorrhage generally
indicate an increased risk of further bleeding that is likely to be severe or protracted and to require
surgical treatment.
(1189)Figure 266. Endoscopic views of ulcers with stigmata of recent hemorrhage. A,

Duodenal ulcer with a visible vessel. B, Gastric ulcer with a red spot in the center of the crater.
C, Duodenal ulcer with a red spot in the center of the crater. D, Purplish clot adherent to a
gastric ulcer. (A-D, From the collection of Dr. M. V. Sivak, Jr.)
Foster et al.84(1190) observed stigmata of recent hemorrhage in 47% of 233 patients with presenting
symptoms of hematemesis or melena. For 89 patients with peptic ulcer, they found that those whose
ulcers had stigmata had a significant increase in recurrent bleeding and required surgery significantly
more often than patients without these findings. These investigators concluded that the endoscopic
stigmata of recent hemorrhage were better predictors of a less favorable outcome than was age,
shock, or serious underlying disease.
Fresh bleeding from an ulcer, clot adherent to an ulcer (see Figure 266D), and the visible vessel were
all considered stigmata of recent hemorrhage in the study of Harris and Heap.85(1191) Stigmata were
present in 52 of 69 patients with duodenal ulcer. Among those with stigmata, bleeding was recurrent in
23%, and 29% required operation. The mean transfusion requirement was 7.2 units. Among the 17
patients with duodenal ulcer and no stigmata, bleeding was recurrent in 12%, the mean transfusion
requirement was 3.2 units, and none required operation. These investigators observed stigmata of
recent hemorrhage in 44 of 57 patients with gastric ulcer. Among patients with stigmata, bleeding was
recurrent in 31%, surgery was performed in 25%, and the mean transfusion requirement was 6.6 units.
Among the 13 patients with gastric ulcers and no stigmata, none had recurrent bleeding or required
surgery, and the mean transfusion requirement was 1.9 units.
Of the several stigmata of recent hemorrhage, a visible vessel seen during endoscopy in the base or
rim of an ulcer is probably the most significant (see Figure 266A). Although some
investigators86(1192) have found it to be unreliable as a predictor of recurrent bleeding, most studies
emphasize its value as an indicator of further hemorrhage.85,8789(1193)
In a retrospective study by Griffiths et al.,87(1194) 28 (18.5%) of 157 ulcer patients had a visible vessel
associated with an ulcer. These 28 patients subsequently experienced uncontrollable or recurrent
bleeding that required surgery. Among the other 129 ulcer patients without visible vessels, only 26%
had uncontrollable or recurrent hemorrhage, a statistically significant difference. In the several
subsequent, prospective trials of endoscopic hemostatic therapy, the frequency of a visible vessel in an
ulcer base has ranged from 5.6 to 57.8%.10,18,88,89,9095(1195)
Peptic ulcer was found to be the cause of bleeding at endoscopy in 402 (49%) of 826 consecutive
patients with acute upper gastrointestinal bleeding in the study of Swain et al.96(1196) In 82% of
cases, the ulcer crater could be fully examined, and a visible vessel was identified in 156 (47%) of
these. Other stigmata of recent hemorrhage were observed in 66 patients; none were found in 107
patients. One hundred twenty-nine patients with stigmata were observed for further bleeding. Of the 93
patients with a visible vessel, 58% had recurrent bleeding; only 6% of those with other stigmata had
further bleeding. In patients with no stigmata, there was no recurrent bleeding.
Bornman et al.89(1197) prospectively studied the incidence of recurrent bleeding in 177 patients with
acute gastrointestinal hemorrhage from peptic ulceration. Bleeding was recurrent in none of 40
patients with only a black spot in the ulcer crater, 11 (23%) of 48 patients with a clot in the crater, and 5
(50%) of 10 patients with a visible vessel. However, bleeding was recurrent in only 4 of 79 patients with
no stigmata of recent bleeding.
Peptic ulcer was diagnosed in 132 patients with upper gastrointestinal bleeding who underwent
endoscopy within 24 hr of admission in the prospective study of Storey et al.88(1198) The ulcer crater
was fully visualized in 117 patients, a visible vessel was identified in 56 patients (48%), other stigmata
of recent bleeding were noticed in 21 patients (18%), and no stigmata were present in 40 patients
(34%). Forty-seven patients were randomly selected for observation without treatment for evidence of
further bleeding, which was observed in 19 (56%) of the 34 patients with a visible vessel in an ulcer
crater. Only 1 (8%) of 13 patients with other stigmata had recurrent bleeding, and none of those
without stigmata had further bleeding.
The degree of active bleeding at endoscopy correlates with outcome. In the ASGE survey, patients in
whom blood was found by endoscopy to be oozing or spurting from a lesion had a more than twofold
increase in mortality and need for surgery compared with patients who did not have these
findings.7(1199) Active bleeding at endoscopy was associated with a transfusion requirement of more
than five units of blood. These correlations with poor outcome were especially strong for patients with
actively bleeding duodenal ulcers.
The study of Bornman et al.89(1200) of endoscopic stigmata of hemorrhage also compared the degree
of hemorrhage with outcome. In this prospective study of 177 patients with bleeding peptic ulcers,
bleeding reoccurred in 2 (2%) of 114 patients without shock, 7 of 38 patients with tachycardia in whom
the systolic blood pressure remained above 100 mm Hg, and in 12 (48%) of 25 patients with shock
(i.e., a systolic blood pressure of less than 100 mm Hg). The incidence of recurrent bleeding in shock
patients was significantly higher in those with a clot or visible vessel in the ulcer crater. Bornman et
al.89(1201) emphasized that the combination of shock and stigmata of recent hemorrhage appears to
be a stronger predictor of recurrent bleeding than either of these parameters alone.
The frequency with which a visible vessel is found at endoscopy may depend on differences in the
endoscopic practices of various investigators. A visible vessel may be more apparent when an ulcer
crater is washed with a jet of water delivered through a catheter passed through the endoscope.
Washing may remove blood, mucus, and other debris that may otherwise obscure the presence of the
vessel. In practice, most endoscopists do not routinely attempt vigorous removal of clots from the base
of every ulcer crater. However, this may be done in a patient with severe or recurrent bleeding when a
decision has been made to treat a bleeding vessel if one can be found. Timing of endoscopy can also
influence whether a visible vessel is identified. Chung et al.97(1202) observed complete
disappearance of ulcer vessels in 62 patients who underwent endoscopy every day for 3 consecutive
days.
The incidence of recurrent bleeding in patients with visible vessels in their ulcers approaches 50% in
studies of upper gastrointestinal bleeding due to peptic ulcer. Other identified stigmata of recent
hemorrhage, such as flat red or black spots in an ulcer base and adherent clot, are associated with
lesser risks of recurrent bleeding. A representative spectrum of the risk of recurrent bleeding
associated with different endoscopic ulcer appearances is illustrated in Figure 267. These data from
Laine et al.98(1203) represent prospective observations over a 6-year period of patients presenting
with ulcer bleeding and various stigmata identified during endoscopy.
(1204)Figure 267. Outcome in patients with gastrointestinal bleeding in relation to

endoscopic findings. (From Laine L, Cohen H, Brodhead J, et al. Prospective evaluation of
immediate versus delayed refeeding and prognostic value of endoscopy in patients with upper
gastrointestinal hemorrhage. Gastroenterology 1992; 102:3146.)
Freeman et al.99(1205) reported a more detailed characterization of the changes in ulcer bases and
their predictive value for recurrent bleeding. Lesions were categorized as a pale, mound-like lesion that
appeared to push through the ulcer floor (Type I-A), a similar lesion capped with a small clot (Type I-B),
or a discrete pigmented plug that appears attached to the ulcer crater and with a distinct color
transition (Type II). The investigators studied 23 patients with ulcer bleeding and found Type I-A lesions
in 6, Type I-B lesions in 4, and Type II lesions in 13. Bleeding reoccurred in 9 of 10 patients with Type
I-A or I-B but in only 1 of 13 patients with Type II lesions. Further studies are needed to confirm the
predictive value of these observations.
It appears that the presence of active bleeding at endoscopy and the observation of a visible vessel in
an ulcer base are the most important predictors of recurrent bleeding. Adherent clots, flat spots of
various colors, and clean, white bases are less frequently associated with recurrence of bleeding.
Timing of Endoscopy
Several factors influence the timing of endoscopy after it has been decided that the procedure is
required. Optimally, endoscopy should be performed in a hemodynamically stable patient. Expert
assistance from supporting medical and technical personnel is mandatory.
The goal of endoscopy in the diagnosis of upper gastrointestinal bleeding is the identification of the
most likely source of hemorrhage. Active hemorrhage found by endoscopy, the presence of a clot, or a
visible vessel in a lesion have been interpreted as evidence that the lesion in question is the actual
source of bleeding. Several investigators have evaluated timing in relation to yield of endoscopy for
signs of active or recent bleeding. In a group of 109 patients with hematemesis or melena, Forrest et
al.100(1206) found a significantly higher number of actively bleeding lesions by endoscopy when the
procedure was performed within the first 24 hr after admission compared with endoscopy at 24 to 48
hr.
Leinicke et al.101(1207) compared the results of endoscopy in 276 bleeding patients in whom most
endoscopies were performed within 12 hr of admission or at the onset of bleeding that arose in hospital
with the results of endoscopy in 103 patients for whom endoscopy was delayed for 24 hr. Active
bleeding or a clot on the lesion was seen in 62% of the group undergoing early endoscopy compared
with 14% of the late endoscopy group. Foster et al.84(1208) found a similarly high incidence of signs of
recent hemorrhage or active bleeding when endoscopy was performed within 12 hr of initial
hematemesis or melena. Among patients who underwent endoscopy within 12 hr, 34 (69%) of 49 had
these endoscopic signs compared with 75 (24%) of 181 who had endoscopy after 12 hr.
In the ASGE survey, the frequency of finding a lesion at endoscopy from which blood is oozing or
spurting decreased significantly as the interval between admission and endoscopy increased from 24
to 96 hr (Table 268).28(1209) As in other reports, endoscopy during the first 12 hr after admission
yielded the highest frequency of actively bleeding lesions. However, the yield was not significantly
higher if endoscopy was performed during the first 3 to 6 hr after admission.
bleeding or a clot on the lesion was seen in 62% of the group undergoing early endoscopy compared
with 14% of the late endoscopy group. Foster et al.84(1208) found a similarly high incidence of signs of
recent hemorrhage or active bleeding when endoscopy was performed within 12 hr of initial
hematemesis or melena. Among patients who underwent endoscopy within 12 hr, 34 (69%) of 49 had
these endoscopic signs compared with 75 (24%) of 181 who had endoscopy after 12 hr.
In the ASGE survey, the frequency of finding a lesion at endoscopy from which blood is oozing or
spurting decreased significantly as the interval between admission and endoscopy increased from 24
to 96 hr (Table 268).28(1209) As in other reports, endoscopy during the first 12 hr after admission
yielded the highest frequency of actively bleeding lesions. However, the yield was not significantly
higher if endoscopy was performed during the first 3 to 6 hr after admission.
American Society of Gastrointestinal

Endoscopy Bleeding Survey: Influence of Interval From
Admission to Endoscopy on Detecting Active Bleeding
TABLE 268
ENDOSCOPICALLY DISCOVERED
ACTIVE BLEEDING
INTERVAL
(HR)
NUMBER OF
PATIENTS
Patients
(n)
Incidence
(%)
012
663
275
41.5
1324
340
100
29.4
2536
115
37
32.3
3748
135
30
22.2
4960
36
9
25.0
6172
68
7
10.3
7384
31
5
16.1
8596
29
5
17.2
>96
180
36
20.0
From Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE survey
on upper gastrointestinal bleeding. III. Endoscopy in upper gastrointestinal
bleeding. Gastrointest Endosc 1981; 27:94102.
From these data, it may be concluded that the best chance to observe signs of active or recent upper
gastrointestinal bleeding by endoscopy occurs within the first 12 hr after admission or onset of bleeding
in hospital. These signs may be especially useful in ascribing the cause of bleeding to a particular
lesion when other lesions without such stigmata are found by endoscopy. These stigmata may change
over time and may be completely absent 3 or 4 days after the initial bleeding episode.97(1210)
From a practical standpoint, the data support the use of immediate endoscopy for patients who are
admitted and stabilized during daytime and early evening hours. Conversely, examination of patients
during nighttime hours may be deferred until early the next morning. This policy ensures the availability
of appropriate medical and technical personnel during the procedure. However, endoscopy during off
hours (i.e., nighttime and weekends) cannot be deferred if important and urgent management
decisions will be based upon endoscopic findings. This is the situation for patients with torrential
hemorrhage, continued bleeding, or recurrent bleeding.
Effect of Diagnostic and Therapeutic Endoscopy on Patient Outcome

The issue of whether diagnostic endoscopy affects the outcome in patients with gastrointestinal
bleeding has been debated at considerable length since the "vigorous diagnostic approach" to this
problem was first proposed.102,103(1211) The results of studies on this question are conflicting, but
most demonstrate little or no benefit from diagnostic endoscopy alone in terms of a decrease in
mortality, blood transfusions, or duration of hospitalization.43,48,104110(1212)
Hoare111(1213) reported the results of a representative study supporting the thesis that endoscopy
may benefit patients. In this study, 156 patients with acute upper gastrointestinal bleeding were divided
into two groups. The endoscopy group of 51 was managed by a physician whose policy was to perform
endoscopy within 48 hr of admission. The 105 control patients had an upper gastrointestinal contrast
radiography within this same period after admission. The mortality rate was 5.7% for the endoscopy
group and 15.2% for the x-ray study group. Surgery, when indicated, was performed in both groups by
the same surgeons and with the same frequency. However, surgery was performed significantly earlier
and mortality from surgery was less for the patients who underwent endoscopy.
One study frequently cited as evidence against the beneficial effect of early endoscopy on patient
outcome is that of Peterson et al.112(1214) In this prospective study, 206 bleeding patients who were
stabilized within 6 hr of admission were randomly assigned to a routine endoscopy or nonendoscopy
group. All patients were treated with antacids. No significant differences were observed in hospital
deaths, recurrence of bleeding, transfusion requirements for recurrent bleeding, or duration of hospital
stay. Long-term follow-up revealed 1 gastric lymphoma in the nonendoscopy group that was
overlooked during the initial evaluation.
Most of the available studies and data on the benefits of diagnostic endoscopy and upper
gastrointestinal bleeding are limited. Many are retrospective or have included groups of patients
managed by different physicians; in some studies, patient entry is not consecutive; in others, cohorts
managed several years apart are compared. In some series, patients in both groups underwent
endoscopy and x-ray studies; in others, there is a high proportion of incomplete endoscopies. Often, as
in the study by Peterson et al.,112(1215) patients were not stratified according to underlying illnesses
or causes of bleeding, and therapy was not modified by the information gained at endoscopy.
Some of the difficulty in substantiating the efficacy of endoscopy for diagnosing the cause of upper
gastrointestinal bleeding is explained by the fact that such bleeding stops spontaneously in most
patients. If the mortality rate from upper gastrointestinal bleeding is assumed to be 10%, it would be
necessary to enroll about 6000 patients in a randomized trial to demonstrate a 1% reduction in
mortality.113(1216) According to Erickson and Glick,113(1217) only 1948 patients had been enrolled in
all prospective trials of diagnostic endoscopy up to 1986.
Another major factor that confounds analysis of the earlier data pertaining to endoscopy and upper
gastrointestinal bleeding is therapy. Few medical treatments of the various causes of upper
gastrointestinal bleeding are of proven effectiveness.1,114(1218) If the accurate information obtained
at endoscopy cannot be used to direct specific effective therapy, an improved outcome based on
accurate diagnosis alone is unlikely. In the study by Peterson et al.,112(1219) all patients were treated
with an antacid regimen whether they were assigned to endoscopy or control groups. It is not
surprising that the outcomes for the two groups were similar.
The advent of endoscopic hemostatic therapy has altered the influence of endoscopy in the diagnosis
of patients with upper gastrointestinal bleeding. The endoscopist can take advantage of the accurate
diagnostic information provided by endoscopy to make decisions about the application of any of
several endoscopic techniques for control of bleeding. There is growing evidence that these
endoscopic methods for achieving hemostasis are effective and that they improve patient outcome,
including favorable effects on morbidity and mortality.7,115(1220) Endoscopic techniques for obtaining
hemostasis are discussed in detail in following chapters (see Chapter 28: Thermal Contact Methods for
Endoscopic Hemostasis; Chapter 29: Laser Therapy; and Chapter 30: Injection Therapy for Ulcer
Bleeding).
A Rational Approach to Endoscopy in Assessing Upper Gastrointestinal

Bleeding
In view of the documented diagnostic accuracy and potential therapeutic benefit of upper
gastrointestinal endoscopy, most gastroenterologists, endoscopists, and surgeons think that
endoscopy should be performed in the most patients with bleeding from the proximal gastrointestinal
tract. Endoscopy usually provides an accurate diagnosis. Its use establishes the level and site of the
bleeding, assesses the degree of active bleeding, and is of predictive value in terms of risk for further
bleeding after hemorrhage has stopped. This information may be critical for immediate or long-term
management decisions. Portosystemic shunt surgery, for example, would be disastrous in a patient
with portal hypertension if the patient was bleeding from a duodenal ulcer. Referral of a patient with
variceal bleeding for ulcer surgery is equally unacceptable. Endoscopy can affect the decision to
operate, the timing of surgery, and the type of surgery to be performed.116(1221) For a patient who
has bled from a peptic ulcer, an accurate endoscopic diagnosis can affect plans for short-term or
long-term medical therapy or surgery if recurrent bleeding becomes problematic. Endoscopic therapy
may be highly beneficial in certain clinical circumstances.
Although most patients with upper gastrointestinal bleeding should undergo endoscopy,45(1222) it may
be appropriate to omit endoscopy under certain circumstances. Endoscopy may not be necessary, for
example, in the postoperative patient with a minuscule amount of "coffee-ground" material in the
nasogastric aspirate or in the young individual with minimal hematemesis during an acute viral illness.
Our approach may be summarized as follows. When a patient is first examined, an orogastric tube
usually is passed into the stomach during or after resuscitation to determine if there is ongoing
bleeding. If the bleeding has stopped, the patient is stabilized, appropriate laboratory tests are
performed, and blood is replaced to raise the hematocrit to the 25 to 30% range. Appropriate
consultations are obtained. Endoscopy is performed 6 to 12 hr after admission. If bleeding does not
stop, as manifested by continued red drainage from the orogastric tube during gentle lavage, unstable
vital signs, or persistent bloody stools, immediate endoscopy is performed as soon as the patient is
reasonably stable. At this point, a diagnosis is usually confirmed and decisions concerning therapy can
be made. Another situation in which immediate endoscopy may be performed arises when a patient
has stabilized but then has a recurrence of significant bleeding. When an aortoduodenal fistula is
suspected, the patient is also a candidate for urgent endoscopy.
The patient with torrential hemorrhage presents a special situation. Immediate exsanguination is a
direct threat. Bleeding of this magnitude may occur in hemorrhage from esophageal varices, a
posterior duodenal ulcer that has eroded the gastroduodenal artery, a gastric ulcer (lesser curvature)
that has eroded the left gastric artery, or an aortoduodenal fistula. The correct therapy often is
immediate operative intervention. Endoscopy before operation may delay surgery and increase the risk
to the patient.116(1223) With this type of excessive bleeding, it may be impossible to perform
endoscopy adequately for diagnosis of the bleeding lesion. For these reasons, immediate surgical
consultation is essential.
The patient may then be taken to the operating room, and while the patient is under general anesthesia
with an endotracheal tube in place, upper gastrointestinal lavage and endoscopy may be performed.
This can be done as preparations are being made for surgery. With this approach, all the essential
elements for diagnosis and management are brought together quickly, and the surgeon and
endoscopist can plan therapy based on the endoscopic observations.
Preparation for Endoscopy

Endoscopy in the patient with upper gastrointestinal bleeding is one of the most difficult endoscopic
examinations, and it should be performed by an experienced endoscopist. Several aspects of the
procedure merit further discussion.

Premedication
The frequency with which various drugs are used as endoscopic premedication in patients with upper
gastrointestinal bleeding, as reported in the ASGE survey, is shown in Table 269.28(1224) Although a
short-acting benzodiazepine such as diazepam or midazolam is commonly administered, it may be
hazardous to use heavy sedation in the unstable, bleeding patient.39(1225) Meperidine may cause a
fall in blood pressure that can confuse clinical interpretation of vital signs and estimation of
intravascular volume. Patients with underlying liver disease may tolerate sedative drugs poorly.
However, in some patients, a small dose can make the procedure easier for the pa-tient and the
endoscopist. Benzodiazepine-reversal agents (e.g., flumazenil) and opiate antagonists (e.g., naloxone)
are important adjuncts that should be available for use in patients who receive endoscopic
premedication.
American Society of Gastrointestinal Endoscopy

Bleeding Survey: Endoscopic Premedication for 2097 Patients
TABLE 269
DRUG
NUMBER OF
PATIENTS (%)
MEAN DOSE* (mg)
RANGE (mg)
Diazepam
1474 (70.2)
10 6.9
0.590
Meperidine
821 (39.2)
57 22.0
5150
Topical anesthetic
1186 (56.6)
Atropine
317 (15.1)
0.6 0.4
0.020.6
Glucagon
40 (1.9)
1.0 0.2
0.21.0
Other
171 (8.2)
From Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE survey on
upper gastrointestinal bleeding. III. Endoscopy in upper gastrointestinal bleeding.
Gastrointest Endosc 1981; 27:94102.
* Values are expressed as the mean 1 standard deviation.
Most patients in the ASGE survey received a topical pharyngeal anesthetic before endoscopy (see
Table 269).28(1226) However, the use of this type of premedication in bleeding patients is debatable.
Although they may help the patient tolerate the procedure, topical anesthetic agents have
cardiovascular and central nervous system effects and may increase the risk of aspiration during
endoscopy.117(1227) This is especially true if sedation has also been administered. For most patients,
our practice is to use a topical pharyngeal anesthetic and small intravenous doses of a benzodiazepine
before the procedure.
The role of monitoring of patients receiving conscious sedation for gastrointestinal endoscopy has
been reviewed.118(1228) During endoscopy in patients presenting with upper gastrointestinal bleeding,
especially those who have been hemodynamically unstable, maintenance of good intravenous access
and meticulous attention to oropharyngeal suction is mandatory. Monitoring of cardiac rate and rhythm
by continuous electrocardiography, blood pressure (by manual or automated methods), and oxygen
saturation by pulse oximetry is appropriate for bleeding patients. These measures are particularly
pertinent to patients undergoing more prolonged therapeutic endoscopic procedures. In some patients
whose mental status may impair their protective airway reflexes, endotracheal intubation before upper
endoscopy may be advisable to prevent aspiration pneumonia.
Lipper et al.119(1229) studied the occurrence of pulmonary aspiration during emergency endoscopy in
patients with upper gastrointestinal bleeding. In 6 (20%) of the 30 patients in the study, new pulmonary
infiltrates were evident on chest radiographs obtained within 4 hr of endoscopy; they were absent on
chest radiographs obtained within 12 hr before endoscopy. Significant oxygen desaturation was
observed in 5 of these 6 patients during endoscopy; these 5 patients also developed fever,
leukocytosis, or both conditions.
Endoscope Components and Accessories
The endoscope selected for use in the bleeding patient must function properly. Excellent light and
visual resolution are essential, as is mechanical capability for such maneuvers as retroflexion. The air
and water insufflation functions of the instrument should be checked, including the capability to clean
the lens. Leaks in the air insufflation system can produce a frustrating and potentially hazardous
situation for the patient if it is impossible to adequately inflate the esophagus and stomach for
examination. Similarly, suction is essential to remove gas and secretions. The accessory channel must
be patent so that endoscopic therapeutic devices can be passed. All endoscope functions should be
carefully checked before the procedure.
There is a trend in endoscopy toward use of instruments with the smallest possible diameter for all
types of procedures. These instruments are easy to pass and highly maneuverable. However, for
upper gastrointestinal bleeding, larger-diameter instruments may be preferable, because they have
larger accessory channels. Increased channel diameter is necessary for passage of certain therapeutic
devices and for the removal of blood and clots. Double-channel endoscopes are preferred by some
endoscopists for these reasons. These instruments typically have an insertion tube diameter of
approximately 1.2 cm with accessory (suction) channels as large as 3.7 mm in diameter. Additional
advantages are excellent visual resolution and the incorporation of an excellent light delivery bundle.
These factors enhance the quality of diagnostic and therapeutic endoscopy in the bleeding patient.
The advent of video endoscopy offers some additional technical advantages in the management of
patients with upper gastrointestinal bleeding. These include improved operator safety, because the
endoscope is held at arm's length rather than near the endoscopist's face, thereby reducing exposure
to blood and secretions. Physicians and medical personnel other than the endoscopist can view the
endoscopic images during the procedure, enhancing their ability to provide assistance with various
diagnostic and therapeutic maneuvers.
Complications of Upper Gastrointestinal Endoscopy
The risks of upper gastrointestinal endoscopy in bleeding patients are higher than those encountered
in routine diagnostic endoscopy. Comprehensive data are contained in a prospective ASGE
survey.120(1230) In this study, 274 physicians reported complications from 2320 endoscopies in
bleeding patients (Table 2610). Endoscopic complications occurred in 21 patients, or 0.9% of all
examinations. In 3 patients, 3 complications occurred during repeat examinations. Twelve
complications were considered major; 9 of these occurred during emergency endoscopy. The major
complications were perforation in 5 patients, aspiration pneumonia in 4, and hemorrhage in 3. One
death (0.1%) attributable to the endoscopic complication occurred in each of these 3 subgroups.
American Society of
Gastrointestinal Endoscopy Bleeding Survey:
Complications of 2320 Endoscopies
TABLE 2610
COMPLICATION
Major
Perforation
Aspiration
Hemorrhage
Minor
NUMBER
5*
4*
3*
American Society of
Gastrointestinal Endoscopy Bleeding Survey:
Complications of 2320 Endoscopies
TABLE 2610
COMPLICATION
NUMBER
Mucosal tear
3
Medication reaction
3
Hypotension
1
Atrial fibrillation
1
Anoxic episode
1
From Gilbert DA, Silverstein FE, Tedesco FJ, et al.
National ASGE survey on upper gastrointestinal
bleeding: Complications of endoscopy. Dig Dis Sci 1981;
26:55s-9s, with permission from Plenum Publishing.
* One death in each of these groups.
This ASGE survey120(1231) is the third of the ASGE membership concerning complications of upper
endoscopy. Both earlier reports were retrospective. In 1967, the experience of 128 ASGE members
was surveyed;121(1232) in 35,448 examinations, there were 31 complications (0.09%) and 6 deaths
(0.02%). A minority of the ASGE members in the survey used flexible fibergastroscopes for these
examinations; most used semiflexible instruments. In 1974, an ASGE survey indicated a complication
rate of 0.13% in 211,410 endoscopic examinations of the upper gastrointestinal tract. Cardiopulmonary
events were the most common type of complication, followed by perforation and
bleeding.122,123(1233)
The complication rate of 0.9% in the third ASGE survey is higher than that in either of the two previous
studies.120(1234) In 2320 procedures, a major complication occurred in 1 of every 193 examinations
and a minor complication occurred in 1 of 257 procedures. The mortality rate associated with
endoscopy was approximately 1 death per 700 patients. The prospective nature of the third survey and
the fact that the patient population was composed of seriously ill and actively bleeding patients explains
to some extent the higher incidence of complications. In other endoscopic series of patients with upper
gastrointestinal bleeding, the complication rate ranges from 0.7 to 8.0%.11,35,39,124,125(1235) The
latter figure was reported in a study by Paul and Huchzermeyer125(1236) that included a group of 98
patients who had upper gastrointestinal bleeding while in intensive care units. Cotton et al.39(1237)
reported 1 perforation in a series of 196 patients undergoing diagnostic endoscopy for upper
gastrointestinal bleeding, and Allan and Dykes11(1238) reported 1 fatal case of aspiration pneumonia
in a series of 100 bleeding patients who underwent endoscopy.
Based on a review of 26 published reports, Katon117(1239) estimated the mortality rate for endoscopy
in upper gastrointestinal bleeding to be approximately 0.02%. Noel et al.124(1240) found that the
complication rate for patients who were 65 years and older was 5%, with a mortality rate of 2%. The
risk of aspiration is emphasized in the study of Lipper et al.119(1241) In a survey of French
endoscopists, emergency endoscopy for upper gastrointestinal bleeding was the most hazardous type
of endoscopy.120(1242)
The complication rate as determined in the ASGE survey appears consistent with other observations
on the risks of endoscopy in this patient population. All available data on endoscopy for upper
gastrointestinal bleeding confirm that the risk of the procedure exceeds that of routine diagnostic
esophagogastroduodenoscopy.
Aside from the fact that the morbidity and mortality associated with gastrointestinal bleeding is
influenced by the presence of other serious illnesses, these concomitant conditions can have a bearing
on the performance of endoscopy. Associated cardiac, pulmonary, and liver diseases may represent
significant problems for the performance of endoscopy. Cappell et al.126(1243) reported the results of
a retrospective study of 82 patients with overt and 14 patients with occult gastrointestinal bleeding that
developed in the setting of acute myocardial infarction (i.e., 1.2% and 0.2%, respectively, of all patients
with myocardial infarction) in two medical centers. Upper gastrointestinal endoscopy was performed on
42 occasions in 34 of these patients at a mean of 6.2 days after myo-cardial infarction. Diagnostic
information was provided by endoscopy in 79% of patients; this was considered clinically helpful for 4
patients (12%). No complications occurred in 26 patients who were relatively stable. However,
endoscopy in 8 unstable patients resulted in complications in 3: fatal ventricular tachycardia, near
respiratory arrest, and hypotension. Cappell et al.126(1244) emphasized the need for careful
monitoring and support when endoscopy is performed in this clinical setting.
REFERENCES
1.
Gilbert DA, Silverstein FE, Auth DC, Rubin CE. Nonsurgical management of acute
nonvariceal upper gastrointestinal bleeding. Prog Hemost Thromb 1978;4:34995.
2. Gostout C. Acute gastrointestinal bleedingA common problem revisited. Mayo Clin Proc
1988;63:596604.
3.
Peterson WL. Gastrointestinal bleeding. In Sleisenger MH, Fordtran JS (eds). Gastrointestinal
Disease: Pathophysiology, Diagnosis, Management. 4th ed. Vol I. Philadelphia: WB
Saunders, 1989;397427.
4.
Elta G. Approach to the patient with gross gastrointestinal bleeding. In Yamada T (ed).
Textbook of Gastroenterology. Vol I. Philadelphia: JB Lippincott, 1991;591616.
5.
Freeman ML. The current endoscopic diagnosis and intensive care unit management of
severe ulcer and other nonvariceal upper gastrointestinal hemorrhage. Gastrointest Endosc
Clin North Am 1991;1:20939.
6. NIH Consensus Development Conference. Endoscopy. What is its role in upper GI bleeding?
Dig Dis Sci 1981;26:1s-5s.
7. NIH Consensus Conference. Therapeutic endoscopy in bleeding ulcers. JAMA
1989;262:136972.
8. Cutler JA, Mendeloff AI. Upper gastrointestinal bleeding. Nature and magnitude of the
problem in the U.S.. Dig Dis Sci 1981;26(suppl 7):90S-6S.
9.
Schiller KFR, Truelove SC, Williams DG. Haematemesis and melaena with special reference
to factors influencing the outcome. BMJ 1970;2:714.
10. Gilbert DA. Epidemiology of upper GI bleeding. Gastrointest Endosc 1990;36:S8-S13.
11. Allan R, Dykes P. A study of the factors influencing mortality rates from gastrointestinal
haemorrhage. Q J Med 1976;45:53350.
12. Morgan AG, Clamp SE. O.M.G.E. International Upper Gastrointestinal Bleeding Survey
19781982. Scand J Gastroenterol Suppl 1984;95:4158.
13. Bordley DR, Mushlin AI, Dolan JG, et al. Early clinical signs identify low-risk patients with
acute upper gastrointestinal hemorrhage. JAMA 1985;253:32825.
14. Katschinski BD, Logan RF, Davies J, Langman MJ. Audit of mortality in upper gastrointestinal
bleeding. Postgrad Med J 1989;65:9137.
15. Larson G, Schmidt T, Gott J, et al. Upper gastrointestinal bleeding: Predictors of outcome.
Surgery 1986;100:76573.
16. Provenzale D, Sandler RS, Wood DR, et al. Development of a scoring system to predict
mortality from upper gastrointestinal bleeding. Am J Med Sci 1987;294:2632.
17. Supe AN, Soonawala PF, Mathur SK, Samsi AB. Prognostic markers in upper gastrointestinal
hemorrhage. Indian J Gastroenterol 1989;8:2336.
18. Matthewson K, Pugh S, Northfield TC. Which peptic ulcer patients bleed?. Gut 1988;29:704.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
FDC Reports. Prescription and OTC Pharmaceuticals. Chevy Chase, MD: F-D-C Reports,
Nov 29, 1987;89.
Soll AH, Weinstein WM, Kurata J, et al. Nonsteroidal anti-inflammatory drugs and peptic ulcer
disease. Ann Intern Med 1991;114:30719.
Levy M, Miller DR, Kaufman DW, et al. Major upper gastrointestinal tract bleeding. Relation to
the use of aspirin and other nonnarcotic analgesics. Arch Intern Med 1988;148:2815.
Hirschowitz BI, Lanas A. Intractable and recurrent post surgical peptic ulceration is due to
aspirin abuse, much of it surreptitious [Abstract]. Gastroenterology 1992;102:A84.
Piper JM, Ray WA, Daugherty JR, et al. Corticosteroid use and peptic ulcer disease: Role of
nonsteroidal anti-inflammatory drugs. Ann Intern Med 1991;114:73540.
Wilcox CM, Truss CD. Gastrointestinal bleeding in patients receiving long-term anticoagulant
therapy. Am J Med 1988;84:68390.
Silverstein FE, Gilbert DA, Tedesco FJ, et al. The national ASGE survey on upper
gastrointestinal bleeding. II. Clinical prognostic factors. Gastrointest Endosc 1981;27:8093.
Koff RS. Benefit of endoscopy in upper gastrointestinal bleeding in patients with liver disease.
Luk GD, Bynum TE, Hendrix TR. Gastric aspiration in localization of gastrointestinal
hemorrhage. JAMA 1979;241:5768.
Gilbert DA, Silverstein FE, Tedesco FJ, et al. The national ASGE survey on upper
gastrointestinal bleeding. III. Endoscopy in upper gastrointestinal bleeding. Gastrointest
Endosc 1981;27:94102.
Cuellar RE, Gavaler JS, Alexander JA, et al. Gastrointestinal tract hemorrhage. The value of a
nasogastric aspirate. Arch Intern Med 1990;150:13814.
Gilbert DA, Saunders DR. Iced saline lavage does not slow bleeding from experimental
canine gastric ulcers. Dig Dis Sci 1981;26:10658.
Herlinger H. Other diagnostic approaches to upper gastrointestinal bleeding: Utility of contrast
radiology. Dig Dis Sci 1981;26:76s-7s.
Winzelberg GG, McKusick KA, Strauss HW, et al. Evaluation of gastrointestinal bleeding by
red blood cells labeled in vivo with technetium-99m. J Nucl Med 1979;20:10806.
Nusbaum M, Baum S. Radiographic demonstration of unknown sites of gastrointestinal
bleeding. Surg Forum 1963;14:3745.
Athanasoulis CA, Waltman AC, Novelline RA, et al. AngiographyIts contribution to the
emergency management of gastrointestinal hemorrhage. Radiol Clin North Am
1976;14:26580.
Palmer ED. The vigorous diagnostic approach to upper-gastrointestinal tract hemorrhage.
JAMA 1969;207:147780.
Allen HM, Block MA, Schuman BM. Gastroduodenal endoscopy. Management of acute upper
gastrointestinal hemorrhage. Arch Surg 1973;106:4505.
McCray RS, Martin F, Amir-Ahmadi H, et al. Erroneous diagnosis of hemorrhage from
esophageal varices. Am J Dig Dis 1969;14:75560.
Dagradi AE, Mehler R, Tan DTD, Stempien SJ. Sources of upper gastrointestinal bleeding in
patients with liver cirrhosis and large esophagogastric varices. Am J Gastroenterol
1970;54:45863.
Cotton PB, Rosenberg MT, Waldram RLP, Axon ATR. Early endoscopy of oesophagus,
stomach, and duodenal bulb in patients with haematemesis and melaena. BMJ 1973;2:5059.
Allan RN, Dykes PW, Toye DKM. Diagnostic accuracy of early radiology in acute
gastrointestinal haemorrhage. BMJ 1972;4:2814.
Stevenson GW, Cox RR, Roberts CJ. Prospective comparison of double-contrast barium
meal examination and fibreoptic endoscopy in acute upper gastrointestinal haemorrhage.
BMJ 976;2:7234.
Katon RM, Smith FW. Panendoscopy in the early diagnosis of acute upper gastrointestinal
bleeding. Gastroenterology 1973;65:72834.
Morris DW, Levine GM, Soloway RD, et al. Prospective, randomized study of diagnosis and
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
outcome in acute upper-gastrointestinal bleeding: Endoscopy versus conventional

radiography. Am J Dig Dis 1975;20:11039.
McGinn FP, Guyer PB, Wilken BJ, Steer HW. A prospective comparative trial between early
endoscopy and radiology in acute upper gastrointestinal haemorrhage. Gut 1975;16:70713.
Morrissey JF. Clinical approach to diagnostic endoscopy in patients with upper gastrointestinal
bleeding. Dig Dis Sci 1981;26(suppl 7):6S-11S.
Thoeni RF, Cello JP. A critical look at the accuracy of endoscopy and double-contrast
radiography of the upper gastrointestinal (UGI) tract in patients with substantial UGI
hemorrhage. Radiology 1980;135:3058.
Voeller GR, Bunch G, Britt LG. Use of technetium-labeled red blood cell scintigraphy in the
detection and management of gastrointestinal hemorrhage. Surgery 1991;110:799804.
Hellers G, Ihre T. Impact of change to early diagnosis and sur-gery in major upper
gastrointestinal bleeding. Lancet 1975;2:12501.
Paul F, Seifert E, Otto P. In Maratka Z, Setka J (eds). Urgent Endoscopy of Digestive and
Abdominal Diseases. Basel: Karger, 1972;64.
Sugawa C, Werner MH, Hayes DF, et al. Early endoscopy. A guide to therapy for acute
hemorrhage in the upper gastrointestinal tract. Arch Surg 1973;107:1337.
Fleischer D. Etiology and prevalence of severe persistent upper GI bleeding.
Sugawa C, Benishek D, Walt AJ. Mallory-Weiss syndrome. A study of 224 patients. Am J
Surg 1983;145:303.
Clain JE, Novis BH, Barbezat GO, Bank S. The Mallory-Weiss syndrome. A prospective study
in 130 patients. S Afr Med J 1978;53:5967.
Paquet KJ, Mercado-Diaz M, Kalk JF. Frequency, significance and therapy of the
Mallory-Weiss syndrome in patients with portal hypertension. Hepatology 1990;11:87983.
Ansari A. Mallory-Weiss syndrome. Experience in a community hospital. Postgrad Med
1984;76:18995, 198.
Baker MS, Fisher JH, van der Reis L, Baker BH. The endoscopic diagnosis of an
aortoduodenal fistula. Arch Surg 1976;111:304.
Bedogni G, Meinero M, Barbieri I, et al. Gastrointestinal hemorrhage from an aortoduodenal
fistula: Endoscopic considerations. Ital J Surg Sci 1985;15:2136.
Brady PG. Aortoduodenal fistula: Role of endoscopy in diagnosis. Am J Gastroenterol
1978;69:7057.
Konigsrainer A, Pointner R, Reissigl H, Weimann S. Endoscopic aspects of aortointestinal
fistula. Surg Endosc 1987;1:3740.
Mark AS, Moss AA, McCarthy S, McCowin M. CT of aortoenteric fistulas. Invest Radiol
1985;20:2725.
Juler GL, Labitzke HG, Lamb R, Allen R. The pathogenesis of Dieulafoy's gastric erosion. Am
J Gastroenterol 1984;79:195200.
Veldhuyzen Van Zanten SJO, Bartelsman JFW, Schipper MEI, Tytgat GNJ. Recurrent
massive haematemesis from Dieulafoy vascular malformationsA review of 101 cases. Gut
1986;27:21322.
Baettig B, Haecki W, Lammer F, Jost R. Dieulafoy's disease: Endoscopic treatment and
follow-up. Gut 1993;34:141821.
Pointner R, Schwab G, Konigsrainer A, Dietze O. Endoscopic treatment of Dieulafoy's
disease. Gastroenterology 1988;94:56366.
Quintero E, Pique JM, Bombi JA, et al. Upper gastrointestinal bleeding caused by
gastroduodenal vascular malformations. Incidence, diagnosis, and treatment. Dig Dis Sci
1986;31:897905.
Clouse RE, Costigan DJ, Mills BA, Zuckerman GR. Angiodysplasia as a cause of upper
gastrointestinal bleeding. Arch Intern Med 1985;145:45861.
Chu DZ, Shivshanker K, Stroehlein JR, Nelson RS. Thrombocytopenia and gastrointestinal
hemorrhage in the cancer patient: Prevalence of unmasked lesions. Gastrointest Endosc
1983;29:26972.
68. Shivshanker K, Chu DZ, Stroehlein JR, Nelson RS. Gastrointestinal hemorrhage in the cancer
patient. Gastrointest Endosc 1983;29:2735.
69. Zuckerman GR, Cornette GL, Clouse RE, Harter HR. Upper gastrointestinal bleeding in
patients with chronic renal failure. Ann Intern Med 1985;102:58892.
70. Sharon P, Stalnikovicz R, Rachmilewitz D. Endoscopic diagnosis of duodenal neoplasms
causing upper gastrointestinal bleeding. J Clin Gastroenterol 1982;4:358.
71. Ashkenazi S, Sharon P, Levij IS, et al. Giant cell carcinoma of the pancreas. Report of a case
with upper gastrointestinal bleeding diagnosed by endoscopic brush cytology. Am J
72. Theodors A, Sivak MV Jr, Carey WD. Hypernephroma with metastasis to the duodenum:
Endoscopic features. Gastrointest Endosc 1980;26:4851.
73. Lee P, Sutherland D, Feller ER. Massive gastrointestinal bleeding as the initial manifestation
of pancreatic carcinoma. Int J Pancreatol 1994;15:2237.
74. DeVault KR, Miller LS, Yaghsezian H, et al. Acute esophageal hemorrhage from a vagal
neurilemoma. Gastroenterology 1992;102:105961.
75. Petersen JM, Ferguson DR. Gastrointestinal neurofibromatosis. J Clin Gastroenterol
1984;6:52934.
76. Morse JM, Reddy KR, Thomas E. Hemosuccus pancreaticus: A cause for obscure
gastrointestinal bleedingDiagnosis by endoscopy and successful management by total
parenteral nutrition. Am J Gastroenterol 1983;78:5724.
77. Rsch W, Schaffner O, Frhmorgen P, Koch H. Massive gastrointestinal hemorrhage into the
pancreatic duct diagnosed by duodenoscopy and ERCP. Endoscopy 1977;8:936.
78. Wolfe MW, Horton ML 3d. Endoscopic observation of ampullary hemorrhage secondary to
occult pancreatitis. Gastrointest Endosc 1989;35:3446.
79. Balldin G, Cronstedt J. Bleeding birth of a gallstone from a hanging papilla of Vater.
80. Kunert H, Ottenjann R. Endoscopy in bleeding duodenal varices: A rare cause of massive
upper GI hemorrhage. Endoscopy 1977;8:99101.
81.
Kunisaki T, Someya N, Shimokawa Y, Okuda K. Varices in the distal duodenum seen by a
fiber-duodenoscope. Endoscopy 1973;5:1014.
82. Rappazzo JA, Kozarek RA, Altman M. Duodenal varices: Endoscopic diagnosis of an unusual
source of upper gastrointestinal hemorrhage. Gastrointest Endosc 1981;27:2278.
83. Levy DJ, Franklin GO, Rosenthal WS. Gastrointestinal bleeding and amyloidosis. Am J
84.
Foster DN, Miloszewski KJA, Losowsky MS. Stigmata of recent haemorrhage in diagnosis
and prognosis of upper gastrointestinal bleeding. BMJ 1978;1:11737.
85. Harris DCH, Heap TR. Significance of signs of recent hemorrhage at endoscopy. Med J Aust
1982;2:3573.
86. Wara P, Hojsgaard A, Amdrup E. Endoscopic electrocoagulationAn alternative to operative
hemostasis in active gastroduodenal bleeding?. Endoscopy 1980;12:23740.
87. Griffiths WJ, Neumann DA, Welsh JD. The visible vessel as an indicator of uncontrolled or
current gastrointestinal hemorrhage. N Engl J Med 1979;300:14113.
88. Storey DW, Bown SG, Swain CP, et al. Endoscopic prediction of recurrent bleeding in peptic
ulcers. N Engl J Med 1981;305:9156.
89.
Bornman PC, Theodorou NA, Shuttleworth RD, et al. Importance of hypovolaemic shock and
endoscopic signs in predicting recurrent haemorrhage from peptic ulceration: A prospective
evaluation. BMJ 1985;291:2457.
90.
MacLeod IA, Mills PR, MacKenzie JF, et al. Neodymium yttrium aluminium garnet laser
photocoagulation for major haemorrhage from peptic ulcers and single vessels: A single blind
controlled study. BMJ (Clin Res Ed) 1983;286:3458.
91. Vallon AG, Cotton PB, Laurence BH, et al. Randomised trial of endoscopic argon laser
photocoagulation in bleeding peptic ulcers. Gut 1981;22:22833.
92. Wara P. Endoscopic prediction of major rebleedingA prospective study of stigmata of

hemorrhage in bleeding ulcer. Gastroenterology 1985;88:120914.
93. Swain CP, Kirkham JS, Salmon PR, et al. Controlled trial of Nd:YAG laser photocoagulation in
94. O'Brien JD, Day SJ, Burnham WR. Controlled trial of small bipolar probe in bleeding peptic
ulcers. Lancet 1986;1:4647.
95. Laine L. Multipolar electrocoagulation in the treatment of peptic ulcers with nonbleeding visible
vessels. A prospective, controlled trial. Ann Intern Med 1989;110:514.
96. Swain CP, Storey DW, Bown SG, et al. Nature of the bleeding vessel in recurrently bleeding
gastric ulcers. Gastroenterology 1986;90:595608.
97.
Chung SCS, Leung JUC, Lo KK, et al. Natural history of the sentinel clot: An endoscopic study
[Abstract]. Gastroenterology 1990;98:31.
98. Laine L, Cohen H, Brodhead J, et al. Prospective evaluation of immediate versus delayed
refeeding and prognostic value of endoscopy in patients with upper gastrointestinal
hemorrhage. Gastroenterology 1992;102:3146.
99.
Freeman ML, Cass OW, Brunsvold NJ, et al. Evolution of the non-bleeding visible vessel in
severe ulcer hemorrhage: Endoscopic and clinical predictors of rebleeding. Gastroenterology
1991;100:A67.
100. Forrest JAH, Finlayson NDC, Shearman DJC. Endoscopy in gastrointestinal bleeding. Lancet
1974;2:3947.
101.
Leinicke JA, Shaffer RD, Hogan WJ, Geenen JE. Emergency endoscopy in active upper GI
bleeding (UGB): Does timing affect the significance of diagnostic yield? [Abstract].
102.
Palmer ED. Observations on the vigorous diagnostic approach to severe upper
gastrointestinal hemorrhage. Ann Intern Med 1952;36:148491.
103.
Eastwood GE. Does the patient with upper gastrointestinal bleeding benefit from endoscopy?.
104. Winans CS. Emergency upper gastrointestinal endoscopy: Does haste make waste?. Am J
Dig Dis 1977;22:53640.
105. Keller RT, Logan GM Jr. Comparison of emergent endoscopy and upper gastrointestinal
series radiography in acute upper gastrointestinal haemorrhage. Gut 1976;17:1804.
106. Dronfield MW, McIllmurray MB, Ferguson R, et al. A prospective randomized study of
endoscopy and radiology in acute upper gastrointestinal tract bleeding. Lancet
1977;1:11679.
107. Sandlow LJ, Becker GH, Spellberg MA, et al. A prospective randomized study of the
management of upper gastrointestinal hemorrhage. Am J Gastroenterol 1974;61:2829.
108. Allan R, Dykes P. A comparison of routine and selective endoscopy in the management of
acute gastrointestinal hemorrhage. Gastrointest Endosc 1974;20:1545.
109. Graham DY. Limited value of early endoscopy in the management of acute upper
gastrointestinal bleeding. Am J Surg 1980;140:28490.
110. Logan RF, Finlayson ND. Death in acute upper gastrointestinal bleeding. Can endoscopy
reduce mortality? Lancet 1976;1:11735.
111.
Hoare AM. Comparative study between endoscopy and radiology in acute upper
gastrointestinal haemorrhage. BMJ 1975;1:2730.
112. Peterson WL, Barnett CC, Smith HJ, et al. Routine early endoscopy in
upper-gastrointestinal-tract bleeding: A randomized, controlled trial. N Engl J Med
1981;304:9259.
113. Erickson RA, Glick ME. Why have controlled trials failed to demonstrate a benefit of
esophagogastroduodenoscopy in acute upper gastrointestinal bleeding? A probability model
analysis. Dig Dis Sci 1986;31:7608.
114. Zuckerman G, Welch R, Douglas A, et al. Controlled trial of medical therapy for active upper
gastrointestinal bleeding and prevention of rebleeding. Am J Med 1984;76:3616.
115. Cook DJ, Guyatt GH, Salena BJ, Laine LA. Endoscopic therapy for acute nonvariceal upper
gastrointestinal hemorrhage: A meta-analysis. Gastroenterology 1992;102:13948.

116. Schrock TR. Does endoscopy affect the surgical approach to the patient with upper
gastrointestinal bleeding?. Dig Dis Sci 1981;26(suppl 7):27S-30S.
117. Katon RM. Complications of upper gastrointestinal endoscopy in the gastrointestinal bleeder.
Dig Dis Sci 1981;26(suppl 7):47S-54S.
118. Fleischer DE. Monitoring the patient receiving conscious sedation for gastrointestinal
endoscopy: Issues and guidelines. Gastrointest Endosc 1989;35:2626.
119. Lipper B, Simon D, Cerrone F. Pulmonary aspiration during emergency endoscopy in patients
with upper gastrointestinal hemorrhage. Crit Care Med 1991;19:3303.
120.
Gilbert DA, Silverstein FE, Tedesco FJ, et al. National ASGE survey on upper gastrointestinal
bleeding: Complications of endoscopy. Dig Dis Sci 1981;26.
121. Katz D. Morbidity and mortality in standard and flexible gastrointestinal endoscopy.
122. Silvis SE, Nebel O, Rogers G, et al. Endoscopic complications. Results of the 1974 American
Society for Gastrointestinal Endoscopy Survey. JAMA 1976;235:92830.
123. Mandelstam P, Sugawa C, Silvis SE, et al. Complications associated with
esophagogastroduodenoscopy and with esophageal dilation. Gastrointest Endosc
1976;23:169.
124. Noel D, Delage Y, Liguory C, et al. Hemorrhagies digestives d'origine haute chez les sujets de
plus de 65 ans. Apport de l'Endoscopie. Nouv Presse Med 1979;8:58991.
125.
Paul F, Huchzermeyer H. Results and complications of emergency endoscopy for acute
gastrointestinal bleeding in patients on intensive care units. In Abstracts of the IV European
Congress of Gastrointestinal Endoscopy. Stuttgart: Georg Tieme Verlag. 1980.
126. Cappell MS. The safety and clinical utility of esophagogastroduodenoscopy for acute
gastrointestinal bleeding after myocardial infarction: A six-year study of 42 endoscopies in 34
consecutive patients at two university teaching hospitals. Am J Gastroenterol
1993;88:34450.
Chapter 27 Lower Gastrointestinal Bleeding

(1245)
(1246)
THEODORE R. SCHROCK, M.D.
The diagnosis and treatment of lower gastrointestinal (GI) bleeding has changed dramatically since the
mid-1960s.17(1247) New as well as previously unrecognized lesions have been described, new
diagnostic methods have been developed, and new therapeutic approaches have been defined.
Angiography was the major technologic advance that first influenced the care of patients with bleeding
from the small bowel or colon; colonoscopy followed soon thereafter. The clinician now must select
from among colonoscopy, angiography, radionuclide scans, and various operative maneuvers to
localize and control lower GI bleeding as quickly, safely, and inexpensively as possible.
Diagnostic strategy hinges to a great extent on the rate of bleeding, and the amount of blood lost at the
moment of decision is an important related factor. A classification of lower GI bleeding based on rate
and volume is given in Table 271. These categories overlap, and patients shift from one to another as
their clinical course unfolds. No set of definitions, including these, is universally accepted, but this
framework is useful to the discussion of diagnosis and treatment. GI bleeding has also been
categorized as primary (patient hospitalized with a primary diagnosis of GI bleeding) and secondary
(bleeding develops during hospitalization for another problem).8(1248)
Lower Gastrointestinal
Bleeding According to Rate and Volume
TABLE 271
categorized as primary (patient hospitalized with a primary diagnosis of GI bleeding) and secondary
(bleeding develops during hospitalization for another problem).8(1248)
Lower Gastrointestinal
Bleeding According to Rate and Volume
TABLE 271
CHRONIC
ACUTE
Occult
Moderate
Gross
Severe
Melena
Hematochezia
From Schrock T. Colonoscopic diagnosis and treatment of
lower gastrointestinal bleeding. Surg Clin North Am
1989; 69:130925.
Numerous bleeding lesions in the upper, middle, and lower GI tract may result in the passage of blood
per rectum. Conditions that may cause occult or gross, chronic or acute lower GI bleeding are listed in
Table 272.
TABLE 272
Causes of Gastrointestinal Bleeding*
ESOPHAGUS, STOMACH, DUODENUM
JEJUNUM, ILEUM
COLORECTUM
AV malformations
Angiodysplasia
Angiodysplasia
AV malformations
Ulcers
Ulcerative colitis
Anastomotic
Diverticulosis
Simple
Cancer
Diverticula
Polyps
Meckel's
Hemorrhoids
Acquired
Anal fissure
Crohn's disease
Stomal varices
Varices
Postoperative
Ischemic ulcer
Postpolypectomy
Tuberculosis
Anastomotic
Arteritis
Trauma
Blind loop
Ulcers
Angioma
Simple
Stercoral
Leiomyoma
Cancer
Typhoid
Sarcoma
Amebic
Polyps
Uremic ulcer
Stomal varices
Lymphoid hyperplasia
Trauma
From Moncure A, Tompkins R, Athanasoulis C, et al. Occult gastrointestinal bleeding: Newer techniques of
diagnosis and therapy. Adv Surg 1989; 22:14177.
Esophageal varices
Esophagitis
Gastritis
Gastric varices
Mallory-Weiss tears
Peptic ulcer
AV malformations
Cancer
Polyps
Leiomyoma
Sarcoma
Brunner's adenoma
Angiodysplasia
Pancreatic rest
Trauma
Postoperative
Retained ulcer
Residual gastritis
Anastomotic ulcer
TABLE 272
Causes of Gastrointestinal Bleeding*
ESOPHAGUS, STOMACH, DUODENUM
JEJUNUM, ILEUM
COLORECTUM
AVarteriovenous.
* Lesions in boldface are usually manifested by acute severe bleeding but also may cause chronic blood loss,
anemia, and positive tests for blood in the stool. The most common lesions are italicized.
Chronic Bleeding
Occult Bleeding
The term occult bleeding implies that blood is hidden rather than visible in the stool, although, in a
sense, all GI bleeding is occult until the site is located.9(1249) As usually defined, occult lower GI
bleeding is chronic, the rate is slow, and the stool is grossly normal. Occult bleeding is detected during
screening or case-finding, or symptoms related to anemia cause a patient to seek medical
attention.10,11(1250) Screening refers to the testing of large populations of asymptomatic individuals.
Case-finding refers to the diagnosis of disease at an early stage in patients, whether symptomatic or
asymptomatic, who consult a physician.11(1251) Analysis of the benefits of occult blood testing must
take this distinction into account. The value of screening as a public health measure is difficult to
prove, but case-finding is standard practice.
Test for Occult Fecal Blood
The most common fecal occult blood test (FOBT) uses guaiac-impregnated (Hemoccult) cards.
Results range widely. Positive tests are found in 3% of participants on average in large screening
programs, with a range of 1.1 to 6.6%.12(1252) Colorectal cancer is detected in approximately 6% of
subjects with a positive test, although the predictive value of the test has a range of 1.1 to 13.4%;
adenomas are detected in 10 to 43% of subjects.13,14(1253) The false-positive rate in large screening
studies varies from 3 to 5%; the false-negative rate is thought to be approximately 30%.10,15(1254)
Kimmig et al.16(1255) obtained FOBT (Hemoccult) in 534 patients prior to colonoscopy. The findings
at colonoscopy were classified as those expected to result in sustained bleeding (e.g., polyps >1 cm
diameter, carcinoma), those expected to bleed intermittently (e.g., polyps <1 cm diameter), and those
with no source of bleeding. Based on this classification, false-positive and false-negative results
occurred, respectively, in 4.3 and 8.6% of patients.
The outcome of screening programs for fecal occult blood is influenced by a variety of factors,
including diet, age of the population to be screened, use of medication, and patient compliance. Other
factors influence the inherent sensitivity and specificity of the test itself, one of the most controversial
being rehydration of guaiac-impregnated cards. This changes the overall results when groups of
patients are studied. In a representative study, positivity rose from 2.4 to 9.8% and sensitivity increased
from 80.8 to 92.2%, but specificity decreased from 97.7 to 90.4% and positive predictivity declined from
5.6 to 2.2%.17(1256) Positive predictivity increased with age from 1.6% for those under 60 years to
3.6% for those over 70 years in the study of Mandel et al.17(1257)
False-negative tests are caused by sampling errors, or they reflect the failure of neoplasms to shed
enough blood to be detected.18(1258) Most carcinomas ulcerate before they bleed. Polypoid
adenomas less than 1 cm in diameter are unlikely to bleed, and even polyps as large as 2 cm do not
bleed in every instance.19(1259) Neoplasms in the left colon are more often positive for occult blood
than are tumors in the right colon.20(1260)
Evidence indicates that mass screening for fecal occult blood reduces the mortality from colorectal
cancer. In the Minnesota Colon Cancer Control Study, mortality from colorectal cancer was reduced by
33% in subjects who underwent annual FOBT screening.21(1261) Furthermore, earlier-stage lesions
and improved survival were also noted for patients who developed colorectal cancer among those who
underwent yearly screening. However, Lang and Ransohoff22(1262) have suggested that a portion of
the "benefit" found in the study of Mandel et al.21(1263) might be attributable to a "chance" selection of
individuals to undergo colonoscopy. Lang and Ransohoff22(1264) argue that the results of the
Minnesota Study can be explained in part by the fact that the sensitivity of the FOBT was increased by
rehydration of the Hemoccult slides and that most of the individuals with a positive test underwent
colonoscopy. In the ongoing trial of Hardcastle et al.,14(1265) more than 107,000 asymptomatic
people have undergone screening. FOBT was positive in 2.3% of subjects; of these, 10.2% had
colorectal cancers and 43.0% had one or more adenomas.14(1266) The incidence of cancer in the
control group was 0.72 per 1000 person-years. Among cancers detected in the test group, 52% were
Stage A, compared with 10.6% that were Stage A in the control group. Although cancers detected by
screening were at a less advanced stage, no effect of screening on mortality from colorectal cancer
has been seen to date.
The reliability of FOBT performed on samples obtained by digital rectal examination is a matter of
considerable debate.23,24(1267) Many explanations other than colorectal neoplasia are possible for
blood in stool sampled in this fashion. It therefore seems prudent to avoid an expensive and
uncomfortable evaluation with the potential, albeit low, for a complication based on this result alone. A
positive FOBT obtained during digital rectal examination should be confirmed under proper
conditions.23(1268)
Several alternative methods of testing for the presence of occult blood in stool samples have been
developed in an effort to circumvent some of the problems inherent to the standard
guaiac-impregnated cards.18, 25, 26(1269) Whether these more complex and more expensive tests
are sufficiently advantageous to replace the older techniques remains to be seen. Further information
on the FOBT as a method of screening for colorectal cancer may be found in a recent review.12(1270)
Diagnostic Evaluation of Patients with Occult Bleeding
The presence of other GI disorders, diet, use of medication (e.g., aspirin), and the manner in which the
test is performed are a few of the factors known to influence the outcome of FOBT. Although it is a
cause for concern when large populations undergo screening, the false-negative FOBT has little
relevance in the evaluation of occult bleeding because individual asymptomatic patients do not come to
medical attention. The false-positive test result, however, is problematic in a number of respects.
A positive FOBT has a relatively low predictive value for colorectal cancer (about 10%). This means, in
effect, that most patients with a positive test undergo one or more negative diagnostic studies. This
situation may affect the perceptions of the public, which may in turn influence compliance. Few data
are available on this point, although a survey by Mant et al.27(1271) found that the need to endure
negative diagnostic tests because of a false-positive FOBT did not change the attitude of patients
toward the test. However, the false-positive FOBT has significant implications with respect to the cost
effectiveness of screening.
The fact that the false-positive FOBT undoubtedly increases the overall costs of colorectal cancer
screening does not mean that the ensuing evaluation is without benefit. Based on data from
colonoscopic studies of asymptomatic subjects, the percentage of individuals with "silent" adenomas
may be as high as 25%.28(1272) It is reasonable to conclude, therefore, that a negative colonoscopy
in a patient in the age range of 50 to 65 years indicates an average risk for the development of
colorectal cancer.
Repeating the FOBT is one possible strategy to reduce the number of negative diagnostic studies.
Patients with a negative result would enter a program of follow-up screening, and those with a positive
result on retesting would undergo further investigation. Flexible sigmoidoscopy (60 cm) and a
double-contrast barium enema (DCBE) were obtained in subjects who had a positive FOBT
(Hemoccult II, rehydrated) on initial screening in the prospective randomized trial of Kewenter et
al.29(1273) FOBT screening was positive in 210 patients, that is, 5.9% of the total number of subjects
who underwent screening. Prior to undergoing further studies, the FOBT was repeated in 184 of the
patients with an initially positive test; the second FOBT was positive in 68 patients. One of 7 patients
with two positive tests was found to have cancer. For those with one positive and one negative FOBT,
the rate of detection for cancer was 1 in 100 patients.
Repeating the FOBT is problematic for several reasons. It is well established that the FOBT may be
negative in patients with cancer.16,28,3032(1274) One may conclude that bleeding associated with
colonic cancer is not constant and therefore does not result in consistently detectable levels of blood in
the stool over the course of time. A second FOBT in a patient who has already tested positive is not a
diagnostic test but an attempt to increase the specificity of the screening procedure, with the tacit
acknowledgment that some cancers are missed. Compliance is an additional problem, as is being
certain that all patients understand the implications of positive and negative test results.
Assuming that nothing is found in the history or physical examination to suggest bleeding in the
proximal GI tract, two approaches to the initial evaluation of patients with a positive FOBT are possible:
flexible sigmoidoscopy plus a DCBE, and colonoscopy. For the individual patient with a positive FOBT,
selection of a diagnostic strategy must take into account a number of factors, including diagnostic
sensitivity, availability, the experience and skill of examiners, patient comfort and convenience,
potential complications, cost, and the potential need for therapeutic intervention.
Unfortunately, few accurate data are available with respect to advantages and disadvantages of the
two strategies. One or more of the factors that may change outcome in the individual case are not
addressed in many studies; only a few prospective trials attempt to control the many variables that can
influence the selection of strategy.3335(1275)
Barium Enema Plus Sigmoidoscopy
The barium enema is reasonably accurate in detecting colorectal cancer and polyps. In 98 patients
with a positive FOBT, Feczko and Halpert36(1276) found that the DCBE detected adenomatous polyps
1 cm or more in diameter in 21 patients and colorectal cancer in 6 patients. Depending on the results
of the radiographic study, either colonoscopy or a repeat FOBT was obtained. Apparently, 25 polyps
were found in 20 patients who underwent colonoscopy, all but 2 of which were detected by barium
enema; 2 polyps identified by radiography were not discovered at colonoscopy. Whether patients with a
positive FOBT and a negative DCBE had colonoscopy is not indicated in this report, and the results of
the repeat FOBT are not provided.
The sigmoid colon and cecum can be especially problematic with regard to the potential for missed
lesions during contrast radiography.37(1277) Although this is usually attributed to the presence of
redundant loops of bowel, poor bowel preparation, or diverticulosis, some studies of missed lesions
suggest that diagnostic errors are often perceptive in nature; that is, the radiographic films were of
acceptable technical quality and an abnormality was evident in retrospect.38,39(1278) Radiologic
studies of the colon, most of the single-contrast type, in 798 patients with carcinoma were reviewed
retrospectively by Bolin et al.39(1279) The sensitivity of the examination was 90%, but 64 cancers were
overlooked in 74 examinations, with a larger percentage than anticipated of the missed lesions being in
the right colon. Most missed lesions were evident in retrospect on the radiographic films. The reasons
for a missed diagnosis of colorectal cancer in the similar but more recent study of Brady et al.40(1280)
included errors of perception in 15% of cases (i.e., the cancer was evident in retrospect), technical
error in 15%, and interpretive errors in 25% (i.e., the lesion was evident but incorrectly diagnosed as
benign).
The barium enema/sigmoidoscopy strategy has been evaluated in a few studies. Flexible
sigmoidoscopy (60 cm) and DCBE were performed by Jensen et al.41(1281) in 530 patients with a
positive FOBT. Findings included an adenoma of 1 cm or greater in diameter in 71 patients and cancer
in 26 patients. An FOBT was positive a second time in 67 of 323 patients in whom the test was
repeated. Colonoscopy obtained in 55 of the patients with a second positive FOBT disclosed 2
adenomas of 1 cm or greater diameter and 1 Dukes A carcinoma. Further screening was conducted
for the entire group of patients for up to 2 years, and 2 more cancers and 1 adenoma were found. In
this study, the combination of flexible sigmoidoscopy and DCBE had an overall sensitivity and
specificity of 94 and 99%, respectively. Barry et al.42(1282) found that sigmoidoscopy plus DCBE
detected 64 to 80% of polyps greater than 5 mm in diameter and 84 to 90% of localized cancers. This
indicates that 20% of adenomas of significant size and 10% of malignant tumors are overlooked when
this strategy is used.
Studies of the accuracy of the barium enema usually refer to the detection of large as opposed to
diminutive polyps. The significance of the latter lesion with regard to potential for malignant
transformation is debated. For example, Hoff et al.43(1283) found that diminutive adenomas did not
increase in size and sometimes regressed if left untreated for periods of up to 2 years. These lesions
have therefore been considered to be less important in relation to the risk of malignant
transformation.44(1284) However, carcinoma has also been described in small adenomas.45,46(1285)
It is highly unlikely that a diminutive polyp is ever the explanation for a positive FOBT, but these lesions
are often discovered in patients who undergo evaluation for bleeding and are frequently considered to
be a positive finding at colonoscopy (the diminutive polyp is discussed in Chapter 85: Polyps and
Tumors of the Colon).
Colonoscopy
For detection of colorectal neoplasia, colonoscopy is more accurate than DCBE alone. Colonoscopy
frequently detects significant abnormalities in patients with bleeding per rectum when the DCBE
discloses no abnormalities other than diverticulosis.47,48(1286) Guillem et al.48(1287) found in a
retrospective study of patients with various forms of rectal bleeding that 10% of those with a normal
barium enema had a malignant lesion at colonoscopy. Despite the superior accuracy of colonoscopy,
significant lesions may be overlooked. Glick et al.49(1288) collected 18 examples from six institutions
of colonic neoplasms overlooked by colonoscopy. The rectum and cecum were the most common
locations for missed lesions. None was near a flexure. Missed lesions tended to be flat and somewhat
nodular and ranged in size from 2 to 8 cm in diameter (average 4.4 cm). Histologic diagnoses,
available in 13 cases, indicated 1 invasive cancer, with the remainder of the lesions being
adenomatous polyps, including 9 with villous components and 5 with foci of malignancy.
Cost is a concern when colonoscopy is the initial procedure for colonic bleeding. Although the cost of
the DCBE alone is likely to be less than that of colonoscopy, the total cost of the DCBE/flexible
sigmoidoscopy strategy is probably equal to that of colonoscopy, depending on which outcome is
considered, and especially if the cost of colonoscopy for management of lesions found by the
sigmoidoscopy/DCBE strategy is included.
The cost-effectiveness ratio was found by Barry et al.42(1289) to be lowest for the DCBE alone. Seven
possible strategies for the evaluation of patients with a positive FOBT were evaluated in this
decision-analysis model. The cost-effectiveness ratio for rigid sigmoidoscopy plus barium enema was
better than for colonoscopy alone, but only if removal of polyps was excluded. Colonoscopy alone was
better than flexible sigmoidoscopy plus barium enema.
It is generally accepted that the combination of DCBE and flexible sigmoidoscopy has a lesser risk of
complications than does colonoscopy. Few data are available on the risks associated with the DCBE,
but complications can occur.50,51(1290) Complications, albeit rare, can also occur with flexible
sigmoidoscopy (see Chapter 91: Flexible Sigmoidoscopy). Eddy52(1291) estimated perforation rates
for these various procedures as follows: 35-cm flexible sigmoidoscopy 0.02%, 60-cm flexible
sigmoidoscopy 0.045%, colonoscopy 0.2%, and DCBE 0.02%.
No accurate data are available on patient comfort, acceptance, and convenience for colonoscopy
alone versus DCBE plus flexible sigmoidoscopy. Discomfort is generally assumed to be greater with
colonoscopy. This may be true when the procedure is technically difficult, but colonoscopy performed
by an experienced endoscopist is reasonably comfortable and well accepted by patients. Because
conscious sedation is usually used for colonoscopy, many patients may regard DCBE and
sigmoidoscopy as more vexatious. One procedure may be more acceptable than two, especially if the
maneuvers for preparing the colon are considered. Furthermore, a third procedurecolonoscopyis
required in at least 15% of patients evaluated initially with the combination of DCBE and flexible
sigmoidoscopy.
Esophagogastroduodenoscopy
If nothing of significance is found in the colon, the second consideration is whether to investigate the
upper GI tract, usually by esophagogastroduodenoscopy (EGD). The role of EGD in the evaluation of
patients with occult fecal bleeding is controversial.
EGD and colonoscopy were performed in 100 patients with occult fecal blood, iron deficiency anemia,
or both in the prospective study of Zuckerman and Benitez.53(1292) Colonoscopy yielded a possible
source of bleeding in 26% of patients, EGD provided a positive diagnosis in 36%, and one or the other
test was positive in 53%.53(1293) Causes of bleeding are listed in Table 273. The diagnostic yield
was significantly higher for EGD than for colonoscopy in patients with both anemia and guaiac-positive
stools (45 vs. 26%). Colonoscopy had a higher cancer detection rate than EGD, yet EGD detected the
origin of occult blood in 68% of patients in whom a bleeding source was found. EGD resulted in a
change of therapy in 30 of the 100 patients studied.53(1294) Zuckerman and Benitez concluded that
when cost is a consideration and the goal of screening is to reduce the mortality of colorectal cancer,
patients with a positive FOBT and no symptoms referable to the upper digestive tract should have
colonoscopy first; if colonoscopy is negative, further investigation is probably unnecessary.53(1295)
However, if fecal occult blood is found on repeat testing, another colonoscopy is indicated, and EGD
should be performed for persistent occult bleeding.
TABLE 273
Endoscopic Findings for the Source of Occult Gastrointestinal
Bleeding
ESOPHAGOGASTRODUODENOSCOPY
(36 PATIENTS)
Esophagitis
Gastric ulcer
Gastric erosions
Duodenal ulcer
Duodenal erosions
Total peptic lesions
Gastric angiodysplasia
Duodenal angiodysplasia
Both gastric and duodenal
Total angiodysplasia
Esophageal varices
Gastric varices
Duodenal polyp
COLONOSCOPY
(26 PATIENTS)
6(17)*
6(17)
12(33)
1
2
27(75)
1
4
3
8(22)
1
1
1
Colon polyps
Colon cancer
Angiodysplasia
Nonspecific ulcers
14(53)
6(26)
5
2
TABLE 273
Endoscopic Findings for the Source of Occult Gastrointestinal
Bleeding
(36 PATIENTS)
Gastric cancer
Total lesions
COLONOSCOPY
(26 PATIENTS)
1
Total lesions
39
27
Stigmata of bleeding
8
Stigmata of bleeding
7
From Zuckerman G, Benitez J. A prospective study of bidirectional endoscopy (colonoscopy and upper
endoscopy) in the evaluation of patients with occult gastrointestinal bleeding. Am J Gastroenterol 1992;
87:626, American College of Gastroenterology.
* Numbers in parentheses are percentages.
Three patients had two lesions each.
One patient had two lesions.
Hsia and al-Kawas54(1296) recommended that EGD be obtained in all asymptomatic patients with
fecal occult blood and a negative evaluation for a colonic source. Significant and previously
undiagnosed sources of bleeding were found in 19 of 70 (27%) such patients in this consecutive
series.54(1297) However, 13 patients were anemic and in 38% of this group a significant finding was
discovered at EGD. Furthermore, 15 patients in this study were taking nonsteroidal anti-inflammatory
agents. Erosions or peptic ulcer disease made up about one half of the positive endoscopic diagnoses.
Findings that could be considered unquestionably significant but clinically silent, such as vascular
ectasias, varices, and Barrett's esophagus, were relatively rare. In the retrospective study of Chen et
al.,55(1298) EGD disclosed a positive finding in 43% of 117 asymptomatic patients with a positive
FOBT and a negative colonoscopy. However, relatively few patients had significant findings at EGD,
this being defined as a diagnosis that resulted in a change in therapy. For example, no cases of
malignancy were found by EGD.
The study of Hsia and al-Kawas54(1299) may be compared with that of Thomas and
Hardcastle,56(1300) in which 447 patients (2.6% of subjects screened) had occult fecal blood. In 283
of these patients (63%), neoplastic disease was not found in the colon. Fourteen of these patients had
symptoms suggestive of upper GI disease, and EGD in this group revealed a significant lesion in 6
patients, including 1 gastric cancer. The remaining 269 patients (asymptomatic, negative colon
evaluation) were followed for a median of 5 years (range 2 to 8 years), during which time various upper
GI disorders, but no malignancies, were found in 5 patients.
Endoscopy of the small intestine (enteroscopy) is performed alone or as part of an abdominal
operation in a few unusual patients with chronic occult (or gross) bleeding (see Chapter 50: Endoscopy
of the Small Intestine).5764(1301) In the study of Lewis et al.60(1302) that compared preoperative
enteroscopy with intraoperative enteroscopy, the results were the same in 77% of 23 patients with
chronic obscure GI bleeding. These authors recommend enteroscopy alone before undertaking
intraoperative enteroscopy because operation is not beneficial if bleeding is due to multiple ectasias
throughout the bowel.60(1303) Ress et al.58(1304) performed intraoperative enteroscopy in patients
who had bled for 1 to 247 months (mean, 15 months); specific mucosal abnormalities were found in
70% of patients, 61% underwent resection of intestinal lesions, and 59% of survivors after resection
had recurrent bleeding. In still another study, recurrent bleeding was encountered in 45% of
patients.60(1305)
The availability of recombinant erythropoietin to stimulate the bone marrow has introduced a new
therapeutic alternative for patients with persistent low-volume occult GI bleeding that cannot be
localized and treated by currently available methods.58(1306) Such patients may be given
erythropoietin and iron to stabilize the red cell mass and avoid blood transfusions instead of performing
diagnostic laparotomy. Few data have been reported on this strategy, but it has theoretical appeal for
highly selected patients.
Choice of Diagnostic Strategy
Occult fecal blood requires evaluation by history, physical examination, laboratory studies, and
anoscopy. Rigid proctosigmoidoscopy is useful in the evaluation of anorectal disease, but it has limited
value in patients with occult bleeding.20(1307) The entire colon must be examined by colonoscopy or a
combination of flexible sigmoidoscopy and DCBE.42(1308) Both the initial colonoscopy and the flexible
sigmoidoscopy/DCBE strategies have their proponents. The most compelling argument in favor of the
latter strategy is cost, but this must be weighed against the additional advantages of colonoscopy,
which include greater diagnostic accuracy for all potentially significant lesions (e.g., vascular ectasias,
small adenomas), greater patient convenience and comfort (one procedure compared with two or three
with the combined approach), and the possibility of histologic diagnosis and immediate therapy (e.g.,
polypectomy, coagulation of vascular ectasias). It is doubtful that two procedures are more convenient
or comfortable for patients, especially if performed on separate occasions. Furthermore, if the DCBE
discloses no lesion in a patient with occult fecal blood, colonoscopy is required in any case. If a lesion
is discovered on barium contrast radiography, colonoscopy is needed to obtain biopsy specimens or
remove the lesion and to look for synchronous lesions.65(1309)
For the diagnosis of cancer and polyps larger than 5 mm in diameter, the diagnostic accuracies for
DCBE plus flexible sigmoidoscopy and colonoscopy are approximately equal. Colonoscopy is clearly
superior in the recognition of diminutive polyps and vascular ectasias, whereas the DCBE plus flexible
sigmoidoscopy is more accurate for the diagnosis of diverticulosis. The diagnostic accuracy of
colonoscopy is enhanced by the ability to provide a histologic diagnosis in situations in which
endoscopic findings are equivocal.
The complication rate for colonoscopy as an initial strategy in the evaluation of patients with fecal
occult blood is higher than that associated with the flexible sigmoidoscopy/DCBE approach, although
the number of serious complications with either approach is exceedingly small. In some patients, the
information provided by colonoscopy is incomplete because of inability to reach the cecum in all cases.
However, colonoscopy permits direct inspection of the entire colonic mucosa in more than 90% of
patients. Overall, it is more sensitive and more specific for the diagnosis of neoplasm than is the
competing combination of procedures. In the evaluation of patients with fecal occult blood, an
important determining factor in the choice between colonoscopy and flexible sigmoidoscopy/DCBE
strategies is the therapeutic capability offered by the former approach.
In summary, colonoscopy accomplishes everything that can be accomplished by flexible
sigmoidoscopy/DCBE and more. It is therefore the logical initial procedure in the evaluation of most
patients with fecal occult blood.
Gross Bleeding
Chronic gross rectal bleeding is separated arbitrarily into melena and hematochezia (see Table 271).
Because the color of blood changes with the length of time it resides in the gut lumen, melena is most
likely to come from an upper tract source; lesions in the small bowel or colon that produce melena are
encountered occasionally. Hematochezia (red blood in the stool) may arise from any part of the
intestine, but patients with chronic bleeding from the upper tract seldom pass red blood per rectum.
EGD is required if upper GI bleeding is suspected.
Some patients delay seeking medical attention after seeing blood in the stools. In one study, 29% of
patients who delayed 14 days or more were not worried by the bleeding, they took some other action
before visiting the physician, or they gave other reasons.66(1310)

The anorectum is the most common site of origin of chronic gross rectal bleeding, and hemorrhoids
are the most prevalent anorectal lesions responsible for this category of bleeding. The finding of an
anorectal lesion is not proof that it has been bleeding, however, and the presence of a more proximal
lesion in the colon must be considered.6769(1311) In the study of Goulston et al.,67(1312) a colonic
or proximal rectal site of bleeding was found in 17% of patients in whom a physician had suspected an
anal source based on the patient's history. In the study of Mant et al.,68(1313) blood mixed with feces
predicted the finding of colorectal cancer in 21% of patients with rectal bleeding, but no other clinical
features were helpful in deciding whether to proceed with complete colorectal evaluation of patients 40
years of age or older. The presence of excessive mucus on or within stools and altered bowel habit are
other symptoms that should raise a concern about the presence of colorectal cancer.70(1314)
Melena
Melena (black stools) is the result of oxidation of hematin in the GI tract. Because melena most often
arises in the proximal digestive tract, EGD usually is the first invasive test performed, and, if it is
unrevealing, colonoscopy should be performed. Colorectal abnormalities are found in about 30% of
patients with melena; some of the colonic lesions are carcinomas, usually on the right side of the
colon.1(1315)
Hematochezia
Red blood on the toilet tissue after defecation is observed at some time by one in seven
adults.71(1316) Blood is seen in the toilet or mixed in the stool of 2 to 3% of apparently healthy
people.71(1317) A malignant neoplasm is present in 10% of the latter group, and a neoplastic polyp is
found in 30% of the same individuals.67,72(1318) If the history, physical examination, anoscopy, and
rigid proctosigmoidoscopy do not conclusively establish the source of bleeding, further investigation is
warranted.69(1319) A variety of studies can be used alone or in combination: barium enema
radiographs, other radiographic studies, flexible sigmoidoscopy, colonoscopy, and enteroscopy.
Barium Contrast Radiography
The barium enema was the study of choice for the diagnosis of colonic diseases for many
decades.73(1320) Compared with colonoscopy, barium contrast radiography of the colon probably has
fewer associated complications, the study is more often complete, and this method of imaging is less
expensive. The barium enema misses 6% or fewer of colorectal carcinomas, and it has a high
sensitivity for polyps larger than 1 cm in diameter.74(1321) For these reasons, some have argued that
barium enema should be the initial diagnostic test for patients with chronic hematochezia.7376(1322)
The single- and double-contrast barium enemas both have advantages and disadvantages, but the
consensus is that double-contrast study is preferable in adults.20(1323) This demonstrates fine
mucosal detail and detects small lesions, and therefore the yield is higher in patients with
hematochezia.74(1324) Single-contrast radiographs are more reliable in evaluation of the sigmoid
colon and the cecum, two common sites of neoplastic lesions.1,77(1325)
Other Radiographic Studies
The small intestine can be examined by enteroclysis. In a series of 125 patients with suspected small
intestinal bleeding, the overall positive yield was relatively low at 10%, but the lesions found were of
clinical importance.78(1326) The highest yield of positive studies (20%) was in patients with normal
evaluations of the proximal digestive tract and colon. No correlation was found between the type of
bleeding and the results of the enteroclysis.78(1327) Computed tomography has been used
successfully in evaluation of patients who bleed from diffuse cavernous hemangioma of the
rectosigmoid;79(1328) similar beneficial results have been obtained with MRI.80(1329)
Flexible Sigmoidoscopy
Flexible sigmoidoscopy detects two to six times the number of lesions found on rigid
proctosigmoidoscopy because the length of bowel examined in the former procedure is two to three
times greater.81,82(1330) Flexible sigmoidoscopy has a sensitivity with respect to neoplasms of the
distal colon and rectum of nearly 94%; the positive predictive value in this location is 100% (see
Chapter 91: Flexible Sigmoidoscopy).83(1331)
The main limitation of the flexible sigmoidoscope in the evaluation of patients with rectal bleeding is the
length of the instrument. Only two thirds of polyps and a similar percentage of cancers of the large
intestine are within reach of a 65-cm sigmoidoscope.1(1332) Further, although more lesions are found
with flexible than rigid sigmoidoscopes, fully 50% of colonic cancers that are too proximal to be seen
with the rigid instrument are also too proximal for detection with the flexible sigmoidoscope. Therefore,
the entire colon must be examined by barium enema or colonoscopy in patients who have chronic
hematochezia, no convincing anorectal source, and negative findings on flexible
sigmoidoscopy.1(1333) Fortunately, only a few patients with chronic passage of bright red blood per
rectum have a bleeding site in the proximal colon.84(1334)
Colonoscopy
A lack of suitable control patients is a problem in some studies that compare colonoscopy and barium
enema in the investigation of rectal bleeding. For example, colonoscopy by experts may be contrasted
with radiographic studies of lesser quality. This seems to be the case with one study in which 30 to
40% of significant lesions found at colonoscopy had been missed by barium enema studies.85(1335)
In another study, colonoscopy was preferred over barium enema, although patients received sedation
for the former but not the latter study.86(1336) Even if methodologic flaws such as these are taken into
consideration, however, the data are persuasive that colonoscopy has a higher sensitivity for tumors
larger than 5 mm than does the combination of flexible sigmoidoscopy and double-contrast barium
enema.33(1337) Colonoscopy permits examination of the entire colon in most instances, small flat
lesions such as vascular ectasias can be diagnosed, a tissue diagnosis can be established by
obtaining biopsies, and polypoid lesions can be removed.87(1338) If a neoplasm is seen on flexible
sigmoidoscopy or barium enema, a colonoscopic search for synchronous lesions is nevertheless
required.48, 88(1339) Many lesions discovered by barium enema, such as diverticula, may not be the
source of bleeding, but this determination cannot be made by barium enema.
Several comparison trials of DCBE plus sigmoidoscopy versus colonoscopy have been done in
patients with rectal bleeding.3335(1340) Colonoscopy was performed after DCBE/flexible
sigmoidoscopy in the prospective evaluation of Irvine et al.33(1341) of 71 patients with overt rectal
bleeding. With regard to neoplastic lesions larger than 5 mm in diameter, colonoscopy was found to
have a higher sensitivity, whereas DCBE plus sigmoidoscopy was more sensitive for diverticulosis.
Vascular ectasias, present in three cases, were diagnosed only by colonoscopy. Colonoscopy was
performed to the cecum in 83% of cases. Side effects and complication rates were approximately
equal.
In a prospective trial of DCBE plus flexible sigmoidoscopy (35 cm) versus colonoscopy in patients with
rectal bleeding, Hixson et al.34(1342) performed sigmoidoscopy first, followed by DCBE 1 to 3 weeks
later. Colonoscopy was performed within 4 weeks of the DCBE. Physicians in training performed all
three procedures. Most patients had overt (87.2%) rather than occult bleeding. Only 87 of the 129
patients who entered the study underwent all three procedures. The combination of DCBE and
sigmoidoscopy detected all 7 malignant tumors, but 36% of the polyps 1 cm or more in diameter and
60.5% of those less than 1 cm in diameter were not detected by this strategy.
The diagnostic yield and cost effectiveness of DCBE/flexible sigmoidoscopy and colonoscopy in the
initial evaluation of 388 patients with colonic bleeding were compared in the randomized trial of Rex et
al.35(1343) The combination of DCBE and flexible sigmoidoscopy strategy was more reliable in the
diagnosis of diverticulosis, whereas colonoscopy found more vascular ectasias and polyps less than 9
mm in diameter. For diagnosis of larger polyps, the two approaches were about equal. Colorectal
cancer was more common in patients 55 years of age and older in both arms of the trial. For patients
less than 55 years old, DCBE plus flexible sigmoidoscopy was more cost effective, whereas initial
colonoscopy was more cost effective in patients aged 55 and older.
Although the issue remains controversial, most authorities agree that colonoscopy is superior to the
combination of flexible sigmoidoscopy and DCBE for evaluation of chronic hematochezia not clearly
arising in the anorectum.1,84(1344)
Enteroscopy
The source of bleeding remains obscure despite standard diagnostic tests in about 5% of patients with
chronic GI bleeding.64(1345) These patients are candidates for enteroscopy (see Chapter 50:
Endoscopy of the Small Intestine).5760,89(1346)
Acute Bleeding
The lower digestive tract was the site of bleeding in 18% of the 417 patients with acute GI hemorrhage
in the report of Gostout et al.90(1347) The classification of acute bleeding as moderate or severe (see
Table 271) is arbitrary, but patients tend to separate into those who bleed small amounts and those
who bleed a great deal. The term massive bleeding is sometimes applied to the latter
circumstance.91(1348) In patients with moderate bleeding, time permits bowel preparation before
obtaining diagnostic studies, and more than one type of study may be possible. With severe bleeding,
however, diagnosis must be prompt and usually without benefit of colonic cleansing. The upper GI tract
can be the source of acute severe bleeding per rectum, a possibility that must be considered in the
diagnostic strategy.90,91(1349) In general, if the patient is passing bright red blood and is not in shock,
the bleeding site probably lies in the distal small bowel or colon.
Colonic Causes of Acute Lower Gastrointestinal Bleeding

Diverticular Disease
Between 3 and 5% of patients with colonic diverticular disease are estimated to have noticeable
bleeding per rectum.85,92(1350) Although diverticula are more prevalent in the left colon, 50% of
diverticular bleeding complications occur in the right colon.9,93(1351) Diverticular bleeding arises when
a fecalith erodes a vessel over the apex of a diverticulum or a vessel ruptures at the neck of a
diverticulum.1(1352) The propensity of right-sided diverticula to bleed has not been explained.93(1353)
Diverticular bleeding is arterial and frequently severe.92(1354) Hemorrhage stops spontaneously in 80
to 90% of patients and recurs in about 25% of patients.85,94(1355)
Diverticula are the most common cause of severe colonic bleeding in some reports. In the study of
Farrands and Taylor,94(1356) 28% of patients who required more than 2 units of transfused blood
because of lower tract bleeding had diverticular disease as the source. In most series, however, other
causes are more common.9,93,95(1357) One difficulty with these statistics is the uncertainty that
frequently accompanies the diagnosis of a colonic bleeding site, especially a bleeding diverticulum, in
the absence of direct colonoscopic, angiographic, or surgical proof.2,93(1358)
Vascular Lesions
A confusing array of terms for vascular lesions of the intestines appears in the literature.
Angiodysplasias (vascular ectasias), hereditary telangiectasias, and hemangiomas are considered
here as distinct entities (see also Chapter 84: Vascular Disorders of the Colon). A more detailed
classification based on clinical and pathologic factors has been proposed.96(1359)

Angiodysplasia is a focal submucosal vascular ectasia acquired with aging.95,96(1360) An estimated
25 to 50% of individuals over the age of 60 years have angiodysplasias. The cecum and ascending
colon are the sites of greatest prevalence, but angiodysplasias may be distributed throughout the colon
and also occur in the small bowel and stomach.95,9799(1361) Lesions are multiple in at least 25% of
patients. An association between angiodysplasia and aortic stenosis has been reported, but the validity
of this observation is disputed.95,100,101(1362) Angiography of resected specimens reveals
spider-like dilated blood vessels, and tortuous, dilated veins and venules are seen microscopically in
the submucosa.102,103(1363) This angiographic appearance can be mimicked by primary or
metastatic carcinoma of the colon.104(1364)
Most angiodysplasias never cause symptoms.105107(1365) Bleeding from angiodysplasias is seldom
life threatening because it is venous in nature. A typical acute episode of bleeding requires transfusion
of 2 to 4 units of blood and is not associated with hypotension.1(1366) Angiodysplasia may cause
melena initially, or iron deficiency anemia with a positive FOBT may be the first clinical manifestation.
Bleeding is typically intermittent; the chance of recurrence after an acute episode is approximately
85%.85(1367) The proportion of patients with acute rectal hemorrhage who are bleeding from
angiodysplasias ranges from 2 to 75% in various reports.1,94,97(1368) This enormous variation likely
stems from differences in the usage and interpretation of diagnostic tests rather than real differences in
patient populations. Endoscopic overdiagnosis may occur if intramucosal capillaries that are not true
angiodysplasias are erroneously regarded as vascular lesions.97(1369) The suggestion has been
made that biopsies be obtained from these lesions as a routine measure.108(1370) A positive
histologic diagnosis was obtained in 60% of patients in one study,108(1371) but others maintain that
endoscopic biopsies often provide disappointing results.109(1372)
Hereditary hemorrhagic telangiectasias (Osler-Weber-Rendu syndrome) result in mucosal lesions
throughout the GI tract.96(1373) Multiple hemangiomas of the skin and gut, including the colon and
small bowel, make up the blue rubber bleb nevus syndrome.110(1374) GI bleeding is typical in this
rare condition. Cavernous hemangiomas of the large intestine bleed severely.79,111(1375)
Neoplasms
Bleeding that occurs from colorectal cancers and benign polyps is chronic in nature. These lesions
may bleed to the point that transfusions are required, but it is very unusual for colonic neoplasms to
cause acute severe hemorrhage.1,94(1376) Bleeding can be a complication of colonoscopic
polypectomy.110(1377)
Inflammatory Bowel Disease
Inflammatory bowel disease accounted for 16% of patients with lower GI bleeding in the series of
Farrands and Taylor.94(1378) Bloody diarrhea is characteristic of ulcerative colitis, but sudden severe
hemorrhage is uncommon.1,112(1379) A third of patients with Crohn's colitis have gross blood in the
feces. Acute severe hemorrhage was encountered in only 1.9% of patients with Crohn's colitis and
0.7% of those with Crohn's ileitis in the series of 1526 patients of Robert et al.113(1380) The precise
bleeding points were located in only 2 of the 26 bleeding episodes, and both were in the ileocecal area.
Bleeding subsided without surgery in 48% of patients, but it recurred in 30% of this group.
Ischemic Colitis
Spontaneous or postoperative colonic ischemia may cause bloody diarrhea or gross blood in the
stools. Endoscopic examination reveals blood in the lumen and a gray discoloration of the mucosa in
more severe cases.114116(1381) Severe bleeding is rare.76(1382) Abdominal pain, tenderness, and
systemic manifestations are clues to the diagnosis (see also Chapter 84: Vascular Disorders of the
Colon).
Ulcers
Ulcers in the colon or rectum from any cause may bleed acutely. These uncommon lesions include
solitary rectal ulcer, cecal ulcer,117(1383) stercoral ulcer, and ulcers associated with infectious
diseases such as typhoid and amebiasis.9,94,118(1384) Renal transplant recipients develop idiopathic
colonic ulcers.119(1385) Dieulafoy's ulcer of the colon has been reported to cause acute
hemorrhage.120(1386)
Radiation Bowel Injury
Acute or chronic radiation injury to the large or small bowel may cause acute lower GI
hemorrhage.121,122(1387) Mucosal erosions may be focal or diffuse.123(1388)
Other Colorectal Lesions
Aortoenteric and aortocolonic fistulas are rare, but bleeding is catastrophic.1(1389) A ruptured splenic
artery aneurysm can bleed massively into the splenic flexure of the colon.124(1390) The splenic artery
also is the source of bleeding in rare cases of erosion of a pancreatic pseudocyst into the
colon.125(1391) Mucosal injury from chemotherapeutic agents, diffuse bleeding from anticoagulation,
postoperative anastomotic hemorrhage, and neoplastic lesions such as angiolipoma account for a few
cases.92,126(1392) Immunodeficiency as a consequence of disease or therapy places patients at risk
of mucosal lesions that can bleed. About 1% of renal transplant recipients bleed from infectious colitis,
uremic colitis, or other causes.119(1393) Colonic or rectal varices bleed in patients with portal
hypertension,127,128(1394) possibly more frequently after sclerosis of esophageal
varices.92,129(1395) Varices around an ileostomy or colostomy are a special case of this general
problem.92,130(1396) Kozarek et al.131(1397) performed colonoscopy in 20 patients with portal
hypertension and lower tract bleeding, including 10 patients with hematochezia and 9 patients with
anemia plus a positive FOBT. The most common finding, noted in 14 patients, was described as
mucosal abnormalities that resembled angiodysplasia; colorectal varices were relatively uncommon.
Small Bowel Causes
Rare abnormalities of the small intestine produce acute severe lower GI bleeding (see Table
271).132(1398) Vascular lesions are described earlier; angiodysplasias can occur in the small
bowel.95,97,133(1399) Dieulafoy's ulcer has been described in the small intestine.134(1400) Jejunal
diverticula cause hemorrhage in some patients. Ectopic gastric tissue in an ileal duplication or a
Meckel's diverticulum is a cause of bleeding in young patients.135(1401) Kaposi's sarcoma of the
ileum may hemorrhage massively.136(1402) Polyps and other tumors are responsible for some cases
of bleeding, and leiomyosarcoma of the small intestine has been reported to cause recurrent
bleeding.137(1403) Iliac arterial-enteric fistulas are rare complications of pelvic irradiation.138(1404)
Diagnostic Evaluation of Acute Bleeding

The speed and thoroughness of the diagnostic evaluation for acute lower GI bleeding depend on the
rate of bleeding, the volume of blood lost, and whether bleeding appears to have stopped. It is
sometimes difficult to determine whether a patient is still bleeding; in this situation it should be
assumed that bleeding persists.
Diagnostic criteria based on history, physical examination, and barium enema studies are unreliable
with regard to precise localization of a bleeding site in the lower GI tract. Diagnosis depends on other
maneuvers, including endoscopy and angiography. A plan for management of lower GI bleeding is
shown in Figure 271. Similar schemes have been proposed.139(1405)
(1406)Figure 271. Plan for the management of acute lower gastrointestinal bleeding.
EGDesophagogastroduodenoscopy; UGIupper gastrointestinal. (Modified from Schrock
T. Colonoscopic diagnosis and treatment of lower gastrointestinal bleeding. Surg Clin North
Am 1989; 69:130925.)
A patient with acute lower GI bleeding should be resuscitated and stabilized if possible before
undergoing endoscopy or radiographic studies. Occasionally, bleeding is so severe that the patient
cannot be stabilized, and an investigation into the cause of bleeding must be undertaken despite the
volatile nature of the patient's condition.1(1407) Intravenous fluids and transfusions are given as
needed. Coagulation parameters should be measured and deficits corrected, and associated medical
conditions should be identified and treated.
Digital rectal examination and anoscopy are required. If an anorectal lesion is seen, it is necessary to
be certain that it is the only source of bleeding before discontinuing the evaluation. Rigid
sigmoidoscopy usually is available in the emergency room; it requires no bowel preparation and can be
performed immediately for whatever information it may provide. Large clots are evacuated, and the
rectal mucosa is inspected. Often, the only finding is blood coming from above, but this is a useful
observation.
The GI tract proximal to the ligament of Treitz is the source of hemorrhage in about 10% of patients
with acute severe rectal bleeding.140(1408) A nasogastric tube is inserted, and the aspirate is
inspected for blood and bile and tested for occult blood if none is present grossly. It is established that
the gastric aspirate may not contain blood in patients with upper GI bleeding. The presence of bile in
the aspirate offers some reassurance that duodenal contents are being sampled. No patient with a
guaiac-negative, bile-containing gastric aspirate had an upper GI tract bleeding site in the study of
Jensen and Machicado.140(1409) EGD should be performed on an urgent basis if blood is found in the
gastric aspirate or if no bile is present; in other patients, endoscopy is indicated if the history or findings
suggest an upper tract source of bleeding. In some hospital populations, esophageal, gastric, and
duodenal lesions are so common that endoscopy should be performed routinely in nearly every patient
with acute severe bleeding per rectum before the evaluation is focused on the lower intestines.1(1410)
Slow Bleeding
If bleeding appears to stop or if the bleeding rate is slow, the best approach to diagnosis is cleansing of
the colon followed by urgent colonoscopy.140(1411) Polyethylene glycol solution is given by mouth or
through a nasogastric tube to prepare the colon.140(1412) This procedure requires an average of 3
hours (2 to 7 hours), and it should be continued until the rectal effluent is clear. Metoclopramide (10 mg
orally or intravenously) may promote gastric emptying and permit administration of the entire required
volume of solution (4 to 15 L; mean 5.5 L). Colonoscopy is performed in an endoscopy unit if the
patient is stable enough for transfer, but it also can be done at the bedside if necessary. Often it is
possible to schedule the procedure for a convenient time (e.g., first thing in the morning) when
experienced personnel are available to assist.
The bleeding site was identified using colonoscopy and EGD in 94% of patients with acute lower GI
bleeding in the large series of Jensen and Machicado (Table 274).140(1413) Other investigators
report success rates of 25 to 90%.1,85,94,141(1414) The finding alone of a lesion at colonoscopy does
not prove that it is the bleeding site.93(1415) Suggested criteria for diagnosis of a bleeding lesion are
(1) active bleeding; (2) adherent clot in a single diverticulum or on an ulcerative lesion, resistant to
washing, with fresh blood nearby and no other lesions present; and (3) a nonbleeding visible vessel in
an ulcer in the absence of other lesions that could be responsible for bleeding.140(1416)
TABLE 274
Final Diagnosis for 80 Patients With Severe
Hematochezia
SITE OF LESIONS
NO. OF PATIENTS (%)
Colonic
59(74)
Angiomata
24(30)
Diverticulosis
13(17)
Active bleeding
6(8)
Adherent clot
7(9)
Polyps or cancer
9(11)
Focal colitis or ulcers
7(9)
Rectal lesions
3(4)
Bleeding polyp stalk
2(2)
Endometriosis
1(1)
Upper gastrointestinal
9(11)
Small bowel*
7(9)
No site found
5(6)
Modified from Jensen D, Machicado G. Diagnosis and treatment of severe
hematochezia: The role of urgent colonoscopy after purge. Gastroenterology 1988;
95:156974.
* A diagnosis of presumed small bowel site of bleeding was made when
panendoscopy and colonoscopy were negative but fresh blood or clots or both were
coming through the ileocecal valve.
Bleeding stops spontaneously in 90% of patients before transfusion requirements exceed 2

units.94(1417) Preparation of the colon and urgent colonoscopy are therefore feasible in most
patients.142(1418) Treatment of certain types of bleeding lesions may be possible at colonoscopy, an
important additional benefit of this procedure. Complications of urgent colonoscopy for bleeding include
fluid overload from the polyethylene purge; in one report, 4% of patients had clinically significant fluid
overload using earlier formulations of polyethylene glycol.140(1419) Perforation and other
colonoscopic complications are rare in the hands of experienced endoscopists.143(1420)
If colonoscopy is performed to the cecum, no blood is present in the colon, no lesions are seen, and
clinically bleeding appears to have stopped, the diagnostic evaluation probably can be terminated. A
barium enema is sometimes performed if colonoscopy does not reach the cecum and a repeat attempt
is unlikely to be more successful. However, it is necessary to be certain that bleeding has ceased lest
the presence of barium in the colon interfere with the performance of other studies. Furthermore, the
barium enema may demonstrate lesions but does not identify which one(s) has been bleeding. The
barium enema does not detect angiodysplasias because they are flat. Claims that barium has a
hemostatic role in colonic bleeding have never been confirmed.1(1421)
The small bowel can be examined by preoperative enteroscopy in patients with slow bleeding,
particularly intermittent or recurrent bleeding (see Chapter 50: Endoscopy of the Small
Intestine).5961,89,144(1422)
Rapid Bleeding
Alternatives for evaluation of patients who are bleeding rapidly from the lower GI tract are colonoscopy
(usually without preparation of the colon), angiography, and exploratory laparotomy with intraoperative
endoscopy.145(1423) The role of radionuclide scans has declined in recent years. EGD has been
recommended in this group of patients.90(1424)
Colonoscopy
Concerns about restricted visibility and the risk of complications make some endoscopists reluctant to
perform colonoscopy in the presence of severe active bleeding.20(1425) It is possible, however, for a
skillful and tenacious endoscopist to perform colonoscopy safely and beneficially in these
circumstances.85,141(1426) Anoscopy must be done first, and rigid sigmoidoscopy usually is
performed in the emergency room, but little is to be gained by preliminary flexible sigmoidoscopy.
Bowel preparation is precluded by profuse bleeding, but blood is a cathartic and the colon may be free
of stool. A colonoscope with a large accessory channel should be used because it will be necessary to
remove stool, blood, and clot from the colon by suction.1(1427) The presence of blood demands more
cautious advancement of the instrument than usual. Overinflation of the colon should be avoided
because it may precipitate or aggravate bleeding.
The diagnostic findings when emergency colonoscopy is performed for acute severe bleeding fall into
five categories (Table 275): (1) Bright red blood issuing from a specific point in the colon or from a
particular lesion is the most conclusive evidence of a bleeding site. (2) Mucosa covered in patches with
clotted blood may indicate that bleeding has stopped. (3) When the mucosal surface is found to be free
of blood but the colon contains fresh blood distal to this point, this suggests that the site of bleeding is
distal to the region in which the mucosal surface is clean. (4) Clinically significant lesions may be
discovered that are not necessarily the site of recent hemorrhage. (5) Diffuse mucosal bleeding from
colitis or other causes should be obvious. The bleeding site can be localized in 50% of patients
undergoing emergency colonoscopy for acute severe bleeding; in 70% of patients the region of the
colon from which bleeding arises can be approximated.146(1428)
Diagnostic Categories in
Emergency Colonoscopy for Acute Severe
Bleeding
TABLE 275
Acutely bleeding lesions

Fresh blood in an area, nonbloody contents proximally
Fresh adherent clot
Fresh blood in an area, unable to prove nonbloody contents
proximally
Failure to localize bleeding
Modified from Forde K, Treat M. Colonoscopy for lower
gastrointestinal bleeding. In Dent T, Strodel W, Turcotte J, et
al. (eds): Surgical Endoscopy. Chicago: Year Book Medical,
1985; 26174.
Radionuclide Scans
Radionuclide bleeding scans have earned a reputation for misleading results, and many clinicians no
longer include them in the algorithm for evaluation of patients with acute lower GI bleeding.147(1429)
Radionuclide scans require intravenous injection of 99mTc sulfur colloid or 99mTc-labeled autologous
red blood cells.148,149(1430)
Sulfur colloid given intravenously is cleared rapidly by the reticuloendothelial system, but extravasation
into the lumen may be detectable with bleeding rates as slow as 0.5 to 0.1 ml/min.105(1431) Whether
bleeding arises in the colon or small intestine cannot be determined with certainty from the scan,
although observed movement of the extravasated radioactive material on sequential images may be
helpful. Bleeding that arises in the hepatic and splenic flexures may be masked by uptake in the liver
and spleen, respectively.
The red cell scan uses autologous erythrocytes that are labeled in vitro with 99mTc and injected into
the patient.147(1432) Sequential images of the anterior abdomen are obtained at 5-min intervals for
the first 30 min and every few hours thereafter up to 24 hr. Focal activity indicates a bleeding site. Only
the early images, those obtained in the first hour, are important in patients with active
hemorrhage.1(1433) The 99mTc-labeled red blood cell scan is less sensitive than the sulfur colloid
scan when the rate of bleeding is slow, but it appears to be more reliable.105(1434) Technetium sulfur
colloid can be injected through a selectively positioned angiography catheter; this method can be very
useful for detection of sites of slow bleeding.
Radionuclide scans have an accuracy of 30 to 90%.85(1435) The ability of the 99mTc red cell scan to
identify the site of severe colonic bleeding was studied retrospectively by Ryan et al.126(1436) The
bleeding site was localized in 12 of 29 patients (41%) bleeding from the lower GI tract. It was correctly
identified in all 9 patients who continued to bleed and required emergency surgery. In the study of
Hunter and Pezim,147(1437) 26% of red cell scans were positive, but 42% of patients who underwent
a surgical procedure based on localization of bleeding by the scan had the wrong operation performed.
Some radiologists prefer that a positive radionuclide scan be obtained before performing angiography
because the scan is more sensitive and less invasive.9(1438) Clinicians who have little faith in the
radionuclide scan often request angiography immediately for the patient in whom colonoscopy cannot
be performed or is unsuccessful.150(1439) A compromise position might be to draw blood for 99mTc
red cell labeling as the patient enters the hospital. Initial resuscitation and evaluation can proceed, and,
when the labeled cells are ready an hour later, they are injected and the scan is performed if
appropriate.1(1440) This approach is feasible only if the patient is stable enough to be moved to the
nuclear medicine department.
Angiography
Visceral angiography offers the potential for both diagnosis and therapy. Depending in part on the
selectivity of the catheter position, extravasation of contrast material may be seen if the arterial
bleeding rate is at least 0.5 ml/min.9(1441) Moncure et al.9(1442) found that the bleeding site could be
localized in 69% of patients with colonic bleeding and 50% of those bleeding from the jejunum or ileum.
Other favorable studies report that angiography is diagnostic in 40 to 90% of
patients.94,108,151153(1443) Poor results also have been reported; the rate of success was only
14% in the study of Jensen and Machicado140(1444) and 17% in that of Gostout et al.154(1445)
Patient characteristics and the enthusiasm and skill of the radiologist account for some of the disparity
in these reports.
An important limitation of angiography is that bleeding usually must be active if the source is to be
identified.9(1446) Injection of anticoagulants or fibrinolytics at the time of angiography has been
reported to improve results, but this extreme measure has not been adopted widely.155(1447) Some
lesions are apparent on angiography when they are not bleeding; angiodysplasia, for example, may
appear as a vascular tuft and an early-filling vein in the right colon.9(1448) However, it is always
possible that a nonbleeding lesion is a coincidental finding.156(1449)
Complications of emergency angiography include arterial thrombosis, emboli, and renal failure related
to the use of the contrast medium. Complication rates vary, but figures as high as 9% have been
reported.140(1450) The need to mobilize the angiography team and to transport the patient to the
radiology department is a disadvantage of angiography. Monitoring of seriously ill patients is not always
optimum in these circumstances.
Colonoscopy and angiography are complementary. Most clinicians, certainly most endoscopists, prefer
colonoscopy if they or a colleague has experience with the procedure in this clinical setting.
Angiography is an alternative if colonoscopy cannot be performed or is unsuccessful, provided, of
course, that a skilled and experienced radiology team is readily available.
Therapy of Acute Bleeding

Acute lower GI hemorrhage stops spontaneously in 75 to 90% of patients.94,157(1451) If active
bleeding is observed during colonoscopy or a site is demonstrable on angiography, treatment is
indicated. Therapy may be accomplished by colonoscopic, angiographic, and surgical methods, alone
or in combination.
Colonoscopic Methods
Endoscopic procedures for control of hemorrhage were performed in 39% of patients who underwent
urgent colonoscopy for bleeding in the series of Jensen and Machicado.140(1452) Colonoscopic
methods widely used to treat acute or chronic bleeding lesions include bipolar electrocoagulation
(BICAP probe), heater probe application, and laser phototherapy (see Chapter 28: Thermal Contact
Methods for Endoscopic Hemostasis, and Chapter 29: Laser Therapy).158(1453) Reapplication of the
polypectomy snare is frequently used to control postpolypectomy hemorrhage.110(1454)
Monopolar electrocoagulation may cause deep necrosis, perforation, or delayed bleeding.159(1455)
The so-called hot biopsy forceps is a monopolar device that is used to obtain a tissue sample for
histologic examination as the biopsy site is simultaneously coagulated. Because air insufflation during
colonoscopy further stretches the thin colonic wall, particularly in the right colon, the risk of perforation
when the hot biopsy forceps is used at colonoscopy is substantial.160(1456) Use of the hot biopsy
forceps cannot be recommended because of this danger.
The BICAP and heater probes (Olympus America, Inc.) are equally safe and effective in controlling
arterial bleeding from the gut.158,159,161(1457) Although the neodymium:yttrium-aluminum-garnet
(Nd:YAG) and argon lasers can be used for this purpose, the majority of endoscopists prefer
electrocoagulation methods for most situations.154,158,159,162164(1458)
Angiodysplasias are treated readily by any of the previously mentioned methods.97,98,164166(1459)
In the series of Trudel et al.,98(1460) the overall success rate for endoscopic coagulation of colonic
vascular ectasias was 68%. Half of the patients in this series had recurrent bleeding after one
treatment session; a third of these episodes were controlled at a second colonoscopic session. The
cumulative probability of remaining free of recurrent bleeding was 61% at 54 months after Nd:YAG
laser photocoagulation of angiodysplasias in various parts of the GI tract in the study of Naveau et
al.99(1461) In this study, the number of angiodysplasias, the presence of coagulation disorders, and
age were independent prognostic parameters.
Asymptomatic (i.e., nonbleeding) angiodysplasias do not require treatment if they are encountered
when colonoscopy is performed for reasons other than bleeding. If they are found in a patient who has
bled recently, endoscopic treatment is probably appropriate, but often the site of bleeding is
elsewhere.140(1462) Administration of estrogen or progesterone has been suggested as effective
treatment for severely bleeding vascular malformations that cannot be managed by direct
means.167(1463)
Endoscopic coagulation of bleeding colonic diverticula has been described,168(1464) but few
endoscopists are willing to attempt this maneuver because the risk of perforation of the thin-walled
diverticulum is great.1(1465) Injection therapy for colonic diverticular bleeding has also been
reported.169(1466) Bleeding due to diffuse radiation injury of the mucosa responds well to treatment
with the argon laser.121,122(1467)

Angiographic Methods
If a bleeding site is demonstrated by selective angiography, infusion of vasopressin results in control of
bleeding in 35 to 90% of cases.1,151(1468) Bleeding reoccurs in 50% of patients when infusion of the
vasoconstrictor is stopped. Some patients cannot tolerate the intravenous infusion of vasopressin. The
feeding artery can be embolized at angiography with microcoils, polyvinyl alcohol particles, or
combinations of these and other materials.170(1469) Necrosis of the intestinal wall by occlusion of end
arteries is minimized by highly selective cannulation of the bleeding artery.
Surgical Methods
Operation should be performed if the patient continues to bleed despite other measures.171(1470) If
transfusion requirements exceed 4 units in the first 24 hours, a 50% chance exists that operation will
be required.94(1471) In patients with hemorrhage from Crohn's disease, bleeding subsides without
surgery in 50% of patients but recurs in 30%, so removal of diseased bowel at the time of the first
episode of severe hemorrhage is advised.113(1472)
The patient should be positioned on the operating table with the legs in stirrups to facilitate
intraoperative colonoscopy. The entire abdomen should be explored. With regard to the site of
bleeding, the surgeon must be wary of incorrect preoperative information from angiography or even
from colonoscopy. Also, multiple lesions are possible.156(1473) Intraoperative methods for
identification of the source of hemorrhage are listed in Table 276.171(1474) These steps are divided
arbitrarily into primary maneuvers (used most of the time) and secondary methods (used occasionally).
Sometimes the bleeding lesion is obvious. If the problem is not readily apparent, the small intestine
should be transilluminated to look for small vascular lesions.171(1475) An operating room light or a
headlamp on an assistant may be sufficient.
TABLE 276
Intraoperative Localization of Intestinal
Bleeding
PRIMARY METHODS
Gross inspection
Transillumination
Endoscopy
Colonoscopy and enteroscopy
Peroral enteroscopy
SECONDARY METHODS
Compression between glass slides
Enterotomy
Angiography
Scintigraphy
Doppler ultrasonography
Venous oxygen measurement
Intraoperative Endoscopy
If the cause of bleeding has not been identified after gross inspection and transillumination,
intraoperative endoscopy should be performed.57,58,144,172,173(1476) The colon is usually
examined first. The colonoscope is inserted per rectum and advanced with the help of the surgeon.
The mucosa should be inspected on insertion rather than withdrawal because trauma is inevitable and
the resultant mucosal tears and contusions may mimic genuine lesions.57(1477) If the view of the
lumen is obscured by stool or blood, the colon can be cleansed by instillation of saline through a Foley
catheter inserted directly into the cecum or through the appendix.174(1478) Three liters is sufficient in
most situations. The distal small bowel is examined if a colonic source is not found. The instrument is
guided through the ileocecal valve into the ileum in 10-cm segments as the mucosa is
examined.57(1479) A suture is placed to mark the proximal extent of endoscopy from below.
Peroral endoscopy of the upper tract and small bowel is performed if the preceding maneuvers fail to
disclose the cause of bleeding.60,62,63,144(1480) The surgeon pleats the intestine on the instrument.
As in the colon, the mucosa should be inspected as the instrument is advanced. Otherwise, small
areas of trauma may be mistaken as vascular lesions. With the operating room darkened, endoscopic
transillumination may demonstrate actual vascular lesions in the wall of the small intestine.
Enteroscopy also can be performed by placing a colonoscope into the small bowel through one or
more enterotomies.58,171(1481) A sheet of plastic excludes the nonsterile instrument from the
surgical field. This type of enteroscopy may allow complete examination of the midsegment of the
small bowel that could not be reached from above or below.123(1482) Intraoperative enteroscopy may
be applied to patients with chronic as well as acute bleeding, as discussed earlier.57,58,132(1483)
Other Methods
Secondary intraoperative diagnostic methods include compression of small intestine between two
glass slides, a maneuver that has been largely replaced by intraoperative enteroscopy. Unguided
enterotomy or colotomy is seldom helpful and contaminates the operative field with bacteria.
Intraoperative angiography may be effective if a bleeding site has been demonstrated
preoperatively.171(1484) The angiographic catheter is left in place, and during operation an angiogram
is obtained with the patient on the operating table with segments of small intestine marked by metal
clips. Alternatively, methylene blue can be injected intraoperatively through a highly selective
angiographic catheter that has been placed preoperatively.132(1485) Another technique involves direct
puncture of a segmental arterial branch with a 21-gauge butterfly needle; an x-ray cassette in a sterile
cover is placed behind the segment of bowel to be imaged.171(1486) Intraoperative scintigraphy has
met with some success.175(1487) Doppler ultrasonography and selective measurement of mesenteric
venous oxygen to detect arteriovenous malformations are occasionally of benefit.171(1488)
Surgical Procedures
If a lesion is seen during abdominal exploration, the portion of small bowel or colon that contains the
lesion is resected. Nearly all patients who undergo operation for Crohn's disease are treated by colonic
resection; only 9% of such patients have recurrent hemorrhage.113(1489) Multiple angiodysplasias in
different portions of the small bowel can be suture-ligated from the serosal side.176(1490) Clusters of
lesions in the small bowel may be resected. Colonic angiodysplasias should be resected if endoscopic
methods do not control bleeding. The presence of angiodysplasias on the right side of the colon and
diverticula on the left can present a dilemma, but, in general, if the angiodysplasias appear to be the
cause of bleeding, an extended right colectomy is sufficient.177,178(1491) Total colectomy and
ileorectal anastomosis for bleeding from a colonic source that cannot be localized are seldom required
today.179(1492)
Mortality Rates
Mortality rates from lower GI hemorrhage vary according to the severity of bleeding, and most reports
do not provide sufficient detail to allow accurate determinations. Estimates are as high as 12%, and in
series of patients subjected to operation, as many as 30% may succumb to bleeding or treatment
measures.1,2,111,179(1493) Bender et al.179(1494) found that the overall mortality rate was 27%
among 49 patients who underwent total colectomy for lower GI bleeding7% of those performed
electively or urgently and 30% of those performed on an emergency basis. Mortality rates since the
wide application of colonoscopy, angiography, and other diagnostic maneuvers are in the range of 3 to
5%.1,126(1495) Gostout et al.90(1496) reported no deaths among 29 patients with bleeding
diverticula. In the series of Lau et al.132(1497) of patients with small intestinal sources of bleeding, the
mortality rate was 6%. A specialized team of physicians, surgeons, nurses, and GI assistants can
reduce the morbidity and mortality of patients with acute lower GI bleeding.90(1498)
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Schrock T. Colonoscopic diagnosis and treatment of lower gastrointestinal bleeding. Surg Clin
North Am 1989;69:130925.
Forde K. Lower gastrointestinal bleeding: The magnitude of the problem. In Sugawa C,
Schuman B, Lucas C (eds). Gastrointestinal Bleeding. New York: Igaku-Shoin, 1992;1325.
Waye J. Diagnosis and treatment of lower intestinal bleeding. In Tytgat G, Blankenstein MV
(eds). Current Topics in Gastroenterology and Hepatology. Stuttgart, New York: Thieme,
1990;2331.
Potter G, Sellin J. Lower gastrointestinal bleeding. Gastroenterol Clin North Am
1988;17:34156.
Rosen A, Fleischer D. Lower GE bleeding: Updated diagnosis and management. Geriatrics
1989;44:4960.
Reinus J, Brandt L. Upper and lower gastrointestinal bleeding in the elderly. Gastroenterol
Clin North Am 1990;19:2293318.
Richardson J. Management of acute and chronic gastrointestinal bleeding distal to the
ligament of Treitz. Perspect Gen Surg 1990;1:324.
Gendler S, Faisal M, Holt P. Evaluation of primary and secondary gastrointestinal bleeders.
Arch Intern Med 1989;149:16346.
Moncure A, Tompkins R, Athanasoulis C, et al. Occult gastrointestinal bleeding: Newer
techniques of diagnosis and therapy. Adv Surg 1989;22:14177.
Decosse J. Early cancer detection: Colorectal cancer. Cancer 1988;62:178790.
Winawer S, Kerner J. Sigmoidoscopy: Case finding versus screening [Editorial].
Van Dam J, Bond JH, Sivak MV Jr. Fecal occult blood screening for colorectal cancer: The
current state of the art. Arch Intern Med 1995;155:2389402.
Simon JB. Occult blood screening for colorectal carcinoma: A critical review.
Hardcastle JD, Thomas WM, Chamberlain J, et al. Randomised, controlled trial of faecal
occult blood screening for colorectal cancer. Results for first 107,349 subjects. Lancet
1989;1(8648): 11604.
Bertario L, Spinelli P, Gennari L, et al. Sensitivity of Hemoccult test for large bowel cancer in
high-risk subjects. Dig Dis Sci 1988;33:60913.
Kimmig JM, Strauch M, Hallen M. Negative Haemoccult test in malignant and premalignant
lesions of the colon. Validation of the Haemoccult test with total colonoscopy. Endoscopy
1989;21:13640.
Mandel JS, Bond JH, Bradley M, et al. Sensitivity, specificity, and positive predictivity of the
Hemoccult test in screening for colorectal cancers. The University of Minnesota's Colon
Cancer Control Study. Gastroenterology 1989;97:597600.
Ahlquist DA, McGill DB, Fleming JL, et al. Patterns of occult bleeding in asymptomatic
colorectal cancer. Cancer 1989;63:182630.
Foutch P, Manne R, Sanowski R, et al. Risk factors for blood loss from adenomatous polyps
of the large bowel: A colonoscopic evaluation with histopathological correlation. J Clin
Schrock T. Colon and rectum: Diagnostic techniques. In Condon R (ed). Shackelford's
Surgery of the Alimentary Tract. 3rd ed. Vol IV: Colon and Anorectum. Philadelphia: WB
Saunders, 1991;2237.
21. Mandel JS, Bond JH, Church TR, et al. Reducing mortality from colorectal cancer by
screening for fecal occult blood. N Engl J Med 1993;328:136571.
22. Lang CA, Ransohoff DF. Fecal occult blood screening for colorectal cancer: Is mortality
reduced by chance selection for screening colonoscopy?. JAMA 1994;271:10113.
23. Longstreth G. Checking for "the occult" with a finger: A procedure of little value. J Clin
24. Nord HJ, Zuckerman G. Controversies, dilemmas, and dialogues. The workup of a patient
with a positive fecal occult blood test: One procedure? Two procedures? Am J Gastroenterol
1991;86:5425.
25. Goldschmeidt M, Ahlquist D, Wieand H, et al. Measurement of degraded fecal
hemoglobin-heme to estimate gastrointestinal site of occult bleeding: Appraisal of its clinical
utility. Dig Dis Sci 1988;33:6058.
26. Miyoshi H, Ohshiba S, Asada S, et al. Immunological determination of fecal hemoglobin and
transferrin levels: A comparison with other fecal occult blood tests. Am J Gastroenterol
1992;87:6773.
27. Mant D, Fitzpatrick R, Hogg A, et al. Experiences of patients with false positive results from
colorectal cancer screening. Br J Gen Pract 1990;40:4235.
28. Rex DK, Lehman GA, Hawes RH, et al. Screening colonoscopy in asymptomatic average-risk
persons with negative fecal occult blood tests. Gastroenterology 1991;100:647.
29. Kewenter J, Engaras B, Haglind E, Jensen J. Value of retesting subjects with a positive
Hemoccult in screening for colorectal cancer. Br J Surg 1990;77:134951.
30. Adamsen S, Kronborg O, Hage E, Fenger C. Reproducibility and diagnostic value of
Hemoccult-II test. A colonoscopic evaluation in asymptomatic patients. Scand J Gastroenterol
1985;20:10737.
31. Ahlquist DA, Klee GG, McGill DB, Ellefson RD. Colorectal cancer detection in the practice
setting. Impact of fecal blood testing. Arch Intern Med 1990;150:10415.
32. Reilly JM, Ballantyne GH, Fleming FX, et al. Evaluation of the occult blood test in screening
for colorectal neoplasms. A prospective study using flexible endoscopy. Am Surg
1990;56:11923.
33. Irvine E, O'Connor J, Frost R, et al. Prospective comparison of double contrast barium enema
plus flexible sigmoidoscopy v colonoscopy in rectal bleeding: Barium enema vs colonoscopy
in rectal bleeding. Gut 1988;29:118893.
34. Hixson LJ, Sampliner RE, Chernin M, et al. Limitations of combined flexible sigmoidoscopy
and double contrast barium enema in patients with rectal bleeding. Eur J Radiol
1989;9:2547.
35. Rex DK, Weddle RA, Lehman GA, et al. Flexible sigmoidoscopy plus air contrast barium
enema versus colonoscopy for suspected lower gastrointestinal bleeding. Gastroenterology
1990;98:85561.
36.
Feczko PJ, Halpert RD. Reassessing the role of radiology in Hemoccult screening. AJR
1986;146:697701.
37. Farrands PA, Vellacott KD, Amar SS, et al. Flexible fiberoptic sigmoidoscopy and
double-contrast barium-enema examination in the identification of adenomas and carcinoma
of the colon. Dis Colon Rectum 1983;26:7257.
38. Kewenter J, Jensen J, Boijsen M, et al. Perception errors with double-contrast enema after a
positive guaiac test. Gastrointest Radiol 1987;12:7982.
39. Bolin S, Franzen L, Nilsson E, Sjodahl R. Carcinoma of the colon and rectum. Tumors missed
by radiologic examination in 61 patients. Cancer 1988;61:19992008.
40. Brady AP, Stevenson GW, Stevenson I. Colorectal cancer overlooked at barium enema
examination and colonoscopy: A continuing perceptual problem. Radiology 1994;192:3738.
41. Jensen J, Kewenter J, Asztely M, et al. Double contrast barium enema and flexible
rectosigmoidoscopy: A reliable diagnostic combination for detection of colorectal neoplasm.
Br J Surg 1990;77:2702.
42. Barry MJ, Mulley AG, Richter JM. Effective workup strategy on the cost-effectiveness of fecal
occult blood screening for colorectal cancer. Gastroenterology 1987;93:30110.

43. Hoff G, Foerster A, Vatn MH, et al. Epidemiology of polyps in the rectum and colon. Recovery
and evaluation of unresected polyps 2 years after detection. Scand J Gastroenterol
1986;21:85362.
44.
Grossman S, Milos ML, Tekawa IS, Jewell NP. Colonoscopic screening of persons with
suspected risk factors for colon cancer: II. Past history of colorectal neoplasms. 96:299306
1989;Gastroenterology.
45. Urbanski SJ, Haber G, Kortan P, Marcon NE. Small colonic adenomas with adenocarcinoma.
A retrospective analysis. Dis Colon Rectum 1988;31:5861.
46. Church JM, Fazio VW, Jones IT. Small colorectal polyps. Are they worth treating? Dis Colon
Rectum 1988;31:503.
47. Aldridge MC, Sim AJ. Colonoscopy findings in symptomatic patients without x-ray evidence of
colonic neoplasms. Lancet 1986;2:833834.
48. Guillem J, Forde K, Treat M, et al. The impact of colonoscopy on the early detection of colonic
neoplasms in patients with rectal bleeding. Ann Surg 1987;206:60611.
49.
Glick SN, Teplick SK, Balfe DM, et al. Large colonic neoplasms missed by endoscopy. AJR
1989;152:5137.
50. Masel H, Masel JP, Casey KV. A survey of colon examination technique in Australia and New
Zealand with a review of complications. Australas Radiol 1971;15:1407.
51.
Seaman WB, Wells WJ. Complications of the barium enema. Gastroenterology
1965;48:72837.
52. Eddy DM. Screening for colorectal cancer. Ann Intern Med 1990;113:37384.
53. Zuckerman G, Benitez J. A prospective study of bidirectional endoscopy (colonoscopy and
upper endoscopy) in the evaluation of patients with occult gastrointestinal bleeding. Am J
54. Hsia PC, al-Kawas FH. Yield of upper endoscopy in the evaluation of asymptomatic patients
with Hemoccult-positive stool after a negative colonoscopy. Am J Gastroenterol
1992;87:15714.
55. Chen YK, Gladden DR, Kestenbaum DJ, Collen MJ. Is there a role for upper gastrointestinal
endoscopy in the evaluation of patients with occult blood-positive stool and negative
colonoscopy?. Am J Gastroenterol 1993;88:20269.
56. Thomas WM, Hardcastle JD. Role of upper gastrointestinal investigations in a screening
study for colorectal neoplasia. Gut 1990;31:12947.
57. Flickinger E, Stanforth A, Sinar D, et al. Intraoperative video panendoscopy for diagnosing
sites of chronic intestinal bleeding. Am J Surg 1989;157:13744.
58. Ress AM, Benacci JC, Sarr MG. Efficacy of intraoperative enteroscopy in diagnosis and
prevention of recurrent, occult gastrointestinal bleeding. Am J Surg 1992;163:948;
discussion 989.
59. Foutch P, Sawyer R, Sanowski R. Push-enteroscopy for diagnosis of patients with
gastrointestinal bleeding of obscure origin. Gastrointest Endosc 1990;36:33741.
60. Lewis B, Wenger J, Waye J. Small bowel enteroscopy and intraoperative enteroscopy for
obscure gastrointestinal bleeding. Am J Gastroenterol 1991;86:1714.
61. Bowden TJ. Endoscopy of the small intestine. Surg Clin North Am 1989;69:123747.
62. Bowden TJ. Intraoperative endoscopy of the gastrointestinal tract: Clinical necessity or lack of
preoperative preparation?. World J Surg 1989;13:1869.
63. Desa L, Ohri S, Hutton K, et al. Role of intraoperative enteroscopy in obscure gastrointestinal
bleeding of small bowel origin. Br J Surg 1991;78:1925.
64. Szold A, Katz LB, Lewis BS. Surgical approach to occult gastrointestinal bleeding. Am J Surg
1992;163:902; discussion 923.
65. Fleischer DE, Goldberg SB, Browning TH, et al. Detection and surveillance of colorectal
cancer. JAMA 1989;261:5805.
66. Dent O, Goulston K, Tennant C, et al. Rectal bleeding. Patient delay in presentation. Dis
Colon Rectum 1990;33:8517.
67. Goulston K, Cook I, Dent O. How important is rectal bleeding in the diagnosis of bowel cancer
and polyps?. Lancet 1986;2:2613.
68. Mant A, Bakey E, Chapus P, et al. Rectal bleeding: Do other symptoms aid in diagnosis?. Dis
69.
Newstead GL. Logical clinical decisions in the investigation and management of non-urgent
rectal bleeding. Gastroenterol Jpn 1991;3:10710.
70.
Schrock T. Examination of anorectum and diseases of anorectum. In Sleisenger M, Fordtran
J (eds). Gastrointestinal Disease: Pathophysiology, Diagnosis, Management. 5th ed. Vol II.
Philadelphia: WB Saunders, 1993;1494516.
71. Dent O, Goulston K, Zubrzycki J, et al. Bowel symptoms in an apparently well population. Dis
72. . Investigation of rectal bleeding [Editorial]. Lancet 1989;1:1957.
73. Simpkins K. What use is barium?. Clin Radiol 1988;39:46973.
74. Margulis A, Thoeni R. The present status of the radiologic examination of the colon.
Radiology 1988;167:15.
75. Isabel-Martinez L, Chapman A, Hall R. The value of a barium enema in the investigation of
patients with rectal carcinoma. Clin Radiol 1988;39:5313.
76. Spencer N, Richards D, Bartlett P, et al. Colorectal polyps: A correlation of radiological and
pathological findings. Clin Radiol 1988;39:40711.
77. Rex D, Lehman G, Lappes J, et al. Sensitivity of double contrast barium study for left colon
polyps. Radiology 1986;158:6972.
78. Rex DK, Lappas JC, Maglinte DD, et al. Enteroclysis in the evaluation of suspected small
intestinal bleeding. Gastroenterology 1989;97:5860.
79. Aylward C, Orangio G, Lucas G, et al. Diffuse cavernous hemangioma of the
rectosigmoidCT scan, a new diagnostic modality and surgical management using
sphincter-saving procedures: Report of three cases. Dis Colon Rectum 1988;31:797802.
80. Lupetin A. Diffuse cavernous hemangioma of the rectum: Evaluation and MRI. Gastrointest
Radiol 1990;15:3435.
81. Dunaway M, Webb W, Rodning C. Intraluminal measurement of distance in the colorectal
region employing rigid and flexible endoscopes. Surg Endosc 1988;2:813.
82. Wilking N, Petrelli N, Herrera-Ornelas L, et al. A comparison of the 25-cm rigid
proctosigmoidoscope with the 65-cm flexible endoscope in the screening of patients for
colorectal carcinoma. Cancer 1986;57:66971.
83. Rozen P, Ran E, Fireman Z, et al. The relative value of fecal occult blood tests and flexible
sigmoidoscopy in screening for large bowel neoplasia. Cancer 1987;60:25538.
84. Church JM. Analysis of the colonoscopic findings in patients with rectal bleeding according to
the pattern of their presenting symptoms. Dis Colon Rectum 1991;34:3915.
85.
Forde K, Webb W. Acute lower gastrointestinal bleeding. Persp Colon Rectal Surg
1988;1:10512.
86. Van Ness N, Chobanian S, Winters CJ, et al. A study of patient acceptance of double-contrast
barium enema and colonoscopy. Arch Intern Med 1987;147:21756.
87. Gane EJ, Lane MR. Colonoscopy in unexplained lower gastrointestinal bleeding. N Z Med J
1992;105:313.
88. Schrock T. Conceptual developments through colonoscopy. Surg Endosc 1988;2:2404.
89. Gostout C, Schroeder K, Burton D. Small bowel enteroscopy: An early experience in
gastrointestinal bleeding of unknown origin. Gastrointest Endosc 1991;37:58.
90. Gostout C, Wang K, Ahlquist D, et al. Acute gastrointestinal bleeding. Experience of a
specialized management team. J Clin Gastroenterol 1992;14:2607.
91.
Waye J. Diagnostic endoscopy in lower intestinal bleeding. In Sugawa C, Schuman B, Lucas
C (eds). Gastrointestinal Bleeding. New York: Igaku-Shoin, 1992;40718.
92.
Harford F, Pickleman J. Large bowel sources of bleeding. In Sugawa C, Schuman B, Lucas C
(eds). Gastrointestinal Bleeding. New York: Igaku-Shoin, 1992;189202.
93. Spencer J. Lower gastrointestinal bleeding. Br J Surg 1989;76:34.
94. Farrands P, Taylor I. Management of acute lower gastrointestinal hemorrhage in a surgical

unit over a 4-year period. J R Soc Med 1987;80:7982.
95. Boley S, Brandt L. Vascular ectasias of the colon. Dig Dis Sci 1986;31(suppl 9):26S-42S.
96. Richardson J. Vascular lesions of the intestines. Am J Surg 1991;161:28493.
97. Danesh B, Spiliadis C, Williams C, et al. AngiodysplasiaAn uncommon cause of colonic
bleeding: Colonoscopic evaluation of 1050 patients with rectal bleeding and anaemia. Int J
Colorectal Dis 1987;2:21822.
98. Trudel J, Fazio V, Sivak M. Colonoscopic diagnosis and treatment of arteriovenous
malformations in chronic lower gastrointestinal bleeding. Clinical accuracy and efficacy. Dis
99. Naveau S, Aubert A, Poynard T, Chaput JC. Long-term results of treatment of vascular
malformations of the gastrointestinal tract by neodymium:YAG laser photocoagulation. Dig Dis
Sci 1990;35:8216.
100. Imperiale T, Ransohoff D. Aortic stenosis, idiopathic gastrointestinal bleeding, and
angiodysplasia: Is there an association?. Gastroenterology 1988;95:16706.
101. Mehta P, Heinsimer J, Bryg R, et al. Reassessment of the association between
gastrointestinal arteriovenous malformations and aortic stenosis. Am J Med 1989;86:2757.
102. Aldabagh S, Trujillo Y, Taxy J. Utility of specimen angiography in angiodysplasia of the colon.
103. Shah I. Angiodysplasia: Morphological diagnosis. Endoscopy 1988;20:14951.
104. Belli A-M, Hemingway A. Malignant 'angiodysplasia'. Clin Radiol 1991;44:313.
105.
Kelvin F, Gardiner R. Clinical Imaging of the Colon and Rectum. New York: Raven Press.
1987.
106. Hartford W. Gastrointestinal angiodysplasia: Clinical features. Endoscopy 1988;20:1448.
107. Richter J, Christensen M, Colditz G, et al. Angiodysplasia. Natural history and efficacy of
therapeutic interventions. Dig Dis Sci 1989;34:15426.
108. Roberts P, Schoetz DJ, Coller J. Vascular ectasia. Diagnosis and treatment by colonoscopy.
Am Surg 1988;54:569.
109.
Swain D. Pathology of the bleeding lesions. In Sugawa C, Schuman B, Lucas C (eds).
Gastrointestinal Bleeding. New York: Igaku-Shoin, 1992;40718.
110. Rex D, Lewis B, Waye J. Colonoscopy and endoscopic therapy for delayed post-polypectomy
hemorrhage. Gastrointest Endosc 1992;38:1279.
111.
Lyon D, Mantia A. Large-bowel hemangiomas. Dis Colon Rectum 1987;27:40414.
112. Robert J, Sachar D, Greenstein A. Management of severe hemorrhage in ulcerative colitis.
Am J Surg 1990;159:5505.
113. Robert J, Sachar D, Greenstein A. Severe gastrointestinal hemorrhage in Crohn's disease.
Ann Surg 1991;213:20711.
114. Chou Y-H, Hsu S-C, Wang C-Y, et al. Ischemic colitis as a cause of massive lower
gastrointestinal bleeding and peritonitis. Report of five cases. Dis Colon Rectum
1989;32:106570.
115. Scherpenisse J, van Hees PAM. The endoscopic spectrum of colonic mucosal injury following
aortic aneurysm resection. Endoscopy 1989;21:1746.
116. Zelenock G, Strodel W, Knol J, et al. A prospective study of clinically and endoscopically
documented colonic ischemia in 100 patients undergoing aortic reconstructive surgery with
aggressive colonic and direct pelvic revascularization, compared with historic controls.
Surgery 1989;106:77180.
117. Blundell CR, Earnest DL. Idiopathic cecal ulcer. Diagnosis by colonoscopy followed by
nonoperative management. Dig Dis Sci 1980;25:494503.
118. Testart J, Maupas J, Metayer J, et al. Rectal peptic ulceration: A rare cause of rectal bleeding:
Report of a case. Dis Colon Rectum 1988;31:8035.
119. Stylianos S, Forde K, Benvenisty A, et al. Lower gastrointestinal hemorrhage in renal
transplant recipients. Arch Surg 1988;123:73944.
120. Richards W, Grove-Mahoney D, Williams L. Hemorrhage from a Dieulafoy type ulcer of the
colon: A new cause of lower gastrointestinal bleeding. Am Surg 1988;54:1214.

121. Buchi K, Dixon J. Argon laser treatment of hemorrhagic radiation proctitis. Gastrointest
Endosc 1987;33:2730.
122. Alexander TJ, Dwyer RM. Endoscopic Nd:YAG laser treatment of severe radiation injury of
the gastrointestinal tract: Long term follow up. Gastrointest Endosc 1988;34:40711.
123. Kuroki F, Iida M, Matsui T, et al. Intraoperative endoscopy for small intestinal damage in
radiation enteritis. Gastrointest Endosc 1992;38:1967.
124. Hong G, Wong C, Nambiar R. Massive lower gastrointestinal haemorrhage from a splenic
artery pseudoaneurysm. Br J Surg 1992;79:174.
125. Santos JJ, Feres O, Rocha J, et al. Massive lower gastrointestinal hemorrhage caused by
pseudocyst of the pancreas ruptured into the colon. Report of two cases. Dis Colon Rectum
1992;35:757.
126.
Ryan P, Styles C, Chmiel R. Identification of the site of severe colon bleeding by
technetium-labeled red-cell scan. Dis Colon Rectum 1991;35:21922.
127. Pickens CA, Tedesco FJ. Colonic varices. Unusual cause of rectal bleeding. Am J
128. Waxman JS, Tarkin N, Dave P, Waxman M. Fatal hemorrhage from rectal varices. Report of
two cases. Dis Colon Rectum 1984;27:74950.
129. Orozco H, Takahasi T, Mercado M, et al. Colorectal variceal bleeding in patients with
extrahepatic portal vein thrombosis and idiopathic portal hypertension. J Clin Gastroenterol
1992;14:13943.
130. Beck D, Fazio V, Grundfest-Broniatowski S. Surgical management of bleeding stomal varices.
131. Kozarek RA, Botoman VA, Bredfeldt JE, et al. Portal colopathy: Prospective study of
colonoscopy in patients with portal hypertension. Gastroenterology 1991;101:11927.
132. Lau W, Yuen W, Chu K, et al. Obscure bleeding in the gastrointestinal tract originating in the
small intestine. Surg Gynecol Obstet 1992;174:11924.
133.
Thompson J, Salem R, Hemingway A, et al. Specialist investigation of obscure
gastrointestinal bleeding. Gut 1987;28:4751.
134. Vetto J, Richman P, Kariger K, et al. Cirsoid aneurysms of the jejunum. An unrecognized
cause of massive gastrointestinal bleeding. Arch Surg 1989;124:14602.
135. Staunton D, Jacobson A, Thorning D, et al. Coumadin-induced gastrointestinal hemorrhage
associated with an ileal duplication. J Clin Gastroenterol 1990;12:6859.
136. Neff R, Kiremer S, Voutsinas L, et al. Primary Kaposi's sarcoma of the ileum presenting as
massive rectal bleeding [Letter to Editor]. Am J Gastroenterol 1987;82:2767.
137. Hansen C, Colstrup H, Mogensen A. Recurrent gastrointestinal bleeding due to
leiomyosarcoma of the small intestine. Case report. Eur J Surg 1991;157:2279.
138. Vetto J, Culp S, Smythe T, et al. Iliac arterial-enteric fistulas occurring after pelvic irradiation.
Surgery 1987;101:6437.
139.
Roberts P, Veidenheimer M. Diverticulosis and hemorrhage. In Fazio V (ed). Current Therapy
in Colon and Rectal Surgery. Philadelphia: BC Decker, 1990;2358.
140. Jensen D, Machicado G. Diagnosis and treatment of severe hematochezia: The role of urgent
colonoscopy after purge. Gastroenterology 1988;95:156974.
141. Caos A, Benner K, Mannier J, et al. Colonoscopy after Golytely preparation in acute rectal
bleeding. J Clin Gastroenterol 1986;8:469.
142. Rossini F, Ferrari A, Spandre M, et al. Emergency colonoscopy. World J Surg 1989;13:1902.
143.
Keeffe D, Schrock T. Complications of gastrointestinal endoscopy. In Sleisenger M, Fordtran
J (eds). Gastrointestinal Disease: Pathophysiology, Diagnosis, Management. 5th ed. Vol I.
Philadelphia: WB Saunders, 1993;30110.
144. Lau WY. Intraoperative enteroscopyIndications and limitations. Gastrointest Endosc
1990;36:26871.
145. Berry A, Campbell W, Kettlewell M. Management of major colonic haemorrhage. Br J Surg
1988;75:63740.
146.
Forde K, Treat M. Colonoscopy for lower gastrointestinal bleeding. In Dent T, Strodel W,

Turcotte J, et al. (eds). Surgical Endoscopy. Chicago: Year Book Medical, 1985;26174.
147. Hunter J, Pezim M. Limited value of technetium 99m-labeled red cell scintigraphy in
localization of lower gastrointestinal bleeding. Am J Surg 1990;159:5046.
148. Bentley D, Richardson J. Role of tagged red blood cell imaging in the localization of
gastrointestinal bleeding. Arch Surg 1991;126:8214.
149. Voeller GR, Bunch G, Britt LG. Use of technetium-labeled red blood cell scintigraphy in the
detection and management of gastrointestinal hemorrhage. Surgery 1991;110:799804.
150. Fiorito JJ, Brandt LJ, Kozicky O, et al. The diagnostic yield of superior mesenteric
angiography: Correlation with the pattern of gastrointestinal bleeding. Am J Gastroenterol
1989;84:87881.
151. Leitman I, Paull E, Shires GI. Evaluation and management of massive lower gastrointestinal
hemorrhage. Ann Surg 1989;209:17580.
152. Uden P, Jiborn H, Johnson K. Influence of selective mesenteric arteriography on the outcome
of emergency surgery for massive lower gastrointestinal hemorrhage: A 15 year experience.
153. Lau W, Ngan H, Chu K, et al. Repeat selective visceral angiography in patients with
gastrointestinal bleeding of obscure origin. Br J Surg 1989;76:2269.
154. Gostout C, Bowyer B, Ahlquist D, et al. Mucosal vascular malformations of the gastrointestinal
tract: Clinical observations and results of endoscopic neodymium:yttrium-aluminum-garnet
laser therapy. Mayo Clin Proc 1988;63:9931003.
155. Rosch J, Kozak B, Keler F, et al. Interventional angiography in the diagnosis of acute lower
gastrointestinal bleeding. Eur J Radiol 1986;6:13641.
156. Steger A, Galland R, Hemingway A, et al. Gastrointestinal haemorrhage from a second
source in patients with colonic angiodysplasia. Br J Surg 1987;74:7267.
157. Krishnan J, Berce M, Sweeney J, et al. A retrospective study on the management of rectal
bleeding. Aust N Z J Surg 1987;57:6358.
158.
Jensen D, Machicado G. Endoscopic diagnosis and treatment of bleeding colonic angiomas
and radiation telangiectasia. Persp Colon Rectal Surg 1989;2:99113.
159. Johnston JH, Jensen DM, Auth D. Experimental comparison of endoscopic
yttrium-aluminum-garnet laser, electrosurgery, and heater probe for canine gut arterial
coagulation: The importance of vessel compression and avoidance of tissue erosion.
160. Wadas DD, Sanowski RA. Complications of the hot biopsy forceps technique. Gastrointest
Endosc 1988;34:327.
161.
Williams C, Fairclough P. Colonoscopy. Curr Opin Gastroenterol 1990;6:5464.
162. Helmrich G, Stallworth J, Brown J. Angiodysplasia: Characterization, diagnosis, and advances
in treatment. South Med J 1990;83:14503.
163. Hunter J. Endoscopic laser applications in the gastrointestinal tract. Surg Clin North Am
1989;69:114766.
164. Buchi K. Vascular malformations of the gastrointestinal tract. Surg Clin North Am
1992;72:55970.
165. Lanthier P, D'Harveng B, Vanheuverzwyn R, et al. Colonic angiodysplasia: Follow-up of
patients after endoscopic treatment for bleeding lesions. Dis Colon Rectum 1989;32:2968.
166. Cottone C, Disclafani G, Genova G, et al. Use of BICAP in a case of colon angiodysplasia.
167. van Cutsem E, Rutgeerts P, Vantrappen G. Treatment of bleeding gastrointestinal vascular
malformations with oestrogen-progesterone. Lancet 1990;335:9535.
168. Savides TJ, Jensen DM. Colonoscopic hemostasis of recurrent diverticular hemorrhage
associated with a visible vessel: A report of three cases. Gastrointest Endosc 1994;40:703.
169. Bertoni G, Conigliaro R, Ricci E, et al. Endoscopic injection hemostasis of colonic diverticular
bleeding: A case report. Endoscopy 1990;22:1545.
170. Encarnacion C, Kadir S, Beam C, et al. Gastrointestinal bleeding: Treatment with
gastrointestinal arterial embolization. Radiology 1992;183:5058.

Schrock T. Surgical management of intestinal bleeding. In Sugawa C, Schuman B, Lucas C
(eds). Gastrointestinal Bleeding. New York: Igaku-Shoin, 1992;40718.
172. Apelgren KN, Vargish T, Al-Kawas F. Principles for use of intraoperative enteroscopy for
hemorrhage from the small bowel. Am Surg 1988;54:858.
173. Whelan R, Buls J, Goldberg S. Intraoperative endoscopy: University of Minnesota experience.
Am Surg 1989;55:2816.
174. Scott H, Lane I, Glynn M, et al. Colonic haemorrhage: A technique for rapid intra-operative
bowel preparation and colonoscopy. Br J Surg 1986;73:3901.
175. Biener A, Palestro C, Lewis BS, et al. Intraoperative scintigraphy for active small intestinal
bleeding. Surg Gynecol Obstet 1990;171:38892.
176. Aalders G, Baeten C, Loffeld R. Suturing angiodysplastic lesions of the small intestine. Surg
177.
Dickstein G, Boley S. Vascular lesions of the colon. In Condon R (ed). Shackelford's Surgery
of the Alimentary Tract. 3rd ed. Vol IV: Colon and Anorectum. Philadelphia: WB Saunders,
1991;7083.
178. Milewski P, Schofield P. Massive colonic hemorrhage: The case for right hemicolectomy. Ann
R Coll Surg Engl 1989;71:2539.
179. Bender J, Wiencek R, Bouwman D. Morbidity and mortality following total abdominal
colectomy for massive lower gastrointestinal bleeding. Am Surg 1991;57:53640.
171.
Chapter 28 Thermal Contact Methods for Endoscopic

Hemostasis*(1499)
(1500)
(1501)
DENNIS M. JENSEN, M.D.
Endoscopic control of nonvariceal bleeding has evolved since the mid-1980s to become the treatment
of choice for upper, jejunal, and lower gastrointestinal (GI) bleeding.13(1502) For many different
lesions, both control of active bleeding and prevention of recurrent bleeding from high-risk nonvariceal
lesions are feasible.4(1503) Since the mid-1980s, endoscopic treatment in the intensive care setting
using contact probes has largely supplanted emergency surgery for control of active and recurrent
bleeding.1,4(1504)
The purposes of this chapter are (1) to discuss thermal contact methods for nonvariceal hemostasis,
(2) to review the principles and goals of hemostasis with these techniques, and (3) to summarize
recent results of endoscopic hemostasis with thermal contact methods.
The thermal contact methods to be discussed are monopolar electrocoagulation, bipolar
electrocoagulation, and use of the heater probe. Patients with nonvariceal hemorrhage of the upper GI
tract from different types of lesions have benefited from endoscopic treatment with these
methods.1,2,4(1505)
Upper GI bleeding (and the lesions from which bleeding arises) can be regarded as arterial and
nonarterial in nature. The arterial lesions include peptic ulcers with active arterial bleeding and
nonbleeding visible vessels,1(1506) Mallory-Weiss tears with active bleeding,5(1507) Dieulafoy's
lesions,6(1508) and some upper GI cancers.7(1509) On histopathology, arteries are the underlying
source of severe upper GI blood loss from these lesions and should be the target of endoscopic
hemostasis.1,48(1510) In contrast to arterial bleeding from peptic ulcers, oozing of blood from ulcers,
tumors, or Mallory-Weiss tears in the absence of stigmata of ulcer hemorrhage (such as a visible
vessel or adherent clot) indicates nonarterial bleeding or a severe coagulopathy. Patients with these
latter findings have not been shown to benefit any more from endoscopic therapy than from medical
treatments alone because the risk of continued or recurrent bleeding is low.18(1511)

Angioma syndromes of the upper GI tract are a common source of hemorrhage, particularly in elderly
patients. The outcome for patients with this type of bleeding is improved by endoscopic compared with
medical treatment.1,9(1512) The histopathology of angiomata is an arteriovenous malformation; this is
the source of hemorrhage rather than an artery, as is the case with ulcers. Watermelon stomach and
Osler-Weber-Rendu syndrome are examples of angioma syndromes. Patients with GI bleeding from
angiomata often have chronic or occult hemorrhage rather than massive, arterial-type
bleeding.9,10(1513)
Several types of nonvariceal lesions of the colon that cause hemorrhage are amenable to endoscopic
hemostasis with contact probes, whether severe arterial bleeding (such as a bleeding polyp stalk) or
oozing hemorrhage (such as from arteriovenous malformations) is present.3,11(1514) Lower GI
lesions amenable to endoscopic therapy include lower GI angiomata,3,11(1515) diverticulosis with
active bleeding or visible vessel,12(1516) bleeding polyp stalk,3(1517) radiation
telangiectasia,11(1518) and ulcerated cancers with focal bleeding.7(1519)
Contact Devices and Principles for their Application

Monopolar Electrocoagulation
Commercially available radio frequency generators operate at frequencies above 100 kHz to induce a
current that passes through the body and produces heat within tissue but no neuromuscular stimulation
(see Chapter 9: Principles of Electrosurgery).
For endoscopic hemostasis, a variety of monopolar electrodes have been applied.13(1520) Electric
current flows from the electrode tip through the tissue to a distant patient plate (indifferent or return
electrode). Current density is highest at the electrode tip (the active electrode) when held directly in
contact with the tissue; the resultant heating effect is proportional to the square of the current density.
When the electrode tip is in contact with tissue, the heating effect at the active monopolar electrode
tends to confine the zone of thermal damage to a region about 4 to 5 mm deep adjacent to the active
electrode.1416(1521) Depending on the waveform of the generator, either coagulation or cutting of
tissue may result from monopolar electrocoagulation.13,14(1522) Undesirable effects of monopolar
electrocoagulation used for endoscopic purposes include tissue erosion and excessive sticking
between electrode and tissue, which can be hazardous during coagulation of arterial
lesions.13,14(1523) Nevertheless, controlled monopolar electrocoagulation is commonly used for
certain specific purposes such as polypectomy and papillotomy and is very effective in these
applications. By contrast, monopolar electrocoagulation is rarely used for endoscopic hemostasis.
Fulguration is produced with large electric voltages (about 1000 V) when sparking occurs between a
monopolar electrode and the tissue. Usually the electrode tip is held just above the surface of the
tissue to avoid sticking. In the laboratory, sealing of gut arteries up to about 0.5 mm in diameter is
possible with electrofulguration.13,14(1524) Hazards of electrofulguration for endoscopic hemostasis
are tissue cutting and erosion.14,16(1525)
Bipolar or Multipolar Electrocoagulation

A bipolar electrode consists of two active electrodes incorporated into a single probe; that is, the return
or indifferent electrode plate is eliminated. Electric current passes from one electrode through the
tissue to the opposite electrode. Therefore, bipolar electric fields are much more confined than those
produced by the monopolar configuration. The depth of coagulation in tissue is about equal to the
center-to-center spacing of the electrodes. For the large-diameter (3.2-mm) probe, this is
approximately 1 to 2 mm deep when light apposition pressure is applied before coagulation. Firmer
apposition pressure with tissue compression before applying the current increases the depth of
coagulation.13(1526) Because bipolar electrodes have metal contacts, touching the probe to heated
tissue and structures containing protein can result in sticking. However, sticking can be reduced by
using broad-area electrodes and deliberately keeping the current density (and the resultant
temperature) in the vicinity of the electrodes low enough that adhesion or severe sticking does not
occur.1,3,13(1527)
Multipolar probes consist of pairs of bipolar electrodes arranged around and over the end of a
cylindrical ceramic core.13,17(1528) Figure 281 illustrates the BICAP probe, the first multipolar probe
to be marketed (Circon-ACMI, Stamford, CT), and a newer configuration of the bipolar electrode (Gold
Probe, Microvasive, Natick, MA). These probes were designed with broad electrodes and narrow
interspaces so that severe sticking is reduced, particularly when low power settings are used.
(1529)Figure 281. Bipolar probes. BICAP (top) and Gold (bottom) probes in standard (2.3
mm diameter) and large (3.2 mm diameter) sizes.
The multipolar probe is the preferred configuration for a bipolar electrosurgical device for the following
reasons: (1) coagulation is omnidirectional; (2) probes operate at low power and with a large electrode
contact area to minimize sticking when firmly pressed against tissue to coapt a bleeding artery; and (3)
the probes have small insulating gaps to assist in rapid production of uniform coagulation in the tissue
tangential to the probe tip. In addition, systems that incorporate the multipolar probes have the
following advantages: (1) The generator automatically limits the amount of energy delivered because of
tissue desiccation, and (2) the unit incorporates a water irrigation system and a timer for regulation of
the electric exposure. A bipolar generator that operates under foot-switch control for coagulation and
irrigation is shown in Figure 282. Irrigation is done by a jet of fluid in the center of the probe tip. The
rate and force of irrigation are adjustable in the newest bipolar generators.
(1530)Figure 282. Bipolar generator, foot pedal, water bottle, and large-diameter probe.
(Courtesy of Circon ACMI.)
When coaptive pressure is applied with the probe prior to current delivery, the efficiency of coagulation
and sealing of an artery are greatly increased.1,1316(1531) When coaptive pressure collapses the
intima of an artery onto itself before coagulation, the artery is efficiently welded shut when current is
applied, thereby providing a thermal ligature.1,13,14(1532) Sparking and erosion into the vessels are
prevented by firm contact and the use of low voltages from the power generator. In the laboratory, gut
arteries up to 2.5 mm in diameter can be effectively welded using the multipolar technique of firm
tamponade and coaptation.14(1533) The depth of tissue coagulation with bipolar probes can be
modulated by changing the pressure applied with the probe, the power setting, or the pulse duration.
Independent of other variables, firm pressure compresses tissue, may thin out the bowel wall, and
increases depth of tissue coagulation.14(1534) Application of a multipolar probe at low power output
(settings of 20 watts on a BICAP II generator or 3 or 4 on a BICAP I 50-watt generator) and with long
pulses (5 to 10 sec), when accompanied by firm tamponade, increases the depth of tissue coagulation
compared with application of short pulses (1 to 2 sec) and high power output (settings 9 to
10).18,19(1535) This relates to rapid tissue desiccation at the higher setting and automatically limits
the amount of energy delivered to the underlying artery by the low-voltage bipolar generator. Lower
power and longer pulses of coagulation are preferred for bleeding ulcers or upper GI tumors, whereas
short pulses and low power are recommended for treatment of GI angiomata.1,3,7,9,11(1536)
Heater Probe
The computer-controlled thermal cautery probe, referred to as the heater probe, delivers
predetermined pulses of heat to an endoscopic catheter (Figures 283 and 284). A silicone chip in the
probe tip generates the heat and controls the temperature to 250 C.20,21(1537) The chip is
surrounded by a low-heat-capacity metal envelope that reaches operational temperature in less than
0.2 sec. The heater probe surface is coated with Silverstone, a state-of-the-art, nonstick, durable
coating. Heat is conducted from the probe into tissue; no electric current flows into the tissue. When
held in firm contact with tissue, the probe cools in less than 0.5 sec after activation.20,21(1538)
(1539)Figure 283. Heater probes of two diameters: large (3.2 mm) (top) and standard (2.4
mm) (bottom). (Courtesy of Olympus America, Inc., Melville, New York.)
(1540)Figure 284. Heater probe generator, foot pedal, water bottle, and large-diameter
probe. (Courtesy of Olympus America, Inc., Melville, New York.)
In order to achieve reproducible coagulation in tissue, the heater probe is linked to an analog computer
that controls the temperature and total energy delivered to the tissue (see Figure 284). The
endoscopist programs the computer to deliver a desired amount of energy for efficient coagulation,
from 5 to 30 J. When the foot pedal is engaged, the preselected amount of heat is delivered to the
tissue or surrounding media. The delivery time depends on the joule setting and the thermal
conductivity of the media. When the probe is applied to an ulcer base with firm tamponade, a 30-J
pulse requires 5 to 8 sec.22(1541) Poor contact between the probe and the tissue or a large volume of
blood or fluid in an ulcer crater increases the time required to deliver the desired thermal energy from
the tip of the heater probe.
In the laboratory, 1-mm-diameter gut arteries required three to four pulses of 30 J each (total 90 to 120
J) to achieve coaptive coagulation.3,13,14(1542) Two-millimeter arteries required 5 or 6 pulses (total
150 to 180 J) for coaptive coagulation.14(1543) Large probes (3.2 mm diameter) are preferred to small
probes (2.4 mm diameter) for coaptive coagulation of bleeding arterial lesions. In the laboratory, gut
arteries as large as 2.5 mm can be effectively coagulated using firm tamponade and enough heat for
coaptive coagulation.
The heater probe has three water jets spaced around the periphery of the distal tip of the probe that
are recessed from the end so that washing can be performed when the probe is firmly pressed into the
tissue (Figure 285). This enables the endoscopist to irrigate the target during tamponade and to verify
effective coaptation of an actively bleeding artery before initiating coagulation. However, suctioning of
pooled water or blood from the area of the probe tip before endoscopic treatment facilitates accurate
visualization, effective coaptation, and successful hemostasis. Simultaneous suctioning and
coagulation require a therapeutic endoscope with two accessory channels (Figure 286).
(1544)Figure 285. Water irrigation via the tip of a heater probe. A large-diameter probe has
been passed through the accessory channel of a therapeutic endoscope.
(1545)Figure 286. Distal end of the insertion tube of a therapeutic video endoscope with two
accessory channels. In one channel is a large-diameter heater probe. The other accessory
channel can be used for simultaneous suctioning.
Methods and Techniques

Endoscopes, Accessories, and General Endoscopic Techniques for Hemostasis
A therapeutic endoscope with two suction channels greatly facilitates diagnostic and therapeutic
endoscopy in patients with severe ulcer bleeding.1,4,13,23(1546) Several commercially available video
endoscopes have 3.7-mm and 2.8-mm-diameter accessory (suction) channels or smaller channels
(see Figure 286). When a probe or catheter is placed in one channel, the other channel is still
available to suction fluid and blood. Although single-channel video endoscopes are adequate for
diagnosis in patients with active bleeding, they are often inadequate for treatment of actively bleeding
ulcers, Mallory-Weiss tears, or Dieulafoy's lesions.23(1547) With a large-diameter probe in the
accessory channel of a single-channel endoscope, suction is inadequate during endoscopic treatment
of arterial bleeding. The actively bleeding lesion and the precise bleeding point are rapidly obscured by
blood, clots, or both.
Visualization is facilitated by suctioning excess fluid or blood from the region of the bleeding lesion,
positioning the patient so that the blood flows away from the lesion, and tamponading the bleeding
point while suctioning at the active bleeding site. When clots obscure the point of active bleeding, it is
necessary to visualize and determine the nature of the bleeding point (e.g., visible vessel, spurting
artery, oozing) prior to tamponade and coagulation. Clots can be removed by irrigation, changing the
position of the patient, suctioning, pushing them away with the probe, or, on occasion, guillotining with
a polypectomy snare. Blind treatment with contact probes around the ulcer, Dieulafoy's lesion, or
Mallory-Weiss tear without tamponade of the bleeding point is usually not effective for coagulation of
active arterial bleeding or prevention of recurrent hemorrhage from a nonbleeding visible vessel.
Endoscopic Hemostasis Techniques

The equipment and techniques of application are important determinants of outcome for patients with
nonvariceal bleeding who are treated with endoscopic thermal devices. Lack of standardization of
techniques among endoscopists accounts for variable outcomes.4(1548) We have been able to train
skilled endoscopists and gastroenterology fellows to use similar techniques of endoscopic hemostasis.
Originally standardized in laboratory studies, these techniques were then applied in randomized,
controlled studies of hemostasis (Tables 281 and 282). More recently, they have been advocated for
the general gastroenterologist. In this section, the clinically relevant details are reviewed and
summarized.
Center for Ulcer Research and Education (CURE) Hemostasis Research Group: E
Technical Parameters for Endoscopic Bipolar Electrocoagulation of Different Gastrointestinal Lesion
TABLE 281
PEPTIC ULCER
Active
Bleeding
Nonbleeding
Visible
Vessel (VV)
Probe size,
Large
Large
Pressure
Power setting
BICAP I
Very firm
3
MALLORYWEISS
TEAR
GASTROINTE
ANGIOMA SYN
DIEULAFOY'S LESION
Active
Bleeding
Active
Bleeding
Nonbleeding
Visible Vessel
(VV)
Active
Bleeding
Large
Large
Very firm
Large or
small
Moderate
Firm
Firm
Large vs.
small
Light
Center for Ulcer Research and Education (CURE) Hemostasis Research Group: E
Technical Parameters for Endoscopic Bipolar Electrocoagulation of Different Gastrointestinal Lesion
TABLE 281
PEPTIC ULCER
BICAP II
Pulse duration/
tamponade
station
Endpoint
MALLORYWEISS
TEAR
GASTROINTE
ANGIOMA SYN
DIEULAFOY'S LESION
Active
Bleeding
Nonbleeding
Visible
Vessel (VV)
20 W
20 W
20 W
20 W
20 W
15 W
10 sec
10 sec
4 sec
10 sec
10 sec
2 sec
Active
Bleeding
Active
Bleeding
Nonbleeding
Visible Vessel
(VV)
Active
Bleeding
Bleeding
VV flattened
Bleeding
Bleeding
VV flattened
Bleeding
stops
and white
stops
stops
and white
stops
* General standardized guidelines based on laboratory and randomized clinical endoscopic studies. Power, pressure, and p
settings must be reduced for small, acute, or very deep upper GI bleeding lesions. Probes should be checked for proper o
to endoscopic application. Patients should be resuscitated before and during endoscopy. Surgical consultation on all pa
severe upper GI hemorrhage is recommended before endoscopic coagulation. Should endoscopic therapy fail to control a
bleeding or prevent recurrent bleeding, continued resuscitation of the patient and surgery are recommended.
Large-channel endoscopes (single or double) are recommended for all emergency endoscopic procedures performed for
GI hemorrhage. Large-diameter thermal probes (approximately 3.2 mm diameter) are recommended for all nonvariceal bl
lesions or nonbleeding visible vessels except for small arteries (spurts) in Mallory-Weiss tears (<0.5 mm diameter) or sma
angiomata (<3 mm diameter).
Small endoscopic thermal probes (approximately 2.4 mm diameter) have less capability for tamponade, washing ca
of coagulation than large probes and are recommended for coagulation via large-channel endoscopes only for small spurt
associated with Mallory-Weiss tears (<0.5 mm, stream) or small upper GI angiomata (<3 mm diameter).
Pressure is the tamponade pressure exerted en face or tangentially via the contact probe, directly on the bleeding lesion o
nonbleeding visible vessel. For active bleeding, it is recommended that enough pressure be applied to stop the bleeding b
initiating coagulation.
Pulse duration/tamponade station refers to the duration (in seconds) of coagulation to the bleeding point, after tamponade
changing the position of the probe. Note that pulse duration is selected on the power unit for bipolar electrocoagulation bu
upon the power setting and probe size for the heater probe.
Center for Ulcer Research and Education (CURE) Hemostasis Research Group:
Endoscopic Technical Parameters for Endoscopic Heater Probe Coagulation of Different Gastrointe
Lesions*
TABLE 282
PEPTIC ULCER
Active
Bleeding
Nonbleeding
Visible
Vessel (VV)
MALLORYWEISS
TEAR
Active
Bleeding
GAS
DIEULAFOY'S LESION
Active
Bleeding
Nonbleeding
Visible Vessel
(VV)
Bleeding
Center for Ulcer Research and Education (CURE) Hemostasis Research Group:
Endoscopic Technical Parameters for Endoscopic Heater Probe Coagulation of Different Gastrointe
Lesions*
TABLE 282
PEPTIC ULCER
Active
Bleeding
Nonbleeding
Visible
Vessel (VV)
Probe size,
Large
Large
Pressure
Power setting
No. of pulses/
tamponade station
Endpoint
Very firm
30 J
4 sec
MALLORYWEISS
TEAR
Active
Bleeding
GAS
DIEULAFOY'S LESION
Active
Bleeding
Nonbleeding
Visible Vessel
(VV)
Bleeding
Large
Large
Very firm
Large or
small
Moderate
Firm
Firm
Large vs.
small
Light
30 J
4 sec
20 J
3 sec
30 J
4 sec
30 J
4 sec
15 J
23 sec
Bleeding
VV flattened
Bleeding
Bleeding
VV flattened and Bleeding
stops
and white
stops
stops
white
stops
* General standardized guidelines based on laboratory and randomized clinical endoscopic studies. Power, pressure, and p
duration settings must be reduced for small, acute, or very deep upper GI bleeding lesions. Probes should be checked for
operation prior to endoscopic application. Patients should be resuscitated before and during endoscopy. Surgical consulta
patients with severe upper GI hemorrhage is recommended before endoscopic coagulation. Should endoscopic therapy f
active bleeding or prevent recurrent bleeding, continued resuscitation of the patient and surgery are recommended.
Large-channel endoscopes (single or double) are recommended for all emergency endoscopic procedures performed for
upper GI hemorrhage. Large diameter thermal probes (approximately 3.2 mm diameter) are recommended for all nonvari
bleeding lesions or nonbleeding visible vessels except for small arteries (spurts) in Mallory-Weiss tears (<0.5 mm diamete
upper GI angiomata (<3 mm diameter).
Small endoscopic thermal probes (approximately 2.4 mm diameter) have less capability for tamponade, washing capacity
volume of coagulation than large probes and are recommended for coagulation via large-channel endoscopes only for sm
arteries associated with Mallory-Weiss tears (<0.5 mm, stream) or small upper GI angiomata (<3 mm diameter).
Pressure is the tamponade pressure exerted en face or tangentially via the contact probe, directly on the bleeding lesion o
nonbleeding visible vessel. For active bleeding, it is recommended that enough pressure be applied to stop the bleeding b
initiating coagulation.
No. of pulses/tamponade station refers to the number of coagulation pulses applied to the bleeding point, after tamponade
changing the position of the probe.
Successful endoscopic hemostasis of bleeding ulcers by monopolar electrocoagulation with
commercially available probes, modified probes, or the hydrothermal probe requires more skill than
any other thermal modality. This is due to the problems of sticking and poor coagulation when these
devices are applied tangential to the bleeding point.14(1549) Also, efficacy and tissue damage are
much less predictable and controllable than with newer techniques that use bipolar or heater probes. In
laboratory studies of monopolar electrocoagulation, an analog computer did not alter patterns of tissue
injury or facilitate endoscopic application.24(1550) Monopolar probes often stick to the tissue during
coagulation. Pulling the probe away from a lesion after coagulation often results in resumption of
bleeding. Although constant water or saline irrigation facilitates visualization of bleeding points, tissue
coagulation and hemostasis are often more difficult to achieve.1,24,25(1551) Nevertheless,

experienced endoscopists have reported good results with the use of monopolar electrocoagulation in
the emergency treatment of bleeding from ulcers and Mallory-Weiss tears.2632(1552)
Specific guidelines and recommendations for endoscopic monopolar electrocoagulation are as follows:
(1) 50 to 100 watts/pulse and short (0.5-sec) coagulation pulses; (2) target washing to identify the
exact bleeding point, tamponade of the bleeding point, and target coagulation; (3) to the maximum
possible extent, application of the probe en face; and (4) simultaneous suctioning of blood and
irrigation solution. When the hydrothermal probe is used for coagulation, a two-channel endoscope is
required.
When specific guidelines are used with monopolar electrocoagulation, efficacy for hemostasis of
actively bleeding ulcers or Mallory-Weiss tears is good to excellent.2632(1553) Although tamponade
at endoscopy is possible with monopolar probes to some extent, controlling bleeding from ulcers that
are oriented tangentially in the endoscopic field and those in which the rate of blood loss is high is
difficult. Also, tissue damage may be extensive.13,14(1554)
Endoscopic hemostasis with monopolar probes is technically more demanding than with the other
thermal probes. Training experienced endoscopists in monopolar electrocoagulation for hemostasis is
difficult. Of all methods of thermal coagulation, monopolar electrocoagulation to achieve hemostasis of
actively bleeding ulcers or Mallory-Weiss tears was the most difficult to teach to fellows in
gastroenterology in the experience of the Center for Ulcer Research and Education (CURE).
Monopolar electrocoagulation is not used extensively for coagulation of GI angiomata. The real or
perceived increase in risk of transmural injury and the availability of devices that produce less tissue
damage, such as bipolar or heater probes, have limited the use of monopolar electrocoagulation in the
treatment of these lesions.33,34(1555)
Bipolar or Multipolar Electrocoagulation
For chronic ulcers with major stigmata of hemorrhage (i.e., active arterial bleeding or a nonbleeding
visible vessel) or Dieulafoy's lesion with active arterial bleeding, we recommend firm tamponade
directly on the visible vessel with a 3.2-mm-diameter probe, suctioning of excess blood and fluid, and
coagulation with long-duration pulses and low power settings (see Table 281). Repositioning and
coagulation are performed until hemostasis occurs and the visible vessel is flattened (Figure 287).
Endoscopic treatment is not recommended for ulcers that are oozing blood but have no associated
stigmata (such as visible vessels or clots) because, in the absence of coagulopathies, the bleeding
stops spontaneously in all cases; recurrent bleeding with medical treatment occurs in less than 10% of
patients at CURE.
(1556)Figure 287. Coagulaton of a giant duodenal ulcer with a nonbleeding visible vessel
using a multipolar (Gold) probe. A, Before application of the probe. B and C, During
tamponade and coagulation. D, After coagulation/tamponade is complete.
The large-diameter (3.2-mm) probe has several advantages over the smaller (2.4-mm) one in the
treatment of bleeding ulcers or Dieulafoy's lesion. The large device is more effective for initial
hemostasis (93 vs. 66% in patients with active arterial bleeding from ulcers at CURE). Tamponade is
more effective because of the greater stiffness of the larger probe, washing is more efficient, and the
coagulation volume is greater.4,13(1557) However, a large-channel endoscope (accessory channel of
3.5 to 3.7 mm in diameter) is required for passage of the large probe.
Coagulation with the side of the probe (tangentially) is at least as effective as coagulation with the
probe oriented en face to the lesion. The Gold probe is more effective than the BICAP probe for
tamponade because of the greater rigidity of the former device.19,35(1558) For tangential treatment of
bleeding ulcers via endoscopy, this difference is significant.19(1559) In an experimental study to
determine the optimum techniques for application of the BICAP and Gold probes, Laine36(1560) found
the following: (1) In the majority of cases, less force could be exerted with the Gold probe; (2) the
maximum force exerted with either probe was less with greater extension of the probe beyond the end
of the endoscope and with greater degrees of angulation of the probe from the perpendicular en face
position; (3) mean depths of coagulation (in millimeters) increased with increases in force of
application, power output of the generator, and duration of treatment; and (4) sticking paralleled
increasing depth of coagulation, increasing with the force and duration of application. Laine36(1561)
concluded that optimum bipolar electrocoagulation of bleeding gastroduodenal lesions was achieved
by (1) an en face approach to the bleeding lesion, (2) application of the maximum possible force and a
low-watt power setting, and (3) prolonged periods of coagulation.
For bleeding Mallory-Weiss tears, the site of active bleeding is moderately tamponaded and two or
three pulses of 1 sec are delivered before changing the position of the probe (see Tables 281 and
282 for technical details).5(1562) Otherwise, the endoscopic maneuvers are similar to those for
treatment of bleeding ulcers. Treatment is continued until active bleeding is controlled.
Whereas firm tamponade may be necessary for endoscopic artery coaptation and successful
coagulation of bleeding peptic ulcers,22,23(1563) this technique should not be used for GI angiomata,
as these lesions are associated with a reduced thickness of the bowel wall (see Tables 281 and
282). Power settings of multipolar probes for treatment of upper GI angiomata should be lower than
for ulcers because of the risk of transmural coagulation.13,37(1564) The transmural thickness of the
normal canine stomach uninflated is 4 to 5 mm; the thickness of the wall of the duodenum and jejunum
is 3 to 4 mm; and that of the right colon is 2 to 3 mm.13(1565) When the lumen of the gut is distended
at endoscopy by air insufflation, the thickness of wall in the upper and lower GI tract may be only 2 or 3
mm or less in elderly patients with chronic diseases. For treatment of angiomata with bipolar
electrocoagulation (BICAP or Gold probe), we recommend low power and short pulses. A
small-diameter probe is adequate for coagulation of angiomata less than 3 to 4 mm in diameter.
However, the large probe should be used for treatment of larger lesions.
Heater Probe
Based on laboratory and clinical experience, the treatment of an actively bleeding ulcer or Dieulafoy's
lesion should include firm tamponade at the site of bleeding and then coagulation with four 30-J pulses
before repositioning the probe (see Table 282). Firm tamponade and coagulation are feasible with
both the tip (en face) and the side (tangential) of the heater probe; use of a large probe (3.2 mm
diameter) is recommended (Figures 288 and 289). Suctioning of pooled water and blood from the
area of treatment facilitates visualization of the bleeding site and increases the effectiveness of
endoscopic hemostasis. For severe active bleeding from ulcers or other upper GI lesions, a
two-channel therapeutic video endoscope is necessary for simultaneous coagulation and suctioning.
Guidelines for treatment of the patient with recent, severe upper GI hemorrhage and a nonbleeding
visible vessel in a chronic ulcer are similar to those for treatment of a patient with an actively bleeding
ulcer.
(1566)Figure 288. Coagulation of a bleeding gastric ulcer using the heater probe. A, Before
application of the probe. B, During tamponade and coagulation. C, After treatment.
(1567)Figure 289. A, Extensive, shallow ulceration of the distal esophagus in a patient who
developed gastrointestinal bleeding after heart surgery. Note the slightly raised area in the
center of the ulceration. B, Raised area proves to be a vessel as bleeding begins while the
lesion is being observed. C, Bleeding becomes heavier as the heater probe, visible at lower
right, is applied. D, Tamponade with the heater probe reduces blood flow. E, Bleeding is
stopped after three applications. (A-E, From the collection of Dr. M. V. Sivak, Jr.)
Bleeding arteries within Mallory-Weiss tears are often smaller (<0.5 mm diameter) than those found in
bleeding ulcers, and the thickness of the esophageal wall is decreased as a result of the
tear.5,13(1568) For treatment of this bleeding lesion, moderate tamponade and application of 20-J
pulses in a sequence of two to three pulses until bleeding is controlled are recommended. For GI
telangiectasia or angioma, a lower power setting (5 to 10 J) and light contact pressure rather than firm
tamponade are recommended. This relates to the thinness of the bowel wall associated with
angiomata in contrast to the thickened wall associated with most chronic peptic ulcers. Endoscopic
treatment with the heater probe of watermelon stomach, a condition that has multiple angiomatous
lesions in a linear array, is illustrated in Figure 2810.38(1569)
(1570)Figure 2810. Heater probe coagulation of a watermelon stomach. A and B, Before

treatment. C, During treatment. D, After endoscopic treatment.
Results of Clinical Trials

Ulcers
Monopolar electrocoagulation is not commonly used for emergency hemostasis of nonvariceal GI
lesions because of its limitations and difficulty of use compared with bipolar electrocoagulation or
treatment with the heater probe. However, three randomized controlled trials of monopolar
electrocoagulation were conducted in which either the liquid or dry probe techniques were
studied.26,28,30(1571) Lesions with differing stigmata of hemorrhage were included in these trials.
The results are summarized in Table 283.
Randomized Controlled Trial of Monopolar Electrocoagulation for

Ulcers With Different Stigmata of Hemorrhage
TABLE 283
PATIENTS
Active bleeding*
Moreto
Freitas
Nonbleeding VV
Moreto
Papp
Freitas
MP
MED
11
HEMOSTASIS
MP
MED
REBLEEDING
MP
0%
0%
10
100%
82%
40%
36%
10
16
15
16
14
17
MED
SURGERY
MP
DEATHS
MED
MP
MED
67%
40%
17%
33%
9%
20%
100%
60%
0%
10%
6%
33%
0%
6%
33%
81%
56%
0%
0%
7%
6%
21%
53%
14%
47%
7%
18%
0%
7%
36%
Flat spots
Freitas
11
15
0%
13%
0%
7%
MPmonopolar electrocoagulation-treated patients.
MEDmedical treatment with H2 receptor antagonists.
* Active bleeding includes both oozing and spurting bleeding with or without a visible vessel.
Randomized Controlled Trial of Monopolar Electrocoagulation for

Ulcers With Different Stigmata of Hemorrhage
TABLE 283
PATIENTS
MP
MED
HEMOSTASIS
MP
MED
REBLEEDING
MP
MED
SURGERY
MP
MED
DEATHS
MP
MED
Nonbleeding VV is visible vessel without active bleeding.

Flat spots are nonbleeding, centrally located areas.
Indicates a significant difference (p < .05) between medically and endoscopically treated patients.
Papp28(1572) randomized patients with severe upper GI hemorrhage and nonbleeding visible vessels
in ulcers to dry monopolar electrocoagulation or medical treatment. Both gastric (n = 17) and duodenal
(n = 15) ulcers were included. Eighty-one percent of the medically treated patients had recurrent
bleeding (vs. 6% of patients treated by monopolar electrocoagulation), and 56% had emergency
surgery (vs. 6% for monopolar). Compared with the other controlled trials, this is a very high rate of
recurrent bleeding for ulcers with nonbleeding visible vessels in patients treated medically. Papp
reported significant differences in rates of recurrent bleeding, the need for emergency surgery, and the
direct cost and length of hospitalization. One death occurred in the medically treated group. No
complications were reported.
Frietas et al.26(1573) treated 78 patients in a randomized study of liquid monopolar electrocoagulation
versus medical therapy. Patients had different stigmata of hemorrhage, including active bleeding (21
patients), nonbleeding visible vessels (31 patients), or flat spots (26 patients). Only in the subgroup
with nonbleeding visible vessels was the benefit of monopolar electrocoagulation significant in terms of
rates of recurrent bleeding and the need for emergency surgery (see Table 283). For patients with
active bleeding from ulcers, no significant differences were seen in outcomes such as recurrent
bleeding and emergency surgery. Nor was there any benefit for patients with central flat spots who had
a very low risk of recurrent bleeding (see Table 283). No complications of monopolar
electrocoagulation were reported. This randomized controlled study confirmed the results of
Papp28(1574) for treatment of ulcers with nonbleeding visible vessels.
Moreto et al.30(1575) randomized patients with bleeding gastric ulcers to liquid monopolar
electrocoagulation or medical management. Patients with duodenal ulcers were excluded, as well as
patients with deep gastric ulcers and those with "large visible vessels." Patients with visible vessels and
either spurting or "nonspurting" bleeding in either gastric (81%) or stomal (19%) ulcers were included.
For the small subgroup of patients with spurting visible vessels (n = 12), significant reductions in
recurrent bleeding and emergency surgery were seen, but no difference in mortality was found (see
Table 283). The group with nonspurting visible vessels had no significant differences in recurrent
bleeding or mortality. When all endoscopically versus medically treated patients were considered,
Moreto et al.30(1576) noted significant differences in the frequency of recurrent bleeding, the need for
emergency surgery, and postendoscopy transfusion requirements. Paradoxically, most statistical
differences in outcomes for medically versus monopolar probe-treated patients were in the subgroup
with spurting bleeding. No complications of monopolar electrocoagulation were reported.
The study of Moreto et al.30(1577) has several major limitations. Exclusion of all duodenal ulcers, of
deep gastric ulcers, and of large visible vessels makes interpretation difficult and limits the ability to
draw general conclusions. Also, these investigators excluded from the analysis two patients
randomized in the study. Furthermore, the definition of a nonspurting visible vessel was not provided.
This is not standard terminology and may refer to a nonbleeding visible vessel, oozing from an ulcer
crater, or an ulcer partially covered by clot. Among the technical difficulties, the authors were unable to
coagulate tangentially with the monopolar probe and frequently used a side-viewing endoscope to
obtain en face views of bleeding lesions. For the patients analyzed, only those with bleeding defined as
spurting appeared to benefit significantly from endoscopic treatment. This result contrasted with those
of the other two randomized controlled trials of monopolar electrocoagulation, both of which reported
an improvement only for subgroups of patients with nonbleeding visible vessels.26,28(1578)
Two of the three randomized trials indicate good results for monopolar electrocoagulation of
nonbleeding visible vessels, and two report variable results for active bleeding (see Table 283).
These pertain to major stigmata of ulcer hemorrhage, and patients with these lesions have been
demonstrated to benefit from other endoscopic treatments, including use of the heater probe, bipolar
electrocoagulation, and injection therapy. As expected based on the natural history of flat spots as
minor stigmata of hemorrhage, the subgroup of patients with this finding did not benefit from
monopolar electrocoagulation or any other endoscopic treatment.1,39(1579)
Concerns for precipitation of bleeding (from large vessels), transmural injury and perforation (from
deep gastric or duodenal ulcers), and the difficulty of tangential coagulation substantially limit the
applicability of endoscopic monopolar electrocoagulation in the treatment of ulcers or other nonvariceal
lesions with major stigmata of hemorrhage. Since the mid-1980s, use of the heater probe, bipolar
probes, and injection therapy has supplanted monopolar electrocoagulation despite the fact that
excellent results with the latter method have been reported in the past.2632(1580)
Bipolar Electrocoagulation
Differing results have been reported from the several randomized controlled studies of therapy with the
multipolar BICAP probe. Four found a significant benefit over medical treatment,22,4042(1581)
whereas three others did not.4345(1582) Laine40,41(1583) found in two studies that treatment with
the BICAP was safe and also effective for control of active ulcer bleeding and for prevention of
recurrent bleeding from nonbleeding visible vessels. The large probe was used in these studies, which
included hospitalized patients with severe upper GI bleeding. In one study, patients with actively
bleeding ulcers, Mallory-Weiss tears, or angiomata were randomized to treatment with the BICAP or
medical therapy without endoscopic treatment.40(1584) Compared with medical management, BICAP
therapy significantly reduced postrandomization transfusions, emergency surgery rates, hospital days,
and direct cost. In another study, patients with nonbleeding visible vessels in ulcers were also
randomized by Laine41(1585) to medical management or BICAP treatment. Significant improvements
were seen with BICAP treatment in rates of recurrent bleeding, number of transfusions after
randomization, need for emergency surgery, hospital days, and direct costs. No major complications
were reported (Table 284).
Randomized Controlled Trials of Bipolar Electrocoagulation for Ulcers With

Different Stigmata of Hemorrhage
TABLE 284
PATIENTS
Active bleeding
O'Brien*
Laine*
Jensen (CURE)
Nonbleeding VV
PRIMARY
HEMOSTASIS
REBLEEDING
SURGERY
DEATHS
BP
MED
BP
MED
BP
MED
BP
MED
BP
40
10
14
21
14
14
N
80%
93%
N
14%
14%
15%
30%
47%
62%
64%
93%
N
30%
29%
N
64%
57%
N
0%
7%
Randomized Controlled Trials of Bipolar Electrocoagulation for Ulcers With

Different Stigmata of Hemorrhage
TABLE 284
PATIENTS
O'Brien
Laine
Jensen (CURE)
Clot
PRIMARY
HEMOSTASIS
REBLEEDING
SURGERY
DEATHS
BP
MED
BP
MED
BP
MED
BP
MED
BP
43
38
27
43
37
27
16%
18%
35%
37%
41%
52%
N
8%
28%
N
30%
33%
N
0%
5%
O'Brien
18
39
22%
13%
Nthe authors did not report these data for the subgroup listed.
BPbipolar (BICAP) electrocoagulation.
MEDmedically treated (H2 receptor antagonists) patients.
* Active bleeding refers to a spectrum of bleeding from oozing to spurting.
Active bleeding in this study refers to spurting bleeding.
Nonbleeding VV is visible vessel without active bleeding.
Clot is a nonbleeding adherent clot in the ulcer base.
Indicates a significant difference (p < .05) between the bipolar and medical treatment groups.
Treatment with a small BICAP probe resulted in a significant improvement in the rate of recurrent
bleeding in some patients. In the randomized controlled trial of O'Brien et al.,42(1586) patients were
stratified by stigmata of hemorrhage at initial endoscopy into subgroups with active bleeding from a
visible vessel (61 patients), a nonbleeding visible vessel (86 patients), or an adherent clot (57 patients).
Compared with medical therapy, significantly fewer BICAP-treated patients had recurrent bleeding (17
vs. 33% for medically treated patients). Differences in rates of recurrent bleeding were noted in the
subgroups of patients with active bleeding and nonbleeding visible vessels (see Table 284). However,
BICAP treatment did not reduce emergency surgery or mortality rates and did not significantly reduce
the rate of recurrent bleeding in patients with adherent clot (22% BICAP vs. 13% medical
management).
The results of a randomized controlled trial of BICAP versus medical treatment conducted at CURE
are given in Table 284.2123(1587) In this study of patients with severe ulcer bleeding and active
arterial bleeding or nonbleeding visible vessels at endoscopy, only the rates of initial hemostasis and
recurrent bleeding in the subgroup with active ulcer bleeding were significantly different for BICAP
versus medical treatment. No significant differences were found in other outcomes for BICAP versus
medical therapy in the subgroups of patients with either active bleeding or nonbleeding visible
vessels.22(1588) In this study, however, patients with active arterial bleeding treated with the heater
probe had a significantly higher rate of initial hemostasis, a lower rate of recurrent bleeding, fewer red
cell transfusions on average, and fewer days in intensive care after randomization compared with
patients treated by medical means. For patients with nonbleeding visible vessels, those treated with the
heater probe had significantly lower rates of recurrent bleeding and emergency surgery for recurrence
of bleeding as well as fewer red cell transfusions on average.2123(1589)
Three other randomized controlled trials did not demonstrate a benefit for BICAP coagulation
compared with medical treatment.4345(1590) Major problems with these latter studies included (1)
inclusion in all three studies of many patients with minor stigmata (such as flat spots), (2) lack of
standardization of entry criteria and endoscopic hemostasis techniques, (3) utilization of the small (2.4
mm diameter) rather than the large (3.2 mm) probe in two studies, and (4) application of less total
energy with the BICAP probe compared with the studies of Laine40,41(1591) and CURE. Study design
and technical issues may account for differences in outcome among these several studies.
Several prospective, randomized trials of multipolar electrocoagulation versus other endoscopic
methods for achieving hemostasis have been undertaken.4648(1592) Laine46(1593) compared
multipolar electrocoagulation with injection therapy with absolute alcohol in patients with bleeding
peptic ulcers. Entry criteria included evidence of upper GI bleeding (bloody nasogastric aspirate,
melena, or hematochezia in the presence of unstable vital signs), transfusion of more than two units of
blood in 12 hr, or a 6% or greater decrease in hematocrit over 12 hr. Twenty-six patients with an ulcer
with active bleeding at endoscopy and 34 patients with a nonbleeding visible vessel were randomized.
Bleeding was controlled by multipolar electrocoagulation in all of 14 patients with active bleeding,
whereas injection therapy resulted in hemostasis in 10 of 12 patients (83%) with active bleeding.
During hospitalization, no significant differences were seen between the two methods of endoscopic
hemostasis in any outcome parameter, including rates of further bleeding, units of blood transfused,
need for surgery because of bleeding, number of days in hospital, cost (of hospitalization), mortality, or
treatment-induced bleeding in patients with visible vessels. Bleeding precipitated by treatment of visible
vessels was controlled in all cases by further treatment. One complication occurred (perforation at 9
days after treatment) in the group of patients treated with multipolar electrocoagulation.
Waring et al.48(1594) found that multipolar electrocoagulation and injection therapy with 98% ethanol
were equally effective in a randomized study of men with peptic ulcers and active bleeding or stigmata
of hemorrhage. Initial hemostasis was achieved in 95% of patients in each group, and bleeding
recurred in approximately 25% of patients treated with either modality. Also, no difference was noted in
mortality, although outcome was poorer in general for patients over age 70 years and those with an
ulcer greater than 2 cm in diameter.
Rutgeerts et al.47(1595) randomized 100 patients with peptic ulcers and active bleeding (spurting or
oozing) or a nonbleeding visible vessel to treatment by multipolar electrocoagulation (large probe; n =
50) or neodymium-yttrium-aluminum-garnet (Nd:YAG) photocoagulation (n = 50). All lesions were first
injected with epinephrine. Hemostasis was achieved after one treatment in 72% of patients in both
groups; after two treatments, the cumulative success rate was 86% for BICAP-treated patients versus
88% for those treated with the laser. Major complications consisted of one perforation in each
treatment group.
Heater Probe
Four randomized controlled studies of therapy with the heater probe have been
reported.22,4951(1596) Three found a significant benefit22,50,51(1597) and one did not.49(1598) In
the trial of Matthewson et al.,49(1599) patients with severe ulcer bleeding were randomized to receive
treatment with the heater probe, Nd:YAG laser therapy, or medical-surgical management. At initial
endoscopy, patients were stratified for active bleeding, nonbleeding visible vessel, overlying clot, and
red or black spots. One hundred and forty-three consecutive patients with stigmata of recent
hemorrhage accessible to endoscopic diagnosis and treatment were randomized to Nd:YAG laser
treatment (44 patients), heater probe coagulation (57 patients), and no endoscopic treatment (42
patients). In the analysis of the results of this trial, patients with minor and major stigmata of
hemorrhage were grouped together. For all treatments (minor and major stigmata taken together),
laser photocoagulation was associated with a lower frequency of recurrent bleeding than medical
treatment (20 vs. 42%), whereas heater probe treatment was not (28% heater probe vs. 42% control).
When subgroups of patients with major or minor stigmata were analyzed separately, the different
treatments showed no significant differences in outcomes. The rates of recurrent bleeding for 31
patients with major stigmata of recent hemorrhage were heater probe 13%, Nd:YAG laser 8%, and
medical treatment 30%. One ulcer perforation occurred among the heater probe-treated patients. For
laser photocoagulation, bleeding was precipitated during treatment of about 20% of the nonbleeding
stigmata, such as nonbleeding visible vessels and flat spots. In every case, the active bleeding was
controlled by further laser treatment. No perforations were noted with laser therapy.
In the randomized controlled trial of Fullerton et al.50(1600) in 43 patients with either active ulcer
bleeding or nonbleeding visible vessels, treatment with the heater probe significantly lowered rates of
recurrent bleeding. Twenty patients were treated with the heater probe and 23 with medical therapy. No
deaths occurred in either group.
Chung et al.52(1601) compared epinephrine injection and heater probe treatments in a prospective,
randomized trial that included 132 patients with active ulcer bleeding. Initial hemostasis was achieved
by means of injections in 96% versus 83% of patients treated with the heater probe (p < .05). With
regard to transfusion requirements, emergency surgery, hospital days, and mortality, no significant
differences were found between the treatment groups. A perforation occurred in 2 patients treated by
heater probe coagulation.
In a large randomized controlled study at CURE, patients with active spurting bleeding or nonbleeding
visible vessels were randomized to either heater probe, BICAP, or medical-surgical therapy.22(1602)
Initial hemostasis rates with heater probe and BICAP for spurting bleeding were 93 to 95%, compared
with 14% for medical-surgical therapy. Compared with the medical-surgical treatment subgroup,
however, significant reductions in postrandomization bleeding, units of red blood cells transfused,
mean number of days in intensive care, and rate of emergency surgery were noted only for the
subgroup of patients treated with the heater probe. For the subgroup with nonbleeding visible vessels,
heater probe treatment significantly decreased postrandomization rates of recurrent bleeding, red cell
transfusion, and emergency surgery. No major complications of treatment were found with either the
heater probe or BICAP (Table 285). Compared with our previous experience with laser
photocoagulation, we observed that both the heater probe and BICAP represented a significant
advance for attaining hemostasis on an emergency basis. Specific advantages include portability,
capability for tamponade, ability to coagulate tangentially, capacity for washing, efficacy, and safety. All
of these characteristics of heater probe and BICAP also represent significant advantages over
monopolar electrocoagulation.
Randomized Controlled Trials of

Heater Probe Versus Other Therapies for Ulcers With
Active Bleeding or Nonbleeding Visible Vessels
TABLE 285
JENSEN
Active
Bleed*
Patients
HP
MED
Other
Primary hemostasis
HP
MED
Other
More bleeding
HP
VV
LIN
Active
Bleed
VV
18
14
15
22
27
27
29
24
28
16
22
18
95%
14%
93%
98%
38%
67%
23%
23%
18%
12%
Randomized Controlled Trials of

Heater Probe Versus Other Therapies for Ulcers With
Active Bleeding or Nonbleeding Visible Vessels
TABLE 285
JENSEN
Active
Bleed*
VV
LIN
Active
Bleed
VV
MED
93%
52%
62%
50%
Other
47%
35%
39%
17%
Surgery
HP
11%
0%
(7%*)
MED
57%
33%
(26%)
Other
29%
28%
(4%**)
Death
HP
5%
0%
(2%)
MED
21%
0%
(15%)
Other
7%
5%
(0%**)
VVa nonbleeding visible vessel.

HPheater probe treatment.
MEDmedical treatment, including an H2 receptor antagonist.
* Active bleed in this study is arterial bleeding, either spurting or
moderate.
Active bleed in this trial includes equal numbers of patients with
oozing and spurting.
Other includes other endoscopic treatment groups, which were
electrocoagulation (BICAP) for the Jensen study and alcohol
injection for the Lin trial.
Indicates a significant difference (p < .05) between heater probe
and medical treatment.
The results for all patients with active bleeding and nonbleeding
visible vessels together.
**Indicates a significant difference (p < .05) between the other
treatment and the medical treatment.
Lin et al.51(1603) randomized 137 patients with ulcers and major stigmata of hemorrhage to receive
medical treatment, alcohol injection, or treatment with the heater probe. Compared with medical
treatment, heater probe treatment had significantly higher rates of permanent hemostasis, lower
mortality, lower rates of emergency surgery, and fewer hospital days. Compared with medically treated
patients, those treated with the heater probe also had a significantly lower mortality and rate of
emergency surgery (see Table 285).
Standardization of the endoscopic treatment parameters such as probe size, tamponade pressure, site
of application, and total energy applied may have influenced the results of heater probe studies as
much as inclusion of high-risk patients, experience and skill of the endoscopists, and size of the
study.22(1604) For example, Matthewson et al.49(1605) coagulated around the bleeding site or
nonbleeding stigmata first (using firm pressure), whereas CURE investigators applied very firm
tamponade first followed by coagulation directly on the visible vessel or the site of active bleeding. The
small-diameter probe was used in 26% of Matthewson's cases and less than 5% of CURE patients.
The joules per tamponade station varied from 60 to 9O J for Matthewson et al.,49(1606) compared
with 120 to 150 J (CURE). These technical differences are significant and influence treatment results,
especially with contact probes.
Other Gastrointestinal Lesions

Upper GI Angioma Syndromes
No randomized controlled studies of endoscopic hemostasis for bleeding upper GI angiomata,
Osler-Weber-Rendu telangiectasia, or watermelon stomach have been reported. However, several
prospective studies have been reported for heater probe and bipolar electrocoagulation. The CURE
Hemostasis Group reported the largest prospective study of endoscopic treatment for bleeding upper
GI angiomata.9,11(1607) Both upper GI angiomata and Osler-Weber-Rendu telangiectasia were
included. The criterion for a definitive endoscopic diagnosis of a vascular lesion as the bleeding site
was active bleeding or stigmata of recent bleeding associated with a vascular lesion (clot, erosion, or
spot), or hematemesis with a nonbleeding angioma without stigmata and no other upper GI bleeding
source. In the absence of these criteria, an angioma noted at endoscopy in a patient with chronic GI
bleeding was considered an incidental lesion. Outcomes were compared for groups of patients before
and after endoscopic treatment; compared with prior medical management (mean 2.5 years) treatment
with BICAP or heater probe significantly improved outcomes. Significant increases in hematocrit were
seen, and transfusion requirements and hospitalizations for bleeding decreased significantly. The
estimated cost of hospitalization and blood transfusions was reduced at least twofold by endoscopic
hemostasis compared with medical management. No major complications were seen. Endoscopic
treatments with heater probe or BICAP were technically easy, safe, and efficacious.
Watermelon Stomach, Radiation Telangiectasia, and Lower GI Angiomata
More recently, the CURE hemostasis group has applied standardized techniques with heater probe
and bipolar electrocoagulation to treat patients with GI bleeding due to other GI angioma syndromes.
These include patients with watermelon stomach,38(1608) radiation telangiectasia,13(1609) and lower
GI angiomata (see Figure 2810 and Tables 281 and 282).3(1610) In all of these syndromes,
patients have had significant improvements in outcomes such as transfusion requirements,
hospitalization for GI bleeding, and hematocrits (statistical comparisons, 2 or more years on medical
therapy versus 2 or more years on endoscopic treatment). No complications have been noted except
for bleeding right colon angiomata. In the 10-year prospective study of See et al.,37(1611) which
included more than 100 patients with right colon angiomata treated by endoscopic coagulation, the
complication rate for bipolar coagulation was 3.7%, versus 8.4% for heater probe coagulation. The
major complications were postcoagulation syndrome (due to transmural coagulation) and delayed
bleeding from ulcers (induced by treatment of right colon angiomata).
Petrini and Johnston53(1612) reported a series of 12 patients with watermelon stomach who
underwent endoscopic treatment with the heater probe. Four sessions were required on average, after
which the appearance of the antrum was said to have returned to normal. No bleeding from the antral
vascular ectasias was evident during follow-up (mean 20.9 months). After treatment, none of the
patients required transfusions, although transfusions for other conditions were required in 2 of the 10
patients who had received transfusions prior to endoscopic therapy.
REFERENCES
1.
2.
Jensen DM. Severe Non-Variceal Upper Gastrointestinal Hemorrhage. Philadelphia: WB

Saunders. 1991.
Kovacs TOG, Jensen DM. Electrothermal modalities in the treatment of non-variceal upper
gastrointestinal bleeding. Eur J Gastroenterol Hepatol 1990;2:8793.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
Jensen DM. Endoscopic control of nonvariceal upper gastrointestinal hemorrhage. In Yamada

T, Alpers D, et al. (eds). Textbook of Gastroenterology. Philadelphia: JB Lippincott,
1991;261834.
Jensen DM. New developments in the diagnosis and treatment of severe upper
gastrointestinal bleeding. Gastrointestinal emergencies. In Tytgat GNJ, Van Blankenstein M
(eds). Current Topics in Gastroenterology and Hepatology. Stuttgart: Georg Thieme Verlag,
1990;422.
Kovacs TOG, Jensen DM. Endoscopic diagnosis and treatment of bleeding Mallory Weiss
tears. Gastrointest Endosc Clin North Am 1991;1:387400.
Miko TL, Thomazy VA. The caliber persistent artery of the stomach. A unifying approach to
gastric aneurysm, Dieulafoy's lesion, and submucosal arterial malformation. Hum Pathol
1988;19:91421.
Savides TJ, Jensen DM, Cohen J, et al. Severe upper gastrointestinal tumor bleeding:
Endoscopic findings, treatment, and outcome. Endoscopy 1996;28:24458.
Swain CP, Storey DW, Bown SG, et al. Nature of the bleeding vessel in recurrently bleeding
Machicado GA, Jensen DM. Upper gastrointestinal angiomata: Diagnosis and treatment.
Gastrointest Endosc Clin North Am 1991;1:24162.
Jensen DM, Machicado GA. Upper gastrointestinal angiomata: Diagnosis and endoscopic
treatment. InJensen DM, Brunetaud JM (eds). Medical Laser Endoscopy. Dordrecht,
Netherlands: Kluwer Academic, 1990;7191.
Jensen D, Machicado G. Endoscopic diagnosis and treatment of bleeding colonic angiomas
and radiation telangiectasia. Persp Colon Rectal Surg 1989;2:99113.
Savides TJ, Jensen DM. Colonoscopic hemostasis of recurrent diverticular hemorrhage
associated with a visible vessel: A report of three cases. Gastrointest Endosc 1994;40:703.
Jensen DM. GI endoscopic hemostasis and tumor palliation Experimental results and
techniques. In Jensen DM, Brunetaud JM (eds). Medical Laser Endoscopy. Dordecht,
Netherlands: Kluwer Academic, 1990;4570.
Johnston JH, Jensen DM, Auth D. Experimental comparison of endoscopic
Sigel B, Dunn MR. The mechanism of blood vessel closure by high frequency
electrocoagulation. Surg Gynecol Obstet 1965;121:82331.
Sigel B, Hatke FL. Physical factors in electrocoaptation of blood vessels. Arch Surg
1967;95:548.
Line L. Bipolar-multipolar electrocoagulation of bleeding ulcers. Gastrointest Endosc Clin
North Am 1991;1:291302.
Jensen D, Hirabayashi K, CURE Hemostasis Research Group. A study of coagulation depths
with BICAP and heater probe to improve endoscopic hemostasis of bleeding peptic ulcers
[Abstract 131]. Gastrointest Endosc 1989;35:181.
Jensen DM, Hirabayashi K. A comparative study of coagulation depths and efficacy of arterial
coagulation for Gold probe [Abstract 90]. Am J Gastroenterol 1989;84:1161.
Protell RI, Rubin CE, Auth DC, et al. The heater probe: A new endoscopic method for
stopping massive gastrointestinal bleeding. Gastroenterology 1978;74:25762.
Jensen DM. Heater probe for endoscopic hemostasis of non-variceal UGI bleeding. InSugawa
C, Schuman B, Lucas C (eds). Gastrointestinal Bleeding. New York: Igaku-Shoin,
1992;298313.
Jensen DM. Heat probe for hemostasis of bleeding peptic ulcers: Techniques and results of
randomized controlled trials. Gastrointest Endosc 1990;36:S429.
Jensen DM. Heater probe for endoscopic hemostasis of bleeding peptic ulcers. Gastroenterol
Piercey JRS, Auth DC, Silverstein FE, et al. Electrosurgical treatment of experimental
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
bleeding gastric ulcer: Development of a computer control and better electrode.

Swain CP, Mills TN, Shemesh E, et al. Which electrode? A comparison of four endoscopic
methods of electrocoagulation in experimental bleeding ulcers. Gut 1984;25:14249.
Frietas D, Donato A, Monteiro JG. Controlled trial of liquid monopolar electrocoagulation in
bleeding peptic ulcers. Am J Gastroenterol 1985;80:8537.
Papp JP. Endoscopic treatment of gastrointestinal bleeding: Electrocoagulation. In Sivak MV
(ed). Gastrointestinal Endoscopy. Philadelphia: WB Saunders, 1987;14357.
Papp JP. Endoscopic electrocoagulation in the management of upper gastrointestinal tract
bleeding. Surg Clin North Am 1982;62:797806.
Sugawa C, Shier M, Lucas CE, Walt AJ. Electrocoagulation of bleeding in the upper part of
the gastrointestinal tract: A preliminary experimental clinical report. Arch Surg
1975;110:9759.
Moreto M, Zaballa M, Ibanez S, et al. Efficacy of monopolar electrocoagulation in the
treatment of bleeding gastric ulcer: A controlled trial. Endoscopy 1987;19:546.
Gaisford WD. Endoscopic electrohemostasis of active upper gastrointestinal bleeding. Am J
Surg 1979;137:4753.
Wara P. Endoscopic electrocoagulation of major bleeding from peptic ulcer. Acta Chir Scand
1985;151:2935.
Wadas DD, Sanowski RA. Complications of the hot biopsy forceps technique. Gastrointest
Endosc 1988;34:327.
Rogers BHG. Endoscopic electrocoagulation of vascular abnormalities of the gastrointestinal
tract in 51 patients [Abstract]. Gastrointest Endosc 1982;28:142.
Jensen DM, Kovacs TOG, Freeman M, et al. A multicenter randomized prospective study of
Gold probe vs. heater probe for hemostasis of very severe ulcer of Mallory-Weiss bleeding
[Abstract]. Gastroenterology 1991;100:A92.
Laine L. Determination of the optimal technique for bipolar electrocoagulation treatment. An
experimental evaluation of the BICAP and Gold probes. Gastroenterology 1991;100:10712.
See JA, Jensen DM, Jutabha R, et al. Complications of endoscopic coagulation of bleeding
colonic angiomata: Post coagulation syndrome and delayed bleeding. A 10 year prospective
study [Abstract]. Gastrointest Endosc 1993;39:304.
Jensen DM, Kovacs TOG, Randall GM, et al. Prospective randomized study of patients with
watermelon stomach vs. other UGI angioma syndromes treated with bipolar or heater probes
[Abstract]. Gastroenterology 1994;106:A241.
Freeman ML. The current endoscopic diagnosis and intensive care unit management of
severe ulcer and other nonvariceal upper gastrointestinal hemorrhage. Gastrointest Endosc
Laine L. Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract
hemorrhage. A prospective controlled trial. N Engl J Med 1987;316:16137.
Laine L. Multipolar electrocoagulation in the treatment of peptic ulcers with nonbleeding visible
O'Brien JD, Day SJ, Burnham WR. Controlled trial of small bipolar probe in bleeding peptic
ulcers. Lancet 1986;I:4647.
Kernohan RM, Anderson JR, McKelvey ST, Kennedy T. A controlled trial of bipolar
electrocoagulation in patients with upper gastrointestinal bleeding. Br J Surg 1984;71:88991.
Goudie BM, Mitchell KG, Birnie GG, Mackay C. Controlled trial of endoscopic bipolar
electrocoagulation in the treatment of bleeding peptic ulcers [Abstract]. Gut 1984;25:A1185.
Brearley S, Hawker PC, Dykes PW, Keighley MR. Per-endoscopic bipolar diathermy
coagulation of visible vessels using a 3.2 mm probeA randomized clinical trial. Endoscopy
1987;19:1603.
Laine L. Multipolar electrocoagulation versus injection therapy in the treatment of bleeding
peptic ulcers. A prospective, randomized trial. Gastroenterology 1990;13036.
Rutgeerts P, Vantrappen G, Van Hootegem Ph, et al. Neodymium-YAG laser
48.
49.
50.
51.
52.
53.
photocoagulation versus multipolar electrocoagulation for the treatment of severely bleeding

ulcers: A randomized comparison. Gastrointest Endosc 1987;33:199202.
Waring JP, Sanowski RA, Sawyer RL, et al. A randomized comparison of multipolar
electrocoagulation and injection sclerosis for the treatment of bleeding peptic ulcer.
Matthewson K, Swain CP, Bland M, et al. Randomized comparison of Nd:YAG laser, heater
probe, and no endoscopic therapy for bleeding peptic ulcers. Gastroenterology
1990;98:123944.
Fullerton GM, Birnie GG, MacDonald A, Murray WR. Controlled study of heater probe in
bleeding peptic ulcers. Br J Surg 1989;76:5414.
Lin HJ, Tsai YT, Lee SD, et al. A prospectively randomized trial of heat probe
thermocoagulation versus pure alcohol injection in nonvariceal peptic ulcer hemorrhage. Am J
Chung SCS, Leung JWC, Sung JY, et al. Injection or heat probe for bleeding ulcer.
Petrini JL Jr, Johnston JH. Heat probe treatment for antral vascular ectasia. Gastrointest
Endosc 1989;35:3248.
Chapter 29 Laser Therapy

(1613)
(1614)
CHRISTOPHER PAUL SWAIN, M.D.
Studies using endoscopic laser treatment have generated a new sense of optimism that endoscopic
treatment can improve outcomes for patients with bleeding peptic ulcers. Most of the 15 randomized,
controlled trials of neodymium:yttrium-aluminum-garnet (Nd:YAG) and argon laser treatment compared
with no endoscopic therapy showed significant improvements in outcome with laser treatment.
Metaanalysis of all laser trials shows highly significant improvement for all major endpoints, including
recurrent bleeding, urgent surgery, and mortality. Lasers are more expensive than alternative
endoscopic devices, such as injection treatment and the multipolar or heater probe, for the treatment
of peptic ulcer bleeding, but a large body of evidence from clinical, controlled, and comparative trials
supports the efficacy and safety of laser therapy compared with any other known treatment.
Lasers have had a pivotal role in the development of endoscopic treatment for gastrointestinal
bleeding. Laser therapy was the first endoscopic therapy for hemostasis to be assessed in large
numbers of randomized, controlled trials. The evidence of efficacy in the treatment of bleeding is
greater than that for any other endoscopic treatment method. Larger numbers of patients have been
studied in randomized trials. Cumulative analysis of all trials suggests that laser-treated patients are at
significantly lower risk of further bleeding, have a significantly reduced requirement for urgent surgery,
and have a lower mortality rate than those treated with other methods. No other therapy for
gastrointestinal bleeding has been as rigorously tested. Although other, nonlaser endoscopic methods,
such as injection and thermal probe methods, are used more frequently in the treatment of acute
gastrointestinal bleeding because they are cheaper and more portable, endoscopic laser therapy
retains several theoretical and practical advantages.
Although Einstein1(1615) suggested the possibility of laser emission in a Nobel Prize-winning paper on
the physical basis of light in 1917, the first laser was not constructed until 1960 by Maiman.2(1616) In
1970, Goodale et al.3(1617) controlled bleeding from gastric erosions using a CO2 laser light through a
rigid cystoscope passed through a gastrotomy. Nath et al.4(1618) demonstrated the feasibility of
passing argon and Nd:YAG laser energy through a flexible fiberoptic delivery system within a
gastroscope in 1973. By the mid-1970s, several groups in Europe and the United States had
experimented with the efficacy and safety of argon and Nd:YAG lasers in animal models of
gastrointestinal bleeding and had used endoscopic laser therapy successfully in the treatment of
bleeding patients.
Figure 291 shows the red helium-neon aiming beam of the Nd:YAG laser waveguide that is protruding
from the distal end of an endoscope; Figure 292 shows the blue-green aiming beam of the argon ion
laser.
(1619)Figure 291. The red helium-neon aiming beam emerging from the
neodymium:yttrium-aluminum-garnet (Nd:YAG) laser waveguide in a flexible endoscope.
(1620)Figure 292. The blue-green aiming beam of an argon laser in an endoscope.
Tissue Interactions During Photocoagulation

Photons of laser light striking a tissue surface may be reflected, absorbed, or scattered and then
absorbed, or the photons may pass directly through the tissue. For hemostasis, the most important of
these mechanisms are absorption and scattering within the tissue, with subsequent conversion of laser
light energy to heat within the volume of tissue in which the beam is scattered.
The effect of the heat developed within the tissue varies with the temperature achieved and duration of
time during which the tissue is heated. Heating between 37 and 60 C accelerates and alters
membrane permeability characteristics. Between 60 and 80 C, protein denaturation occurs, with the
loss of cellular metabolic and membrane functions and degeneration of structural proteins. Hemoglobin
pigment oxidizes and turns black at 80 C. At 100 C, water boils and causes cells to explode and
thereby produce superficial tissue erosions. At higher temperatures, charring, carbonization, and
smoking occur.
Because of varied absorption of light by tissue components at different wavelengths, lasers can
produce a variety of tissue effects. The three lasers most widely used in medicine, the argon, Nd:YAG,
and CO2, produce very different tissue effects, which affect their hemostatic qualities. These
differences can be explained by considering the absorption in hemoglobin and water of the
wavelengths of light produced by these lasers (Figure 293A).
(1621)Figure 293. A, Absorption characteristics in water and hemoglobin of the argon,

Nd:YAG, and CO2 lasers. B and C, Comparison of the immediate and the delayed tissue
effects of the argon, Nd:YAG, and CO2 lasers at optimum hemostatic settings.
Water does not absorb the blue-green light of the argon laser; it is almost completely absorbed by
blood at a depth of penetration of only 1 mm. The argon laser tends to coagulate to a depth of 1 mm at
optimum settings and without an intervening pool of blood to absorb its energy. The light from the
Nd:YAG laser in the near-infrared region of the electromagnetic spectrum penetrates water less well
than that of the argon laser, but its penetration of blood is much better; less than a fourth of its energy
is absorbed by 1 mm of blood. At optimum hemostatic settings, the Nd:YAG laser can produce a
coagulative effect to a depth of 5 mm within tissue. Because the CO2 laser light is emitted in the
far-infrared electromagnetic spectrum and is intensely absorbed by water, this laser tends to coagulate
vessels to a depth of only about 0.5 mm.
Because of wide scattering of photons within tissue, all of the described lasers produce tissue damage
lateral to the small area where the beam strikes tissue (see Figure 293B). This effect is more obvious
histologically at 1 week rather than immediately after treatment (see Figure 293C). Scattering within
tissue explains why laser light retains its hemostatic effects when tissue is approached tangentially and
the angle of incidence of the laser beam is acute rather than perpendicular. The Nd:YAG laser
penetrates more deeply than the argon laser and produces a wider tissue effect because of greater
lateral scattering. Almost 10 times more energy is required to achieve an optimum hemostatic effect,
because the Nd:YAG laser heats a much larger volume of tissue than the argon laser. Laser energy is
selectively absorbed by the red color of hemoglobin and is transmitted through structures that contain
less of this pigment. One theoretical advantage of laser hemostasis over other methods is that it is
"blood seeking."
Vessel shrinkage is the single most important mechanism involved in the immediate effect of the
laser.5(1622) Vessel contraction mainly results from alterations in structural protein, especially
collagen, in the vessel wall. Histologic examination of ulcers treated with argon or Nd:YAG lasers
suggests that intravascular thrombosis is not an important mechanism of hemostasis.6,7(1623) Edema
around the bleeding point has been suggested as a hemostatic mechanism.8,9(1624) However, this
seems improbable, because injection of water around vessels in experimental ulcers or vessels within
tissue has no hemostatic effect10(1625) and lasers can effectively coagulate isolated
vessels.1113(1626)
The size of a vessel limits occlusion by a beam of laser light or by any thermal or injection method.
Small vessels of up to 0.5 mm are easily sealed by the Nd:YAG laser.11,13(1627) It becomes
increasingly more difficult to seal vessels of 1 mm and greater diameters.1214(1628) In one series of
patients coming to gastrectomy after recurrent bleeding from a gastric ulcer, 5 (19%) of 27 patients
had arteries that were greater than 1 mm in diameter,15(1629) and 47% of patients dying of bleeding
duodenal ulcers had arteries that were larger than 1 mm in diameter.16(1630) It is possible to stop
bleeding from small holes in large vessels using the Nd:YAG laser without occluding the
vessel.12,17,18(1631)
Neodymium:Yttrium-Aluminum-Garnet Laser
The lasing material of the Nd:YAG laser is a synthetic crystal rod of yttrium, aluminum, and garnet
(YAG) that holds a small concentration of neodymium atoms (Nd). Blue-green light, emitted by a
krypton lamp and focused into the rod by an elliptical reflector, excites (pumps) the neodymium atoms
into a band of high energy states, which then decay. Lasing, predominantly at the near-infrared
1064-nm wavelength, results from transition of metastable higher energy states to lower energy states,
which then decay rapidly through nonradiative transitions to the ground state (see Chapter 10: Laser
Physics and Laser-Tissue Interactions). The laser cavity is completed by mirrors positioned at each
end of the crystal rod. The output mirror is usually multicoated to allow only the desired wavelength to
lase. This laser emits in the near-infrared part of the spectrum at 1064 nm; alternative emission at
1318 nm is also possible but is used less commonly.
Efficacy and Safety

Animal Models of Ulcer Bleeding
Studies by Silverstein et al.19(1632) published in 1979 demonstrated that the Nd:YAG laser was
relatively ineffective at a power output of 15 watts (W) in stopping bleeding from experimental ulcers,
but it became very effective at 50 W, although these higher power settings caused full-thickness tissue
damage in two thirds or more of experimentally produced ulcers. Several groups performed studies,
especially in the standard canine model of bleeding gastric ulcer, to establish optimum settings for
laser photocoagulation. Johnston et al.20(1633) found optimum settings to be 70 W at a pulse duration
of 0.5 sec, although full-thickness tissue damage occurred in more than half the ulcers treated. Coaxial
gas flow (CO2) around the laser delivery fiber, which had been shown to be essential to enhanced
photocoagulation with the argon laser, did not enhance the high efficacy of the Nd:YAG laser. Bown et
al.21(1634) found optimum pulse duration to be between 300 and 500 ms and optimum pulse energy
to be 20 to 40 Joules (J). Rutgeerts et al.22(1635) demonstrated that short pulses of Nd:YAG laser
radiation (60 W at 1 sec or 70 W at 0.5 sec) resulted in a lesser depth of tissue damage without loss of
efficacy compared with continuous irradiation (60 W).
Uncontrolled Clinical Trials for Bleeding Peptic Ulcer
Beginning in 1977, Kiefhaber,23,24(1636) working with Nath, popularized the use of the Nd:YAG laser
as an endoscopic method for treatment of bleeding peptic ulcer. In 1990, Kiefhaber et al.,25(1637)
again using the Nd:YAG laser, reported the largest uncontrolled series of endoscopically treated
patients with gastrointestinal bleeding; 996 (94%) of 1058 episodes of active bleeding were
successfully treated (i.e., primary hemostasis was achieved). Perforations occurred in fewer than 2%
of patients and occurred more commonly in patients with acute than those with chronic ulcers.
Randomized, Controlled Trials for Bleeding Peptic Ulcer
The results of 11 randomized, controlled trials of endoscopic Nd:YAG laser therapy have been
reported (Table 291).2637(1638) Seven of these trials demonstrated significant benefit for laser
treatment in at least one major parameter: episodes of further bleeding, requirement for urgent
surgery, or death.26,29,30,32,33,35,37(1639) Three studies showed a significant improvement in two
endpoints.30,32,33(1640) In 7 trials, a significant reduction in further bleeding was demonstrated, 3
found a reduction in the requirement for urgent surgery, and 2 found a reduction in mortality (Table
291). One trial demonstrated significant improvements in all endpoints.33(1641) A significant
reduction in transfusion requirement was found in 4 trials. The apparently negative results of some
trials with small numbers of patients can be interpreted as showing trends toward improvement in
outcome parameters with laser therapy.
TABLE 291
Results of Randomized, Controlled Trials of Nd:YAG Laser in the Treatment of Bl

REBLEEDING (%)
STUDY (YEAR)
TOTAL [n]
SUBGROUP (n)
SURGERY (%)
DEATH (%)
Control
Treatment
Control
Treatment
Control
Rhode et al. 26
(1980)
Ihre et al. 27
(1981)
Active bleed (105)
57
59
41
13
27
Total [25]
38
42
38
42
15
Escourrou et al. 28
(1981)
Total [83]
17
29
15
26
15
Rutgeerts et al. 29
(1982)
MacLeod et al. 30
(1983)
Homer et al. 31
(1985)
Trudeau et al. 32
(1985)
Active (67)
VV/clot (36)
38
32
13
23
3
14
16
14
Arterial bleed (20)

Spots (25)
100
0
100
0
13
0
25
0
VV/clot (42)
32
5
21
25
0
18
VV (33)
40
11
26
8
5
33
TABLE 291
Results of Randomized, Controlled Trials of Nd:YAG Laser in the Treatment of Bl

REBLEEDING (%)
STUDY (YEAR)
Swain et al. 33
(1986)
Krejs et al. 34
(1987)
Buset et al. 35
(1988)
TOTAL [n]
SUBGROUP (n)
SURGERY (%)
DEATH (%)
Control
Treatment
Control
Treatment
Control
Total [138]
Spurting (29)
VV (50)
Minor SRH & clot
(59)
Total [174]
Active bleed (32)
40
80
48
16
10
20
14
3
35
10
12
20
33
22
29
15
20
14
18
1
0
VV in GU (42)
VV in DU (46)
41
32
15
33
20
Total [86]
42
32
20
2
Matthewson et al.
37 (1990)
VVvisible vessel; SRHstigmata of recent hemorrhage.
* In some studies, patients with bleeding were subgrouped according to nature of bleeding or SRH. Subgroups are indicate
p < .05 versus control.
No figures for this trial are found in the abstract of Escourrou et al. (1981) 28, but they have appeared in a chapter by Fleis
study did not randomize the patients but compared patients at one hospital where one operator used the laser and anothe
No trial data suggest that laser treatment adversely affected outcome. A perforation occurred in only 2
of more than 900 patients randomized to Nd:YAG laser treatment or a control group in these trials.
One occurred in a patient randomized to the control arm of one study;33(1642) the other patient had
undergone laser treatment.34(1643)
Reasons for Apparent Differences between Trials
It is possible that there are no real differences in outcomes between these trials and that a type II or
error explains the negative results of trials (i.e., that a positive result was missed simply because
numbers of patients randomized were insufficient to show a benefit). Most randomized, controlled trials
of laser therapy for peptic ulcer bleeding have included far too few patients. One way to overcome this
difficulty is metaanalysis or cumulative analysis of published results.
Metaanalysis of Trials
A metaanalysis of trials of laser photocoagulation for bleeding ulcers has been calculated by Henry and
White.38(1644) This analysis incorporated eight trials in which a total of 448 patients were randomized
to laser photocoagulation and 458 to no endoscopic treatment. Endoscopic photocoagulation was
performed with the Nd:YAG and argon lasers in five and three trials,
respectively.21,26,31,3943(1645) The overall results strongly favored laser photocoagulation over no
endoscopic therapy for all endpoints at a high level of statistical significance: recurrent bleeding (p <
.025), surgery (p < .005), and death (p < .005). Henry and White38(1646) concluded that endoscopic
therapy, particularly with the laser, had been shown, more rigorously than any other conventional
treatment including surgery, to be effective treatment for bleeding peptic ulcer. Their results also
suggested that endoscopic laser therapy, unlike thermal probe methods, significantly reduced
mortality, perhaps because more than twice the numbers of patients had been included in laser trials
than in trials of thermal probe methods. This analysis has been confirmed by two subsequent
metaanalyses.39,44(1647)
Technical Explanations for Different Trial Results
Other technical concepts may explain differences in the results of trials of laser therapy for bleeding
peptic ulcer. Early trials used Nd:YAG lasers with a maximum power output of 50 W and no built-in
power meter. Animal studies, not available when some of these trials were started, showed that short
pulses (0.5 to 1.0 sec) and higher power levels (70 to 80 W) were optimum for safe
hemostasis.2022(1648) It is probable that much lower power levels were used in the early studies.
The importance of the visible vessel as the target for treatment was not understood until the early
1980s. Trials that have stratified patients into this subgroup have generally shown significant benefit. It
is likely that laser energy directed indiscriminately around an ulcer will not be as effective as that
focused directly on the bleeding point of the vessel.
Comparisons with Other Endoscopic Therapy for Bleeding Peptic Ulcer

Rutgeerts et al.45(1649) compared Nd:YAG laser photocoagulation and treatment with a 3.2-mm
(10-French) multipolar probe in 100 patients presenting at endoscopy with bleeding peptic ulcers (i.e.,
spurting or oozing vessel) or ulcers with nonbleeding visible vessels. All were pretreated with an
injection of 4 to 8 ml of 1:10,000 epinephrine. Twenty patients had a second treatment because of
recurrent bleeding. There were no significant differences between the two treatment groups.
In a small, randomized study, Goff46(1650) treated 8 patients with Nd:YAG laser (1-sec pulses at 50
W) and 11 patients with multipolar electrocoagulation. No further bleeding was encountered in 37.5%
of laser-treated patients, compared with 47.4% of multipolar probe-treated patients, a difference that
was not statistically significant.
Rutgeerts et al.36(1651) compared the efficacy of Nd:YAG laser plus epinephrine injections,
epinephrine plus polidocanol injections, and injections of epinephrine alone (Table 292). This study
included a control group with nonbleeding visible vessels, but it did not include a control group with
bleeding visible vessels for ethical reasons. One hundred and forty patients with peptic ulcers and
bleeding or nonbleeding visible vessels were randomized. This study had some new design features. If
there was further bleeding, the initial endoscopic treatment at randomization was repeated at a second
session before considering surgery. Patients with recurrent bleeding in the control group also
underwent one endoscopic treatment before consideration of surgery. With respect to reducing further
bleeding, the results of this study suggest that epinephrine plus polidocanol injections (18 of 20
patients, p = .002) and epinephrine plus Nd:YAG laser (16 of 20 patients, p = .012) were both
significantly superior to the results in the control group (8 of 20 patients). Patients treated with
epinephrine plus polidocanol injections fared about the same as those treated by epinephrine injection
plus Nd:YAG laser photocoagulation.
Randomized, Controlled Trial of Different Combinations of Therapies for

Ulcer Hemostasis
TABLE 292
RATE OF RECURRENT BLEEDING*

TREATMENT
Control
Epinephrine alone
Epinephrine + polidocanol
Epinephrine + Nd:YAG laser
NUMBER OF PATIENTS
First Session
(%)
Second Session
(%)
20
40
40
40
60
40
18
28
10
8
13
Randomized, Controlled Trial of Different Combinations of Therapies for

Ulcer Hemostasis
TABLE 292
RATE OF RECURRENT BLEEDING*

TREATMENT
NUMBER OF PATIENTS
First Session
(%)
Second Session
(%)
Adapted from Rutgeerts P, Vantrappen G, Broeckaert L, et al. Comparison of endoscopic polidocanol injection
and YAG laser therapy for bleeding peptic ulcers. 1(8648):11647,
by The Lancet Ltd, 1989.
* For this randomized study of different endoscopic therapies for hemostasis of ulcer patients with major
stigmata, the rates of recurrent bleeding are shown after one or two treatment sessions.
p < .05 of this treatment versus control.
A small, randomized series was reported by Carter and Anderson.47(1652) Forty-four patients with a
visible vessel (23 with active bleeding) were randomized to undergo treatment with Nd:YAG laser or
epinephrine injection. Recurrent bleeding was encountered in 1 of the 21 Nd:YAG laser-treated
patients and in 3 of the 23 patients treated by epinephrine injection, a difference that was not
statistically significant. One patient died in the laser-treated group; there were no deaths in the injection
group.
Experiments using animals have shown that injection methods do not stop bleeding from ulcers or
isolated mesenteric or serosal vessels, but bleeding in these models can be easily stopped by thermal
methods (i.e., Nd:YAG laser, multipolar and heater probes).4850(1653) Extensive surgical experience
indicates that injection methods may be unreliable as a technique for hemostasis and are inferior to
thermal or mechanical measures.
The randomized, controlled trial of Matthewson et al.37(1654) comparing Nd:YAG laser, heater probe,
and no endoscopic therapy included 143 consecutive patients with stigmata of recent hemorrhage
accessible to laser therapy (Table 293). Bleeding recurred in 9 (21%) of 44 Nd:YAG laser-treated
patients compared with 16 (28%) of 57 patients treated with the heater probe and 18 (43%) of 42
patients who received no endoscopic therapy. Compared with the control group, there was a significant
reduction in recurrent bleeding (p < .05) for laser-treated patients. Although a trend suggested that
heater probe-treated patients had a lower rate of recurrent bleeding compared with that for the control
group, this result did not reach statistical significance.
Results of a Randomized, Controlled Trial

Comparing Nd:YAG Laser, Heat Probe, and No Endoscopic
Therapy in the Treatment of Patients With Bleeding Peptic
Ulcer
TABLE 293
REBLEEDING
Nd:YAG laser
Heat probe
Control
DEATH
NUMBER OF
PATIENTS
44
57
42
9*
16
18
21
28
43
1
6
4
2
11
10
Results of a Randomized, Controlled Trial

Comparing Nd:YAG Laser, Heat Probe, and No Endoscopic
Therapy in the Treatment of Patients With Bleeding Peptic
Ulcer
TABLE 293
REBLEEDING
NUMBER OF
PATIENTS
DEATH
Adapted from Matthewson K, Swain CP, Bland M, et al. Randomized

comparison of Nd:YAG laser, heater probe, and no endoscopic therapy for
bleeding peptic ulcers. Gastroenterology 1990; 98: 123944.
* p < .05 when compared with the control (i.e., no endoscopic treatment).
These studies confirmed Nd:YAG laser treatment as superior to no endoscopic treatment and
suggested that laser therapy was superior to heater probe treatment. The observation that results
equivalent to those of the previous trial of Nd:YAG laser therapy by Swain et al.33(1655) could be
obtained using the same protocol but with most patients being treated by a different operator
(Matthewson) is strong evidence that Nd:YAG laser is of benefit in the treatment of patients with
bleeding peptic ulcer.
Firm conclusions probably cannot be drawn from available studies comparing laser therapy with other
endoscopic methods. A type II error in comparative studies is more likely than in conventional
randomized, controlled trials; if both methods are somewhat effective compared with control treatment,
even larger numbers will be required to show significant differences between treatments being
compared. Available data appear to indicate that Nd:YAG laser is better or equal in efficacy in these
comparative studies.
Esophageal Varices
Experimental studies using a canine model of bleeding esophageal varices demonstrated that Nd:YAG
laser could be highly effective in arresting variceal bleeding induced by needle puncture. Laser
treatment was slightly more effective than sclerotherapy in terminating acute bleeding, but it did not
produce subsequent thrombosis of the varix, as did sclerotherapy.18(1656) Although there have been
enthusiastic clinical reports on the efficacy of Nd:YAG laser treatment of esophageal variceal bleeding
in humans,7,8(1657) Kiefhaber,24(1658) the most experienced operator, recommended that
sclerotherapy, a more definitive treatment, be performed early in the course of a bleeding episode.
Kiefhaber et al.24(1659) were able to terminate 160 of 174 acute episodes of variceal bleeding using
the Nd:YAG laser, but the rate of recurrent bleeding was 30%. In an early, controlled trial that included
patients with esophageal varices, Ihre et al.27(1660) found no benefit for Nd:YAG laser treatment.
Fleischer51(1661) conducted the only randomized, controlled trial of Nd:YAG laser treatment of varices
in patients with active bleeding (Table 294). Initial hemostasis was significantly greater in the
laser-treated patients than in the control patients, but the rate of recurrent bleeding was equal.
Available data suggest that laser therapy may be useful for control of acute variceal bleeding and of
some interim benefit, but it does not represent definitive treatment.
Results of a Randomized Trial of Nd:YAG Laser

for Treatment of Actively Bleeding Esophageal Varices
TABLE 294
in patients with active bleeding (Table 294). Initial hemostasis was significantly greater in the
laser-treated patients than in the control patients, but the rate of recurrent bleeding was equal.
Available data suggest that laser therapy may be useful for control of acute variceal bleeding and of
some interim benefit, but it does not represent definitive treatment.
Results of a Randomized Trial of Nd:YAG Laser

for Treatment of Actively Bleeding Esophageal Varices
TABLE 294
NUMBER OF
PATIENTS
INITIAL
HEMOSTASIS
REBLEEDING
DEATH
Laser
10
7
7
4
Control
10
0
7
7
p value
<.01
NS
NS
From Hashimoto D, Miyohara T, Yoshimura K. A lateral radiation probe in Nd:YAG
laser therapy. Gastrointest Endosc 1986; 32:1245; adapted from Fleischer D.
Endoscopic Nd:YAG laser therapy for active esophageal variceal bleeding.
Gastrointest Endosc 1985; 31:49.
NSnot significant.
Modifications of Laser Delivery Systems to Enhance Hemostasis

Lateral Fiber
Hashimoto et al.52(1662) described the use of a lateral radiation fiber in a study of Nd:YAG laser for
prophylactic treatment of esophageal and gastric varices inaccessible to treatment with standard
fibers. This device has a prism mounted in the distal tip of the waveguide that allows the laser beam to
exit at angles of 90 and 45 degrees relative to the long axis of the fiber.
Contact Tips
Materials that transmit Nd:YAG laser light without burning, melting, or sticking to tissue can be placed
on the distal end of the laser waveguide as an endcap. Commercially available contact laser endcaps
use a ceramic material of artificial sapphire, an oxide of aluminum.53(1663) This in effect converts the
laser waveguide to a contact probe that can be pressed against a vessel.14(1664) In animal
experiments, this lowers the power levels needed for hemostasis, because back-scattering is markedly
reduced and the area being heated is much smaller than that heated when laser energy is applied
without contact. Converting a laser waveguide to a contact probe also permits coaptation (i.e.,
application of pressure to a vessel to flatten the endothelial surfaces together before heat welding).
Coaptation allows for coagulation of larger vessels.54(1665) No randomized clinical studies have
suggested that contact laser methods offer advantages over the noncontact method for ulcer
hemostasis.
Liquid Light Guide
The standard Nd:YAG laser waveguide offers coaxial gas flow (CO2) around a central laser delivery
fiber. Sander et al.40, 55(1666) have substituted water for CO2 in the coaxial waveguide; because light
can be totally internally reflected at an air-water interface, the water can be used to contain the laser
radiation. Such a water-guided transmission system may offer some advantages; blood can be washed
away, the tip of the delivery fiber and the target surface are cooled, abdominal distention is minimized
without coaxial gas flow, and the "plume" of vaporized tissue is reduced. However, nonlaminar flow of
the water may cause asymmetric distribution of laser energy.
Sander et al.55(1667) presented data from a randomized study that suggest that ulcer hemostasis is
improved by the water jet technique. Twenty-eight patients with peptic ulcers and bleeding or
nonbleeding visible vessels were treated with conventional noncontact Nd:YAG laser, and 29 were
treated with water jet-guided Nd:YAG laser. Eleven (55%) of 20 patients with nonbleeding visible
vessels in the former group had intermittent arterial hemorrhage during treatment, compared with 6
(25%) of 24 with visible vessels in the latter group, a statistically significant difference (p < .05). Primary
hemostasis was achieved in 71% of the patients undergoing conventional treatment and in 90% of
those treated with the water jet method. Two (7%) patients in the water jet group had recurrent
bleeding and required surgery, compared with 5 patients (18%) operated for recurrent bleeding in the
conventional treatment group, a nonsignificant difference. Two deaths occurred in the noncontact
group, and none occurred in the water jet group.
Treatment Techniques for Bleeding Peptic Ulcer

Active Arterial Spurting
Active spurting hemorrhage is seen about 5% of patients with gastrointestinal bleeding (Figure
294).33,56,57(1668) This group has an 80% incidence of further bleeding and a high mortality rate.
The principles of laser treatment are similar for active spurting hemorrhage and the nonbleeding visible
vessel.
(1669)Figure 294. A, Spurting arterial bleeding from an ulcer on the proximal lesser
curvature of the stomach. B, Treatment with the Nd:YAG laser (note the position of the
aiming beam relative to the jet of blood). C, Ulcer after treatment.
Spurting can be something of a shock to the endoscopist. However, provided the patient was not
hypotensive at the start of the endoscopic procedure and intravenous access is available and adequate
for replacement of blood and fluids, there is no need to rush. The endoscopist should define the exact
point of bleeding, wash the ulcer to remove clot, and distort the jet of blood to pinpoint its source. As
with a nonbleeding visible vessel, laser pulses are first delivered in a tight circle around the bleeding
point, followed by pulses directly on the bleeding point. If this seems to produce little effect, increasing
power levels and pulse length (e.g., 90 J for 0.7 or 1 sec) may be helpful if blood flow is carrying the
thermal energy away from the bleeding point. Bleeding is usually intermittent and often stops
spontaneously before treatment. If bleeding obscures the endoscopic view, patience is often rewarded
by improvement in the view, because bleeding may slow after a few minutes. It is probably more
important to heat damage the intact vessel on either side of the bleeding point than to hit the opening
from which the blood is coming. The jet of blood sometimes ricochets off a bump or spot in the floor or
edge of the ulcer, giving a false impression that the bump or spot is the source of bleeding. For safety
and efficacy, it is important to move the pulses of laser energy around the target rather than hitting the
same spot repeatedly. Rutgeerts et al.36,48(1670) have recommended pretreatment of the spurting
ulcer with injections of 1:10,000 epinephrine before laser therapy.
Nonbleeding Visible Vessel
On most occasions, endoscopy does not reveal active hemorrhage in patients who present with
gastrointestinal bleeding. Treatment is given to those at high risk for further bleeding because of the
presence of a nonbleeding visible vessel. The latter feature is usually obvious as a raised mound of
darker color than the surrounding ulcer floor and with fresh blood nearby in the crater.58,59(1671)
Sometimes it is partially obscured by loose clot, mucus, or debris that must be cleared by gentle
washing.
The bleeding ulcer tends to be larger than the nonbleeding ulcer,60(1672) and complete examination
of the lesion may be challenging. Sometimes it is difficult to differentiate a visible vessel from a small
adherent clot within the ulcer. If in doubt, the endoscopist should check that another ulcer has not been
missed, especially at the sites known to be associated with major bleeding (e.g., posterior duodenal
bulb, proximal lesser curvature of the stomach), determine that the feature thought to be a visible
vessel is actually raised, and then inspect the ulcer to see if there is any other site that could be the
source of bleeding.
After establishing that a visible vessel is present, setting the laser at optimum power, and verifying that
the coaxial gas flow is adequate, the laser waveguide is passed through the accessory channel of the
endoscope. The tip of the waveguide should be visible in the endoscopic field and held about 1 cm
from the tissue. The aiming beam should be moved over the whole area that requires treatment to
ensure that all aspects are accessible. It may be helpful to check the laser effect biologically by firing a
sighter shot onto normal tissue at about 2 mm from the edge of the ulcer crater. This should produce a
clean white circle of thermocoagulated mucosa. If the power is too high, a circular erosion that breaks
the mucosal surface appears; if it is too low, the white circle of treated mucosa is difficult to see.
About eight pulses of laser energy should be delivered in a tight circle within 1 or 2 mm of the center of
the vessel. Treatment should be initiated at the distal aspect of the vessel, because this region is
usually the most difficult to target and becomes more difficult if bleeding occurs. Much of the target
area may be missed, but this is inconsequential provided several pulses are not placed at the same
spot. After eight pulses have been delivered in a circle, the center of the visible vessel should be
treated. Because this sometimes causes bleeding, it is advantageous that the waveguide be in the best
possible alignment with the visible vessel so that several pulses can be delivered quickly before the
view becomes obscured. Bleeding commonly arises from a previously nonbleeding visible vessel
during treatment, but it usually can be controlled by further treatment, and it often stops spontaneously.
The previously described technique is similar to that recommended by Fleischer and Bown61(1673)
and Rutgeerts et al.48(1674)
Bleeding From Under an Adherent Clot
If a large-volume floppy clot obscures the view of an ulcer, it is usually possible to move the clot away
by gentle, persistent washing. The aim is to wash without disturbing the underlying vessel; the roof of
an aneurysm or pseudoaneurysm should not be removed, because it is easier to treat a nonbleeding
vessel than a bleeding vessel. The force exerted by endoscopic washing should be gentle. If
performed properly, washing is unlikely to provoke active bleeding. Some experienced endoscopists
recommend pulling clot away from the crater with various grabbing accessories, but this occasionally
precipitates active bleeding or traumatizes the edge of the ulcer and causes it to ooze. Sometimes, it is
possible to see that a large, floppy clot narrows to become a plug in a vessel. Treatment is often
possible in this situation, but it may be difficult to target laser pulses in a circle around the vessel.
Washing is particularly helpful in clearing clot and oozing from an ulcer and establishing precisely the
source of the bleeding. If there is active bleeding but clot prevents treatment, it is reasonable to attempt
removal by mechanical means, because the clot is not contributing to the hemostasis, although the
prospect of hemostasis with laser therapy exists. Laser photocoagulation through a clot without being
able to identify the bleeding vessel is not useful, because blood tends to absorb the laser energy. In
trials of laser therapy, outcome was not improved when clots could not be cleared by
washing.33,37(1675) Occasionally, a clot is mistaken for a vessel, and treatment removes the clot to
reveal a clean ulcer floor.
Selection of Patients
The strongest case can be made for treatment of patients with spurting, oozing, or nonbleeding visible
vessels at initial diagnostic endoscopy. Among patients who have bled from ulcers, these parameters
encompass almost all who are at risk of further bleeding and death. A strong argument can be made
against treating patients without these stigmata of hemorrhage because they have a negligible
incidence of further bleeding and death.33,59(1676) A case could be made for treating all patients with
any stigmata of recent hemorrhage. In several trials, the overall results of treatment based on this
policy were superior to no endoscopic therapy.33,37,56(1677) Employing this approach, however,
would mean that patients with minor stigmata of recent hemorrhage (e.g., black spots) and low risk of
further bleeding would undergo unnecessary treatment, although this would be relatively safe.
Some investigators treat only patients with active bleeding.62(1678) This approach has certain
disadvantages. More than half of the patients at risk of further bleeding and death in most series do not
have active bleeding at initial endoscopy, and effective treatment for this group, which includes patients
with a nonbleeding visible vessel, is delayed. Other investigators recommend that treatment be
deferred until a further episode of substantial bleeding occurs in hospital.63,64(1679) This approach is
likely to be associated with increased transfusion requirements, morbidity, extra endoscopic
procedures in high-risk patients, and delays in the decision for definitive surgery if the treatment fails.
Some investigators recommend that treatment be avoided in patients with bleeding duodenal ulcers in
the posterior position, because erosion into the gastroduodenal artery is not amenable to endoscopic
therapy.65(1680) It is difficult to predict at endoscopy which duodenal ulcers with visible vessels are
likely to respond poorly to treatment, and it therefore seems reasonable to attempt to treat all such
lesions.66(1681) The precipitation of uncontrolled bleeding at endoscopy is not necessarily bad. The
patient would probably have bled again in any case, a precise diagnosis will have been established,
and the patient will be referred for surgery early in his or her course.
How many attempts at endoscopic therapy are acceptable before surgery is undertaken? In our trials,
all patients with evidence of recurrent bleeding after the initial therapeutic or control endoscopy were
considered for urgent surgery. Although endoscopic retreatment may have improved results in some
series,48(1682) it is nevertheless prudent to seek early surgical consultation for all patients with
bleeding that recurs in hospital after one attempt at endoscopic therapy. If the surgeon considers that
the risks of surgery are unacceptably high and further endoscopic therapy is regarded as technically
feasible and potentially helpful, further attempts to control bleeding by endoscopic means are
reasonable. However, the likelihood of a poor outcome increases when definitive surgery is postponed
by several unsuccessful attempts at endoscopic hemostasis.
Argon Laser
The argon laser is an example of the ion gas type of instrument. The laser medium is ionized argon
gas held in a tube sealed at each end by glass windows. A high-voltage current, constrained by a
magnetic field and flowing between electrodes positioned within each end of the tube, ionizes the gas
and excites the ions to higher energy states. Transitions between high-energy states and lower-energy
levels release photons in the blue-green region of the spectrum. The output of this laser is a beam of
light made up of a number of spectral lines or wavelengths between 347 and 527 nm. Eighty percent of
the power output is equally divided between two bands, one at 488 nm and the other at 514.5 nm, in
the blue-green region of the visible spectrum.
Studies of Ulcer Bleeding in Animal Models

The first report of argon laser photocoagulation through a flexible endoscope was that of Nath et
al.67(1683) in 1973. Initial studies of argon laser photocoagulation in animal models used low-power
output (1 W).6870(1684) Subsequent studies suggested that low power was ineffective for
hemostasis, although higher power outputs (5 to 7 W) were highly effective in stopping bleeding from
ulcers in animal models.41(1685) The ability to stop bleeding with an argon laser is impaired when the
bleeding is rapid because the laser energy is absorbed superficially by blood. To clear blood from the
ulcer surface, the waveguide with coaxial flow of CO2 gas was developed independently by two groups
of investigators.71,72(1686) Coaxial gas flow significantly decreased the number of pulses required to
stop bleeding and increased the number of ulcers effectively coagulated.15(1687) Tissue injury could
be minimized if the total energy delivered was limited to 50 J per ulcer,73(1688) by holding the
waveguide tip 1 to 2 cm from the ulcer (closer proximity caused tissue erosion and sometimes made
the bleeding worse), and by avoiding gaseous overdistention of the stomach, which thinned the area to
be treated and increased tissue damage.15(1689) When experiments were carried out with the use of
endoscopy, about twice as many pulses were required to achieve hemostasis than during laparotomy
because of the increased technical difficulty.15(1690)
Uncontrolled Clinical Trials

After the demonstration that argon laser photocoagulation was possible through a flexible endoscope
by Nath et al. in 197367(1691) and experimental animal studies indicating its hemostatic potential, the
first report of the use of the argon laser in the human gastrointestinal tract came from Frhmorgen et
al. in 1975.74(1692) The patient had bled from a colonic angioma. Four uncontrolled studies of argon
laser photocoagulation have been reported. In that of Brunetaud et al.,42(1693) bleeding was stopped
in 86% of the 33 patients treated. Bleeding was controlled in 90% of 100 patients in the series of
Frhmorgen et al.,75(1694) in 80% of 60 patients in that of Laurence et al.,76(1695) and in 76% of 67
patients treated by Bown et al.77(1696)
Randomized, Controlled Trials for Photocoagulation of Bleeding Peptic Ulcer

The first report of a randomized, controlled trial of argon laser photocoagulation was that of Vallon et
al.57(1697) Of 322 patients hospitalized with gastrointestinal bleeding, 136 had peptic ulcers; active
bleeding was found at endoscopy in 28 of these patients, who were randomized to laser treatment or a
control group. Permanent hemostasis was achieved in 10 (67%) of 15 of the laser-treated patients, but
only 4 (31%) of 13 of control patients stopped bleeding spontaneously, a statistically significant
difference (p < .05). Five laser-treated patients died, compared with 10 in the control group, a
difference that was not significant. There was little evidence that laser treatment prevented recurrent
bleeding in patients with a nonbleeding peptic ulcer and a red spot or a visible vessel, although a
nonsignificant trend favored laser treatment.
In the controlled trial of Storey et al.,58(1698) 76 patients with stigmata of recent hemorrhage were
randomized to no endoscopic treatment or to argon laser photocoagulation if the lesion was
accessible. Of the 52 patients with a visible vessel, 8 of 20 patients treated by laser and 17 of 28
control patients had further bleeding (p < .05). There were no deaths among laser-treated patients, but
seven control patients died (p < .05).
In the trial of Jensen et al.,43(1699) 16 patients with spurting or nonbleeding visible vessels were
randomized to argon laser photocoagulation or to a control group. Outcome was assessed using a
scoring system for adverse events such as episodes of further bleeding, requirement for urgent
surgery, and death. Jensen et al.43(1700) found an outcome score of 1.43 for argon laser-treated
patients, compared with 2.44 for control patients, which was a significant difference (p < .005). The
mean cost for care was less for argon laser-treated patients than for controls. The method of analysis
in this study is unusual, claiming to show a statistical benefit when assessed on a ranked outcome
analysis; larger numbers of patients are probably required for convincing demonstration of benefit.
Clinical Use of the Argon Laser

Despite reasonably convincing evidence of its efficacy in randomized, controlled trials, the argon laser
is no longer widely used for treating bleeding peptic ulcers, because the Nd:YAG laser has proved to
be safe and more effective. The Nd:YAG laser also can be used to treat obstructing gastrointestinal
tumors. The argon laser is more difficult to maintain so that it consistently delivers a high power output
(12 W) compared with the Nd:YAG laser, which is relatively rugged and has a maximum power output
of about 100 W.
Where argon and Nd:YAG lasers are both available, the former is preferred for the treatment of
vascular malformations and multiple polyps, especially in the colon, because of its greater margin of
safety.9(1701) There is a larger body of literature on the efficacy of argon laser photocoagulation for
vascular malformations than for any other endoscopic therapeutic modality.
Vascular Malformations
Endoscopic treatment of angiodysplasia can be beneficial. The best results are achieved when a
particular lesion accessible to endoscopic therapy is known to be the source of bleeding. The
intermittent chronic blood loss associated with multiple vascular malformations in the bowel remains
difficult to treat. Some patients require repeated hospitalization and transfusions. Vascular
malformations are particularly amenable to laser therapy, because their bright red color selectively
absorbs laser light, and because they are small and round, it is possible to coagulate many of them in
a single treatment session.
The terms angiodysplasia, vascular malformation, and angioma are used interchangeably in following
sections.
Patient Assessment
Patients fall into various categories according to the rate of bleeding, transfusion requirement, and
diagnostic difficulty. They may have continuous bleeding and major or minor transfusion requirements.
Some patients have intermittent melena or rectal bleeding. Others have no manifest bleeding but have
occult blood loss or recurrent iron deficiency anemia. Some patients have an anemia that can be
controlled with iron therapy despite continued occult blood loss. Some patients have incidental
angiodysplasia at colonoscopy or gastroscopy without evidence of anemia or bleeding. Patients with
incidental angiodysplasias without bleeding and those who can be maintained on iron therapy should
not necessarily undergo endoscopic treatment.
Preparation for Endoscopic Therapy

At the time of endoscopic therapy, it is helpful if the patient is not markedly anemic. A normal
hematocrit is manifested by a degree of greater color contrast between an angiodysplasia and the
surrounding mucosa, making it easier to identify and treat the lesion.
Accessories for washing the mucosal surface should be available. Washing makes it easier to identify
lesions that are potential bleeding sources. It helps to differentiate loose blood from angiodysplasia and
to locate the bleeding point in a lesion that is actively bleeding. If washing causes a vascular
malformation to bleed, it is likely the lesion is a significant source of blood loss. Normal mucosa does
not bleed on washing. The ability to wash the mucosal surface is also useful during active bleeding to
precisely identify a source.
Recording the Findings

The record of the procedure should provide enough information to allow others to estimate the degree
of confidence that the lesion seen and treated was an angiodysplasia that had caused significant
bleeding. The information provided should be adequate to allow another endoscopist to identify the
area where a lesion was found. In addition to site, the report should include the number of lesions
present, estimated sizes, and notations of any unusual features (e.g., depressed, umbilicated,
satellite). When numbers of vascular malformations are inaccurately recorded, it can be difficult to
determine if there are more or fewer lesions at a follow-up endoscopy. The presence or absence of
spontaneous bleeding and any bleeding induced by washing or treatment should be recorded.
Photographs are useful and provide additional information.
Technique
It can be helpful to fire a pulse of laser energy off target to verify that the equipment is working and that
power levels are satisfactory. If one of several vascular malformations is seen to bleed spontaneously,
that lesion should be treated immediately while the view is clear, because it is probably the major
single source of bleeding. If there are several vascular malformations, the largest should be treated
first, because it is most likely the one responsible for the bleeding. Because most large lesions have a
feeding vessel near the center, this region should be treated first. After bleeding starts, it is often
difficult to identify the center. Enough laser energy should be placed to produce a burn that
encompasses the entire lesion; ideal treatment results in a superficial ulcer at one day after treatment
that is equal in size to the angioma. Large angiodysplasias usually bleed regardless of which method is
used for treatment. This usually identifies a lesion as the probable source of significant bleeding.
Treatment-induced bleeding is common and of less clinical importance than is generally supposed.
The endoscopist should not continue to pulse energy into the same point if it does not stop bleeding,
because this increases the risk of perforation. Induced bleeding usually stops spontaneously. The
clinical pattern of relentless, life-threatening, recurrent bleeding seen with peptic ulcers is rare in
patients with angiodysplasia, even those with large transfusion requirements and large angiomas that
bleed on treatment.
Trials of Endoscopic Laser Treatment

Argon Laser Trials
The superficial mucosal effect of the blue-green argon ion laser makes it exceptionally safe for the
treatment of angiodysplasia. However, cost and major maintenance problems have confined its use to
only a few endoscopy units. Nonetheless, experience with endoscopic treatment of angiodysplasia, in
terms of numbers of patients treated and reported in the world literature, is greatest with the argon
laser (Table 295).9,7882(1702)
TABLE 295
Argon Laser Treatment of Vascular Malformations
STUDY
NUMBER OF
PATIENTS
Bowers and Dixon78
11
Waitman et al. 79
50
Bown et al. 80
18
Buchi and
Brunetaud9
126
Brunetaud et al. 81
53
Machicado and
Jensen82
26
OUTCOME
Reduction in bleeding and transfusion
requirements
Two thirds had no further bleeding; one
third had diminished bleeding and
transfusion requirements
Fourteen had marked reduction of
bleeding; average number of treatment
sessions was 5 for OWR and 2 for
non-OWR; 3 required surgery
Bleeding was controlled in 95% of
"sporadic" cases; significant reduction in
transfusion requirement in OWR
All 43 non-OWR patients stopped
bleeding; average number of treatment
sessions was two; 80% with further
bleeding were stopped with additional
treatment; all 10 OWR patients stopped
bleeding
Reduced transfusion requirements and
hospital admissions; mean follow-up of
33 months
OWROsler-Weber-Rendu syndrome.
COMPLICATIONS
None
None
None
None
None
None
Neodymium:Yttrium-Aluminum-Garnet Laser Trials

The Nd:YAG laser, which produces a more deeply penetrating emission in the near-infrared region of
the electromagnetic spectrum, has also been used extensively in the treatment of angiodysplasia
(Table 296).24,80,8388(1703)
TABLE 296
STUDY
Fleischer84
Johnston85
Nd:YAG Laser Treatment of Angiodysplasia

NUMBER OF
PATIENTS
OUTCOME
COMPLICATIONS
Bleeding reduced or stopped in all
None
22
Nineteen had marked reduction or

abolition of transfusion requirements
Kiefhaber et al.
24
30
Twenty-nine had marked reduction or

abolition of transfusion requirements
Three had significant

bleeding after treatment;
1 required surgery.
One required surgery for
further bleeding.
Rutgeerts et al.
83
56
Abolition or marked reduction of

transfusion requirements for 90%
Two patients had cecal

perforations.
Bown et al. 80
13
One had surgery.
Cello and
Grendell86
42
Naveau et al. 87
60
Sargeant et al.
88
41
Twelve had significant reductions in

transfusion requirements
Significant reduction in transfusion
requirements; results better for lower
than upper GI tract lesions
Fifteen had further bleeding; best results
were produced in patients with fewer
than four lesions and older than 75
years
Transfusion requirements reduced in
61%; recurrent bleeding in 22%
None
None
Death from perforation in

1 patient
The nontouch method of laser treatment and the selective effect of laser energy on blood containing
epithelium offer effective and precise therapy, especially for vascular lesions with large, deep
submucosal ramifications. The greater depth of penetration of Nd:YAG laser energy may be
advantageous for the treatment of deeper feeding vessels.81(1704) In two reported cases of cecal
perforation, angiodysplasia was treated with the Nd:YAG laser.83(1705)
Good results have been reported for laser treatment of Osler-Weber-Rendu syndrome and "sporadic"
angiomas, especially in the stomach, proximal jejunum, and right colon (see Table 296). Similar
results are reported with less common vascular abnormalities.
Gastric antral vascular ectasia (i.e., watermelon stomach) has been treated with excellent results.
Bleeding was stopped or markedly reduced in 13 patients treated with the Nd:YAG laser by Gostout et
al.89(1706) Cessation of bleeding was described in two case reports.90,91(1707) Bjorkman and
Buchi92(1708) treated seven patients with watermelon stomach, four with Nd:YAG laser and three with
the argon laser. The mean number of treatment sessions required to eliminate the need for
transfusions was significantly higher with the argon laser (5.75) than with Nd:YAG (2.33). Further
treatment was required for recurrent lesions in three patients treated with the argon laser, compared
with one patient treated by Nd:YAG. Buchi and Brunetaud9(1709) also reported effective control of
bleeding in three of four patients with blue rubber bleb nevus syndrome. Endoscopic laser treatment of
Von Willebrand's disease has been described.93(1710)
Compared with other vascular abnormalities, bleeding from radiation-induced gastrointestinal vascular
anomalies is more difficult to treat, although some patients respond to laser treatment. Alexander and
Dwyer94(1711) treated nine patients with chronic radiation proctocolitis with the Nd:YAG laser. Other
groups have reported successful treatment using Nd:YAG95, 96(1712) and argon97(1713) lasers.
Viggiano et al.98(1714) used the Nd:YAG laser to treat 47 patients with refractory rectal bleeding due
to radiation-induced vascular abnormalities in the rectum and sigmoid colon. Improvement was noted
in 41 patients. At 3 to 6 months after laser therapy, results showed a significant decrease in the
number of patients with daily bleeding per rectum and a significant increase of 2 gm/dl in the median
hemoglobin level. Complications occurred in three patients.
It is difficult to assess the value of laser treatment of vascular malformations, because the natural
history of this lesion is variable; with some patients, years may elapse between episodes of manifest
bleeding. Controlled trials have not been performed, and the number of patients required makes such
studies difficult. Comparison of transfusion requirements before and after treatment is poor evidence of
efficacy because bleeding may stop spontaneously. Vascular malformations may be found without
clinical evidence of bleeding.97(1715) Treatment is probably not required in such cases. New lesions
may appear, especially with the Osler-Weber-Rendu syndrome, and patients with vascular
malformations require careful follow-up.
Laser treatment of vascular malformations commonly precipitates bleeding. Bleeding occurred in 50%
of patients in one series.80(1716) However, induced bleeding usually stops spontaneously. The aim of
treatment is not to stop observed bleeding but rather to damage the abnormal vasculature so that
subsequent healing and scarring prevent reformation of these abnormal vessels. A few vascular
lesions are too large for successful or safe endoscopic laser therapy; these are better dealt with
surgically.
A vascular abnormality found and treated at endoscopy may not be the only source of persistent
bleeding. In one series, a range of associated abnormalities was found in patients with angiomas,
including bleeding colonic cancer, duodenal ulcer, and hemorrhoids.99(1717) Machicado and
Jensen82(1718) published data that suggest that endoscopic treatment is cost effective in the
management of patients with bleeding angiodysplasia (Table 297).
Cost Estimates of
Endoscopic Treatment of Angiomas
TABLE 297
TIME RELATIVE TO
COAGULATION
DETERMINING
FACTORS
Mean years of follow up
OWR
Non-OWR
Mean hospital costs
OWR
Non-OWR
Mean transfusion costs
Before
After
2
2
2
2
$10,400
7,400
2,895
$4,600*
2,800*
1,170*
Cost Estimates of
Endoscopic Treatment of Angiomas
TABLE 297
TIME RELATIVE TO
COAGULATION
DETERMINING
FACTORS
Before
After
Non-OWR
1,650
675*
Adapted from Machicado GA, Jensen DM. Upper
gastrointestinal angiomata: Diagnosis and treatment.
Gastrointest Endosc Clin North Am 1991; 1:24162.
OWROsler-Weber-Rendu syndrome.
* p < .05 for values after and before treatment.
Safety
Studies using animals during the late 1970s raised a concern that use of the Nd:YAG laser in the
gastrointestinal tract might be dangerous, because it usually produced full-thickness tissue damage in
the canine stomach when used at optimum power levels for hemostasis.19(1719) This led to initial
enthusiasm for the argon laser and the abandonment of monopolar electrocoagulation methods, with
the consequent development of safer multipolar and heater probe thermal devices. However, studies
using animals by Dixon et al.100(1720) demonstrated that an order of magnitude more power was
needed to produce perforation with the Nd:YAG laser than was required to secure hemostasis, and
that full-thickness tissue damage was rarely associated with perforation. Clinical studies of the Nd:YAG
laser showed that it was safe. For example, the perforation rate in Kiefhaber et al.'s huge experience of
treating more than 1000 bleeding episodes was 1%, with perforations occurring mainly with treatment
of acutely bleeding gastric ulcers, more commonly with a second course of treatment.23(1721) In the
11 randomized, controlled trials of Nd:YAG laser therapy (see Table 291), only one perforation was
reported in a laser-treated patient.34(1722)
In the comparative study of Rutgeerts et al.,45(1723) the incidence of perforation with Nd:YAG and
multipolar treatments was identical (i.e., 1 [2%] of 50 patients). In the study of Matthewson et
al.,37(1724) 1 (2%) of 57 patients sustained a perforation after treatment with the heater probe, and
none of 44 Nd:YAG laser-treated patients had a perforation. A perforation may develop spontaneously
in patients with bleeding peptic ulcer. One such patient was entered into the control arm of one
randomized trial.33(1725) Had she been randomized to the treatment arm of this trial, the perforation
would have been ascribed to laser treatment.
Bleeding from a previously nonbleeding visible vessel can be precipitated by laser treatment. In two
trials, the incidence was 15%57(1726) and 20%.33(1727) In all but one case, induced bleeding was
stopped by further treatment, or it stopped spontaneously. Two such patients had further bleeding that
was treated surgically. In the trial of Buset et al.,35(1728) the incidence of bleeding induced by Nd:YAG
laser treatment of nonbleeding visible vessel was reported as 50%, and in the study of Sander et
al.,55(1729) it was 55%, although bleeding was stopped with further treatment in most cases.
Cost and Efficacy

Lasers are expensive. The costs for patients who have recurrent bleeding from peptic ulcers after
hospital admission and require transfusion, urgent surgery, and prolonged periods in intensive care are
also high. If lasers prevent recurrent bleeding and thus convert the course of some patients with
bleeding peptic ulcer from a prolonged hospitalization to the short-term uncomplicated admission
experienced by most patients who have no further bleeding, the financial savings may be considerable.
A laser costs 8 to 10 times more than a thermal probe, and if these devices were of equivalent efficacy
to the Nd:YAG laser, the thermal probe would be 8 to 10 times more cost effective. Available evidence
suggests that the heater probe, which appears in experimental animal101(1730) and clinical
studies102(1731) to be the most effective of the thermal probes, is not of equivalent efficacy compared
with the Nd:YAG laser.37(1732) Injection methods are probably even cheaper, although the use of
disposable needles raises the cost somewhat.
Studies of the cost implications of other endoscopic methods of hemostasis (i.e., argon laser,
monopolar and multipolar electrocoagulation, and injection methods) suggest that the effective
reductions in the rate of recurrent bleeding and requirement for surgery achieved by endoscopic
means are associated with substantial savings in overall hospital costs.103107(1733)
Advantages of Lasers for Hemostasis

Laser energy produces tissue damage that is more predictable than with contact thermal methods.
Penetration through blood and tissue is greater with the Nd:YAG laser than with a multipolar or heater
probe. The latter devices were designed to produce a safe superficial burn and are therefore limited in
efficacy by inferior tissue penetration. Because laser energy is delivered at the speed of light, an
optimum treatment pulse requires only 0.3 to 0.5 sec. The time required to deliver optimum hemostatic
energy with the heater probe may be 8 to 10 sec because thermal conduction into tissue is relatively
slow. During the time required to deliver adequate energy, the heater probe tends to slide around the
ulcer, causing linear streak burns. In experimental studies, thermal methods, including laser, are far
superior to any injection method with respect to the ability to terminate bleeding from vessels of 0.5 to
2.0 mm in diameter.
Unlike other devices for endoscopic hemostasis, the laser can be easily converted from a noncontact
to a tissue contact device. Compared with other methods, thermal damage is more predictable at
acute angles of approach to ulcers with lasers. It produces less tissue damage than injection methods
of hemostasis at optimum hemostatic dose.
The laser is easier to use than other methods of endoscopic hemostasis. Bipolar and heater probes
are more effective with large-diameter (3.2-mm) probes and therefore require the use of nonstandard
endoscopes with large accessory channels. A laser can be used with any endoscope with a
conventional 2.8-mm diameter channel. Waveguides are also available for use with small-diameter
(e.g., pediatric) endoscopes. Because tissue contact is not required, laser treatment can be delivered
more precisely and easily, requiring less endoscopic skill. The aiming beam can be maneuvered
around the vessel before firing. There is no need to attempt to press on the vessel for inordinate
lengths of time as with thermal contact probes, which tend to slide about on the tissue surface during
coagulation. The laser produces less risk of trauma to the vessel with precipitation of bleeding, which
happens commonly with thermal probe and injection methods, and the endoscopists can easily deliver
several pulses of laser energy quickly if bleeding does start.
The reported body of clinical data on use of the laser in patients with gastrointestinal bleeding is far
larger than for any other endoscopic modality. The laser has an excellent safety record in clinical trials.
It has been extensively tested in animal studies and shown to be highly effective in terminating
bleeding from experimental ulcers, which is not the case for injection methods. The evidence of
efficacy in reducing further bleeding, the requirement for urgent surgery, and mortality in randomized,
controlled clinical trials is greater for laser therapy than for any other endoscopic modality.
Improvements in cavity design have increased efficiency, eliminated the need for a three-phase
electrical current, and allowed air cooling, removing the need to supply water and plumbing to the
laser. These developments have made the modern Nd:YAG laser more portable.108(1734) The
improved Nd:YAG laser is among the most rugged and reliable of lasers.
The laser has played an important role in the extraordinary changes that have taken place in the
therapy of gastrointestinal bleeding from peptic ulcer. In 1986, reviews of the problem stressed that no
therapy had been shown to alter outcome and that the mortality rate had not changed for 30 years. It is
mainly the result of studies of laser treatment that a more optimistic viewthat therapeutic endoscopy
offers significant benefitcurrently prevails.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Einstein A. Zur Quanten Theories der Strahlung. Phys Zeit 1917;18:1218.

Maiman TE. Stimulated optic radiation in ruby. Nature 1960;187:4934.
Goodale RL, Okada A, Gonzales R, et al. Rapid endoscopic control of bleeding gastric
erosions by laser radiation. Arch Surg 1970;101:2114.
Nath G, Gorisch W, Kiefhaber P. First laser endoscopy via a fiberoptic transmission system.
Endoscopy 1973;5:20813.
Gorisch W, Boergen K-P. Heat induced contraction of blood vessels. Lasers Surg Med
1982;2:113.
Kelly DF, Bown SG, Salmon PR, et al. Nature and extent of histological changes induced by
argon laser photocoagulation in canine gastric mucosa. Gut 1980;21:10417.
Kelly DF, Bown SG, Calder BM, et al. Histological changes following Nd:YAG laser
photocoagulation of canine gastric mucosa. Gut 1983;24:91420.
Dwyer RM. The technique of gastrointestinal laser endoscopy. In Goldman L (ed). The
Biomedical Laser. New York: Springer, 1981;2028.
Buchi KN, Brunetaud JM. Endoscopic laser therapy. In Dixon JA (ed). Surgical Applications of
Lasers. Chicago: Year Book Medical, 1987;95118.
Swain CP, Mills TN, Dark JM, et al. A comparative study of the safety and efficacy of liquid
and dry monopolar electrodes in experimental bleeding ulcers using computerized energy
monitoring. Gastroenterology 1984;86:93103.
Dennis MB, Silverstein FE, Gilbert DA, et al. Evaluation of Nd:YAG photocoagulation using a
new experimental ulcer model with a single bleeding artery. Gastroenterology 1981;80:1522.
Mills TN, Swain CP, Dark JM, et al. The "hot squeeze" bipolar forceps. A more effective
endoscopic method for stopping bleeding from large vessels in the gastrointestinal tract
[Abstract]. Gastrointest Endosc 1983;29:184A.
Johnston JH, Jensen DM, Auth D. Experimental comparison of endoscopic
Swain CP, Mills TN, Matthewson K, et al. Development and testing of contact endoscopic
thermal methods which are more effective than conventional lasers and diathermy electrodes
in controlling bleeding from the gastrointestinal tract [Abstract]. Gut 1984;25:1186A.
Johnston JH, Jensen DM, Mautner W, et al. Argon laser treatment of bleeding canine lesions:
Limitations and guidelines for endoscopic use. Gastroenterology 1981;80:70816.
Swain CP, Storey DW, Bown SG, et al. Nature of the bleeding vessel in recurrently bleeding
MacLeod IA, Bow CR, Joffe SN. A quantifiable bleeding gastric ulcer model in dogs for
assessing the neodymium YAG laser. Endoscopy 1982;14:910.
Jensen DM, Silpa ML, Tapia JI, et al. Comparison of different methods of endoscopic
hemostasis of bleeding canine esophageal varices. Gastroenterology 1983;84:145561.

19. Silverstein FE, Protell RL, Gilbert DA, et al. Argon vs neodymium YAG laser photocoagulation
of experimental canine ulcers. Gastroenterology 1979;77:4916.
20.
Johnston JH, Jensen DM, Mautner W, et al. YAG laser treatment of experimental bleeding
canine gastric ulcers. Gastroenterology 1980;79:125661.
21. Bown SG, Salmon PR, Storey DW, et al. Nd:YAG laser photocoagulation in the dog stomach.
Gut 1980;21:81825.
22.
Rutgeerts P, Vantrappen G, Geboes K, Broeckaert L. Safety and efficacy of neodymium-YAG
laser photocoagulation: An experimental study in dogs. Gut 1991;22:3844.
23. Kiefhaber P, Nath G, Moritz K. Endoscopical control of massive gastrointestinal haemorrhage
by irradiation with a high power neodymium-YAG laser. Prog Surg 1977;15:14055.
24.
Kiefhaber P, Kiefhaber K, Huber F, et al. Endoscopic applications of neodymium-YAG laser
radiation in the gastrointestinal tract. In Joffe SN (ed). Neodymium-YAG Laser in Medicine
and Surgery. New York: Elsevier, 1983;614.
25.
Kiefhaber P, Kiefhaber K, Huber F, et al. Ten years endoscopic neodymium-YAG laser
coagulation in gastrointestinal hemorrhage. In Jensen DM, Brunetaud JM (eds). Medical
Laser Endoscopy. Dordrecht, The Netherlands: Kluwer Academic Publishers, 1990;10918.
26.
Rhode H, Thon K, Fischer M, et al. Results of a defined concept of endoscopic
neodymium-YAG laser therapy in patients with upper gastrointestinal bleeding [Abstract]. Br J
Surg 1980;67:360A.
27. Ihre T, Johansson C, Seligson U, et al. Endoscopic YAG-laser treatment in massive
gastrointestinal bleeding. Scand J Gastroenterol 1981;16:63340.
28.
Escourrou J, Frexinos J, Bommelaer G, et al. Prospective randomized study of YAG
photocoagulation in gastrointestinal bleeding. In Atsumi K, Nimsakul N (eds). Proceedings of
Laser Tokyo '81. Vol 5. Tokyo: Inter Group Corp, 1981;30.
29. Rutgeerts P, Vantrappen G, Broeckaert L, et al. Controlled trial of YAG laser treatment of
upper digestive hemorrhage. Gastroenterology 1982;83:4106.
30. MacLeod IA, Mills PR, MacKenzie JF, et al. Neodymium yttrium aluminium garnet laser
photocoagulation for major haemorrhage from peptic ulcers and single vessels: A single blind
controlled study. Br Med J (Clin Res Ed) 1983;286:3458.
31. Homer AC, Powell S, Vicary FR. Is Nd:YAG laser treatment for upper gastrointestinal bleeds
of benefit in a district general hospital?. Postgrad Med J 1985;61:1922.
32.
Trudeau W, Siepler JK, Ross K, et al. Endoscopic Nd:YAG laser photocoagulation of bleeding
ulcers with visible vessels [Abstract]. Gastrointest Endosc 1985;31:138A.
33. Swain CP, Kirkham JS, Salmon PR, et al. Controlled trial of Nd-YAG laser photocoagulation in
34. Krejs GJ, Little KH, Westergaard H, et al. Laser photocoagulation for the treatment of acute
peptic-ulcer bleeding. A randomized controlled clinical trial. N Engl J Med 1987;316:161821.
35.
Buset M, Des Marez B, Vandermeeren A, et al. Laser therapy for nonbleeding visible vessels
in peptic ulcer hemorrhage: A prospective randomized study. Gastrointest Endosc
1988;34:A1734.
36. Rutgeerts P, Vantrappen G, Brocckaert L, et al. Comparison of endoscopic polidocanol
injection and YAG laser therapy for bleeding peptic ulcers. Lancet 1989;1:11647.
37. Matthewson K, Swain CP, Bland M, et al. Randomized comparison of Nd:YAG laser, heater
probe, and no endoscopic therapy for bleeding peptic ulcers. Gastroenterology
1990;98:123944.
38.
Henry DA, White I. Endoscopic coagulation for gastrointestinal bleeding. N Engl J Med
1988;318:1867.
40. Sander R, Poesl H, Prank P, et al. A Nd:YAG laser with a water-guided laser beamA new
transmission system. Gastrointest Endosc 1988;34:3369.
41. Silverstein FE, Auth DC, Rubin CE, Protell RL. High power argon laser treatment via standard
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
endoscopes. I. A preliminary study of efficacy in control of experimental erosive bleeding.

Brunetaud JM, Enger A, Herjot A, et al. La photocoagulation per endoscopique des lesion
digestives hemorrhagiques (nos premiers results chez l'homme a l'aide d'un laser a argon de
moyenne puissance). Gastroenterol Clin Biol 1975;2:5578.
Jensen DM, Machicado GA, Tapia JL, et al. Controlled trial of endoscopic argon laser for
severe ulcer hemorrhage [Abstract]. Gastroenterology 1984;86:1125.
Sacks HS, Chalmers TC, Blum AL, et al. Endoscopic hemostasis: An effective therapy for
bleeding peptic ulcers. JAMA 1990;264:4949.
Rutgeerts P, Vantrappen G, Van Hootegem PH, et al. Neodymium-YAG laser
photocoagulation versus multipolar electrocoagulation for the treatment of severely bleeding
ulcers: A randomized comparison. Gastrointest Endosc 1987;33:199202.
Goff JS. Bipolar electrocoagulation versus Nd:YAG laser photocoagulation for upper
gastrointestinal bleeding lesions. Dig Dis Sci 1986;31:9068.
Carter R, Anderson JR. Adrenaline injection or laser photocoagulation for bleeding peptic
ulcersA randomised study. Br J Surg 1994;81:86971.
Rutgeerts P, Geboes K, Vantrappen G. Experimental studies of injection therapy for severe
nonvariceal bleeding in dogs. Gastroenterology 1989;97:61021.
Randall GM, Jensen DM, Hirabayashi K, et al. Controlled study of different sclerosing agents
for coagulation of canine gut arteries. Gastroenterology 1989;96:127481.
Kalabakas A, Xourgias B, Van Someren N, et al. A randomised controlled comparison of
injection of adrenaline, polidocanol, alcohol, hypertonic saline and bipolar electrocoagulation
in treatment of standard experimental bleeding ulcers and bleeding mesenteric vessels
[Abstract]. Gut 1990;31:A1181.
Fleischer D. Endoscopic Nd:YAG laser therapy for active esophageal variceal bleeding.
Hashimoto D, Miyohara T, Yoshimura K. A lateral radiation probe in Nd:YAG laser therapy.
Daikuzono N, Joffe SN. Artificial sapphire probe for contact photocoagulation and tissue
vaporization with the Nd:YAG laser. Med Instrum 1985;3:173.
Sigel B, Hatke FL. Physical factors in electrocoaptation of blood vessels. Arch Surg
1967;95:548.
Sander R, Poesl H, Zuern W, et al. The waterjet-guided Nd:YAG laser in the treatment of
gastroduodenal ulcer with a visible vessel. A randomized controlled and prospective study.
Endoscopy 1989;21:21720.
Swain CP, Bown SG, Storey DW, et al. Controlled trial of argon laser photocoagulation in
bleeding peptic ulcer. Lancet 1981;2:131316.
Vallon AG, Cotton PB, Laurence BH, et al. Randomised trial of endoscopic argon laser
photocoagulation in bleeding peptic ulcers. GUT 1981;22:22833.
Storey DW, Bown SG, Swain CP, et al. Endoscopic prediction of recurrent bleeding in peptic
Wara P. Endoscopic prediction of major rebleedingA prospective study of stigmata of
hemorrhage in bleeding ulcer. Gastroenterology 1985;88:120914.
Matthewson K, Pugh S, Northfield TC. Which peptic ulcer patients bleed?. Gut 1988;29:704.
Fleischer D, Bown SG. Endoscopic Nd:YAG laser therapy for acute upper gastrointestinal
bleeding. In Jensen DM, Brunetaud JM (eds). Medical Laser Endoscopy. Dordrecht, The
Netherlands: Kluwer Academic Publishers, 1990;149.
Chung SCS, Leung JWC, Steele RJC, et al. Endoscopic injection of adrenaline for actively
bleeding ulcers: A randomised trial. Br Med J 1988;296:16313.
Johnston JH, Sones JQ, Long BW, et al. Comparison of the heater probe and YAG laser in
endoscopic treatment of major bleeding from peptic ulcers. Gastrointest Endosc
1985;31:17581.
Jensen DM. Heater probe for endoscopic hemostasis of bleeding peptic ulcers. Gastroenterol

65. Soehendra N, Grimm H, Stenzel M. Injection of nonvariceal bleeding lesions of the upper
gastrointestinal tract. Endoscopy 1985;17:12932.
66.
Swain CP. Nd:YAG laser for treatment of bleeding peptic ulcers. Techniques and results of
randomized prospective trials. Gastrointest Endosc Clin North Am 1991;1:34166.
67.
Nath G, Gorisch W, Kreitmair A, et al. Transmission of a powerful argon beam through a
fiberoptic flexible gastroscope for operative gastroscopy. Endoscopy 1973;5:2135.
68.
Frhmorgen P, Reidenbach HD, Bodem F, et al. Reports on new instruments and new
methods. Experimental examinations of laser endoscopy. Endoscopy 1974;6:11622.
69. Dwyer RM, Haverback BJ, Bass M, et al. Laser induced hemostasis in the canine stomach.
Use of a flexible fiberoptic delivery system. JAMA 1975;231:4869.
70. Spira I, Chryssanthou CP, Stenger RJ. Fiberoptic-coupled argon laser in the control of
experimentally produced gastric bleeding. Gastrointest Endosc 1975;22:7881.
71.
Frhmorgen P, Reidenbach HD, Bodem F, et al. Report on preliminary experimental and
clinical experiences with a new type of endoscopic argon laser photocoagulator [Abstract].
Endoscopy 1977;9:191A.
72. Silverstein FE, Protell RL, Gulsik C, et al. Endoscopic laser treatment III. Development and
testing of a gas assisted argon laser waveguide in control of bleeding experimental ulcers.
73. Bown SG, Salmon PR, Kelly DF, et al. Argon laser photocoagulation in the dog stomach. Gut
1979;20:6807.
74.
Frhmorgen P, Bodem F, Kaduk B. The first endoscopic laser coagulation in the human GI
tract. Endoscopy 1975;7:1567.
75. Frhmorgen P, Bodem F, Reidenbach HD. Endoscopic photocoagulation by laser irradiation
in the gastrointestinal tract of man. Acta Hepatogastroenterol (Stuttg) 1978;25:15.
76. Laurence BH, Vallon AG, Cotton PB, et al. Endoscopic laser photocoagulation for bleeding
peptic ulcers. Lancet 1980;1:1247.
77.
Bown SG, Storey DW, Swain CP, et al. Argon laser photocoagulation for upper
gastrointestinal haemorrhage: Is technique the key to success?. Digestion 1981;11:91820.
78.
Bowers JH, Dixon JA. Argon laser photocoagulation of vascular malformations in the GI tract:
Short term results [Abstract]. Gastrointest Endosc 1982;28:126A.
79.
Waitman AM, Grant DZ, Chateau F. Argon laser photocoagulation treatment of patients with
acute and chronic bleeding secondary to telangiectasia [Abstract]. Gastrointest Endosc
1982;28:153A.
80. Bown SG, Swain CP, Storey DW, et al. Endoscopic laser treatment vascular anomalies of the
upper gastrointestinal tract. Gut 1985;26:133848.
81.
Brunetaud JM, Maunoury V, Cochelard D, et al. Laser treatment for vascular malformations of
the digestive tract. In Jensen DM, Brunetaud JM (eds). Medical Laser Endoscopy. Dordrecht,
The Netherlands: Kluwer Academic Publishers, 1990;17.
82.
Machicado GA, Jensen DM. Upper gastrointestinal angiomata: Diagnosis and treatment.
83. Rutgeerts P, Van Gompel F, Geboes K, et al. Long term results of treatment of vascular
malformations of the gastrointestinal tract by neodymium-YAG laser photocoagulation. Gut
1985;26:58693.
84.
Fleischer D. Nd:YAG laser photocoagulation for upper gastrointestinal angiodysplasia
[Abstract]. Gastrointest Endosc 1981;27:122A.
85.
Johnston JH. Complication following endoscopic laser therapy. Gastrointest Endosc
1982;28:135.
86. Cello JP, Grendell JH. Endoscopic laser treatment for gastrointestinal vascular ectasias. Ann
Intern Med 1986;104:3524.
87. Naveau S, Aubert A, Poynard T, Chaput JC. Long-term results of treatment of vascular
malformations of the gastrointestinal tract by neodymium-YAG laser photocoagulation. Dig
Dis Sci 1990;35:8216.
88. Sargeant IR, Loizou LA, Rampton D, et al. Laser ablation of upper gastrointestinal vascular
ectasias: Long term results. Gut 1993;34:4705.
89. Gostout CJ, Ahlquist DA, Radford CM, et al. Endoscopic laser therapy for watermelon
stomach. Gastroenterology 1989;96:1462- 5.
90. Frager JD, Brandt LJ, Frank MJ, et al. Treatment of a patient with watermelon stomach using
endoscopic laser photocoagulation. Gastrointest Endosc 1988;34:1347.
91. Tsai HH, Smith J, Danesh BJ. Successful control of bleeding from gastric antral vascular
ectasia (watermelon stomach) by laser photocoagulation. Gut 1991;32:934.
92. Bjorkman DJ, Buchi KN. Endoscopic laser therapy of the watermelon stomach. Lasers Surg
Med 1992;12:47881.
93.
Etienne J, Rambert P, Dorme N. Successful Nd:YAG laser photocoagulation in Rendu Osler's
and Willebrand's diseases. In Atsumi K, Nimsakul N (eds). Proceedings of Laser Tokyo '81.
Vol 5. Tokyo: Inter Group Corp, 1981;2627.
94. Alexander TJ, Dwyer RM. Endoscopic Nd:YAG laser treatment of severe radiation injury of
the gastrointestinal tract: Long term follow-up. Gastrointest Endosc 1988;34:40711.
95. Ahlquist DA, Gostout CJ, Viggiano TR, Pemberton JH. Laser therapy for severe
radiation-induced rectal bleeding. Mayo Clin Proc 1986;61:92731.
96. Berken CA. Nd:YAG laser therapy for gastrointestinal bleeding due to radiation colitis. Am J
97. Boley SJ, Sammartano R, Brandt LJ, et al. Vascular ectasias of the colon. Surg Gynecol
Obstet 1979;149:3539.
98. Viggiano TR, Zighelboim J, Ahlquist DA, et al. Endoscopic Nd:YAG laser coagulation of
bleeding from radiation proctopathy. Gastrointest Endosc 1993;39:5137.
99.
Jensen DM, Machicado GA. Bleeding colonic angiomata: Endoscopic coagulation and
follow-up [Abstract]. Gastroenterology 1985;88:1433A.
100. Dixon JA, Berenson MM, McCloskey DW. Neodymium-YAG laser treatment of experimental
canine gastric bleeding. Acute and chronic studies of photocoagulation, penetration and
perforation. Gastroenterology 1979;77:64751.
101. Swain CP, Mills TN, Shemesh E, et al. Which electrode? A comparison of four endoscopic
methods of electrocoagulation in experimental bleeding ulcers. Gut 1984;25:14249.
102.
Jensen DM, Machicado GA, Kovacs TOG, et al. Controlled randomized study of heater probe
and BICAP for hemostasis of severe ulcer bleeding [Abstract]. Gastroenterology
1988;94:A208.
103. Papp JP. Endoscopic electrocoagulation in the management of upper gastrointestinal tract
bleeding. Surg Clin North Am 1982;62:797806.
104.
Jensen DM, Machicado GA. Argon laser for severe ulcer hemostasis: Health and economic
considerations. In Jensen DM, Brunetaud JM (eds). Medical Laser Endoscopy. Dordrecht,
The Netherlands: Kluwer Academic Publishers, 1990;119.
105. Laine L. Multipolar electrocoagulation in the treatment of active upper gastrointestinal tract
hemorrhage. A prospective controlled trial. N Engl J Med 1987;316:16137.
106. Laine L. Multipolar electrocoagulation in the treatment of peptic ulcers with nonbleeding visible
107.
peptic ulcers. A prospective, randomized trial. Gastroenterology 1990;13036.
108. Swain CP. Nd:YAG laser: The pro approach. Gastrointest Endosc 1990;36:S301.
Chapter 30 Injection Therapy for Ulcer Bleeding

(1735)
(1736)
S. C. SYDNEY CHUNG, M.D.
JOSEPH W. C. LEUNG, M.D.
Although injection therapy has been used extensively for variceal hemorrhage and bleeding
hemorrhoids, its use at endoscopy in the treatment of nonvariceal upper GI hemorrhage has become
popular only in the last decade. The technique has gained favor very rapidly, however. Injection therapy
is simple, and the results are at least comparable to those obtained with other more complicated and
expensive endoscopic techniques of ulcer hemostasis. Provided that the bleeding point can be
identified, control of active bleeding is achieved in almost every case. Injection therapy is now the most
commonly used endoscopic hemostatic technique for ulcer bleeding.
Although used for hemostasis with a variety of lesions including the Mallory-Weiss tear,1(1737) benign
esophageal ulcer,2(1738) and the Dieulafoy lesion,3(1739) endoscopic injection is considered in this
chapter mainly in relation to hemostasis of peptic ulcers.
Technique
The technique of injection for nonvariceal bleeding is relatively simple and should be within the
competence of endoscopists who carry out emergency endoscopy for bleeding. Because no special
equipment is necessary, injection can proceed when an actively bleeding lesion (or a visible vessel) is
identified during emergency endoscopy. In a patient with active bleeding, the stomach may contain
large amounts of blood and clot. With the patient in the left lateral position, blood and clot gravitate
toward the fundus and the greater curvature. The lesser curvature, antrum, and duodenum, where
most bleeding lesions are found, lie above the pool of blood. By following the lesser curvature into the
antrum and duodenum and staying above the pool of blood, the bleeding point can be identified even
when a large volume of blood is present in the stomach. Gastric lavage is seldom necessary.
The importance of adequate resuscitation before endoscopy and monitoring during the procedure
cannot be overemphasized. We consider continuous blood pressure monitoring, using an automatic
blood pressure monitor, and transcutaneous pulse oximetry mandatory during therapeutic endoscopy
for bleeding. The patient may lose a significant amount of blood during the procedure; facilities for
rapid blood transfusion must be readily available.
Precise identification of the bleeding point is essential before hemostatic maneuvers can be carried
out. It is often necessary to clear the ulcer of blood and clots by irrigation. Unlike multipolar and heater
probes, injection needles do not have a built-in mechanism for washing. Small clots may be washed
away using a syringe of saline connected to the accessory channel. For massive bleeding, a
mechanical water pump is indispensable. We use a modified water tooth brush (Water-Pik) connected
to the accessory channel of the endoscope via a three-way tap.4(1740) The water pump is activated by
a foot switch. A forceful jet of water (up to 500 ml/min) can be directed at the bleeding lesion to wash
away any blood or clot, while both hands of the endoscopist remain free to manipulate the endoscope
and injection needle. If a double-channel endoscope is used, the Water-Pik can be passed down one
channel and the injection needle down the other, so that irrigation and injection can be carried out
simultaneously.
Needles designed for sclerotherapy of esophageal varices are suitable for injection of bleeding ulcers.
These devices consist of a needle that can be retracted into a flexible sheath so that the device can be
passed through the accessory channel of the endoscope without creating a puncture within the
bending section. We use 23-gauge needles with a 5-mm shaft for ulcer injection. Extension of the
needle from the distal end of the sheath is controlled by a nurse-assistant. Close cooperation between
endoscopist and assistant is essential. The endos-copist and the assistant must use the same
vocabulary and understand what is meant by needle-in and needle-out as well as the volume and
nature of each aliquot of injection. Use of video endoscopes has made coordination much easier.
Because of the high pressure needed to inject fluid into tissues, disconnection of the syringe from the
hub of the needle may occur and may lead to splashing of fluid. Eye injury to the endoscopist,
nurse-assistant, and patient may result if sclerosant solutions are used. A syringe with a screw-on
Luer-Lok should be used to minimize the danger of splashing. The endoscopist and assistant should
wear protective eye goggles or shields if sclerosant solutions are used.

After the bleeding point (or the nonbleeding visible vessel) is identified, multiple aliquots of the
hemostatic agent (epinephrine, sclerosant, or a combination of these agents) are injected through
separate punctures into the ulcer base close to the bleeding point. Some authors advocate placing the
injection 1 to 2 mm away from the bleeding point; others suggest surrounding the bleeding point with
several aliquots before depositing the final aliquot into the region of the bleeding vessel. During active
bleeding, precise targeting of the bleeding vessel may be difficult or even impossible; suffice it to say
that injection close to the bleeding point arrests the hemorrhage. In patients with chronic ulcers, the
base of the lesion may be very fibrotic. The nurse-assistant may have to exert substantial force to
inject. If no resistance is felt, the needle is likely off target and the injected material has spilled.
Injection is continued until the bleeding is controlled (Figure 301). The volume of each injection and
the total volume used vary according to the agent (Table 301). If sclerosants are used, care must be
exercised not to exceed the recommended volume, or extensive tissue damage may result in
perforation or gastric wall necrosis.
TABLE 301
Agents for Injection Therapy of Bleeding Ulcers
AUTHOR
AGENT
Leung and Chung8

Steele et al.11,12
Epinephrine
Asaki et al.20
Alcohol
CONCENTRATION
ALIQUOT
TRUE VOLU
1/10,000
0.5 ml
10 ml*
1/10,000
1.0 ml
10 ml
Absolute
0.10.2 ml
0.60.8 ml
Sugawa et al.21
Lin et al.24,25
98.0%
0.10.2 ml
0.60.8 ml
99.8%
0.30.5 ml
12 ml
Laine23
Absolute
0.10.2 ml
Up to 2 ml
1.0%
1015 ml
1.5%
35 ml
Kortan et al.31
1/10,000
1.0%
1/10,000
1.0%
1/10,000
510 ml
5 ml
310 ml
612 ml
510 ml
Balanz et al.32
1.0%
1/10,000
12 ml
35 ml
~10 ml
1.0%
1/10,000
3.0%
3.6/7.2% saline with
0.005% epinephrine
1/10,000
30 IU/ml
0.5 ml
0.5 ml
3 ml
35 ml
1020 ml
1 ml
Up to 12 ml
12 ml
12 ml
~10 ml
510 ml
Wrdehoff and Gros19

Soehendra and Werner18
Soehendra et al.33
Chung et al.42
Polidocanol
Epinephrine +
polidocanol
Epinephrine +
sodium tetradecyl sulfate
34
Hypertonic
Hirao et al.
saline-epinephrine
Epinephrine
Balanz et al.13
+ thrombin
* Up to 30 ml has been used with no complications.
(1741)Figure 301. Bleeding duodenal ulcer with an arterial spurter before (A), during (B and
C), and after (D) injection of epinephrine. (A-D, Steele RJC, Chung SCS, Leung JWC.
Practical management of
Butterworth-Heinemann, 1993.)
acute
gastrointestinal
bleeding.
Oxford,
Boston:
After all bleeding has ceased, the ulcer is observed for a few minutes before the endoscope is
removed. The patient is returned to the ward for intensive monitoring. Endoscopy is routinely
performed 24 hr later and injection is repeated if active bleeding is again seen.
An algorithm for managing patients with ulcer bleeding is shown in Figure 302. Endoscopic
hemostasis is but one facet of overall management. Good results and low mortality are achieved by
aggressive resuscitation, careful monitoring, and timely surgery for recurrent bleeding. Despite
enthusiasm for application of endoscopic methods of hemostasis, the endoscopist should remember
that surgery remains the most definitive method of controlling ulcer bleeding and that emergency
surgery is safe provided that patients are referred early. Clear and unequivocal guidelines for operation
are helpful, and all members of the team should be aware of these endpoints. We recommend
immediate surgery if (1) arterial spurting cannot be controlled by endoscopy, (2) clinically significant
recurrent bleeding occurs in the form of hematemesis and/or melena with shock, or (3) total
transfusion exceeds 8 units of blood.
(1742)Figure 302. Algorithm for the management of patients with bleeding ulcers.
GIgastrointestinal.
Agents
Epinephrine, various sclerosants, or combinations have been used for injection therapy (see Table
301). No consensus has been reached as to the best agent. Epinephrine is effective in controlling
active hemorrhage, and large volumes may be used without fear of causing tissue damage.
Sclerosants cause more inflammation and tissue damage but may be more effective in causing
vascular thrombosis.
The mechanism of action of the various agents has been studied in the laboratory.5,6(1743) Because
of the lack of an adequate animal model of a bleeding ulcer, the assessments have been performed on
artifactual ulcers created with the Quinton device (Quinton Instruments, Seattle, WA)5(1744) or
transected mesenteric vessels.6(1745) Whittle et al.7(1746) compared the relative efficacy of various
hemostatic agents in a canine model. The hemostatic effects of the following agents on blood flow rate
from transected gastric serosal vessels (diameter 1.6 to 2.2 mm) were studied: normal saline (5 ml),
3% hypertonic saline (5 ml), 1:10,000 epinephrine in normal saline (5 ml), 1:10,000 epinephrine in
hypertonic saline (5 ml), 1:20,000 epinephrine in hypertonic saline (5 ml), 2 ml of "old thrombin
cocktail" (thrombin, cephapirin, 1% sodium tetradecyl sulfate), and 2 ml of fresh thrombin cocktail.
Except for the old thrombin cocktail, all agents reduced blood flow in this model. The best results were
achieved with the epinephrine, reduction in blood flow being greatest for epinephrine in hypertonic
saline. Tissue necrosis was least with epinephrine and greatest with the cocktail solutions.
The animal models of ulcer bleeding do not mimic the clinical situation closely because no fibrosis is
present in the ulcer base to retain the injected solution. Conclusions from these studies must be
interpreted with caution, as extrapolation to the clinical situation in humans may not be valid.
Epinephrine
Epinephrine, in a concentration of 1:10,000, has been used extensively in the treatment of bleeding
ulcers, postpapillotomy bleeding, and Mallory-Weiss tears.813(1747) Given in aliquots of 0.5 to 1 ml
around the bleeding point, epinephrine injection is very effective in stopping active bleeding; success
rates at the time of endoscopy range from 96 to 100%.
The exact mechanism by which injected epinephrine stops bleeding is conjectural. Possible
mechanisms that have been suggested are vasoconstriction,14(1748) compression of the vessel due
to a volume effect of the injected fluid, and platelet aggregation.15(1749)
Scant data are available on the maximum safe dose of epinephrine that can be injected submucosally
in the stomach and the duodenum. The recommended maximum dose is 10 ml; volumes of up to 30
ml have been used with no untoward effects, although tachycardia may occur. Stevens and
Lebwohl16(1750) reported the case of a 48-year-old man who developed a hypertensive crisis and
ventricular tachycardia after injection of 10 ml of 1:10,000 epinephrine for control of bleeding from a
Mallory-Weiss tear. The patient presented in shock 2 days after a drinking binge, and a 1.5-cm tear,
not actively bleeding, was found at the gastroesophageal junction at initial endoscopy. The patient had
recurrent bleeding, and the injection procedure was performed at a third endoscopic procedure.
Epinephrine has a large first-pass effect, and most of the submucosally injected drug is metabolized by
the liver. Nevertheless, significant increases in systemic blood levels are measurable during
injection.17(1751) Caution should be exercised, particularly for patients with impaired liver function and
when epinephrine injection is used for bleeding lesions in the esophagus.
Because of the low potential for tissue damage, it is safe to use relatively large volumes of epinephrine.
In a patient with active bleeding, precisely identifying the bleeding point is sometimes difficult because
the bleeding vessel may be covered by a pool of blood or adherent clot. Provided that the bleeding
point can be located approximately, epinephrine may be injected close to it. The first few aliquots
decrease the bleeding, allowing a clearer view, and subsequent injections may be placed more
accurately.
It is also safe to carry out repeat injections on the same patient without the risk of causing gastric
necrosis or perforation. We routinely repeat the endoscopic examination 24 hr after the initial injection
procedure. If active bleeding is again seen, repeat injection is carried out, this being necessary in about
10% of cases.
Sclerosants
The most popular sclerosants for injection of bleeding ulcers are polidocanol18,19(1752) and
ethanol.2025(1753) The efficacy and tissue effects of these sclerosants have been studied in animal
experiments. Polidocanol causes hemostasis by bowel wall spasm and acute edema, followed by
inflammation and sclerosis. Alcohol stops bleeding by rapid dehydration and tissue fixation, which
causes arterial coagulation. Alcohol is the most effective agent for vessel obliteration and also causes
the most tissue damage.6(1754) All sclerosants cause tissue necrosis at the site of injection and
ulceration of the overlying mucosa.
The tissue damage produced by sclerosants is determined by the injected volume. All authors
advocating sclerosant injection have cautioned against injecting an excessive volume. Asaki et
al.20(1755) recommend that alcohol be injected in 0.1 to 0.2 ml aliquots using 1-ml tuberculin syringes.
The alcohol is injected at two to four sites, 1 to 2 mm away from the vessel, with the total dose limited
to 0.6 to 0.8 ml. Lin et al.24,25(1756) believe that this volume may be too small for a spurting bleeder;
these investigators used 0.3- to 0.5-ml aliquots up to a total of 1 to 2 ml for each bleeding vessel. Up to
3.5 ml have been used on two occasions without perforation by Lin et al.24(1757) Laine23(1758) also
used up to 2 ml of pure ethanol, injected in 0.1- to 0.2-ml aliquots. No complications attributable to
injections were reported in any of these series. Polidocanol is less irritating, and larger volumes may be
used. Wrdehoff and Gros19(1759) used 10 to 15 ml of 1% polidocanol in 36 patients. Perforation or
delayed hemorrhage did not occur, but ulcer extension and delayed healing were noted when large
volumes were used.
Perforation and gastric wall necrosis have been reported after injection sclerotherapy of bleeding
ulcers. These are infrequent complications that occur in less than 1% of patients. In the original series
of Asaki et al.,20(1760) one perforation occurred in 81 cases. Levy et al.26(1761) reported fatal
full-thickness necrosis of the gastric antrum, lesser curvature, and duodenum after the injection of 5 ml
of 1:10,000 epinephrine followed by 6 ml of polidocanol. Full-thickness necrosis of the stomach was
also reported after injecting 12 ml of ethanolamine into an ulcer on the lesser curvature of the
stomach.27(1762) Loperfido et al.28(1763) reported a patient who developed extensive necrosis and
ulceration of the stomach after the injection of 7 to 8 ml of epinephrine and 4 ml of polidocanol. The
ulcer gradually healed over 3 months, but the cicatrization caused by the healing resulted in gastric
stenosis that necessitated a Billroth type II gastrectomy. Vallieres et al.29(1764) have described a case
of segmental infarction of the duodenum and head of the pancreas, necessitating
pancreatoduodenectomy, that was attributed to endoscopic injection of cyanoacrylate to control arterial
bleeding from a duodenal ulcer.
Combinations of Agents
One of the most popular techniques for injecting actively bleeding ulcers is the so-called Soehendra
technique.3032(1765) Epinephrine is injected first until bleeding is controlled. A small volume of
sclerosant (1% polidocanol) is then injected at the vessel site to cause thrombosis. With the bleeding
already controlled by epinephrine, a clearer endoscopic view of the bleeding vessel can be obtained for
precise targeting of the sclerosant injection. The volume of sclerosant can therefore be limited, thereby
decreasing the risk of severe tissue damage.
Results of Injection Treatment

Nonrandomized Series
Injection treatment for ulcer bleeding was first reported by Soehendra and Werner in 1976.18(1766)
Two patients with arterial bleeding were successfully treated by injection of 3 to 5 ml of 1.5%
polidocanol. Wrdehoff and Gros19(1767) used 10 to 15 ml of 1% polidocanol in 36 patients with
arterial spurters, oozing, and other stigmata. Hemostasis was achieved in 33. Seven patients had
recurrent bleeding that was controlled in 3 cases by repeat injection. No perforations were reported,
although injection of larger volumes caused large ulcers and retarded ulcer healing.
Using the Soehendra technique (epinephrine followed by polidocanol), Soehendra et al.33(1768)
treated 56 patients with active bleeding and nonbleeding stigmata. Definitive hemostasis was achieved
in all 22 patients with arterial bleeding and 27 of 28 patients with oozing, although 12 patients required
a second session of injection for rebleeding. Using the same technique, Kortan et al.31(1769) reported
equally good results in 56 patients with active bleeding. Bleeding was controlled in 90% of arterial
spurters and 100% of patients with oozing. Recurrent bleeding developed in 22% of patients, but only 1
required emergency surgery.
The largest series of patients treated by injection therapy was reported from Japan. Using 98% alcohol,
Asaki et al.20(1770) treated 332 patients with active bleeding and adherent clots in a multicenter study.
Bleeding was controlled in all but 2 patients, 20 rebled, and only 10 required emergency surgery. Using
the same technique, Sugawa et al.21(1771) achieved permanent hemostasis in 34 of 40 patients with
active bleeding.
Using a mixture of hypertonic saline and epinephrine (3.6 or 7.2% sodium chloride and 0.005%
epinephrine), Hirao et al.34(1772) reported overall success in 98.1% of 158 patients. Only 1 patient
required emergency surgery. Chen et al.35(1773) treated 67 patients with a variety of bleeding lesions,
including peptic and marginal ulcers, esophageal and gastric cancers, and polyps, with injections of
either a solution of hypertonic saline plus epinephrine or pure ethanol. In this uncontrolled study,
hemostasis was achieved in 87.9% of patients injected with the hypertonic saline/epinephrine solution.
Hemostasis was achieved in 91.2% of those who underwent ethanol injections.

The successful use of epinephrine alone in arresting active ulcer bleeding was reported first by Leung
and Chung in 37 patients.8(1774) The efficacy of this technique was confirmed by Steele et
al.11,12(1775) in two other series. Epinephrine causes no tissue damage, and large volumes can be
used without risk of stomach wall necrosis. The success rate in controlling bleeding at the time of
endoscopy is close to 100%. Leung and Chung8(1776) recommend that endoscopy be repeated
routinely 24 hr after injection; if further active bleeding is seen, injection is repeated. This is required in
about 10% of patients. Eight to 20% of patients develop clinically significant recurrent bleeding and
require emergency surgery.
The results of some of the published series of injection therapy, classified according to the severity of
bleeding into arterial spurting, oozing, and nonbleeding visible vessels, are tabulated in Table 302.
Overall, injection therapy is remarkably effective at controlling active bleeding. Success rates range
from 70 to 100% for spurting arterial bleeding and approach 100% for nonarterial bleeding. Recurrent
bleeding occurs in 15 to 20% of patients and may be controlled by repeat injection in some of them.
Emergency surgery is required in approximately 10% of patients overall.
TABLE 30-2 Injection of Bleeding Ulcers: Nonrandomized Series (Part i of ii )

ARTERIAL BLEEDING
AUTHOR
Initial
Control
Rebleeding
Emergency
Surgery
37
12
100%
17%
5%
15
15
20%
13%
53
26
19%
Polidocanol
36
11
Alcohol
48
10
70%
10%
40%
Epinephrine
58
22
27%
0%
Polidocanol
56
32
90%
17%
12.5%
AGENT
Epinephrine
Leung and Chung8

11
Steele et al.
12
Steele et al.
Wrdehoff and Gros
19
21
Sugawa et al.
33
Soehendra et al.
31
Kortan et al.
(Continued in Part ii)
TABLE 30-2 Injection of Bleeding Ulcers: Nonrandomized Series (Part ii of ii )

OOZING
NONBLEEDING
Initial
Control
Rebleeding
Emergency
Surgery
Rebleeding
Emergency
Surgery
25
100%
12%
8%
Steele et al.11
12
27
15%
AUTHOR
Leung and
Chung8
Steele et al.
Rebleeding
TABLE 30-2 Injection of Bleeding Ulcers: Nonrandomized Series (Part ii of ii )

OOZING
NONBLEEDING
Initial
Control
Rebleeding
Emergency
Surgery
Rebleeding
Emergency
Surgery
Wrdehoff and
Gros19
15
10
Sugawa et al.21
Soehendra et
al.33
30
100%
7.4%
0%
0%
28
7%
7%
0%
0%
Kortan et al.31
24
100%
29%
0%
AUTHOR
Rebleeding
These results are impressive. Reports of successful control of bleeding must, however, be interpreted
in the context that up to 80% of patients with ulcer bleeding stop bleeding spontaneously. Important
prognostic factors in patients with nonvariceal upper GI bleeding include the type of lesion (ulcer,
erosions, Mallory-Weiss tear), bleeding status (spurting, oozing, nonbleeding vessel or clot), and
coexisting medical illnesses. Outcome depends as much on patient selection as on treatment modality.
Different criteria for emergency surgery and for the diagnosis of recurrent bleeding make comparison
of reports from different investigators difficult. Validation of the usefulness of an endoscopic technique
in ulcer bleeding depends on well-conducted randomized clinical trials.
Randomized Trials
Injection Versus Controls
Results of seven randomized trials comparing injection treatment of bleeding ulcers with no
endoscopic treatment have been published (Table 303). In a trial involving 113 patients with active
bleeding and nonbleeding visible vessels, Pans et al.36(1777) demonstrated that injection of
epinephrine followed by 1% polidocanol reduced the rate of recurrent bleeding from 43 to 5%.
Transfusion requirements and length of hospital stay were also significantly reduced, although no
significant reduction in mortality was seen.
TABLE 30-3 Randomized Trials Comparing Injection With Conventional Therapy (Part i of ii )
REBLEEDING
Injection
Control
Epinephrine alone
72
Epinephrine + 1%
polidocanol
7/36(19%)
15/36(44%)
145
Absolute alcohol
2/65(3%)
AUTHOR
INCLUSION
Chung et al. 9
Active bleeding
68
Balanz et al.32
Active bleeding,
visible vessel,
or red clot
Active bleeding,
nonbleeding
stigmata
AGENT(S)
p < .05
22
Pascu et al.
TABLE 30-3 Randomized Trials Comparing Injection With Conventional Therapy (Part i of ii )
REBLEEDING
AUTHOR
INCLUSION
Pans et al.36
Active bleeding
vessel
113
Epinephrine + 1%
polidocanol
Oxner et al.
Visible vessel
(a few active
bleeding)
93
Epinephrine + 5%
ethanolamine
Rajgopal and
Palmer38
Active bleeding,
nonbleeding
stigmata
109
Epinephrine + 5%
ethanolamine
37
39
Rutgeerts et al.
Visible vessel
25
25
AGENT(S)
Injection
Control
3(5%)
25(43.1%)
p < .001
8/48(16.7%)
21/45(46.7%)
p = .011
7/56(12.5%)
25/53(47%)
p < .001
10/25(40%)
11/25(44%)
Epinephrine + 1%
polidocanol
Absolute alcohol
5/25(20%)
11/25(44%)*
*p = .07
TABLE 30-3 Randomized Trials Comparing Injection With Conventional Therapy (Part ii of ii )
TRANSFUSION
(UNITS)
EMERGENCY SURGERY
AUTHOR
Injection
Control
Chung et al.9
5/34(15%)
14/34(41%)
Balanz et
al.32
7/36(19%)
p < .02
12/36(36%)
1/65(1.5%)
NS
17/78(22%)
22
Pascu et al.
HOSPITAL DAYS
Injection
Control
Injection
3(0-12)
5(0-25)
6(2-27)
1.63
p < .05
2.86
Control
8(3-103)
MORTALITY
Injection
3/34(9%)
p < .005
Not studied
p < .05
Not stated
Not stated
2/65(3%)
16.3
2(4%)
p < .05
36
Pans et al.
37
Oxner et al.
Rajgopal and
Palmer38
Rutgeerts et
al.39
3(5%)
20(34%)
0.42
2.8
p < .001
7.5
11.7
p < .001
4/48(8.3%)
8/45(17.8%)
6/56(10.7%)
13/53(24.5%)
8/25(32%)
3/25(12%)
6.0
6.0
Not stated
4/25(16%)
2/25(8%)
3/25(12%)
3.9
6.0
Not stated
1/25(4%)
9.5
8
p < .001
10
8
4/48(8.3%)
3/56(5.4%)
TABLE 30-3 Randomized Trials Comparing Injection With Conventional Therapy (Part ii of ii )
EMERGENCY SURGERY
AUTHOR
Injection
Control
TRANSFUSION
(UNITS)
Injection
Control
HOSPITAL DAYS
Injection
Control
MORTALITY
Injection
p = .02
NSnot statistically significant.
* Statistical significance refers to the Rutgeerts' absolute alcohol study; rebleeding treated by second injection session; hem
increased to 68% for patients treated with epinephrine plus polidocanol and 88% for those treated with absolute alcohol.
Statistical significance refers to the Rutgeerts' absolute alcohol study.
In a prospective randomized trial using the same injection technique as Pans et al.,36(1778) Balanz
et al.32(1779) demonstrated significant reductions in the rate of recurrent bleeding (44 to 19%) and
transfusion requirement (2.86 to 1.63 units) in the group of patients undergoing injection. Although a
significant difference was found in rates of recurrent bleeding, the difference in numbers of patients
undergoing emergency surgery did not reach statistical significance; three patients with recurrent
bleeding underwent injection rather than surgery because of contraindications to operation. No
significant difference in mortality was seen.
Pascu et al.22(1780) were able to demonstrate a reduction in mortality in the injection group in a
randomized trial using absolute alcohol as the sclerosant. The increased mortality in the control group
resulted from a high emergency surgery rate and a high operative mortality for emergency operations.
The criteria for selecting patients for emergency surgery were not stated.
Chung et al.9(1781) included only patients with active bleeding (arterial spurting or active oozing) in a
randomized trial using epinephrine alone. Because of the difficulties in clearly defining recurrent
bleeding, these investigators did not use recurrent bleeding as an endpoint but adhered to strict criteria
of surgery, which included unequivocal recurrence of bleeding (hematemesis, or shock plus melena)
for both groups. Significant reductions in emergency surgery (41 to 15%), transfusion requirement, and
length of hospital stay were shown in the injection group. No difference in mortality was found.
Ninety-three of 98 patients with an ulcer with a visible vessel found at endoscopy performed for upper
GI bleeding were randomized to undergo injection (n = 48) or conventional therapy (n = 45) in the study
of Oxner et al.37(1782) Endoscopic treatment consisted of injection of 1 to 2 ml of 1:10,000
epinephrine at four to six sites around the ulcer, followed by injection of epinephrine and 5%
ethanolamine oleate (1 to 2 ml) directly into the vessel. The difference in the rate of recurrent bleeding
was statistically significant; this occurred in 8 of the 48 patients who underwent injection treatment
(16.7%) versus 21 of 45 patients (46.7%) in the control group (p = .011). Mortality, the need for
surgery, and mean blood-transfusion requirement were also lower in the treated group.
Rajgopal and Palmer38(1783) randomized 109 patients with severe bleeding peptic ulcers (active
bleeding or stigmata of recent hemorrhage) to injection treatment with a combination of 1:100,000
epinephrine and 5% ethanolamine oleate or conventional treatment. Only those patients with ulcers
accessible at endoscopy were treated. A significant reduction in the rate of recurrent bleeding occurred
in the group that underwent endoscopic treatment (12.5 vs. 47%, p < .001). Although the need for
emergency surgery and transfusion requirement was lower in the treated group, these differences did
not reach statistical significance.
Patients with severely bleeding peptic ulcers who were found at endoscopy to have visible vessels
were randomized to undergo injection with epinephrine followed by polidocanol (n = 25), injection of
absolute ethanol (n = 25), or sham injection (n = 25) in the comparative trial of Rutgeerts et al.39(1784)
Rates of recurrent bleeding were, respectively, 40, 20, and 44%. These differences were not
statistically significant, although the difference between the ethanol-treated and the sham groups
nearly reached significance. Patients with recurrent bleeding underwent a second endoscopic injection
treatment, which increased the rates of hemostasis to 68% for the epinephrine/polidocanol group and
88% for patients injected with absolute ethanol, a difference that was not statistically significant. Mean
transfusion requirements were 6.0 units for the sham group, 6.0 for the epinephrine/polidocanol-treated
group, and 3.9 for the group that received absolute ethanol injections, with the difference between the
ethanol and the sham groups being statistically significant at p = .02.
Injection Versus Other Endoscopic Methods
Injection treatment has been compared with other forms of endoscopic ulcer hemostasis in several
randomized trials (Table 304). In a trial comparing epinephrine, epinephrine plus polidocanol, and
epinephrine followed by neodymium:yttrium-aluminum-garnet (Nd:YAG) laser treatment, Rutgeerts et
al.30(1785) found that epinephrine plus polidocanol is at least as effective as epinephrine plus laser.
The rate of recurrent bleeding with epinephrine alone was higher, but repeat injection improved its
efficacy. One perforation occurred in the group of laser-treated patients.
Comparative Trials of Injection Treatment Versus Other Endoscopic Hemostatic M
TABLE 304
AUTHOR
Rutgeerts et al.30
140
Chung et al.10
132
Laine23
INCLUSIONS
AGENT
INJECTED
COMPARED WITH
CONCLUSIONS
Active bleeding
and
non-bleeding
visible vessel
Active bleeding
only
Epinephrine
alone/epinephrine
+ polidocanol
Nd:YAG laser
Epinphrine +
polidocanol at least as
good as epinephrine
plus laser
Injection and heater
probe are equally
effective. Injection is
easier to perform.
Multipolar
electro-coagulation
and ethanol injection
are of comparable
efficacy.
Heat probe is more
effective than ethanol
injection.
Epinephrine
Heater probe
60
Active bleeding
and visible
vessel
Ethanol
Multipolar
electro-coagulation
Lin et al.25
137
Ethanol
Heater probe
Pans et al.40
127
Epinephrine +
polidocanol
Microwave
Injection and
microwave equally
effective
Waring et al.41
60
Active bleeding
and
nonbleeding
visible vessel
Active bleeding,
nonbleeding
visible vessels,
and adherent
clots
Active bleeding +
stigmata
Ethanol
Multipolar
electro-coagulation
Injection and
electro-coagulation
equally effective
Laine23(1786) compared multipolar electrocoagulation with injection of absolute ethanol in a

randomized trial involving patients with active bleeding and nonbleeding visible vessels. The two
methods were found to be equally effective. Treatment-induced bleeding occurred in patients with
nonbleeding visible vessels in 35% of each group, but this was controlled with continued treatment in
all patients. One perforation occurred in the multipolar electrocoagulation group.
Lin et al.25(1787) compared absolute alcohol injection with heater probe treatment in 137 patients with
active bleeding or nonbleeding visible vessels. The heater probe was significantly better than alcohol
injection at controlling bleeding at the time of endoscopy (98 vs. 67%) and at achieving permanent
hemostasis (91 vs. 67%). Only 4 of 13 patients with spurting hemorrhage could be controlled by
alcohol injection. Significant reductions in length of hospitalization, number of patients requiring
emergency surgery, and mortality were also demonstrated in the heater probe group.
In contrast to the results of Lin et al.,25(1788) Chung et al.10(1789) found epinephrine injection to be
superior to heater probe for initial control of bleeding at the time of endoscopy. In a trial involving only
patients with actively bleeding ulcers, epinephrine injection stopped the bleeding at the time of
endoscopy in 96%, whereas the heater probe was successful in only 83%. However, no difference was
found in patient outcome as measured by transfusion requirements, emergency surgery, length of
hospital stay, or mortality. A perforation was noted in two patients in the heater probe group, both of
whom had undergone repeat electrocoagulation for recurrent bleeding.
The major difference in results obtained by Lin et al.25(1790) and Chung et al.10(1791) lies in the
control of arterial bleeding by injection. The differing results obtained by these two groups of
investigators may be explained at least in part by the volume of the solution used. Chung et
al.10(1792) used up to 20 ml of epinephrine, whereas Lin et al.25(1793) used 1 to 2 ml of ethanol only.
A larger volume of solution may be more effective at arresting spurting hemorrhage.
Pans et al.40(1794) compared injection sclerosis using the Soehendra technique (epinephrine
followed by polidocanol) and microwave coagulation in 127 patients with ulcers and active bleeding,
nonbleeding visible vessels, and adherent clots. The rate of recurrent bleeding, requirement for
emergency surgery, transfusions, and mortality were all numerically higher in the microwave-treated
group, although statistical significance was not reached with regard to any of these parameters. Pans
et al.40(1795) concluded that microwave and injection treatment are equally effective.
Sixty men with actively bleeding ulcers or ulcers with stigmata of recent hemorrhage were randomized
to receive either injection treatment with 98% ethanol (1 to 2 ml) or multipolar probe electrocoagulation
in the trial of Waring et al.41(1796) No differences in outcome were found with respect to initial
hemostasis, rates of recurrent bleeding, transfusion requirement, need for emergency surgery, or
mortality.
Comparison between Different Agents

Relatively few studies have compared different methods of injection (Table 305). Balanz et
al.13(1797) compared epinephrine injection versus epinephrine followed by 5 to 10 ml of thrombin (30
IU/ml) and found no differences in hemostasis, the need for emergency surgery, or number of blood
transfusions. Chung et al.42(1798) compared epinephrine injection alone with injection of epinephrine
followed by 3% sodium tetradecyl sulfate. Patient outcomes were identical in both groups, and no
additional benefit was conferred by the additional injection of sodium tetradecyl sulfate. Chiozzini et
al.43(1799) compared epinephrine injection, alcohol injection, and no endoscopic treatment. Patients in
the injection groups fared better than those in the control group, but no significant differences were
seen between epinephrine and alcohol injection.
TABLE 305
Comparative Trial Between Different Methods of Injection
al.43(1799) compared epinephrine injection, alcohol injection, and no endoscopic treatment. Patients in
the injection groups fared better than those in the control group, but no significant differences were
seen between epinephrine and alcohol injection.
Comparative Trial Between Different Methods of Injection
TABLE 305
AUTHOR
INCLUSIONS
INJECTED
AGENT
COMPARED WITH
Balanz et al.13
64
Active bleeding +
visible vessel
Epinephrine
Epinephrine + thrombin
Chung et al.42
200
Epinephrine
Chiozzini et
al.43
Lin et al.44
53
Active bleeding
only
Active bleeding +
visible vessel
Epinephrine
Epinephrine + sodium
tetradecyl sulfate
Alcohol
Active bleeding
Epinephrine
Epinephrine + alcohol
Lin et al.45
200
Ethanol
Normal saline, 3%
NaCl, 50% glucose
64
Active bleeding +
visible vessel
* Published as an abstract.
CONCLUSIONS
No difference in
rebleeding,
emergency
surgery, or
transfusion
No difference in
outcome
No difference
Epinephrine +
alcohol was better
for spurting
No difference
Lin et al.44(1800) studied the efficacy of alcohol injection after injection of epinephrine in 64 patients
with active peptic ulcer bleeding. Initial control of bleeding was achieved in 97% of patients treated with
epinephrine alone versus all of those treated by epinephrine followed by alcohol injection. However,
bleeding recurred in 36% of patients treated with epinephrine alone versus 16% of those who received
both agents (p > .05). With regard to spurting hemorrhage, epinephrine plus absolute alcohol achieved
hemostasis in 9 of 10 cases versus 5 of 11 for epinephrine alone (p < .05). No difference was found
between the two treatment groups in terms of the need for emergency surgery and blood transfusions,
although the length of time in hospital was significantly less for patients who underwent treatment with
two agents.
The hemostatic effects of endoscopic injection with normal saline, hypertonic (3%) saline, glucose
(50% in water), and absolute ethanol were compared in patients (n = 50 for each group) with peptic
ulcers and active bleeding or nonbleeding visible vessels by Lin et al.45(1801) No statistically
significant differences were seen between any of the patient groups with respect to initial or ultimate
hemostasis, rates of recurrent bleeding, need for emergency surgery, average length of stay in
hospital, or transfusion requirement, although a trend was noted toward a lower transfusion
requirement for patients who underwent injections with ethanol. One perforation occurred at 5 days
after treatment in a patient who underwent injection of 3 ml of 50% glucose. Lin et al.45(1802)
considered these several solutions to be virtually equivalent for injection treatment.
Summary
Ample evidence is now available that injection treatment is effective in controlling bleeding from peptic
ulcers and that endoscopic injection can favorably alter the clinical course of patients with active
bleeding and visible vessels. Injection treatment is at least as effective as laser photocoagulation or
coaptive coagulation using thermal probes such as the heater or multipolar probe. No agreement has
yet been reached on the best agent for injection treatment. Epinephrine causes the least tissue
damage and may be as effective as sclerosants.
A comparison of laser photocoagulation, coaptive thermal coagulation (heater probe/multipolar probe),
and injection therapy is shown in Table 306. All three methods are effective in stopping ulcer
bleeding. The great advantage of injection treatment is that it does not require specialized equipment
and is therefore universally available to all endoscopists. Complex machines, which require
maintenance and are prone to break down, are not required. A multiuse metal injector costs around US
$100 and a disposable injection catheter around US $30. This is less than 1% of the cost of a heater
probe or multipolar unit and 0.1% of the cost of a laser unit. Injection therapy is remarkably easy and
can be performed at the bedside.
TABLE 306
Mechanism
Use with 2.8-mm
channel scope
Application
Comparison of Different Hemostatic Methods

LASER
HEATER
PROBE/BICAP
Thermal
coagulation
Coaptive thermal
coagulation
INJECTION
Vasoconstriction,
compression,
thrombosis
+
En face
En face and
En face and
tangential
tangential
Ease of use
+
Built-in irrigation
+
Tissue sticking
+
Risk of perforation
+
+
*
Maintenance required
++
+
Portable
+
++
Cost
US $100,000
US $10,000
US $100
Effective hemostasis
+
+
+
* Perforation and stomach necrosis reported with sclerosants.
Ulcer bleeding is a common medical emergency. Most patients suffering from this condition are
admitted to a hospital near their community. Those patients who require intervention to control
bleeding are likely to be unstable and therefore cannot be transferred to specialist centers for
treatment. For endoscopic hemostasis to have an impact on the outcome of ulcer bleeding overall, a
simple, effective, and inexpensive method must be widely available in most hospitals. Available
evidence indicates that injection treatment comes closest to this ideal.
REFERENCES
1.
Bataller R, Llach J, Salmeron JM, et al. Endoscopic sclerotherapy in upper gastrointestinal

bleeding due to the Mallory-Weiss syndrome. Am J Gastroenterol 1994;89:214750.
2. Park KG, Steele RJ, Masson J. Endoscopic injection of adrenaline for benign oesophageal
ulcer haemorrhage. Br J Surg 1994;81:13178.
3. Pointner R, Schwab G, Konigsrainer A, Dietze O. Endoscopic treatment of Dieulafoy's
disease. Gastroenterology 1988;94:5636.
4. Chung SC, Leung JW. Improving the view in emergency endoscopy. Endoscopy 1987;19:47.
5. Rutgeerts P, Geboes K, Vantrappen G. Experimental studies of injection therapy for severe
nonvariceal bleeding in dogs. Gastroenterology 1989;97:61021.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Randall GM, Jensen JM, Hirabayashi K, et al. Controlled study of different sclerosing agents
for coagulation of canine gut arteries. Gastroenterology 1989;96:127481.
Whittle TJ, Sugawa C, Lucas CE, et al. Effect of hemostatic agents in canine gastric serosal
blood vessels. Gastrointest Endosc 1991;37:3059.
Leung JWC, Chung SCS. Endoscopic injection of adrenalin in bleeding peptic ulcers.
Chung SCS, Leung JWC, Steele RJC, et al. Endoscopic injection of adrenaline for actively
bleeding ulcers: A randomised trial. Br Med J 1988;296:16313.
Steele RJ, Logie JR, Munro A, Nichols DM. Endoscopic haemostasis in non-variceal upper
gastrointestinal haemorrhage using adrenaline injection. J R Coll Surg Edinb 1989;34:1336.
Steele RJC, Park KGM, Crofts TJ. Adrenaline injection for endoscopic haemostasis in
non-variceal upper gastrointestinal haemorrhage. Br J Surg 1991;78:4779.
Balanz J, Villanueva C, Sainz S, et al. Injection therapy of bleeding peptic ulcer. A
prospective, randomized trial using epinephrine and thrombin. Endoscopy 1990;22:1579.
Chung SCS, Leung FW, Leung JWC. Is vasoconstriction the mechanism of hemostasis in
bleeding ulcers injected with epinephrine? A study using reflectance spectrophotometry.
O'Brien JR. Some effects of adrenaline and anti-adrenaline compounds on platelets in vitro
and in vivo. Nature 1963;200:763.
Stevens PD, Lebwohl O. Hypertensive emergency and ventricular tachycardia after
endoscopic epinephrine injection of a Mallory-Weiss tear. Gastrointest Endosc 1994;40:778.
Chung SCS, Low JM, Cocks R, et al. Systemic absorption of epinephrine after endoscopic
submucosal injection in patients with bleeding peptic ulcers. Gastrointest Endosc
1993;39:202.
Soehendra N, Werner B. A new technique for endoscopic treatment of bleeding gastric ulcer.
Wrdehoff D, Gros H. Endoscopic haemostasis by injection-therapy in high-risk patients.
Endoscopy 1982;14:1969.
Asaki S, Nishimura T, Satoh A, et al. Endoscopic hemostasis of gastrointestinal hemorrhage
by local application of absolute ethanol: A clinical study. Tohoku J Exp Med 1983;141:37383.
Sugawa C, Fujita Y, Ikeda T, Walt AJ. Endoscopic hemostasis of bleeding of the upper
gastrointestinal tract by local injection of ninety-eight percent dehydrated ethanol. Surg
Pascu O, Draghici A, Acalovehi I. The effect of endoscopic hemostasis with alcohol on the
mortality rate of nonvariceal upper gastrointestinal hemorrhage. A randomized prospective
study. Endoscopy 1989;21:535.
peptic ulcers. A prospective, randomized trial. Gastroenterology 1990;99:13036.
Lin HJ, Tsai YT, Lee SD, et al. A prospectively randomized trial of heat probe
thermocoagulation versus pure alcohol injection in nonvariceal peptic ulcer hemorrhage. Am J
Lin HJ, Lee FY, Kang WM, et al. Heat probe thermocoagulation and pure alcohol injection in
massive peptic ulcer haemorrhage: A prospective, randomised controlled trial. Gut
1990;31:7537.
Levy J, Khakoo S, Barton R, et al. Fatal injection sclerotherapy of a bleeding peptic ulcer.
Lancet 1991;337:504.
Chester F, Hurley PR. Gastric necrosis: A complication of endoscopic sclerosis for bleeding
peptic ulcer. Endoscopy 1990;22:287.
Loperfido S, Patelli G, La Torre L. Extensive necrosis of gastric mucosa following injection
sclerotherapy of a bleeding peptic ulcer. Endoscopy 1990;22:2856.
Vallieres E, Jamieson C, Haber GB, Mackenzie RL. Pancreatoduodenal necrosis after
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
endoscopic injection of cyanoacrylate to treat a bleeding duodenal ulcer: A case report.

Surgery 1989;106:9013.
Rutgeerts P, Vantrappen G, Brocckaert L, et al. Comparison of endoscopic polidocanol
injection and YAG laser therapy for bleeding peptic ulcers. Lancet 1989;1:11647.
Kortan P, Haber G, Marcon N. Endoscopic injection therapy for non-variceal bleeding lesions
of the upper gastrointestinal tract. Gastrointest Endosc 1986;32:145.
Balanz J, Sainz S, Such J, et al. Endoscopic hemostasis by local injection of epinephrine and
polidocanol in bleeding ulcer. A prospective randomized trial. Endoscopy 1988;20:28991.
Soehendra N, Grimm H, Stenzel M. Injection of nonvariceal bleeding lesions of the upper
gastrointestinal tract. Endoscopy 1985;17:12932.
Hirao M, Kobayashi T, Masuda K, et al. Endoscopic local injection of hypertonic
saline-epinephrine solution to arrest hemorrhage from upper gastrointestinal tract.
Chen P-C, Wu C-S, Liaw Y-F. Hemostatic effect of endoscopic local injection with hypertonic
saline-epinephrine solution and pure ethanol for digestive tract bleeding. Gastrointest Endosc
1986;32:31923.
Pans J, Viver J, Forne M, et al. Controlled trial of endoscopic sclerosis in bleeding peptic
ulcers. Lancet 1987;2:12924.
Oxner RBG, Simmonds NJ, Gertner DJ, et al. Controlled trial of endoscopic injection
treatment for bleeding from peptic ulcers with visible vessels. Lancet 1992;339:9668.
Rajgopal C, Palmer KR. Endoscopic injection sclerosis: Effective treatment for bleeding peptic
ulcer. Gut 1991;32:7279.
Rutgeerts P, Gevers AM, Hiele M, et al. Endoscopic injection therapy to prevent rebleeding
from peptic ulcers with a protruding vessel: A controlled comparative trial. Gut
1993;34:34850.
Pans J, Viver J, Forne M. Randomized comparison of endoscopic microwave coagulation
with endoscopic sclerosis in the treatment of bleeding peptic ulcer. Gastrointest Endosc
1991;37:6116.
Waring JP, Sanowski RA, Sawyer RL, et al. A randomized comparison of multipolar
electrocoagulation and injection sclerosis for the treatment of bleeding peptic ulcer.
Chung SCS, Leung JWC, Leong HT, et al. Does adding a sclerosant improve the results of
epinephrine injection in actively bleeding ulcers? A randomised trial. Gastrointest Endosc
1993;39:6115.
Chiozzini G, Bortoluzzi F, Pallini P, et al. Controlled trial of absolute ethanol vs epinephrine as
injection agent in gastroduodenal bleeding. Gastroenterology 1989;96:A86.
Lin H-J, Perng C-L, Lee S-D. Is sclerosant injection mandatory after an epinephrine injection
for arrest of peptic ulcer haemorrhage? A prospective, randomised, comparative study?. Gut
1993;34:11825.
Lin H-J, Perng C-L, Lee F-Y, et al. Endoscopic injection for the arrest of peptic ulcer
hemorrhage: Final results of a prospective, randomized comparative trial. Gastrointest
Endosc 1993;39:159.
Chapter 31 Variceal Bleeding

(1803)
(1804)
R. A. J. SPENCE, M.D., F.R.C.S.
The most important of the many potential portosystemic collaterals are those that give rise to
esophageal and gastric varices. Varices have been described in the duodenum,13(1805) small
intestine,4(1806) colon,5(1807) and also around stomata,6(1808) but usually variceal hemorrhage
results from the effects of portal hypertension on the esophagus or the stomach. Furthermore,
although the current results of liver transplantation in selected patients in good centers are
excellent,7,8(1809) the worldwide majority of patients with bleeding varices are treated with injection
sclerotherapy, some form of local transection devascularization surgery, or, less commonly, a
portosystemic venous shunt operation.
The venous anatomy of the stomach and esophagus has significant implications for direct approaches
to management of esophageal and gastric varices, surgery as well as injection sclerotherapy. Since
the mid-1970s, many advances have been made in knowledge of the anatomy of esophageal and
gastric varices; in addition, the pathophysiology of variceal bleeding, although still somewhat
controversial, is becoming clearer. Endoscopic appearances of esophageal and gastric varices are
now of increasing importance, and various classifications have been proposed in an effort to predict
bleeding so as to concentrate resources on those patients who are at high risk of hemorrhage.
This chapter discusses the anatomy and histology of esophageal and gastric varices, current concepts
of the pathophysiology of bleeding, the role of currently available endoscopic classifications of varices,
and the value of those classifications in the prediction of bleeding.
Esophageal Venous Anatomy

Until the mid-1980s, the detailed venous anatomy of the lower esophagus had been little studied in
detail. Kegaries9(1810) in 1934 studied eight normal specimens and found three or four longitudinal
venous trunks in the esophagus with few cross-anastomoses. He demonstrated a fine capillary
anastomosis between the portal and systemic circulations.
The next major contribution was that of Butler,10(1811) who described three components: intrinsic
veins consisting of a subepithelial and a submucosal plexus, venae comitantes arranged along the
vagi, and extrinsic veins. Butler's studies, however, were mainly on fetuses, and most of his fixed
cadaver and fresh postmortem specimens were proximal esophagi; only two were adult stomach and
esophagus. MacBeth11(1812) in 1955 confirmed Butler's findings and emphasized the importance of
the subepithelial veins in variceal bleeding. This finding formed the pathologic basis of his injection
sclerotherapy. Beswick and Butler12(1813) further emphasized the importance of subepithelial veins
based on observations in a patient who had bled from varices. They found that the dilated subepithelial
vessels extended only 2 inches above the cardia; these appeared to be the site of bleeding.
DeCarvalho13(1814) published his very detailed anatomic studies in 1966. Using injection techniques,
he found that the veins in the distal esophagus seemed to lie mainly in the mucosa in regular palisades
extending 1.5 to 3.5 cm from the gastroesophageal junction, whereas in the stomach and proximal
esophagus the veins seemed to lie mostly in the submucosa. Similar findings were reported by
Stelzner and Lierse14(1815) in studies using material mainly from the rhesus monkey.
A reexamination of the venous anatomy of the lower esophagus using modern techniques of image
analysis has shown three distinct zones in the proximal stomach and esophagus.15,16(1816) These
data clearly demonstrate that the veins in the distal esophagus lie superficially in the lamina
propria.16(1817) In the stomach (Zone 1), the relative area of the lamina propria occupied by veins
was 2.6% in normal specimens and 3.7% in specimens from patients with varices (Figure 311). In the
distal esophagus (Zone 2), the mean area occupied by veins in the lamina propria was 19.8% (normal)
and 32.8% (varices). Zone 2 extended to 2 to 5 cm from the gastroesophageal junction. Zone 3 was
the remainder of the esophagus, and there the area of the lamina propria occupied by veins was 4.9%
(normal) and 6.1% (varices). A reciprocal pattern was found in the submucosa. On an anatomic basis,
this work indicates that varices are prone to rupture in the distal 2 to 5 cm of the esophagus. The
subsequent demonstration, using Doppler ultrasonography, of a constant perforating vein at 36 cm in
the esophagus is further evidence of this unique zone in the distal esophagus.17(1818)
(1819)Figure 311. Graphic representation of the venous anatomy of the lower esophagus (of
normal and variceal subjects). The relative areas occupied by veins within the lamina propria
and the submucosa were determined using image analysis techniques. (From Spence RAJ. The
venous anatomy of the lower oesophagus in normal subjects and in patients with varicesAn
image analysis study. Br J Surg 1984; 71:73944. By permission of Blackwell Science Ltd.,
publisher.)
Teleologically, the explanation for the unique venous distribution in the lower esophagus may be the
high-pressure zone immediately above the cardia. A definite area of thickened muscle in the lower
esophagus corresponding to this zone of high pressure has been shown by Liebermann-Meffert et
al.18(1820)
Further evidence of the unique nature of the gastroesophageal transition zone has been reported by
Fujimura and Fereaz de Carvalho.19(1821) Using India ink techniques, three venous nets in the
muscular layer of the gastroesophageal transition zone were foundperimuscular, intramuscular, and
submuscular. These three nets of veins, especially the perimuscular and intramuscular, were shown to
converge at either extremity of the transition zone to large confluent trunks that proceed via perforating
veins to tributaries of the portal vein.
Noda20(1822) injected contrast medium into esophageal veins in 25 autopsy specimens from patients
with varices and four specimens from subjects without varices. The area of rupture, termed the critical
area, was identified as the orad end of longitudinal veins, called sudare-like veins by Noda, in the
lamina propria of the distal esophagus. By means of morphometric techniques, dilation of these
sudare-like veins was found to be most obvious in patients who had severe portal hypertension with
large varices. Furthermore, these veins penetrated the muscularis mucosa to connect with the
submucosal veins at the so-called critical area. Noda emphasized that this critical area appears to be
the region of greatest importance in the pathogenesis of spontaneous variceal rupture.
Further complementary work is that of Kitano et al.,21(1823) who studied venous anatomy in fresh
cadaver specimens of the lower esophagus and proximal stomach using high-resolution resin casts.
This study, which provides three-dimensional information regarding the venous anatomy, describes
four layers of veins in the esophagi of both normal subjects and patients with portal hypertension.
Intraepithelial channels drained into a superficial venous plexus that connected to larger deep intrinsic
veins. Both the superficial plexus and the deep intrinsic veins communicated directly with counterpart
veins in the stomach. As demonstrated in previous studies, perforating veins were found to connect the
deeper veins with the adventitial plexus, which was considered to be the fourth layer. In patients with
portal hypertension, all of these veins were significantly dilated. Kitano et al.21(1824) proposed that
typical large esophageal varices arise from the main trunks of the deep intrinsic veins, which
communicate directly with gastric varices. The deep intrinsic veins, enlarged in portal hypertension,
appear to represent the large varices seen at endoscopy. These have no direct cross-connections but
are linked to the superficial plexus of veins in which intercommunications are abundant so that varices
are actually interconnected but via an indirect route.
Intraepithelial channels had been previously described by Spence et al.22,23(1825) These dilated
blood-filled channels within the squamous epithelium probably correspond to the dilated intraepithelial
channels found by Kitano et al.21(1826) The additional findings of Spence et al.22(1827) of large
subepithelial channels eroding through the epithelium appear to be compatible with the description of
Kitano et al.21(1828) in which both the superficial venous plexus and the underlying variceal channels
extend up to the epithelium.
Based on the report of Kitano et al.21(1829) and earlier work of Spence et al.,22,23(1830) it can be
postulated that a major episode of variceal bleeding occurs as a result of rupture of the large deep
intrinsic variceal channels that lie adjacent to the epithelial surface, or, alternatively, from a major
branch of the superficial venous plexus at a point near a direct connection to a larger varix. On the
other hand, minor variceal bleeding, which can stop spontaneously, may occur from a branch of the
superficial venous plexus via a nearby connection to a large varix or, alternatively, from the smaller
dilated intraepithelial channels. The intraepithelial vascular channels or indeed the subepithelial
blood-filled channels, previously described by Spence et al.22(1831) and confirmed in the anatomic
studies of Kitano et al.,21(1832) may correlate with the so-called cherry-red spots described at
endoscopy by Japanese investigators (see later).
Presumably, successful sclerotherapy causes thrombosis and eradicates the main vascular channels
along with their communications with the superficial venous plexus. This explains the efficacy of
intravariceal injection techniques. Presumably the communications between the venous channels are
responsible for recurrence of varices after sclerotherapy. Paravariceal sclerotherapy24(1833) may
work by inducing thickening and fibrosis of the overlying mucosa and thrombosis of the smaller
superficial venous plexus; it may not affect the main vascular channels.25(1834)
Vianna et al.26(1835) studied the normal anatomy of the distal esophagus using a combination of
radiographic techniques, corrosion casting, and morphometry. Again, four distinct zones of venous
drainage were identified (Figure 312): a gastric zone characterized by longitudinal venous distribution;
a palisade zone composed of parallel vessels arranged in groups lying mainly within the lamina
propria; a perforating zone characterized by treble clef-shaped veins that seem to collect and channel
blood into intrinsic veins; and a truncal zone composed of four or five deep veins. These investigators
suggested that this anatomic pattern implies that venous flow is bidirectional at the palisade zone and
this zone acts as a high-resistance watershed region between the portal and the azygos systems. This
work, using a different technique, complements that of Spence.16(1836) It appears therefore that the
area of maximum resistance to blood flow through the gastric into the esophageal intrinsic veins is at
the palisade zone. It could be assumed that the high-pressure zone occurs below this; again, this may
explain the propensity for varices to bleed from the distal esophagus and indeed may also explain the
appearance of gastric varices.
(1837)Figure 312. A, Radiograph of a specimen of the distal esophagus and proximal

stomach injected with barium-gelatine. Specimen opened along the greater curvature. The four
zones of normal venous drainage identified are the gastric zone (GZ), the palisade zone (PZ),
the perforating zone (PFZ), and the truncal zone (TZ). The transition between stratified
squamous and columnar epithelium is demarcated by a radiopaque wire (GOJ). B,
Magnification (3) of radiograph of specimen injected with barium-gelatine demonstrating the
palisade zone (PZ). Longitudinal veins anastomose with large venous trunks at the proximal
perforating zone (PFZ) and the distal gastric zone. C, Detail of the perforating zone showing
"treble-clef" veins (TCV) and the perforating vein (Pfv). (A-C, From Vianna A, Hayes PC,
Moscoso G, et al. Normal venous circulation of the gastroesophageal junction.
Gastroenterology 1987; 93:8769.)
The detailed histology and microanatomy of the lower esophagus in patients with esophageal varices
have been studied in detail since the mid-1980s. In the late 1970s and early 1980s, esophageal
transection for bleeding varices became increasingly popular because of encouraging short-term
results.27,28(1838) This operation excises a ring of full-thickness distal esophagus from the usual site
of variceal bleeding, followed immediately by an anastomosis made with the circular stapling gun. This
provided the opportunity to study fresh tissue from the most common site of esophageal bleeding.
During a study to investigate the incidence of esophagitis in patients undergoing esophageal
transection for varices, dilated vascular channels were noted within and immediately below the
esophageal epithelium (Figure 313).22(1839) In a detailed investigation, these vascular channels
were found in all esophageal rings removed at transection for varices.23,29(1840) These were
compared with rings removed using the circular gun for reanastomosis during resections for
esophageal and gastric tumors. Although a small number of nonvariceal patients had epithelial
vascular channels, these were significantly larger and more numerous in patients with varices. The
number and the area of dilated epithelial and subepithelial vascular channels were found to be greater
in the variceal patients.23,29(1841) The depth of the esophageal papillae were also found to be greater
in the variceal patients.
(1842)Figure 313. Photomicrograph of a section from an esophageal transection ring in a

patient with varices demonstrating intraepithelial and subepithelial vascular channels.
(Courtesy of Dr. J. Sloan.)
By electron microscopy, the epithelial vascular channels were found to be lined with flattened cells that,
although not typical of endothelial cells, stained positively for Factor VIII-related antigen using indirect
immunofluorescence. The latter antigen has been localized only on endothelial cells, platelets, and
megakaryocytes.30(1843) Electron microscopy showed no evidence of platelet lining of these vascular
channels; it therefore seems that these flattened lining cells were functioning as endothelial cells.
Burnaud et al.31(1844) demonstrated an increase in the size of the skin capillary bed in response to
sustained local venous hypertension in the dog. Similar findings have been reported in the human skin
of the lower leg in the presence of venous hypertension.32(1845) It is therefore possible that prolonged
portal venous hypertension may similarly induce an increase in the size of the capillary bed of the lower
esophagus; the increased length of the papillae may be a manifestation of this rather than a feature of
esophagitis.
We suggested that these histologic findings of epithelial vascular channels may correlate with the
so-called cherry-red spots and red wale markings, described at endoscopy by Japanese
investigators,33,34(1846) which are now thought to correlate with bleeding from varices (see later).
Furthermore, these vascular channels may give a clue to the pathogenesis of variceal rupture (see
later).23(1847)
The existence of epithelial vascular channels has been confirmed by Kitano et al.21(1848) and more
recently by the work of Hashizume et al.,35(1849) in which specimens from nine patients with portal
hypertension and five normal patients were studied using injection techniques. These investigators
identified four venous channels: intraepithelial, subepithelial superficial, deep submucosal, and
adventitial. In patients with portal hypertension, so-called giant esophageal varices form 2 to 3 cm
proximal to the gastroesophageal junction, and these fuse with numerous superficial and deep
submucosal veins. Gastric varices were found in the submucosa, where branches of the left gastric
vein penetrated the gastric wall 2 cm distal to the gastroesophageal junction. Hashizume et al.35(1850)
subdivided the distal esophageal varices into two further types: palisading and bar. The palisading type
had dilated intraepithelial vascular channels and many small superficial collateral veins extending in a
longitudinal arrangement. The bar type had triple dilated subepithelial superficial veins and deep
mucosal veins that had eroded the epithelium. This latter type was associated with gastric varices.
The propensity for variceal recurrence after sclerotherapy is evident from the extensive network of
small superficial collateral veins that lie below the epithelial vascular channels. It seems reasonable,
therefore, that the combined approach of intravariceal and paravariceal sclerotherapy may be more
efficacious than either alone (see Chapter 32: Technique of Endoscopic Sclerotherapy). For example,
Kitano et al.36(1851) found that sclerotherapy to produce a marked fibrotic thickening of the
submucosa in the lower esophagus and deliberately to produce shallow ulcers that subsequently heal
with fibrosis to be very effective.
The anatomic and the histologic studies of esophageal varices during the 1980s are summarized by
Arakawa and Kage.37(1852) Many of these are complementary and indicate that the distal 2 to 5 cm of
the esophagus is the so-called danger or critical area where the veins lie superficially in the lamina
propria. These data in conjunction with the new histologic findings of epithelial vascular channels have
clinical correlations in terms of the site of bleeding, the pathogenesis of rupture, and endoscopic
findings.
The modern view of the venous anatomy of the lower esophagus explains the observation by many
workers that varices tend to rupture within 2 to 5 cm of the gastroesophageal junction. In an autopsy
study from 1953, Chiles et al.38(1853) noted that variceal bleeding arose in most instances from the
distal third of the esophagus. This was confirmed in a later study of operative specimens by Orloff and
Thomas.39(1854) Many other investigators with extensive experience in the field of portal hypertension
have commented that bleeding nearly always occurs from within the distal 5 cm of the
esophagus.4042(1855) Terblanche43(1856) has pointed out that pressure exerted with the rigid
esophagoscope at the gastroesophageal junction usually results in immediate albeit temporary control
of variceal bleeding. Other possible mechanisms that have been proposed to account for the
predilection for variceal bleeding in the distal esophagus include acid reflux-induced esophagitis and
the pressure changes that occur in the region of the cardia. However, no evidence indicates that
esophagitis is a precipitating factor in the rupture of varices,29(1857) and it is now generally accepted
that the venous anatomy of the lower esophagus accounts for the fact that this area is almost always
the site of initial rupture. Furthermore, the modern view of the venous anatomy of the lower esophagus
provides the anatomic and pathologic rationale for procedures such as sclerotherapy and esophageal
transection-devascularization procedures that directly treat the critical danger area in the distal
esophagus.28,44(1858)
Natural History of Esophageal Varices

The natural course of patients with esophageal varices is unpredictable. The mortality associated with
an index episode of bleeding is relatively high, and patients are at increased risk for further bleeding.
The clinical symptoms of esophageal varices relate only to hemorrhage. An accurate assessment of
the prevalence of esophageal varices is probably impossible because hemorrhage is not invariable.
About two thirds of cirrhotic patients have varices in autopsy series.45(1859) At endoscopy, between
35 and 84% of such patients have varices. Furthermore, the incidence of hemorrhage for patients with
varices differs among various studies because of differences in design. In the prospective study of the
North Italian Endoscopic Club (NIEC)46(1860) of 321 patients with esophageal varices and no prior
bleeding, 26.5% bled during the study period (median 23 months). About two thirds of patients bled
during the first year of the study. By contrast, the incidence of variceal bleeding was 66% in the study
of Paquet,47(1861) although this refers to patients with large varices and other endoscopic findings
known to be associated with an increased likelihood of hemorrhage.
The mortality for variceal bleeding was greater than 30% in a survey of the membership of the
American Society for Gastrointestinal Endoscopy (ASGE).48(1862) Graham and Smith49(1863) found
a mortality rate of 42% at 6 weeks after an index episode of variceal hemorrhage, with most deaths
being related to bleeding. In the absence of some form of intervention, more than 60% of patients who
survive an index episode of variceal hemorrhage experience recurrent bleeding.50,51(1864)
Pathogenesis of Esophageal Variceal Rupture

The exact precipitating cause of esophageal variceal rupture is still unclear. However, several factors
have been identified that probably play a role in the pathogenesis of rupture.
Gastroesophageal reflux with mucosal erosions on the surface of large, thin-walled varices was once
considered to be a primary factor in variceal hemorrhage. It is now clear that erosive esophagitis plays
virtually no role. In early studies, esophageal ulceration was noted in 43 to 56% of patients who died as
the result of acute variceal bleeding.38,52(1865) Later studies, however, disclosed no histologic
evidence of esophagitis.29,53(1866) Esophageal function in cirrhotic patients with esophageal varices
has also been found to be normal.54(1867)
That rupture tends to occur mainly in the distal esophagus is well established. Currently, the main
factors thought to play a role include intravariceal pressure and portal pressure, variceal size, and varix
wall tension, as evidenced by the cherry-red spots described by the endoscopist and
anatomist.55(1868)
Portal Pressure
Increased portal pressure is required for the development and rupture of varices; a pressure of
approximately 12 mm Hg is needed before significant varices can develop and bleed.56(1869)
However, whether a linear correlation exists between bleeding and pressure at portal pressures
greater than 12 mm Hg is controversial.57(1870)
Blood volume expansion in patients with portal hypertension may increase portal pressure greatly and
precipitate variceal hemorrhage.58(1871) Furthermore, significant increases in portal venous pressure
can occur during coughing and the Valsalva maneuver. These maneuvers can also affect variceal
intraluminal pressure.59(1872)
Although the existence of a linear relationship between portal hypertension and initial variceal bleeding
is controversial, increasing evidence indicates that the degree of portal pressure elevation has some
relation to the risk of recurrent bleeding.60,61(1873) The degree of elevation may have some
predictive value in terms of the risk of early recurrent bleeding, provided that the pressure is measured
within 24 hr of an initial hemorrhage.62(1874) A multivariate analysis has found portal pressure to be
an important factor in predicting recurrent bleeding.63(1875) Physiologic stress theoretically can also
enhance portal pressure and may play a minor role.64(1876)
Intravariceal Pressure
Intravariceal pressure was recorded by Rigau et al.61(1877) using an endoscopically directed pressure
gauge in 70 cirrhotic patients, including 47 with documented variceal bleeding and 23 who had no prior
history of bleeding. Intravariceal pressure was significantly higher in the patients with prior variceal
bleeding. Intravariceal pressure measurements were compared with wedged and free hepatic venous
pressures, and for both groups of patients intravariceal pressure was less than portal pressure.
Variceal pressure was significantly higher in patients with large varices. For both groups of patients, the
hepatic venous pressure gradient was elevated, but the degree of elevation was not significant.
By performing fine-needle variceal punctures in 40 patients with cirrhosis, Kleber et al.65(1878) found
intravariceal pressure to be significantly higher in varices with red color signs. However, no relationship
was found between the transmural variceal pressure and the number or diameter of esophageal
varices or the severity of liver dysfunction. In a similar study of 40 patients, Beppu and
Hashizume66(1879) also found that intravariceal pressure, as determined by fine-needle puncture at
endoscopy, was significantly higher for varices with red color signs. However, correlation with variceal
diameter, the number of variceal columns, or severity of liver disease (Child's classification) was not
found.
Variceal Size
It is now recognized that variceal size, measured endoscopically, is important in the assessment of risk
of hemorrhage. The size of the varix is determined by the intravariceal pressure and the local anatomic
and histologic features of the critical zone (see earlier). A correlation probably also exists between the
azygos vein blood flow and the size and extent of esophageal varices.67(1880)
Patients with large varices tend to bleed more frequently than patients with small varices, including
those who have not previously bled as well as those at risk for recurrent bleeding.34,68,69(1881) By
reference to Laplace's Law, Polio and Groszmann70(1882) have emphasized the interactions between
varix wall tension, thickness, and transmural pressure as determinants of variceal rupture. Large
varices are surrounded by less supporting tissue than small varices and may therefore more easily
reach the wall tension necessary to trigger rupture at a given portal pressure. The endoscopic findings
described by Beppu et al.34(1883) and subsequently assessed by the NIEC46(1884) are thought to
reflect the microvasculature of esophageal varices as outlined earlier. These histologic and endoscopic
findings indicate varices at risk of rupture. It was previously believed that the so-called white nipple or
Mount St. Helen's sign was a further endoscopic finding in patients who had recently bled. This was
thought to be a fibrin platelet plug at the site of recent variceal hemorrhage.71(1885) However, a
prospective study has shown that this "sign" has no predictive value with respect to bleeding.72(1886)
Nakaniski et al.73(1887) stressed the importance of angiography in determining the size of varices and
indicated that more than two thirds of patients with large varices have hepatofugal variceal flow, with a
narrow left gastric artery and a large left gastric vein. Patients with small varices seem to have a
greater proportion of hepatopetal flow. It therefore seems that the hemodynamics in patients with large
varices differ from those in patients with small varices. Also, Nakaniski et al.73(1888) believe that
patients with gastric varices have a different flow type than those with esophageal varices. Worldwide,
relatively little emphasis has been placed on such angiographic findings.
Hepatic Dysfunction
Multivariate analysis has shown that patients with the worst hepatic dysfunction are at increased risk
for an initial episode of bleeding as well as episodes of recurrent bleeding.63(1889) The reason for this
is unclear but may be associated with higher portal pressure, the presence of ascites, and peripheral
vasodilation.
By correlating the detailed microanatomy of esophageal varices with endoscopic findings, Hashizume
et al.74(1890) have shown that the pattern of veins described as palisading35(1891) are characterized
by a number of fine longitudinal venous channels seen through the mucosa. Patients with this finding
seem to have less elevation of portal vein pressure and better liver function than patients with the
so-called bar type of veins, which are apparent in the subepithelial superficial venous plexus.
In the study of Siringo et al.,75(1892) the severity of liver disease correlated with the presence of
esophageal varices and red color signs but lacked correlation with variceal size. However, portal blood
flow velocity, as determined by Doppler ultrasonography, did correlate with the presence and size of
esophageal varices.
Endoscopic Diagnosis and Classification of Esophageal Varices

Endoscopic diagnosis of esophageal variceal bleeding is essential, as upper gastrointestinal (GI)
bleeding in a patient with varices not infrequently arises from some other source. Bleeding in patients
with noncirrhotic causes of portal hypertension, such as schistosomiasis and portal vein thrombosis,
usually arises from varices.76(1893) However, in patients with cirrhosis, especially those with an
alcoholic background, up to 50% of bleeding episodes are caused by other lesions apart from varices,
either gastric erosions, peptic ulceration, or portal hypertensive gastropathy. The latter cause is being
diagnosed with increasing frequency.51(1894)
It is generally recommended that endoscopy be performed within the first 12 hr after the patient
presents with bleeding.77(1895) At least 80% of variceal bleeding arises within the distal 2 to 5 cm of
the esophagus.76(1896) Variceal bleeding is frequently intermittent so that active variceal bleeding or
evidence of recent bleeding (e.g., fibrin plug or clot) may not be seen at the time of endoscopy.
Furthermore, other potential sources of bleeding may be present at endoscopy, leading to uncertainty
about the source of bleeding as well as delays in arriving at a correct plan of management.
Many retrospective and prospective studies have emphasized that a varix is the usual source of
bleeding in patients with upper GI hemorrhage and portal hypertension.7883(1897) In the prospective
study of Mitchell et al.,80(1898) 90 consecutive initial endoscopic examinations performed within 24 hr
of the onset of upper GI bleeding in patients with portal hypertension revealed variceal bleeding in only
21 (23.3%) cases. Although other lesions were pres-ent in 38.8% of the procedures, only 5 were noted
to be actively bleeding. Furthermore, bleeding was recurrent in 39 of the 64 cases (60.9%) in which
active bleeding was not observed at initial endoscopic evaluation. Endoscopy repeated within 1 hr of
the onset of recurrent bleeding in 27 cases disclosed active variceal bleeding in 20 (74.1%) of cases.
Endoscopic Classification
The endoscopic appearances of varices have assumed greater importance with the advent of
prophylactic endoscopic treatment of patients with varices. Earlier classifications of the appearance of
esophageal varices at endoscopy focused on size, that is, diameter and linear extent. Endoscopic
grading of esophageal varices is by no means standardized. Similar terms (e.g., Grade 1, Grade 2) are
used in many classification systems, but their meaning can be different. Proposed classifications have
also incorporated the appearance of varices, including such observations as varices upon varices and
cherry-red spots.34(1899)
Dagradi40,84(1900) suggested one of the first classifications of esophageal varices by dividing varices
into Grades 1 to 5 (Table 311). This system was devised using the rigid esophagoscope, whereas
most endoscopists now use flexible instruments. In an early study of the natural history of esophageal
varices in patients with alcoholic liver disease, Dagradi84(1901) classified varices endoscopically into
five grades based on diameter. In Grade 1, small varices (2 mm diameter) were noted only by
compression of the esophageal wall with the endoscope. Diameters of 2 mm, 3 to 4 mm, and 5 mm or
more were classified as Grades 2, 3, and 4, respectively. Variceal diameter was equal for Grades 4
and 5, the difference between these two grades being the presence of "small varices on top of
varices." These were characteristic bright red, small, telangiectatic vessels also described as
cherry-red spots. Varices first appeared on the anterior wall of the distal esophagus, always at the level
of the gastroesophageal junction. When fully developed, there were usually four large columns,
although Dagradi found that the anterior column was always the most prominent. Variceal caliber was
always greatest in the distal esophagus and decreased progressively in a step-wise fashion in an orad
direction. Varices of a given caliber usually increased in length to varying degrees before their diameter
increased to the next highest grade. The most highly developed and extensive variceal columns
usually terminated at the 24-cm level.
Modified Dagradi Classification of

Esophageal Varices
TABLE 311
Grade 1
Grade 2
Grade 3
Grade 4
Blue or red varices <2 mm in diameter

Blue varices 23 mm in diameter
Elevated blue veins 34 mm in diameter
Tortuous blue varices \gr4 mm in diameter, almost
meeting in the midline
Modified Dagradi Classification of

Esophageal Varices
TABLE 311
Grade 5
Grapelike varices occluding the lumen and showing the

presence of small cherry-red varices overlying
blue-gray varices
Data from Dagradi AE. The natural history of esophageal varices in
patients with alcoholic liver cirrhosis. An endoscopic and clinical
study. Am J Gastroenterol 1972; 57:52040; and Dagradi AE,
Arguello JF, Weingarten ZG. Failure of endoscopy to establish a
source for upper gastrointestinal bleeding. Am J Gastroenterol 1979;
72:395402.
Other investigators have proposed endoscopic classifications that refer to the size of varices and in
some cases their appearance. For example, Palmer and Brick85(1902) measured the length and
diameter of varices at endoscopy in 201 patients with proven cirrhosis and varices. Variceal diameter
was classified as mild (<3 mm diameter), moderate (3 to 6 mm diameter), and severe (>6 mm
diameter). The longitudinal extent of variceal columns was measured as a fraction of the total length of
the esophagus. In the classification of Baker et al.,86(1903) variceal size was classified as Grade 0 (no
definite varices visible), 1+ (one or more varices under 4 mm diameter and under 4 cm in length), 2+
(multiple varices 4 to 10 cm in extent), and 3+ (multiple varices over 10 cm long). Paquet47(1904)
devised a system that attempts to identify patients who are likely to bleed from their varices and are
therefore candidates for prophylactic sclerotherapy. Paquet's classification uses four grades. The
Japanese Research Society for Portal Hypertension proposed a very detailed and complicated
classification for esophageal varices in 1981 (Table 312).34(1905) The purpose of this classification
was to identify patients at high risk of bleeding who may be treated by prophylactic surgery or
sclerotherapy. This complex classification involves five factors: fundamental color, red color signs,
form, location, and adjunctive findings (Table 312).
General Rules for Recording Endoscopic

Findings on Esophageal Varices (Japanese Research Society
for Portal Hypertension)
TABLE 312
1. Fundamental color
a. White (Cw)
b. Blue (Cb)
2. Red color signs (RCS)* (small dilated vessels or microtelangiectasia on
varix surface)
a. Red wale marking (RWM)
b. Cherry-red spot (CRS)
c. Hematocystic spot (HCS)
d. Diffuse redness (DR)
3. Form
a. Small, straight varices (F1)
b. Enlarged tortuous varices occupy less than one third of lumen (F2)
c. Largest coil-shaped varices occupy more than one third of lumen (F3)
4. Location (longitudinal extent)
a. Lower one third (Li)
b. Mid-one third below tracheal bifurcation (Lm)
General Rules for Recording Endoscopic

Findings on Esophageal Varices (Japanese Research Society
for Portal Hypertension)
TABLE 312
c. Upper one third tracheal bifurcation (Ls)

5. Adjunctive findings
a. Erosion (E+)
Data from Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal
hemorrhage by esophageal endoscopy. Gastrointest Endosc 1981; 27:2138.
* The RWM and CRS are graded from 1+ to 3+ depending on number and
extent.
The fundamental color of varices may be classified as either white or blue.

The term red color sign refers to the presence of small dilated vessels on the variceal surface.
Dagradi84(1906) and also Paquet47(1907) had previously described similar signs and indicated that
these represented a major risk factor for bleeding. However, the Japanese classification subdivides the
red color sign into (1) red wale markings, dilated venules arranged longitudinally on the varices; (2)
cherry-red spots, small red dots about 2 mm in diameter on the surface of the varices; (3)
hematocystic spots, round dark red blood blisters about 4 mm or larger in diameter; and (4) diffuse
redness, a diffuse red discoloration of the variceal surface (see Table 312).
Most classifications of varices include estimates of variceal size, although classifications based on
observations with the rigid esophagoscope may not be pertinent in the era of flexible endoscopes. In
the Japanese classification, form refers to the size and also the shape of varices. Three forms (grades)
from 1 to 3 are recognized, form 1 being small and straight, form 2 being slightly enlarged tortuous
varices occupying less than one third of the esophageal lumen, and form 3 being coil-shaped varices
occupying more than one third of the esophageal lumen.
Of all the various systems for describing varices, only the Japanese classification employs a category
of location to define the upper limit of the varices. Three subtypes are used (see Table 312). The
category of adjunctive findings refers to the presence or absence of esophagitis, which is termed
positive or negative.
The classification of varices proposed by the Japanese Research Society for Portal Hypertension is
generally regarded as highly complex but potentially very useful. Other classifications have been
proposed that attempt to better define variceal size. The Italian Liver Cirrhosis Project divided varices
into small, medium, and large sizes, these designations corresponding to less than 30%, 30 to 60%,
and 60% of the maximum potential sizea varix that occupies half of the esophageal lumen.87(1908)
The NIEC uses a similar definition of small, medium, and large varices. Because all variceal columns
are not necessarily of the same size in a given patient, the NIEC took the size of the largest varix as
the size for the purpose of their studies.46(1909)
Reproducibility of Endoscopic Observations

Concern has been expressed that use of the very complex variceal grading system proposed by the
Japanese Research Society for Portal Hypertension would result in marked variation in the description
of varices between observers. This has important implications for studies that incorporate data from
multiple centers and for comparison of data from different centers.46,87,88(1910)
De Franchis et al.89(1911) addressed the problem of interobserver variation. The agreement between
observers (kappa index) was fair to good with respect to size of varices and red color sign, but
significant variation was seen for the other parameters described in the Japanese classification.
Intraobserver variability in the endoscopic assessment of variceal features was studied by Cales et
al.88(1912) in 100 consecutive cirrhotic patients. Variceal size, color, and extent and the presence of
red color signs were assessed by four independent observers. Agreement was poor for variceal color
and extent but good for variceal size and red color signs, suggesting that only certain features in the
endoscopic appearance of esophageal varices are distinguished consistently by experienced
endoscopists.
Bendtsen et al.90(1913) found considerable variability in the diagnosis and grading of esophageal
varices among 22 endoscopists who reviewed videotapes of esophagoscopies from 28 cirrhotic
patients who were not known to have varices and who had no history of prior variceal bleeding. With
regard to the presence or absence of varices, endoscopists agreed in 70% of cases.
Endoscopic and Histologic Correlations

The endoscopic findings described by the various classifications of esophageal varices correlate with
the detailed morphology of the lower esophagus and the modern view of variceal anatomy and
histology.
The red color sign explains the relative ease with which esophageal varices rupture. These dilated
vessels are adjacent to the epithelial papillae on a superficial layer of varices that is close to the
surface. Clearly, they can rupture with ease and cause at least minor variceal hemorrhage. The red
color sign has been termed intraepithelial microvarices by Kumagai et al.91(1914) This finding
corresponds closely to the observations of Spence.16(1915) In the classification of Dagradi,84(1916)
the term varices upon varices probably corresponds to the cherry-red spots of the Japanese
classification. Arakawa et al.92(1917) suggest that the larger cherry-red spots may be due to thinning
of the epithelium, causing the subepithelial varices to lie very close to the surface and to be clearly in
danger of rupture, as has also been pointed out by Spence et al.22(1918) In the classification of
Paquet,47(1919) it is likely that the term erosions corresponds to the cherry-red spots.
Prediction of Variceal Bleeding

In 1959, Baker et al.86(1920) demonstrated that the tendency for patients with large varices to bleed
was substantially greater than that for patients with small to medium-sized varices. These investigators
studied the natural history of varices in 115 cirrhotic patients who had no prior variceal bleeding.
Variceal size was classified into four grades (see earlier) from 0 (no definite varices visible) to 3+
(multiple columns over 10 cm in length). The incidence of bleeding was approximately the same in
patients with 1+ and 2+ varices, but was twice as frequent among those patients with 3+ varices.
However, the severity of bleeding was inversely proportional to the extent of the varices. One of 11
patients with 3+ varices died as a result of the initial episode of bleeding, and 8 of 13 patients with 1+
varices died as a result of the first hemorrhage. Only 28 patients had follow-up endoscopy, but varices
disappeared in 9 patients, the grade decreased in 7, the varices remained unchanged in 6, and they
became more extensive in 6. The authors concluded that the endoscopic classification of varices in a
given patient is not static.
In 1956, Palmer and Brick85(1921) measured the length and diameter of varices at endoscopy in 201
patients with proven cirrhosis and varices. Some patients had a history of variceal bleeding but others
did not. Variceal diameter was classified as mild, moderate, and severe; the length of variceal columns
was expressed as a fraction of the length of the esophagus. Three fourths of the patients examined
during active hemorrhage had varices classified as severe, and in half of the patients the entire length
of the esophagus was involved. If endoscopy was performed shortly after an episode of bleeding, there
was a tendency for vessel diameter and extent to be greater than similar determinations in patients
with a longer time interval between bleeding and endoscopy. Among 90 patients who were examined
more than once (intervals of 1 week to 1 year), spontaneous disappearance of the varices was
observed in 14%, and varices appeared in 10% of patients in whom the esophagus was previously
normal. Palmer and Brick concluded that most patients have large-diameter varices that extend over a
substantial length of the esophagus at the time of variceal bleeding. Small varices tended to enlarge
just before the onset of bleeding, and variceal size decreases during the interval between episodes of
bleeding. Because of variability in the graded severity of varices, Palmer and Brick believe that small
size alone offers no assurance that the chance for bleeding is small.
Dagradi84(1922) performed follow-up endoscopic examinations in alcoholic patients who continued to
drink and noted a tendency toward variceal enlargement, both in diameter and in longitudinal extent,
after initial appearance. On average, about 50 months were required to reach grades 4 and 5. Varices
never disappeared in patients who continued to consume alcohol, and in 65% of them there was a
transition to a higher grade. However, a significant decrease in caliber and extent was noted in 80% of
patients who stopped drinking. In a prospective serial endoscopic study of alcoholic cirrhotic patients,
Dagradi et al.40(1923) noted that cherry-red spots were found only on large-diameter varices.
Subsequently, all patients who bled had been shown at first endoscopy to have large varicesgreater
than 5 mm in diameterwith cherry-red spots.
Lebrec et al.69(1924) studied variceal size and the risk of bleeding in 100 patients. Forty-seven
patients without prior bleeding were evaluated because of signs of liver disease. The other 53 patients
had bleeding that was due to gastric erosions in 27 patients and varices in 26 patients. Barium swallow
radiography was performed in all patients within 2 weeks of entry, and the varices were classified
according to size by two independent radiologists. Varices were graded as large or small according to
whether a varix was more or less than 5 mm in diameter. The radiologists agreed about 94 patients.
Both radiographic and endoscopic variceal gradings were obtained in the 53 patients with prior
bleeding, and the radiologists and the endoscopist agreed in 52 cases. Patients were followed for 1
year, during which time bleeding occurred in 17 patients, 14 of whom were in the group with prior
bleeding and 3 were from the group with only signs of liver disease at entry. No correlation was found
between the variceal size, whether measured by radiography or endoscopy, and the gradient between
the wedged hepatic pressure and the free hepatic venous pressure. Furthermore, no statistically
significant difference was seen in this gradient between patients with and without bleeding on entry.
The prevalence of recent bleeding (i.e., prior to entry), whether due to varices or to gastritis, was
significantly higher in those patients with large varices. Also, the incidence of GI bleeding during
follow-up was significantly greater in patients with large varices. Variceal size had been measured
endoscopically in most patients with bleeding during follow-up, although in some cases the grading
was done radiographically. All patients who bled during follow-up had large varices. However, in some
cases recurrent bleeding was due to acute gastritis and in others to varices.
The study of Lebrec et al. has several problems.69(1925) The radiographic diagnosis of esophageal
varices is generally not as reliable as reported by these authors. Furthermore, the correlation in this
study between endoscopic and radiographic assessment of varices is far better than reported by
others.93,94(1926) The authors found that the risk of variceal bleeding was significantly higher if the
varices were large, but they also found that the risk of bleeding from gastric erosions was also
significantly increased if the varices were large. Lebrec et al.69(1927) speculated that this could be
explained by the general increase in collateral circulation through gastric and esophageal veins, which
would favor bleeding from established gastric erosions.
The endoscopic appearance of esophageal varices was analyzed according to the classification of the
Japanese Research Society for Portal Hypertension (see Table 312) in the retrospective study of
Beppu et al.34(1928) Data from 172 patients were analyzed in an effort to correlate a history of
bleeding with the endoscopic appearances of the varices according to the Japanese classification. For
varices with forms F1, F2, and F3 (see later), 15%, 31.8%, and 67.6% of patients, respectively, had a
history of variceal bleeding. These differences were statistically significant, although most F3 varices
also had red color signs. No significant difference was seen in bleeding in relation to location (linear
extent) of varices. Only 9.1% of patients without red color signs had a history of bleeding; 58.7% of
patients with red color signs had a history of bleeding, this difference being statistically significant. For
patients with both red wale markings and cherry-red spots, the severity of bleeding tended to be
greater, whereas all patients with hematocystic spots had a history of bleeding. Most patients with
diffuse redness had bled, but they also had other red color signs, and the presence of diffuse redness
alone did not correlate with bleeding. Of the patients with bluish varices, 79.4% had bled, versus 45.7%
of those with whitish varices, a significant difference. Beppu et al.34(1929) concluded that varices with
positive red color signs were more likely to bleed, especially those with higher grades of red wale
markings, cherry-red spots, or hematocystic spots. Patients with bluish varices are at greater risk for
bleeding irrespective of the presence or absenceor even the extentof red color signs. The linear
extent of the variceal columns did not correlate with bleeding, nor did the diameter of the varices.
Modified classification systems for esophageal varices have been developed for use in prospective
studies that are less complex than the Japanese system. The NIEC has formulated a prognostic index
that uses a combination of Child's classification, size of varices, and the absence or "degree of
presence" of red wale markings.46(1930) This classification was developed in a multicenter study of
321 patients with cirrhosis and no prior history of variceal hemorrhage who were followed for a median
of 23 months. Bleeding, largely variceal in origin, occurred in 26.5% of patients. Red color signs
(especially red wale markings as opposed to cherry-red spots), variceal size, and the degree of liver
dysfunction were found to be independent, additive risk factors for variceal bleeding. These
parameters were then used to predict the risk of bleeding using an index equation in another group of
75 cirrhotic patients and varices but no prior bleeding. The difference between the observed and the
predicted incidence of bleeding was not statistically significant. The NIEC study also demonstrated that
the prevalence of red wale markings increases with larger varices.
Rigo et al.95(1931) attempted to assess prospectively the predictive value of the Dagradi, Japanese,
and NIEC classifications for variceal hemorrhage in a series of 320 patients without prior bleeding. The
classification of Dagradi was considered to be most sensitive, although the Japanese and NIEC
classifications were more specific and had a higher predictive value.
Essentially, the NIEC index is derived from the Japanese classification of esophageal varices. These
data from the NIEC indicate that large variceal diameter and the presence of red wale markings are
especially important risk factors (in addition to Child's classification). The NIEC index has been
validated in further studies that show that the index provides accurate information regarding the risk of
bleeding.96,97(1932) Many centers are now using the NIEC index, which, although not ideal, is
sufficiently simple to be user friendly.
The Stomach in Portal Hypertension

The importance of the effects of portal hypertension on the stomach and gastric mucosa has been
recognized in recent years. In general, two major abnormalities may occur: gastric varices and portal
hypertensive gastropathy.
Gastric Varices
Gastric varices arise because of either generalized or segmental portal hypertension. The latter is due
to splenic vein thrombosis causing collateral venous channels to open via the short gastric veins and
the left gastric vein. This results in varices in the body and fundus of the stomach but usually not in the
esophagus. Gastric varices arising in patients with generalized portal hypertension are usually
accompanied by esophageal varices. Gastric varices bleed in about 60% of cases if left untreated.
Hosking and Johnson98(1933) classified gastric varices into three types. Type 1 varices appear as an
inferior extension of esophageal varices across the squamocolumnar junction. Sometimes termed
junctional varices, these are usually easy to treat by sclerotherapy. Type 2 are gastric varices, usually
accompanied by esophageal varices, located in the fundus; if they bleed, type 2 varices usually require
devascularization or shunt surgery. Type 3 are gastric varices located in the fundus or body of the
stomach in the absence of esophageal varices; these are usually treated by splenectomy. Gastric
varices were seen more frequently along the lesser curvature of the stomach than in the fundus in 104
patients with portal hypertension studied by Mathur et al.99(1934) using endoscopy and
splenoportography.
Hashizume et al.100(1935) proposed an endoscopic classification of gastric varices based on data
from a retrospective study. This classification uses three main parameters: form, location, and color.
With respect to form, varices may be tortuous and winding (F1), nodular (F2), or "huge" (F3). Five
areas are recognized for location, ranging from the anterior wall of the cardia through to fundus, lesser
curvature posterior wall, and greater curvature. The color of the mucosa overlying the gastric varices is
described and further subdivided into white and red varices, red spots, clots and erosions, and
hematocystic spots. This classification is not widely used because of its complexity. Cales et
al.88(1936) simply divide gastric varices into suspicious and definite categories.
Data on the reproducibility of endoscopic classification of gastric varices are very limited. The little
information available suggests that interobserver agreement may be very poor.101(1937) No
prospective studies on the classification of gastric varices have correlated endoscopic appearances
with the risk of bleeding. In the future, newer diagnostic modalities such as endoscopic
ultrasonography may prove to be useful in assessing gastric varices. 102,103(1938)
Portal Hypertensive Gastropathy

Although many terms have been used to describe the gastric lesions associated with portal
hypertensiongastric erosions, erosive gastritis, alcoholic gastritis, gastric mucosal ectasia, portal
vasculopathythe term preferred now is portal hypertensive gastropathy (PHG). The term congestive
gastropathy is also acceptable. Terminology that uses the word gastritis is misleading because
inflammatory changes are not a major feature in PHG. The unique feature is dilated mucosal and
submucosal blood vessels, usually in the absence of significant inflammation.
The endoscopic appearances of PHG have been classified as either mild or severe. Mild gastropathy
has been subdivided into categories as follows:104,105(1939) (1) fine speckling, (2) superficial
reddening, especially on the tips of the rugae, and (3) a fine white reticular pattern separating areas of
pink edematous mucosa, causing the so-called snake skin or mosaic appearance. The latter is now
thought to be the most common finding in portal hypertension. One prospective study found this
mosaic pattern in 94 of 100 patients with cirrhosis but only 1 of 300 control subjects and in none of 100
alcoholic patients without liver disease.106(1940) Lin et al.107(1941) found the mosaic appearance
associated with PHG to be significantly more common in 100 patients with cirrhosis than in 100 normal
control subjects.
Severe gastropathy is evidenced by the presence of discrete cherry-red spots arising in islands of
mucosa within the mosaic appearance described earlier. The red spots often progress to become
confluent areas of diffuse weeping mucosa with bleeding. These cherry-red spots in the stomach are
considered to be similar to those described in the esophagus. In the stomach, as in the esophagus,
their presence seems to indicate a high risk of bleeding. They represent ectatic vessels.108(1942)
At endoscopy, the gastric mucosal abnormalities of portal hypertension tend to be seen more often in
the fundus; this area should always be examined by retroflexing the endoscope.
Misra et al.109(1943) compared the endoscopic appearances of the stomach in 50 patients with portal
hypertension and 1323 control subjects. Evidence of mild gastritis, including a mosaic appearance,
was found more frequently in the former patients, although this latter finding was highly nonspecific
with respect to portal hypertension. Mucosal vascular congestion was noted in biopsies of the stomach
in most patients with portal hypertension, but similar changes were seen in control subjects and the
difference was not statistically significant. No correlation was found between the endoscopic and
histologic findings of PHG.
PHG has been noted at endoscopy in children with portal hypertension.110(1944)
Histology and Pathophysiology
In the normal stomach, 90% of blood flow goes to the mucosa, which indicates a very high blood flow
requirement for adequate function.111(1945) In cirrhosis, both in the experimental model and in
humans, there appears to be a significant reduction in gastric mucosal blood flow, particularly in the
antrum, as well as significant opening of arteriovenous shunts.112(1946) A relative ischemia of the
gastric mucosa therefore occurs in portal hypertension and is accompanied by increased submucosal
blood flow, changes that give the appearance of gastritis.
Histologically, the stomach in the patient with portal hypertension contains dilated veins in the
submucosa that are tortuous and irregular. In some areas, the intima is thickened. The morphology of
the stomach in patients with portal hypertension has been well studied by Hashizume et al.113(1947)
Using a silicone casting technique, these investigators demonstrated arteriovenous anastomoses with
decreased numbers of spiral arterioles and an increased number of straight arterioles. The
precapillaries, capillaries, submucosal veins, and subserosal veins were all dilated. They found many
focal areas of abnormal mucosal vessels, primarily capillary and venous ectasia. Very few
inflammatory cells were seen, but even when some were present the vascular changes were far more
pronounced than expected in patients with "gastritis."
McCormack et al.105(1948) studied full-thickness surgical (n = 7) and autopsy specimens (n = 11) of
the gastric wall as well as endoscopic biopsies in 14 patients. Tortuous, dilated submucosal veins were
demonstrated in the former specimens and "vascular ectasia" in the latter.
The reluctance for obvious reasons to obtain gastric biopsies in patients with portal hypertension has
limited the information available on the pathology of PHG. However, Quintero et al.114(1949)
described histologic findings in biopsy specimens obtained with a diathermy snare in 42 patients.
Morphometric quantitative differences were found between endoscopically mild and severe
gastropathy. The mean mucosal capillary cross-sectional area in antral biopsy specimens in patients
with gastric red spots was greater than in patients without red spots. In turn, this area was significantly
greater than that observed in noncirrhotic control subjects. Staining for Factor VIII-related antigen, a
specific marker for endothelial cells, may be of some benefit in the interpretation of histologic
specimens. Using this technique, Foster et al.115(1950) found mucosal capillary dilation to be
significantly greater in biopsy specimens from 23 patients with portal hypertension compared with
specimens from 25 patients with nonulcer dyspepsia.
Functional studies indicate that the gastric mucosa in PHG has a hyperdynamic blood flow. Guslandi et
al.116(1951) studied gastric mucosal blood flow by laser Doppler flowmetry at endoscopy in 15
patients with PHG, 15 patients with chronic gastritis, and 15 normal subjects. Blood flow was
significantly increased in patients with PHG and reduced in patients with chronic gastritis. Increased
mucosal blood flow has also been demonstrated by Sarfeh et al.117(1952)
Physiologic studies demonstrate that the gastric mucosa, in addition to being hypoxic, is also
functionally impaired.118(1953) In addition to increased mucosal blood flow, Sarfeh et al.117(1954)
have found reduced acid secretion and reduced levels of pepsinogen I in association with PHG.
However, Lin et al.107(1955) found no significant differences in serum gastrin level or serum
pepsinogen I levels in a study of cirrhotic patients with PHG (21 patients) and without PHG (25
patients). Vigneri et al.119(1956) found hypergastrinemia to be present in cirrhotic patients with and
without congestive gastropathy.
Sarfeh et al.120(1957) demonstrated impaired oxygenation of the gastric mucosa in patients with portal
hypertension and hypothesized that this factor is partially responsible for the increased susceptibility to
injury.
Although Helicobacter pylori may have a causal role in gastritis, initial studies have failed to show any
obvious causal relationship between Helicobacter infection and PHG.116,121,122(1958)
Risk of Bleeding
PHG is now well recognized in patients with cirrhosis and is thought to be responsible for about one
fifth of the episodes of bleeding associated with this condition. Many of the patients reported to have
so-called alcoholic gastritis may in fact have PHG.104(1959) This condition is now of increasing
importance because many of the modern methods of managing esophageal varices do not affect
portal pressure per se. Hence, after eradication of esophageal varices by either esophageal
transection or injection sclerotherapy, it is possible that redirection of the portal pressure toward the
stomach may precipitate gastric bleeding. Bleeding following esophageal transection may in fact arise
in many cases from sources within the stomach.28(1960)
The incidence of PHG, which is virtually never associated with dyspepsia, varies between 30 and 60%.
With increasing liver decompensation, the incidence of PHG seems to increase,123,124(1961)
although a direct relation to the functional status of the liver has not been found by all
investigators.105,119(1962) Mild gastropathy (see earlier) rarely seems to present any clinically
significant problem, whereas the severe type is associated with a very high incidence of GI
hemorrhage. More than 70% of patients who present with incidental gastric cherry-red spots bleed
during the next 2 years.125(1963)
Pathogenesis of Bleeding
The exact mechanism responsible for the development of PHG in either the mild or severe form is
essentially unknown. Furthermore, the precise precipitating factors that cause bleeding from the gastric
mucosa are again unknown, although bleeding appears to be associated with a number of factors.
An invariable association of PHG with portal hypertension is seen. Furthermore, PHG usually
disappears following a decompression shunt procedure. Although elevated portal venous pressure
must play a role in bleeding, it has no exact correlation with wedged hepatic vein pressure in patients
with or without gastropathy.105,107(1964) Therefore, gastric mucosal factors per se must play a role in
bleeding in addition to elevated portal venous pressure.
Groszmann and Colombato126(1965) have proposed that the dilated thin-walled vessels in the gastric
mucosa with an increased internal radius and possibly a thinner wall may be more prone to rupture in
response to stimuli that otherwise would not cause bleeding. Bleeding gastric mucosa in patients with
portal hypertension may be due to both explosive and erosive factors.126(1966)
Many experimental studies have demonstrated that the gastric mucosa of patients with portal
hypertension is more susceptible to injury than the mucosa of normal subjects. In the rat model, injury
easily occurs following the administration of bile acids, alcohol, and aspirin.127(1967) Patients with
cirrhosis are often prone to repeated episodes of endotoxemia, and Shibayama128(1968) has shown
in the rat model that intravenous administration of endotoxin causes bleeding. This may be a factor in
humans, although it does not explain mucosal bleeding in patients with extrahepatic portal
hypertension.
Repeated injection sclerotherapy could redirect portal pressure toward the stomach and eventually
precipitate bleeding from the gastric mucosa. However, data to confirm or refute this supposition are
limited.118(1969) One study by Kotzampassi et al.129(1970) offers some evidence that endoscopic
sclerotherapy may aggravate the changes of congestive gastropathy. Theoretically, a similar situation
could arise following esophageal transection.28(1971) Staple-line erosion is being documented
increasingly as a cause of recurrent bleeding following stapled esophageal transection.130(1972)
Sarin et al.124(1973) prospectively studied 107 patients with portal hypertension due to various
causes, including cirrhosis, noncirrhotic portal fibrosis, extrahepatic portal vein obstruction, and
Budd-Chiari syndrome. Prior to sclerotherapy, PHG was present in only 4 of the patients with cirrhosis.
After initiation of sclerotherapy, 21 additional patients developed gastropathy; most of these patients
had portal hypertension caused by cirrhosis. However, a correlation was also found between the
development of gastropathy and the severity of liver disease in this study. Tanoue et al.131(1974)
studied the effect of sclerotherapy in patients with PHG (n = 35) and without PHG (n = 102). In patients
with gastropathy, the endoscopic findings were scored according to severity into four grades. The
severity score was significantly worse after sclerotherapy, especially at 6 to 9 months after eradication
of the varices.
D'Amico et al.132(1975) found the risk of congestive gastropathy to be related to the presence of large
esophageal varices and sclerotherapy. In this prospective study of 212 cirrhotic patients followed up for
a mean of 46 months, congestive gastropathy was noted in 83% of the 75 patients who had
sclerotherapy, versus 59% of patients in whom sclerotherapy was not performed. At the outset of the
study, severe gastropathy (granular mucosa with red spots) was present in 20 patients, and mild
changes (mosaic pattern) were noted in 110 patients. Anemia, presumably from chronic blood loss,
and outright GI bleeding occurred more often in patients with severe gastropathy than in those with
mild or no gastropathy.
REFERENCES
1.
Olson RW, Hodgson JR, Adson MA. Significance of duodenal deformity in patients with
extrahepatic portal obstruction. Radiology 1963;80:63640.
2.
Richter RM, Pochaczevsky R. Duodenal varices. Arch Surg 1967;95:26973.
3.
fiber-duodenoscope. Endoscopy 1973;5:1014.
4.
Bloor K, Orr W. A case of haemorrhage from varices in the small intestine due to portal
hypertension. Br J Surg 1961;48:4234.
5. Saitoh M, Tsai HF, Futagawa S. Colonic varices in portal hypertension. J Gastroenterol
Hepatol 1989;4(suppl 1):28990.
6.
Watkins RM. Variceal haemorrhage from a colostomy due to portal hypertension secondary to
intrahepatic metastases from rectal carcinoma. BMJ 1981;282:18990.
7. Starzl TE, Demetris AJ, Van Thiel D. Liver transplantation (1). N Engl J Med
1989;321:101422.
8. Starzl TE, Demetris AJ, Van Thiel D. Liver transplantation (2). N Engl J Med
1989;321:10929.
9.
Kegaries DL. The venous plexus of the oesophagusIts clinical significance. Surg Gynecol
Obstet 1934;58:4651.
10.
Butler H. The veins of the oesophagus. Thorax 1951;6:27696.
11.
MacBeth R. Treatment of oesophageal varices in portal hypertension by means of sclerosing
injections. BMJ 1955;2:87780.
12.
Beswick TSL, Butler H. Fatal haematemesis from oesophageal varices in presence of large
portal-caval anastomosis. BMJ 1951;2:5225.
13.
DeCarvalho CAF. Sur l'angio-architecture veineuse de la zone de transition
oesophago-gastrique et son interpretation fonctionelle. Acta Anat (Basel) 1966;64:12562.
14.
Stelzner F, Lierse W. Der Angiomuskulare Dehnverschlub der terminalen Speisohre.
Langenbecks Arch Klin Chir 1968;321:3564.
15.
Spence RAJ. Venous drainage of the oesophagus. Gullet 1991;1:1635.
16. Spence RAJ. The venous anatomy of the lower oesophagus in normal subjects and in
patients with varicesAn image analysis study. Br J Surg 1984;71:73944.

17. McCormack TT, Smith PM, Rose JD, Johnson AG. Perforating veins and blood flow in
oesophageal varices. Lancet 1983;2:14424.
18.
Liebermann-Meffert D, Heberer M, Martinoli S, Allgower M. Are there muscular structures
which may contribute to closure of the gastro-oesophageal junction?. Scand J Gastroenterol
1981;67:123.
19. Fujimura I, Fereaz de Carvalho CA. Venous angioarchitecture of the tunica muscularis in the
eosophagogastric transition segment in man. Acta Anat (Basel) 1977;98:40112.
20.
Noda T. Angioarchitectural study of esophageal varices with special reference to variceal
rupture. Virchows Arch Pathol Anat Histopathol 1984;404:38191.
21. Kitano S, Terblanche J, Kahn D, Bornman PC. Venous anatomy of the lower oesophagus in
portal hypertension: Practical implications. Br J Surg 1986;73:52531.
22. Spence RAJ, Sloan JM, Johnston GW, Greenfield A. Oesophageal mucosal changes in
patients with varices. Gut 1983;24:10249.
23. Spence RAJ, Sloan JM, Johnston GW. Histological factors of the oesophageal transection
ring as clues to the pathogenesis of bleeding varices. Surg Gynecol Obstet 1984;159:2539.
24. Paquet KJ. Endoscopic paravariceal injection sclerotherapy of the oesophagusIndications,
technique, complications: Results of a period of 14 years. Gastrointest Endosc
1983;29:3105.
25. Spence RAJ, Terblanche J. Venous anatomy of the lower oesophagus: A new perspective on
varices. Br J Surg 1987;74:65960.
26. Vianna A, Hayes PC, Moscoso G, et al. Normal venous circulation of the gastroesophageal
junction. Gastroenterology 1987;93:8769.
27. Spence RAJ, Johnston GW. Stapled oesophageal transection for varices. Results in 100
consecutive patients. Surg Gynecol Obstet 1985;160:3239.
28.
Spence RAJ. Prevention of recurrent bleedingNon-shunt surgery. In Okuda K, Benhamou
JP (eds). Portal Hypertension. Clinical and Physiological Aspects. Tokyo, New York:
29. Spence RAJ, Sloan JM, Johnston GW. Oesophagitis in patients undergoing oesophageal
transection for varicesA histological study. Br J Surg 1983;70:3324.
30. Jaffe EA. Synthesis of Factor VIII-related antigen by cultured human endothelial cells. Ann N
Y Acad Sci 1975;240:629.
31. Burnaud KG, Clemenson G, Whimister I, et al. The effect of sustained venous hypertension
on the skin capillaries to the canine hind limb. Br J Surg 1982;69:414.
32.
Dodd H, Cockett FB. The Pathology and Surgery of the Veins of the Lower Limb. Edinburgh:
Churchill Livingstone, 1956;362.
33. Iuokuchi K. The general rules for recording endoscopic findings on oesophageal varices. Jpn
J Surg 1980;10:847.
34. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by esophageal
35. Hashizume M, Kitano S, Sugimachi K, Sueishi K. Three dimension view of the vascular
structure of the lower oesophagus in clinical portal hypertension. Hepatology 1988;8:14827.
36. Kitano S, Koyanagi N, Iso Y, et al. Prevention of recurrence of esophageal varices after
endoscopic injection sclerotherapy with ethanolamine oleate. Hepatology 1987;7:8105.
37.
Arakawa M, Kage M. The anatomy and pathomorphology of esophageal varices. In Okuda K,
Benhamou J-P (eds). Portal Hypertension. Clinical and Physiological Aspects. Tokyo, New
York: Springer-Verlag, 1991;41528.
38.
Chiles NH, Baggenstoss AH, Butt HR, Olsen AM. Oesophageal varices: Comparative
incidence of ulceration and spontaneous rupture as a cause of fatal haemorrhage.
39.
Orloff MJ, Thomas HS. Pathogenesis of oesophageal varix rupture. A study based on gross
and microscopic examination of the oesophagus at the time of bleeding. Arch Surg
1963;87:3017.
40. Dagradi AE, Arguello JF, Weingarten ZG. Failure of endoscopy to establish a source for
upper gastrointestinal bleeding. Am J Gastroenterol 1979;72:395402.
41.
Warren WD, Delaney JP. Surgery of the Liver, Pancreas and Biliary Tract. New York: Stratton
Intercontinental Medical Book Corp. 1975;63742.
42. Paquet K-J, Oberhammer E. Sclerotherapy of bleeding oesophageal varices by means of
43. Terblanche J. Treatment of esophageal varices by injection sclerotherapy. Adv Surg
1981;15:25791.
44.
Spence RAJ. Oesophageal transection for varices: Rationale, indications, technique and
results. In Idezuki Y (ed). Treatment of Oesophageal Varices. Amsterdam: Excerpta Medica
1988;12339.
45. Olsson R. The natural history of esophageal varices. A retrospective study of 224 cases with
liver cirrhosis. Digestion 1972;6:6574.
46. The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices.
esophageal varices. N Engl J Med 1988;319:9839.
47. Paquet K-J. Prophylactic endoscopic sclerosing treatment of the esophageal wall in
varicesA prospective controlled randomized trial. Endoscopy 1982;14:45.
48. Silverstein FE, Gilbert DA, Tedesco FJ, et al. The national ASGE survey on upper
gastrointestinal bleeding. II. Clinical prognostic factors. Gastrointest Endosc 1981;27:8093.
49. Graham DY, Smith JL. The course of patients after variceal hemorrhage. Gastroenterology
1981;80:8009.
50.
Jackson FC, Perrin EB, Degradi AE, et al. Clinical investigation of the portacaval shunt. I.
Study design and preliminary survival analysis. Arch Surg 1965;91:4354.
51. Christensen E, Fauerholdt L, Schlichting P, et al. Aspects of the natural history of
gastrointestinal bleeding in cirrhosis and the effect of prednisone. Gastroenterology
1981;81:94452.
52.
Wagenknecht TW, Noble JF, Barnofsky ID. Nature of bleeding in esophageal varices.
Surgery 1953;33:86974.
53. Ponce J, Froufe A, de la Morena E, et al. Morphometric study of the esophageal mucosa in
cirrhotic patients with variceal bleeding. Hepatology 1981;1:6416.
54. Eckardt VF, Grace ND. Gastroesophageal reflux and bleeding esophageal varices.
55.
MacMathuna P. The pathogenesis of variceal rupture in gastrointestinal endoscopy.
56. Garcia-Tsao G, Groszmann RJ, Fisher RL, et al. Portal pressure, presence of
gastro-oesophageal varices and variceal bleeding. Hepatology 1985;5:41924.
57. Mahl TC, Groszmann RJ. Pathophysiology of portal hypertension and variceal bleeding in
management of variceal haemorrhage. Surg Clin North Am 1990;70:25166.
58. Boyer JL, Chatterjee C, Iber FL, et al. Effect of plasma volume expansion on portal
hypertension. N Engl J Med 1966;275:7505.
59.
Reding P, Urbain D, Grivegnee A, et al. Portal venous esophageal luminal pressure gradient
in cirrhosis. Hepatology 1986;6:9801000.
60. de Dombal FT, Clarke JR, Clamp SE, et al. Prognostic factors in upper G.I. bleeding.
61. Rigau J, Bosch J, Bordas JM, et al. Endoscopic measurement of variceal pressure in
1989;96:87380.
62.
Ready JB, Rector WG. Stratification of risk of continued bleeding and early rebleeding from
oesophageal varices by measurement of portal pressure within 24 hours of index
haemorrhage. Gastroenterology 1989;96:A649.
63.
Vlavianas P, Gimson A, MacMathuna P, et al. The prognostic significance of systemic and
portal haemodynamic parameters in patients with cirrhosis and previous variceal bleeding.
Gut 1990;31:A592.
64. Thornton JR, Dean HG, Losowsky MS. Do increased catecholamines and plasma methionine
enkephalin in cirrhosis promote bleeding oesophageal varices. Q J Med 1988;68:54151.
65. Kleber G, Sauerbruch T, Fisher G, et al. Pressure of intraoesphageal varices assessed by
fine needle puncture: Its relation to endoscopic signs and severity of liver disease in patients
with cirrhosis. Gut 1989;30:22832.
66. Beppu K, Hashizume M. Endoscopic markers of impending variceal hemorrhage: Correlation
with the pressure in the varices. Hepatology 1990;11:1478.
67. Bosch J, Mastai R, Kravetz D, et al. Measurement of azygos venous blood flow in the
evaluation of portal hypertension in patients with cirrhosis. Clinical and haemodynamic
correlation in 100 patients. J Hepatol 1985;1:12539.
68. Kleber G, Sauerbruch T, Ansari H, et al. Prediction of variceal hemorrhage in cirrhosis. A
prospective followup study. Gastroenterology 1991;100:13327.
69. Lebrec D, De Fleury P, Rueff B, et al. Portal hypertension, size of esophageal varices, and
risk of gastrointestinal bleeding in alcoholic cirrhosis. Gastroenterology 1980;79:113944.
70. Polio J, Groszmann RJ. Haemodynamic factors involved in the development and rupture of
oesophageal varices. A pathophysiological approach to treatment. Semin Liver Dis
1986;6:31831.
71. Chung RS, Lewis JW. The platelet fibrin plug in oesophageal variceal haemorrhage. The
Mount St. Helen's sign. Gastrointest Endosc 1984;30:270.
72. Siringo S, McCormick PA, Mistry P, et al. Prognostic significance of the white nipple sign in
variceal bleeding. Gastrointest Endosc 1991;37:515.
73. Nakaniski P, Nishimura Y, Futagawa S. Angiographic study of haemodynamics in portal
hypertension. J Gastroenterol Hepatol 1989;4(suppl 1):2424.
74. Hashizume M, Kitano S, Yamaga H, et al. Vascular architecture of the lower oesophagus in
portal hypertension. J Gastroenterol Hepatol 1989;4(suppl 1):2013.
75. Siringo S, Bolondi L, Gaiani S, et al. The relationship of endoscopy, portal Doppler ultrasound
flowmetry, and clinical and biochemical tests in cirrhosis. J Hepatol 1994;20:118.
76.
Westaby D. The management of an episode of variceal bleeding. In Okuda K, Benhamou J-P
(eds). Portal Hypertension. Clinical and Physiological Aspects. Tokyo, New York:
77. Cotton PB. Severe upper gastrointestinal bleeding. Part IV: Problems for emergency
endoscopic treatment. Clin Gastroenterol 1981;10:5964.
78. Dave P, Romeu J, Messer J. Upper gastrointestinal bleeding in patients with portal
hypertension: A reappraisal. J Clin Gastroenterol 1983;5:1135.
79. Mitchell CJ, Jewell DP. The diagnosis of the site of upper gastrointestinal haemorrhage in
patients with established portal hypertension. Endoscopy 1977;9:1315.
80. Mitchell KJ, MacDougall BR, Silk DB, Williams R. A prospective reappraisal of emergency
endoscopy in patients with portal hypertension. Scand J Gastroenterol 1982;17:9658.
81. Sutton FM. Upper gastrointestinal bleeding in patients with esophageal varices. What is the
most common source? Am J Med 1987;83:2735.
82. Tabibian N, Graham DY. Source of upper gastrointestinal bleeding in patients with
esophageal varices seen at endoscopy. J Clin Gastroenterol 1987;9:27982.
83. Buccino RV, Bogliolo G, Ferrara M, et al. Endoscopic approach to patients with portal
hypertension: A complex diagnosis. A retrospective study based on 10 years' experience.
84. Dagradi AE. The natural history of esophageal varices in patients with alcoholic liver cirrhosis.
An endoscopic and clinical study. Am J Gastroenterol 1972;57:52040.
85.
Palmer ED, Brick IB. Correlation between the severity of esophageal varices in portal cirrhosis
and their propensity toward hemorrhage. Gastroenterology 1956;30:8590.
86.
Baker LA, Smith C, Lieberman G. The natural history of esophageal varices: A study of 115
cirrhotic patients in whom varices were diagnosed prior to bleeding. Am J Med
1959;26:22837.
87. The Italian Liver Cirrhosis Project. Reliability of endoscopy in the assessment of variceal
features. J Hepatol 1987;4:938.
88. Cales P, Zabotto P, Meskens C, et al. Gastroesophageal endoscopic features in cirrhosis:
Observer variability, interassociations and relationship to hepatic dysfunction.
89.
de Franchis R, Primignani M, Cosentino F. Endoscopic diagnosis and classification of varices
in variceal bleeding. Gastrointest Endosc Clin North Am 1992;2:1730.
90. Bendtsen F, Skovgaard LT, Sorensen TIA, Matzen P. Agreement among multiple observers
on endoscopic diagnosis of esophageal varices before bleeding. Hepatology 1990;11:3417.
91.
Kumagai P, Tsuzuki T, Makuuchi H. Endoscopic classification of esophageal varices and its
clinical significance. Stomach Intestine (Jpn) 1976;11:74150.
92.
Arakawa M, Noda T, Fukuda K, et al. Clinico-pathological studies of esophageal
varicesHistological findings of ruptured esophageal varices and its relationship to
endoscopical findings. Gastroenterol Endosc 1985;27:1919.
93. Conn HO. The varix volcano connection. Gastroenterology 1980;79:13336.
94.
Conn HO, Mitchell JR, Brodoff MG. A comparison of the radiologic and esophagoscopic
diagnosis of esophageal varices. N Engl J Med 1961;265:1604.
95. Rigo GP, Merighi A, Chahin NJ, et al. A prospective study of the ability of three endoscopic
classifications to predict hemorrhage from esophageal varices. Gastrointest Endosc
1992;38:4259.
96.
Piai G, Minieri M, Catalano M, et al. Prospective validation of two incidences of first variceal
bleeding. Gastroenterology 1991;100:A784.
97.
Prada A, Bortoli A, Minoli G, et al. The prediction of esophageal variceal bleeding. NIEC score
validation by an independent group. Gastroenterology 1990;98:A109.
98. Hosking SW, Johnson AG. Gastric varices: A proposed classification to management. Br J
Surg 1988;75:1956.
99. Mathur SK, Dalvi AN, Someshwar V, et al. Endoscopic and radiological appraisal of gastric
100. Hashizume M, Kitano S, Yamaga H, et al. Endoscopic classification of gastric varices.
101.
Theodossi A, Westaby D, Spiegel Halter JL, et al. Observer variation in the assessment of
oesophageal and gastric varices [Abstract]. Gut 1984;25:A577.
102. Ziegler K, Gregor M, Zeitz M, et al. Evaluation of endosonography in sclerotherapy of
esophageal varices. Endoscopy 1991;23:2845.
103. Caletti GC, Brocchi E, Ferrari A, et al. Value of endoscopic ultrasonography in the
management of portal hypertension. Endoscopy 1992;24(suppl 1):3426.
104.
Triger DR. Portal hypertensive gastropathy. In Okuda K, Benhamou J-P (eds). Portal
Hypertension. Clinical and Physiological Aspects. Tokyo, New York: Springer-Verlag,
1991;44152.
105. McCormack TT, Simms J, Eyre-Brook I, et al. Gastric lesions in portal hypertension:
Inflammatory gastritis or congestive gastropathy?. Gut 1985;26:122632.
106. Papazian A, Braillon A, Dupas JL, et al. Portal hypertensive gastric mucosa. An endoscopic
study. Gut 1986;27:11991203.
107. Lin WJ, Lee FY, Lin HC, et al. Snake skin pattern gastropathy in cirrhotic patients. J
Gastroenterol Hepatol 1991;6:1459.
108. Van Vliet ACM, Ten Kate FJV, Dees J, Van Blankenstein M. Abnormal blood vessels of the
prepyloric antrum in cirrhosis of the liver as a cause of chronic gastrointestinal bleeding.
Endoscopy 1978;10:8994.
109. Misra SP, Dwivedi M, Misra V, et al. Endoscopic and histologic appearance of the gastric
mucosa in patients with portal hypertension. Gastrointest Endosc 1990;36:5759.
110. Hyams JS, Treem WR. Portal hypertensive gastropathy in children. J Pediatr Gastroenterol
Nutr 1993;17:138.
111. Landgren O. Gastric mucosal haemodynamics. Gastroenterol Clin Biol 1985;9:269.
112. Manabe T, Suzuki T, Honjo I. Changes of gastric mucosal blood flow in experimentally
induced cirrhosis of the liver. Surg Gynecol Obstet 1978;147:7537.
113. Hashizume M, Tawaka M, Inokuchi K. Morphology of gastric micro-circulation in cirrhosis.
Hepatology 1983;3:100816.
114. Quintero E, Pique JM, Bombi JA, et al. Gastric mucosal vascular ectasias causing bleeding in
cirrhosis. A distinct entity associated with hypergastrinemia and low serum levels of
pepsinogen I. Gastroenterology 1987;93:105461.
115. Foster PN, Wyatt JI, Bullimore DW, Losowsky MS. Gastric mucosa in patients with portal
hypertension: Prevalence of capillary dilatation and Campylobacter pylori. J Clin Pathol
1989;42:91921.
116. Guslandi M, Sorghi M, Foppa L, Tittobello A. Congestive gastropathy versus chronic gastritis:
A comparison of some pathophysiological aspects. Digestion 1993;54:1602.
117. Sarfeh IJ, Tarnawski A, Malki A, et al. Portal hypertension and gastric mucosal injury in rats.
Effects of alcohol. Gastroenterology 1983;84:98793.
118.
Hosking SW. Portal hypertensive gastropathy in gastrointestinal endoscopy in variceal
bleeding. Gastrointest Endosc Clin North Am 1992;2:11120.
119. Vigneri S, Termini R, Piraino A, et al. The stomach in liver cirrhosis. Endoscopic,
morphological, and clinical correlations. Gastroenterology 1991;101:4728.
120. Sarfeh IJ, Soliman H, Waxman K, et al. Impaired oxygenation of gastric mucosa in portal
hypertension. The basis for increased susceptibility to injury. Dig Dis Sci 1989;34:2258.
121. D'Amico G, Traina M, Montalbano L, et al. Congestive gastropathy. Prevalence, incidence and
morphologic features [Abstract]. J Hepatol 1989;9(suppl 1):141.
122. Parikh SS, Desai SB, Prabhu SR, et al. Congestive gastropathy: Factors influencing
development, endoscopic features, Helicobacter pylori infection, and microvessel changes.
123. Sacchetti C, Capello M, Rebecchi P, et al. Frequency of upper gastrointestinal lesions in
patients with liver cirrhosis. Dig Dis Sci 1988;33:121822.
124. Sarin SK, Sreenivas DV, Lahoti D, Saraya A. Factors influencing development of portal
hypertensive gastropathy in patients with portal hypertension. Gastroenterology
1992;102:9949.
125. Triger DR, Hosking SW. The gastric mucosa in portal hypertension. J Hepatol 1989;8:26772.
126. Groszmann RJ, Colombato LA. Gastric vascular changes in portal hypertension. Hepatology
1988;8:170810.
127. Sarfeh IJ, Tarnawski A, Hajduczek A, et al. The portal hypertensive gastric mucosa.
Histologic, ultrastructural and functional analysis after aspirin induced damage. Surgery
1988;104:7985.
128. Shibayama Y. An experimental study into the cause of acute haemorrhagic gastritis in
cirrhosis. J Pathol 1986;149:30713.
129. Kotzampassi K, Eleftheriadis E, Aletras H. The "mosaic-like" pattern of portal hypertensive
gastric mucosa after variceal eradication by sclerotherapy. J Gastroenterol Hepatol
1990;5:65963.
130. Kage GL, McCormick A, Siringo S, et al. Staple-line erosionA common source of recurrent
bleeding following stapled oesophageal transection. Br J Surg 1991;78:13557.
131. Tanoue K, Hashizume M, Wada H, et al. Effects of endoscopic injection sclerotherapy on
portal hypertensive gastropathy: A prospective study. Gastrointest Endosc 1992;38:5825.
132. D'Amico G, Montalbano L, Triana M. Natural history of congestive gastropathy in cirrhosis.
Chapter 32 Technique of Endoscopic Sclerotherapy

(1976)
(1977)
JACOB KORULA, M.D.
Endoscopic variceal sclerotherapy has evolved since the mid-1980s to become the definitive therapy
for management of variceal hemorrhage.1(1978) During this period, the technique of variceal
sclerotherapy also evolved based on clinical and experimental studies that balanced its effects on
recurrent bleeding and related mortality against complications and morbidity. This chapter evaluates
published information relating to technique of variceal sclerotherapy and summarizes current practice.
Equipment
Endoscopes
Worldwide, most gastroenterologists perform variceal sclerotherapy using flexible fiberoptic
endoscopes, although flexible video endoscopes are gaining in popularity. Sclerotherapy is performed
with the patient under sedative analgesia (conscious sedation). In endoscopy units in the United
States, pulse oximetry and frequently electrocardiographic and blood pressure monitoring are usually
required. In some surgical units, however, the rigid esophagoscope is used and sclerotherapy is
performed with the patient under general anesthesia.2(1979) Both single-channel and two-channel
flexible endoscopes are used, particularly the latter when active bleeding occurs; however, suctioning
of blood and clots eventually clogs the accessory channel. We prefer a single-channel endoscope and
wash the visual field during the examination with minimum use of suction. This method should suffice
in most cases of active bleeding and will allow the facile performance of sclerotherapy.
Needle Injectors
A number of needle injectors are commercially available for variceal sclerotherapy (Table 321). Most
are disposable and have minor differences and advantages. Ultimately, the choice of injector is
predicated on cost. Needles are available in 23- and 25-gauge and in lengths of 4 or 5 mm. The use of
so-called long needles (i.e., 7 mm) is discouraged because depth of injection cannot be controlled; in
one study, the incidence of bacteremia was increased when long needles were used.3(1980)
TABLE 321
Selected Examples of Sclerotherapy Needle
Injectors
INJECTOR
MODEL/TYPE
SIZE
Mill-Rose
Hobbs
Clear
Clear
5.5 mm, 25 g
4.0 mm, 25 g
Microvasive
Olympus
Contrast (black)
Variject
NM 5L-15L
4.0 mm, 25 g
4.0 mm, 25 g
46 mm, 21, 23 g
Reusable
Autoclavable
Bard
NM 13K
NM 5L-9L
Clear
46 mm, 21, 23 g
46 mm, 21, 23 g
46 mm, 25 g
High-contrast
Wilson-Cook
Yellow tip
Contrast (blue)
46 mm, 25 g
4.0 mm, 23 g
MANUFACTURER
Mill-Rose, Mentor, OH
Hobbs Medical, Stafford
Springs, CT
Microvasive, Natick, MA
Olympus America, Inc.,
Melville, NY
Bard Interventional,
Billerica, MA
Wilson-Cook,
Winston-Salem, NC
The injector needle offered by Mill-Rose (Mentor, Ohio) is not as rigid as those manufactured by
Microvasive (Natick, Massachusetts) or Hobbs Medical (Stafford Springs, Connecticut). Also, the
catheter has an inner metal tip at the distal end that prevents shearing of the needle as it is advanced
or withdrawn. The hub mechanism is cumbersome. The catheter is transparent, like the Hobbs
injector, and this allows visualization of blood aspirated into the injector.
The injector available from Microvasive has a black outer sheath that makes it difficult to visualize
when there is brisk bleeding; a transparent catheter is now available (Variject). Problems with shearing
of the catheter while the needle is advanced that occurred with earlier models seem to have been
overcome by the addition of a metal tip to the catheter. In addition, the stop handle allows better control
when the needle is advanced.
Newer models of flexible injector needles offered by Olympus America (Melville, New York) replace the
NMK series and are very good, although expensive. Sheaths that may be autoclaved are available in
addition to disposable needle injectors.
Needles manufactured by Bard, Inc. (Billerica, Massachusetts), are transparent and have a distal metal
inner hub to prevent shearing; catheters are available with a yellow tip that enhances visibility. The
handle has a thumb ring and locking mechanism. The Bard Hemoject needle has two channels, one
for the needle and the other for washing. However, we have found this injector to be more useful for
injection of bleeding ulcers than for sclerotherapy of esophageal varices.
Needles manufactured by Wilson-Cook, Inc. (Winston-Salem, North Carolina), are similar and have no
specific advantages.
Sclerotherapy of Esophageal Varices

Sclerotherapy is usually performed on an outpatient basis in patients who are stable and
ambulatory.4(1981) In patients with active bleeding, special measures are required as outlined in the
section on Sclerotherapy During Active Hemorrhage.
Technique of Injection
The two basic techniques of injection are intravariceal and paravariceal. The objective of intravariceal
injections is to initiate thrombosis and necrotizing inflammation within varices.5(1982) The purpose of
paravariceal injections is to generate fibrosis of the esophageal mucosa overlying the tortuous
esophageal varices, thus preventing rupture of the thin-walled vessels.6,7(1983)
Intravariceal Versus Paravariceal Injection
Intravariceal Injection
When, at endoscopy, esophageal varices are deemed the cause of hemorrhage, a needle injector is
inserted through the accessory channel of a single-channel endoscope or one of the channels of a
two-channel endoscope. The injector is advanced until the distal tip is within the endoscopic field of
view. The endoscope is maneuvered, so that the injector is directly opposite the variceal column to be
injected. The needle is then advanced within the injector sheath, so that its tip is visible prior to
insertion into the crest of a variceal column. The direction of approach is usually perpendicular to the
varix or at a 45-degree angle (Figures 321 and 322). Some endoscopists prefer to insert the needle
into the varix by first indenting the variceal column with the injector sheath and then advancing the
needle. Since the needle tip is not visualized, the disadvantage of this method is that the needle may
not advance properly within the sheath, in which case the endoscopist may not be aware that injection
of the varix has not occurred. Some endoscopists insert the needle into the varix with a sharp thrust.
This has the disadvantage that an excessively deep injection may result.
(1984)Figure 321. Intravariceal injection of a varix at the gastroesophageal junction. (From

the collection of Dr. M. V. Sivak, Jr.)
(1985)Figure 322. Intravariceal injection of an esophageal varix. The injection needle is

placed about 2 cm beyond the end of the endoscope. In this position, any sudden motion is
absorbed by the injector's flexibility, and tearing of the varix is more easily avoided. (From the
collection of Dr. M. V. Sivak, Jr.)
There are several ways to determine whether an injection of sclerosant is actually within a varix.
Usually, intravariceal injection can be assumed if there is no resistance to injection, bleb formation, or
blanching of the variceal wall (Figures 323 and 324). Formation of a bleb or submucosal swelling is
indicative of a paravariceal injection (Figure 325). If this is not intended, injection of the sclerosant
should be halted. Methylene blue can also be mixed with the sclerosant, in which case the variceal
column proximal to the puncture site is stained blue as the sclerosant is injected. In a prospective
study, Waring et al.8(1986) did not find the addition of methylene blue to be helpful.
(1987)Figure 323. Intravariceal injection of a large varix near the gastroesophageal junction.
Note the characteristic grayish color change. (From the collection of Dr. M. V. Sivak, Jr.)
(1988)Figure 324. A, Swollen distal esophageal varices after injection. B, Appearance of

midesophageal varices a few moments after injection. Note distention and change to a
dirty-white color. (A and B, From the collection of Dr. M . V. Sivak, Jr.)
(1989)Figure 325. Esophageal bleb raised as a result of paravariceal injection, in this case
inadvertently. (From the collection of Dr. M. V. Sivak, Jr.)
The volume and the type of sclerosant injected intravariceally usually depend on the technique and
preferences of the endoscopist (see later in this chapter). In reports of sclerotherapy, the volumes
injected into each puncture site range between 0.5 and 3.0 ml,9,10(1990) 5.0 ml in some
studies,11,12(1991) and rarely as much as 10 ml.13(1992) We inject 0.5 to 1.0 ml of sclerosant into
each injection site.
Each variceal column is injected in a similar manner, taking care to rotate or torque the insertion tube
of the endoscope, so that the variceal column to be injected is directly across from the needle injector.
Injections are usually begun at the gastroesophageal junction working in a circumferential pattern and
proceeding in a cephalad direction. The same variceal column can be injected proximally, provided
large volumes are not injected distally near the gastroesophageal junction. Only varices in the distal
third of the esophagus need be injected, as sclerosant is expected to flow in a cephalad direction in
most instances. Infrequently, varices in the stomach and near the gastroesophageal junction, termed
junctional gastric varices, will be sclerosed if sclerosant flows retrograde during injection of esophageal
varices.5(1993) Of interest, however, is a study by Barsoum et al.14(1994) in which 15 patients with
schistosomiasis underwent sclerotherapy using a mixture of 5% ethanolamine and a contrast agent.
During injections of esophageal varices, fluoroscopic monitoring disclosed that the sclerosant-contrast
solution did not enter the gastric varices. Junctional varices can be injected by means of either an
antegrade or a retrograde (retroflexed) approach.
After the initial procedure, sclerotherapy is performed at intervals of 1 week unless there is early
recurrence of bleeding.
Paravariceal Injection
This technique, introduced by Paquet,15(1995) involves the submucosal injection of a small volume
(0.5 ml) of sclerosant next to variceal columns beginning at the gastroesophageal junction and
concluding in the middle third of the esophagus. Volumes of 30 to 40 ml are usually injected at each
session.
It is not clear that the paravariceal injection technique has any distinct advantages. One rationale for its
use is that collateral flow is maintained within the esophagus. Obliteration of esophageal varices and
thereby loss of collateral flow may lead to formation of varices in other anatomic locations such as the
stomach, duodenum, and colon. However, this appears to be only a theoretical risk because varices in
other sites are not observed with any frequency in studies in which injections are predominantly
intravariceal.
Sarin et al.16(1996) compared paravariceal with intravariceal sclerotherapy in a randomized study of
54 consecutive patients with esophageal variceal bleeding. Sclerotherapy was performed every 3
weeks. Intravariceal injection technique proved to be significantly more effective for control of active
bleeding (91% compared with 16.7% for paravariceal injections). The mean time to obliteration of
varices was approximately 2 weeks in patients undergoing intravariceal injectionsthat is, about twice
as fast as eradication in the paravariceal injection group. Also, fewer treatment sessions were required
for eradication when the varices were injected directly. Recurrent bleeding was significantly more
frequent in patients who underwent paravariceal injections (38.5% vs. 14.3% for the intravariceal
treatment group). The complication rate was the same for both groups, except that retrosternal pain
was noted with greater frequency in patients who underwent paravariceal injections. Variceal
recurrence was the only parameter for which paravariceal injection proved to be superior: 25% of
patients who underwent intravariceal injections experienced this complication compared with 3.9% of
those in the paravariceal treatment group.
It is important to recognize that even when the intravariceal injection technique is precise, sclerosant
can be found in the lamina propria,5(1997) that is, many intravariceal injections have a paravariceal
component even if this was not intended. Furthermore, it is reasonable to assume that some
sclerosant appears within a varix during paravariceal injections.
Use of the paravariceal injection method varies widely throughout the world. Although it is not popular
in the United States, the paravariceal injection technique is utilized in parts of Europe. However, the
number of studies in which the paravariceal injection technique is employed has decreased since the
early 1980s. This probably indicates a preference for the intravariceal method and attests to its clinical
utility.
Combined Intravariceal and Paravariceal Injection
Both intravariceal and paravariceal injections can be performed at endoscopy as a combined injection
technique. Although the combined technique is preferred in some centers,17(1998) no controlled
studies of the specific benefits of this approach are available.
Kitano et al.18(1999) reported their experience with combined therapy in an uncontrolled study that
included 155 patients. Forty patients were treated during active bleeding, 49 underwent variceal
sclerotherapy electively, and 66 were offered prophylactic therapy. A transparent overtube was used
during the first and second sessions to perform intravariceal injections of 5% ethanolamine oleate.
Once thrombosis was achieved, submucosal injections were made with the same sclerosant to create
superficial ulcers. These healed with epithelialization and led to the "removal" of the submucosa
including the varices. An average of four sessions over a mean duration of 5 weeks were needed to
obtain the desired effect. Mean follow-up for the group was 14 months. Bleeding did not occur after the
first 2 months. Two patients died as the result of bleeding, in both cases from gastric varices. Although
these results are impressive, similar results have not been reported from other centers.
Freehand Versus Overtube Injection Technique
Williams and Dawson19(2000) first described the use of an overtube (Williams overtube, KeyMed,
Southend-on-Sea, Essex, UK) to achieve sclerotherapy in the 1970s. This tubular device, also known
as a sheath, is designed to slip over the rigid or flexible endoscope and is advanced into the
esophagus after intubation with the endoscope. A slot in the distal aspect of the overtube allows a varix
to protrude into the lumen. A single varix is therefore isolated within the overtube, a stable condition
that is thought to facilitate injection. After injection of the varix in the slot, the overtube is rotated to
allow an adjacent varix to protrude into the lumen for injection. This process is continued until all
varices are injected. The advantage of the overtube is the apparent ease with which injections can be
made during active bleeding.
Kitano et al.20(2001) recommended the use of a transparent overtube (unlike the Williams overtube,
which is black) that was said to improve visibility and facilitate variceal injections even during active
bleeding. In a controlled study of sclerotherapy using 5% ethanolamine oleate in 102 patients who were
randomly assigned to so-called freehand injections or injections made with the aid of an overtube,
Kitano et al.21(2002) found that use of the latter technique was associated with a reduction in the rate
of variceal recurrence. After variceal eradication, 100 patients were followed (mean duration, 30.8
months) at intervals of 1 month. At 1 month, scarring and ulceration were found in the distal esophagus
in 16 of 50 (32%) and 42 of 50 (84%) patients in the overtube and freehand injection groups,
respectively. Variceal recurrence was noted in 5 patients (10%) in the overtube group versus 14
patients (28%) in the group who underwent freehand injections. In the latter group, 2 patients had
recurrent bleeding.
Westaby et al.22(2003) also reported the results of a randomized, controlled study of the overtube and
freehand injection techniques in 40 patients. For the group treated using the overtube, there was
significantly less bleeding within 24 hr after the injection procedure, significantly fewer episodes of fatal
bleeding (although the rate of recurrent bleeding was the same), and a trend toward earlier obliteration
of varices. However, the overtube technique was associated with significantly more postinjection pain
and esophageal stricture formation.
Small-diameter varices may not protrude into the slot in the overtube for injection. Some authors have
addressed this problem by creating a small negative pressure within the overtube.2325(2004) As
devised by Kawano et al.,25(2005) this method is said to allow successful injection of all grades and to
increase the accuracy of intravariceal injection in small-diameter varices.
Although the overtube is attractive and reasonable in concept, it is our view that its benefits are
outweighed by poor patient tolerance. The device is also somewhat cumbersome in use. It appears,
therefore, that the overtube is not used in most centers, particularly those in the United States, since
satisfactory injections can be achieved by the freehand method.
Manometric Method
Hosking et al.26(2006) in 1986 described a manometric technique that can be used to determine
variceal patency and to ensure intravariceal injections. In this method, a polytetrafluorethylene variceal
injector (21-gauge) is connected through a flow regulator to a pressure transducer and infused with 5%
dextrose at a flow rate of 0.05 ml/min. A syringe containing sclerosant (5% ethanolamine oleate in the
initial description) is connected to the variceal injector using a three-way tap. Sclerosant is injected
after intravariceal pressure is measured to determine whether a varix is patent or thrombosed. The
criterion for varix thrombosis is a pressure greater than 80 mm Hg; pressures less than 50 mm Hg
indicate patency. An intravariceal pressure of 50 mm Hg appears to be high for patent varices; since
Hosking et al.26(2007) do not provide a zero reference, the significance of these pressures is not
known.
In a prospective controlled study, Hosking et al.27(2008) used the manometric technique to determine
variceal patency and found that this resulted in a significant increase in the rate of variceal obliteration
and decreases in the recurrence of bleeding and ulceration. Similar results using the manometric
method of variceal injection have not been reproduced by other investigators.
For endoscopists with experience in sclerotherapy, the technique of measuring intravariceal pressure
is not difficult. Although a needle inserted into a varix may appear to be in good position, a recording of
a consistent and satisfactory waveform, rather than a specific pressure measurement (Figure 326), is
necessary to confirm an intravariceal location. In over 200 intravariceal pressure measurements, we
have not encountered values of 50 mm Hg. In the study of Hosking et al.,27(2009) only 11 of 33
patients treated using the manometric method had obliteration of varices at the end of the study (mean
follow-up, 13 months); this is significantly lower than the 67 to 75% obliteration rate observed in most
controlled trials. Experienced endoscopists will find the manometric technique tedious and without
significant effect on the results of therapy.
(2010)Figure 326. Tracing made during intravariceal pressure measurement. The esophageal
luminal pressure is 1 mm Hg. After placement of the needle in the varix at endoscopy, a small
volume of saline solution is injected to flush the needle; a consistent waveform indicates that
the needle is within the varix. The intravariceal pressure is 17 mm Hg; the esophageal
variceal-luminal pressure gradient is 16 mm Hg. (From Korula J. Investigative studies in
portal hypertension. Gastroenterol Nurs 1993; 15[6]:2339.)
Balloon Tamponade
A variety of balloons for tamponade were used in some early trials of sclerotherapy with fiberoptic
endoscopes with the objectives of facilitating injection, prolonging the effect of the sclerosant within the
varix, and preventing postinjection bleeding.28(2011) There are several possible approaches to
balloon tamponade. A balloon can be positioned distal to the endoscope during injection by attaching it
to a long tube inserted through one accessory channel of a two-channel endoscope. When inflated at
the gastroesophageal junction prior to variceal injection (performed via the second accessory channel),
the balloon was expected to obstruct cephalad variceal flow from the stomach into the esophagus and
thereby facilitate injection. Several balloons, similar in concept, were fashioned for proximal placement
on the insertion tube of the endoscope. These were designed to compress varices after injection, to
reduce postinjection bleeding, and to impede cephalad variceal flow so that the sclerosant would
remain within the varix long enough to result in thrombosis.29(2012)
By mixing radiographic contrast with sclerosant and injecting varices under fluoroscopy, Grobe et
al.30(2013) found no impediment to flow when a balloon was inflated proximal to the injection site. It is
unlikely that variceal compression immediately after injection of a sclerosing agent has any appreciable
benefit. In a randomized controlled trial of variceal compression, Rhodes et al.31(2014) found no
demonstrable benefit for postinjection variceal compression in terms of time required for variceal
eradication, total numbers of patients with recurrent bleeding, and deaths owing to early recurrent
variceal bleeding.
Tamponading balloons for use in conjunction with sclerotherapy have largely become obsolete and are
not considered useful. Satisfactory injections are achieved using the freehand method, and the
frequency of postinjection bleeding is low.
Sclerosing Agents
A number of sclerosing agents are available (Table 322). As experience with endoscopic
sclerotherapy evolved, personal preferences, availability, and cost all played a role in the choice of
sclerosant(s). The largest experience worldwide is with 5% ethanolamine oleate (Ethamolin, Schwarz
Pharma, Inc., Milwaukee, Wisconsin). The choice of a sclerosing agent is determined by certain
specific properties: its ability (1) to control active bleeding from a varix, (2) to produce thrombosis, (3)
to induce sufficient necrotizing inflammation without severe ulceration, and (4) to lead ultimately to
obliteration of varices.
TABLE 322
Sclerosing Agents
EFFECTS
AGENTS
Alcohol
Sodium tetradecyl
(STD) sulfate 3%
Sodium morrhuate
Polidocanol
Ethanolamine oleate
(EO) 5%
Phenol
Mixtures:
Concentration
(%)
Ulcers
(%)
Obliteration
(%)
95
47
1.0
1.5
5.0
0.5
3.0
80
10
40
60
0
90
Elkins-Sinn
30
51
7
80
82*
33
Glenwood-Palisades
Kreussler GmbH
Schwarz Pharma
14
88
MANUFACTURER
STD + saline solution

10
70
STD + alcohol
0
50
STD + thrombin
8
33
Data from Jensen DM, Machicado GA, Silpa M. Esophageal varix hemorrhage and
sclerotherapyAnimal studies. Endoscopy 1986; 18:1822.
* Composite data from all studies using polidocanol.
Complete data not available for STD + cefazolin.
Each sclerosant has specific characteristics. For example, absolute alcohol has the property of being
the most ulcerogenic by comparison with other sclerosants. Other commonly used agents arranged in
approximate order of decreasing propensity for ulcer formation, are sodium tetradecyl sulfate
(Sotradecol, Elkins-Sinn, Cherry Hill, New Jersey), sodium morrhuate (Scleromate,
Glenwood-Palisades Pharmaceuticals, Tenafly, NJ), and ethanolamine oleate. Sodium morrhuate and
ethanolamine oleate result in significant thrombosis, perhaps owing to their fatty acid composition.
Mechanism of Action
Relatively little information is available on the mechanism(s) of action of the various sclerosing agents.
Earlier work in animal models probably has limited relevance to the treatment of esophageal variceal
hemorrhage. Sodium tetradecyl sulfate was found in animal experiments by Reiner32(2015) to be
more thrombogenic than sodium morrhuate. Blenkinsopp33(2016) compared 5% ethanolamine to
solutions of tetradecyl sulfate in an animal model (rat) and found the latter to be superior with respect
to venous occlusion. Tetradecyl sulfate was found to be superior to ethanolamine oleate and sodium
morrhuate in a study in animals because its action was more localized, an effect attributed to its neutral
pH compared with the alkaline pH of the fatty acid solutions.34(2017)
Paravariceal Injection
Histopathologic studies have shown that initially edema develops within the subepithelium and
submucosa. This is followed by an intense infiltration of macrophages and fibroblasts. After several
weeks to months, fibrosis develops along with occasional ectatic vessels within the sclerosed
submucosa.6(2018) Rarely, the inflammatory infiltrate can extend into the periesophageal tissue and
lead to mediastinitis.6(2019) The fibrotic reaction is thought to constitute a "protective" layer that
prevents rupture of the varices, and the risk of bleeding does not decrease until the fibrotic layer is fully
developed. There is no protective effect against bleeding from gastric varices because these vessels
are not injected using this technique.
Intravariceal Injection
Intravariceal injection of a sclerosing agent is thought to result in thrombus formation within the varix
and an inflammatory reaction that eventually obliterates the variceal channel (Figure 327).
(2020)Figure 327. Photomicrograph of a section of the esophageal wall showing

inflammatory reaction and fibrous obliteration of variceal channels as a result of
sclerotherapy. (From the collection of Dr. M. V. Sivak, Jr.)
Kang et al.35(2021) found that plasma fibrinopeptides were transiently elevated in 20 patients who
underwent sclerotherapy with ethanolamine oleate. This observation suggests that there is transient
activation of coagulation and fibrinolysis in response to injection of this sclerosing agent. However,
studies in vitro by these same investigators indicate that ethanolamine also inhibits fibrin clot formation,
while the oleate fraction of this sclerosing agent exhibits procoagulant activity. Clinically, Kang et
al.35(2022) found that the overall response to intravariceal injection of ethanolamine is local activation
of the coagulation system that is accelerated by an acute inflammatory reaction to the oleate fraction of
the agent. Musso et al.36(2023) suggested that the contact phase of blood coagulation may be directly
activated by polidocanol during sclerotherapy. These authors speculated that this may have desirable
as well as adverse systemic effects such as complement activation.
Comparative Studies
Jensen et al.37(2024) compared the effects of sclerosants in a canine model of portal hypertension
and found that 95% ethanol caused excessive damage to the esophagus. These investigators also
found 1.5% sodium tetradecyl sulfate to be as effective as 5% ethanolamine oleate; both of these were
more effective than 5% sodium morrhuate. In this study, the combination of 1.5% sodium tetradecyl
sulfate, 50% ethanol, and normal saline solution seemed to produce the best sclerosant effect.
There are a number of clinical studies comparing sclerosants. Frequently, the results of these studies
are divergent.
Sarin et al.38(2025) initially reported good results with absolute alcohol. Later, however, these
investigators retrospectively analyzed the effects of absolute and 50% alcohol in 79 patients.39(2026)
The only differences in response between the two concentrations were significant increases in the
occurrence of fever and retrosternal pain in the patients treated using absolute alcohol. The rate of
ulceration associated with each concentration of alcohol was similar and was reported to be very low at
about 10 to 13%.
Sarin et al.40(2027) compared the efficacy and safety of 5% ethanolamine oleate with that of absolute
alcohol in a prospective randomized trial in which 23 patients were treated using the former agent and
24 patients with the latter. Sclerotherapy was performed every 3 weeks. Variceal obliteration was
significantly faster with absolute alcohol, eradication being accomplished at about 13 weeks compared
with more than 22 weeks for patients treated with ethanolamine. The mean volume of sclerosant and
number of treatment sessions were significantly less in the alcohol-treated patients. The frequency of
recurrent variceal bleeding was 20.8% for the alcohol-treated group compared with 30.4% for the
ethanolamine-treated patients. Although this difference was not statistically significant, there were
fewer episodes of recurrent bleeding among patients who underwent injections of absolute alcohol.
Two patients (8.1%) in the ethanolamine-injected group died owing to recurrent variceal bleeding; none
died owing to bleeding in the group treated by alcohol injections. Complication rates were similar for
both groups of patients. The authors emphasized that absolute alcohol is inexpensive and readily
available.
Kochhar et al.41(2028) compared the effects of absolute alcohol, 3% sodium tetradecyl sulfate, and
5% ethanolamine oleate and also found no significant differences in the rate of recurrent bleeding,
ulceration, or stricture formation among the patient groups. In their experience, the effects of
sclerotherapy with absolute alcohol appeared not to be influenced by etiology of the underlying liver
disease.42(2029)
In a randomized trial that included 52 patients, Chawla and Dilawari43(2030) compared 3% sodium
tetradecyl sulfate with absolute alcohol. The number of treatment sessions required for variceal
obliteration was about the same in both groups. Although ulceration was more common in patients who
underwent sclerotherapy with 3% sodium tetradecyl sulfate and severe substernal pain was more
commonly associated with absolute alcohol, these differences were not statistically significant. By
contrast, Paoluzi et al.44(2031) and Atmakuri et al.45(2032) noted severe ulceration and retrosternal
pain in their patients who received absolute alcohol compared, respectively, with those who underwent
sclerotherapy with polidocanol and with ethanolamine.
A controlled study by Kitano et al.46(2033) comparing ethanolamine oleate and polidocanol
(Aethoxysklerol, Kreussler Co. GmbH, Wiesbaden, Germany) demonstrated apparent superiority for
ethanolamine oleate, which was associated with significantly lower rates of recurrent bleeding and
ulceration and a shorter time to obliteration. These investigators also noted an apparent benefit when
human thrombin was added to ethanolamine oleate.47(2034) The results of the study comparing
ethanolamine oleate and polidocanol by Balanzo et al.48(2035) correspond to the findings of Kitano et
al.46(2036) Polidocanol was found to be somewhat more effective when compared with sodium
tetradecyl sulfate in a study by Bhargava et al.49(2037) of 61 patients with differing causes of portal
hypertension.
The results of a double-blind, randomized, and controlled trial of 3% sodium tetradecyl sulfate and 5%
ethanolamine oleate were reported by Chan et al.50(2038) Intravariceal sclerotherapy was performed
within 24 hr of admission in 95 patients with active variceal bleeding or signs of recent variceal
hemorrhage, the endoscopist being unaware of which agent was being used. Subsequent treatments
were performed at weekly intervals until variceal obliteration was achieved. Essentially, there were no
differences between the treatment groups with respect to control of the acute bleeding episode and the
occurrence of major local complications (esophageal ulceration, stricture formation, and perforation).
There was, however, a greater frequency of postsclerotherapy fever in the group treated with sodium
tetradecyl sulfate. Variceal obliteration was achieved after a mean of 3.3 treatment sessions in the
group that underwent sclerotherapy with sodium tetradecyl sulfate versus 4.5 sessions for those
treated with ethanolamine oleate, this being the only major difference in outcome between the two
treatment groups.
Sodium tetradecyl sulfate and sodium morrhuate were found by McClave et al.51(2039) to be
equivalent in a prospective randomized trial in which 21 patients underwent sclerotherapy with 0.75%
sodium tetradecyl sulfate and 20 patients were treated using 1.6% morrhuate diluted with 50%
dextrose. Essentially, no significant differences were found with respect to control of acute bleeding,
transfusion requirements, variceal obliteration, ulceration, stricture formation, and overall mortality.
Variceal obliteration was achieved in only 33% of the tetradecyl sulfate-treated group and 25% of the
morrhuate-treated patients. However, the obliteration rate was higher (greater than 80%) for patients in
both groups who remained in the study for more than 3 months. Although not statistically significant,
there was a trend toward a higher rate of recurrent bleeding for patients who underwent injections of
tetradecyl sulfate.
Noncomparative Trials
Sclerotherapy has been performed with 3% aqueous phenol. Mathur et al.52(2040) reported on the
use of this agent in 301 patients with portal hypertension owing to various causes. In the 72 patients in
whom sclerotherapy was performed as an emergency procedure, bleeding was controlled in 87% of
cases. The in-hospital mortality rate was 9.6% with approximately half of these deaths being
attributable to bleeding; for patients undergoing emergency sclerotherapy, the mortality rate was
approximately 10 times higher than that for patients whose initial treatment was considered elective.
Variceal eradication was achieved in 84% of patients after a mean of 5.14 treatment sessions. The rate
of recurrent bleeding in 290 surviving patients was 35%. A complication, including superficial ulceration
and stricture formation, occurred in 14% of patients. The complication rate for patients treated on an
emergency basis was more than twice that for patients in whom the procedure was regarded as
elective.
The use of sclerosant mixtures consisting of several agents with presumably different properties and
mechanisms of action has been proposed by a number of investigators. In the study of Lyons et
al.,53(2041) the addition of thrombin and cefazolin to sodium tetradecyl sulfate seemed to be
somewhat more beneficial for the early control (fewer hospital days and transfusions) of variceal
bleeding when compared with the use of sodium tetradecyl sulfate alone. Nakamura et al.54(2042)
also found the combination of sodium tetradecyl sulfate, bovine thrombin, and cefazolin to be effective
for control of the initial episodes of variceal bleeding in 101 patients. No systemic allergic or thrombotic
events related to the use of bovine thrombin were observed in this study.
Despite the differing results of the various studies of sclerosing agents, it is apparent from clinical
practice that the use of absolute alcohol for sclerotherapy can result in serious esophageal ulceration,
chest pain, and fever. Therefore, its use as a sclerosing agent has largely been abandoned, even
though it is inexpensive and readily available. It does not appear that any one of the commonly used
agents such as ethanolamine oleate, sodium tetradecyl sulfate, and sodium morrhuate has any clear
advantage over the others. It is likely that the clinical response to injections of a sclerosing agent is
determined as much by the volume and concentration of the agent as its particular chemical
composition. In effect, the difference between variceal obliteration and a serious complication is
determined by the amount of active agent placed within or adjacent to a varix.
Injection Volume
Wide variations in the volume of sclerosant are found in early reports of variceal sclerotherapy
because there were no data as to the optimum volume. Only a few controlled studies have evaluated
the effect of different volumes of sclerosant on recurrent bleeding and obliteration of varices.
Unfortunately, these studies are not comparable because different sclerosants were used. For practical
purposes, the volume of sclerosant per injection, per treatment session, and the total volume required
for eradication of varices are determined empirically by observation at endoscopy. In general, however,
smaller volumes of sclerosant are required for small varices, larger volumes for larger columns.
Iso et al.55(2043) compared the effects of intravariceal sclerotherapy using large and small volumes of
5% ethanolamine oleate. In each group, the volume of sclerosant injected at the initial procedure was
the largest, with decremental volumes being injected at subsequent treatments. The small-dose patient
group received approximately half the volume of the large-volume injection group. Although variceal
eradication was complete in both groups, fewer treatments were required over a shorter time span in
the large-volume group. Mucosal ulceration was significantly greater in the large-volume injection
group, but no other differences in complications or survival were noted between the two groups.
Akriviadis et al.56(2044) compared the efficacy of large-volume (greater than 10 ml) weekly injections
of 1.5% sodium tetradecyl sulfate with small-dose (less than 10 ml) injections performed twice weekly.
Although time to obliteration, the rate of recurrent bleeding, and numbers of patients developing ulcers
were the same in both groups, a trend toward an increase in esophageal perforation in the patients
who underwent frequent small-volume injections necessitated termination of the study. Furthermore, in
the small-volume injection group, more injections were made during active bleeding where visibility was
diminished. This may have contributed to the higher frequency of life-threatening complications in the
group of patients that received small-volume injections.
Currently, we inject small volumes of sclerosant during weekly treatment sessions into multiple variceal
sites along their proximal linear course and into junctional gastric varices. Although not confirmed by a
controlled study, this technique appears to result in significantly lower ulceration rates, negligible
stricture formation, and infrequent episodes of recurrent bleeding.
Schedule of Injections
The interval between treatment sessions in the early reports of variceal sclerotherapy using fiberoptic
instruments varies between 2 and 4 weeks. Intervals between procedures, as with the volumes of
sclerosant injected, were chosen empirically. However, this had the objective of allowing a period of
time after the initial thrombosis and inflammatory reaction for fibrosis and obliteration of the varices to
develop. In addition, the observation that large areas of esophageal ulceration might develop, in some
instances leading to sloughing of the mucosa and even destruction of the wall of the esophagus,
indicated a need for a cautious, measured approach toward the desired goal of variceal obliteration.
Several studies have attempted to determine the optimum frequency and timing of treatments.
Westaby et al.57(2045) compared weekly injection treatments with injections performed at intervals of
3 weeks using 5% ethanolamine oleate in 55 patients. Obliteration was achieved in a shorter time, but
mucosal ulceration was more frequent with the weekly injection schedule. Nevertheless, this was not
associated with any increase in pain after the procedure or the development of strictures. No difference
was observed in the number of treatments required for obliteration or the rate of recurrent bleeding.
In a similar study but using absolute alcohol, Sarin et al.58(2046) found that injections performed
weekly resulted in earlier eradication of varices (7.1 vs. 14.9 weeks for patients undergoing treatment
every 3 weeks) and less recurrent bleeding. The latter occurred in 8.5% of the 47 patients undergoing
weekly injections versus 26.5% of 49 patients who underwent sclerotherapy at intervals of 3 weeks.
Mucosal ulceration was observed in 68% in the weekly treatment group compared with 12% in the
group treated every 3 weeks. Ulceration necessitated postponement of treatment in about a quarter of
the patients. However, there was no difference between the groups for stricture formation, chest pain,
or fever. Based on these data, Sarin et al.58(2047) concluded that weekly sclerotherapy sessions with
injection of small volumes at each variceal site was the most effective schedule.
Mathur et al.59(2048) studied 125 patients in which injection groups were composed as follows: 3 to 5
ml of aqueous phenol injected at intervals of 3 weeks, 1 week, and 3 days, and injections of 2 to 3 ml
of phenol every 3 days. Thereafter, injections were performed at intervals of 4 weeks in all patients.
The results were similar to those obtained in other studies, that is, the time required for variceal
obliteration was shortest when the interval between injection sessions was brief. There was no
significant difference in the rate of recurrent bleeding for patients treated every 3 days versus those
treated at 1- or 3-week intervals. However, the mortality owing to recurrent bleeding was significantly
less in patients treated every 3 days (nil) versus those treated weekly (13.3%) and those treated every
3 weeks (10.7%). As with the other studies of similar design, mucosal ulceration and stricture
formation were significantly more common in patients injected every 3 days.
The value of continuing injection sessions until obliteration of varices is achieved was studied by
Moreto et al.60(2049) In this study, 56 patients underwent injections only for control of ongoing
bleeding (referred to as short-term sclerotherapy by these investigators). Another 50 patients
underwent short-term sclerotherapy but also had continued sclerotherapy sessions. No significant
difference in the overall cumulative proportion of patients with recurrent bleeding was found. However,
there were fewer episodes of variceal hemorrhage in the group that had both short-term and long-term
sclerotherapy. This approach reduced the rate of recurrent bleeding from esophageal varices but not
from junctional varices (i.e., those within the cardia distal to the gastroesophageal junction).
Endpoints of Therapy
A desired objective of sclerotherapy is elimination of the risk of recurrent bleeding. Intravariceal
technique with obliteration of varices is associated with a significant reduction in recurrent bleeding and
an increase in bleeding-free interval. Thus, obliteration of varices appears to be a satisfactory endpoint
(Figure 328).9,61(2050) However, it is sometimes difficult to judge whether the degree of variceal
eradication is adequate. To determine how much of a reduction in variceal size is necessary before it
can be concluded that sclerotherapy has reached an endpoint, de Franchis et al.62(2051) carried out a
prospective study of bleeding risk in patients whose varices were either obliterated or remained small.
They concluded that recurrence of varices and bleeding risk were significantly reduced if the varices
were either obliterated or persisted as small white columns. However, persistent small bluish varices
were associated with a continued risk of bleeding. Agrawal et al.63(2052) demonstrated that
persistence of small varices is an acceptable endpoint. The rate of recurrent bleeding associated with
small-diameter vessels was significantly lower than that for larger varices.
(2053)Figure 328. A, Appearance of the distal esophagus after six sclerotherapy sessions. B,
Note the residual vasa vasorum of the varices. (A and B, From the collection of Dr. M. V.
Sivak, Jr.)
With paravariceal technique, varices are in theory not obliterated, and the risk of recurrent bleeding is
diminished by the development of a fibrotic protective layer. The endpoint using this technique is
somewhat more difficult to discern. Histologic studies have demonstrated that thickening and fibrosis
develop in the subepithelium and submucosa several weeks to months following paravariceal injection,
although ectatic vessels may still be present.6(2054)
Sclerotherapy of Gastric Varices

The natural history of gastric varices is not known. However, bleeding from these abnormal vessels is
associated with high morbidity and mortality, especially in patients with advanced liver disease (Figure
329).
(2055)Figure 329. A, Bleeding gastric varix. B, Sclerotherapy. Bleeding was controlled, but
it was necessary to inject a relatively high volume of the sclerosant (1% sodium tetradecyl
sulfate). (A and B, From the collection of Dr. M. V. Sivak, Jr.)
Stray et al.64(2056) reported good results with sclerotherapy of gastric varices in an uncontrolled study
of a small group of patients. Sclerotherapy of gastric varices in controlled trials has been associated
with severe ulceration and high rates of recurrent bleeding.9(2057)
In a prospective nonrandomized study of 92 patients, Trudeau and Prindiville65(2058) identified gastric
varices in 10% of patients, a subgroup with a higher rate of recurrent bleeding and lower survival.
Korula et al.66(2059) demonstrated two subsets of patients in whom gastric varices developed during
variceal sclerotherapy; although gastric varices occurred in 15% of patients, fundal varices developed
in only 4% but these were associated with increased rates of recurrent bleeding and mortality. The
clinical behavior of junctional gastric varices, that is, those present distal to the gastroesophageal
junction, was similar to that of esophageal varices with regard to recurrent bleeding, complications, and
survival. Sarin et al.67(2060) attempted a combination of intravariceal and paravariceal injection of
gastric varices and reported success in a small uncontrolled series of patients.

We inject junctional varices in a manner similar to esophageal varices, using either an antegrade or a
retrograde approach (Figure 3210). In most instances, we prefer to make injections from the
antegrade direction because difficulty may be encountered in advancing the needle injector within the
accessory channel when the endoscope is retroflexed. When junctional varices are not easily
accessible with the antegrade approach, injection is performed from a retroflexion position within the
stomach. In the latter instance, it is important to advance the injector so that the tip is within the
endoscopic field of view prior to the retroflexion maneuver; the needle can then be inserted into the
varix after a satisfactory position is achieved. It may be necessary to torque the endoscope to obtain a
satisfactory position, especially for varices that are located posteriorally or along the lesser curvature.
(2061)Figure 3210. Bleeding "junctional" varix just distal to the gastroesophageal

squamocolumnar junction. (From the collection of Dr. M. V. Sivak, Jr.)
Soehendra et al.68(2062) demonstated that injection of a tissue adhesive, N-butyl 2-cyanoacrylate
(Histoacryl, Braun-Melsungen, Germany), into large fundal varices is both feasible and effective. This
technique has led to decreasing enthusiasm for injection of sclerosants in the management of bleeding
gastric varices (see Chapter 14: The Use of Histoacryl Tissue Adhesive for the Treatment of Variceal
Bleeding).
Sclerotherapy During Active Hemorrhage

The mettle and skill of an endoscopist are best demonstrated during active bleeding from esophageal
varices. It is almost impossible to control hemorrhage from esophageal varices by endoscopic
sclerotherapy when bleeding is torrential and precludes endoscopy, when the rent in the variceal wall is
large and the varices are collapsed, or when brisk bleeding prevents adequate visualization during
endoscopy (Figure 3211). For satisfactory endoscopy whenever there is active bleeding, the patient
must be hemodynamically stable, the airway should be protected, and lavage prior to endoscopy
should be adequate to clear the stomach of blood and clots.
(2063)Figure 3211. Large, actively bleeding varix in the distal esophagus. (From the
Although immediate sclerotherapy was slightly more effective than sclerotherapy preceded by balloon
tamponade in the control of hemorrhage and associated with lower complication rate in the trial of Lo et
al.,69(2064) it may nevertheless be necessary to resort temporarily to balloon tamponade to control or
lessen the intensity of torrential bleeding. Any of the available tubes (Linton-Nachlas,
Sengstaken-Blakemore, Minnesota) can be used provided the duration of tamponade is not over 6
hours in the usual case and 12 hours when bleeding is exceptionally severe. The longer the duration,
the more severe the mucosal injury in the distal esophagus, which increases the difficulty of
subsequent sclerotherapy and the potential for a complication. Endoscopy and sclerotherapy should be
performed immediately on deflation and removal of the balloon device.
A quick but complete endoscopic examination of the upper gastrointestinal tract should be performed
prior to sclerotherapy. Approximately 50% of patients with portal hypertension and gastroesophageal
varices have associated lesions such as peptic ulcer, gastroduodenal erosions, hemorrhagic "gastritis,"
or antral ectasias. Thus, exclusion of these lesions should be attempted prior to the performance of
sclerotherapy. Although the enigmatic entity of portal hypertensive gastropathy (congestive
gastropathy) is considered a cause of bleeding, it does not appear to be an important source of
significant bleeding. Antral vascular ectasias and portal hypertensive gastropathy are both difficult to
diagnose when large volumes of blood are present in the stomach. During endoscopy, we prefer to
wash the lumen constantly using water jets from a separate catheter or the irrigation attachment of the
BICAP probe (Endostat, Microvasive, Natick, Massachusetts). Constant suctioning may lead to a
clogged endoscope; use of a two-channel instrument may be helpful, especially when bleeding is brisk,
but in most instances a single-channel endoscope should suffice.
Bleeding from varices occurs mostly within 2 to 5 cm of the gastroesophageal junction (Figure 3212);
consequently, careful examination of the distal esophagus and lesser curvature of the stomach is
critical to diagnosis and treatment. When bleeding is not active, the varices should be observed for
several minutes because bleeding can be intermittent. The presence of a "white nipple" sign70(2065)
or clot adherent to a varix provides an additional clue to the site of hemorrhage (Figure 3213).
Cherry-red spots or red wale markings on varices are considered to be harbingers of
bleeding;71(2066) recent studies suggest that these features are also seen in patients without a history
of bleeding.72(2067)
(2068)Figure 3212. A, Actively bleeding varix near the gastroesophageal junction. B,

Close-up view of a bleeding varix. (A and B, From the collection of Dr. M. V. Sivak, Jr.)
(2069)Figure 3213. A, Clot protruding from a varix in a patient with recent severe
hemorrhage. B, Injection of a varix proximal to the clot. C, Severe bleeding begins as the clot
is dislodged from the varix. Bleeding is subsequently controlled by injection of a large volume
of sclerosant solution using both paravariceal and intravariceal punctures. (A-C, From the
Bleeding from gastric fundal varices can be difficult to recognize because large amounts of blood and
clots collect in the fundus when a patient is in the left lateral position; however, visualization of the
fundus is possible if gastric lavage is meticulous. We sometimes place a patient in the right lateral
position to clear blood from the fundus. In every patient with variceal bleeding, it is essential to search
for gastric fundal varices; the high failure rate of sclerotherapy in patients with fundal varices may make
it necessary to consider alternative treatments such as a transjugular intrahepatic portacaval shunt
(TIPS) or portosystemic shunt surgery. In countries where tissue adhesives are available (e.g.,
cyanoacrylate), injection of these agents is another alternative, especially in poor-risk patients.
Several maneuvers are available to achieve hemostasis when blood is observed to be spurting from a
varix (Figure 3214). Injection of sclerosant into the bleeding varix adjacent to the site will result in
immediate hemostasis, provided the rent is relatively small. If the bleeding does not stop, sclerosant
can be injected paravariceally on either side of the bleeding point (or proximal and distal to it). The
resulting localized bleb usually transforms the spurt to a trickle or ooze. The change in bleeding rate
improves visibility and allows injection of other varices. Because of the presence of interconnecting
channels, injection of adjacent varices can control bleeding. Care must be exercised not to inject
inordinate volumes of sclerosant, particularly in the stomach, because severe ulceration may result,
especially when liver disease is advanced or severely decompensated (Figures 3215 and 3216).
Oozing from the multiple injection sites can obscure the endoscopic field and may be a source of
anxiety for the endoscopist; such bleeding usually resolves and does not require additional injections
(Figure 3217).
(2070)Figure 3214. A, Blood spurting from a varix near the gastroesophageal junction. B,
Injection needle (bottom) placed near the bleeding point. C, Injection of sclerosant solution.
D, Bleeding slows. E, Bleeding controlled after slow injection of 2 ml of sclerosant solution.

(A-E, From the collection of Dr. M. V. Sivak, Jr.)
(2071)Figure 3215. Barium swallow radiograph of the same patient as shown in Figure
3213 demonstrates ulceration with extensive necrosis and walled-off perforation. (From the
(2072)Figure 3216. Esophageal ulceration and necrosis 7 days after sclerotherapy for acute
sclerotherapy-induced bleeding. Same patient as shown in Figure 3213. (From the collection
of Dr. M. V. Sivak, Jr.)
(2073)Figure 3217. Ooze of blood from a varix immediately after injection of sclerosant and
removal of the injection needle. Such bleeding usually stops spontaneously after 1 or 2
minutes; further injections are usually not required. (From the collection of Dr. M. V. Sivak,
Jr.)
If the initial effort to control bleeding is unsuccessful, the procedure should be terminated. At this stage,
experienced endoscopists may sometimes attempt multiple additional paravariceal injections at the
bleeding site to achieve hemostasis. However, a large number of injections at one site, even if injected
volumes of sclerosant are small, increases the risk of severe necrotizing inflammation and perforation.
Under these circumstances, it may be prudent to offer alternative treatments, temporizing with
measures such as balloon tamponade, while consideration is given to a TIPS procedure or
portosystemic shunt surgery.
Complications Related to Technique

A major factor contributing to the development of complications is the volume of sclerosant at the
injection site. Even if intravariceal injections are made correctly, sclerosant extravasates into the
lamina propria.4(2074) Using a radiographic technique, Rose et al.73(2075) demonstrated that
endoscopic assessment of intravariceal injection is accurate in only 75% of cases, the likelihood of
paravariceal injections (36%) being greatest when varices are small. It is not clear whether
extravascular sclerosant is due to leakage consequent to damage to the vein wall or inadvertent
paravariceal injection. However, the degree of necrotizing inflammation is approximately proportional to
the volume of sclerosant present in the lamina propria. It is therefore essential to minimize the volume
of sclerosant in extravariceal locations. With parivariceal technique, only small volumes should be
injected.
We prefer to add methylene blue to the sclerosant mixture, as this makes it easier to identify and
estimate the volume of sclerosant in blebs and localized areas of blanchingobservations that indicate
the need to avoid further injections. Even when intravariceal injection is satisfactory, large volumes are
best avoided to minimize extravasation of sclerosant. "Blind" injections and injection of large volumes
of sclerosant must be avoided, even if there is active bleeding, because of the risk of serious and
life-threatening complications.74(2076) Although ulceration almost always develops at injection sites
and is considered by some to be an accompaniment rather than a complication of
sclerotherapy,75(2077) the occurrence of deep ulcers must be avoided, as these increase the risks of
recurrent bleeding and perforation (Figures 3218, 3219, 3220, 3221, 3222). Patients with very
advanced liver disease are particularly prone to severe ulceration. Control of bleeding in these patients
is usually difficult and may require multiple injections. Furthermore, the severe coagulopathy that
commonly arises in these patients may result in significant oozing of blood following injections. This
may lead the inexperienced endoscopist to make additional and unnecessary injections that perpetuate
the bleeding and increase the risk of a complication. For sclerotherapy to be successful, the number
and volume of sclerosant injections made for control of bleeding must be balanced against the risk of
severe esophageal ulceration and necrosis (Figure 3223).
(2078)Figure 3218. Yellow plaque over a variceal injection site 5 days after sclerotherapy.
(2079)Figure 3219. A, Yellow plaque over an ulcerated variceal injection site 5 days after
sclerotherapy. B, Removal of yellow exudate reveals an ulcer over the varix. (A and B, From
(2080)Figure 3220. Esophageal ulcer 1 week after sclerotherapy. (From the collection of Dr.
M. V. Sivak, Jr.)
(2081)Figure 3221. Postsclerotherapy bleeding from the region of a yellow plaque over an
ulcerated varix. (From the collection of Dr. M. V. Sivak, Jr.)
(2082)Figure 3222. Gross specimen of the esophagus and stomach showing a

sclerotherapy-induced esophageal ulcer in an area of partially thrombosed varices. Patient
died because of exsanguination. (From the collection of Dr. M. V. Sivak, Jr.)
(2083)Figure 3223. A, Healed esophageal ulcer defect after 6 sclerotherapy sessions. B,

Residual bridge of tissue after sclerotherapy. (A and B, From the collection of Dr. M. V.
Sivak, Jr.)
Failure of Sclerotherapy
Endoscopic variceal sclerotherapy should be discontinued when treatment fails. However, there is no
precise definition for failure of sclerotherapy. In controlled trials, this is defined as the inability to control
acute hemorrhage from varices, recurrent bleeding after variceal obliteration, and death related to
bleeding or treatment complications. Inability to obliterate varices, nonfatal complications of the
treatment, bleeding from sclerotherapy-induced ulceration, and recurrent bleeding prior to obliteration
are less definite criteria for failure. For instance, inability to obliterate varices may be due to
sclerotherapy technique that is not sufficiently aggressive or to technical problems. The number of
bleeding episodes prior to obliteration, the severity of the bleeding episode, the relationship of recurrent
bleeding to sclerotherapy-induced ulceration, and the severity of liver disease are some of the factors
that must be considered in the decision to stop or continue treatment. We regard the occurrence of two
bleeding episodes while a patient is undergoing sclerotherapy by an experienced endoscopist as a
reasonable basis to consider alternative therapy.
Recurrence of Varices and Obliteration

The factors responsible for the re-formation of varices after obliteration are not clearly understood.
Kitano et al.18(2084) proposed that the removal of the esophageal mucosa lowers the rate of
recurrence. In the few controlled studies that address the problem of recurrence, varices were
observed to recur after obliteration in from 15 to 60% of cases.11,61,76(2085) The wide range in
frequency may reflect the timing of follow-up endoscopy and the onset of recurrent bleeding. Westaby
et al.61(2086) observed that recurrence usually takes place within 12 months after initiation of
treatment. These investigators also found that despite the high frequency of recurrence, only a third of
patients with newly formed varices experienced bleeding. Sarin et al.77(2087) found that varices
recurred in 19% of their patients with either cirrhosis or noncirrhotic portal fibrosis, but there was no
evidence of recurrence in their patients with extrahepatic portal vein obstruction. Although the
mechanisms underlying variceal recurrence after sclerotherapy are unclear, these newly formed
vessels respond promptly to further sclerotherapy. Usually, only one or two treatments are required.
Varices can develop in the proximal third of the esophagus in patients in whom distal esophageal
varices have been obliterated by sclerotherapy. Although the risk of bleeding from these varices is not
known precisely, we have seen serious hemorrhage from these varices. It is therefore important to
inject and obliterate these vessels. However, sclerotherapy in this location must be performed
cautiously because the aorta, trachea, carina, and main stem bronchi are in close proximity and injury
to these structures must be avoided; deep and large-volume injection may predispose to localized
abscess formation and the consequent life-threatening complications of mediastinitis and perforation.
Factors that Predict Failure of Sclerotherapy

It is not clear why from 10 to 30% of patients who undergo sclerotherapy have recurrent bleeding and
fail treatment. Bleeding from gastric fundal varices occurs only in 4%,66(2088) and hemorrhage from
varices in the colon, duodenum, and other unusual locations is rare. The exact frequency of bleeding
owing to portal hypertensive gastropathy is unknown, the main problem being the lack of established
and uniform criteria for the diagnosis of this entity. In our experience, however, significant bleeding
from this lesion is rare.
The relationship of portal pressure to the efficacy of sclerotherapy has not been carefully investigated.
In a preliminary study at our center of 22 patients, all of whom had portal pressure measurements at
the initiation of therapy, no significant differences in portal pressures were observed in patients who
developed recurrent bleeding and in those who did not (unpublished observations). To look for a
relationship between intravariceal pressures and success of sclerotherapy, 33 patients who had
intravariceal pressure measurements at the inception of sclerotherapy were followed for a mean period
of 12 months. There was no statistically significant difference in variceal pressure or gradient between
patients who developed recurrent bleeding following sclerotherapy and those who did not.78(2089)
Thus, neither portal pressure nor intravariceal pressure appears to influence the efficacy of
sclerotherapy.
REFERENCES
1.
2.
3.
4.
Terblanche J, Burroughs AK, Hobbs KEF. Controversies in the management of bleeding

esophageal varices [Second of two parts]. N Engl J Med 1989;320:146975.
Bornman PC, Kahn D, Terblanche J, et al. Rigid versus fiberoptic endoscopic injection
sclerotherapy. Ann Surg 1988;208:1758.
Snady H, Korsten MA, Waye JD. The relationship of bacteremia to the length of injection
needle in endoscopic variceal sclerotherapy. Gastrointest Endosc 1985;31:2436..
Korula J, Baydur A, Sassoon C, Sakimura I. Effect of esophageal variceal sclerotherapy
(EVS) on lung function. Arch Intern Med 1986;146:151720.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Kage M, Korula J, Harada A, et al. Effects of sodium tetradecyl sulfate endoscopic

sclerotherapy on the esophagus. A prospective clinical and histopathological study. J Clin
Helpap B, Bollweg L. Morphological changes in the terminal oesophagus with varices
following sclerosis of the wall. Endoscopy 1981;13:22933.
Paquet KJ, Oberhammer E. Sclerotherapy of bleeding oesophageal varices by means of
Waring JP, Sanowski RA, Pardy K, et al. Does the addition of methylene blue to the
sclerosant improve the accuracy of injections during variceal sclerotherapy?. Gastrointest
Endosc 1991;37:15960.
Korula J, Balart LA, Radvan G, et al. A prospective, randomized controlled trial of chronic
esophageal variceal sclerotherapy. Hepatology 1985;5:5849.
The Copenhagen Esophageal Varices Sclerotherapy Project: Sclerotherapy after first variceal
hemorrhage in cirrhosis. A randomized multicenter trial. N Engl J Med 1984;311:15941600.
Soderlund C, Ihre T. Endoscopic sclerotherapy v. conservative management of bleeding
Barsoum MS, Bolous FI, El-Rooby AA, et al. Tamponade and injection sclerotherapy in the
Yassin YM, Sherif SM. Randomized controlled trial of injection sclerotherapy for bleeding
oesophageal varicesAn interim report. Br J Surg 1983;70:202.
Barsoum MS, Khattar NY, Risk-Allah MA. Technical aspects of injection sclerotherapy of
acute oesophageal variceal haemorrhage as seen by radiography. Br J Surg 1978;5889.
Paquet KJ. Endoscopic paravariceal injection sclerotherapy of the oesophagusIndications,
technique, complications: Results of a period of 14 years. Gastrointest Endosc
1983;29:3105.
Sarin SK, Nanda R, Sachdev G, et al. Intravariceal versus paravariceal sclerotherapy: A
prospective, controlled, randomised trial. Gut 1987;28:65762.
Hashizume M, Kitano S, Koyanagi N, et al. Endoscopic injection sclerotherapy for 1000
patients with esophageal varices: A nine-year prospective study. Hepatology 1992;15:6975.
Kitano S, Koyanagi N, Iso Y, et al. Prevention of recurrence of esophageal varices after
Williams KGD, Dawson JL. Fibreoptic injection of oesophageal varices. BMJ 1979;2:7667.
Kitano S, Sugimachi K, Iso Y, Koyanagi N. A new overtube for endoscopic injection
sclerotherapy. In Burroughs AK (ed). Methodology and Reviews of Clinical Trials in Portal
Hypertension. Amsterdam, New York, London:. Excerpta Medica 1987;12930..
Kitano S, Yamaga H, Wada H, et al. Over-tube is preferable to free-hand technique to avoid
recurrence of varices after endoscopic injection sclerotherapy. Prospective randomized trial.
Int Surg 1992;77:13740.
Westaby D, MacDougall BRD, Melia W, et al. A prospective randomized study of two
sclerotherapy techniques for esophageal varices. Hepatology 1983;3:6814.
Aizawa T. Endoscopic injection sclerotherapy of esophageal varices utilizing over-tube (EIS
tube). Nippon Rinsho 1990;48:694700.
Nakamura H, Inoue H, Kawano T, Endo M. Advantages of the endoscopic injection
sclerotherapy under the negative pressure in a transparent over-tube. Nippon Rinsho
1990;48:7015.
Kawano T, Nakamura H, Inoue H, Endo M. Endoscopic injection sclerotherapy using a
transparent overtube with intraluminal negative pressure (np-EIS) for esophageal varices.
Hosking SW, Robinson P, Johnson AG. Can follow-up sclerotherapy for oesophageal varices
be improved? Early results of a new technique. Endoscopy 1986;18:1713.
Hosking SW, Robinson P, Johnson AG. Usefulness of manometric assessment of varices in
maintenance sclerotherapy. A controlled trial. Gastroenterology 1987;93:84651.
Sciaretta G, Verri A, Malaguti P. A new endoscopic balloon for sclerotherapy in active
bleeding from esophageal varices. Endoscopy 1985;17:1134.

29. Hughes RW, Larson DE, Vigiano TR, et al. Endoscopic variceal sclerosis: One-year
experience. Gastrointest Endosc 1982;28:626.
30. Grobe JI, Kozarek RA, Sanowski RA, et al. Venography during endoscopic injection
sclerotherapy of esophageal varices. Gastrointest Endosc 1984;30:68.
31. Rhodes JM, Dawson J, Cockel R, et al. A randomized controlled trial of variceal compression
as an adjunct to endoscopic variceal sclerosis. Scand J Gastroenterol 1986;21:121720.
32.
Reiner L. The activity of anionic surface active compounds in producing vascular obliteration.
Proc Soc Exp Biol Med 1946;62:4954.
33.
Blenkinsopp WK. Comparison of tetradecyl sulphate of sodium with other sclerosants in rats.
Br J Exp Pathol 1967;49:197201.
34.
Cooper WM. Clinical evaluation of sotradecol, a sodium alkyl sulfate solution, in the injection
therapy of varicose veins. Surg Gynecol Obstet 1946;383:64752.
35. Kang JH, Kambayashi J, Sakon M, et al. Mechanism of the haemostatic effect of
ethanolamine oleate in the injection sclerotherapy for oesophageal varices. Br J Surg
1987;74:503.
36. Musso R, Longo A, Triolo A, et al. Polidocanol may directly activate the contact phase of
blood coagulation during sclerotherapy. Gastrointest Endosc 1987;33:4001.
37. Jensen DM, Machicado GA, Silpa M. Esophageal varix hemorrhage and
sclerotherapyAnimal studies. Endoscopy 1986;18:1822..
38. Sarin SK, Sachdev GK, Nanda R, et al. Endoscopic sclerotherapy using absolute alcohol. Gut
1985;26:1204.
39. Sarin SK, Nanda R, Sachdev G. Relative efficacy and safety of absolute alcohol and 50%
alcohol as variceal sclerosants. Gastrointest Endosc 1987;33:3627.
40. Sarin SK, Mishra SP, Sachdev GK, et al. Ethanolamine oleate versus absolute alcohol as a
variceal sclerosant: A prospective, randomized, controlled trial. Am J Gastroenterol
1988;83:52630.
41. Kochhar R, Goenka MK, Mehta S, Mehta SK. A comparative evaluation of sclerosants for
esophageal varices: A prospective, randomized, controlled study. Gastrointest Endosc
1990;36:12730.
42. Kochhar R, Goenka M, Mehta SK. Outcome of injection sclerotherapy using absolute alcohol
in patients with cirrhosis, non-cirrhotic portal fibrosis and extrahepatic portal venous
obstruction. Gastrointest Endosc 1991;37:4604.
43. Chawla Y, Dilawari JB. Studies in sclerotherapy: I. Comparison of sodium tetradecyl sulphate
(STS) with absolute alcohol (AA) as sclerosants in the treatment of esophageal varices. J Clin
44. Paoluzi P, Petroiusti A, Ferrar S, et al. Absolute alcohol in esophageal vein sclerosis.
45. Atmakuri SP, Bhargava DK, Sharma MP. Absolute alcohol in variceal sclerosis. Gastrointest
Endosc 1989;35:4667.
46. Kitano S, Iso Y, Koyanagi N, et al. Ethanolamine oleate is superior to polidocanol
(aethoxysklerol) for endoscopic injection sclerotherapy of esophageal varices: A prospective
randomized trial. Hepatogastroenterology 1987;34:1923.
47. Kitano S, Hashizume M, Yamaga H, et al. Human thrombin plus 5% ethanolamine oleate
injected to sclerose oesophageal varices: A prospective randomized trial. Br J Surg
1989;76:7158.
48. Balanzo J, Sainz S, Espinos ES, et al. Efficacy of ethanolamine and polidocanol in the
eradication of esophageal varices. A prospective randomized trial. Endoscopy 1989;21:2513.
49. Bhargava DK, Singh B, Dogra R, et al. Prospective randomized comparison of sodium
tetradecyl sulfate and polidocanol as variceal sclerosing agents. Am J Gastroenterol
1992;87:1826.
50. Chan AC, Chung SC, Sung JY, et al. A double-blind randomized controlled trial comparing
sodium tetradecyl sulphate and ethanolamine oleate in the sclerotherapy of bleeding
oesophageal varices. Endoscopy 1993;25:5137.

51. McClave SA, Kaiser SC, Wright RA, et al. Prospective randomized comparison of esophageal
variceal sclerotherapy agents: Sodium tetradecyl sulfate versus sodium morrhuate.
52. Mathur SK, Naik SR, Supe AN, et al. Endoscopic esophageal variceal sclerotherapy using 3%
aqueous phenol. Gastrointest Endosc 1992;38:1527..
53. Lyons SD, Sugawa C, Geller ER, Vandenberg DM. Comparison of 1% sodium tetradecyl
sulfate to a thrombogenic sclerosant cocktail for endoscopic sclerotherapy. Am Surg
1988;54:814.
54. Nakamura R, Bucci LA, Sugawa C, et al. Sclerotherapy of bleeding esophageal varices using
a thrombogenic cocktail. Am Surg 1991;57:22630.
55. Iso Y, Kitano S, Iwanaga T, et al. A prospective randomized study comparing the effects of
large and small volumes of the sclerosant 5% ethanolamine oleate injected into esophageal
varices. Endoscopy 1988;20:2858.
56. Akriviadis E, Korula J, Gupta S, et al. Frequent endoscopic variceal sclerotherapy increases
risk of complications. Prospective controlled study of two treatment schedules. Dig Dis Sci
1989;34:106874.
57. Westaby D, Melia WM, MacDougall BRD, et al. Injection sclerotherapy for oesophageal
varices: A prospective randomised trial of different treatment schedules. Gut 1984;25:12932.
58. Sarin SK, Sachdev G, Nanda R, et al. Comparison of the two time schedules for endoscopic
sclerotherapy: A prospective randomised controlled study. Gut 1986;27:7103.
59. Mathur SK, Naik SR, Plumber ST, et al. Comparative trial of three different schedules for
endoscopic esophageal variceal sclerotherapy [see Comments]. Indian J Gastroenterol
1990;9:2014.
60. Moreto M, Zaballa M, Ojembarrena E, et al. Combined (short-term plus long-term)
sclerotherapy versus short-term only sclerotherapy: A randomised prospective trial. Gut
1994;35:68791.
61. Westaby D, MacDougall BRD, Williams R. Improved survival following injection sclerotherapy
for esophageal varices: Final analysis of a controlled trial. Hepatology 1985;5:82730.
62. de Franchis R, Vitagliano P, Agape D, et al. Eradication of esophageal varices by endoscopic
sclerotherapy: How much is enough?. Gastrointest Endosc 1988;34:3959.
63. Agrawal BK, Suman A, Singh V, Kumar P. End point for sclerosis of esophageal varices.
Indian J Gastroenterol 1993;12:835.
64. Stray N, Jacobsen CD, Rosseland A. Injection sclerotherapy of bleeding oesophageal and
gastric varices using a flexible endoscope. Acta Med Scand 1982;211:1259.
66. Korula J, Chin K, Ko Y, Yamada S. Demonstration of two distinct subsets of gastric varices.
Observations during a seven-year study of endoscopic sclerotherapy. Dig Dis Sci
1991;36:3039.
67.
Sarin SK, Sachdev G, Nanda R, et al. Gastric varicesCourse and treatment with
endoscopic sclerotherapy [Abstract]. Hepatology 1987;7:1062.
68. Soehendra N, Nam V Ch, Grimm H, Kempeneers I. Endoscopic obturation of large
esophagogastric varices with bucrylate. Endoscopy 1986;18:256..
69. Lo GH, Lai KH, Ng WW, et al. Injection sclerotherapy preceded by esophageal tamponade
versus immediate sclerotherapy in arresting active variceal bleeding: A prospective
70. Siringo S, McCormick PA, Mistry P, et al. Prognostic significance of the white nipple sign in
variceal bleeding. Gastrointest Endosc 1991;37:515.
71. Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by esophageal
72.
Kravetz D, Korula J, Mack W. Intravariceal pressure and variceal characteristics between
variceal bleeders and non-bleeders: Preliminary results. Hepatology 1993;104:A930.
73.
Rose JDR, Smith PM, Crane MD. Radiological control of oesophageal sclerotherapy: An
improved technique. Gut 1982;23:915.
74. Korula J, Pandya K, Yamada S. Perforation of esophagus after endoscopic variceal
sclerotherapy. Incidence and clues to pathogenesis. Dig Dis Sci 1989;34:3249.
75. Sarin SK, Sachdev G, Nanda R. Follow-up of patients after variceal eradication. Ann Surg
1986;204:7882.
76.
Terblanche J, Kahn D, Campbell JAH, et al. Failure of repeated injection sclerotherapy to
improve long-term survival after oesophageal variceal bleeding. Lancet 1983;II:132832.
77. Sarin SK, Nanda R, Vij JC, Anand BS. Oesophageal ulceration after sclerotherapyA
complication or an accompaniment?. Endoscopy 1986;18:445.
78.
Korula J, Kravetz D. Are intravariceal pressures useful in predicting efficacy of endoscopic
sclerotherapy?. Gastroenterology 1993;104:A929..
Chapter 33 Indications and Results of Variceal Injection Sclerotherapy

(2090)
(2091)
DR. DAVID WESTABY, M.A., F.R.C.P.
Endoscopic injection of a sclerosing agent is the most widely used method for treatment of bleeding
esophageal varices. The technique has been used since the 1940s, but almost 40 years were required
before injection sclerotherapy became firmly established.1(2092) The technique has been applied to
the management of active variceal bleeding, for the prevention of recurrent bleeding, and as primary
prophylaxis against a first variceal hemorrhage.
Many controlled trials of injection sclerotherapy have been carried out, perhaps more than in any other
area of endoscopy or hepatology. The quality of such trials has improved considerably in recent years
and has been further enhanced by the application of the statistical technique of metaanalysis, which
facilitates pooling of data.2(2093)
This chapter describes injection sclerotherapy for active variceal bleeding and for the prevention of
recurrent hemorrhage. Prophylactic use of the technique is considered in Chapter 34: Prophylactic
Sclerotherapy.
Active Variceal Bleeding

Two large, uncontrolled series were the first to provide substantial support for the use of injection
sclerotherapy for variceal bleeding.3,4(2094) Bleeding was controlled for the period of admission in
more than 90% of the patients in both studies. However, many of those included had stopped bleeding
spontaneously or had received balloon tamponade as initial therapy. The study results reflect the
combined approach to treatment and do not provide specific information about the value of injection
sclerotherapy alone, particularly in patients who continue to bleed actively.
Subsequent controlled trials compared injection sclerotherapy with conservative measures such as
balloon tamponade, vasoconstrictor therapy, or both methods.512(2095) In most trials, injection
sclerotherapy was shown to be superior to conservative measures for the period of admission. This
was not reflected in a significant survival benefit, although a trend was evident.
Sung et al.13(2096) compared emergency sclerotherapy (3% sodium tetradecyl sulfate) with infusion
of octreotide (50 g intravenous bolus plus 50 g/hr intravenous infusion for 48 hr) in 100 evenly
matched patients with acute variceal hemorrhage. At 48 hr after admission, patients treated with
octreotide underwent sclerotherapy. There were no significant differences between the two groups with
respect to initial control of bleeding (90% for sclerotherapy vs. 84% for octreotide), recurrent bleeding
during the 48 hr after admission, blood transfusion requirements, length of hospital stay, or in-hospital
mortality. In the randomized trial of Planas et al.,14(2097) 35 patients with acute variceal bleeding were
randomized to treatment with somatostatin, and 35 similar patients underwent sclerotherapy. There
were no significant differences between the two groups with respect to numbers of patients with
persistent bleeding or recurrent bleeding during the 48-hr trial period or in 6-week mortality rates. Major
complications, including two that were fatal, occurred in 5 patients in the sclerotherapy group but in
none of those who received somatostatin.
Immediate sclerotherapy was compared with initial balloon tamponade followed by sclerotherapy in the
randomized study of Lo et al.15(2098) that included 60 patients with acute esophageal variceal
hemorrhage. Immediate sclerotherapy proved to be impossible in 3 patients (10%) with acute
hemorrhage. Although the two approaches to treatment were similar with respect to initial hemostasis,
recurrent bleeding, and mortality, transfusion requirements were significantly reduced for patients who
underwent immediate sclerotherapy. Major complications occurred with significantly less frequency in
the immediate-sclerotherapy group.
Although the many trials of injection sclerotherapy compared with conservative methods of treatment
document the efficacy of sclerotherapy, the comparison with conservative measures, such as balloon
tamponade and vasoconstrictor therapy, may not be strictly valid. These conservative measures are
only effective during the period of application, unlike injection sclerotherapy, which almost certainly
reduces the early recurrence of bleeding. When comparisons are extended over the full period of
admission, a bias toward injection sclerotherapy is inevitably seen.
Immediate Versus Delayed Sclerotherapy

The initial controlled trials of sclerotherapy did not answer the important question of how to optimize the
use of injection sclerotherapy. Of particular importance is the timing of intervention with respect to the
active bleeding episode. The options lie between immediate treatment, with the associated risk of
complications and technical difficulties, and initial delay to allow vasoconstrictor drugs or balloon
tamponade to provide temporary hemostasis.
The first evidence in support of immediate injection sclerotherapy came from a study comparing this
approach with a historical control group in which sclerotherapy was delayed to allow time to establish
hemostasis using conservative measures.16(2099) Bleeding was controlled for the first 48 hr of
admission in almost 90% of those receiving immediate sclerotherapy, compared with 64% of the
control group (p = .05).
Subsequently, the results of several randomized controlled trials comparing immediate with delayed
sclerotherapy were reported. In the first of these, immediate sclerotherapy was compared with initial
treatment with a regimen of vasopressin and nitroglycerin.17(2100) Control of bleeding was assessed
over the first 12 hr after randomization, and the patients from both groups were then entered into a
long-term program of injection sclerotherapy. Immediate treatment was associated with an almost 90%
control rate, compared with 65% for those treated initially by the vasoconstrictor regimen (p = .05). The
frequency of recurrent bleeding during the period of admission was similar for both groups (31%),
reflecting the use of long-term sclerotherapy in all patients. Blood transfusion requirements were less in
the immediate sclerotherapy group, but overall admission mortality was similar for both groups.
In another controlled trial encompassing a total of 134 bleeding episodes, patients were randomized to
receive immediate or delayed sclerotherapy.18(2101) The latter patients were initially managed with
the use of vasoconstrictor drugs, balloon tamponade, or both approaches. Immediate injection
sclerotherapy controlled active bleeding in 95% of patients, compared with 83% when it was delayed (p
= .03), but no survival benefit resulted. This failure to improve survival by treatment measures that
enhance the control of bleeding is disappointing, but this finding has been universal in similar trials of
other treatment measures for active variceal bleeding.
Sclerotherapy for immediate control of active bleeding by injection around a bleeding point was
compared with immediate plus long-term sclerotherapy in the study of Moreto et al.,19(2102) in which
56 patients were randomized to the former treatment and 50 to the latter. There was no significant
difference in the overall cumulative proportion of patients with recurrent bleeding, although there were
fewer episodes of hemorrhage in the combined-therapy group. Immediate plus long-term sclerotherapy
was more effective in preventing recurrent bleeding from an esophageal source than from a
gastroesophageal (i.e., junctional) source. Mortality rates were the same for both groups.
Most deaths associated with an episode of variceal bleeding occur in patients with decompensated
liver disease, and it appears that it is the bleeding event itself that poses the major risk of subsequent
mortality, outweighing any beneficial influence of early therapy. Although immediate injection
sclerotherapy cannot be justified on the grounds of influencing mortality, it may still remain the optimum
treatment. Injection sclerotherapy offers the advantage of a single therapy that may be considered as
an integral part of the diagnostic endoscopy with the expectation of minimizing the resources
necessary for resuscitation, particularly blood transfusion requirements.
If injection sclerotherapy is used as the primary management of an episode of variceal bleeding, it is
important to define failure of treatment. Successful sclerotherapy is apparent almost immediately
(certainly within the first 2 to 3 hr after the procedure), and any ongoing bleeding thereafter should be
considered as failed therapy. The bleeding may signify inadequate treatment of the bleeding point or
the failure to recognize the correct source of hemorrhage. Fundal gastric varices are a common cause
of the latter, and any further endoscopic intervention must include a thorough diagnostic endoscopy. If
active bleeding continues after a second attempt at therapy, an alternative measure should be sought.
To minimize morbidity and mortality, the whole process of defining failure and directing subsequent
intervention should be carried out over a period of no more than 6 to 8 hr.20(2103)
It is much more difficult to define failure of injection sclerotherapy over the subsequent days and weeks
after the presenting variceal hemorrhage has been controlled. Early recurrence of bleeding is observed
in as many as 50% of patients after injection sclerotherapy; most of these episodes can be managed
by further endoscopic treatment. In such circumstances, failure of treatment is usually restricted to
those who have had a single life-threatening episode or multiple smaller episodes of recurrent
bleeding.
Sclerotherapy Versus Surgery

Several studies have attempted to investigate the relative merits of injection sclerotherapy and surgical
measures for the management of an episode of variceal bleeding.
Portacaval Shunt
Only one trial has compared injection sclerotherapy with the portacaval shunt, and it was restricted to
patients with decompensated liver disease. Sixty-four patients with Child class C liver disease and
acute variceal bleeding requiring six or more units of blood were randomized in the trial of Cello et
al.21(2104) to undergo intravariceal sclerotherapy with 5% sodium morrhuate (n = 32) or portacaval
shunt (n = 32). Treatment was initiated at from 1 to 6 hr after randomization. Transfusion requirements
were significantly lower, and the length of hos-pitalization was significantly shorter for sclerotherapy
patients during the initial hospitalization, although the rate of recurrent hemorrhage was higher. Early
mortality was extremely high (almost 55% in both groups), although survival at 30 days was similar.
The advantages of sclerotherapy during the early phase of the trial were less evident after the index
hospitalization. The rate of rehospitalization, total days of hospitalization, and units of blood transfused
were significantly higher for sclerotherapy patients. Sclerotherapy had a significant cost benefit during
the initial hospitalization, but the long-term overall costs of the two treatment methods equalized
because of the costs of treating recurrent episodes of variceal bleeding in the sclerotherapy group.
Seven of 14 sclerotherapy patients who were discharged from the hospital after treatment had severe
or repeated episodes of bleeding and were considered treatment failures. All 7 patients underwent
surgery (30-day mortality rate of 14%). At a mean follow-up of 530 days, survival was the same in both
treatment arms of this trial. Cello et al.21(2105) concluded that sclerotherapy and portacaval shunt
surgery were equally effective in poor-risk patients with massive bleeding and that elective shunt
surgery should be considered for patients who fail sclerotherapy.
Esophageal Transection
Three trials compared injection sclerotherapy with esophageal transection and devascularization
procedures.2224(2106) In two of these trials, esophageal transection significantly reduced the risk of
early recurrence of bleeding,22,23(2107) and in the third, results were similar for both treatment
groups.24(2108) No survival benefit was observed for either treatment group.
In the trial of Burroughs et al.,22(2109) 101 patients were randomized to receive sclerotherapy or
surgical transection of the esophagus. Patients were enrolled in the study if conservative measures
failed to control bleeding within 5 days of the index episode. At 6 and 12 months, there was no
significant difference in survival. The severity of encephalopathy, ascites, prolongation of prothrombin
time, and balloon tamponade were independently associated with death at 6 weeks. With surgical
intervention, significantly more patients remained free of bleeding for a period of 5 days after initiation
of therapy. Sclerotherapy failed to control bleeding (three sessions) in nine patients, five of whom
underwent esophageal transection. Although bleeding was controlled in all five, only one of these
patients survived. Burroughs et al.22(2110) thought that esophageal transection was superior to a
single sclerotherapy session for prevention of recurrent bleeding. They also recommended that
surgical salvage be attempted after a second episode of recurrent bleeding.
Huizinga et al.23(2111) reported a prospective, randomized trial of sclerotherapy in 37 patients and
esophageal transection in 39 patients. In the group of patients who underwent sclerotherapy, 48.6%
had early recurrence of bleeding, compared with 2.5% of patients who had surgery, a highly significant
difference. None of the operated patients died as a result of bleeding, but 6 patients who underwent
sclerotherapy died of bleeding. The patients who underwent surgery had no episodes of bleeding
during the later phases of the trial, but the 13 patients who underwent sclerotherapy had 22 episodes
of bleeding. The 30-day mortality rates were similar: 33% for surgically treated patients and 24.3% for
those who had sclerotherapy. For the 15 patients with liver disease classified as Child class B, there
were no operative deaths. However, 13 of 23 patients with Child class C disease died as a result of the
surgery.
Consecutive patients with persistent or recurrent variceal bleeding in whom the operative risk was
considered high were randomized to undergo sclerotherapy (n = 15) or esophageal transection with
stapler anastomosis (n = 17) in the trial of Teres et al.24(2112) Control of bleeding and long-term
survival were similar for both groups, although sclerotherapy was associated with fewer complications.
Staple transection of the esophagus has been employed as a "salvage" procedure when sclerotherapy
fails to control bleeding. McCormick et al.25(2113) performed this operation on an emergency basis in
22 patients with persistent variceal bleeding despite injection therapy. Bleeding was controlled in 90%
of these patients. In-hospital mortality was high, although 4 of 10 patients with functionally advanced
liver disease survived.
The interest in transjugular intrahepatic portosystemic stent shunting (TIPS) has led to speculation on
the relative merits of this treatment compared with sclerotherapy.26(2114) Controlled trials comparing
these two treatments are lacking, but the lack of availability of TIPS probably represents a major
restraint on the use and value of TIPS for the management of variceal hemorrhage.
Recurrent Variceal Bleeding

The use of repeated courses of injection sclerotherapy to obliterate esophageal varices and prevent
recurrent bleeding has been extensively investigated in a number of controlled
trials.6,7,12,2732(2115) On an individual basis, all of the trials have shown a benefit for injection
sclerotherapy with respect to recurrent bleeding, although the median rate was in excess of 50%,
underlining one of the major limitations of this approach.
The multicenter, prospective, single-blind, randomized trial of the Veterans Affairs Cooperative Variceal
Sclerotherapy Group32(2116) compared initial and long-term sclerotherapy to sham sclerotherapy in
122 and 131 similar patients, respectively, with alcoholic liver disease. Patients had a history of variceal
bleeding or were bleeding from esophageal varices at randomization. The rate of recurrent bleeding,
number of episodes of bleeding due to esophageal varices, mean transfusion requirement, and need
for shunt surgery were significantly higher for the sham-sclerotherapy group of patients. Mortality rates
were the same for both groups.
In an attempt to clarify the data from these numerous, divergent trials, the statistical tool of
metaanalysis has been applied.2(2117) Two such analyses have been report-ed.33,34(2118) Both
analyses confirmed a significant benefit for injection sclerotherapy with respect to recurrent bleeding
and survival; both of these endpoints improved by approximately 50% in the sclerotherapy-treated
groups (Table 331 and Figure 331).
Controlled Trials of Long-Term Injection

Sclerotherapy Included in the Metaanalysis*
TABLE 331
FIRST AUTHOR
1.
2.
3.
4.
5.
6.
7.
8.
*
Barsoum
Terblanche
Copenhagen project
Westaby
Soderlund
Paquet
Korula
Burroughs
See also Figure 331.
YEAR OF PUBLICATION
1982
1983
1984
1985
1985
1985
1985
1989
REFERENCE
27
28
29
30
7
6
12
31
(2119)Figure 331. Metaanalysis of the eight controlled trials of long-term injection

sclerotherapy listed by number in Table 331. The relative risk and 95% confidence intervals
are expressed for the outcome. A, Recurrent bleeding. B, Death for each trial and also for the
pooled data. (A and B, From Pagliari L, Burroughs AK, Sorensen TA, et al. Therapeutic
controversies and randomised controlled trials: Prevention of bleeding and rebleeding in
cirrhosis. Gastroenterol Int 1989; 2:7184.
Despite the evidence offered by metaanalyses in favor of long-term injection sclerotherapy, an
important area of controversy remains concerning treatment measures permitted for patients in control
groups who presented with recurrent bleeding. In most trials, this therapy was restricted to
conservative measures, such as vasoconstrictor drugs or balloon tamponade. All the trials showing a
significant survival benefit for injection sclerotherapy used a protocol encompassing these features. In
contrast, no survival benefit was observed in the two trials in which a single session of injection
sclerotherapy was permitted for recurrent bleeding episodes in the control groups.28,31(2120) It
therefore was suggested that restricting sclerotherapy to episodes of recurrent bleeding alone has the
same survival advantage as a long-term regimen. The available data cannot definitively answer this
question, and researchers are reevaluating long-term treatment regimens. If a survival benefit does not
accrue from long-term injection sclerotherapy, justification for its use would depend on the extent of the
reduction in the frequency of recurrent bleeding and associated cost analysis.
The influence of sclerotherapy on survival has been less easily defined, with only four of the eight trials
confirming a significant benefit.6,2730(2121) The problematic nature of assessing survival is
illustrated by the Copenhagen randomized trial of sclerotherapy versus medical therapy.29(2122)
Overall results showed no difference in survival between the two patient groups. However, survival was
significantly improved for the group of patients undergoing sclerotherapy if other factors were
accounted for, such as the presence of ascites or encephalopathy. Mortality was somewhat lower for
patients undergoing sclerotherapy during the first 10 days after randomization, but it was higher
between days 10 and 40. The differences in mortality during the two intervals were not statistically
significant. The survival curves did not disclose any overall difference in mortality, although late survival
was significantly improved in the patients undergoing sclerotherapy.
In the study of Westaby et al.,30(2123) 60 patients received medical therapy, and 56 patients
underwent sclerotherapy. A statistically significant reduction in mortality was found (32% vs. 18%) after
a median follow-up of 37 months. Death from hemorrhage occurred significantly more often in the
group that received medical therapy (25% vs. 5%). However, no significant difference was found in
survival in the trial of Korula et al.12(2124) in which 63 patients were treated by sclerotherapy and 57
patients received conservative treatment. No improvement in survival was found in the trial of
Terblanche et al.28(2125) in which 37 patients underwent long-term sclerotherapy and 38 patients
were managed by conservative means. In this trial, however, acute bleeding episodes in the latter
group of patients were treated by sclerotherapy.
Sclerotherapy Versus Shunt Surgery

Portosystemic shunt surgery has elicited considerable interest throughout the last 2 decades. Six trials
have compared shunt surgery with long-term injection sclerotherapy. Four of these used the selective
distal splenorenal shunt,3538(2126) and the remaining two used a portacaval
anastomosis.21,39(2127) All these studies confirmed shunt surgery as more effective than
sclerotherapy for preventing recurrent bleeding. Although only one trial showed a significant increase in
encephalopathy after shunt surgery compared with injection sclerotherapy, a trend was evident in the
majority.
Eighty-two patients with Child-Campbell class A and B liver dysfunction were randomized to undergo
end-to-side portacaval shunt or sclerotherapy in the trial of Planas et al.39(2128) Hemodynamic
stability for at least 3 days was required before randomization, and 31 patients with persistent or
recurrent hemorrhage were excluded for this reason. Recurrent hemorrhage was significantly more
frequent in sclerotherapy patients (40% vs. 2.9%), but encephalopathy was significantly higher for
those who underwent surgery (40% vs. 12%). However, neurologic symptoms were incapacitating in
8.8%. Cost, complication rates, days of hospitalization, and long-term mortality were equivalent.
In contrast to the trial of Cello et al.,21(2129) the 112 patients in the randomized trial of Teres et
al.36(2130) had Child class A or B liver dysfunction, and all were stable at randomization, which
occurred 10 to 15 days after the index episode of variceal bleeding. Fourteen of the 57 patients
randomized to surgery (i.e., distal splenorenal shunt) were excluded for a variety of reasons (e.g.,
venous thrombosis, inadequate splenic vein size, medical contraindications). Four of 55 patients
initially randomized to sclerotherapy were excluded because of bleeding gastric varices or massive
esophageal bleeding. Survival at 2 years was similar: 71% for surgery and 68% for sclerotherapy.
Recurrent bleeding was significantly more frequent in the sclerotherapy group (37.5% vs. 14.3%), but
encephalopathy was significantly more common in those who underwent surgery (24% vs. 8%). As
observed in other trials, days in the hospital as a consequence of the index episode of bleeding were
significantly fewer for patients undergoing sclerotherapy, but on a long-term basis, days in the hospital
were equal because of the higher rates of recurrent bleeding and complications in the sclerotherapy
group. Teres et al.36(2131) concluded that the applicability of sclerotherapy was wider based on the
numbers of patients excluded from both treatment arms of the trial.
The final results of a randomized, controlled trial comparing sclerotherapy and distal splenorenal shunt
were reported by Henderson et al.37(2132) In contrast to the trial of Teres et al.,36(2133) patients with
Child class C liver disease were included, and most were randomized within 2 weeks of the index
episode of bleeding. The rate of recurrent bleeding was significantly higher in the sclerotherapy group
(39% vs. 3%). Surgery was performed in seven sclerotherapy patients who were classified as
treatment failures. There was no increase in operative mortality for these patients. The estimated
4-year survival was significantly better for patients undergoing sclerotherapy (median follow-up of 61
months). This survival advantage remained significant for poor-risk and good-risk patients. For patients
with alcoholic cirrhosis, sclerotherapy was associated with a significant survival advantage; for patients
with nonalcoholic liver disease, neither method of treatment provided a survival advantage. The
investigators regarded so-called salvage or rescue surgery in patients who failed sclerotherapy as an
important factor in the survival advantage found for patients in the sclerotherapy group.
The trial of Rikkers et al.35(2134) included 30 patients who underwent elective shunt surgery (i.e.,
distal splenorenal in 26, nonselective in 4) and 30 patients who were randomized to receive long-term
sclerotherapy. Most patients (86%) had alcoholic cirrhosis, and one third had Child class C liver
dysfunction. Recurrent bleeding was significantly less frequent in patients who underwent surgery
(17%) compared with those treated by sclerotherapy (60%). The development of hepatic
encephalopathy was similar for both groups.
Mortality data are conflicting. Four of the six controlled studies showed a trend toward lower mortality
after shunt surgery or no difference whatsoever.21,36,38,39(2135) Survival at 6 years of follow-up was
significantly better for shunt patients (53%) than for those who underwent sclerotherapy (26%) in the
trial of Rikkers et al.35(2136) The trial of Henderson et al.37(2137) demonstrated a clear survival
benefit for the sclerotherapy group, which was attributed to the maintenance of liver function in patients
undergoing the endoscopic technique. Henderson et al.37(2138) emphasized the importance of shunt
surgery as a rescue procedure for patients who fail sclerotherapy. From the same study, it was
possible to identify patients with alcoholic cirrhosis as those most likely to benefit from sclerotherapy
compared with the distal splenorenal shunt. Salvage shunt surgery was possible in relatively few
patients in the trial of Rikkers et al.35(2139) owing to wide geographic distribution.
Transjugular Intrahepatic Shunt

Data are now available from four randomized, controlled trials of TIPS versus endoscopic variceal
sclerotherapy.40,41,41a,41b(2140)
In the trial of elective therapy of Cabrera et al.,40(2141) in which 63 patients were randomized within 3
days of the onset of bleeding to either TIPS (n = 31) or sclerotherapy (n = 32), there was a significant
difference in the rate of recurrent variceal bleeding in favor of TIPS (23% vs. 51.6%) at a mean
follow-up of 15 months. Episodes of bleeding had occurred before randomization in 32 patients. In 10
patients in the sclerotherapy group, recurrent bleeding was considered uncontrollable and 9 of these
patients underwent TIPS. One third of patients who underwent TIPS developed encephalopathy, or in
some cases encephalopathy worsened, versus 13% of those who had sclerotherapy. There was no
significant difference in survival between the two groups.
Cello et al.41(2142) randomized 49 patients to TIPS (n = 24) or sclerotherapy (n = 25) within 24 hours
of the onset of bleeding. At respective mean follow-up periods of 575 and 567 days, the frequency of
recurrent bleeding was significantly less for patients who underwent TIPS (12.5% vs. 48%). Six
patients (24%) in the sclerotherapy group underwent TIPS for recurrent variceal bleeding. Although a
trend toward improved survival for TIPS-treated patients was noted, the difference was not statistically
significant. In contrast to the other trials, there were no significant differences in the new onset or
worsening of encephalopathy that existed before randomization. Other outcome parameters assessed
in this trial for which there were no significant differences between the two treatment groups included
total days of hospitalization for variceal bleeding and total health care costs.
At approximately 10 days after an episode of bleeding, 41 patients were randomized to TIPS and 39 to
sclerotherapy in the trial of Sanyal et al.41a(2143) At a follow-up of approximately 1000 days, the rates
of recurrent variceal bleeding were approximately the same (22% TIPS vs. 20.5% sclerotherapy). Six
patients (15%) in the sclerotherapy group underwent TIPS for recurrent bleeding. Encephalopathy
developed in 29.3% of patients who underwent TIPS. The mortality rate was significantly higher in the
TIPS group. In contrast to the studies by Cabrera et al.40(2144) and Rossle et al.41b(2145) in which
there were no deaths as a result of variceal bleeding in the TIPS treatment group, 5 patients in the
TIPS group (vs. 3 who underwent sclerotherapy) died as a result of bleeding in the study of Sanyal et
al.41a(2146)
In the trial of Rossle et al.,41b(2147) 61 patients were randomized to undergo TIPS and 65 to have
sclerotherapy or band ligation, or both, combined with propranolol therapy within 24 hours of the onset
of bleeding. Episodes of bleeding had occurred prior to randomization in 61% of patients in the study.
The cumulative 1-year and 2-year rates of recurrent bleeding were significantly less in the TIPS group
compared with sclerotherapy (15% vs. 41%, and 21% vs. 52%, respectively). At 1 year the rate of
"clinically significant" encephalopathy was considerably higher in the TIPS group (36% vs. 18%), but
there was no significant difference in mortality between the two groups.
There are both similarities and differences in the results of these four trials.41c(2148) It is reasonable
to conclude that the treatments provided to patients, TIPS as well as sclerotherapy, were not strictly
comparable because of differences in level of operator skill and experience, techniques, and the
additional use of other treatments such as variceal banding and prophylactic pharmacotherapy. From
32 to 61% of patients had survived prior episodes of bleeding. Presumably, many of these episodes
had been treated by endoscopic means. In all four studies, therefore, each treatment group was
composed of selected patients for whom the natural history of variceal bleeding is likely to be different
from that for a group of patients with new-onset variceal bleeding. In all of these studies, recurrent
bleeding was the primary endpoint and all except that of Sanyal et al.41a(2149) demonstrated a
significantly higher rate in patients who underwent sclerotherapy. Jalan et al.41c(2150) offer several
possible explanations for the result reported by Sanyal et al.,41a(2151) including a selection bias as a
result of a long lead time between randomization and treatment and a failure to adequately reduce
portal pressure in some patients in the TIPS group.
New-onset encephalopathy or evidence of worsening encephalopathy was observed in about one third
of patients undergoing TIPS compared with about 15% of patients treated by sclerotherapy. The
difference between the two treatment groups with regard to encephalopathy was significant in all
studies except that of Cello et al.41(2152) As pointed out by Jalan et al.,41c(2153) this result from the
trial of Cello et al. might be explained by a failure to detect milder degrees of encephalopathy at the
point of randomization. In fact, the only trial in which strict criteria for assessing encephalopathy were
utilized was that of Rossle et al.41b(2154)
In all four studies, TIPS was performed when uncontrolled bleeding developed in patients randomized
to sclerotherapy. The definition of "uncontrolled" bleeding is not always clear in these studies and is
likely to vary from study to study. Sclerotherapy in expert hands is very effective for control of active
variceal bleeding. The fact that recurrent bleeding could not be controlled by sclerotherapy in up to
28% of patients appears unusual and therefore unexpected. Nevertheless, the use of TIPS in patients
in whom sclerotherapy had failed to control bleeding makes it difficult to compare survival between
both groups of patients. It is possible that some of the patients with uncontrolled variceal bleeding who
had been randomized to sclerotherapy would have died as a result of hemorrhage if TIPS had not
been performed.
Percutaneous Transhepatic Variceal Obliteration

Endoscopic intravariceal sclerotherapy was found to be superior to percutaneous transhepatic variceal
obliteration for prevention of recurrent variceal bleeding and mortality in patients with nonalcoholic
cirrhosis treated electively in the trial of Terabayashi et al.42(2155)
-Receptor Blockade
Initial reports of the use of propranolol in good-risk patients with cirrhosis and portal hypertension
indicated that this agent reduced the frequency of recurrent variceal bleeding.43,44(2156) A favorable
effect on variceal bleeding has been demonstrated in some45(2157) but not all subsequent
studies.46,47(2158) Nevertheless, early reports of the efficacy of -adrenergic receptor blockade as
long-term management for recurrent variceal bleeding prompted a number of studies comparing this
pharmacotherapy to sclerotherapy.4858(2159) In general, these trials can be divided into two types:
those comparing sclerotherapy alone to propranolol alone and those that compare sclerotherapy plus
propranolol to propranolol alone. The second of these groups is considered in a following section.
Seventy patients with at least one episode of variceal bleeding were randomized by Fleig et al.48(2160)
to receive treatment with propranolol or sclerotherapy (1% polidocanol) until variceal eradication was
achieved. At a mean follow-up of 14 months, recurrent bleeding had occurred in 28% of the
sclerotherapy-treated patients and in 29% of the propranolol-treated group at a mean follow-up of 9.2
months. No significant difference in mortality rates were observed for the two groups.
Rossi et al.54(2161) randomized 79 patients with cirrhosis (Child-Pugh class A 22%, class C 38%) to
receive propranolol, to receive weekly sclerotherapy (1% polidocanol) sessions, or to a control group.
No significant differences were observed in the endpoints of death or recurrent bleeding for the three
groups. The cumulative percentages of patients free of recurrent bleeding at 1-year follow-up for the
propranolol, sclerotherapy, and control groups were 81%, 64%, and 54% respectively, differences that
did reach reaching statistical significance. The authors noticed a trend toward reducing recurrent
bleeding in favor of propranolol and attributed the lack of statistical significance to the poor statistical
power of their study.
Long-term sclerotherapy and propranolol were compared in another trial in which 108 patients
(Child-Pugh class A or B) were randomized after spontaneous cessation of an index episode of
variceal bleeding.50(2162) One session of sclerotherapy was permitted if active variceal hemorrhage
occurred in a patient receiving propranolol. A series of sclerotherapy sessions was commenced if a
patient in the -receptor blockade group experienced an episode of variceal hemorrhage that required
more than 10 units of blood or if the number of episodes of bleeding exceeded three while receiving
propranolol. Analysis of data was based on intention-to-treat criteria. No significant differences
between the two groups of patients were found for recurrent bleeding or mortality during follow-up
periods that ranged from 1 to more than 5 years. One session of sclerotherapy was performed in 31%
of patients receiving propranolol; 26% underwent a series of sclerotherapy sessions.
The trial of Qureshi et al.56(2163) is similar to that of Rossi et al.54(2164) in that patients with portal
hypertension and variceal bleeding were randomly allocated to three groups: oral propranolol (n = 47),
sclerotherapy (n = 51), and conservative management that served as the control (n = 41). No
significant differences in recurrent bleeding were seen in the three groups at 6 weeks' follow-up. At the
6-month and 1-year follow-up evaluations, however, the frequency of recurrent bleeding was
significantly less for patients receiving propranolol than for the other two treatment groups; at 2 years'
follow-up, the frequency of recurrent bleeding was significantly better for the propranolol and the
sclerotherapy groups compared with controls. The results of this study, which clearly demonstrate
efficacy for propranolol in preventing recurrent bleeding, differ from those of other studies. The
investigators speculated that this could be attributable to substantial differences in the causes of portal
hypertension.
In the trial of Teres et al.,57(2165) consecutive patients with cirrhosis (predominantly Child-Pugh class
A and B) and variceal bleeding were randomized to undergo sclerotherapy (intravariceal 5%
ethanolamine at weekly intervals) until variceal obliteration or to receive propranolol. Only
hemodynamically stable patients were randomized, in all cases within 1 week of the initial bleeding
event. Twenty-six of 58 patients undergoing sclerotherapy had recurrent bleeding, compared with 37 of
58 patients who received propranolol. Although the probability of recurrent bleeding was lower for
patients who underwent sclerotherapy, no differences between the two groups were observed for
hospitalization, survival, cause of death, and bleeding index. The latter calculation (i.e., months of
follow-up divided by number of episodes of bleeding plus one) reflected the time free of recurrent
bleeding during follow-up. The value of sclerotherapy in terms of bleeding was further offset by a
higher frequency and severity of complications.
In the prospective, randomized controlled trial of Kanazawa et al.,58(2166) patients with cirrhosis and
variceal bleeding, evenly matched with respect to severity of liver disease, size of esophageal varices,
and degree of portal hypertension, were randomized to undergo sclerotherapy alone (n = 23) or oral
propranolol after elective sclerotherapy (n = 20). Patients were randomized after initial control of
bleeding. At 1 and 2 years, mortality was the same for both groups, but the cumulative percentages of
patients who had not had recurrent episodes of bleeding were significantly higher in the group
receiving propranolol.
A somewhat different approach was taken by Lo et al.,59(2167) who studied the effect of propranolol
on the recurrence of varices after obliteration by sclerotherapy. Patients were randomized to
propranolol (n = 30) or control (n = 29) groups. After a mean follow-up of 28 months, no significant
differences between the two groups were found for recurrence of varices, variceal bleeding, survival, or
death from massive variceal hemorrhage. Gastric (cardia) varices, a common source of recurrent
bleeding, developed in 15% of patients receiving propranolol, compared with 11% of the control group.
Although -receptor blockade with propranolol is uniformly associated with a high frequency of
recurrent bleeding, the incidence was substantially higher than that seen in the sclerotherapy group in
only one trial.52(2168) In this trial, carried out specifically with patients who had more severe liver
disease (Child-Pugh class B and C), the number of patients who experienced a recurrent episode of
bleeding, the number of such episodes overall, the number of hospitalizations for recurrent bleeding,
and the blood transfusion requirements were all significantly higher in the group of patients (n = 46)
who received propranolol than those who underwent sclerotherapy (n = 45). A survival benefit was also
observed for the sclerotherapy group; this was the only trial to do so.
Noncirrhotic Portal Hypertension

Extrahepatic Portal Venous Obstruction
The most common causes of portal hypertension and variceal bleeding in pediatric patients are
thrombosis, fibrosis, and congenital malformation of the portal venous system (see Chapter 53:
Esophagogastroduodenoscopy in the Pediatric Patient). However, variceal bleeding due to
extrahepatic portal vein obstruction may appear in some patients only after they reach adulthood.
Surgical treatment of this condition has met with only limited success. Shunting procedures are often
precluded in children owing to the small size of vascular structures, and Fonkalsrud et al.60(2169)
found that more than half of their 164 operations for this condition were performed to correct problems
created by previous operations.
Injection sclerotherapy is reported to be especially effective in patients with portal hypertension and
variceal bleeding due to obstruction of the portal vein of various causes.6163(2170) These results are
undoubtedly influenced by the fact that most of these patients have normal liver function.
Kahn et al.62(2171) reported a 15-year experience of 362 sclerotherapy sessions using ethanolamine
oleate in 55 adult and teenage patients with extrahepatic portal venous obstruction and variceal
hemorrhage. Variceal eradication was achieved in 44 patients after a median of six injection sessions
over an average of 12.5 months. Before eradication, 18 episodes of variceal bleeding occurred in 11
patients. Mean follow-up for this series of patients was 6.8 years, with a range of 1.1 to 14.6 years.
Seven bleeding episodes attributed to variceal recurrence after eradication were observed in 6
patients. There were 4 deaths during the study period. Similar favorable results were reported by
Bhargava et al.63(2172) in 43 patients who underwent intravariceal sclerotherapy with 1% polidocanol.
Extrahepatic portal vein obstruction accounts for a high percentage of episodes of variceal bleeding in
certain countries such as India. Chawla et al.64(2173) reported their experience with sclerotherapy
(intravariceal injections of 3% sodium tetradecyl sulfate) in 122 such patients with a mean age of 17.2
years (range 1.5 to 55 years). All patients presented with bleeding from large-diameter varices.
Bleeding stopped spontaneously in about three fourths, and balloon tamponade was employed to
achieve initial hemostasis in the remaining patients. Eighteen patients were lost to follow-up, 3
underwent operation, and 6 died. Variceal obliteration was achieved in 95 patients and required a
mean of 5.4 sessions. Before eradication, 17 episodes of bleeding occurred in 15 patients. Esophageal
varices recurred in 15 patients; large gastric varices developed in 10 patients after eradication of
esophageal varices. Serious complications occurred more frequently in children. Sclerotherapy was
regarded as having a favorable effect on variceal bleeding because the number of such episodes after
initiation of injections was less than that before sclerotherapy.
Pregnancy is generally considered to increase the risk of bleeding in patients with portal hypertension
due to extrahepatic portal venous obstruction. Although this condition typically arises in childhood,
increasing numbers of patients are surviving into adulthood. Few data are available concerning the
outcome of sclerotherapy in these patients before conception or during gestation. Most available
information is derived from case reports in which the outcome was usually satisfactory.6568(2174)
Noncirrhotic Portal Fibrosis

Noncirrhotic portal fibrosis is a common cause of portal hypertension and variceal hemorrhage in
certain parts of the world such as the subcontinent of Asia. Bhargava et al.69(2175) reported results of
sclerotherapy (intravariceal 1% polidocanol) in 60 such patients. Eradication was achieved in 88% of
patients after a mean of 8.43 injection sessions; the condition recurred in 15% of patients at a mean
follow-up of 19 months. The complication rate was relatively low (12%). A favorable response to
treatment was indicated by comparisons of parameters of bleeding before and after sclerotherapy.
Survival was related to liver function status.
Comparison Studies
The results of sclerotherapy in patients with portal hypertension due to a variety of causes was
reported by Sarin et al.70(2176) In addition to patients with cirrhosis, this study included 31 patients
with noncirrhotic portal fibrosis and 16 patients with extrahepatic portal venous obstruction. After
variceal eradication, the latter group with extrahepatic obstruction underwent endoscopic surveillance
procedures on a monthly basis for an average of 17.9 months. Recurrence of varices was not seen in
these patients, in contrast to the recurrence rate of 22% for patients with cirrhosis or noncirrhotic portal
fibrosis. Complications attributable to sclerotherapy (i.e., injections of absolute alcohol) occurred in
about 15% of patients in each group.
A similar study of Bhargava et al.71(2177) included 83 patients with noncirrhotic portal fibrosis, 234
with cirrhosis, and 87 with extrahepatic portal venous obstruction. Response to sclerotherapy in terms
of the ability to eradicate varices was better in patients with venous obstruction and noncirrhotic portal
fibrosis than in those with cirrhosis. After variceal eradication, the frequency of recurrent bleeding was
significantly less and the median bleeding-free period significantly longer in patients with portal venous
obstruction. However, recurrent bleeding was more common in patients with poor liver function
regardless of cause. The probability of 7-year survival was lowest for patients with cirrhosis and similar
for those with portal venous obstruction or noncirrhotic fibrosis, although survival was also related
directly to liver function status regardless of the cause of portal hypertension.
Schistosomiasis
Sclerotherapy has been used for the treatment of bleeding esophageal varices due to
schistosomiasis.7274(2178) The study of Sakai et al.72(2179) found the response to sclerotherapy to
be better in patients with recurrent variceal bleeding after decompressive surgery than in those with
bleeding and no prior surgical treatment. The sclerotherapy technique included intravascular injections
of ethanolamine oleate in some patients and paravariceal injections in others. The disparity in control
of recurrent bleeding was thought to result from differences in the level of portal pressure.
Mohamed et al.73(2180) compared the results of sclerotherapy in patients with chronic liver disease
due to hepatitis B, schistosomiasis, and unknown factors. With respect to eradication of varices,
recurrent bleeding, and mortality, the patients with schistosomiasis had better results. The study of
Cordeiro et al.74(2181) compared the results of paravariceal sclerotherapy using ethanolamine oleate
in 32 patients with variceal hemorrhage with a control group composed of 18 patients with hemorrhage
who refused sclerotherapy. Although the incidence of recurrent bleeding was 28.1% for the group
undergoing sclerotherapy and 44.5% for the control group, this difference did not reach statistical
significance. However, a statistically significant difference between the two groups was observed for
mortality due to variceal hemorrhage (3.1% vs. 27.7%)
Hepatocellular Carcinoma
The management of patients with cirrhosis and hepatocellular carcinoma who develop bleeding as a
result of portal hypertension and esophageal varices is problematic because of the associated poor
prognosis. Nakashima et al.75(2182) reported the results of sclerotherapy with intravariceal
ethanolamine oleate in 13 such patients. The response to sclerotherapy in this small series of patients
appeared to depend more on liver function status than the stage of the cancer. Lo et al.76(2183)
reported a larger series in which the results of sclerotherapy (i.e., intravariceal injections of
ethanolamine oleate or 1.5% sodium tetradecyl sulfate) in 37 patients with unresectable hepatocellular
carcinoma were compared with those in 33 patients managed without sclerotherapy. Although survival
was short in both groups, with no difference in the number of deaths from bleeding, mortality due to the
initial episode of bleeding was significantly less for those patients who underwent sclerotherapy. For
patients without portal vein thrombosis, the frequency of fatal hemorrhage was significantly reduced by
sclerotherapy.
In the trial of Cheng et al.,77(2184) sclerotherapy performed for acute variceal bleeding in 65 patients
with advanced hepatocellular carcinoma was compared with conservative treatment in 60 similar
patients. Although sclerotherapy was more effective for control of acute bleeding and reduced the
frequency of recurrent bleeding, survival time was short for patients in both groups, especially those
with poor liver function.
Virtually all studies of sclerotherapy for variceal bleeding due to extrahepatic portal venous obstruction
report favorable results. Variceal eradication is relatively easy, recurrence is relatively infrequent, and
complication rates are relatively low for patients with this condition. The differences in outcome for
sclerotherapy are related to the cause of portal hypertension. Although these data are not extensive,
patients with cirrhosis evidently fare worse than those with portal hypertension due to other causes
such as portal venous obstruction, noncirrhotic portal fibrosis, and schistosomiasis. Liver function
status plays an important role in the favorable outcome of these patients, particularly those with
extrahepatic portal venous obstruction.
Sclerotherapy Combined with Other Measures

A high rate of recurrent hemorrhage is a consistent finding in all controlled trials of long-term injection
sclerotherapy. Most episodes occur in the first few weeks to months after the commencement of
treatment, which is during the period before variceal obliteration. Several measures have been adopted
with the specific aim of reducing the frequency of early recurrence of bleeding. Many of these involve
the development of new techniques or schedules of injections (see Chapter 32: Technique of
Endoscopic Sclerotherapy), but a number of pharmacologic agents have been assessed in this role.
One approach to pharmacologic therapy is based on the observation that many of the early episodes of
recurrent bleeding are secondary to sclerotherapy-induced mucosal damage. Although H2-receptor
antagonists have been employed in this setting, no objective evidence supports their use. However, the
mucosal protectant sucralfate has been shown in a large, controlled trial to significantly reduce the
frequency of early recurrent bleeding, particularly from esophageal ulceration.78(2185) The major
benefit appeared to be restricted to patients with well-compensated cirrhosis. Of particular interest was
the failure of this agent to prevent ulcers or hasten healing of the mucosal ulcers. This provides some
indirect evidence that sucralfate may have a local hemostatic effect.
A small percentage of sclerotherapy-induced ulcers, particularly those straddling the gastroesophageal
junction, may become chronic and be the source of recurrent bleeding and pain. In a small study, a
number of such ulcers that were resistant to H2-receptor antagonists were shown to heal within 1 to 2
weeks when treated with use of the proton pump inhibitor omeprazole.79(2186) There was a high
frequency of ulcer relapse after treatment was discontinued, and a number of patients had to use
omeprazole on an indefinite basis. Omeprazole has also been shown to heal the mucosal ulceration
associated with esophageal transection.80(2187)
Two preliminary reports describe the use of the somatostatin analog octreotide given subcutaneously
in combination with long-term injection sclerotherapy.81,82(2188) The first of these81(2189) showed
considerable benefit for this combination, but this result was not confirmed in the second
report.82(2190) This combination, which incorporates an expensive drug that requires parenteral
administration, cannot be recommended without further support from a controlled investigation.
The effectiveness of isosorbide-5-mononitrate in combination with long-term sclerotherapy was studied
in the double-blind, multicenter trial of Bertoni et al.83(2191) Sclerotherapy (intravariceal and
paravariceal 1% polidocanol) was performed at weekly intervals until varices were obliterated in 37
patients who received the drug and 39 who received a placebo. Additional sclerotherapy sessions were
performed for recurrent bleeding. The number of patients with recurrent hemorrhage and total number
of episodes of recurrent bleeding were significantly less in the group that received the drug. No
difference was seen in the severity of bleeding or median transfusion requirement per bleeding
episode. However, the cumulative number of units of blood transfused was three times greater in the
placebo group. Survival and the number of sclerotherapy sessions required to achieve variceal
eradication were not significantly different in the two groups.
The addition of long-term, oral -adrenergic receptor blockade to a regimen of injection sclerotherapy
has been evaluated in several studies, with somewhat conflicting results that overall provide little
support for adopting such combined therapy.53,8486(2192)
Vinel et al.53(2193) randomized 75 patients with cirrhosis, largely alcohol induced, and endoscopically
proven variceal hemorrhage to treatment by sclerotherapy (paravariceal plus intravariceal 1%
polidocanol) or sclerotherapy plus propranolol. Initial treatment of variceal bleeding included
intravenous vasopressin and sclerotherapy within 24 hr of admission. Only those patients regarded as
hemodynamically stable were considered for randomization, and 16 patients were thereby eliminated.
Sclerotherapy sessions were performed until variceal eradication was achieved. More than twice as
many patients experienced recurrent variceal bleeding in the group treated by sclerotherapy alone. The
number of episodes of recurrent bleeding and the mean number of units of blood transfused were
significantly less for patients who received propranolol in addition to sclero-therapy. Mortality rates
were similar for both treatment groups. Vinel et al.53(2194) concluded that administration of
propranolol favorably influenced the frequency and severity of recurrent episodes of bleeding during
the interval between initial treatment and variceal eradication.
The randomized trial of Lundell et al.55(2195) compared sclero-therapy alone (monthly injections of
1% polidocanol injection sessions as needed for recurrent bleeding) with sclerotherapy plus
propranolol. This trial included 41 patients who were admitted with a first episode of bleeding.
Injections were performed until variceal eradication was achieved; the time until obliteration was the
same for both treatment groups. Although the cumulative number of episodes of recurrent bleeding
attributable to sclerotherapy-induced esophageal ulceration and frank variceal bleeding was the same
for both groups of patients, bleeding was somewhat less severe in patients who received sclerotherapy
plus propranolol.
Sclerotherapy plus propranolol, oral propranolol alone, and transhepatic sclerotherapy plus propranolol
were compared in a trial by O'Connor et al.85(2196) Mortality was significantly higher, and there was a
trend toward more serious episodes of bleeding for patients treated with pro-pranolol alone. Mortality
rates and the frequency of variceal bleeding tended to be lower for patients who received propranolol in
addition to sclerotherapy.
Ink et al.87(2197) compared the combination of sclerotherapy and propranolol against propranolol
alone for the prevention of recurrent bleeding in a randomized trial that included 131 patients, most of
whom had alcoholic cirrhosis. The distribution was approximately even between Child-Pugh class B
and C cirrhosis. No additional benefit for the combination was found for overall recurrence of bleeding
or for survival. More than twice as many patients in the group treated with propranolol alone had
recurrent variceal bleeding compared with those who had combined therapy.
Although the numbers of patients randomized in the trial of Jensen and Krarup86(2198) to
sclerotherapy plus placebo or sclerotherapy plus propranolol were small (n = 16 and n = 15,
respectively), the investigators found that the addition of propranolol to a regimen of long-term
sclerotherapy (paravariceal 2% polidocanol) reduced the frequency of recurrent variceal bleeding
during the interval between initial treatment and variceal eradication.
Vickers et al.88(2199) compared sclerotherapy alone in 34 patients with sclerotherapy plus propranolol
in 39 patients. No significant differences between the two groups were observed for the cumulative
percentage of patients without recurrent episodes of bleeding, survival, mean time to eradication of
varices, and rate of recurrence of varices. These results are similar to those from the trial of Westaby
et al.,84(2200) in which 27 patients underwent sclerotherapy alone and 26 patients received
propranolol in addition to sclerotherapy. Before variceal obliteration, 8 patients in the former and 7 in
the latter group had recurrent bleeding. Hemorrhage was uncontrollable and was the cause of death of
2 patients. Resuscitation of 1 patient during an episode of bleeding was said to be compromised by the
use of propranolol.
The high frequency of recurrent variceal bleeding represents an important limitation of injection
sclerotherapy and pharmacotherapy by -receptor blockade. Although available data are limited, the
addition of propranolol to a regimen of sclerotherapy sessions aimed at eradication of esophageal
varices appears to have a favorable effect on recurrent variceal bleeding during the interval between
initial treatment and eradication. Data from trials of -receptor blockade versus sclerotherapy alone are
not strictly comparable. Data on the efficacy of propranolol for the prevention of recurrent variceal
hemorrhage without reference to sclerotherapy are also conflicting. The value of sclerotherapy is offset
by associated complications. The -receptor blockade offers the advantage of simple administration
but is critically dependent on compliance and has also been associated with a significant complication
rate.89(2201)
Gastric and Ectopic Varices

The incidence of gastric varices has been poorly documented, reflecting frequent difficulties in
diagnosis, particularly for patients with active variceal bleeding. As many as 50% of all bleeding
episodes arising from gastric varices may remain undiagnosed after the first diagnostic
endoscopy,90(2202) and this is particularly true for fundal gastric varices, which are inevitably covered
by the gastric pool when bleeding has been considerable. The diagnostic yield may be enhanced by a
high index of suspicion and by tipping and rolling the patient to displace the fundal pool.
Classification of gastric varices should provide a framework to guide diagnostic endoscopy. Most
workers have adopted a simple classification based on the anatomic distribution of gastric varices
(Figure 332).91,92(2203) The most common type is that found on the lesser curve, which is in direct
continuity with esophageal varices, crossing the gastroesophageal junction. Fundal varices may be
associated with esophageal varices, crossing the cardia, or may occur in isolation. Isolated varices are
a common finding in patients with splenic vein thrombosis. Gastric varices may rarely be found within a
hiatus hernia, at which site their failure to protrude into the gastric lumen may make diagnosis difficult.
Studies in which gastric variceal bleeding has been assessed prospectively suggest that 20 to 30% of
all episodes of variceal bleeding arise from within the gastric lumen.93,94(2204)
(2205)Figure 332. The grading of gastric varices according to anatomic site. A (left to right),
Lesser curve, cardia, fundus. B, Hiatus hernia.
Sclerotherapy has been used for each of the gastric variceal sites. Available evidence suggests that
varices arising on the lesser curve or within a hiatus hernia can be managed by this technique, with
success rates comparable to that observed for esophageal varices.92,94(2206) Although there are
reports of successful management of fundal varices by injection sclerotherapy,95,96(2207) the largest
series reported only a 44% success rate for controlling active bleeding and a 50% associated mortality
rate (Table 332).92(2208)
Results Obtained Using Injection of a Sclerosing

Agent (5% Ethanolamine Oleate) for Gastric Variceal Bleeding
TABLE 332
SITE (%)
BLEEDING EPISODE
Lesser Curve
Fundus
Hiatus Hernia
Admission control of bleeding 11/13 (85)

12/27 (44)
6/7 (86)
Recurrent bleeding*
7/13 (58)
10/14 (77)
4/6 (67)
Data from Gimson A, Westaby D, Williams R. Endoscopic therapy in the
management of gastric variceal haemorrhage. J Hepatol 1991; 13:2748. 1991
Munksgaard International Publishers, Ltd., Copenhagen, Denmark.
* Refers to those surviving the index admission.
The major concern about sclerotherapy for treating fundal varices is the development of mucosal
ulceration and the associated risk of recurrent bleeding. This has prompted trials of alternative
injectates, particularly the cyanoacrylate tissue adhesives (see Chapter 14: The Use of Histoacryl
Tissue Adhesive for the Treatment of Variceal Bleeding).97(2209) These agents solidify almost
instantaneously when brought into contact with blood, and when injected into a varix, they can produce
total luminal occlusion. Only a small number of reports are available to assess the efficacy of tissue
adhesives, but these provide considerable evidence of benefit.98,99(2210) The injection of the
cyanoacrylate adhesives requires considerable attention to detail to avoid polymerization on the tip of
the endoscope itself, which can cause irreparable damage.
The complications associated with tissue adhesive injection require more complete documentation and
increased experience in their use. Mucosal damage is well recognized endoscopically and in
postmortem studies.100(2211) Of greatest concern is the report of two patients who developed
neurologic damage after bucrylate injection.101(2212)
Bovine or human thrombin represent further alternative injectates.102(2213) These agents have been
used alone or with sclerosing agents to enhance local intravarix thrombosis. Concerns that thrombin
might precipitate portal vein thrombosis or disseminated intravascular coagulation have not been
substantiated.103(2214) Our own preliminary experience with pure bovine thrombin has been
encouraging and without overt complications (Table 333). The major advantage of this approach is
the total absence of mucosal damage and associated recurrence of bleeding from this type of lesion.
Results Obtained With the

Injection of Human Thrombin for Gastric Variceal
Bleeding
TABLE 333
SITE*
BLEEDING EPISODE
Admission control of bleeding
Recurrent bleeding
Lesser Curve
Fundus
2/2 (100)
1/2 (50)
8/9 (89)
1/9 (11)
Previously unpublished data are supplied by Westaby.

* Values are expressed as the number of successful interventions
per site, with the success rate within parentheses.
Median follow-up was 6 months.
Bleeding from varices other than those arising in the esophagus and proximal stomach is rare, but
several case reports describe injection sclerotherapy used for varices in ectopic sites104(2215) such
as the duodenum, gastric antrum, anorectum, stoma,105(2216) and colon.106(2217) As in the case of
gastric fundal varices, the major concern is the frequency of sclerotherapy-induced mucosal ulceration,
which may precipitate further bleeding. Little evidence is available to support a possible role for other
injectates, such as tissue adhesives or thrombin.
Clinical Use of Injection Sclerotherapy

The popularity of injection sclerotherapy for the management of variceal bleeding has been based on
widespread applicability and proven efficacy, particularly for the management of an episode of active
variceal bleeding, a role in which injection sclerotherapy can be considered the treatment of choice.
This status is not based solely on efficacy (i.e., surgical techniques have been shown to be equally
beneficial) but includes the logical use of an endoscopic technique during the mandatory diagnostic
endoscopy. Even in the most experienced hands, injection sclerotherapy fails to control bleeding in 10
to 15% of cases, and in such circumstances, successful salvage has been possible by
devascularization and shunt surgery. Such an integrated management protocol depends on the early
recognition of failed treatment, and no more than two courses of sclerotherapy are justified in a patient
who continues to bleed. In 20 to 30% of patients with active variceal bleeding, the condition arises from
within the gastric lumen. Evidence suggests that injection sclerotherapy is a justified primary measure
when gastric variceal bleeding arises from the lesser curve or within a hiatus hernia. In the case of
fundal varices, the success rate has been considerably less, with high associated morbidity and
mortality rates. Newer injectates, such as tissue adhesives or thrombin, may prove to be better
alternatives.
Despite the collective evidence from a number of controlled trials, the use of long-term injection
sclerotherapy is being critically assessed. The major concern is the consistently high frequency of early
recurrent bleeding, which is in many cases secondary to sclerotherapy-induced mucosal damage.
Despite a general decline in the reported complications of injection sclerotherapy over the last 2
decades, serious morbidity and mortality rates are associated with the technique, particularly in
inexperienced hands. The reported absence of a survival benefit compared with a single session of
sclerotherapy for each episode of recurrent bleeding places an important question mark on the cost
effectiveness of long-term regimens. It is likely that these will be rationalized to a number of early
treatment sessions, perhaps with the addition of a long-term pharmacologic agent such as a
receptor-blocking drug. Long-term surveillance programs appear to have little justification.
Two techniques appear to have the potential to replace long-term injection sclerotherapy. TIPS is
undoubtedly effective for the prevention of recurrent bleeding in the short term, although long-term
efficacy and complications are poorly documented. The major limitation of this technique is its
restricted availability, requiring highly specialized equipment and operators. However, a new local
endoscopic treatment, called variceal banding (see Chapter 36: Endoscopic Elastic Band Ligation for
Bleeding Esophageal Varices) appears to provide the greatest prospect for replacing injection
sclerotherapy. Although this technique has not been shown to have any advantages over sclerotherapy
for the management of active bleeding,107(2218) it appears to have the potential to obliterate varices
with fewer treatment sessions and with a very low complication rate.
Over a period of 2 decades, long-term injection sclerotherapy and shunt surgery have competed for the
predominant place in the management of variceal bleeding. We now have an endoscopic technique
and a nonsurgical shunt ready to extend the period of comparison of these two philosophic
approaches.
REFERENCES
1.
Westaby D, Williams R. History and development of sclerotherapy. Gastrointest Endosc

1983;29:3037.
2. L'Abbe K, Detsky A, O'Rourke K. Meta-analysis in clinical research. Ann Intern Med
1987;107:22433.
3. Johnston GW, Rodgers HW. A review of 15 years experience in the use of sclerotherapy in
the control of acute haemorrhage from esophageal varices. Br J Surg 1973;60:797800.
4. Terblanche J, Yakoob HI, Bornman PC, et al. A five-year prospective evaluation of
tamponade and sclerotherapy. Ann Surg 1981;194:52130.
5. Yassin YM, Sherif SM. Randomized controlled trial of injection sclerotherapy for bleeding
oesophageal varicesAn interim report. Br J Surg 1983;70:202.
6. Paquet KJ, Feussner H. Endoscopic sclerosis and esophageal balloon tamponade in acute
hemorrhage from esophagogastric varices: A prospective controlled randomized trial.
7. Soderlund C, Ihre T. Endoscopic sclerotherapy versus conservative management of bleeding
8. Larson AW, Cohen H, Zweiban B, et al. Acute esophageal variceal sclerotherapy. JAMA
1986;255:497500.
9. Moreto M, Zaballa M, Bernal A, et al. A randomized trial of tamponade or sclerotherapy as
immediate treatment for bleeding esophageal varices. Surg Gynecol Obstet 1988;167:3314.
10. Clark AW, Westaby D, Silk DBA, et al. Prospective controlled trial of injection sclerotherapy in
patients with cirrhosis and recent variceal haemorrhage. Lancet 1980;2:5524.
11. MacDougall BRD, Westaby D, Theodossi A, et al. Increased long-term survival in variceal
haemorrhage using injection sclerotherapy. Lancet 1982;1:1247.
12. Korula J, Balart LA, Radvan G, et al. A prospective, randomized controlled trial of chronic
13. Sung JJY, Chung SCS, Lai C-W, et al. Octreotide infusion of emergency sclerotherapy for
variceal haemorrhage. Lancet 1993;342:63741.
14. Planas R, Quer JC, Boix J, et al. A prospective randomized trial comparing somatostatin and
sclerotherapy in the treatment of acute variceal bleeding. Hepatology 1994;20:3705.
15. Lo GH, Lai KH, Ng WW, et al. Injection sclerotherapy preceded by esophageal tamponade
versus immediate sclerotherapy in arresting active variceal bleeding: A prospective
16. Prindiville T, Trudeau W. A comparison of immediate versus delayed endoscopic injection
sclerosis of bleeding esophageal varices. Gastrointest Endosc 1986;32:3858.
17. Westaby D, Hayes P, Gimson A, et al. Controlled clinical trial of injection sclerotherapy for
active variceal bleeding. Hepatology 1989;9:2747.
18. Shemesh E, Czerniak A, Klein E, et al. A comparison between emergency and delayed
endoscopic injection sclerotherapy of bleeding esophageal varices in nonalcoholic portal
hypertension. J Clin Gastroenterol 1990;12:59.
19. Moreto M, Zaballa M, Ojembarrena E, et al. Combined (short-term plus long-term)
sclerotherapy versus short-term only sclerotherapy: A randomised prospective trial. Gut
1994;35: 68791.
20. Terblanche J. Has sclerotherapy altered the management of patients with variceal bleeding?.
Am J Surg 1990;160:3742.
21. Cello JP, Grendell JH, Crass RA, et al. Endoscopic sclerotherapy versus portacaval shunt in
patient with severe cirrhosis and acute variceal hemorrhage. Long-term follow-up. N Engl J
Med 1987;316:115.
22. Burroughs AK, Hamilton G, Phillips A, et al. A comparison of sclerotherapy with staple
transection of the esophagus for the emergency control of bleeding from esophageal varices.
N Engl J Med 1989;321:85762.
23. Huizinga W, Angorn P, Baker L. Esophageal transection versus injection sclerotherapy in the
management of bleeding esophageal varices in patients at high risk. Surg Gynecol Obstet
1985;160:53946..
24. Teres J, Baroni R, Bordas JM, et al. Randomized trial of portacaval shunt, stapling transection
and endoscopic sclerotherapy in uncontrolled variceal bleeding. J Hepatol 1987;4:15967.
25. McCormick PA, Kaye GL, Greenslade L, et al. Esophageal staple transection as a salvage
procedure after failure of acute injection sclerotherapy. Hepatology 1992;15:4036.
26. . Bleeding oesophageal varices: IST, EVL or TIPS [Editorial]. Lancet 1992;340:515.
27. Barsoum MS, Bolous FI, El-Rooby AA, et al. Tamponade and injection sclerotherapy in the
28. Terblanche J, Kahn D, Campbell JAH, et al. Failure of repeated injection sclerotherapy to
improve long-term survival after oesophageal variceal bleeding. Lancet 1983;2:132832.
29. The Copenhagen Esophageal Varices Sclerotherapy Project. Sclerotherapy after first variceal
30.
Westaby D, MacDougall BRD, Williams R. Improved survival following injection sclerotherapy
for esophageal varices: Final analysis of a controlled trial. Hepatology 1985;5:62731.
31.
Burroughs AK, McCormick PA, Siringo S, et al. Prospective randomized trial of long-term
sclerotherapy for variceal rebleeding using the same protocol to treat rebleeding in all patients
[Abstract]. J Hepatol 1989;9:S12.
32. The Veterans Affairs Cooperative Variceal Sclerotherapy Group. Sclerotherapy for male
alcoholic cirrhotic patients who have bled from esophageal varices: Results of a randomized,
multicenter clinical trial. Hepatology 1994;20:61825.
33.
Pagliari L, Burroughs AK, Sorensen TA, et al. Therapeutic controversies and randomised
controlled trials: Prevention of bleeding and rebleeding in cirrhosis. Gastroenterol Int
1989;2:7184.
34. Infante-Rivard C, Esnaola S, Villeneuve JP. Role of endoscopic variceal sclerotherapy in the
long-term management of variceal bleeding: A meta-analysis. Gastroenterology
1989;96:108792.
35. Rikkers LF, Jin G, Burnett DA, et al. Shunt surgery versus endoscopic sclerotherapy for
variceal hemorrhage: Late results of a randomized trial. Am J Surg 1993;165:2733.
36. Teres J, Bordas JM, Bravo D, et al. Sclerotherapy versus distal splenorenal shunt in the
elective treatment of variceal hemorrhage: A randomized controlled trial. Hepatology
1987;7:4306.
37. Henderson JM, Kutner MH, Millikan WJ, et al. Endoscopic variceal sclerosis compared with
distal splenorenal shunt to prevent recurrent variceal bleeding in cirrhosis. A prospective,
randomized trial. Ann Intern Med 1990;112:2629.
38. Spina GP, Santambrogio R, Opocher E, et al. Distal splenorenal shunt versus endoscopic
sclerotherapy in the prevention of variceal rebleeding. First stage of a randomized, controlled
trial. Ann Surg 1990;211:17886.
39. Planas R, Boix J, Broggi M, et al. Portacaval shunt versus endoscopic sclerotherapy in the
elective treatment of variceal hemorrhage. Gastroenterology 1991;100:107886.
40. Cabrera J, Maynar M, Granados R, et al. Transjugular intrahepatic portosystemic shunt
versus sclerotherapy in the elective treatment of variceal hemorrhage. Gastroenterology
1996;110:8329.
41. Cello JP, Ring EJ, Olcott EW, et al. Endoscopic sclerotherapy compared with percutaneous
transjugular intrahepatic portosystemic shunt after initial sclerotherapy in patients with acute
variceal hemorrhage. A randomized, controlled trial. Ann Intern Med 1997;126:85865.
41a. Sanyal AJ, Freedman AM, Luketic VA, et al. Transjugular intrahepatic portosystemic shunts
compared with endoscopic sclerotherapy for the prevention of recurrent variceal hemorrhage.
A randomized, controlled trial. Ann Intern Med 1997;126:84957.
41b. Rossle M, Deibert P, Haag K, et al. Randomised trial of
transjugular-intrahepatic-portosystemic shunt versus endoscopy plus propranolol for
prevention of variceal rebleeding. Lancet 1997;349: 10439.
41c. Jalan R, Finlayson ND, Hayes PC. TIPSS trials: Design determines outcome. Hepatology
1997;26:13615.
42. Terabayashi H, Ohnishi K, Tsunoda T, et al. Prospective controlled trial of elective endoscopic
sclerotherapy in comparison with percutaneous transhepatic obliteration of esophageal
varices in patients with nonalcoholic cirrhosis. Gastroenterology 1987;93:12059.
43. Lebrec D, Poynard T, Hillon P, Benhamou J-P. Propranolol for prevention of recurrent
gastrointestinal bleeding in patients with cirrhosis. N Engl J Med 1981;305:13714.
44. Lebrec D, Poynard T, Bernuau J, et al. A randomized controlled study of propranolol for
prevention of recurrent gastrointestinal bleeding in patients with cirrhosis: A final report.
45. Garden OJ, Mills PR, Birnie GC, et al. Propanolol in the prevention of variceal hemorrhage in
cirrhotic patients. A controlled trial. Gastroenterology 1990;98:18590.
46. Villeneuve J-P, Pomier-Layrargues G, Infante-Rivard C, et al. Propranolol for the prevention of
recurrent variceal hemorrhage: A controlled trial. Hepatology 1986;6:123943.
47. Burroughs AK, Jenkins WJ, Sherlock S, et al. Controlled trial of propranolol for the prevention
of recurrent variceal hemorrhage in patients with cirrhosis. N Engl J Med 1983;309:153942.
48. Fleig WE, Stange EF, Hunecke R, et al. Prevention of recurrent bleeding in cirrhotics with
recent variceal haemorrhage: Prospective randomized comparison of propranolol and
sclerotherapy. Hepatology 1987;7:35561.
49. Alexandrino PT, Alves MM, Pinto Correia J. Propranolol or endoscopic sclerotherapy in the
prevention of recurrence of variceal bleeding. A prospective, randomized controlled trial. J
Hepatol 1988;7:17585.
50. Westaby D, Polson RJ, Gimson AE, et al. A controlled trial of oral propranolol sclerotherapy
for the long-term management of variceal bleeding. Hepatology 1990;11:3539.
51. Dollet JM, Champigneulle B, Patris A, et al. Sclerotherapie endoscopique contre propranolol
apres hemorragie par rupture de varices oesophagiennes chez le cirrhotique. Resultats a 4
ans d'une etude randomisee. Gastroenterol Clin Biol 1988;12:2349.
52. Dasarathy S, Dwivedi M, Bhargava DK, et al. A prospective randomized trial comparing
repeated endoscopic sclerotherapy and propranolol in decompensated (Child class B and C)
cirrhotic patients. Hepatology 1992;16:8994.
53. Vinel JP, Lamouliatte H, Cales P, et al. Propranolol reduces the rebleeding rate during
endoscopic sclerotherapy before variceal obliteration. Gastroenterology 1992;102:17603.
54. Rossi V, Cales P, Burtin P, et al. Prevention of recurrent variceal bleeding in alcoholic cirrhotic
patients: Prospective controlled trial of propranolol and sclerotherapy. J Hepatol
1991;12:2839.
55. Lundell L, Leth R, Lind T, et al. Evaluation of propranolol for prevention of recurrent bleeding
from esophageal varices between sclerotherapy sessions. Acta Chir Scand 1990;156:7115.
56. Qureshi H, Zuberi SJ, Alam E. Efficacy of oral propranolol and injection sclerotherapy in the
long-term management of variceal bleeding. Digestion 1990;46:1938.
57. Teres J, Bosch J, Bordas JM, et al. Propranolol versus sclerotherapy in preventing variceal
rebleeding: A randomized controlled trial. Gastroenterology 1993;105:150814..
58. Kanazawa H, Watari A, Matsusaka S, et al. Prospective controlled study of elective
sclerotherapy plus oral propranolol for prevention of recurrent bleeding in cirrhotics with recent
variceal hemorrhage. Nippon Shokakibyo Gakkai Zasshi 1991;88:13418.
59. Lo GH, Lai KH, Lee SD, et al. Does propranolol maintain post-sclerotherapy variceal
obliteration? A prospective randomized study. J Gastroenterol Hepatol 1993;8:35862.
60. Fonkalsrud EW. Treatment of variceal hemorrhage in children. Surg Clin North Am
1990;70:47587.
61. Thatcher BS, Sivak MV Jr, Petrini JL. Endoscopic sclerotherapy for bleeding esophageal
varices secondary to extrahepatic portal vein obstruction. Gastrointest Endosc
1987;33:2149.
62. Kahn D, Krige JE, Terblanche J, et al. A 15-year experience of injection sclerotherapy in adult
patients with extrahepatic portal venous obstruction. Ann Surg 1994;219:349.
63. Bhargava DK, Dwivedi M, Dasarathy S, Arora A. Endoscopic sclerotherapy for portal
hypertension due to extrahepatic obstruction: Long-term follow-up. Gastrointest Endosc
1989;35:30911.
64. Chawla YK, Dilawari JB, Ramesh GN, et al. Sclerotherapy in extrahepatic portal venous
obstruction. Gut 1990;31:2136.
65. Pauzner D, Wolman I, Niv D, et al. Endoscopic sclerotherapy in extrahepatic portal
hypertension in pregnancy. Am J Obstet Gynecol 1991;164:1523.
66. Augustine P, Joseph PC. Sclerotherapy for esophageal varices and pregnancy. Gastrointest
Endosc 1989;35:4678.
67. Salena BJ, Sivak MV Jr. Pregnancy and esophageal varices. Gastrointest Endosc
1988;34:4923.
68. Iwase H, Morise K, Kawase T, Horiuchi Y. Endoscopic injection sclerotherapy for esophageal
varices during pregnancy. J Clin Gastroenterol 1994;18:803.
69. Bhargava DK, Dwivedi M, Dasarathy S, Sundaram KR. Sclerotherapy after variceal
hemorrhage in non-cirrhotic portal fibrosis. Am J Gastroenterol 1989;84:12358.
70. Sarin SK, Sachdev G, Nanda R. Follow-up of patients after variceal eradication. Ann Surg
1986;204:7882.
71. Bhargava DK, Dasarathy S, Sundaram KR, Ahuja RK. Efficacy of endoscopic sclerotherapy
on long-term management of oesophageal varices: A comparative study of results in patients
with cirrhosis of the liver, non-cirrhotic portal fibrosis (NCPF) and extrahepatic portal venous
obstruction (EHO). J Gastroenterol Hepatol 1991;6:4715.
72. Sakai P, Boaventura S, Ishioka S, et al. Sclerotherapy of bleeding esophageal varices in
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
schistosomiasis. Comparative study in patients with and without previous surgery for portal
hypertension. Endoscopy 1990;22:57.
Mohamed AE, al Karawi MA, al Otaibi R, Hanid MA. Results of sclerotherapy in 100 patients
comparison of the outcome between schistosomiasis and hepatitis B. Hepatogastroenterology
1989;36:3336.
Cordeiro F. Variceal sclerosis in schistosomotic patients: A 5-year follow-up study.
Nakashima M, Yoshida K, Izuno K, et al. Clinical evaluation of endoscopic injection
sclerotherapy for bleeding esophageal varices in liver cirrhosis with hepatocellular carcinoma.
Gastroenterol Jpn 1991;26(suppl 3):368.
Lo GH, Lin CY, Lai KH, et al. Endoscopic injection sclerotherapy versus conservative
treatment for patients with unresectable hepatocellular carcinoma and bleeding esophageal
varices. Gastrointest Endosc 1991;37:1614.
Cheng CY, Chen GH, Chang CS, et al. Sclerotherapy for esophageal variceal bleeding in
advanced hepatocellular carcinoma: An 8-year experience in Taiwan. J Gastroenterol
1994;29:4748.
Polson RJ, Westaby D, Gimson AES, et al. Sucralfate for the prevention of early rebleeding
following injection sclerotherapy for esophageal varices. Hepatology 1989;10:27983.
Gimson A, Polson R, Westaby D, Williams R. Omeprazole in the management of intractable
esophageal ulceration following injection sclerotherapy. Gastroenterology 1990;99:182931.
Kaye G, McCormick A, Siringo S, et al. Bleeding from staple line erosion after oesophageal
transection. Effect of omeprazole. Hepatology 1992;6:10315.
Jenkins SA, Ellenbogen S, Baxter JN, et al. Sandostatin in the long-term management of
portal hypertensionA preliminary prospective randomised controlled clinical trial [Abstract].
Gut 1991;32:A838.
de Franchis R, Andreoni B, Primignani M, et al. Octreotide does not reduce the incidence of
early rebleeding in cirrhotics submitted to variceal sclerotherapy [Abstract]. Hepatology
1992;16:A150.
Bertoni G, Sassatelli R, Fornaciari G, et al. Oral isosorbide-5-mononitrate reduces the
rebleeding rate during the course of injection sclerotherapy for esophageal varices. Scand J
Westaby D, Melia W, Hegarty J, et al. Use of Propranolol to reduce the rebleeding rate during
injection sclerotherapy prior to variceal obliteration. Hepatology 1986;6:6735.
O'Connor KW, Lehman G, Yune H, et al. Comparison of three nonsurgical treatments for
bleeding esophageal varices. Gastroenterology 1989;96:899906.
Jensen L, Krarup N. Propranolol in prevention of recurrent bleeding from oesophageal varices
during the course of endoscopic sclerotherapy. Scand J Gastroenterol 1989;24:33945.
Ink O, Martin T, Poynard T, et al. Does elective sclerotherapy improve the efficacy of
long-term propranolol for prevention of recurrent bleeding in patients with severe cirrhosis? A
prospective multicentre randomised trial. Hepatology 1992;16:9129.
Vickers C, Rhodes J, Chesner I, et al. Prevention of rebleeding from oesophageal varices:
Two-year follow up of a prospective controlled trial of propranolol in addition to sclerotherapy.
J Hepatol 1994;21:817.
Hayes PC. Safety of propranolol in portal hypertension: Conventional long acting formulations.
Drugs 1989;37(suppl 2):703.
Madsen MS, Patersen TH, Sommer H. Segmental portal hypertension. Ann Surg
1986;204:727.
Hosking SW, Johnson AG. Gastric varices: A proposed classification to management. Br J
Surg 1988;75:1956.
Gimson A, Westaby D, Williams R. Endoscopic therapy in the management of gastric variceal
haemorrhage. J Hepatol 1991;13:2748.
Mitchell KJ, MacDougall BR, Silk DB, Williams R. A prospective reappraisal of emergency
endoscopy in patients with portal hypertension. Scand J Gastroenterol 1982;17:9658.
94.
Sarin SK, Kumar A. Gastric varices: Profile, classification, and management. Am J

96. Sarin SK, Sachdev G, Nanda R, et al. Endoscopic sclerotherapy in the treatment of gastric
97. Gotlib JP, Demma I, Fonsecca A, et al. Resultats a 1 an du traitement endoscopique electif
des hemorragies par rupture de varices oesophagiennes chez le cirrhotique [Abstract].
Gastroenterol Clin Biol 1984;8:133A..
98. Soehendra N, Nam VCH, Grimm H, Kempeneers I. Endoscopic obturation of large
esophagogastric varices with bucrylate. Endoscopy 1986;18:256.
99. Ramond M-J, Valla D, Mosnier J-F, et al. Successful endoscopic obturation of gastric varices
with butyl cyanoacrylate. Hepatology 1989;10:48893.
100.
Fabiani B, Degott C, Ramond MJ, et al. Obturation endoscopique des varices oesogastriques
par le Bucrylate. Gastroenterol Clin Biol 1986;10:5803.
101.
See A, Florent C, Lamy P, et al. Accidents vasculaires cerebraux apres obturation
endoscopique des varices oesophagiennes par 1(MathematicalPi-One
)9-isobutyl-2-cyanoacrylate chez deux malades. Gastroenterol Clin Biol 1986;10:6047.
102. Kitano S, Hashizume M, Yamaga H, et al. Human thrombin plus 5% ethanolamine oleate
injected to sclerose oesophageal varices: A prospective randomized trial. Br J Surg
1989;76:7158.
103. Kujii Y, Sugawa C, Ozawa C. Haemostasis activation during esophageal variceal
sclerotherapy with thrombin in cirrhotics. Ann Surg 1991;57:2225.
104. Ramage JK. Ectopic variceal bleeding. In Westaby D (ed). Variceal Bleeding. Philadelphia:
WB Saunders, 1992;95110.
105.
Wolfsen HC, Kozarek RA, Bredfeldt JE, et al. The role of endoscopic injection sclerotherapy
in the management of bleeding peristomal varices. Gastrointest Endosc 1990;36:4724.
106.
Kozarek RA, Botoman VA, Bredfeldt JE, et al. Portal colopathy: prospective study of
colonoscopy in patients with portal hypertension. Gastroenterology 1991;101:11927.
107.
Stiegmann GV, Goff JS, Michaletz-Onody PA, et al. Endoscopic sclerotherapy as compared
with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992;326:152732..
Chapter 34 Prophylactic Sclerotherapy

(2219)
(2220)
KARL-JOSEPH PAQUET, M.D., F.I.C.S.
Bleeding from gastroesophageal varices is a major cause of death among patients with cirrhosis and
portal hypertension. For many decades, interest in the prevention of variceal bleeding has been
increasing. The opinion that bleeding could be prevented by decompression shunt surgery was tested
in a series of controlled trials during the early 1970s. Although the incidence of bleeding was effectively
reduced, no substantial improvement in survival could be achieved. The benefit of a reduction in the
rate of bleeding was balanced by an increased incidence of severe encephalopathy, liver failure, and
death.1(2221) Four alternative modalities are available for the prevention of variceal hemorrhage in
addition to shunt surgery: abstinence from alcohol, drug treatment, endoscopic variceal
sclerotherapy,24(2222) and nondecompression surgery.5(2223)
Data from controlled trials of portal decompression surgery performed in unselected patients with liver
cirrhosis and portal hypertension indicate that the 2-year probability of bleeding from varices is 20 to
30%.1(2224) In evaluating prophylactic treatment modalities, identifying groups of patients in whom the
risk of bleeding is highgreater than 60%is therefore essential. Otherwise, if all patients with liver
cirrhosis and portal hypertension were treated prophylactically, two thirds of them would be treated
unnecessarily.
Severe hemorrhage from gastroesophageal varices remains one of the most serious complications in
patients with cirrhosis and portal hypertension. The mortality rate associated with a first episode of
hemorrhage ranges from 30 to 80%.6(2225) Therefore, it is logical to attempt to prevent the initial
hemorrhage in an effort to improve the survival of patients with gastroesophageal varices. Knowledge
of the pathophysiology of variceal hemorrhage and of which varices are at high risk of bleeding is
essential to this effort and forms the basis for deciding which modality will be most effective in
preventing variceal hemorrhage, including the role of endoscopic sclerotherapy.
The topics of variceal anatomy and pathophysiology are covered in detail in Chapter 31: Variceal
Bleeding. Because these are extremely important considerations in the selection of patients for
prophylactic treatment, the essential features of pathologic variceal anatomy are summarized in this
chapter. The modalities for prevention of variceal bleeding are then described, concentrating on
prophylactic endoscopic sclerotherapy.
Variceal Anatomy
Butler7(2226) described the esophageal venous system as having three main components: intrinsic
veins, associated veins, and extrinsic veins. The extrinsic and intrinsic veins are interconnected at
frequent intervals along the length of the esophagus by perforating veins, which have valves that
normally prevent the flow of blood from extrinsic veins into the intrinsic plexus. In patients with portal
hypertension, the muscularis mucosae proximal to the esophagocardiac junction is replaced by
submucosal varices that connect via subepithelial or intraepithelial channels to the surface epithelium,
these channels being visible endoscopically as so-called red color signs (see Chapter 31: Variceal
Bleeding). The submucosal varices are connected to the extrinsic veins via perforating veins.
Figure 341 illustrates the venous system at the esophagocardiac junction in portal
hypertension.79(2227) The intraesophageal collaterals are located mainly in the lamina propria of the
esophageal wall. Proximal to the cardia for a distance of 2 or 3 cm in a circumferential segment of the
esophagus, the varices perforate the muscularis mucosae.9(2228) These submucosal varices are
separated from the esophageal lumen only by the epithelial layer, which becomes progressively more
atrophic as portal hypertension develops.10,11(2229) Small subepithelial and intraepithelial channels,
sometimes called minivarices, originate from the submucosal veins. The subepithelial and
intraepithelial channels are filled with blood and are identical to the lesions described originally by
Dagradi et al.12(2230) as cherry-red spots, by our group as telangiectasias,13,14(2231) and by Beppu
at el.15(2232) as red wale markings or hematocystic spots. These lesions are prone to rupture.
(2233)Figure 341. Schematic illustration of esophagocardiac junction in portal hypertension.

The submucosal veins replace the muscularis mucosa above the esophagocardiac junction
(thick arrows). Subepithelial and intraepithelial channels (triangles) represent red color signs
or minivarices. Perforating veins (thin arrows) connect extrinsic and intrinsic systems of
veins. Eepithelium; Lplamina propria; Mmmuscularis mucosae; Submsubmucosa;
Mpmuscularis propria; Sserosa with periesophageal varices. (Modified from Noda T.
Angioarchitectural study of oesophageal varices with special reference to variceal rupture.
Virchows Arch A Pathol Anat Histopathol 1984; 404:38192.)
The submucosal and subepithelial veins, mainly in the distal third of the esophagus, are connected via
perforating veins with the periesophageal veins (see Figure 341). These communicating veins are
important with respect to the hemodynamics of variceal blood flow.10(2234) They have valves that
normally direct blood flow away from the lumen of the esophagus. In the case of portal hypertension,
the valves of the perforating veins are incompetent and permit flow from the extrinsic to the intrinsic
veins, as shown in studies using contrast medium and others using endoscopic Doppler
sonography.10,16(2235) During respiratory expiration, blood flow is directed in the opposite direction
toward the lumen of the esophagus. Blood volume and pressure can therefore change rapidly within
the wall of the esophagus; this and the associated turbulence are predisposing factors in variceal
rupture.
Endoscopic Features
Histologic, radiologic, and Doppler flow studies10,11,16(2236) have demonstrated that the red color
signs detectable at endoscopy correlate with blood flow through dilated subepithelial communicating
veins that connect the periesophageal extrinsic veins with the more superficial submucosal veins
(Figure 342A). Dilated subepithelial veins, which appear as raised red areas (cherry-red spots,
telangiectasias, minivarices, red wale markings), are generally about 1 to 2 mm in diameter.17(2237)
When present, these lie at the top of large dilated subepithelial or submucosal veins, which are usually
more than 5 mm in diameter. (Figure 342B and C).
(2238)Figure 342. A, Diagram of the venous circulation in the distal esophagus in portal
hypertension. CRScherry-red spots; RWMwale marking; HCShemocystic spots. B,
Large varices without red color signs. C, Large varices that also have very prominent red color
signs including cherry-red spots, telangiectasias, red wale markings, and a hemocystic spot. (B
and C, From Snady H. The role of sclerotherapy in the treatment of esophageal varices:
Personal experience and a review of randomized trials. Am J Gastroenterol 1987;
82[9]:81322.)
Under certain conditions, the muscularis mucosae of the distal esophagus may atrophy in some areas,
resulting in a confluence of lamina propria and submucosa. Enlarged submucosal veins can then
appear above the muscularis mucosae. The hematocystic spot, a slightly raised bleb approximately 4
mm in diameter, appears to represent blood flow coming from the deeper extrinsic veins straight
inward toward the esophageal lumen via communicating veins at a point where the submucosal veins
are most superficial in relation to the esophageal lumen. Doppler flow studies10,18(2239) have
demonstrated that this arrangement may result in stasis at the site of communicating veins, with the
possibility of eruption and massive hemorrhage. This sequence of events has been likened to the
eruption of a volcano.
When red color signs are present, the frequency distribution of the various lesions is as follows:
cherry-red spots, minivarices, or telangiectasias can be seen in 90%, red wale markings in 50%, and
hematocystic spots in 8%. These endoscopic signs may represent channels in the epithelium rather
than the subepithelium.11,19(2240) Endoscopic variceal sclerotherapy therefore appears to work by
obliterating these abnormalities of the variceal wall and submucosal veins, thus diverting blood flow to
deeper collateral veins.18(2241)
The relationships among endoscopic observations in portal hypertension (e.g., red color signs,
minivarices, telangiectasias), the esophageal venous anatomy, and the portal hemodynamics, as
described by numerous authors, are illustrated conceptually in Figure 342. In the distal 5 cm of
esophagus, the large varices in the submucosa tend to be very close to the lumen, even in the lamina
propria rather than only the submucosa under certain conditions. More perforating or communicating
veins are also seen in this area. In 90% of cases, variceal bleeding occurs in the distal 6 to 8 cm of the
esophagus.14,20(2242)
That red color signs are an important prognostic indicator for variceal bleeding has been confirmed in
numerous retrospective and prospective studies.13,2025(2243) However, variceal diameter is also a

prognostic factor; the risk of recurrent bleeding after a first variceal hemorrhage is higher with
large-diameter varices.20(2244) That variceal size is a prognostic indicator of bleeding and survival of
patients with mainly alcoholic liver cirrhosis has recently been confirmed by several
groups.26,27(2245)
Snady and Feinman28(2246) described a numerical scoring system for esophageal varices that is
similar to that of The Japanese Research Society for Portal Hypertension.15(2247) Elements of this
system include shape and size of varices, location, and fundamental color. Using these parameters,
varices were given an aggregate grade of 1 to 10; those with a score of greater than 8 were considered
high grade and those with a score of less than 7 low grade. The predictive value of this grading system
was studied prospectively in 52 patients who received a variety of treatments, including sclerotherapy,
as well as 15 patients who received no treatment. The frequency of bleeding was significantly high
during the study (mean duration 26 months) for treated as well as untreated patients. The predictive
value of endoscopic findings, mainly variceal size and red color signs, has been verified in numerous
studies over the years.1315,1922,2932(2248)
Portal Pressure
A relationship no doubt exists between elevated portal pressure and variceal hemorrhage, although the
exact nature of this relationship remains controversial. A certain level of portal pressure, specifically a
portohepatic gradient of 10 to 12 mm Hg, is required for the development of esophageal varices. At
levels lower than this, the presence of varices is unlikely.20,33(2249) In patients with early-stage liver
disease, the presence of varices correlates with the degree of portal hypertension.34(2250) However,
in patients with established portal hypertension, whether a relationship exists between variceal
hemorrhage and the level of portal venous pressure is still open to question.35,36(2251) That such a
relation exists is doubted for several reasons: methodologic problems exist with the measurement of
portal pressure, including wedged hepatic venous pressure; intravariceal pressure may be of greater
importance than portal pressure with respect to risk of variceal bleeding;37(2252) esophageal varices
occur a distance from the portal vein, and other factors may contribute to vascular resistance.
Rigau et al.24(2253) found that intravariceal pressure in patients with esophageal variceal bleeding
was significantly higher than that in patients without bleeding. Similar results were obtained in earlier
studies by Staritz et al.37,38(2254) These authors compared 16 patients with compensated cirrhosis
following variceal hemorrhage with 13 similar patients without prior bleeding; those with a history of
bleeding had a median intravariceal pressure of 27 mm Hg, versus 16 mm Hg in those without prior
bleeding. We also compared median intravariceal pressure in 14 patients with compensated cirrhosis
and a history of bleeding with that in 15 similar patients who had no prior bleeding [Paquet,
unpublished data]; median intravariceal pressure in the former group of patients was 22 mm Hg,
versus 15 mm Hg in the latter group. Thus, intravariceal pressure in excess of 20 mm Hg indicates a
high risk of variceal bleeding (Figure 343).
(2255)Figure 343. Relationship between variceal pressure, wall tension, and prediction of
variceal hemorrhage in cirrhotic patients with and without a history of variceal bleeding.
(From Paquet K-J, Kuhn R. Prophylactic endoscopic sclerotherapy in patients with liver
cirrhosis, portal hypertension and esophageal varices. Hepatogastroenterology 1997;
44:62536.)
Ready et al.39(2256) assessed the risk of bleeding from esophageal varices with continuous
monitoring of portal pressure by means of an indwelling hepatic vein balloon catheter in patients with
alcoholic cirrhosis during and after episodes of variceal bleeding. Portal pressure was significantly
higher than 16 mm Hg in patients who had persistent or recurrent bleeding, compared with patients
who remained stable. Ready et al.39(2257) concluded that continuous monitoring of portal pressure
was possible and safe in patients with bleeding esophageal varices due to alcoholic cirrhosis and that
the risk of bleeding could be stratified on this basis.
Nearly all investigators who have measured esophageal variceal pressure, albeit by different methods,
agree that pressure is always higher in large than in small varices.37,38,4043(2258) Furthermore,
intravariceal pressure is influenced by a variety of factors, including respiration, the Valsalva maneuver,
and vomiting.3740(2259) According to the Law of Laplace, with elevated pressure in esophageal
varices, the tension and thickness of the variceal wall may be the decisive factors in relation to the risk
of bleeding.44(2260) This law states that the tension (T) of the wall of a vessel must be proportional to
pressure (P) and radius (R); furthermore, the tension (T) is inversely proportional to the thickness (TH)
of the wall:
T = P R: TH
The combination of these factorsvariceal size, wall thickness and tension, and pressure within the
varixhas greater prognostic value than any one of these factors alone.
Patients with poor hepatic function undoubtedly have a higher incidence of bleeding. Factors other than
portal pressure and variceal anatomy must therefore be considered as predisposing to variceal
hemorrhage, such as ascites, coagulation disorders, and localized increases in fibrogenolysis; these
are included in our Heinz-Kalk modified classification of hepatic reserve (Table 341).
TABLE 341
Heinz-Kalk Classification of Hepatocellular Function

POINTS
OBSERVATIONS
Nutritional status
Ascites
1
Excellent
None
Good
Moderate, easily
controlled
I-II
2.03.0
3.03.5
5075%
Poor
Marked, poorly
controlled
III-IV
>3.0
<3.0
<50%
Neurologic disorder (encephalopathy)

None
Serum bilirubin (mg/dl)
<2.0
Serum albumin
>3.5
Prothrombin concentration
>75%
Child A = 68 points
Child B = 911 points
Child C = 12 or more points
Modified from Paquet K-J, Kalk J-F, Klein C-P, Gad HA. Prophylactic sclerotherpy for
esophageal varices in high-risk cirrhotic patients selected by endoscopic and
hemodynamic criteria: A randomized, single-center controlled trial. Endoscopy 1994;
26:73440. 1994 Springer-Verlag.
Prophylactic Treatment Modalities

The prediction of bleeding from varices has been attempted by numerous investigators with varying
degrees of success.3,1315,1821,2325,27,33,34,37,41(2261) These efforts have been based on
size of varices, endoscopic appearance of varices, level of portal or esophageal variceal pressure, and
other clinical factors such as Child classification. Which factors are most essential for assessing the
risk of bleeding? Furthermore, which criteria should be included in trials of prophylactic treatment of
esophageal varices?
From our point of view, the main factors are as described earlier: size of the varices (at least degree III
to IV according to our classification), endoscopic appearance of varices (the presence of
telangiectasias, minivarices, cherry-red spots, hematocystic spots, or red wale markings), and
esophageal variceal or portal pressure (>16 to 20 mm Hg). Because the measurement of portal or
intravariceal pressure is an invasive procedure and because a correlation exists between variceal size
and intravariceal pressure, obtaining actual measurements is not absolutely necessary. Furthermore,
the level of pressure can also be inferred from noninvasive types of measurements such as Doppler
measurements of azygous blood flow.
Abstinence from Alcohol

Dagradi29(2262) was the first to demonstrate that, in patients with alcoholic cirrhosis, continuous
consumption of alcohol leads to sustained elevations of portal pressure and that abstinence results in a
reduction of portal pressure. Thus, the risk of bleeding is influenced by alcohol intake. These
observations have been confirmed by our group.45(2263)
Pharmacologic Therapy
The results of trials of therapy using -blocking agents such as propranolol or nadolol appear
favorable, especially those that included patients with well-compensated liver disease, although the
severity of cirrhosis differed widely between studies (Table 342). Based on data from a number of
trials of -blocking agents, effective prophylaxis of a first episode of upper gastrointestinal hemorrhage
seems to be possible in patients with large esophageal varices.4651 In addition, variceal size and red
color signs have been shown to be independent predictive factors for a first episode of bleeding, and
the first episode usually occurs within 1 year of diagnosis of the esophageal varices.27,51(2264) These
clinical studies show that the continuous administration of blockers decreases the risk of
gastrointestinal bleeding due to portal hypertension and may prolong survival in patients with cirrhosis
with large esophageal varices. Therefore, treatment with -blocking agents can be proposed for
compliant cirrhotic patients without severe ascites.
Results of Six Controlled Trials of Pharmacologic Prevention of the First Bleeding

Patients With Cirrhosis and Large Esophageal Varices
TABLE 342
ANDRANI
ET AL.46
BOSCH ET
AL.47
Blocker
Propranolol
Dose (mg/day)
(mean range)
Number of patients
blocker/control
Alcohol as cause of
cirrhosis (%)
Severity of cirrhosis
Child-Pugh A (%)
Child-Pugh B (%)
Child-Pugh C (%)
92 26
(40320)
84
43/41
80
Long-acting
propranolol
132
(40320)
102
51/51
78
25
50
25
45
49
6
IDO ET
AL.48
Nadolol
80
(80120)
79
30/49
52
7/5
LEBREC
ET AL.49
Nadolol
96
(40160)
106
53/53
74
58
42
25
PASCAL ET
AL.50
Long-acting
propranolol
162 85
(40320)
230
118/112
90
17
37
46
Results of Six Controlled Trials of Pharmacologic Prevention of the First Bleeding

Patients With Cirrhosis and Large Esophageal Varices
TABLE 342
ANDRANI
ET AL.46
BOSCH ET
AL.47
IDO ET
AL.48
LEBREC
ET AL.49
PASCAL ET
AL.50
Bleeding (%)
Control
1 year
2 years
Blocker
1 year
30
39
22
19
30
21
25
61
6*
4*
0*
15
2 year
6*
6*
17 (3)
66
54
78
84
78
80
66
51
95
94
78
76*
72*
Survival rate (%)

Control
1 year
2 years
Blocker
1 year
79
84
2 years
66
* Significantly different from controls.

Significantly different from controls in compliant patients.
26*
Significantly different from controls in patients without large ascites.
Side effects occurred in certain patients receiving blockers, the incidence ranging from 5 to 40%.
Therapy was interrupted in only about 10% of the patients. Dizziness was the main side effect,
occurring in approximately 10% of the treated patients. Raynaud's syndrome and pulmonary or cardiac
complications occurred in less than 5%. Except for hepatic encephalopathy, these side effects were
similar to those observed in patients without liver disease who receive blockers. Although
encephalopathy occurred in some cases, this complication was mild and transient. In all the studies,
the resuscitation of patients receiving blockers who bled was no more difficult than that in untreated
patients.
Pagliaro et al.52(2265) reported a metaanalysis of nine randomized clinical trials of blockers. No
substantial heterogeneity was found in the pooled data. Propranolol and nadolol were assessed,
respectively, in three and two trials. The design of these trials varied from two-arm comparisons
between a blocker and a nonactive agent, a blocker and a control, to a four-arm trial of propran-olol
or sclerotherapy versus respective untreated control groups. Only two of the studies were fully double
blinded. The odds ratio for bleeding was less than 1.00 in eight of the nine trials, indicating a favorable
effect of treatment. This reached statistical significance in four trials. The odds ratio was greater than
1.00 in one small trial. When data from all seven trials were pooled, the odds ratio for bleeding was
0.54 (95% confidence interval [CI], 0.390.74). The effect of therapy with blockers was particularly
evident in patients with large or medium-sized varices and in those with an hepatic vein pressure
gradient above 12 mm Hg. The pooled odds ratio for death was not significant at 0.75 (CI, 0.571.06),
which the authors regarded as indicating only a trend toward reduced mortality.
A variety of other promising pharmacologic approaches to the management of portal hypertension are
currently under evaluation. The therapeutic role of calcium-channel blockers is uncertain.53(2266)

Organic nitrates reduce portal pressure in most patients with alcoholic cirrhosis. The effects are dose
dependent and are achieved by a reduction in intrahepatic resistance.5456(2267)
Isosorbide-5-mononitrate administered orally has been shown to be a safe and effective alternative to
propran-olol in the prophylaxis of bleeding in cirrhosis.57(2268) However, the long-term effects of this
agent in cirrhotic patients are uncertain. Serotonin antagonists (e.g., ketanserin) reduce portal pressure
in humans.58(2269) Ketanserin has been shown to precipitate encephalopathy. Combination therapy
using the agents mentioned earlier or vasoconstrictor agents (to reduce portal inflow) and vasodilator
agents (to reduce portal resistance) are under evaluation. The ideal portal hypotensive agent should
reduce portal vascular resistance, improve hepatic perfusion and function, and reduce the portal
pressure.
Prophylactic Sclerotherapy
The rebirth of endoscopic sclerotherapy as the treatment of choice for acute variceal hemorrhage has
prompted investigations of the use of sclerotherapy for the prevention of both recurrent and first
variceal hemorrhages. The first two controlled trials of prophylactic sclerotherapy, by our group in
198213(2270) and Witzel et al.59(2271) in 1985, indicated that prophylactic sclerotherapy not only
reduces the risk of first variceal hemorrhage but also improves survival. Although the results of these
studies generated considerable interest in prophylactic sclerotherapy, concerns were raised about the
unusually high incidence of variceal hemorrhage in the control groups: 66% in our study13(2272) and
57% in that of Witzel et al.59(2273) Only patients at high risk for variceal hemorrhage (large varices,
cherry-red spots, poor coagulation values) were selected for our study, and this may account for the
high incidence of variceal hemorrhage among control patients. However, Witzel et al.59(2274) included
all patients with esophageal varices without regard for specific risk factors. In addition, the withholding
of sclerotherapy for acute variceal bleeding in control patients might partially explain the
higher-than-usual mortality rate in the control groups in these studies.
Fifteen controlled trials of prophylactic sclerotherapy versus medical therapy have been published to
date; overall results of these studies have been mixed (Table 343).13,5973(2275) A metaanalysis
for both bleeding from esophageal varices and bleeding from all sources has clearly demonstrated the
marked heterogeneity among these studies.72(2276) Although these studies were designed to assess
the effect of sclerotherapy in the prevention of esophageal variceal hemorrhage, increasing evidence
indicates that patients whose esophageal varices are successfully obliterated or protected by scar
tissue by means of sclerotherapy may subsequently bleed from other sites such as gastric varices,
portal hypertensive gastropathy,74(2277) and portal hypertensive colopathy.75(2278) Because
obliteration of esophageal varices may increase the risk of bleeding from other sites, the data are
presented as bleeding from all upper gastrointestinal sources as well as documented esophageal
variceal hemorrhage wherever the data permitted (see Table 343).
TABLE 343
Endoscopic Sclerotherapy (ST) in the Prophylaxis of First Variceal Hemorrhage V

ALCOHOLICS
NO. OF
PATIENTS
STUDY
ST
SELECTION
CRITERIA
Child C
(%)
ST
Mean
Follow-Up
(%)
ST
BLEEDING (%
(months)
Overall
ST
TABLE 343
Endoscopic Sclerotherapy (ST) in the Prophylaxis of First Variceal Hemorrhage V

ALCOHOLICS
NO. OF
PATIENTS
SELECTION
Child C
(%)
BLEEDING (%
Mean
Follow-Up
(%)
Overall
ST
CRITERIA
ST
ST
(months)
ST
Paquet (1982)
32
33
24?
66
Witzel et al.
(1985)
Koch et al. (1994)
56
53
21
17
80
81
25
70
71
10
47
62
56
19
47
Wrdehoff and
Spech (1987)
Santangelo et al.
(1988)
25
24
16
21
48
42
44
20
63
49
46
88
91
13
35
15
Sauerbruch et al.
(1988)
Piai et al. (1988)
68
65
18
22
63
69
22
32
43
71
69
24
32
34
33
13
de Franchis et al.
(1991)
55
51
18
21
27
49
24
36
34.5
Veterans Affairs
Cooperative
Variceal
Sclerotherapy
Group (1991)
Triger et al.
(1991)
143
138
Cirrhosis, large
varices,
coagulation
factors <30%
Cirrhosis, small
and large varices
Cirrhosis, large
varices
Cirrhosis, large
varices
Chronic liver
disease, large
varices
Cirrhosis, large
varices
Cirrhosis, large
varices
High-risk cirrhotics
with endoscopic
bleeding risk
Cirrhosis >3
varices
34
26
100
100
22.5*
22
17
33
35
12
17
46
40
61
70
74
PROVA Study
Group (1991)
73
71
82
87
28
18
18
Paquet et al.
(1994)
44
45
73
69
33
32
78
STUDY
Cirrhosis portal
pressure: >12
mm Hg
Cirrhosis, more
small than large
varices
Cirrhosis, large
varices, portal
pressure >20
mm Hg
30
27
* Duration of study-treatment phase.

Upper gastrointestinal bleeding.
Median.
A significant reduction in the incidence of variceal hemorrhage in patients who underwent

sclerotherapy compared with patients treated medically was found in eight

studies.13,5964,66,68,73(2279) A survival advantage for patients treated with sclerotherapy was
noted in five studies.13,59,64,65,73(2280) One large study was terminated because the increase in
mortality in the sclerotherapy group reached statistical significance.70(2281) The main problem with
the latter study is that it is a multicenter trial, and one must suppose that expertise in endoscopic
sclerotherapy may have been lacking or at least variable among the several participating centers. This
presumption can explain why patients treated to prevent bleeding had a higher frequency of variceal
hemorrhage than did control patients.
Van Ruiswyk and Byrd72(2282) have reported a metaanalysis of prophylactic sclerotherapy for the
prevention of a first variceal hemorrhage, including eight reports of randomized controlled trials in
adults. The pooled estimate of the decrease in the rate for a first episode of variceal hemorrhage was
19% (95% CI, 237%); the pooled estimate of the decrease in death rate due to bleeding was 15%
(95% CI, 326%). Furthermore, the 13-month mortality rate was reduced by 11% (95% CI, 419%) for
patients undergoing prophylactic sclerotherapy, that is, a 41% relative reduction in mortality rate. A
positive correlation was found between reductions in the mortality and bleeding rates, and a negative
correlation was found between mortality and complication rates. Despite these favorable results, van
Ruiswyk and Byrd72(2283) concluded that prophylactic sclerotherapy was not suitable for wide
application because of the availability of less expensive treatment alternatives and the relatively high
complication rate associated with sclerotherapy. Furthermore, these authors asserted that the benefits
of sclerotherapy might be due in part to the more intensive follow-up afforded to endoscopically treated
patients.
Nineteen randomized trials of sclerotherapy were included in the metaanalysis of Pagliaro et
al.52(2284) These trials were found to be very heterogeneous with respect to parameters such as
quality of the trial, baseline characteristics of patients, treatment modalities, rate of bleeding for control
subjects, and outcome measures. However, a favorable effect for sclerotherapy on bleeding was
indicated by an odds ratio for bleeding of less than 1.00 in 14 trials (statistically significant in 5); the
odds ratio was 1.00 in 1 trial and higher than 1.00 (no favorable effect) in 4 trials (statistically significant
in 4). The pooled odds ratio was 0.6 (95% CI, 0.490.74). The odds ratio for risk of death was less
than 1.00 in 11 trials (statistically significant in 3); it was 1.00 in 1 trial and higher than 1.00 in 7 trials.
Statistical significance was achieved in only 1 of the 7 trials, with an odds ratio for death of greater than
1.00, the largest of the group. The pooled odds ratio for death was 0.76 (CI, 0.610.94) However,
Pagliaro et al.52(2285) noted that sclerotherapy was used to treat acute bleeding episodes in the
sclerotherapy groups of patients but not in control groups in 2 of the 3 trials in which the odds ratio for
death was significantly less than 1.00. This was regarded as a bias in favor of the sclerotherapy-treated
groups. Two trials were halted because of an increasing mortality rate in sclerotherapy-treated patients.
Pagliaro et al.52(2286) concluded that the statistically significant reduction in pooled odds ratio was
"meaningless" because of the marked heterogeneity of the pooled trials.
The metaanalysis of the effect of prophylactic sclerotherapy of Fardy and Laupacis74(2287) in 14
randomized controlled trials also found significant heterogeneity. When all 14 trials were taken
together, the odds ratio of death in sclerotherapy-treated versus control patients was 0.74 (95% CI,
0.600.93) in favor of treated patients. This problem of heterogeneity was not evident when trials were
pooled in subgroups according to sclerosing agent. With respect to mortality, sclerotherapy with
polidocanol resulted in a significant benefit for treated patients, whereas treatment with sodium
tetradecyl sulfate had a detrimental effect. When pooled data were analyzed according to Child class,
no benefit of sclerotherapy was found for patients with Child Class A liver disease. A further
metaanalysis based on 15 reports of controlled trials of sclerotherapy versus medical therapy is shown
in Figure 344.
(2288)Figure 344. A metaanalysis of sclerotherapy versus medical therapy for prevention of

first bleeding. Squares represent odds ratios (OR), horizontal bars represent 95% confidence
intervals. The odds ratios were calculated as OR = ad/bc, where a and c are the numbers of
bleeding patients and b and d the numbers of nonbleeding patients in the treatment and control
groups, respectively. VACV SGVeterans Affairs Cooperative Variceal Sclerotherapy
Group; PROVA SGProva Study Group.
Comparisons among these studies are obviously complicated by differences in entry criteria, that is,
large varices versus all varices, cause and severity of liver disease, varying techniques of
sclerotherapy and sclerosing agents, and percentage of patients with endoscopically proven sites of
bleeding. Often, studies showing sclerotherapy to be beneficial did not define a bleeding episode,
whereas studies that failed to show a difference had clear definitions. Two of five studies reporting a
survival benefit for sclerotherapy did not use sclerotherapy for treatment of acute variceal hemorrhage
in patients randomized to medical treatment.13,59(2289) Despite these problems, certain conclusions
can be reached: Although the initial studies tend to support the use of sclerotherapy for the prevention
of first variceal hemorrhage, later studies have failed to confirm the earlier findings. However, the more
recent studies demonstrate again a trend in favor of prophylactic sclerotherapy for the prevention of a
first episode of variceal bleeding.68,69,72,73(2290)
The data on patients with alcoholic liver disease are conflicting. Although the studies of Triger et
al.,69(2291) Ptzi et al.,65(2292) and Sauerbruch et al.63(2293) showed a survival benefit for the
subgroup of patients with alcoholic liver disease, the study conducted by the Veterans Affairs
Cooperative Variceal Sclerotherapy Group67(2294) found a significantly higher mortality in patients
with alcoholic liver cirrhosis treated by sclerotherapy. On the other hand, it is evident from these
studies that prophylactic sclerotherapy does not have a role in patients with nonalcoholic liver disease.
Conversely, the study of Russo et al.66(2295) demonstrates the opposite result. Whether patients with
decompensated (Child-Pugh B and C) alcoholic cirrhosis benefit from prophylactic sclerotherapy is
controversial and requires further investigation.
Because of the heterogeneity of the data and lack of a clear consensus, prophylactic sclerotherapy
should be adopted as a therapy to prevent first variceal hemorrhage only in the framework of a
controlled randomized trial.75(2296) Furthermore, only patients at high risk of bleeding (i.e., large
varices, endoscopic bleeding markers, and variceal or portal pressure greater than 16 mm Hg) should
be selected for such studies. Finally, prophylactic endoscopic sclerotherapy should be performed only
by expert endoscopists.
REFERENCES
1.
2.
3.
4.
5.
Chalmers TH. Prophylactic shunts. In Child CG III (ed). Portal Hypertension. Philadelphia: WB
Klein C-P, Kalk J-Fr, Mting D, Klein C-G. Einfluss von Alkohol auf die Hmodynamik der
Pfortader bei nutritiv-toxischer Leberzirrhose (Influence of alcohol on portal hemodynamics in
alcoholic cirrhosis). Dtsch Med Wochenschr 1993;118:8993.
Lebrec D, Poynard T, Bernuau J, et al. A randomized controlled study of propranolol for
prevention of recurrent gastrointestinal bleeding in patients with cirrhosis: A final report.
Lebrec D, Benhamou J-P. In Okuda K, Benhamou J-P (eds). Portal Hypertension. Clinical and
Physiological Aspects. Tokyo: Springer-Verlag, 1991;54961.
Inokuchi K, Cooperative Study Group of Portal Hypertension of Japan. Improved survival
following prophylactic portal non-decompression surgery for esophageal varices. Hepatology
1990;12:16.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Sherlock S. Diseases of the Liver and Biliary System. Oxford, London, Edinburgh, Boston,
Melbourne: Blackwell. 1989.
Butler H. The veins of the oesophagus. Thorax 1951;6:27696.
MacBeth R. Treatment of oesophageal varices in portal hypertension by means of sclerosing
injections. BMJ 1955;2:87780.
Noda T. Angioarchitectural study of oesophageal varices with special reference to variceal
rupture. Virchows Arch A Pathol Anat Histopathol 1984;404:38192.
McCormack TT, Smith PM, Rose JD, Johnson AG. Perforating veins and blood flow in
Spence RAJ, Sloan JM, Johnston GW, Greenfield A. Oesophageal mucosal changes in
patients with varices. Gut 1983;24:10249.
Dagradi AE, Stempien SE, Owens LK. Bleeding esophagogastric varices. An endoscopic
study of 50 cases. Arch Surg 1966;92:94452.
Paquet K-J. Prophylactic endoscopic sclerosing treatment of the esophageal wall in
varicesA prospective controlled randomized trial. Endoscopy 1982;14:45.
Oberhammer E, Paquet K-J, Distelmaier W. Endoskopische Befunde bei portaler
Hypertension unter Einschluss der Notfallendoskopie (Endoscopic findings in portal
hypertension including emergency endoscopy). Therapiewoche 1978;28:717883.
Beppu K, Inokuchi K, Koyanagi N, et al. Prediction of variceal hemorrhage by esophageal
Kurtz W, Classen M. Messung des Blutflusses in sophagusvarizen mit einem
endoskopischen Ultraschall-Doppler (Measurement of blood flow in esophageal varices using
an endoscopic sonographic Doppler system). Dtsch Med Wochenschr 1984;109:8214.
Snady H. The role of sclerotherapy in the treatment of esophageal varices: Personal
experience and a review of randomized trials. Am J Gastroenterol 1987;82:81322.
Onishi K. Doppler flowmetry and portal hypertension. Gastroenterology 1988;94:12423.
The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices.
Lebrec D, De Fleury P, Rueff B, et al. Portal hypertension, size of esophageal varices, and
risk of gastrointestinal bleeding in alcoholic cirrhosis. Gastroenterology 1980;79:113944.
Palmer ED, Brick IB. Correlation between the severity of esophageal varices in portal cirrhosis
and their propensity toward hemorrhage. Gastroenterology 1956;30:8590.
Baker LA, Smith C, Lieberman G. The natural history of esophageal varices: A study of 115
cirrhotic patients in whom varices were diagnosed prior to bleeding. Am J Med
1959;26:22837.
Garcia-Tsao G, Groszmann RJ, Fisher RL, et al. Portal pressure, presence of
gastro-oesophageal varices and variceal bleeding. Hepatology 1985;5:41924.
Rigau J, Bosch J, Bordas JM, et al. Endoscopic measurement of variceal pressure in
1989;96:87380.
Rector WG, Reynolds TB. Risk factors for hemorrhage from esophageal varices and acute
gastric erosion. Klin Gastroenterol 1985;14:13953.
Merkel C, Bolognesi M, Angeli P, et al. Prognostic indicators of survival in patients with
cirrhosis and esophageal varices, without previous bleeding. Am J Gastroenterol
1989;84:71722.
Cals P, Groupe d'Etude de la Prophylaxie Primaire. Facteurs prditifs de la premire
hmoragie digestive et de la mortalit chez le malades cirrhotiques avec varices
oesphagiennes. Gastroenterol Clin Biol 1989;13:549.
Snady H, Feinman L. Prediction of variceal hemorrhage: A prospective study. Am J
Dagradi AE. The natural history of esophageal varices in patients with alcoholic liver cirrhosis.
An endoscopic and clinical study. Am J Gastroenterol 1972;57:52040.
30. Rigo GP, Merighi A, Chahin NJ, et al. A prospective study of the ability of three endoscopic
classifications to predict hemorrhage from esophageal varices. Gastrointest Endosc
1992;38:4259.
31.
Piai G, Minieri M, Catalano M, et al. Prospective validation of two incidences of first variceal
bleeding. Gastroenterology 1991;100:A784.
32.
Prada A, Bortoli A, Minoli G, et al. The prediction of esophageal variceal bleeding. NIEC score
validation by an independent group. Gastroenterology 1990;98:A109.
33. Viallet A, Marleau D, Huet M, et al. Hemodynamic evaluation of patients with intrahepatic
portal hypertension. Relationship between bleeding varices and the portohepatic gradient.
34. Reynolds TB. Interrelationships of portal pressure, variceal size, and upper gastrointestinal
bleeding. Gastroenterology 1980;79:13323.
35. Vinel JP, Cassigneul J, Louis A, et al. Clinical and prognostic significance of portohepatic
gradient in patients with cirrhosis. Surg Gynecol Obstet 1982;155:34752.
36. Sauerbruch T, Kleber G, Gerbes A, et al. Prophylaxis of first variceal hemorrhage in patients
with liver cirrhosis. Klin Wochenschr 1986;64:126775.
37.
Staritz M, Manns M, Poralla T, Meyer zum Buschenfelde KH. A new method for measurement
of the intravascular oesophageal pressure (IOVP) and assessment of IOVP in cirrhotic
patients with and without recent variceal haemorrhage. Gastroenterology 1985;A88:1697.
38. Staritz M, Meyer zum Buschenfelde KH. The endoscopic measurement of intravascular
pressure and flow in oesophageal varices. J Hepatol 1988;7:12631.
39. Ready JB, Robertson AD, Gow JS, Rector WG. Assessment of the risk of bleeding from
esophageal varices by continuous monitoring of portal pressure. Gastroenterology
1991;100:140310.
40. Bosch J, Bordas JM, Rigau J, et al. Noninvasive measurement of the pressure of esophageal
varices using an endoscopic gauge: Comparison with measurements by variceal puncture in
patients undergoing endoscopic sclerotherapy. Hepatology 1986;6:66772.
41. Gertsch P, Loup P, Diserens H, et al. Endoscopic non-invasive manometry of esophageal
varices: Prognostic significance. Am J Surg 1982;144:52830.
42. Gross E, Erhard J. Endoskopisch gefhrte Druckmessung in distalen sophagusvarizen
(Endoscopically guided pressure measurement in the terminal esophagus). Dtsch Med
Wochenschr 1987;112:1257.
43. Schneider B, Lackner K. Pathogenese der sophagusvarizenblutung. Dtsch Med
44. Polio J, Groszmann RJ. Haemodynamic factors involved in the development and rupture of
oesophageal varices. A pathophysiological approach to treatment. Semin Liver Dis
1986;6:31831.
45. Mting D, Kalk J-F, Fischer R, Wiewel D. Spontaneous regression of oesophageal varices
after longterm conservative treatment. Retrospective study in 20 patients with alcoholic liver
cirrhosis, posthepatitic cirrhosis and haemochromatosis with cirrhosis. J Hepatol
1990;10:15862.
46. Andrani T, Poupon RE, Balkau BJ, et al. Preventive therapy of first gastrointestinal bleeding
in patients with cirrhosis: Results of a controlled trial comparing propranolol, endoscopic
sclerotherapy and placebo. Hepatology 1990;12:14139.
47.
Bosch J, Groszmann RJ, Grace N, et al. Propranolol in the prevention of the first hemorrhage
from esophageal varices: Results of a randomised, double-blind, cooperative clinical trial
[Abstract]. J Hepatol 1988;7(suppl I):S12.
48. Ido G, Bellati G, Feszie E, Grimoldi D. Nadolol can prevent the first gastrointestinal bleeding
in cirrhotics: A prospective, randomised study. Hepatology 1988;8:69.
49. Lebrec D, Poynard T, Capron JP, et al. Nadolol for prophylaxis of gastrointestinal bleeding in
patients with cirrhosis. A randomised trial. J Hepatol 1988;7:11825.
50. Pascal J-P, Cals P, Multicenter Study Group. Propranolol in the prevention of first upper
gastrointestinal tract hemorrhage in patients with cirrhosis of the liver and esophageal varices.
N Engl J Med 1987;317:85661.

51. The Italian Multicenter Project for Propranolol in Prevention of Bleeding. Propranolol prevents
first gastrointestinal bleeding in non-ascitic cirrhotic patients. Final report of a multicenter
randomised trial. J Hepatol 1989;9:7583.
52. Pagliaro L, D'Amico G, Sorensen TI, et al. Prevention of first bleeding in cirrhosis. A
meta-analysis of randomized trials of nonsurgical treatment. Ann Intern Med 1992;117:5970.
53.
Freeman JG, Barton JR, Record CO. Effect of isosorbide dinitrate, verapamil, and labetalol on
portal pressure in cirrhosis. Br J Med 1985;291:5612.
54. Kong C-W, Lay C-S, Tsai Y-T, et al. The hemodynamic effect of verapamil on portal
hypertension in patients with postnecrotic cirrhosis. Hepatology 1986;6:4236.
55. Blei AT. Vasodilator therapy of portal hypertension: Focus on the liver. Hepatology
1989;9:8969.
56. Navasa M, Chesta J, Bosch J, Rodes J. Reduction of portal pressure by
isosorbide-5-mononitrate in patients with cirrhosis. Gastroenterology 1989;96:11108.
57.
Angelico M, Carli L, Piat C, et al. Isosorbide-5-mononitrate versus propranolol in the
prevention of first bleeding in cirrhosis. Gastroenterology 1993;1043:14605.
58. Hadengue A, Morro R, Serini R, et al. Combination of ketanserin and verapamil or propranolol
in patients with alcoholic cirrhosis: Search for an additive effect. Hepatology 1989;9:835.
59. Witzel L, Wolbergs E, Merki H. Prophylactic endoscopic sclerotherapy of oesophageal
varices. A prospective controlled study. Lancet 1985;1:7735.
60. Koch H, Binmoeller KF, Grimm H, et al. Prophylactic sclerotherapy for esophageal varices:
Long-term results of a prospective study. Endoscopy 1994;26:72933.
61. Wrdehoff D, Spech HJ. Prophylaktische sophagusvarizensklerosierung. Ergebnisse einer
prospektiven, randomisierten Langzeitstudie ber 7 Jahre (Prophylactic esophageal
sclerotherapy. Results of a prospective, randomized long-term follow-up study over seven
years). Dtsch Med Wochenschr 1987;112:94751.
62. Santangelo WC, Dueno MI, Estes BL, Krejs GJ. Prophylactic sclerotherapy of large
63. Sauerbruch T, Wotzka R, Kopcke W, et al. Prophylactic sclerotherapy before the first episode
of variceal hemorrhage in patients with cirrhosis. N Engl J Med 1988;319:815.
64. Piai P, Cipolletta L, Claar M, et al. Prophylactic sclerotherapy of high risk esophageal varices:
Results of a multicentric prospective controlled trial. Hepatology 1988;8:14951500.
65. Ptzi R, Bauer P, Schofl R, et al. Prophylactic endoscopic sclerotherapy of esophageal
varices in liver cirrhosislong-term follow-up and final results of a multicenter prospective
controlled randomized trial in Vienna. Endoscopy 1993;25:2879.
66. Russo A, Giannone G, Magnagno A, et al. Prophylactic sclerotherapy in non-alcoholic liver
cirrhosis: Preliminary results of a prospective controlled randomized trial. World J Surg
1989;13:14953.
67. Kobe E, Zipprich B, Schenke KU, Nilius R. Prophylactic endoscopic sclerotherapy of
esophageal varicesA prospective randomized trial. Endoscopy 1990;22:2458.
68. de Franchis R, Primignani M, Arcidiacono PG, et al. Prophylactic sclerotherapy in high-risk
cirrhotics selected by endoscopic criteria. A multicenter randomized controlled trial.
69. Triger DR, Smart HL, Hosking SW, Johnson AG. Prophylactic sclerotherapy for esophageal
varices: Longterm results of a single-center trial. Hepatology 1991;13:11723.
70. Veterans Affairs Cooperative Variceal Sclerotherapy Group. Prophylactic sclerotherapy for
esophageal varices in men with alcoholic liver disease. A randomized, single-blind,
multicenter clinical trial. N Engl J Med 1991;324:177984.
71. The Prova Study Group. Prophylaxis of first hemorrhage from esophageal varices by
sclerotherapy, propranolol or both in cirrhotic patients: A randomized multicenter trial.
Hepatology 1991;14:101624.
72. van Ruiswyk J, Byrd JC. Efficacy of prophylactic sclerotherapy for prevention of a first variceal
hemorrhage. Gastroenterology 1992;102:58797.
73. Paquet K-J, Kalk J-F, Klein C-P, Gad HA. Prophylactic sclerotherapy for esophageal varices
in high-risk cirrhotic patients selected by endoscopic and hemodynamic criteria: A
randomized, single-center controlled trial. Endoscopy 1994;26:73440.
74. Fardy JM, Laupacis A. A meta-analysis of prophylactic endoscopic sclerotherapy for
esophageal varices. Am J Gastroenterol 1994;89:193848.
75.
DeFranchis R. Portal hypertension II. Proceedings of the second Bavano International
Workshop on Definitions, Methodology and Therapeutic Strategies. Oxford, Blackwell
1996;126..
Chapter 35 Complications of Sclerotherapy

(2297)
(2298)
JAKE E. J. KRIGE, F.R.C.S., F.C.S.(S.A.)
PHILIPPUS C. BORNMAN, M.MED., F.R.C.S.
JOHN TERBLANCHE, CH.M.,F.R.C.S., F.C.S.(S.A.)
Endoscopic variceal sclerotherapy is widely accepted as the treatment of choice for control of acute
variceal hemorrhage, and repeated injections have been shown to reduce the frequency of subsequent
episodes of bleeding.13(2299) Although sclerotherapy is perceived as the least invasive definitive
treatment for variceal bleeding, the procedure is not without risk. More than 40 different complications
have been described.48(2300) Important factors influencing the complication rate are the experience
of the endoscopist; injection technique; use of ancillary devices, including overtubes or balloon
tamponade; and whether sclerotherapy is performed as an emergency or elective procedure. Other
interrelated anatomic factors are the proximity of the esophagus to vital mediastinal structures,
repetitive breaching of the mucosa, and the potential for pulmonary and systemic spread of sclerosant
through portal venous collaterals.7,9(2301)
Incidence
There is no uniformity or universal agreement regarding the definition or classification of the
complications that occur after variceal sclerotherapy, and consequently, the incidence varies widely in
reported studies. Comparative analyses of complication rates in many series are hampered by
variations in patient population, type and severity of liver disease, and sclerotherapy technique used.10
The differences in study design introduce a covert selection bias that may influence results.
Complication rates are also higher when carefully documented in prospective studies. Some studies
express complication rates in terms of incidence per patient treated; others describe complications per
procedure performed. Surprisingly, in some prospective studies, details of sclerotherapy complications
are omitted, but others only consider major events. In long-term studies, repeated injections also
increase the cumulative risk of sclerotherapy-induced complications in the individual patient.3,6(2302)
The most reliable data indicate that 10 to 15% of patients undergoing sclerotherapy develop a major
complication (Table 351), but less than 1% of patients die as a direct result of the procedure.
TABLE 35-1 Incidence of Complications After Injection Sclerotherapy (Part i of ii )

STUDY
Kjaergaard et al.206
207
Copenhagen project
YEAR
NUMBER OF
PATIENTS
SCLEROSANT
METHOD OF
INJECTION
1982
61
3% P
PV
1984
93
3% P
PV
INCIDENCE OF
REBLEEDING (%
TABLE 35-1 Incidence of Complications After Injection Sclerotherapy (Part i of ii )

YEAR
NUMBER OF
PATIENTS
SCLEROSANT
METHOD OF
INJECTION
Korula et al.208
1985
63
1.5% STS
IV
Sarles et al.18
1985
50
1.5% STS
IV
Larson et al.19
1986
44
1.5% STS
IV
Terabayashi et al.20
1987
33
5% EO
IV
Fleig et al.209
1987
36
1% P
PV
Alexandrino et al.210
1988
31
5% EO
IV
El-Zayadi et al.211
1988
63
5% EO
PV
Magnano et al.212
1988
104
1% P
IV + PV
Van Hootegem et al.213
1988
133
1% P
IV
Low et al.21
1989
75
1.5% STS
IV
Kahn et al.6
1989
304
5% EO
IV + PV
McKee et al.214
1991
163
5% EO
IV
STUDY
INCIDENCE OF
REBLEEDING (%
(Continued In Part
TABLE 35-1 Incidence of Complications After Injection Sclerotherapy (Part ii of ii )

STUDY
Kjaergaard et al.206
ULCERATION
(%)
ASPIRATION OR
PNEUMONIA (%)
PLEURAL
EFFUSION (%)
STRICTURES
(%)
NA
18
31
NA
NA
NA
208
70
NA
18
78
NA
NA
20
87
15
15
NA
58
NA
NA
11
44
NA
25
16
10
19
49
NA
NA
Van Hootegem et al.
NA
NA
NA
21
61
10
20
41
NA
11
207
Copenhagen project
Korula et al.
Sarles et al.
19
Larson et al.
20
Terabayashi et al.
209
Fleig et al.
210
Alexandrino et al.
211
El-Zayadi et al.
212
Magnano et al.
213
Low et al.
Kahn et al.
9
24
NA
5
McKee et al.
Ppolidocanol; PVparavariceal; IVintravariceal; NAnot applicable; STSsodium tetradecyl sulfate; EOethanolami
oleate.
214
In this chapter, sclerotherapy-induced complications are categorized as local effects involving the
esophagus, regional respiratory and cardiovascular effects, and distant consequences. We define
minor events as those that are self-limiting and do not require specific treatment or interfere with an
injection program. Major complications are serious or life-threatening events that prolong
hospitalization.
Esophageal Complications
Morphologic Changes
Esophageal complications of variceal sclerotherapy are invariably a consequence of excessive
sclerosant-induced submucosal or transmural necrosis. The few studies that have examined the local
histopathologic effects of sclerosants on the esophageal wall in detail have been based on necropsy
findings. Although the injection techniques, type and volume of sclerosant used and intervals between
injections vary in these studies, the histopathologic findings are remarkably similar and provide a
time-dependent morphologic profile of the effects of sclerotherapy on the esophagus.1114(2303)
The earliest changes in the esophageal wall during the initial 48 hr after injection are thrombosis in
superficial veins, submucosal edema, and minor areas of tissue necrosis.11(2304) Mucosal ulceration
is uncommon during this phase, and no significant cellular reaction occurs.15(2305) After 48 hr,
progressive tissue necrosis occurs, predominantly in the superficial layers and to a lesser extent in the
deeper tissues. During the first week, mucosal ulceration and a marked acute polymorphonuclear
leukocyte inflammatory response occurs and is followed by an intense macrophage and fibroblast
infiltration.11,16 Intramural microabscesses may develop as a consequence of localized necrotizing
inflammation.15(2306) In some cases, extravasation of sclerosant into the submucosa and muscle
layers results in a giant cell reaction and focal calcification.14(2307) Some residual varices remain
patent, but others contain thrombi in the early stages of endothelial and fibroblastic
organization.13,17(2308)
The extent of sclerotherapy-induced ulceration varies from small, linear, superficial defects to
extensive, wide-based ulcers.13(2309) Although most ulcers are limited to the submucosa or inner
layer of the muscularis propria, a few extend more deeply into the muscularis propria. A fourth of
necropsy specimens show transmural necrosis that could have progressed to mediastinitis.13(2310)
The chronic reaction is characterized by an evolution from granulation tissue to mature collagen, with
an accompanying chronic inflammatory cell infiltrate that becomes less prominent with time.16(2311)
Necrosis and ulceration may persist for as long as 3 weeks. Organized thrombi and fibrosis become
evident 1 month after injection.12,13(2312) Fibrosis is usually limited to the submucosa and the inner
muscularis propria but may occur as a localized transmural breach in muscle or as diffuse transmural
fibrosis encasing residual varices.16(2313) Marked thickening of the esophageal wall is seen in some
specimens.15(2314)
Ulceration
Small areas of superficial mucosal ulceration are a common finding in the lower esophagus after
sclerotherapy.1821(2315) Some investigators have considered ulceration as an inevitable and
necessary consequence of effective sclerotherapy.22,23(2316) The reported incidence ranges from 9
to 87% (see Table 351). In our detailed prospective study, the incidence was 41% on a per-patient
basis and 19% on a per-procedure basis.3(2317)
The prevalence and extent of ulceration can depend on the type24,25(2318) and volume26,27(2319)
of sclerosant injected, method of injection,28(2320) interval between injections,2931(2321) and size

of varices.32(2322) Robertson et al.,33(2323) in a study using rabbit stomachs, found that 5%
ethanolamine oleate was the least ulcerogenic agent when compared with 2% or 3% sodium tetradecyl
sulfate, 2% or 3% polidocanol, and 5% sodium morrhuate. In a clinical study, 2% sodium tetradecyl
sulfate caused significantly more ulceration than 5% ethanolamine oleate when a combined
paravariceal and intravariceal injection technique was used.25(2324) Ulceration is reported to occur
more frequently in patients with Child class C liver disease and after injection of large varices.32(2325)
There is evidence that increasing volumes of sclerosant may be implicated in the occurrence of deep
ulceration (Figures 351 and 352). Madonia et al.34(2326) endoscopically evaluated 40 patients 1
week after sclerotherapy with 1% sodium tetradecyl sulfate. Those who developed deep ulcers had
received significantly greater total volumes of sodium tetradecyl sulfate than those with shallow
ul-ceration (12.8 vs. 9.3 ml). Balloon tamponade was also associated with an increased incidence of
deep ulceration. Singal et al.,35(2327) in a clinical study using absolute alcohol, found that the
incidence and size of the ulcers was directly related to the volume injected. In a canine model of
esophageal varices, Sugawa et al.36(2328) demonstrated that increasing volumes of 5% sodium
morrhuate for intravariceal injection caused greater necrosis and ulceration.
(2329)Figure 351. Endoscopic photograph of a mucosal ulceration involving one quadrant

of the esophagus.
(2330)Figure 352. Endoscopic photograph of a circumferential esophageal ulceration.

In a randomized study comparing techniques, Sarin et al.28(2331) found no significant difference in the
incidence of ulceration between paravariceal and intravariceal injections using 50% ethanol. The risk of
ulceration may be related more to the intensity of the sclerotherapy program than to the technique of
injection.28,29,37(2332) However, it is not always possible, even in controlled studies, to determine the
individual ulcerogenic potential of various sclerosants or the factors primarily responsible for ulceration
(Table 352).
TABLE 352
Randomized Trials Comparing Sclerosants

SCLEROSANT
(%)
SCHEDULE
(WK)
SUCCESS
(%)
TIME TO VARIX
OBLITERATION
(WK)
REBLEEDING
(%)
Kitano et al.24
Polidocanol (1%)
80
5.4
100
4.7
Kitano et al.25
Ethanolamine
(5%)
STS (2%)
75
4.4
24
100
4.9
Sarin et al.215
Ethanolamine
(5%)
Alcohol (50%)
54
20.1
28
Alcohol (100%)
Ethanolamine
(5%)
Alcohol (100%)
3
3
93
40
14.3
22.3
21
30
100
12.9
21
STUDY
Sarin et al.216
TABLE 352
STUDY
Kochhar et
al.217
Chawla and
Dilawari218
Randomized Trials Comparing Sclerosants

SCHEDULE
(WK)
SUCCESS
(%)
TIME TO VARIX
OBLITERATION
(WK)
REBLEEDING
(%)
Alcohol (10%)
82.6
4.8 Ses
17.4
Ethanolamine
(5%)
STS (3%)
STS (3%)
AA
86.4
10.1 Ses
22.7
3
3
3
73.7
62
54
9.2 Ses
5.2
4.7 Ses
15.8
15
19
SCLEROSANT
(%)
Adapted from Sarin SK, Kumar A. Sclerosants for variceal sclerotherapy: A critical appraisal. Am J Gastroenterol 1990; 85:6
STSsodium tetradecyl sulfate; AAabsolute alcohol; Sessessions.
Superficial ulceration is not necessarily harmful, and it is deliberately produced as part of the injection
technique promoted by Kitano et al.23(2333) to enhance eradication of varices. The initial and second
injection (1 week later) are given intravariceally using a transparent overtube. After variceal thrombosis
has occurred, subsequent injections are placed submucosally to create a circumferential ulcer
involving the lower 5 to 10 cm of the esophagus. In their study, healing by epithelialization occurred,
and although pain, pyrexia, and pleural effusions were common, no serious complications were
encountered. Varices were eradicated in all survivors, without variceal recurrences.
In most instances, minor areas of superficial ulceration are asymptomatic and usually heal rapidly
without the need for specific treatment (Figure 353). Singal et al.35(2334) evaluated the symptoms
associated with ulceration after 0.5- to 2-ml intravariceal injections of absolute alcohol. All 40 patients
had mucosal ulceration when examined endoscopically after sclerotherapy, but only two thirds were
symptomatic, with mild dysphagia (53%), mild to moderate retrosternal pain (28%), and low-grade
fever (15%). Patients with large ulcers (>1 cm in diameter) were more likely to be symptomatic. The
dysphagia and chest pain usually improved rapidly and resolved within days, even though ulcer healing
took longer.
(2335)Figure 353. Endoscopic photograph of a healing ulceration.

To prevent sclerotherapy-induced ulcers and their complications, sucralfate, H2-receptor blocking
agents, and antacids, alone or in combination, have been used.3841(2336) Although sucralfate may
reduce recurrent bleeding from ulceration, the frequency and extent of ulcers are similar in patients
who had not received sucralfate.41(2337) Another controlled trial suggested that ulcer healing may be
accelerated by sucralfate, especially in patients with deep ulceration.39(2338) The ulcers healed more
slowly in patients with a serum albumin level of less than 3 g/dl.35(2339) In the study of Kumar et
al.,42(2340) 31 patients were randomized to receive sclerotherapy plus ranitidine (300 mg per day) or
sclerotherapy alone. Varices were treated until eradication, and the treatment groups were said to be
similar with respect to the mean number of treatment sessions, time and volume of sclerosant required
for eradication. Significant reductions in the frequency of sclerotherapy-associated esophageal
ulceration and episodes of recurrent bleeding were observed in the group that took ranitidine.
A small proportion of ulcers persist, despite prolonged treatment with high-dose H2-receptor
antagonists and sucralfate.43(2341) In a small group of patients with complicated chronic ulcers,
Gimson et al.44(2342) achieved complete healing in all patients after an 8-week course of 40 mg of
omeprazole daily. Similar results for a small number of patients were reported by Johlin et al.45(2343)
The rapid healing of resistant ulcers with omeprazole suggests that such ulcers are perpetuated by
mucosal damage from continuing gastroesophageal reflux. The investigators suggest that
consideration should be given to even earlier use of omeprazole for postsclerotherapy ulcers
complicated by symptoms or hemorrhage. Further studies are needed to determine the role of medical
therapy for the prevention and treatment of sclerotherapy-induced ulceration.
Bleeding
Bleeding from the needle puncture site during variceal sclerotherapy is common and can usually be
controlled without difficulty by an adjacent small volume submucosal injection or by tamponade using
the flexed tip of the endoscope. More severe bleeding can result from variceal laceration or accidental
entry of the hilt and needle sheath into the varix in the restless or heaving patient. This can be avoided
by retracting the needle into the sheath between injections.
Early recurrent bleeding is the most common major life-threatening event after sclerotherapy and
occurs in 18 to 55% of patients (see Table 351). Urgent endoscopy is important to establish whether
recurrent bleeding is from a varix, sclerosant-induced ulceration, esophagitis, or another source. If the
recurrent bleeding is variceal in origin, further sclerotherapy is indicated. Although control of acute
variceal bleeding is usually achieved with a single injection session in 70% of patients, some require
additional injections.46(2344) There is evidence to suggest that somatostatin may be helpful in this
situation.47(2345) If variceal bleeding recurs despite two adequate injections, mortality increases
exponentially, and some other definitive procedure should be employed.48,49(2346)
Bleeding from ulceration after sclerotherapy may be particularly troublesome and occurs in as many as
13.3% of patients.5,50,51(2347) It may be difficult to exclude a variceal component aggravating the
hemorrhage because of the complex venous anatomy of the lower esophagus.52(2348) Repeat
sclerotherapy is inappropriate in cases with deep ulceration or esophagitis may compound the
problem. Bleeding in most ulcers is self-limited or stops with the addition of vasopressin and
sucralfate.6,40(2349) The small group that continue to bleed pose a major management problem.
Balloon tamponade increases the risk of pressure necrosis and perforation. Esophageal transection
may be hazardous after previous injection sclerotherapy, and shunt surgery may be inappropriate in
cirrhotic patients with poor liver function.53(2350) In this difficult situation, Jenkins et al.54(2351) were
able to control severe bleeding in 20 of 22 patients using intravenous somatostatin (250 mg/hr).
Perforation
Contained Perforation
Deep ulceration with transmural necrosis may progress to a localized or contained perforation without
mediastinitis or communication with the pleural cavity.6,55(2352) Such confined perforations should be
suspected in patients who have persistent pain and pyrexia after variceal sclerotherapy, and the
diagnosis is confirmed on Gastrografin swallow (Figure 354). Kahn et al.6(2353) recorded this
complication in 25 (8.2%) of 304 patients. These patients were treated with antibiotics, parenteral
hyperalimentation, or enteral feeding through a fine-bore Silastic nasoduodenal tube. In most patients,
subsequent sclerotherapy was delayed for 3 to 4 months. Seven of the 25 patients died; 2 died after a
devascularization operation for continued bleeding from the ulceration, and 5 died of progressive liver
failure. The remaining 18 patients recovered after conservative treatment.
(2354)Figure 354. Barium swallow x-ray demonstrating a contained esophageal perforation.

Free Perforation
Free perforation occurs in 2 to 5% of patients and has a prohibitive mortality rate, especially for
patients with advanced liver disease. Perforation was more frequent when the rigid esophagoscope
was used for sclerotherapy, and it resulted from instrumental injury.3,6,56(2355) Perforation after
fiberoptic injection sclerotherapy is usually delayed and is the result of deep ulceration and transmural
necrosis.57(2356) The risk of perforation is greatest in patients requiring repeated injections for
uncontrolled or recurrent bleeding during the index admission.58,59(2357) During these sessions,
large cumulative volumes of sclerosant are often used, and the risk of inadvertent misplaced, deep
injections is greatest.60,61(2358) Possible aggravating factors that could predispose the patient to
delayed perforation include concurrent balloon tamponade, impairment of healing secondary to poor
liver function, mucosal ischemia associated with infusion of vasopressin, prolonged nasogastric
intubation, and colonization of the ulcer base by Candida albicans.61,62(2359)
Free perforation generally occurs 10 to 14 days after the index injection session (Figures 355 and
356). Analysis of patients for whom detailed clinical information is available reveals a prodrome with
several features in common.57,59,6366(2360) Most developed deep local ulceration at the injection
site after urgent or emergency sclerotherapy during the index admission. Most patients had severe,
prolonged retrosternal and pleuritic chest pain, fever, an exudative pleural effusion, and worsening
encephalopathy. The effusions were initially sterile, but invariably became infected with a variety of
organisms. Gram-negative septicemia, shock, and deteriorating liver function with multiple organ
failure was a common outcome despite adequate surgical or tube drainage. Some patients may not
manifest the clinical features of an esophageal leak but present only with subtle signs of sepsis,
worsening encephalopathy, or deteriorating liver function, with the diagnosis frequently being made
only at necropsy.59,67(2361)
(2362)Figure 355. Contrast x-ray study demonstrating a free esophageal perforation (arrow)
communicating with the pleural space.
(2363)Figure 356. Radiographic images from a patient with an intraperitoneal esophageal

perforation. A, Barium swallow. Note that the study is apparently normal. B, Chest x-ray in the
same patient demonstrating free air under the diaphragm.
Free esophageal perforation poses a major management problem. Esophageal necrosis, mediastinal
venous collaterals, and sepsis with multiple organ failure preclude conventional treatment for
esophageal perforation. At thoracotomy, the tissues are friable and edematous, making repair difficult
and likely to break down. Most reported perforations were managed conservatively with tube
thoracostomy, and the mortality rate was high for patients who suffered this complication. This reflects
the reluctance to institute major operative treatment for high-risk patients who are already considered
to have a poor prognosis.
Intramural Hematoma
An intramural hematoma is a rare complication of sclerotherapy and should be suspected in a patient
who has retrosternal pain and difficulty swallowing that progresses to total dysphagia soon after
sclerotherapy.6875(2364) Presentation may, however, be delayed for 2 to 4 days after sclerotherapy.
The diagnosis is readily made using contrast studies or endoscopy. A large intraluminal filling defect is
revealed on barium swallow and is best seen in the lateral view.69(2365) Endoscopy shows an
intraluminal bulge of the mucosa with a characteristic dark blue discoloration. Occasionally, bleeding
due to mucosal slough may occur. Treatment is essentially conservative. Dysphagia resolves within
days with complete resolution of the hematoma in 1 to 3 weeks.
Stricture
The incidence of esophageal stricture after sclerotherapy ranges from 11 to 58% (Table 353; see also
Table 351). In 204 patients undergoing long-term sclerotherapy, we found that 1 of 10 patients
developed a stricture.76(2366) It is difficult to determine from published series precisely which factors
contribute to stricture formation (Table 353). Most reports have not found a direct relationship with the
number of sclerotherapy sessions, volume or type of sclerosant, and site of injection.7680(2367)
Sorensen et al.,37(2368) however, demonstrated a clear relationship between frequency and
cumulative volume of injection and an association with preexisting ulceration. The 59% incidence of
stenosis in 20 of 34 patients is among the highest reported. Their technique, however, differed from
other studies. Paravariceal injections extending over the lower 10 cm of the esophagus were
performed every 3 days. Patients who developed strictures had received more injections and larger
volumes of sclerosant, and a significantly greater number had preceding mucosal necrosis.
TABLE 353
Esophageal Strictures After Sclerotherapy

NUMBER
OF
PATIENTS
PERCENTAGE
WITH
STRICTURE
Waring and
Sanowski77
Haynes et al.78
103
Sorensen et
al.37
Snady and
Korsten79
Farrell et al.76
STUDY
Kochhar et
al.80
SCLEROSANT
FOR EVS
TYPE
OF
EVS
VOLUME OF
SCLEROSANT
(ML)
30
1.5% STS
NA
89
61
16
1.5% STS
IV
61
34
59
P + EO
PV
123
31
55
1% STS
IV + PV
NA
205
11
5% EO
IV + PV
58
129
15
NUMBER OF
PROCEDURE
93
16
AA
IV
32
20
15
5% EO
IV
84
16
12
STS
IV
41
EVSendoscopic variceal sclerotherapy; STSsodium tetradecyl sulfate; NAdata not available; IVintravariceal; Ppo
EOethanolamine oleate; PVparavariceal; AAabsolute alcohol.
* Mean.
Interval between initiating sclerotherapy and onset of stricture.
In the trial of Snady et al.,81(2369) patients undergoing sclerotherapy were randomized to receive
sclerotherapy plus acid protection (i.e., antacids, cimetidine, and sucralfate; 31 patients) or
sclerotherapy alone (31 patients). In the former treatment group, 9.7% of patients developed
symptomatic structures during the course of treatment, compared with 38.7% of those who underwent
sclerotherapy alone.
Sclerotherapy-induced strictures are usually short and localized to the lower 5 cm of the esophagus
(Figure 357). Most strictures can be dilated without difficulty. Two to three dilation sessions suffice for
more than 85% of patients.80(2370) Persistent esophageal dysmotility may explain the refractory
dysphagia that occurs in some patients despite adequate dilation. Dilation does not precipitate bleeding
from partially treated varices, and although the stricture may temporarily delay eradication of varices,
the sclerotherapy program can be continued after stricture dilation. For short, symmetric strictures,
Maloney mercury-filled rubber dilators allow easy and safe dilation, although tighter and longer
strictures require fluoroscopically controlled dilation over an endoscopically placed guidewire with
Savary or Eder-Puestow dilators (Figure 358).
(2371)Figure 357. Endoscopic photograph of a high-grade esophageal stricture after

sclerotherapy.
(2372)Figure 358. Endoscopic photograph of guidewire placement through the stricture

before dilation.
Motility Disorders
Several studies have evaluated the short-term and long-term effects of sclerotherapy on esophageal
motor function and gastroesophageal reflux.8284(2373) Serial evaluation of motility patterns in the
esophagus before sclerotherapy, 3 days after sclerotherapy, and 6 months later has demonstrated that
the length of the high-pressure zone, peristaltic velocity, and swallow-wave symmetry are markedly
affected. The length of the high-pressure zone increased significantly after the initial sclerotherapy
session because of intense inflammation in the distal esophagus. The normal waveform pattern and
symmetry are altered considerably by sclerotherapy. Double- and triple-peak waveforms, dropped
swallow waves in the distal esophagus, and simultaneous and spontaneous contractions have been
observed.8386(2374) Esophageal scintigraphy of esophageal function after eradication of varices has
shown increased transit times compared with controls.87(2375) These changes increase after
sequential treatment, and this effect probably is a manifestation of sclerosant-induced esophagitis,
intramural inflammatory response, or fibrotic changes in the esophageal wall.88(2376)
Injection sclerotherapy has had no substantial effect on lower esophageal sphincter pressure in
most83,84,86(2377) but not all studies.89(2378) In the study of Suzuki et al.89(2379) of lower
esophageal sphincter pressure measured in 41 patients before and after sclerotherapy, including
measurements 6 months after treatment, the magnitude of the decrease in pressure was greater in
patients with sclerotherapy-associated ulcers. The results showed correlations among the occurrence
of reflux symptoms, ulceration, and decreases in sphincter pressure. However, these abnormalities
were transient and had usually resolved by the 6-month follow-up examination.
There is some disagreement concerning the incidence and severity of gastroesophageal reflux after
sclerotherapy and its effect on esophageal acid clearance.9,8385,9092(2380) Reilly et al.83(2381)
found that gastroesophageal reflux, as determined by standard reflux tests, becomes more prevalent
after sclerotherapy, and they suggested that gastroesophageal reflux contributes to stricture formation.
In contrast, Ogle et al.85(2382) found no instance of acid reflux into the esophagus, although patients
who received sclerotherapy did have impaired acid clearance. The magnitude of these changes are not
thought to be severe enough to promote pathologic gastroesophageal reflux. Sauerbruch et
al.93(2383) compared the results of long-term pH monitoring of 19 patients who underwent
sclerotherapy and 15 untreated patients with cirrhosis who served as controls. There were no
significant differences between the two groups with respect to percentage of time that esophageal pH
was less than 4 or the mean duration of reflux episodes. In a similar study of 24-hr pH monitoring of 16
patients who underwent variceal sclerotherapy and 21 untreated patients with varices, Kinoshita et
al.94(2384) found a significant increase in the rate of gastroesophageal reflux in treated patients. The
severity of reflux directly correlated with the volume of paravariceally injected sclerosant.
Other Esophageal Complications

A variety of unusual local esophageal complications have been reported after sclerotherapy. Davion et
al.95(2385) described the development of gastric bezoars in five patients undergoing sclerotherapy.
This complication was attributed to possible transient vagal nerve damage, although the possibility of
vagal neuropathy due to other causes was not eliminated. Pneumatosis intestinalis and
pneumoperitoneum may occur because of intramural air entering through a small mucosal tear in the
esophageal wall and dissecting distally into the stomach, small bowel, and colon. Rupture into the
peritoneum produces free intraperitoneal air. The condition is benign and resolves
spontaneously.96(2386) Other rare findings include pseudodiverticula,97(2387) mucosal
bridges,98(2388) and periesophageal granulomas.99(2389) These findings are usually incidental and
require no specific therapy.
Sporadic reports have described esophageal carcinoma, usually the squamous cell variety, developing
in patients undergoing sclerotherapy.100103(2390) These cases are few, and additional risk factors
(e.g., smoking, alcohol intake) can be identified in virtually all instances. Although surveillance of
patients undergoing sclerotherapy has been recommended,104(2391) there are no scientific data to
support claims of a relationship between sclerotherapy and the development of carcinoma of the
esophagus.
Cardiorespiratory Effects
Cardiac complications specifically related to sclerotherapy are rare. Anecdotal accounts of coronary
artery spasm,105(2392) persistent bradyarrhythmia,106(2393) and heart failure due to
polidocanol107,108(2394) have been reported. Seven cases of pericarditis after sclerotherapy have
been described.109,110(2395) Onset is heralded by fever, chest pain, and dyspnea; a pericardial
friction rub is usually heard, and electrocardiographic and echocardiographic evidence point to a
pericardial effusion. If pericarditis remains undiagnosed, progression to cardiac tamponade or
constrictive pericarditis may occur.111,112(2396) No precipitating factors have been clearly identified.
Pulmonary complications are common and range from minor asymptomatic changes found incidentally
on routine chest x-ray films to aspiration or bronchopneumonia, pleural effusions, lobar collapse or
consolidation, and the adult respiratory distress syndrome.99,113(2397) It is often difficult to determine
to what extent respiratory changes are directly attributable to sclerotherapy, because underlying
parenchymal and vascular abnormalities are commonly found in patients with chronic liver
disease.113,114(2398) Sepsis, pulmonary congestion due to fluid shifts after vigorous resuscitation
with crystalloids, massive transfusion, and diaphragmatic splinting by tense ascites are additional
factors that may contribute to a deterioration in pulmonary function.114(2399) In two cases described
by Crawford and Ryan,115(2400) acute respiratory insufficiency occurred after sclerotherapy. Both
patients had ascites, and the respiratory difficulty was reversed by aspirating air from the stomach.
Several studies have investigated the distribution and potentially damaging effects of sclerosant
solutions on the respiratory system.116118(2401) Sclerosant dissemination to the pulmonary and
systemic circulation after intravariceal sclerotherapy occurs through esophagogastric collaterals and
the azygos-hemiazygos systems. Entry of sclerosant into the pulmonary circulation has been
demonstrated by positive uptake on a lung scan of 99mTc-labeled sodium tetradecyl sulfate and
sodium morrhuate solutions when injected into esophageal varices. Systemic dissemination has also
been demonstrated to occur with ethanolamine plus 99mTc sodium pertechnetate, but the frequency
and consequences appear to be minor.116118(2402)
Studies in experimental models, using large-volume infusions of sclerosant solutions containing fatty
acid derivatives (i.e., sodium morrhuate and ethanolamine oleate) injected directly into the right atrium,
have produced pulmonary endothelial damage, transient increases in pulmonary artery pressure, and
hemorrhagic pulmonary edema,119123(2403) with consequent significant worsening of gas
exchange.120(2404) The fact that oleic acid causes an acute pulmonary injury in animals has raised
concern that variceal sclerosants containing fatty acids may cause similar injury in humans.
The pulmonary hemodynamic effects of sclerosant injection have been evaluated in several small
clinical studies. Transient increases in pulmonary artery pressure from normal basal values occur
during sclerotherapy but are not associated with a change in cardiac output or arterial
oxygenation.124(2405) Intravariceal sclerotherapy using absolute alcohol results in significant mean
pulmonary artery pressure increases at 1 and 5 min, with return to basal levels by 15 min. Equivalent
volumes of saline produce similar significant rises in pulmonary artery pressure from basal levels,
suggesting that the effect results from a volume load rather than the sclerosant.125(2406) Although
pulmonary and systemic vascular resistance indices increase significantly from basal values after
sclerotherapy with sodium morrhuate, pulmonary artery and pulmonary capillary wedge pressures
remain stable without alterations in gas exchange.119(2407) The changes in pulmonary
hemodynamics after sclerotherapy in humans are small in magnitude and not sufficient to cause acute
pulmonary capillary injury.
Because premedication and passage of an endoscope may contribute to aspiration pneumonitis or
hypoxemia, the incidence of respiratory dysfunction in patients receiving sclerotherapy should be
compared with those undergoing endoscopy for other reasons. In a controlled study, Korula et
al.126(2408) found no difference in the short-term or long-term effects on lung function and gas
exchange after sclerotherapy in patients with cirrhotic portal hypertension compared to a similar group
undergoing diagnostic endoscopy only. In contrast, Kitano et al.127(2409) found that patients
complaining of post-injection retrosternal pain 24 hr after sclerotherapy had a larger fall in vital capacity
and forced expiratory volume than patients without pain. One third of cirrhotic patients with esophageal
varices were shown to have preexisting pulmonary interstitial edema and arterial hypoxemia (PaO2 <
80 mm Hg). In these patients, injection of 5% ethanolamine oleate may lead to further deterioration of
pulmonary function and a decrease in arterial oxy-gen content.127(2410) Samuels et al.128(2411)
found significant decreases in arterial oxygen tension and vital capacity in cirrhotic patients undergoing
sclerotherapy compared with controls. The change in arterial oxygen tension correlated with the
percentage change in vital capacity in the group that underwent sclerotherapy. Samuels et
al.128(2412) thought these results suggested that sclerotherapy produced a restrictive defect in
pulmonary function, possibly caused by embolization of sclerosant to the lungs.
Pulmonary and mediastinal abnormalities are frequently found on routine chest x-ray and computed
tomography examinations when performed within 48 hr after sclerotherapy. These changes may be
explained by periesophageal inflammation and by the fact that the esophagus does not have the
serosal layer that provides a barrier function for other organs. Saks et al.129(2413) found radiologic
changes in as many as 79% of patients. Pleural effusions and mediastinal soft tissue densities are the
most common findings, and atelectasis, linear lung shadows, and retrocardiac soft tissue densities are
less often demonstrated.129,130(2414)
Chest pain and effusions occur more frequently in patients who develop deep ulceration and are
caused by an intense periesophageal, mediastinal, and pleural inflammatory reaction.12,16,131 Most
effusions are small and resolve spontaneously. Pleural effusions were found by Bacon et al.131(2415)
after 31 (48%) of 65 sclerotherapy sessions performed in 30 patients. There was an approximately

equal distribution between left-sided, right-sided, and bilateral effusions. The total volume of sclerosant
injected was significantly greater in patients who developed effusions than in those who did not. Most
of the effusions were exudative. Parikh et al.132(2416) found pleural effusions in 6 of 31 patients who
underwent sclerotherapy with absolute alcohol. As in the study of Bacon et al.,131(2417) chest pain,
frequently persistent, was found to be significantly more common in patients with pleural effusions.
Most effusions were small, developed within 48 hr of a treatment session, and disappeared without
treatment within 1 week.
Aspiration is the most serious of the respiratory complications and occurs most frequently during
sclerotherapy for acute bleeding. Aspiration pneumonia is avoidable if the stomach is completely
emptied by suction before sclerotherapy and an assistant scrupulously clears the mouth and
hypopharynx with a suction catheter during the injection procedure. Excessive sedation, hepatic
encephalopathy, and a prolonged procedure without adequate or effective airway protection during
active bleeding are contributing factors if bleeding is massive. In this situation, endotracheal intubation
before endoscopy is essential to avoid this potentially lethal complication.
Other uncommon pulmonary complications reported after sclerotherapy are bronchoesophageal
fistula,133,134(2418) pneumothorax,135(2419) subcutaneous emphysema,99(2420)
chylothorax,136,137(2421) and hemothorax.138(2422)
Systemic Complications
Septicemia and Bacteremia
Transient fever due to an acute local inflammatory response or chemical phlebitis occurs in a fourth of
patients after sclerotherapy. If a fever persists for more than 2 days, a search for a septic or local
esophageal complication is mandatory. Anecdotal reports have incriminated sclerotherapy as a cause
of meningococcal and streptococcal pneumoniae septicemia;139(2423) infective
endocarditis;140(2424) pyogenic meningitis;141(2425) subdural empyema;142(2426) abscesses
involving the brain,143146(2427) spleen,147(2428) and perirenal areas;148(2429) empyema with and
without apparent esophageal perforation;149,150(2430) and bacterial peritonitis.151154(2431) These
reports have raised a question about whether the incidence of septic complications is increased as a
consequence of sclerotherapy-induced bacteremia.
Several sources of bacterial contamination are possible during injection sclerotherapy. The spectrum
of organisms associated with bacteremia and the predominance of -hemolytic streptococci strongly
suggest the oropharyngeal flora as the source of contamination. During sclerotherapy, these
organisms may be introduced by the endoscope or injector needle and enter the bloodstream. The
length of the needle injector and a contaminated water supply have been implicated in
sclerotherapy-associated bacteremia.155,156(2432)
The incidence of bacteremia after sclerotherapy ranges from 0 to 50%.155164(2433) A variety of
injection techniques, sclerosant solutions, and different lengths of injection needles were used in these
studies. An increased incidence of bacteremia occurs during and up to 5 min after sclerotherapy.
Because blood cultures were drawn during both these periods in fewer than half of the studies, the
extent of bacteremia in some studies may have been underestimated. Inherent to all studies designed
to search for positive blood cultures is the difficulty of determining true bacteremia from contaminants.
Some investigators have isolated common skin commensals, and in one study, 23% of isolates were
coagulase-negative staphylococci,160(2434) which may originate from the skin during venipuncture.
Most previous data on bacteremia after sclerotherapy have been derived from blood cultures obtained
during elective sclerotherapy. The risk of bacteremia may be higher during the technically more
demanding and traumatic emergency sclerotherapy and in the presence of venous and urine catheters
and endotracheal tubes. In the report of Bac et al.,154(2435) for example, the calculated risk for the
development of bacterial peritonitis was significantly higher after sclerotherapy performed on an
emergency basis than for elective procedures. Patients with alcoholic cirrhosis may develop
bacteremia spontaneously because of decreased reticuloendothelial system function, impaired
neutrophil chemotaxis, low levels of serum complement, and impaired cell-mediated
immunity.164(2436)
The incidence of infection after sclerotherapy was studied by Rolando et al.165(2437) in a trial that
compared sclerotherapy plus antibiotic prophylaxis (i.e., intravenous imipenem or cilastatin) with
sclerotherapy alone in patients with variceal bleeding. No significant difference was found between the
incidence of bacteremia after treatment sessions in the control group and that in the patients who
received antimicrobial prophylaxis (5.6% vs. 1.1%). Most of the episodes of bacteremia were
associated with emergency treatment sessions. In a similar study by Pulanic et al.,166(2438) 30
patients with bleeding esophageal varices underwent sclerotherapy without antibiotic prophylaxis, and
30 similar patients had sclerotherapy plus intravenous infusions of ampicillin with treatment and for 3
days thereafter. No significant differences between the two groups were found in the clinical
parameters of infection (e.g., temperature, white blood cell count, differential blood cell count,
sedimentation rate) during a follow-up period of 3 days.
The clinical importance of blood culture isolates after sclerotherapy remains questionable. In none of
the prospective studies have organisms (other than probable commensals) been isolated more than 30
min after sclerotherapy, suggesting that bacteremia is always transient. No infective complications
have been reported after bacteremia in these studies. Previous recommendations advising routine
antibiotic prophylaxis are no longer valid, and most authorities recommend prophylaxis only for patients
with specific vascular risk factors such as prosthetic valves or previous episodes of
endocarditis.163(2439) Strict attention to cleaning and disinfection of equipment to avoid contamination
of endoscopes and the water supply are essential (see Chapter 8: Disinfection of Endoscopes and
Accessories).
Hemodynamic and Thrombotic Effects

Potential effects of repeated long-term sclerotherapy and obliteration of esophageal varices are an
increase in portal pressure, development of other compensatory collaterals, and bleeding from varices
at remote sites.167169 Korula and Ralls168(2440) demonstrated that, despite improvements in
laboratory and clinical parameters of hepatic function, the portal venous pressure gradient increased
by a third in cirrhotic patients after eradication of esophageal varices. Dilawari et al.170(2441) found
that 6 (40%) of 15 patients with nonalcoholic portal hypertension developed spontaneous
splenoadrenorenal shunts after sclerotherapy. The same mechanism may explain the increased
incidence of portal hypertensive gastropathy after repeated sclerotherapy171(2442) and the
phenomenon of bleeding from varices at other sites, including the umbilical vein (i.e.,
Cruveilhier-Baumgarten syndrome), duodenum, ileum, colon, rectum, and bowel-related
adhesions.167,169,172176(2443)
Changes and direction of flow in the coronary and azygos systems are complex in patients with portal
hypertension. Phasic retrograde esophageal collateral flow has been demonstrated during
sclerotherapy using fluoroscopy and endoscopic Doppler-flow techniques.60,177(2444) Aoki et
al.178(2445) demonstrated by intraoperative portography that flow could be hepatofugal, to and fro, or
hepatopetal. Altered venous flow, endothelial damage, and a hypercoagulable state after repeated
intravariceal sclerotherapy may promote excessive local venous thrombosis, with propagation into the
splanchnic venous system and thrombosis of the portal and splenic veins. Some researchers claim
that a local endothelial inflammatory response after sclerotherapy is the initiating event, although
others have shown that hypercoagulable states may be induced by the sclerosant.179181(2446)
In an umbilical cord model designed to simulate variceal blood flow, brief exposure to even low
concentrations of sodium tetradecyl sulfate produces damage and stripping of endothelium, which
exposes highly thrombogenic Factor VIII-rich subendothelium.181(2447) The effects of sodium
tetradecyl sulfate on coagulation and platelet function depend on sclerosant concentration. A dilute
concentration induces a hypercoagulable state by selective inhibition of protein C and promotion of
platelet aggregation. Activation of systemic blood coagulation in cirrhotic patients after sclerotherapy,
which may be aggravated by vasopressin infusion, may promote venous thrombosis in the splanchnic
bed. In experimental studies, higher concentrations of sodium tetradecyl sulfate inactivate the
coagulation cascade and cause lysis of platelets.181(2448)
Because sclerotherapy may lead to thrombosis of gastric varices,15(2449) it is conceivable that
thrombus formation may extend further and initiate thrombosis in the splanchnic venous system. Portal
vein thrombosis is a well-recognized complication of cirrhosis and portal hypertension. The reported
incidence ranges from 0.5 to 21%.182184(2450) Acute portal or mesenteric venous thrombosis in
association with sclerotherapy is uncommon. Nevertheless, a number of cases of portal or mesenteric
venous infarction have been reported after sclerotherapy or intravenous administration of
vasopressin.185191(2451)
In a small retrospective series of patients undergoing portosystemic shunt surgery, Leach et
al.189(2452) found splanchnic venous thrombosis to be more common in patients who had prior
sclerotherapy than in those who had no prior injection therapy. Korula et al.192(2453) described a case
in which amorphous, eosinophilic material thought to be similar to sclerosant injected at sclerotherapy
was found within a clot encountered at portocaval shunt surgery. Stoltenberg et al.,193(2454) in an
autopsy series, demonstrated extension of thrombus from esophageal varices into the portal and
mesenteric venous systems that resulted in small intestinal infarction and hepatic failure. In two cases,
splenic vein thrombosis and splenic infarction occurred, suggesting propagation of clot through the
coronary and left gastric veins. Rice et al.194(2455) demonstrated clots in the portal venous system by
means of Doppler sonography in patients who had undergone sclerotherapy or a portosystemic shunt
operation, although this finding was relatively infrequent in the endoscopically treated patients
compared with the surgically treated patients (i.e., end-to-side shunt 69%, distal splenorenal shunt
67%, sclerotherapy 5%).
Caletti et al.195(2456) performed ultrasonography of the portal venous system in 25 patients before
initiation of sclerotherapy and after obliteration of varices. They demonstrated no evidence of
thrombosis or alteration in the caliber of the portal venous system. In contrast, Hunter et al.196(2457)
observed portal vein thrombosis in 36% of a small series of patients who had undergone sclerotherapy;
controls had only a 10% incidence of such findings.
An increased incidence of thrombosis of the portal vein or its major tributaries after long-term
sclerotherapy has been disputed. In the controlled trial of Kawasaki et al.197(2458) comparing
sclerotherapy with distal splenorenal shunt, all patients underwent angiographic assessment of the
portal, splenic, and superior mesenteric veins before and after treatment. Those who received chronic
sclerotherapy provided a unique group in which the incidence of thrombosis could be assessed.
Despite frequent injections (mean 6.5) and large volumes (mean 62 ml), no patient developed splenic
or portal vein thrombosis.
Distant histologic effects due to sclerosant, including intimal damage and fibrosis in the portal vein,
have been reported after obliteration of esophageal varices. Hunter et al.196(2459) found substantial
changes when comparing the morphology of portal and splenic veins in patients who had received
sclerotherapy with those who had not. In addition to the loss of smooth muscle and elastin fibers and
medial fibrosis in patients with portal hypertension, those who had received sclerotherapy also had
disruption of normal venous architecture with loss of elastic fibers, smooth muscle bundles, and an
increase in fibrous tissue. Chaudhary et al.198(2460) demonstrated changes in splenic vein histology
in patients undergoing splenorenal shunt after sclerotherapy, including increased fibrosis, intimal and
medial destruction, and microthrombi. Retrograde flow through collateral pathways or abnormal
responses of the perivenous lymphatic vessels to the sclerosant may, alone or in combination, account
for the observed changes.196(2461)
Seidman et al.199(2462) reported the case of a child who developed paraplegia after intravariceal
injections of ethanolamine oleate. At autopsy 2 years later, they found evidence of an infarct of the
spinal cord secondary to occlusion of the anterior spinal artery.
Several studies have demonstrated transient changes in the hemostatic mechanisms of patients
undergoing sclerotherapy.200204(2463) In a study of 45 patients, Yamaga et al.201(2464) found
significant increases in concentrations of various fibrinopeptides and fibrin degradation products and
observed suppression of platelet aggregation after intravariceal sclerotherapy using 5% ethanolamine
oleate. Platelet function gradually returned to normal in approximately 1 week. Ohta et al.202(2465)
found evidence of hemolysis, including hemoglobinuria, in patients undergoing sclerotherapy with 5%
ethanolamine oleate; this finding was significantly more frequent in patients with albumin levels of less
than 3.0 g/dl but was otherwise unrelated to liver function status. Hemolysis was found to increase with
increasing concentrations of ethanolamine oleate and could be inhibited by increases in serum
albumin. Creatinine clearance decreased in patients with hemoglobinuria, and two such patients were
said to have developed acute renal failure. Miyoshi et al.205(2466) found evidence of renal tubular
dysfunction but no significant changes in glomerular filtration rate in patients undergoing sclerotherapy
with ethanolamine oleate. Changes in renal function were suppressed by pretreatment with
haptoglobin.
De Franchis et al.204(2467) found that standard coagulation tests (i.e., prothrombin time, partial
thromboplastin time, platelets, and fibrinogen) were not altered in patients undergoing sclerotherapy
with 5% sodium morrhuate plus thrombin or sodium morrhuate alone. However, an abrupt increase in
plasma fibrinopeptide A was demonstrated after sclerotherapy; in most cases, these levels returned to
baseline within 24 hr. De Franchis et al.204(2468) thought that sclerotherapy-related changes in
coagulation were of no clinical significance. However, Yuki et al.200(2469) felt that mild transient
symptoms such as headache and fatigue could be correlated with changes in coagulation parameters
in patients who underwent sclerotherapy with a combination of hypertonic glucose, thrombin, and 1%
polidocanol.
Conclusions
Although complications after sclerotherapy are common, most are minor and do not interrupt the
injection program. In a small group, however, the success of sclerotherapy is compromised by
recurrent bleeding and serious procedure-related complications.5(2470) Most of the serious
complications related to sclerotherapy occur in patients with severe liver disease in whom control of
bleeding is difficult.
Several precautions are critical to avoid local and systemic complications. Effective resuscitation
should precede endoscopy in patients with evidence of recent major bleeding. Diagnostic and
therapeutic endoscopy should be performed in a well-equipped unit, with competent assistants and
careful monitoring. Meticulous attention should be given to suctioning of the mouth and hypopharynx
by a dedicated assistant to avoid aspiration. Early endotracheal intubation is crucial if massive bleeding
occurs.
Precise and accurately placed injections are essential. To ensure adequate visibility during active
bleeding, a large or double-channel endoscope with vigorous irrigation should be used, and the head of
the patient's bed should be elevated. Uncontrolled, blind, large-volume injections during active bleeding
must be avoided. The sclerotherapy needle's length should not exceed 5 mm, and a short bevel
reduces the risk of deep injections.
Recurrent bleeding after sclerotherapy requires careful evaluation and repeat endoscopy to determine
the source. If variceal bleeding continues or recurs during the index admission despite two adequate
injection sessions, other definitive therapy should be instituted. If ulceration involves more than one
esophageal quadrant, further injections should be delayed until healing has occurred. Treatment with
H2 blockers or sucralfate does not prevent ulceration but may accelerate healing. Omeprazole has
been effective in the treatment of chronic ulcers. Special care should be taken in treating patients with
deep ulceration and persistent pain, fever, an increasing pleural effusion, and deterioration of liver
function, all of which suggest transmural necrosis and impending perforation. Motility abnormalities are
usually transient and of minor clinical consequence. Most symptomatic strictures respond effectively to
dilation.
Mature clinical judgment is necessary in acute problematic or complex cases, and careful supervision
of trainees or assistants by an experienced endoscopist becomes essential when critical decisions are
required. Early and close multidisciplinary consultation is often useful in demanding cases to facilitate
appropriate therapy and optimal management.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Terblanche J, Kahn D, Campbell JAH, et al. Failure of repeated injection sclerotherapy to

improve long-term survival after oesophageal variceal bleeding. Lancet 1983;2:132832.
Terblanche J. The long-term management of patients after an oesophageal variceal bleed:
The role of sclerotherapy. Br J Surg 1985;72:8890.
Terblanche J, Kahn D, Bornman PC. Long-term injection sclerotherapy treatment for
esophageal varices. A 10-year prospective evaluation. Ann Surg 1989;210:72531.
Sanowski RA, Waring JP. Endoscopic techniques and complications in variceal sclerotherapy.
J Clin Gastroenterol 1987;9:50413.
Schuman BM, Beckman JW, Tedesco FJ, et al. Complications of endoscopic injection
sclerotherapy: A review. Am J Gastroenterol 1987;82:82330.
Kahn D, Jones B, Bornman PC, Terblanche J. Incidence and management of complications
after injection sclerotherapy: A ten-year prospective evaluation. Surgery 1989;105:1605.
Baillie J, Yudelman P. Complications of endoscopic sclerotherapy of esophageal varices.
Endoscopy 1992;24:28491.
Muhldorfer SM, Kekos G, Hahn EG, Ell C. Complications of therapeutic gastrointestinal
Schuman BM. The systemic complications of sclerotherapy of esophageal varices.
Lieberman DA. In the eye of the needle. A reappraisal of endoscopic sclerotherapy. J Clin
Helpap B, Bollweg L. Morphological changes in the terminal oesophagus with varices
following sclerosis of the wall. Endoscopy 1981;13:22933.
Evans DMD, Jones DB, Cleary BK, Smith PM. Oesophageal varices treated by sclerotherapy:
A histopathological study. Gut 1982;23:61520.
Papadimos D, Kerlin P, Harris OD. Endoscopic sclerotherapy: Lessons from a necropsy
study. Gastrointest Endosc 1986;32:26973.
Pushpanathan C, Idikio H. Pathological findings in the esophagus after endoscopic
sclerotherapy for variceal bleeding. Am J Gastroenterol 1986;81:913.
Kage M, Korula J, Harada A, et al. Effects of sodium tetradecyl sulfate endoscopic
sclerotherapy on the esophagus. A prospective clinical and histopathological study. J Clin
Ayres SJ, Goff JS, Warren GH. Endoscopic sclerotherapy for bleeding esophageal varices:
Effects and complications. Ann Intern Med 1983;98:9003.

17. Soehendra N, de Heer K, Kempeneers I, Frommelt L. Morphological alterations of the
esophagus after endoscopic sclerotherapy of varices. Endoscopy 1983;15:2916.
18. Sarles HE, Sanowski RA, Talbert G. Course and complications of endoscopic variceal
sclerotherapy: A prospective study of 50 patients. Am J Gastroenterol 1985;80:5959.
19. Larson AW, Cohen H, Zweiban B, et al. Acute esophageal variceal sclerotherapy. JAMA
1986;255:497500.
20. Terabayashi H, Ohnishi K, Tsunoda T, et al. Prospective controlled trial of elective endoscopic
sclerotherapy in comparison with percutaneous transhepatic obliteration of esophageal
varices in patients with nonalcoholic cirrhosis. Gastroenterology 1987;93:12059.
21. Low DE, Kozarek RA, Ball TJ, Beebe HG. Endoscopic variceal sclerotherapy as primary
treatment for bleeding esophageal varices. J Clin Gastroenterol 1989;11:2539.
23. Kitano S, Koyanagi N, Iso Y, et al. Prevention of recurrence of esophageal varices after
24. Kitano S, Iso Y, Koyanagi N, et al. Ethanolamine oleate is superior to polidocanol
(Aethoxysclerol) for endoscopic injection sclerotherapy of esophageal varices: A prospective
randomized trial. Hepatogastroenterology 1987;34:1923.
25. Kitano S, Iso Y, Yamaga H, et al. Trial of sclerosing agents in patients with oesophageal
26. Neeman A, Leiser A, Kadish U. Treatment of esophageal varices: Low versus high dose of
5% ethanolamine oleate. Am J Gastroenterol 1991;86:11824..
27. Iso Y, Kitano S, Iwanaga T, et al. A prospective randomized study comparing the effects of
large and small volumes of the sclerosant 5% ethanolamine oleate injected into esophageal
varices. Endoscopy 1988;20:2858.
28. Sarin SK, Nanda R, Sachdev G, et al. Intravariceal versus paravariceal sclerotherapy: A
prospective, controlled, randomised trial. Gut 1987;28:65762.
30. Sarin SK, Sachdev G, Nanda R, et al. Comparison of the two time schedules for endoscopic
sclerotherapy: A prospective randomised controlled study. Gut 1986;27:7103.
31. Higashi H, Kitano S, Hashizume M, et al. A prospective randomized trial of schedules for
sclerosing esophageal varices: 1 versus 2 week intervals. Hepatogastroenterology
1989;36:33740.
32. Choudhuri G, Agrawal BK, Tantry BV, et al. Post-sclerotherapy esophageal ulcers: A
prospective analysis of their behaviour. Indian J Gastroenterol 1989;8:1921.
33. Robertson CS, Womack C, Robson K, Morris DL. A study of the local toxicity of agents used
for variceal injection sclerotherapy. HPB Surg 1989;1:14954.
34. Madonia S, Traina M, Montalbano L, D'Amico G. Variceal ulceration following sclerotherapy:
Normal consequence or complication?. Gastrointest Endosc 1990;36:767.
35. Singal A, Sarin SK, Sood GK, Broor SL. Ulcers after intravariceal sclerotherapy: Correlation of
symptoms and factors affecting healing. J Clin Gastroenterol 1990;12:2504.
36. Sugawa C, Okumura Y, Lucas CE, Walt AJ. Endoscopic sclerosis of experimental
esophageal varices in dogs. Gastrointest Endosc 1978;24:1146.
37. Sorensen T, Burcharth F, Pedersen ML, Findahl F. Oesophageal stricture and dysphagia after
endoscopic sclerotherapy for bleeding varices. Gut 1984;25:4737.
38. Tamura S, Shiozaki H, Kobayashi K, et al. Prospective randomized study on the effect of
ranitidine against injection ulcer after endoscopic injection sclerotherapy for esophageal
varices. Am J Gastroenterol 1991;86:47780.
39. Paquet KJ, Koussouris P, Keinath R, et al. A comparison of sucralfate with placebo in the
treatment of esophageal ulcers following therapeutic endoscopic sclerotherapy of esophageal
varicesA prospective controlled randomized trial. Am J Med 1991;91:147S-50S.
40. Roark G. Treatment of postsclerotherapy esophageal ulcers with sucralfate. Gastrointest

Endosc 1984;30:910.
41. Polson RJ, Westaby D, Gimson AES, et al. Sucralfate for the prevention of early rebleeding
following injection sclerotherapy for esophageal varices. Hepatology 1989;10:27983.
42. Kumar A, Mehta SR, Joshi V, et al. Ranitidine for the prevention of complications following
endoscopic sclerotherapy for esophageal varices. J Assoc Physicians India 1993;41:584, 589.
43. Subramanyam K, Patterson M. Chronic esophageal ulceration after endoscopic
sclerotherapy. J Clin Gastroenterol 1986;8:5860.
44. Gimson A, Polson R, Westaby D, Williams R. Omeprazole in the management of intractable
esophageal ulceration following injection sclerotherapy. Gastroenterology 1990;99:182931.
45. Johlin FC, Labrecque DR, Neil GA. Omeprazole heals mucosal ulcers associated with
endoscopic injection sclerotherapy. Dig Dis Sci 1992;37:13736.
46. Kahn D, Bornman PC, Terblanche J. A 10-year prospective evaluation of balloon tube
tamponade and emergency injection sclerotherapy for actively bleeding oesophageal varices.
HPB Surg 1989;1:20715 [discussion 2159].
47. Jenkins SA, Shields R, Jaser N, et al. The management of persistent or recurrent variceal
bleeding after injection sclerotherapy by somatostatin. HPB Surg 1992;5:2217.
48. Bornman PC, Terblanche J, Kahn D, et al. Limitations of multiple injection sclerotherapy
sessions for acute variceal bleeding. S Afr Med J 1986;70:346.
49. Burroughs AK, Hamilton G, Phillips A, et al. A comparison of sclerotherapy with staple
transection of the esophagus for the emergency control of bleeding from esophageal varices.
N Engl J Med 1989;321:85762.
50. Sauerbruch T, Weinzierl M, Kopcke W, Paumgartner G. Long-term sclerotherapy of bleeding
esophageal varices in patients with liver cirrhosis. Scand J Gastroenterol 1985;20:518.
51. Palani CK, Abuabara S, Kraft AR, Jonasson O. Endoscopic sclerotherapy in acute variceal
hemorrhage. Am J Surg 1981;141:1648.
52. Kitano S, Terblanche J, Kahn D, Bornman PC. Venous anatomy of the lower oesophagus in
portal hypertension: Practical implications. Br J Surg 1986;73:52531.
53.
Durtschi MB, Carrico CJ, Johansen KH. Esophageal transection fails to salvage high risk
cirrhotic patients with variceal haemorrhage. Am J Surg 1986;150:1823.
54. Jenkins SA, Shields R, Jaser N, et al. The management of gastrointestinal haemorrhage by
somatostatin after apparently successful endoscopic sclerotherapy for bleeding oesophageal
varices. J Hepatol 1991;12:296301.
55. Shemesh E, Bat L. Esophageal perforation after fiberoptic endoscopic injection sclerotherapy
for esophageal varices. Arch Surg 1986;121:2435.
56. Bornman PC, Kahn D, Terblanche J, et al. Rigid versus fiberoptic endoscopic injection
sclerotherapy. Ann Surg 1988;208:1758.
57. McGrew W, Goodin J, Stuck W. Fatal complication of endoscopic sclerotherapy: Serratia
marcescens bacteremia with delayed esophageal perforation. Gastrointest Endosc
1985;31:32931.
58. Pillay P, Starzl TE, Van Thiel DH. Complications of sclerotherapy for esophageal varices in
liver transplant candidates. Transplant Proc 1990;22:214951..
59. Bacon BR, Camara DS, Duffy MC. Severe ulceration and delayed perforation of the
esophagus after endoscopic variceal sclerotherapy. Gastrointest Endosc 1987;33:3115.
60. Grobe JI, Kozarek RA, Sanowski RA, et al. Venography during endoscopic injection
61. Soderlund C, Wiechel KL. Oesophageal perforation after sclerotherapy for variceal
haemorrhage. Acta Chir Scand 1983;149:4915.
62. Barthel JS, Sprouse RF, Dix JD, Sunderrajan EV. Fatal candida esophageal abscess and
sepsis complicating endoscopic variceal sclerosis. Gastrointest Endosc 1987;33:10710.
63. Vickers CR, O'Connor HJ, Quintero GA, et al. Delayed perforation of the esophagus after
variceal sclerotherapy and hepatic transplantation. Gastrointest Endosc 1989;35:45961.
64. Huizinga WKJ, Keenan JP, Marszalek A. Sclerotherapy for bleeding oesophageal varicesA
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
fatal complication. S Afr Med J 1984;65:4368.

Shibuya S, Takase Y, Aoyagi K, et al. Esophageal perforation after endoscopic sclerotherapy
for esophageal varices: A case report with review of the literature. Dis Esophagus
1989;3:2038.
Perino LE, Gholson CF, Goff JS. Esophageal perforation after fibreoptic variceal
Korula J, Pandya K, Yamada S. Perforation of esophagus after endoscopic variceal
sclerotherapy. Incidence and clues to pathogenesis. Dig Dis Sci 1989;34:3249.
Meier JH, Zeitlin JH, Smith MT. Post-sclerotherapy intramural esophageal hematoma:
Endoscopic and radiologic findings. Gastrointest Endosc 1992;38:1023.
Salomez D, Ponette E, Van Steenbergen W. Intramural hematoma of the esophagus after
variceal sclerotherapy. Endoscopy 1991;23:299301.
Low DE, Patterson DJ. Complete esophageal obstruction secondary to dissecting intramural
hematoma after endoscopic variceal sclerotherapy. Am J Gastroenterol 1988;83:4358.
Korula J. Pseudotumor of the esophagus: An unusual complication of esophageal variceal
sclerotherapy. Am J Gastroenterol 1985;80:9546.
Van Steenbergen W, Fevery J, Broeckaert L, et al. Intramural hematoma of the esophagus:
Unusual complications of variceal sclerotherapy. Gastrointest Radiol 1984;9:2935.
Mosimann F, Bronnimann B. Intramural haematoma of the oesophagus complicating
sclerotherapy for varices. Gut 1994;35:1301.
McGrath JP, Walsh TN, Hennessy TP. Total dysphagia from intramural haematoma following
sclerotherapy for oesophageal varices. Br J Surg 1992;79:781.
Bak YT, Yu ZM, Cho JY, et al. Intramural hematoma of the esophagus: Unusual complication
of esophageal variceal sclerotherapy. Korean J Intern Med 1988;3:9294.
Farrell RJ, Goldberg PA, Krige JEJ, et al. Oesophageal stricture following variceal
sclerotherapy in portal hypertension. S Afr Med J 1992;81:989.
Waring JP, Sanowski RA. Food impaction and strictures after sclerotherapy of esophageal
varices. Am J Gastroenterol 1988;83:12457.
Haynes WC, Sanowski RA, Foutch PG, Bellapravula S. Esophageal strictures following
endoscopic variceal sclerotherapy: Clinical course and response to dilatation therapy.
Snady H, Korsten MA. Prevention of dysphagia and stricture formation after endoscopic
sclerotherapy. Gastroenterology 1984;86:1258.
Kochhar R, Goenka MK, Mehta SK. Esophageal strictures following endoscopic variceal
sclerotherapy. Antecedents, clinical profile, and management. Dig Dis Sci 1992;37:34752.
Snady H, Rosman AS, Korsten MA. Prevention of stricture formation after endoscopic
Sauerbruch T, Wirsching R, Leisner B, et al. Esophageal function after sclerotherapy of
bleeding varices. Scand J Gastroenterol 1982;17:74551.
Reilly JJ, Schade RR, Van Thiel DS. Esophageal function after injection sclerotherapy:
Pathogenesis of esophageal stricture. Am J Surg 1984;147:858.
Larson GM, Vandertoll DJ, Netscher DT, Polk HC. Esophageal motility: Effects of injection
sclerotherapy. Surgery 1984;96:7039.
Ogle SJ, Kirk CJC, Bailey RJ, et al. Oesophageal function in cirrhotic patients undergoing
injection sclerotherapy for oesophageal varices. Digestion 1978;18:17885.
Soderlund C, Thor K, Wiechel K-L. Oesophageal motility after sclerotherapy for bleeding
varices. Acta Chir Scand 1985;151:24953.
Spence RA, Smith JA, Isaacs S, Terblanche J. Disturbed oesophageal motility after
eradication of varices by chronic sclerotherapyA scintigraphic study. S Afr Med J
1990;77:13840.
Snady H, Korsten MA. Esophageal acid-clearance and motility after endoscopic sclerotherapy
of esophageal varices. Am J Gastroenterol 1986;81:41922.
Suzuki S, Fujita R, Takahashi H, et al. Effects of endoscopic injection sclerotherapy for
esophageal varices on the reflux prevention mechanisms of the lower esophageal sphincter.
Dig Endosc 1991;3:5161..
90. Cohen LB, Simon C, Korsten MA, et al. Esophageal motility and symptoms after endoscopic
injection sclerotherapy. Dig Dis Sci 1985;30:2932.
91. Shoenut JP, Micflikier AB. Retrosternal pain subsequent to sclerotherapy. Gastrointest
Endosc 1986;32:847.
92. Siemens F, Paquet KJ, Koussouris P, et al. Long-term endoscopic injection sclerotherapy of
bleeding esophageal varices. A prospective analysis of results by endoscopy, manometry and
24-hr pH-monitoring. Surg Endosc 1989;3:13741.
93. Sauerbruch T, Wirsching R, Holl J, et al. Effects of repeated injection sclerotherapy on acid
gastroesophageal reflux. Gastrointest Endosc 1986;32:813.
94. Kinoshita Y, Kitajima N, Itoh T, et al. Gastroesophageal reflux after endoscopic injection
sclerotherapy. Am J Gastroenterol 1992;87:2826.
95. Davion T, Delamarre J, Reix N, et al. Gastric bezoar: Another side effect of endoscopic
variceal sclerotherapy. Scand J Gastroenterol 1989;24:81820.
96.
DeMarino GB, Sumkin JH, Leventhal R, Van Thiel DH. Pneumatosis intestinalis and
pneumoperitoneum after sclerotherapy. AJR 1988;151:9534.
97. Scherl EJ, Fabry TL. Pseudodiverticula secondary to injection sclerotherapy. J Clin
98. Gottfried EB, Goldberg HJ. Mucosal bridge of the distal esophagus after esophageal variceal
sclerotherapy. Gastrointest Endosc 1985;31:2679.
99. Barsoum MS, Abdel-Wahab Mooro H, Bolous FI, et al. The complications of injection
sclerotherapy of bleeding oesophageal varices. Br J Surg 1982;69:7981.
100. Macias Rodriguez MA, Soria de la Cruz MJ, Iglesias Arrabal M, Martin Herrera L. Esophageal
carcinoma after the endoscopic sclerotherapy of varices. Rev Esp Enferm Dig 1992;82:436.
101.
Nakamura R, Watanabe M, Sugimura Y, et al. Early esophageal cancer discovered 6 months
after endoscopic injection sclerotherapy of esophageal varices. Gan No Rinsh
1988;34:11904.
102. Kokudo N, Sanjo K, Umekita N, et al. Squamous cell carcinoma after endoscopic injection
sclerotherapy for esophageal varices. Am J Gastroenterol 1990;85:8614.
103. Bochna GS, Harty RF, Harned RK, Markin RS. Development of squamous cell carcinoma of
the esophagus after endoscopic variceal sclerotherapy. Am J Gastroenterol 1988;83:5648.
104. Dina R, Cassisa A, Baroncini D, D'Imperio N. Role of esophageal brushing cytology in
monitoring patients treated with sclerotherapy for esophageal varices. Acta Cytol
1992;36:4779.
105. Charng MJ, Wang SP, Chang MS. Coronary spasm complicating sclerotherapy of esophageal
varices. Chest 1988;93:2045.
106. Perakos PG, Cirbus JJ, Camara DS. Persistent bradyarrhythmia after sclerotherapy for
esophageal varices. South Med J 1984;77:5312.
107. Paterlini A, Salmi A, Buffoli F, Lombardi C. Heart failure and endoscopic sclerotherapy of
variceal bleeding [Letter]. Lancet 1984;1:1241.
108. Imperiali G, Terruzzi V, Spotti D. Heart failure as a side effect of polidocanol given for
oesophageal variceal sclerosis. Endoscopy 1986;18:207.
109. Knauer CM, Fogel MR. Pericarditis: Complication of esophageal sclerotherapy. Report of
three cases. Gastroenterology 1987;93:28790.
110. Caletti GC, Brocchi E, Labriola E, et al. Pericarditis: A probably overlooked complication of
endoscopic variceal sclerotherapy. Endoscopy 1990;22:1445.
111. Brown DL, Luchi RJ. Cardiac tamponade and constrictive pericarditis complicating
endoscopic sclerotherapy. Arch Intern Med 1987;147:216971.
112. Tabibian N, Schwartz JT, Smith JL, Graham DY. Cardiac tamponade as a result of
endoscopic sclerotherapy: Report of a case. Surgery 1987;102:5467.
113. Baydur A, Korula J. Cardiorespiratory effects of endoscopic esophageal variceal
sclerotherapy. Am J Med 1990;89:47782.
114. Zeller FA, Cannan CR, Prakash UB. Thoracic manifestations after esophageal variceal
sclerotherapy. Mayo Clin Proc 1991;66:72732.
115.
Crawford DC, Ryan DW. Acute respiratory insufficiency after endoscopy for bleeding
oesophageal varices. BMJ (Clin Res Ed) 1984;288:163940.
116. Sukigara M, Omuta R, Tatsuya M. Systemic dissemination of ethanolamine oleate after
injection sclerotherapy for esophageal varices. Arch Surg 1985;120:8336.
117. Connors AF Jr, Bacon BR, Miron SD. Sodium morrhuate delivery to the lung during
endoscopic variceal sclerotherapy. Ann Intern Med 1986;105:53942.
118. DePuey EG, Richards WO, Millikan WJ, Henderson JM. Scintigraphic detection of pulmonary
embolization of esophageal variceal sclerosant. Endoscopy 1988;20:914.
119. Bailey-Newton RS, Connors AF Jr, Bacon BR. Effect of endoscopic variceal sclerotherapy on
gas exchange and hemodynamics in humans. Gastroenterology 1985;89:36873.
120. Jones JG, Minty BD, Beeley JM, et al. Pulmonary epithelial permeability is immediately
increased after embolisation with oleic acid but not with neutral fat. Thorax 1982;37:16974.
121. Spragg RG, Abraham JL, Loomis WH. Pulmonary platelet deposition accompanying acute
oleic-acid-induced pulmonary injury. Am Rev Respir Dis 1982;126:5537.
122. Ashbaugh DG, Uzawa T. Respiratory and hemodynamic changes after injection of free fatty
acids. J Surg Res 1968;8:41723.
123. Griffith BP, Caroll RG, Hardesty RL, et al. Selected lobar injury after infusion of oleic acid. J
Appl Physiol 1979;47:70611..
124. Glauser FL, Fairman RP, Monroe P, Hammond B. Transient pulmonary hypertension
associated with esophageal sclerotherapy. Chest 1984;86:6589.
125. Sarin SK, Sethi KK, Nandi R. Pulmonary hemodynamic changes after intravariceal
sclerotherapy with absolute alcohol. Gastrointest Endosc 1988;34:4036.
126. Korula J, Baydur A, Sassoon C, Sakimura I. Effect of esophageal variceal sclerotherapy
(EVS) on lung function. Arch Intern Med 1986;146:151720.
127. Kitano S, Iso Y, Yamaga H, et al. Temporary deterioration of pulmonary functions after
injection sclerotherapy for cirrhotic patients with esophageal varices. Eur Surg Res
1988;20:298303.
128. Samuels T, Lovett MC, Campbell IT, et al. Respiratory function after injection sclerotherapy of
oesophageal varices. Gut 1994;35:145963.
129. Saks BJ, Kilby AE, Dietrich PA, et al. Pleural and mediastinal changes following endoscopic
injection sclerotherapy of esoph-ageal varices. Radiology 1983;149:63942.
130.
Mauro MA, Jaques PF, Swantkowski TM, et al. CT after uncomplicated esophageal
sclerotherapy. AJR 1986;147:5760.
131. Bacon BR, Bailey-Newton RS, Connors AF Jr. Pleural effusions after endoscopic variceal
sclerotherapy. Gastroenterology 1985;88:191014.
132. Parikh SS, Amarapurkar DN, Dhawan PS, et al. Development of pleural effusion after
sclerotherapy with absolute alcohol. Gastrointest Endosc 1993;39:4045.
133. Carr-Locke DL, Sidky K. Broncho-oesophageal fistula: A late complication of endoscopic
variceal sclerotherapy. Gut 1982;23:10057.
134. Wilborn SL, Rector WG Jr, Schaefer JW. An esophagobronchial fistula after endoscopic
variceal sclerotherapy. J Clin Gastroenterol 1988;10:813.
135. Alwmark A, Bengmark S, Borjesson B, et al. Emergency and long-term transesophageal
sclerotherapy of bleeding esophageal varices. A prospective study of 50 consecutive cases.
Scand J Gastroenterol 1982;17:40912.
136. Gertsch P, Mosimann R. Chylothorax complicating sclerotherapy for bleeding oesophageal
137. Nygaard SD, Berger HA, Fick RB. Chylothorax as a complication of oesophageal
sclerotherapy. Thorax 1992;47:1345.
138. Rajagopalan N, Hoffstein V. Hemothorax following uncomplicated sclerotherapy for
esophageal varices. Chest 1994;106:3145.
139. van Zaanen HCT, Schotborgh RH, Chamuleau RAFM. Meningococcus septicemia: A
complication of endoscopic injection sclerotherapy. Am J Gastroenterol 1990;85:6312.

140. Baskin G. Prosthetic endocarditis after endoscopic variceal sclerotherapy: A failure of
antibiotic prophylaxis. Am J Gastroenterol 1989;84:3112.
141. Kumar P, Mehta SK, Devi BI, et al. Pyogenic meningitis and cerebral abscesses after
endoscopic injection sclerotherapy. Am J Gastroenterol 1991;86:16724.
142. Dewar TN, Thompson CE, Bass NM. Subdural empyema after endoscopic sclerotherapy. Am
J Med 1989;87:5934.
143. Wang WM, Chen CY, Jan CM, et al. Central nervous system infection after endoscopic
injection sclerotherapy. Am J Gastroenterol 1990;85:8657.
144. Robert JY, Raoul JL, Bretagne JF, et al. Unusual presentation of a case of brain abscess after
endoscopic injection sclerotherapy of esophageal varices. Endoscopy 1991;23:2378.
145. Hassig WM, Dekovich AA. Brain abscess after sclerotherapy. Am J Gastroenterol
1992;87:679.
146. Cohen FL, Koerner RS, Taub SJ. Solitary brain abscess following endoscopic injection
sclerosis of esophageal varices. Gastrointest Endosc 1985;31:3313.
147. Nagral A, Nagral SS, Abraham P, Vora I. Splenic abscessA possible complication of
endoscopic variceal sclerotherapy. Indian J Gastroenterol 1993;12:567.
148. Ritchie MT, Lightdale CJ, Botet JF. Bilateral perinephric abscesses: A complication of
endoscopic injection sclerotherapy. Am J Gastroenterol 1987;82:6703.
149. Yatziv I, Katz S, Granot E. Pneumococcal empyema following endoscopic sclerotherapy in a
child with cavernous transformation of the portal vein. Isr J Med Sci 1992;28:3624.
150. Chertow GM, Marcantonio ER, Wells RG. Saccharomyces cerevisiae empyema in a patient
with esophagopleural fistula complicating variceal sclerotherapy. Chest 1991;99:15189.
151. Barnett JL, Elta G. Bacterial peritonitis following endoscopic variceal sclerotherapy.
152. Tam F, Chow H, Prindiville T, et al. Bacterial peritonitis following esophageal injection
sclerotherapy for variceal hemorrhage. Gastrointest Endosc 1990;36:1313.
153. Schembre D, Bjorkman DJ. Post-sclerotherapy bacterial peritonitis. Am J Gastroenterol
1991;86:4816.
154. Bac DJ, de Marie S, Siersema PD, et al. Post-sclerotherapy bacterial peritonitis: A
complication of sclerotherapy or of variceal bleeding?. Am J Gastroenterol 1994;89:85962.
155. Snady H, Korsten MA, Waye JD. The relationship of bacteremia to the length of injection
needle in endoscopic variceal sclerotherapy. Gastrointest Endosc 1985;31:2436.
156. Gerhartz HH, Sauerbruch T, Weinzierl M. Nosocomial septicemia in patients undergoing
sclerotherapy for variceal hemorrhage. Endoscopy 1984;16:12930.
157. Cohen LB, Korsten MA, Scherl EJ, et al. Bacteremia after endoscopic injection sclerosis.
Gastrointest Endosc 1983;29:198200..
158. Camara DS, Gruber M, Barde CJ, et al. Transient bacteremia following endoscopic injection
sclerotherapy of esophageal varices. Arch Intern Med 1983;143:13502.
159. Brayko CM, Kozarek RA, Sanowski RA, Testa AW. Bacteremia during esophageal variceal
sclerotherapy: Its cause and prevention. Gastrointest Endosc 1985;31:102.
160. Sauerbruch T, Holl J, Ruckdeschel G, et al. Bacteriaemia associated with endoscopic
sclerotherapy of oesophageal varices. Endoscopy 1985;17:1702.
161. Low DE, Shoenut JP, Kennedy JK, et al. Infectious complications of endoscopic injection
sclerotherapy. Arch Intern Med 1986;146:56971.
162. Hegnhoj J, Andersen JR, Jarlov JO, et al. Bacteraemia after injection sclerotherapy of
oesophageal varices. Liver 1988;8:16771.
163. Lorgat F, Madden MV, Kew G, et al. Bacteremia after injection of esophageal varices. Surg
Endosc 1990;4:189.
164. Ho H, Zuckerman MJ, Wassem C. A prospective controlled study of the risk of bacteremia in
emergency sclerotherapy of esophageal varices. Gastroenterology 1991;101:16428.
165. Rolando N, Gimson A, Philpott-Howard J, et al. Infectious sequelae after endoscopic
sclerotherapy of oesophageal varices: Role of antibiotic prophylaxis. J Hepatol
1993;18:2904.
166. Pulanic R, Vrhovac B, Jereb B, Jokic N. Controlled trial of the prophylactic administration of
antibiotics in sclerotherapy of esophageal varices. J Chemother 1989;1:2615.
167.
Manzione NC, Das KM, Wolkoff AW, Carnevale N. Unusual sites of upper gastrointestinal
variceal bleeding. J Clin Gastroenterol 1989;9:402.
168. Korula J, Ralls P. The effects of chronic endoscopic variceal sclerotherapy on portal pressure
in cirrhotics. Gastroenterology 1991;101:8005.
169. Foutch PG, Sivak MV. Colonic variceal hemorrhage after endoscopic injection sclerosis of
esophageal varices: A report of three cases. Am J Gastroenterol 1984;79:75660.
170. Dilawari JB, Raju GS, Chawla YK. Development of large spleno-adreno-renal shunt after
endoscopic sclerotherapy. Gastroenterology 1989;97:4216.
171. D'Amico G, Montalbano L, Triana M. Natural history of congestive gastropathy in cirrhosis.
172. Dilawari JB, Chawla YK, Kaur U. Development of the Cruveilhier-Baumgarten syndrome after
endoscopic obliteration of oesopha-geal varices. Report of a case. Acta Radiologica
1989;30:4012.
173. Keane RM, Britton DC. Massive bleeding from rectal varices following repeated injection
sclerotherapy of oesophageal varices. Br J Surg 1986;73:120.
174. Fry RD, Fischer KC, Susman N, et al. Adhesion-related variceal hemorrhage following
sclerosis of esophageal varices. Arch Surg 1988;123:945.
175. Arst HF, Reynolds JDH. Acute ileal variceal hemorrhage secondary to esophageal
176. Eleftheriadis E. Duodenal varices after sclerotherapy for esophageal varices. Am J
177. McCormack TT, Smith PM, Rose JD, Johnson AG. Perforating veins and blood flow in
178.
Aoki H, Hasumi A, Shimazu M. The hemodynamics of esophagogastric varices: Significance
of esophagogastric arterial inflow in their information. In Idezuki Y (ed). Treatment of
Esophageal Varices. International Congress Series 794. New York: Elsevier, 1988;31528.
179. Kang JH, Kambayashi J, Sakon M, et al. Mechanism of the haemostatic effect of
ethanolamine oleate in the injection sclerotherapy for oesophageal varices. Br J Surg
1987;74:503.
180. Musso R, Longo A, Triolo A, et al. Polidocanol may directly activate the contact phase of
blood coagulation during sclerotherapy. Gastrointest Endosc 1987;33:4001.
181. Jacobson BF, Franz RC, Hurly EM, et al. Mechanism of thrombosis caused by sclerotherapy
of esophageal varices using sodium tetradecyl sulphate. Surg Endosc 1992;6:49.
182. Sarfeh IJ. Portal vein thrombosis associated with cirrhosis. Clinical importance. Arch Surg
1979;114:9025.
183. Okuda K, Ohnishi K, Kimura K. Incidence of portal vein thrombosis in liver cirrhosis. An
angiographic study on 708 patients. Gastroenterology 1985;89:27986.
184. Belli L, Sansalone CV, Aseni P. Portal vein thrombosis in cirrhosis. A retrospective analysis.
Ann Surg 1986;203:28691.
185. Thatcher BS, Sivak MV Jr, Ferguson DR, Petras RE. Mesenteric venous thrombosis as a
possible complication of endoscopic sclerotherapy: A report of two cases. Am J Gastroenterol
1986;81:1269.
186. Goldberg H, Fabry TL. Mesenteric thrombosis following sclerotherapy during vasopressin
infusion: Mechanism and therapeutic implications. J Clin Gastroenterol 1989;11:567.
187. Ashida H, Kotoura Y, Nishioka A, et al. Portal and mesenteric venous thrombosis as a
complication of endoscopic sclerotherapy. Am J Gastroenterol 1989;84:30610.
188. Deboever G, Elegeert I, Defloor E. Portal and mesenteric venous thrombosis after endoscopic
injection sclerotherapy. Am J Gastroenterol 1989;84:13367.
189. Leach SD, Meier GH, Gusberg RJ. Endoscopic sclerotherapy: A risk factor for splanchnic
venous thrombosis. J Vasc Surg 1989;10:912 [discussion 123].
190. Terada T, Nakanuma Y, Yonejima M, et al. Portal, mesenteric, and splenic venous
thrombosis after endoscopic injection sclerotherapy. J Clin Gastroenterol 1990;12:2389..
191. Ofek B, Shemesh D, Abramowitz HB. Mesenteric vein thrombosis after injection sclerotherapy
for oesophageal varices. Case report. Eur J Surg 1992;158:1956.
192. Korula J, Yellin A, Kanel GC, Nichols P. Portal vein thrombosis complicating endoscopic
variceal sclerotherapy. Convincing further evidence. Dig Dis Sci 1991;36:11647.
193. Stoltenberg PH, Goodale RL, Silvis SE. Portal vein thrombosis following combined
endoscopic variceal sclerosis and vasopressin therapy for bleeding varices. Am J
194.
Rice S, Lee KP, Johnson MB, et al. Portal venous system after portosystemic shunts or
endoscopic sclerotherapy: Evaluation with Doppler sonography. AJR 1991;156:859.
195. Caletti GC, Brocchi E, Zani L, et al. Sonographic evaluation of the portal venous system after
elective endoscopic sclerotherapy of esophageal varices. Surg Endosc 1987;1:1657.
196. Hunter GC, Steinkirchner T, Burbige EJ, et al. Venous complications of sclerotherapy for
esophageal varices. Am J Surg 1988;156:497501.
197. Kawasaki S, Henderson JM, Riepe SP, et al. Endoscopic variceal sclerosis does not increase
the risk of portal venous thrombosis. Gastroenterology 1992;102:20615.
198. Chaudhary A, Tatke M, Aranya RC. Endoscopic sclerotherapy: The far and near effects. Br J
Surg 1990;77:963.
199. Seidman E, Weber AM, Morin CL, et al. Spinal cord paralysis following sclerotherapy for
esophageal varices. Hepatology 1984;4:9504.
200. Yuki N, Kubo M, Noro Y, et al. Manifestations of temporary symptoms during endoscopic
variceal sclerotherapy using thrombin as a sclerosant. Jpn J Med 1991;30:193201.
201. Yamaga H, Hashizume M, Kitano S, et al. Platelet aggregability after endoscopic intravariceal
injection of 5 per cent ethanolamine oleate into oesophageal varices. Br J Surg
1989;76:93942.
202. Ohta M, Hashizume M, Ueno K, et al. Albumin inhibits hemolysis of erythrocytes induced by
ethanolamine oleate during endoscopic injection sclerotherapy. Hepatogastroenterology
1993;40:658.
203.
Kujii Y, Sugawa C, Ozawa C. Haemostasis activation during esophageal variceal
sclerotherapy with thrombin in cirrhotics. Ann Surg 1991;57:2225.
204. de Franchis R, Cipolla M, Primignani M, et al. Activation of coagulation in cirrhotics after
endoscopic variceal sclerotherapy. Am J Gastroenterol 1987;82:128791.
205. Miyoshi H, Ohshiba S, Matsumoto A, et al. Haptoglobin prevents renal dysfunction associated
with intravariceal infusion of ethanolamine oleate. Am J Gastroenterol 1991;86:163841.
206. Kjaergaard J, Fischer A, Miskowiak J, et al. Sclerotherapy of bleeding oesophageal varices.
Long-term results. Scand J Gastroenterol 1982;17:3637.
207. The Copenhagen Esophageal Varices Sclerotherapy Project: Sclerotherapy after first variceal
208. Korula J, Balart LA, Radvan G, et al. A prospective, randomized controlled trial of chronic
209. Fleig WE, Stange EF, Hunecke R, et al. Prevention of recurrent bleeding in cirrhotics with
recent variceal haemorrhage: Prospective randomized comparison of propranolol and
sclerotherapy. Hepatology 1987;7:35561.
210. Alexandrino PT, Alves MM, Pinto Correia J. Propranolol or endoscopic sclerotherapy in the
prevention of recurrence of variceal bleeding. A prospective, randomized controlled trial. J
Hepatol 1988;7:17585.
211. El-Zayadi A, El-Din SS, Kabil SM. Endoscopic sclerotherapy versus medical treatment for
bleeding esophageal varices in patients with schistosomal liver disease. Gastrointest Endosc
1988;34:3147.
212. Magnano A, Passanisi G, Longo C, et al. Early and late complications of endoscopic
oesophageal varices sclerotherapy. Surg Endosc 1988;2:20912.
213. Van Hootegem PH, Van Besien K, Broeckaert L, et al. Endoscopic sclerotherapy of
esophageal varices. Long-term follow-up, recurrence, and survival. J Clin Gastroenterol

1988;10:36872.
214. McKee RF, Garden OJ, Carter DC. Injection sclerotherapy for bleeding varices: Risk factors
and complications. Br J Surg 1991;78:10981101.
215. Sarin SK, Nanda R, Sachdev G. Relative efficacy and safety of absolute alcohol and 50%
alcohol as variceal sclerosants. Gastrointest Endosc 1987;33:3627.
216. Sarin SK, Kumar A. Sclerosants for variceal sclerotherapy: A critical appraisal. Am J
217. Kochhar R, Goenka MK, Mehta S, Mehta SK. A comparative evaluation of sclerosants for
esophageal varices: A prospective, randomized, controlled study. Gastrointest Endosc
1990;36:12730.
218. Chawla Y, Dilawari JB. Studies in sclerotherapy: I. Comparison of sodium tetradecyl sulphate
(STS) with absolute alcohol (AA) as sclerosants in the treatment of esophageal varices. J Clin
Gastroenterol 1990;12:37880..
Chapter 36 Endoscopic Elastic Band Ligation for Bleeding

Esophageal Varices
(2471)
(2472)
GREGORY STIEGMANN, M.D.
Endoscopic variceal sclerotherapy improves outcome in patients with variceal hemorrhage. In addition
to initial control of bleeding, a significant benefit for sclerotherapy in terms of recurrent bleeding and
survival has been confirmed by metaanalysis.1,2(2473) Furthermore, sclerotherapy has been found to
be equal or superior to shunt surgery in several trials (see Chapter 33: Indications and Results of
Variceal Injection Sclerotherapy).39(2474)
Despite these data, however, sclerotherapy has a significant incidence of associated local and
systemic complications (see Chapter 35: Complications of Sclerotherapy). There is, in fact, an
extremely long list of possible complications. These may arise as a result of the endoscopic procedure
itself or as an effect of the injected sclerosant, so that complications are usually divided into local
esophageal and systemic categories. Local complications include esophageal ulceration, stricture
formation, perforation (localized or free), and development of a hematoma within the wall of the
esophagus. Esophageal ulceration is extremely common, although most ulcers heal spontaneously or
with medical therapy.1012(2475) Larger, deeper ulcers, which usually signify the presence of
esophageal wall necrosis, may lead to perforation or stricture formation, or both. Necrosis and
perforation occur in from 1 to 6% of patients; those with poor liver function are more prone to this
complication.10(2476) Long-term effects of sclerotherapy include abnormal esophageal motility and
lower esophageal sphincter pressures in some but not all postsclerotherapy manometric
studies.1315(2477)
Systemic complications include bacteremia, thrombotic events, and a variety of adverse
cardiorespiratory reactions. Chest x-rays in patients who have undergone sclerotherapy exhibit a wide
range of abnormalities.16(2478) Although most of these findings resolve, more serious pulmonary
problems, including the adult respiratory distress syndrome, have been reported in association with
sclerotherapy.17,18(2479) Bacteremia occurs in from 5 to 10% of patients.1921(2480) Although this
may not represent a hazard for many sclerotherapy patients, those who have ascites appear to be at
increased risk for bacterial peritonitis.22(2481) Thrombotic complications including paraplegia resulting
from thrombosis of the anterior spinal artery and portal and mesenteric venous thrombosis have been
reported in small numbers of patients.2325(2482) Less serious systemic effects caused by injected
sclerosants include chest pain and fever lasting from 24 to 48 hr in up to 40% of patients.
Endoscopic variceal ligation using elastic bands was developed to provide an alternative to endoscopic
sclerotherapy in the treatment of bleeding esophageal varices. This chapter describes the principles
and operation of the endoscopic ligating device and reviews the results of experimental endoscopic
ligation studies as well as the results of ongoing and recently concluded clinical trials of endoscopic
ligation.
Endoscopic Elastic Band Ligation of Esophageal Varices

Injection sclerotherapy was for many years the accepted nonoperative treatment for refractory internal
hemorrhoids. Elastic band ligation of internal hemorrhoids was introduced by Blaisdell26(2483) in 1958
and refined by Barron27(2484) in 1963. Elastic band ligation has almost completely replaced injection
sclerotherapy in North America for the treatment of patients with symptomatic internal hemorrhoids
who fail medical treatment.28,29(2485)
Complications of elastic band ligation for internal hemorrhoids are uncommon and include delayed
bleeding in about 1%, ulceration, and pain.30(2486) Several deaths have resulted from pelvic cellulitis
secondary to infection at the treated site.28(2487) The results of two of the three prospective trials that
compared elastic band ligation with sclerotherapy for treatment of hemorrhoids indicate that band
ligation was superior to sclerotherapy with regard to prevention of recurrent symptoms and
bleeding.2931(2488) Patients treated with elastic band ligation required fewer repeat treatment
sessions than those treated with sclerotherapy.
The device for endoscopic ligation of varices operates on principles similar to those employed for
elastic band ligation of hemorrhoids. This device (Bard Interventional Products, Billerica,
Massachusetts) consists of four parts: a housing (outer) cylinder, a banding (inner) cylinder, a tripwire,
and a latex O ring (Figure 361). The outer housing cylinder attaches to the end of the endoscope by a
friction mount. The smaller banding cylinder is fitted with a clasp that allows for insertion of the tripwire
and is constructed to fit snugly, yet slide smoothly, inside the housing cylinder. The tripwire is a
monofilament strand with a flange at its distal end that fits into the clasp in the banding cylinder. Small
latex O rings are mounted on the banding cylinder.
(2489)Figure 361. The endoscopic ligating device. The device consists of an outer cylinder
that is attached to the endoscope by a friction mount. The elastic band (marker) is stretched
over the inner cylinder. A tripwire (not visible) is connected to the inner cylinder and runs via
the accessory channel of the endoscope to the operator. The operator pulls the tripwire to eject
the elastic band and ligate the varix (see Figure 362).
The device is assembled with the endoscope as follows: The housing cylinder is secured to the end of
the endoscope. The tripwire is passed through the vacuum lock of the accessory channel entry port
and exits the endoscope via the distal port of the accessory channel. The tripwire is secured to the
clasp in the banding cylinder, and the banding cylinder is "backed" into the housing cylinder. The
banding cylinder is positioned such that approximately 1 mm of this cylinder protrudes beyond the O
ring, which is seated on the banding cylinder and against the distal end of the housing cylinder. Slight
tension is applied by means of a handle to the tripwire where it protrudes from the accessory channel
port on the control section of the instrument. This handle also serves as an extraction tool that
facilitates removal of the spent inner cylinder from the housing cylinder after firing.
The endoscopist identifies a target varix and advances the endoscope under direct vision until the
banding cylinder is in full 360-degree contact with the targeted varix. Once full contact is made, the
endoscopic suction is activated to draw the varix into the banding chamber (Figure 362). When the
target varix has filled the chamber, as evidenced by a complete "red-out" of the visual field, the tripwire
is pulled and the latex O ring is thereby securely fixed around the base of the target varix.
(2490)Figure 362. Schematic diagram of esophageal variceal elastic band ligation. A, Full
360-degree contact between the end of the banding cylinder and the target varix. B, By
activating the endoscopic suction, the varix is drawn into the banding cylinder, producing a
red-out. C, The tripwire is pulled, resulting in movement of the banding (inner) cylinder
toward the endoscope and release of the elastic O ring around the base of the target. D, The
ligated varix. (A-D, From Stiegmann GV, Yamamoto M. Approaches to the endoscopic
treatment of esophageal varices. World J Surg 16:103441, 1992.)
Ligation of multiple varices requires removal, reloading, and reinsertion of the endoscope for each
target. An endoscopic overtube, inserted at the outset of the procedure, facilitates multiple ligations
and prevents inadvertent firing of the device during passage through the oropharynx.
Experimental Studies
The initial evaluation of the endoscopic ligating device was performed in portal hypertensive dogs with
gastroesophageal varices.32,33(2491) Treatments were confined to the distal esophagus where two to
four variceal ligations were performed in each animal. Endoscopic follow-up was obtained at 1, 7, 14,
21, and 60 days as well as prior to sacrifice. Animals were sacrificed at 24 hr (n = 2), 3 to 7 days (n =
2), 14 to 21 days (n = 3), and 50 to 60 days (n = 3). After sacrifice, specimens were fixed and sections
were obtained for histologic analysis.
Initially, ligated varices blanched and became cyanotic within 3 to 7 min. Repeat endoscopy at 24 hr
disclosed minimum change in appearance of ligated tissue sites compared with the endoscopic
appearance at the time of initial treatment. All elastic O rings remained in place despite the fact that no
dietary restrictions were imposed on the animals. Animals sacrificed at this time had no evidence of
bleeding or damage to the esophageal wall. Microscopic examination showed ischemic necrosis of
mucosa and submucosa with intact muscularis proper (Figure 363).
(2492)Figure 363. Photomicrograph of a canine varix at 24 hr after ligation (H & E stain,

2). The raised necrotic area is the site of ligation. Note the presence of vascular channels and
the inflammatory response in the submucosa. (From Stiegmann GV, Sun JH, Hammond WS.
Results of experimental endoscopic esophageal varix ligation. Am Surg 1988; 54:1058.)
At days 3 to 7, there was slough at all treatment sites with shallow (1 to 2 mm) ulcerations 8 to 12 mm
in length. All elastic O rings had been displaced at this time. Microscopic inspection revealed
granulation tissue, sloughing necrotic tissue, and intense inflammatory reaction in the ulcer base with
early scar tissue formation (Figure 364).
(2493)Figure 364. Photomicrograph of the ligated varix at 7 days (H & E stain, 2). An
intense inflammatory reaction is present in the submucosa, and early reepithelialization is seen
at the periphery of the ulcer. The muscularis propria is undisturbed. (From Stiegmann GV,
Sun JH, Hammond WS. Results of experimental endoscopic esophageal varix ligation. Am
Surg 1988; 54:1058.)
There were minimum residual varices and no evidence of full-thickness esophageal injury at 14 to 21
days. Sites that had previously been shallow ulcers appeared by gross inspection to have healed with
slight depressions in the esophageal mucosa measuring from 6 to 10 mm in diameter. Microscopy
showed full-thickness replacement of vascular structures in the submucosa with maturing scar tissue;
continuation of the inflammatory response was uniformly present at this time.
At 50 to 60 days, the presence of shallow, smooth depressions in the esophageal mucosa was
maintained at treated sites. Microscopically, reepithelialization was complete and the entire submucosa
at treated sites was replaced by dense mature scar tissue (Figure 365). The underlying muscular wall
was consistently intact. No animals suffered clinically apparent ill effects following endoscopic elastic
band ligation treatment.
(2494)Figure 365. Photomicrograph of endoscopically ligated varix at 52 days ( 2). The

treated site is reepithelialized, and dense scar tissue has replaced the entire submucosa,
obliterating all vascular channels. The underlying muscular wall is intact. (From Stiegmann
GV, Sun JH, Hammond WS. Results of experimental esophageal endoscopic varix ligation.
Am Surg 1988; 54:1058.)
Additional experimental studies comparing endoscopic ligation with sclerotherapy in a portal
hypertensive canine model have been reported by Jensen et al.3436(2495) Active bleeding was
induced by needle puncture of an esophageal or gastric varix after intravenous administration of
heparin. Test animals received one form of endoscopic treatment, and control animals had no therapy.
Active bleeding from both esophageal and gastric varices was effectively controlled with either ligation
or sclerotherapy. The majority of control animals continued to bleed for periods of greater than 15 min.
Ulceration occurred at all treatment sites in the stomach with either ligation or sclerotherapy. Ulceration
occurred at 90% of the ligated esophageal sites in contrast to only 10% of such sites treated with
sclerotherapy. Treatment-induced ulcers at the site of ligation of canine gastric varices were larger and
deeper than those induced by sclerotherapy, but healing was equally rapid for both cases and resulted
in a decrease in size of the varices.
A prospective randomized clinical study by Young et al.37(2496) compared ulcers induced by
sclerotherapy with those caused by ligation. Sclerotherapy-induced ulcers were significantly deeper
than those induced by ligation. However, the ligation-induced ulcers were significantly larger in surface
area and more circular than the linear lesions induced by sclerotherapy. Ligation resulted in ulcers at
all ligation sites at 7 days following treatment, in contrast to a 90% incidence at injection sites.
Ligation-induced ulcers healed at a mean of 14 days compared with 21 days for those resulting from
sclerotherapy. Two of the 13 sclerotherapy-treated patients developed esophageal strictures compared
with none of the 10 patients in the ligation group. Bleeding from treatment-induced ulceration was not
observed in either cohort. These findings confirm earlier clinical and laboratory observations that the
majority of ligated sites (whether in the stomach or esophagus) slough and produce consistent shallow
ulcerations at from 3 to 7 days after treatment. Ligation-induced ulcers in the clinical setting appear to
be associated with little risk of bleeding or stricture formation.
Technique
Endoscopic variceal ligation is performed in the same manner as conventional endoscopic
sclerotherapy. Patients receive intravenous sedation (usually meperidine and midazolam) or topical
pharyngeal anesthesia, or both. A survey examination of the esophagus, stomach, and duodenum is
performed prior to attaching the ligating device to the endoscope. An endoscopic overtube is inserted
during the survey examination except in patients who are undergoing routine follow-up endoscopy after
variceal eradication has been achieved. If it is determined during the survey examination that
esophageal or gastric varices require treatment, the endoscope is removed and the ligating device is
installed. The loaded, ligating endoscope is then inserted through the overtube.
Treatment of esophageal varices begins with ligation of the most distal esophageal veins, usually at or
just caudad to the gastroesophageal junction. Subsequent ligations are performed at the same level or
more cephalad. Patients with actively bleeding varices are treated in the same manner as those
without active bleeding unless the bleeding site is specifically identified, in which case it is ligated first.
Subsequent ligations are performed as described previously until all variceal channels in the distal 5 to
7 cm of the esophagus are ligated at least once. Patients with bleeding gastric fundal varices are best
treated with a combination of ligation and sclerotherapy (see later in this chapter). Following
endoscopic ligation of esophageal varices, patients are instructed to masticate food well but are not
given specific dietary restrictions. Repeat ligation treatments aimed at eradication of varices are
conducted at 10- to 14-day intervals until venous channels in the distal esophagus are obliterated. All
elective treatments are performed on an outpatient basis.
Clinical Results
Initial clinical experience with endoscopic elastic band ligation included 146 consecutive cirrhotic
patients who had endoscopically documented bleeding from esophageal varices (Table
361).38(2497) Elective retreatments after initial control of hemorrhage were performed at 7- to 14-day
intervals until varices in the distal esophagus were eradicated. Thirty-three of the 146 total patients
(23%) were actively bleeding at the index treatment. Control of bleeding was achieved in 31 of these
33 patients (94%) (during index hospitalization or until the time of death). One hundred fourteen
patients survived initial hospitalization, 65 of whom (57%) experienced 72 episodes of recurrent
variceal bleeding. Most such episodes occurred prior to eradication of varices. Overall survival was
73%. Varices were eradicated from the distal esophagus in 80% of the patients who remained in the
study more than 30 days, with the mean number of ligation treatment sessions being 5.5. During the
course of this study, there were 4 nonbleeding complications that required treatment. Two strictures
responded to a single treatment by bougienage. One patient developed a meat impaction that was
managed by endoscopic disimpaction. A fourth patient developed dysphagia that lasted for 36 hours
following an elective treatment; topical anesthetics and antispasmodic drugs were required for relief of
dysphagia.
Initial Results of Endoscopic Variceal Band Ligation

in 146 Consecutive Patients
TABLE 361
Number of Patients
Mean age (years)
Child/Pugh class
A
B
C
Mean follow-up (months)
Active bleeding (%)
Varices controlled
Varices eradicated
Mean number of
treatments
Recurrent bleeding (%)
Survival rate
LANNEC'S CIRRHOSIS
OTHER CIRRHOSIS
93
52
53
53
29
34
30
15.6
20(21)
90%
80%
5.6
25
19
9
15.2
13(24)
100%
78%
5.4
47(50)
74%
10(30)
72%
Goff et al.13(2498) obtained esophageal manometry in patients who had undergone endoscopic
ligation, patients who had been treated with sclerotherapy, and untreated control patients who had
esophageal varices. Although sclerotherapy patients had a greater incidence of stricture formation,
esophageal manometry did not disclose any persistent, long-term differences among the three groups.
Endoscopic ligation was examined in three additional uncontrolled trials. In the trial of Hall et
al.,39(2499) 16 infants and children were treated under general endotracheal anesthesia that obviated
use of the overtube in many cases. Varices were eradicated in 6 patients and 8 remained in treatment
at the time of this report. Saeed et al.40(2500) reported the use of endoscopic ligation to successfully
salvage 10 patients in whom sclerotherapy was considered to have failed because of multiple episodes
of recurrent bleeding. Yamamoto et al.41(2501) reported favorable preliminary results both with
endoscopic ligation and with endoscopic ligation combined with low-volume sclerotherapy.
Data from eight prospective randomized trials of ES versus EVL in the treatment of esophageal
variceal bleeding have been published and are summarized in Table 362.4249(2502) Patients with
acute variceal bleeding were randomized in six of these studies. 4247(2503) In the trial of Avgerinos
et al.,48(2504) patients were randomized after initial treatment by ES; in that of Baroncini et
al.,49(2505) randomization occurred after control of the acute episode of bleeding. ES and EVL were
equivalent in the control of acute variceal bleeding in the six trials that randomized such
patients.4247(2506) Every trial published thus far has demonstrated that variceal eradication can be
achieved with significantly fewer EVL treatments compared with ES. Significant differences in rates of
recurrent bleeding in favor of EVL have been found in six trials.4449(2507) Furthermore, there were
significantly fewer complications in six trials among patients undergoing EVL.42,43,45,46,48,49(2508)
Although there was no difference in mortality in most trials, survival was significantly better for patients
undergoing EVL in two.42,46(2509) A metaanalysis published by Laine and Cook50(2510) in 1995
confirmed that EVL compared with ES has lower rates of recurrent bleeding and complications, has
improved survival, and requires fewer treatment sessions to eradicate varices.
Results From Prospective Randomized Trials Comparing Endoscopic

Ligation With Endoscopic Sclerotherapy
TABLE 362
PATIENTS
(n)
THERAPY
REBLEEDING
(%)
MEAN NO. TX
TO ERADICATE
SURVIVAL
(%)
65
ES
48
55*
Laine et al. 43
64
39
EVL
ES
36
44
4
6*
72*
85
Gimson et al. 44
38
49
EVL
ES
26
53*
4*
4.9*
89
37
Hou et al. 45
54
67
EVL
ES
30*
33*
3.4*
4.6*
52
84
Lo et al. 46
67
59
EVL
ES
28*
36*
3.5*
6.5*
79
67
Sarin et al. 47
61
48
EVL
ES
11*
6*
3.8*
5.2*
84
94
Avgerinos et al. 48
47
40
EVL
ES
21*
48*
4.1*
5.8*
94
79
Baroncini et al. 49
37
54
EVL
ES
27*
19*
3.7*
4.0*
80
78
57
EVL
16*
3.5*
79
STUDY
Stiegmann et al.
42
COMPLICA
Results From Prospective Randomized Trials Comparing Endoscopic

Ligation With Endoscopic Sclerotherapy
TABLE 362
STUDY
PATIENTS
(n)
THERAPY
REBLEEDING
(%)
MEAN NO. TX
TO ERADICATE
SURVIVAL
(%)
COMPLICA
Key: *Denotes significant differences; Rebleedingpatients with recurrent bleeding; Mean No. Tx to Eradicatemean num
treatments to eradicate varices; Complicationsnonbleeding complications requiring active treatment; ESendoscopic
sclerotherapy; EVLendoscopic variceal ligation; Significantly more patients withdrawn from this group; NSnot signifi
Although EVL reduces the rate of recurrent variceal bleeding, the rate of recurrence of nonbleeding
esophageal varices in two trials has been higher among patients undergoing EVL compared with those
treated by ES.47,49(2511) Whether the rate of recurrent bleeding in patients with longer follow-up after
EVL might also be higher than expected is unknown at present. In one trial, there was also a significant
increase in bleeding from sites other than the esophagus among patients undergoing EVL.46(2512)
Jalan et al.51(2513) conducted a randomized trial in which patients with cirrhosis were randomized to
receive either a transjugular intrahepatic portosystemic shunt (TIPS) (n = 31) or EVL (n = 27) at 24 hr
after an episode of bleeding. Patients with active bleeding at endoscopy performed within 6 hr of
admission underwent sclerotherapy. Significantly fewer patients in the TIPS group (9.8%) had recurrent
variceal bleeding compared with the group treated by EVL (51.9%). Furthermore, the severity of
recurrent variceal bleeding was also significantly less for TIPS-treated patients. In this study, no
significant difference in complications was found, including hepatic encephalopathy, between the two
treatment groups.
The greater incidence of pulmonary complications in sclerotherapy-treated patients in the multicenter
trial of Stiegmann et al.42(2514) suggests a beneficial effect of ligation over sclerotherapy in this
regard, or alternatively, fewer pulmonary complications in ligation patients may have resulted from
prevention of tracheal aspiration by use of the endoscopic overtube, which in this trial was employed
only in the group treated with ligation. The incidence of pulmonary complications (pneumonia) at
interim analysis is similar in both the ligation and the sclerotherapy arms of the study of Laine et
al.43(2515) in which both cohorts were treated using an overtube, thus suggesting that the use of the
overtube may prevent some pulmonary problems. In the trial of Laine et al.,43(2516) the frequency of
complications in the group of patients treated by ligation was significantly less than that for patients
treated by sclerotherapy. However, the percentage (33%) of sclerotherapy patients who developed
esophageal strictures was inordinately high in comparison with the great majority of published reports
of sclerotherapy. Esophageal ulcers occurred with about equal frequency in both treatment groups but
tended to be more severe in sclerotherapy-treated patients.
Berner et al.52(2517) studied the short-term risks of bacteremia, changes in pulmonary and
coagulation functions, esophageal motility, and gastroesophageal reflux in a prospective randomized
trial of sclerotherapy versus variceal ligation. Although the numbers of patients were small, there were
no significant differences with respect to pulmonary and coagulation parameters or bacteremia.
However, esophageal dysmotility and evidence of reflux were more common in patients undergoing
sclerotherapy. Patient acceptance of ligation procedures was said to be better than that for
sclerotherapy sessions. The use of an overtube during endoscopic variceal ligation has been
associated with complications including perforation and bleeding. These are described for the most
part as case reports.5356(2518) However, a giant ulcer, noted in the proximal esophagus 2 days after
the initial ligation session, was encountered in the randomized trial of Laine et al.43(2519) and
attributed to the use of the overtube.
Endoscopic Ligation Combined with Low-Volume Sclerotherapy

Recurrent bleeding in patients treated with either sclerotherapy or ligation usually results from varices
that have not been eradicated. Eradication usually requires five to six endoscopic treatments with
conventional sclerotherapy and three to four with elastic band ligation.37,4244(2520) In treating a
small number of patients with bleeding gastric and fundal varices with a combination of band ligation
and low-volume sclerotherapy at the same treatment session, we observed that eradication of the
varices occurred after only one or two such treatments.56(2521) We therefore thought that combining
sclerotherapy with endoscopic ligation could result in even more efficient eradication of varices in the
esophagus than is possible with ligation alone. Furthermore, injection of lesser volumes of sclerosant
might minimize complications associated with sclerotherapy.
Combination treatment of esophageal varices involves intravariceal injection of 1 ml of sclerosant per
varix and ligation of varices at the same endoscopic treatment session (Figure 366). Ligation of
individual varices is performed first at sites near the gastroesophageal junction. Injections are then
made at sites 1 to 3 cm cephalad to the ligation sites. When sclerosant is injected into varices in which
blood flow has been decreased by ligation, there should be minimum systemic dissemination. Also, the
volume of sclerosant required is low, and this should reduce the potential for local esophageal
complications.
(2522)Figure 366. Schematic diagram of combined endoscopic ligation and low-volume

sclerotherapy. After ligation of varices at the gastroesophageal junction, 1 ml of sclerosant is
injected into each variceal channel cephalad to the ligation site. (From Stiegmann GV,
Yamamoto M. Approaches to the endoscopic treatment of esophageal varices. World J Surg
16:103441, 1992.)
Forty-six patients were enrolled in a pilot study of combination endoscopic ligation with low-volume
sclerotherapy.56(2523) Preliminary data indicated eradication of esophageal varices could be
accomplished with a mean of approximately two treatment sessions. Ulcerations at the site of ligation
and sclerosant injection were seen in most patients within the first few days of treatment but resulted in
few adverse effects. The overall rate of recurrent bleeding with mean follow-up of 8 months was 30%,
with only 1 death resulting from hemorrhage. Serious complications associated with combination
ligation sclerotherapy included 1 proximal esophageal perforation and 2 cases of spontaneous
bacterial peritonitis.
Preliminary data from the prospective randomized trial of Koutsomanis57(2524) of ligation versus the
combination of ligation and sclerotherapy confirm our initial findings. Eradication of esophageal varices
could be accomplished with a mean of 1.2 combined treatments in contrast to a mean of 4.4
treatments to eradicate using sclerotherapy alone. Hashizume et al.58(2525) conducted a trial in which
25 patients underwent initial sclerotherapy or variceal ligation followed by sclerotherapy alone. The total
volume of sclerosant used to eradicate varices was significantly less in the group of patients who had
combined therapy. Adverse effects were significantly less common in patients who had ligation plus
sclerotherapy. These data tentatively support the theory that more rapid eradication of varices may be
possible with combined ligation and low-volume sclerotherapy with fewer complications. However,
further confirmation and additional data will be necessary before conclusions can be drawn.
Summary
Endoscopic variceal ligation was developed to provide a safer alternative to sclerotherapy for treatment
of bleeding esophageal varices. Laboratory studies indicate the mechanical ligation technique
eradicates varices by obliterating the submucosal space in which these venous channels reside.
Shallow ulcers occur at each site of mechanical ligation but result in few untoward effects. Clinical
studies, including prospective randomized trials, have shown endoscopic elastic band ligation results in
equal or superior outcomes with regard to recurrent bleeding and survival when compared with
conventional endoscopic sclerotherapy. Ligation-treated patients achieve eradication of varices with
fewer treatments, which may translate into an economic advantage for this type of therapy. The
incidence of nonbleeding complications of endoscopic ligation treatment is consistently lower than
those of sclerotherapy in all trials in which these data are reported. Combination elastic band ligation
and low-volume sclerotherapy may be the optimum method for eradication of esophageal varices.
Preliminary data suggest the two therapies administered simultaneously may give the best results in
terms of variceal eradication while preserving the low complication rate associated with endoscopic
ligation alone.
REFERENCES
1.
Pagliari L, Burroughs AK, Sorensen TA, et al. Therapeutic controversies and randomised
controlled trials: Prevention of bleeding and rebleeding in cirrhosis. Gastroenterol Int
1989;2:7184.
2. Infante-Rivard C, Esnaola S, Villeneuve JP. Role of endoscopic variceal sclerotherapy in the
long-term management of variceal bleeding: A meta-analysis. Gastroenterology
1989;96:108792.
patients with severe cirrhosis and variceal hemorrhage. N Engl J Med 1984;311:158994.
patients with severe cirrhosis and acute variceal hemorrhage. Long-term follow-up. N Engl J
Med 1987;316:115.
5. Planas R, Boix J, Broggi M, et al. Portacaval shunt versus endoscopic sclerotherapy in the
elective treatment of variceal hemorrhage. Gastroenterology 1991;100:107886.
6. Rikkers LF, Burnett DA, Volentine GD, et al. Shunt surgery versus endoscopic sclerotherapy
for long-term treatment of variceal bleeding. Early results of a randomized trial. Ann Surg
1987;206:2617.
7. Spina GP, Santambrogio R, Opocher E, et al. Distal splenorenal shunt versus endoscopic
sclerotherapy in the prevention of variceal rebleeding. First stage of a randomized, controlled
trial. Ann Surg 1990;211:17886.
8. Teres J, Bordas JM, Bravo D, et al. Sclerotherapy versus distal splenorenal shunt in the
elective treatment of variceal hemorrhage: A randomized controlled trial. Hepatology
1987;7:4306.
9. Warren WD, Henderson JM, Millikan WJ, et al. Distal splenorenal shunt versus endoscopic
sclerotherapy for long-term management of variceal bleeding. Ann Surg 1986;203:45462.
10. Choudhuri G, Agrawal BK, Tantry BV, et al. Post-sclerotherapy esophageal ulcers: A
prospective analysis of their behaviour. Indian J Gastroenterol 1989;8:1921.
13. Goff JS, Reveille RM, Stiegmann GV. Endoscopic sclerotherapy versus endoscopic variceal
ligation: Esophageal symptoms, complications, and motility. Am J Gastroenterol
1988;83:12404.
14. Larson GM, Vandertoll DJ, Netscher DT, Polk HC. Esophageal motility: Effects of injection
sclerotherapy. Surgery 1984;96:7039.
15. Snady H, Korsten MA. Esophageal acid-clearance and motility after endoscopic sclerotherapy
of esophageal varices. Am J Gastroenterol 1986;81:41922.
16. Saks BJ, Kilby AE, Dietrich PA, et al. Pleural and mediastinal changes following endoscopic
injection sclerotherapy of esophageal varices. Radiology 1983;149:63942.
17. Monroe P, Morrow CF Jr, Millen JE, et al. Acute respiratory failure after sodium morrhuate
esophageal sclerotherapy. Gastroenterology 1983;85:6939.
18. Kitano S, Iso Y, Yamaga H, et al. Temporary deterioration of pulmonary functions after
injection sclerotherapy for cirrhotic patients with esophageal varices. Eur Surg Res
1988;20:298303.
19. Lorgat F, Madden MV, Kew G, et al. Bacteremia after injection of esophageal varices. Surg
Endosc 1990;4:189.
20. Brayko CM, Kozarek RA, Sanowski RA, Testa AW. Bacteremia during esophageal variceal
sclerotherapy: Its cause and prevention. Gastrointest Endosc 1985;31:102.
21. Camara DS, Gruber M, Barde CJ, et al. Transient bacteremia following endoscopic injection
sclerotherapy of esophageal varices. Arch Intern Med 1983;143:13502.
22. Tank LV, Estay R, Ovalle L, et al. Bacterial peritonitis and sclerotherapy for variceal
hemorrhage. Gastrointest Endosc 1992;38:200.
23. Ashida H, Kotoura Y, Nishioka A, et al. Portal and mesenteric venous thrombosis as a
complication of endoscopic sclerotherapy. Am J Gastroenterol 1989;84:30610.
24. Seidman E, Weber AM, Morin CL, et al. Spinal cord paralysis following sclerotherapy for
25. Thatcher BS, Sivak MV Jr, Ferguson DR, Petras RE. Mesenteric venous thrombosis as a
possible complication of endoscopic sclerotherapy: A report of two cases. Am J Gastroenterol
1986;81:1269.
26.
Blaisdell PC. Office ligation of internal hemorrhoids. Am J Surg 1958;96:4014.
27.
Barron J. Office ligation of internal hemorrhoids. Am J Surg 1963;105:5637.
28. Russell TR, Donohue JH. Hemorrhoidal banding: A warning. Dis Colon Rectum
1985;28:2913.
29. Sim AJ, Murie JA, MacKenzie I. Three-year follow-up study on the treatment of first and
second degree hemorrhoids by sclerosant injection or rubber band ligation. Surg Gynecol
Obstet 1983;157:5346.
30. Greca F, Hares MM, Nevah E, et al. A randomized trial to compare rubber band ligation with
phenol injection for treatment of hemorrhoids. Br J Surg 1981;68:2502.
31. Gartell PC, Sheridan RJ, McGinn FP. Out-patient treatment of haemorrhoids: A randomized
clinical trial to compare rubber band ligation with phenol injection. Br J Surg 1985;72:4789.
32. Jensen DM, Machicado GA, Tapia JL. A reproducible canine model of esophageal varices.
33. Stiegmann GV, Sun JH, Hammond WS. Results of experimental endoscopic esophageal
varix ligation. Am Surg 1988;54:1058.
34.
Jensen DM, Weisz N, Hirabayashi K, et al. Randomized controlled study of rubber band
ligation and sclerotherapy for active esophageal varix bleeding in dogs [Abstract].
Gastroenterology 1991;100:A92.
35.
Jutabha R, Jensen D, Egan J, et al. Randomized prospective study of cyanoacrylate,
sclerotherapy, or rubber band ligation for endoscopic hemostasis of bleeding canine gastric
varices [Abstract]. Gastrointest Endosc 1992;38:235.
36.
Egan J, Jensen DM, Hirabayashi K, et al. Randomized controlled study of endoscopic rubber
band ligation and high-volume sclerotherapy for actively bleeding gastric varices in dogs
[Abstract]. Gastrointest Endosc 1991;37:239.
37. Young MF, Sanowski RA, Rasche R. Comparison and characterization of ulcerations induced
by endoscopic ligation of esophageal varices versus endoscopic sclerotherapy. Gastrointest
Endosc 1993;39:11922.
38.
Goff JS. Endoscopic variceal ligation. Can J Gastroenterol 1990;4:63942.
39. Hall RJ, Lilly JR, Stiegmann GV. Endoscopic esophageal varix ligation: Technique and
preliminary results in children. J Pediatr Surg 1988;23:12223.
40. Saeed ZA, Michaletz PA, Winchester CB, et al. Endoscopic variceal ligation in patients who
have failed endoscopic sclerotherapy. Gastrointest Endosc 1990;36:5724.
41.
Yamamoto M, Suzuki H, Aoki T, et al. Endoscopic variceal ligation (EVL). Digest Endosc
1990;2:26975.
42. Stiegmann GV, Goff JS, Michaletz-Onody PA, et al. Endoscopic sclerotherapy as compared
with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992;326:152732.
43. Laine L, el-Newihi HM, Migikovsky B, et al. Endoscopic ligation compared with sclerotherapy
for the treatment of bleeding esophageal varices. Ann Intern Med 1993;119:17.
44. Gimson AES, Ramage JK, Panos MZ, et al. Randomised trial of variceal banding ligation
versus injection sclerotherapy for bleeding oesophageal varices. Lancet 1993;342:3914.
45. Hou MC, Lin HC, Kuo BI, et al. Comparison of endoscopic variceal injection sclerotherapy and
ligation for the treatment of esophageal variceal hemorrhage: A prospective randomized trial.
Hepatology 1995;21:151722.
46. Lo GH, Lai KH, Cheng JS, et al. TI: A prospective, randomized trial of sclerotherapy versus
ligation in the management of bleeding esophageal varices. Hepatology 1995;22:46671.
47. Sarin SK, Govil A, Jain AK, et al. Prospective randomized trial of endoscopic sclerotherapy
versus variceal band ligation for esophageal varices: Influence on gastropathy, gastric varices
and variceal recurrence. J Hepatol 1997;26:82632.
48. Avgerinos A, Armonis A, Manolakopoulos S, et al. Endoscopic sclerotherapy versus variceal
ligation in the long-term management of patients with cirrhosis after variceal bleeding. A
prospective randomized study. J Hepatol 1997;26:103441.
49. Baroncini D, Milandri GL, Borioni D, et al. A prospective randomized trial of sclerotherapy
versus ligation in the elective treatment of bleeding esophageal varices. Endoscopy
1997;29:23540.
50. Laine L, Cook D. Endoscopic ligation compared with sclerotherapy for treatment of
esophageal variceal bleeding. A meta-analysis. Ann Intern Med 1995;123:2807.
51. Jalan R, Forrest EH, Stanley AJ, et al. A randomized trial comparing transjugular intrahepatic
portosystemic stent-shunt with variceal band ligation in the prevention of rebleeding from
52. Berner JS, Gaing AA, Sharma R, et al. Sequelae after esophageal variceal ligation and
sclerotherapy: A prospective randomized study. Am J Gastroenterol 1994;89:8528.
53.
Saltzman JR, Arora S. Complications of esophageal variceal band ligation. Gastrointest
Endosc 1993;339:1856.
54. Berkelhammer C, Madhav G, Lyon S, Roberts J. "Pinch" injury during overtube placement in
upper endoscopy. Gastrointest Endosc 1993;39:1868.
55. Goldschmiedt M, Haber G, Kandel G, et al. A safety maneuver for placing overtubes during
endoscopic variceal ligation. Gastrointest Endosc 1992;38:399400.
56.
Reveille RM, Goff JS, Stiegmann GV, et al. Combination endoscopic variceal ligation (EVL)
and low-volume sclerotherapy (ES) for bleeding esophageal varices: A faster route to variceal
eradication? [Abstract]. Gastrointest Endosc 1991;37:243.
57.
Koutsomanis D. Endoscopic variceal ligation combined with low-volume sclerotherapy: A
controlled study [Abstract]. Gastroenterology 1992;102:A835.
58. Hashizume M, Ohta M, Ueno K, et al. Endoscopic ligation of esophageal varices compared
with injection sclerotherapy: A prospective randomized trial. Gastrointest Endosc
1993;39:1236..
Chapter 37 Indications, Contraindications, and Complications of

Upper Gastrointestinal Endoscopy
(2526)
(2527)
GREGORY S. COOPER, M.D.
EDMOND W. BLADES, M.D.
Esophagogastroduodenoscopy (EGD), or upper endoscopy, is one of the most commonly performed

procedures in the world. Estimates of the numbers of procedures performed in the United
States1(2528) and the United Kingdom2(2529) are 1.2 examinations per 100 persons age 65 years
and older and 650 to 1000 procedures per 100,000, respectively. In the United States, the cost of
Medicare benefits from 1986 to 1989 increased more in gastroenterology than in any other medical
specialty,3(2530) and this increase is largely due to the expanding volume of endoscopic procedures.
Therefore, the appropriate indications for EGD as well as other invasive procedures have come under
intense scrutiny.4(2531) Although guidelines on appropriate indications for EGD have been
published,57(2532) they are based largely on consensus, and few controlled trials of the effect of
endoscopy on patient outcomes have been performed. Moreover, there is some disparity between
guidelines offered by the American Society for Gastrointestinal Endoscopy (ASGE) and those
presented by other organizations.8(2533) This chapter critically examines the indications and
contraindi-cations for upper endoscopy, using published ASGE guidelines as a framework (Tables
371, 372, 373).5(2534) Complications of the procedure, which can occasionally be life-threatening,
are also discussed in detail. Although all indications are mentioned, conditions that are the subject of
other chapters (e.g., upper gastrointestinal bleeding and foreign body removal) are not discussed at
length.
Indications for Diagnostic

Upper Endoscopy
TABLE 371
Upper abdominal distress that persists despite an

appropriate trial of therapy
Upper abdominal distress associated with symptoms or
signs that suggest serious organic disease
Dysphagia or odynophagia
Esophageal reflux symptoms that are persistent or
recurrent despite appropriate therapy
Persistent nausea and vomiting of unknown cause
Surveillance for malignancy
Gastric or esophageal ulcers
Familial adenomatous polyposis
Adenomatous gastric polyps
Barrett's esophagus
Occult gastrointestinal bleeding
Small bowel biopsy
Cirrhotic patients in whom prophylactic therapy is
considered
After caustic ingestion to assess for acute injury
Adapted from American Society for Gastrointestinal
Endoscopy. Appropriate Use of Gastrointestinal
Endoscopy. Manchester, MA: ASGE, 1992; 68.
Indications for Therapeutic

Upper Endoscopy
TABLE 372
Treatment of bleeding lesions

Sclerotherapy or banding of varices
Removal of foreign bodies
Removal of selected polypoid lesions
Placement of feeding or drainage tubes
Dilation of stenotic lesions
Indications for Therapeutic

Upper Endoscopy
TABLE 372
Palliative treatment of stenosing neoplasms

TABLE 373
Contraindications for Upper
Endoscopy
The risks to patient health or life are judged to outweigh
the most favorable benefits of the procedure
Adequate patient cooperation cannot be obtained
A perforated viscus is known or suspected
There are almost no data concerning the results of EGD in asymptomatic individuals. The study of
Akdamar et al.9(2535) is therefore of particular interest. EGD in 355 healthy men without
gastrointestinal symptoms, age 18 to 45 years, revealed an abnormality in 38%. Erosive changes were
found most commonly in the stomach, followed by the duodenum and esophagus. However, the point
prevalence for both gastric and duodenal ulcers was 2%.
Indications
Dyspepsia
Dyspepsia, broadly defined as upper abdominal pain or discomfort, is estimated to have a prevalence
of 25 to 30% in surveys from both the United States and the United Kingdom.10,11(2536) A working
party has developed a classification scheme for dyspepsia that includes ulcer-like, reflux-like,
dysmotility-like, and irritable bowel syndrome-associated categories.6(2537) Other important causes of
dyspepsia include gastric carcinoma and pancreaticobiliary diseases.
Dyspepsia constitutes one of the most frequent indications for upper endoscopy. However, the
procedure would be used optimally if it added information to that available by history and physical
examination and routine laboratory studies such that patient outcome would be affected.11(2538)
Examples of such information include a diagnosis of gastric carcinoma or complicated
gastroesophageal reflux disease (GERD) (e.g., ulcerations, Barrett's esophagus), in which aggressive
medical or surgical therapy would be a consideration, and possibly peptic ulcer disease. The
importance of a negative study has also been emphasized as reassuring patient and physician that a
serious disorder is not present.12,13(2539)
Unfortunately, symptoms probably do not reliably discriminate organic from functional disease. In a
study of outpatients referred for endoscopy, Talley et al.14(2540) found that categorizing patients
according to types of dyspeptic symptoms (e.g., ulcer-like, reflux-like) had little clinical value. Mansi et
al.15(2541) classified 2253 consecutive dyspeptic patients referred for EGD as having ulcer-like (43%
of patients), reflux-like (38%), or dysmotility-like symptoms (19%). There were no significant findings at
EGD in almost 33% of patients with dysmotility-like symptoms, whereas this occurred in about 25% of
patients in the other two categoriesa statistically significant difference. In the dysmotility group, no
malignant lesions were discovered in patients less than 60 years of age. In patients less than 50 years
old with dysmotility-like symptoms, there were no gastric ulcers, and duodenal ulcer and esophagitis
were encountered only rarely. There was no relation between age and a positive finding at EGD in
patients with ulcer-like or reflux-like symptoms. Although these findings suggest that EGD is usually not
indicated in patients less than 50 years of age with dysmotility-like symptoms, Mansi et al.15(2542)
suggested that these parameters pertain to only a small fraction (7.15%) of their patients with
dyspepsia.
Gastric carcinoma, the most ominous finding in the dyspeptic population, is a relatively uncommon
cancer in the United States, with less than 25,000 new cases reported annually.16(2543) The disease
is also uncommon among dyspeptic patients; the incidence was 2% in a cumulative analysis of four
series from the United Kingdom.17(2544) Moreover, it is probable that most patients with gastric
carcinoma would undergo evaluation because of other findings (e.g., lack of response to empirical
therapy). Because the incidence of gastric cancer increases with age, endoscopic studies of older
patients would be expected to detect a larger proportion of cases. In an endoscopic survey of dyspeptic
patients 70 years of age and older, malignancy was found in 33%,18(2545) whereas a study of patients
aged 25 years and under found no cases of malignancy.19(2546) Williams et al.20(2547) studied 1386
patients with dyspepsia who were referred for EGD or double-contrast barium radiography. Abnormal
findings were encountered at EGD in 40% of patients under the age of 45 years versus 58% of patients
of greater age. Malignancies were not diagnosed in any patient less than 45 years of age. Of 707
patients with malignant disorders diagnosed at endoscopy, Williams et al.20(2548) found that only 13
(1.8%) were under 45 years of age and all had further symptoms in addition to simple dyspepsia.
Proponents of early endoscopy in the elderly patient with dyspepsia argue that EGD should be
performed more readily in this group, particularly if symptoms are of recent onset and not associated
with the use of ulcerogenic drugs.6,12(2549) Others contend that an earlier diagnosis would not
change the outcome for patients with a disease that is metastatic in more than 70% of cases16(2550)
and that is poorly responsive to chemotherapy.11(2551) However, in a large series of patients
undergoing screening endoscopy for dyspepsia, 15 of 57 malignancies found were considered early
gastric cancer, a stage of the disease that has a much more favorable prognosis.21(2552)
Peptic ulcer disease is diagnosed in approximately 20 to 30% of dyspeptic patients undergoing
endoscopy.11,17(2553) However, the effect of establishing the diagnosis on subsequent patient
management is unclear. Approximately 90% of patients with duodenal ulcers and 80% of those with
gastric ulcers would be expected to heal with a 6-week course of an H2-receptor antagonist, the usual
choice for empirical management of dyspeptic patients. However, in selected cases the diagnosis
could lead to avoidance of medications such as nonsteroidal anti-inflammatory drugs (NSAIDs) or
anticoagulants, and a diagnosis of a large ulcer could influence duration or type of therapy.
The strong association of Helicobacter pylori and peptic ulcer disease may result in additional
indications for upper endoscopy in the dyspeptic patient. The presence of active infection is most
readily demonstrated by histology or the urease broth (Campylobacter-like organism [CLO]) test, both
of which require endoscopy with biopsy.22(2554) Most authorities agree that follow-up endoscopy to
document eradication is not routinely needed. Because numerous studies have shown a dramatically
lower ulcer recurrence rate23,24(2555) in patients in whom H. pylori is eradicated, it could be argued
that all dyspeptic patients should undergo endoscopy to establish both the diagnosis of peptic ulcer
disease and the presence of H. pylori; this would result in proper selection of patients for treatment with
antimicrobial agents.
Gastritis or duodenitis, or both, are noted in approximately 20% of endoscopies in dyspeptic
patients,17(2556) but these are often incidental findings and are not believed to correlate with
symptoms.25(2557) However, many of these patients improve both histologically and symptomatically
after empirical therapy for peptic ulcer disease.26(2558)
In the early 1980s, the American College of Physicians Health and Public Policy Committee
commissioned a study of the role of endoscopy in the management of dyspepsia.7(2559) The
committee recommended, largely based on review of the existing body of data and information, that a
course of therapy with antiulcer agents be given empirically to all dyspeptic patients and that all
potentially offending agents such as tobacco, alcohol, and NSAIDs be withdrawn. Exceptions included
patients with systemic signs such as weight loss and those in whom complications such as bleeding or
obstruction develop. It was further recommended that patients with either minimal or no response to
therapy after 7 to 10 days or those in whom symptoms persist or do not resolve after 6 to 8 weeks
should undergo evaluation. These patients were thought to be at higher risk than the overall dyspeptic
population for serious diseases.
There have been attempts to develop algorithms based on symptoms for the evaluation of dyspeptic
patients.13,27(2560) Because variables for patient selection for endoscopy included parameters such
as "pain provoked by berries"27(2561) and "endoscopy performed by Dr. A,"13(2562) and because the
indices were validated neither prospectively nor by other groups of investigators, no conclusions can
be drawn regarding the general use of these algorithms. A study from the United Kingdom attempted
to select dyspeptic patients for endoscopy based on one or more of the following criteria: age 45 years
or older, use of NSAIDs, and a positive serology for H. pylori.28(2563) This model was validated
prospectively and detected 228 of 234 patients with peptic ulcer disease and all 5 patients with
malignancy. Although Sobala et al.28(2564) stated that this strategy would have reduced endoscopies
by approximately 25%, the specificity of the model, or the proportion of patients with insignificant
findings who underwent endoscopy, was not mentioned.
There is a definite need for prospective studies that focus on the selection of dyspeptic patients who
will benefit from endoscopy as well as on the optimal timing of the procedure in relation to the clinical
course of these patients. In addition to reducing unnecessary procedures, such studies could also
define much needed parameters for the identification of patients with serious diseases and thereby
provide an opportunity to affect outcome through early diagnosis and appropriate treatment.
Dysphagia and Odynophagia

Dysphagia, or difficulty swallowing, is a major indication for upper endoscopy because it rarely
represents a functional complaint. In the evaluation of patients, the dysphagia should first be
characterized as most likely due to a preesophageal (oropharyngeal) or an esophageal cause.29(2565)
In the former case, the initial evaluation generally consists of careful examination of the mouth and
pharynx and a modified barium swallow. Among esophageal causes, a presumptive diagnosis of
neuromuscular dysphagia would be best evaluated initially with barium swallow and manometry,
whereas endoscopy would be the best initial test if the supposed cause is anatomic (e.g., malignancy,
peptic stricture).
It is controversial whether barium contrast x-rays should be obtained before endoscopy in patients with
presumed anatomic causes of dysphagia. Proponents claim that radiography eliminates the possibility
of lesions, such as Zenker's diverticulum, that represent an increased risk of perforation during
passage of an endoscope. Others contend that with direct intubation, the risk of a complication is
significantly lower than with blind intubation.12(2566) A barium study also provides information about
the location, length, and luminal caliber of an obstructing lesion that could influence the choice of
endoscope and accessories. Moreover, esophageal strictures of diameter larger than that of the
endoscope may escape detection at endoscopy. In one series, the sensitivity of EGD in this situation
was only 88% compared with 100% for radiography;30(2567) however, barium studies were
considerably less specific. Similarly, Schatzki's rings, the most common cause of episodic dysphagia in
adults, can be missed on EGD if the diameter of the ring is larger than that of the instrument and full
distention of the distal esophageal lumen is not achieved during the procedure. In one report, the
sensitivity of EGD for detecting rings was only 58%, compared with 95% for prone, full-column barium
studies.31(2568)
Odynophagia, or painful swallowing, is usually due to an inflammatory process involving the
oropharynx or esophagus. A thorough examination of the pharynx should be performed first, followed
by endoscopy to evaluate the esophageal mucosa.
In the immunocompromised host, dysphagia and odynophagia are usually due to infectious causes,
although prior radiation and chemotherapy can induce symptomatic esophageal injury. Neutropenic
patients with esophageal symptoms should generally undergo endoscopy promptly and biopsies should
be obtained for fungal, viral, and bacterial cultures. In a retrospective series of bone marrow transplant
patients with esophageal symptoms, approximately half had esophageal infec-tions
documented.32(2569) Furthermore, the causative agent was not predicted by clinical and radiographic
criteria.
In the human immunodeficiency virus (HIV)-infected host, esophageal symptoms are usually due to
either Candida esophagitis or viral infection, including cytomegalovirus, herpes simplex, and primary
HIV infection. Although radiography with barium contrast is sometimes performed in the initial
evaluation of symptoms, this study lacks specificity and does not allow concurrent histologic
evaluation.33(2570) It is generally accepted that in the presence of oral thrush and esophageal
symptoms, an HIV-infected patient should first be empirically treated with antifungal agents and not
undergo endoscopy unless symptoms fail to respond to therapy.33(2571) Using this strategy, the
positive predictive value for Candida esophagitis of oral thrush and symptoms approaches
100%;34(2572) other causes are diagnosed in nonresponders in approximately 45% of
cases.35(2573)
Gastroesophageal Reflux Disease (GERD)

GERD is an extremely prevalent condition; up to 44% of the population has symptoms of the disorder
at least monthly.36(2574) The vast majority of individuals can manage symptoms of GERD by lifestyle
modifications plus antacids as needed without physician intervention. Similarly, patients whose
symptoms are well controlled by a standard dosage of an H2-receptor antagonist and lifestyle
modifications probably do not warrant further investigation, as it is unlikely that the results would alter
management. Moreover, the specificity of upper endoscopy for the diagnosis of GERD approaches
100%, but its overall sensitivity is only 60 to 70%37(2575) and is considerably lower in patients with
milder symptoms. Sensitivity improves with the addition of endoscopic biopsies. If suction biopsies are
obtained via a Rubin tube, the diagnostic yield improves even further; however, this method is
associated with a higher incidence of complications.38(2576) Thus, a negative endoscopy, even with
concurrent biopsies, does not eliminate the diagnosis of GERD, and other modalities, such as 24-hr pH
monitoring, are needed for precise diagnosis.
The subset of patients in whom symptoms persist despite therapy with first-line agents constitutes a
group in which endoscopy is more likely to contribute to management. One benefit is accurate staging
of the severity of disease at baseline, so that the type and duration, whether short-term or long-term, of
therapy can be modified accordingly. These patients are more likely to have erosions or ulcerations at
endoscopy and will thus require more aggressive therapy. Conversely, 60 to 70% of patients with
erosive or ulcerative esophagitis have symptoms that interfere with usual activity or are even
incapacitating.39(2577) Because patients with more severe endoscopic findings are also more likely to
develop recurrent disease off therapy,39(2578) the initial endoscopic staging may also influence the
decision for and choice of maintenance treatment.
Another benefit of diagnostic endoscopy in the patient with persistent symptoms of GERD is the
detection of complications of esophagitis. Barrett's esophagus, which usually develops in the setting of
chronic and severe GERD, is found in up to 15 to 20% of patients referred for endoscopy because of
reflux symptoms. Biopsies can be obtained to determine whether dysplasia is present, and the
management of the patient can be altered by incorporating an appropriate surveillance program based
on histologic findings. Radiographically documented esophageal ulcers in a patient with GERD should
in general be evaluated endoscopically and biopsies should be obtained to rule out concurrent
malignancy. Furthermore, an esophageal ulcer in the setting of severe or persistent symptoms of
GERD is frequently a manifestation of Barrett's epithelium. Peptic strictures are found in 10 to 20% of
patients who undergo endoscopy for reflux symptoms.37(2579) However, most such patients would
also have dysphagia to a variable degree and would therefore undergo upper endoscopy for this
reason. EGD may also detect erosive or ulcerative esophagitis in patients with occult gastrointestinal
bleeding.
Although authorities recommend endoscopy in patients who do not respond to first-line therapy for
GERD,5(2580) no controlled trials have been performed to compare various possible management
strategies. For example, patients in whom H2-receptor antagonist therapy fails could be empirically
treated with a more potent agent such as omeprazole without undergoing endoscopy. The cost of the
procedure would then have to be balanced against the cost of prescribing the drug for an unknown
duration of time and the potential ramifications of uncertainty with regard to the severity of disease and
the presence or absence of a complication of GERD. Data from controlled studies of various strategies
would be useful in planning the most effective and cost-saving approach.
The evaluation of noncardiac chest pain is not included in the ASGE recommended indications for
upper endoscopy.5(2581) Presumably, this is because the yield of the procedure for diagnosing motility
disturbances is low compared with those of barium studies and manometry. However, a significant
number of patients with diffuse esophageal spasm (nutcracker esophagus), the most common
manometric cause of esophageal chest pain, may also suffer from concurrent GERD and improve with
H2-receptor antagonist or proton pump inhibitor therapy.40(2582) In a study of 100 consecutive
patients with noncardiac chest pain, endoscopic abnormalities included grade II or higher esophagitis
in 24, gastritis or duodenitis in 18, and peptic ulcer in 6;41(2583) 30 of 54 patients with a normal
esophageal manometry had endoscopic abnormalities. Thus, in the population with noncardiac chest
pain, there may be one subset with predominant reflux-related symptoms, and another with truly
isolated chest pain, with endoscopy having a major role only in the former group.42(2584)
Persistent Nausea and Vomiting

There are multiple causes for persistent nausea and vomiting.43(2585) Gut motility disorders,
obstructing lesions (e.g., pyloric channel ulcer, gastric carcinoma), and nonobstructive mucosal lesions
(e.g., gastritis, peptic ulcer) account for most of the gastrointestinal causes. However, many
nongastrointestinal conditions including neurologic and metabolic disorders should also be considered.
If a gastrointestinal disorder is suspected, either upper endoscopy or barium contrast radiography
should be performed first to search for a mucosal lesion. If the result of this study is negative, further
evaluation is indicated for motility disorders using gastric emptying studies or other techniques (e.g.,
antral-duodenal motility).
Compared with barium contrast radiography, endoscopy may offer better visualization of the upper
gastrointestinal tract in patients with persistent nausea and vomiting, particularly if retained food or a
bezoar is present. If indicated, biopsies can also be obtained and therapeutic maneuvers can be
performed such as hydrostatic balloon dilation of strictures and endoscopic disruption of a
bezoar.44(2586) However, the diagnostic yield and therapeutic benefit of EGD in this population are
unknown. Despite these benefits, endoscopic evaluation is usually suboptimal and incomplete in the
patient with obstruction of the upper gastrointestinal tract. Ideally, therefore, patients with known or
strongly suspected gastric outlet obstruction should first undergo lavage with a large-bore tube to
remove large food particles, followed by 1 to 2 days of nasogastric decompression before evaluation.
EGD in 37 consecutive patients with bulimia nervosa and chronic, self-induced vomiting revealed only
minor abnormalities in 23 patients.45(2587)
Surveillance for Malignancy

Gastric Ulcers
With the possible exception of patients less than 40 years of age, in whom the prevalence of gastric
cancer is extremely low, patients with gastric ulcers should generally undergo endoscopy with biopsy to
rule out malignancy. Although radiographic criteria have been developed to differentiate benign from
malignant ulcers, the combination of endoscopic assessment and biopsy achieves a diagnostic
accuracy of at least 95%.4648(2588) Diagnosis based on the endoscopic findings alone has an
accuracy closer to 85%.47,48(2589) Four-quadrant directed biopsies are therefore necessary as
adjuncts to endoscopic diagnosis. Endoscopic confirmation of a radiographically demonstrated
duodenal or pyloric channel ulcer crater is indicated only if symptoms fail to respond to appropriate
therapy.5(2590)
Although many gastroenterologists routinely perform endoscopy to document healing of gastric ulcers,
the utility of this practice has been questioned. Several studies have examined the incidence of
unsuspected carcinoma at follow-up endoscopy in patients with benign-appearing, biopsy-negative
gastric ulcers; this has been less than 2% in all reports.4649(2591) The majority of patients found to
have carcinoma at a follow-up endoscopy have also had persistent symptoms and would have
undergone repeat evaluation for this reason alone. Moreover, there is little evidence that delay in
diagnosis influences the natural history of gastric carcinoma. Nevertheless, a follow-up endoscopy is
prudent if there was any question at initial evaluation that a lesion might be malignant, even if the initial
biopsies did not disclose carcinoma. It is also advisable to perform a follow-up procedure if biopsies
were not obtained at the time of diagnosis.
Familial Adenomatous Polyposis
The association of familial adenomatous polyposis (FAP) and colorectal neoplasia is well known, but
many studies have also documented a strong association with upper gastrointestinal
neoplasms.5054(2592) The prevalence of duodenal adenomas in endoscopic series ranges from 25
to 92%.5054(2593) These adenomas arise primarily in the periampullary area where the endoscopic
appearance may be normal despite their presence.50,51(2594) The risk of duodenal neoplasms
appears to be higher in those patients with the Gardner syndrome variant of FAP.53(2595)
Associations with gastric adenomas and fundic gland polyps have also been documented.50,53(2596)
In one study, the relative risk of duodenal and ampullary cancer in patients with FAP was increased
more than 330-fold and 120-fold, respectively, although the risk of gastric and nonduodenal small
bowel carcinoma was similar to that of controls.52(2597)
Based on available data, periodic upper endoscopic surveillance has been recommended for patients
with FAP including those with the Gardner variant.55(2598) In one report, however, no histologic
change was found in adenomas over an average follow-up period of 7.7 years.51(2599) Screening
intervals and the age at which surveillance should be initiated have not been well evaluated, however,
and no study has documented a reduction in mortality as a result of screening. Endoscopy should be
performed with a side-viewing as well as a forward-viewing instrument, so that the periampullary region
is optimally visualized.
Gastric Polyps
Gastric polyps are uncommon. Histologically, most gastric polyps are hyperplastic types that have an
extremely low malignant potential. In contrast, adenomatous polyps have a significantly higher
malignant potential that correlates with size; up to 60% of lesions greater than 2 cm in diameter contain
in situ or invasive cancer. Although differentiation of adenomas from hyperplastic polyps is unreliable
based on histologic assessment of forceps biopsy specimens, gastric polyps can be removed safely at
endoscopy and histopathologic classification of polypectomy specimens is accurate.56(2600) It is
recommended that patients with adenomatous polyps should undergo endoscopic surveillance at
intervals of 2 to 4 years, with removal of the largest lesions.55(2601) There have been no studies of
the long-term efficacy of these measures. Surveillance is not indicated for hyperplastic polyps.
Barrett's Esophagus
Barrett's esophagus is a premalignant condition associated with an estimated 30-fold to 40-fold
increased risk of esophageal adenocarcinoma.57(2602) Because dysplasia is believed to be the
precursor of invasive malignancy, endoscopic surveillance programs have been proposed to detect
dysplastic tissue in random biopsies or biopsies obtained in prescribed patterns. Because Barrett's
epithelium is found in up to 15 to 20% of patients undergoing endoscopy for GERD,57(2603) the cost
of such screening programs would be substantial. Although it is recommended that patients with
Barrett's esophagus undergo periodic endoscopic screening,55(2604) there are no data to indicate that
screening improves long-term survival. There is also no consensus on intervals between screening
examinations, but a consensus conference recommended alternate-year screening for patients with no
dysplasia, intensive medical therapy followed by close surveillance in those with low-grade dysplasia,
and consideration of esophagectomy in those with high-grade dysplasia.58(2605)
Occult Gastrointestinal Bleeding

Although patients with occult gastrointestinal bleeding, as manifested by guaiac-positive stools or iron
deficiency, are generally evaluated initially with colonoscopy or flexible sigmoidoscopy plus a barium
enema, the findings are either normal or unexplanatory in a significant number of cases. The role of
EGD in the evaluation of these patients is controversial (see also Chapter 27: Lower Gastrointestinal
Bleeding).
Several studies have examined the role of EGD in patients with guaiac-positive stools and negative
lower gastrointestinal evaluations;5964(2606) abnormalities have been documented in up to 79% of
cases.59(2607) However, the frequency of gastric cancer in these combined series was only 0.3%. In
the presence of iron deficiency anemia and negative colonic evaluations, the prevalence of gastric
cancer increases to as much as 6%;62,63(2608) presumably, this constitutes a higher-risk population.
A study of patients with iron deficiency anemia by Rockey and Cello65(2609) found no abnormalities at
colonoscopy in patients who had symptoms referable to the upper gastrointestinal tract. These
investigators recommended upper endoscopy as the initial diagnostic test in patients with appropriate
symptoms and that colonoscopy not be performed when there are abnormal findings at EGD.
However, these findings must be corroborated before these recommendations can be accepted for
widespread implementation. In a retrospective review of the records of 170 patients over the age of 50
years with iron deficiency anemia, Gordon et al.66(2610) found that colonoscopy had revealed a
source of bleeding in 30 patients (18%), whereas EGD identified a source in 70 patients (41%); a
cause was not identified in 70 patients (41%). Gastrointestinal symptoms and a history of NSAID use
or alcohol ingestion did not reliably predict diagnostic findings. Furthermore, many patients with upper
gastrointestinal sources of blood loss were asymptomatic. Gordon et al.66(2611) concluded that EGD
(including small bowel biopsies) should be performed in older patients, especially when a source of
bleeding is not evident at colonoscopy. In the prospective study of Zuckerman and Benitez,60(2612)
upper endoscopy and colonoscopy were performed in 100 patients with guaiac-positive stools or iron
deficiency anemia. In patients with both anemia and guaiac-positive stools, the diagnostic yield was
significantly higher for EGD (45% vs. 26%). Therapy was altered in response to the EGD findings in 30
patients. Zuckerman and Benitez60(2613) concluded that EGD is not indicated in asymptomatic
patients with a positive fecal occult blood test (FOBT) and a negative colonoscopy.
Hsia and al-Kawas64(2614) recommended that EGD be obtained in all asymptomatic patients with
fecal occult blood and a negative evaluation for a colonic source. Significant and previously
undiagnosed sources of bleeding were found in 19 of 70 (27%) such patients in this consecutive
series.64(2615) However, 13 patients were anemic and in 38% of this group a significant finding was
discovered at EGD. Furthermore, 15 patients in this study were taking NSAIDs. Erosions or peptic
ulcer disease made up about half of the positive endoscopic diagnoses. Findings that could be
considered unquestionably significant but clinically silent, as for example vascular ectasias, varices,
and Barrett's esophagus, were relatively rare. EGD disclosed a positive finding in 43% of 117
asymptomatic patients with a positive FOBT and a negative colonoscopy in the retrospective study of
Chen et al.61(2616) However, the EGD findings were not significant and did not result in a change in
therapy in the majority of cases. Thomas and Hardcastle67(2617) studied 283 patients with a positive
FOBT and a negative colonoscopy. EGD revealed a significant abnormality in 6 of 14 patients with
symptoms referable to the upper gastrointestinal tract, including 1 gastric carcinoma. The remaining
269 asymptomatic patients were followed for a median of 5 years (range, 2 to 8 years). During
follow-up, various upper gastrointestinal disorders were diagnosed, but no carcinomas.
Endoscopic evaluation of the small intestine (enteroscopy), intraoperatively in some cases, may be
indicated in patients with unexplained or persistent iron deficiency anemia. Enteroscopy is discussed in
Chapter 50: Endoscopy of the Small Intestine.
Small Bowel Biopsy

Procurement of biopsies from the small intestine in the evaluation of malabsorptive or other conditions
is another indication for EGD. Although the traditional approach has been to obtain tissue via
fluoroscopically positioned suction instruments, such as the Rubin tube and the Crosby capsule, most
biopsies are currently obtained endoscopically. This has the advantage that biopsy sites can be
targeted in the case of patchy lesions. Despite initial concerns over adequacy of specimens, both
uncontrolled68(2618) and controlled69(2619) trials have demonstrated that diagnostic yield is
equivalent to that for suction devices. The presence of Brunner's glands in biopsies from the proximal
duodenum could potentially interfere with interpretation, but this was a significant factor in only 3% of
specimens in one series.70(2620) Ladas et al.71(2621) studied the weight, length, and orientation of
biopsy specimens obtained with standard and large ("jumbo") forceps and oriented with and without the
aid of a stereomicroscope. Specimens obtained with the jumbo forceps were larger and longer and
more suitable for stereomicroscopic orientation. When oriented with the latter instrument, 72% of the
large specimens contained at least four villi in a row, whereas this was the case in only 44% of large
specimens oriented without the assistance of stereomicroscopy. Furthermore, less than 39% of
specimens obtained with the standard forceps contained four villi in a row.
Surveillance for Esophageal Varices

Esophageal variceal bleeding is a major cause of morbidity and mortality in patients with cirrhosis.
Studies that have attempted to estimate the likelihood of variceal hemorrhage in patients who have
never bled have found that severity of liver disease (Child class), size of the varices, and the presence
of red wale markings (longitudinally dilated venules) on the varices are all predictive.72(2622) EGD to
define endoscopic parameters may therefore be useful in identifying patients who have the highest
likelihood of bleeding and who would benefit the most from prophylactic therapy. Although prophylactic
sclerotherapy is probably not beneficial in most patients73(2623) and may be associated with an
increased mortality,74(2624) the use of -receptor blocking drugs is effective in preventing the first
episode of bleeding, particularly in patients with larger varices (see also Chapter 31: Variceal Bleeding
and Chapter 34: Prophylactic Sclerotherapy).73,75(2625) For patients with a clear contraindication to
therapy with blockers, the benefit of screening endoscopy would be substantially lower.
Corrosive Ingestion
There are 5000 to 10,000 cases of ingestion of a corrosive substance(s), predominantly alkali, in the
United States each year, primarily accidental ingestion in children and suicide attempts in
adults.76(2626) Immediate sequelae include perforation of the esophagus or stomach and upper
airway injury; late sequelae include esophageal stricture formation, gastric outlet obstruction, and
esophageal carcinoma. In more than 80% of cases in one series, endoscopy disclosed damage to both
the esophagus and the stomach.77(2627) Although concurrent injury to the mouth and pharynx often
occurs, 10 to 30% of patients with esophageal burns have no oropharyngeal damage and there is poor
correlation between the degrees of visceral and oropharyngeal injury.76(2628) Upper endoscopy
should be performed early in the patient's course after ingestion to identify those without injury who can
be discharged and those with evidence of severe injury who require close observation. It has also been
recommended that patients with documented severe injury at an initial endoscopy undergo another
EGD at 2 weeks after ingestion of the corrosive agent to assess healing and detect
complications.76(2629) In the prospective series of Gumaste and Dave,76(2630) the findings at initial
and follow-up endoscopy predicted the development of early complications and subsequent strictures.
There were no procedure-related complications. However, caution is always mandatory when
corrosive-induced injury is suspected. Corrosive injury is also discussed in Chapter 45: Special
Varieties of Esophagitis and Chapter 53: Esophagogastroduodenoscopy in the Pediatric Patient.
Therapeutic Applications
A detailed discussion of therapeutic applications of upper endoscopy is beyond the scope of this
chapter and the many indications are covered elsewhere in this text.
The benefit in terms of patient outcome for therapeutic endoscopy (see Table 372) is in general more
clearly defined than for diagnostic applications. For example, purely diagnostic EGD in patients with
upper gastrointestinal hemorrhage probably offers no benefit in terms of outcome.78(2631) However,
EGD identifies patients with lesions that confer an increased risk of persistent or recurrent bleeding.
The endoscopic modalities of electrocoagulation, injection, and laser therapy are all effective in
decreasing rates of recurrent bleeding, surgery, and mortality in the setting of high-risk lesions such as
visible vessels.79,80(2632) However, the application of the various methods of endoscopic hemostasis
probably does not affect outcome in patients with ulcers that have flat spots or adherent clot.80(2633)
Emergency sclerotherapy controls acute hemorrhage from esophageal varices in 70% of cases with
one session and 90% with two sessions and is more effective than either balloon tamponade or
vasopressin (see Chapter 33: Indications and Results of Variceal Injection Sclerotherapy).81(2634)
Endoscopic band ligation of varices, a newer technique, appears to be associated with fewer
complications and lower mortality when compared with sclerotherapy;82(2635) whether this technique
replaces sclerotherapy remains to be seen (see Chapter 36: Endoscopic Elastic Band Ligation for
Bleeding Esophageal Varices).
Endoscopy is also the preferred modality for removal of foreign bodies, with success rates of up to
98% and minimal or no complications (see Chapter 54: Foreign Body Extraction).83(2636) Other
therapeutic applications of endoscopy include dilation of benign esophageal strictures (see Chapter 44:
Hiatus Hernia and Peptic Diseases of the Esophagus),84(2637) placement of long nasogastric tubes in
the duodenum (decompression, enteral feeding),85(2638) injection of botulinum toxin for treatment of
achalasia (see Chapter 39: Esophageal Motility and Miscellaneous Disorders),86(2639) and palliation
of obstructing esophageal tumors with either laser therapy or placement of endoprostheses (see
Chapter 42: Palliation of Malignant Obstruction: Dilation and Peroral Prosthesis and Chapter 43:
Palliation of Malignant Obstruction: Lasers and Tumor Probes).87(2640)
Comparison with Radiography

The other major diagnostic imaging modality in addition to endoscopy for evaluation of the upper
gastrointestinal tract is barium radiography. Advantages of barium studies include lower cost, fewer
complications, ability to crudely assess motility and emptying, and better visualization of specific
lesions such as webs, rings, and diverticula. These are offset by the difficulties of performing the
examination in immobile, uncooperative patients and interpretation in patients with prior upper
gastrointestinal surgery or diseases.
In general, radiography is not indicated for acute gastrointestinal bleeding because barium in the
gastrointestinal tract interferes with other imaging studies (endoscopy, angiography), because
therapeutic capability is lacking, and because of the inability to recognize those stigmata of
hemorrhage that predict patient outcome. Early studies that demonstrated a greater accuracy of
endoscopy were criticized for biases in patient selection, retrospective analyses, failure to use
state-of-the-art methodology, and the lack of an independent standard for comparison. However, a
subsequent report, which prospectively evaluated double-contrast radiography and upper endoscopy,
with the final diagnosis made by a consensus panel, found that endoscopy was more sensitive (92%
vs. 54%) and more specific (100% vs. 91%) than barium contrast radiography.88(2641) Endoscopy
also had a greater effect on subsequent clinical management. Patient preference also appears to be
higher for endoscopy, with almost twice as many patients favoring endoscopy over radiography in one
study.89(2642) The initial higher cost of endoscopy may be partially balanced by lower subsequent
medical costs due to improved diagnostic accuracy, as shown in one randomized trial.90(2643)
Open-Access Endoscopy
There is a growing trend throughout the world toward the provision of endoscopic services at the
request of a referring physician without prior consultation with a specialist in digestive
diseases.9195(2644) This practice, known as primary diagnostic endoscopy, primary panendoscopy,
or open-access endoscopy, remains controversial.96(2645) Proponents claim that open-access
endoscopy is efficient and cost-effective, whereas opponents maintain that it results in performance of
procedures without appropriate indications, overutilization of endoscopy, and a potential for increased
morbidity owing to a failure to identify factors that increase the risk of the procedure. In some cases,
furthermore, it is asserted that prior consultation with a gastroenterologist might result in alternative
approaches to evaluation that are not only more appropriate and cost-effective but also likely to spare
the patient an unnecessary procedure or suggest a diagnosis that would not be discovered by
endoscopy.97(2646)
Assessment of the appropriateness of referrals for EGD under an open-access system is problematic
because there are differences of opinion as to valid and inappropriate indications for the procedure.
Quine et al.98(2647) prospectively audited 400 referrals for procedures in one health region of the
United Kingdom. The appropriateness of the referral was judged by a panel of seven
gastroenterologists and against a computer software program that utilized criteria proposed by the
Rand Corporation. The panel found 11% of the requests for endoscopy to be inappropriate compared
with 31% as determined by the software program. Quine et al.98(2648) also surveyed physician
opinion regarding appropriate indications. Physicians in a variety of types of practice, including general
practitioners and gastroenterologists, were asked whether endoscopy would be appropriate in several
hypothetical patient scenarios. Endoscopy was selected as an appropriate test in 4.5 to 63.8% of
scenarios in which it was considered inappropriate by the working party.
There appears to be a correlation between the frequency of significant endoscopic findings and the
indication for the procedure in studies of open-access endoscopy. Adang et al.99(2649) found
"relevant" findings at endoscopy in 42.2% of patients with clinical evidence of upper gastrointestinal
bleeding, 40.5% of patients with a history of peptic ulcer disease, and 31.9% of those with dysphagia
compared with 21% of patients in whom the indication was dyspepsia.

Despite the lack of agreement as to accepted indications, several studies have attempted to assess
the appropriateness of open-access referrals for EGD. By comparison with ASGE guidelines, Minoli et
al.94(2650) found that 32% of the requests for EGD from family physicians were inappropriate
compared with 17% of those from internists, 19% from surgeons, and 14% from gastroenterologists, a
significant difference. These results are in contrast to those of the prospective study of Mahajan et
al.,95(2651) which found that referrals for open-access EGD from family physicians and general
internists were more likely to be appropriate (97%) compared with referrals from subspecialists and
surgeons (81.3%). In a study of 2253 dyspeptic patients, Mansi et al.100(2652) found that the
frequencies of positive and negative endoscopic findings were comparable whether patients were
referred for EGD by hospital-based physicians or general practitioners.
There is relatively little information concerning the influence of open-access EGD on patient
management. Hungin et al.101(2653) compared the management of 715 patients during a 1-year
period before open-access referral for EGD with management during 1 year after endoscopy. Major
and minor abnormalities were found at EGD in 34% and 26%, respectively, of cases, whereas the
findings were within normal limits in 36%. Therapy with pharmacologic drugs was either stopped or
substantially altered as a result of the endoscopic findings in 39% of patients overall. But a major
change in pharmacologic management occurred in 60% of patients in whom no abnormality was
discovered at EGD. Conversely, a major finding at EGD was associated with a change to more
aggressive therapy in 58% of patients. Hungin et al.101(2654) also found fewer referrals for specialty
consultation among patients with normal EGD findings.
Contraindications
There are few absolute (see Table 373) and relative contraindications to upper endoscopy.5(2655)
However, before performance of every endoscopy, the perceived diagnostic or therapeutic benefits of
the procedure must be weighed against the risks.
Known or suspected perforation within the gastrointestinal tract is an absolute contraindication to all
endoscopic procedures.5(2656) Insufflation of air will enhance contamination of the peritoneal cavity
and may convert a walled-off perforation to a free perforation. A competent patient who refuses
consent for endoscopy should not be coerced into undergoing the procedure, although the risks of
nonendoscopic management must be presented clearly to the patient. Endoscopy is contraindicated in
patients with severe cardiopulmonary instability and when the risks to the patient outweigh the potential
benefits of the procedure.5(2657)
Relative contraindications to endoscopy include decompensated cardiopulmonary disease and the
presence of a large Zenker diverticulum. This latter abnormality increases the risk of perforation,
particularly if intubation under direct vision is encumbered by the presence of the diverticulum. For
patients with respiratory insufficiency, for whom the risk associated with conscious sedation is
increased, endotracheal intubation and mechanical ventilation may be necessary. Although recent
myocardial infarction is a relative contraindication, a retrospective study found that hemodynamically
stable patients with upper gastrointestinal bleeding can safely undergo endoscopy in the postinfarction
period.102(2658)
Patients with coagulopathies should ideally have their bleeding diathesis corrected before endoscopy,
especially if a therapeutic maneuver or the need for biopsies is anticipated. Whether it is necessary
that all parameters of coagulation be completely within normal limits has not been established. This
may be difficult or impossible, costly, and not without risk in certain patients. For a patient at the point
of exsanguination, delay in endoscopic intervention for the purpose of correcting a bleeding diathesis
may represent a greater hazard. It is also evident that EGD has been performed safely in many
patients who have mild coagulopathies (e.g., thrombocytopenia in association with cirrhosis).
Unfortunately, no studies or data are available that correlate the precise degree or nature of a
coagulopathy with the risk of a bleeding complication as a result of diagnostic or therapeutic EGD.
There are relatively few prospective or retrospective data concerning the risks and benefits of upper
endoscopy in pregnant women. In addition to the small chance of morbidity and mortality in the case of
the mother, there is also the possibility of injury to the fetus. This might occur as a result of mechanical
trauma (e.g., placental abruption), especially during the third trimester, or as a result of the
administration of drugs for conscious sedation. Available information suggests that the risks of the
procedure to both mother and fetus are relatively low. Cappell and Sidhom103(2659) retrospectively
reviewed the outcomes of EGD in 20 consecutive pregnant women. The indication for the procedure in
about two thirds of patients was gastrointestinal bleeding. Conscious sedation was induced by
intravenous administration of drugs commonly used for this purpose (meperidine, diazepam, and
midazolam). Diagnostic yield was high, especially in patients with gastrointestinal bleeding. Except for
transient and unexplained fever at 12 hours after endoscopy in 1 patient, none of the pregnant women
experienced a complication of the procedure. Outcomes of the pregnancies were determined in 19
cases; all of the babies were delivered alive (17 term, 2 premature). There was no increase in the
number of cesarean sections. During the neonatal period, no serious illnesses or developmental
anomalies were detected. Cappell and Sidhom103(2660) also found 17 previous reports where EGD
was performed during pregnancy; no complications attributable to EGD were described. As in every
patient, the risks of a procedure must always be weighed against the potential benefit. It would appear,
however, that EGD can be performed safely in pregnant women when there is a definite indication for
the procedure, such as gastrointestinal bleeding.
EGD may rarely be indicated during the early postoperative period after gastric surgery, usually for
gastrointestinal bleeding (see also Chapter 52: Endoscopy in the Postoperative Upper Gastrointestinal
Tract). A wide range of factors affect the integrity of a surgical anastomosis, including surgical
technique, inflammation, infection, and local blood flow.104(2661) Furthermore, the tensile strength of
an anastomosis is at its lowest during the first few days after surgery. A study in 10 dogs by
Chardavoyne et al.104(2662) of EGD performed at 1 to 7 days after partial gastrectomy found no
evidence of anastomotic leakage of air or infection. However, there are relatively few data concerning
the potential for complications when upper endoscopy is performed in patients with a recent surgical
anastomosis. It is therefore advisable to avoid EGD during the immediate postoperative period.
Complications
Despite the increased use of upper endoscopy, the complication rate has remained low. In fact, Hart
and Classen105(2663) described a temporal trend toward a reduced incidence of complications. This
decrement may be due to improvements in endoscopic training, cardiopulmonary monitoring, and
endoscope disinfection, among other factors. The reported frequency of EGD-related complications is
lower than that of other endoscopic procedures including colonoscopy and endoscopic retrograde
cholangiopancreatography but increases with the use of therapeutic interventions including
sclerotherapy and dilation.106(2664) Complications of sclerotherapy, percutaneous endoscopic
gastrostomy, and laser therapy are discussed, respectively, in Chapters 35, 55, and 43.
Most data on frequency of complications are derived from surveys of practitioners, and hence, there is
the potential for inaccuracy in reporting. Data from four relatively large series, representing experience
from the United States and the United Kingdom, are summarized in Table 374.107110(2665)
Complications of endoscopy have been extensively reviewed by Baillie.111(2666)
TABLE 37-4 Reported Complications of Upper Gastrointestinal Endoscopy (Part i of ii )

the potential for inaccuracy in reporting. Data from four relatively large series, representing experience
from the United States and the United Kingdom, are summarized in Table 374.107110(2665)
Complications of endoscopy have been extensively reviewed by Baillie.111(2666)
TABLE 37-4 Reported Complications of Upper Gastrointestinal Endoscopy (Part i of ii )

COUNTRY
[REF.]
US [107]
US [108]
US [109]
Britain [110]
YEAR
COLLECTED
NUMBER
OF
PROCEDURES
NUMBER
OF
ENDOSCOPISTS
CARDIORESPIRATORY*
211,410
9,875
2,320
23,500
404
55
277
63
129 (0.06)
9 (0.9)
3 (0.1)
NR
1972-1973
1978
1978-1979
1971-1972
INFECTION*
17 (0.008)
1 (0.01)
4 (0.2)
16 (0.07)
TABLE 374 Reported Complications of Upper Gastrointestinal Endoscopy (Part ii of ii )

COUNTRY [REF.]
BLEEDING*
PERFORATION*
MORBIDITY
63 (0.03)
70 (0.03)
0.2
US [107]
US [108]
1 (0.01)
10 (0.1)
0.24
US [109]
3 (0.1)
5 (0.2)
0.6
Britain [110]
6 (0.03)
26 (0.01)
0.2
NRnot reported.
* Values in parentheses represent percentages.
Study included pneumonia in cardiorespiratory group instead of in infection group.
DEATH
0.006
0.03
0.1
0.012
The most common complications were cardiopulmonary in nature and resulted from administration of
sedative drugs or passage and manipulation of the endoscope. Bleeding and perforation have been
less frequently reported. Overall morbidity and mortality rates were 0.25% and 0.009%, respectively;
the higher mortality and morbidity reported in the ASGE survey of procedures for upper gastrointestinal
bleeding is probably due to the emergent nature of the procedures.109(2667) Although not usually
classified as a complication, a missed diagnosis may have a significant adverse effect on patient
outcome. In a survey of an insurance industry database in the United States, Gerstenberger and
Plumeri112(2668) found that about one fourth of malpractice claims alleged diagnostic error, including
missed malignancies.
Pulmonary Complications
Respiratory complications are attributable to several factors, including premedication with
benzodiazepines or narcotics, oropharyngeal intubation, and elevation of the diaphragm owing to
gastric distention. Respiratory distress has been reported owing to air insufflation and compression of
the lung when the stomach is positioned above the diaphragm.113(2669) Iber et al.114(2670) found 10
cases of apnea and cardiopulmonary arrest among 10,326 endoscopic procedures recorded in a
computer database. In 6 cases, the arrest occurred after completion of the procedure. A decrease in
respiratory excursion occurred in 80% of patients who received intravenous mid-azolam and were
monitored by means of a plethysmograph vest during endoscopy in the study of Bell et al.115(2671)
Respiratory rate did not change in most patients except for cyclical periods of apnea in some cases.
Desaturation of arterial oxygen tension, defined as a decrement of greater than 4% in oxygen
saturation or saturation of less than 90%, occurs with a reported frequency of 15 to
54%.116120(2672) Most such episodes are detected by pulse oximetry alone and are of little clinical
consequence. In a survey of ASGE members that included more than 21,000 endoscopic procedures,
only 114 serious cardiopulmonary complications, defined as requiring intervention or resulting in death,
were reported, a rate of 5.43 per 1000 procedures.121(2673) Although there was no difference in the
occurrence of complications whether midazolam or diazepam was administered for sedation, there
was a higher complication rate with concomitant use of narcotic agents and for procedures performed
on an emergency basis.
Several investigators have attempted to predict which patients are at greatest risk for pulmonary
complications. In the study of 82 patients undergoing EGD, a decrease in arterial oxygen saturation of
greater than 4% and a saturation of less than 90% occurred in 54% and 43%, respectively.117(2674)
Risk factors for desaturation included age greater than 65 years, hemoglobin less than 10 gm/dl, and
body mass index of greater than 28/m2. However, the study of Bell et al.115(2675) of intravenous
midazolam for conscious sedation found no correlation between age or gender and the development of
hypoxemia during EGD. Age, sex, and body weight were not predictive of oxygen desaturation in a
study of 34 patients aged 2 months to 18 years in which various combinations of benzodiazepine or
narcotic agents were administered intravenously.122(2676) Body weight alone is probably not the best
parameter for assessment of risk of hypoxemia. When body weight and height were compared with
ideal weight, Block et al.123(2677) found that obesity was associated with a significant increase in risk
of hypoxemia.
Pulmonary function tests were obtained by Rostykus et al.118(2678) in 13 patients before EGD.
Desaturation to less than 90% was observed in 6 of 7 patients with moderate or severe airway
obstruction and in none of the patients without obstructive airway disease. Desaturation to less than
85% occurred in 15% of patients in another study and was found to be more common in patients of
advanced age, those with obstructive airway disease, and those receiving higher doses of sedative
drugs.116(2679) These results are in contrast to those of the study of Bell et al.115(2680) in which
preprocedure respiratory function testing in 85 patients failed to identify those at greater risk for
hypoxemia during the procedure. Similar results were obtained by Whorwell et al.,124(2681) who
obtained pulmonary function tests in 65 patients before endoscopy. Because desaturation is relatively
common during EGD, the use of supplemental oxygen during the procedure has been advocated.
Several studies have demonstrated that this measure abolishes desaturation.123,125127(2682) In
addition to hypoxemia, hypercarbia may also occur during EGD as a result of hypoventilation. Using
transcutaneous sensing, a partial pressure of carbon dioxide (PCO2) of greater than 50 mm Hg was
found in 10 to 15% of patients undergoing upper endoscopy in one study.119(2683) The occurrence of
hypercarbia was independently predicted by narcotic and benzodiazepine doses, dementia, and
oxygen requirement.119(2684)
Although most endoscopists are well aware of the potential for respiratory complications related to the
use of conscious sedation, they may be less cognizant of the risk of pulmonary aspiration.128(2685)
The latter was encountered in 0.07% of examinations in one series;110(2686) risk factors included
gastric distention, inadequate gastric emptying, and impaired airway protective reflexes. The risk of
aspiration increases if the stomach is not empty, as may occur with gastric outlet obstruction or
gastrointestinal bleeding. Lipper et al.129(2687) found that 6 of 30 consecutive patients (20%)
undergoing emergency endoscopy for active gastrointestinal hemorrhage developed new pulmonary
infiltrates when postprocedure chest x-rays were compared with chest radiographs obtained within 12
hours before the procedure.
Cardiac Complications
Minor cardiac dysrhythmias occur during EGD and are usually inconsequential. These include
premature ventricular contractions, atrial ectopic beats, and less commonly, ST segment or t wave
changes.130132(2688) Albeit rarely, arrhythmias other than sinus tachycardia have been observed in
children and adolescents undergoing endoscopy.122(2689)
Although dysrhythmias may be due in part to anxiety and abdominal distention, hypoxemia is also a
major risk factor.118,133(2690) In a study of 13 patients undergoing EGD, cardiac arrhythmias, mostly
frequent premature ventricular contractions, occurred mainly during periods of desaturation.118(2691)
However, one randomized, double-blind trial found that supplemental oxygen (2 L/min) did not reduce
the frequency of clinically significant arrhythmias.130(2692) The results of this study by Bowling et
al.130(2693) of 103 patients over the age of 60 years also found that arrhythmias are common among
individuals in this age category and that there was no evidence that these were induced by EGD.
Cardiac dysrhythmias have been found to be more frequent in patients with a history of cardiac
disease.134136(2694) Preexisting cardiac disease and concomitant therapeutic procedures were
identified as risk factors for adverse events, with frequencies of 72% and 87%, respectively, in a study
by Thompson et al.133(2695) in the United Kingdom. In another study, premedication with
anticholinergic agents was a risk factor for tachycardia and possibly dysrhythmias.137(2696) Although
monitoring of blood pressure and heart rate during procedures often detected unsuspected
hemodynamic aberrations in the prospective study of DiSario et al.,138(2697) these were not clinically
significant. In the prospective study of Wilcox et al.,139(2698) the frequency of arrhythmias in patients
with stable but severe coronary heart disease did not increase during endoscopic procedures by
comparison with baseline periods of monitoring. Although there was electrocardiographic evidence of
myocardial ischemia in 24% of patients during endoscopy, this was not clinically significant in any
patient.
Infectious Complications
Infectious complications of EGD include transmission of infectious agents by contaminated
endoscopes or ancillary equipment,140142(2699) bacteremia, and most commonly, aspiration
pneumonia.107(2700) Endoscopy-related infection is the subject of an extensive review by Schembre
and Bjorkman.143(2701)
Two outbreaks in the United States of infections with Pseudomonas aeruginosa and Mycobacterium
chelonae have been linked to contaminated endoscope washing machines.144(2702) In a review of
published reports, Spach et al.145(2703) found a total of 281 documented cases of infections
transmitted by endoscopes; Salmonella spp.146148(2704) and P. aeruginosa149(2705) were the
most common pathogens. Only one case of endoscopy-related hepatitis B infection has been
reported.150(2706) This was thought to be due to inadequate cleaning of the air/water channel of an
endoscope. Longitudinal follow-up studies of patients undergoing endoscopy in areas where hepatitis B
infection is endemic have failed to document transmission of the virus.151,152(2707) To date, no
episodes of HIV transmission have been reported. Transmission of infectious agents during endoscopy
is discussed in Chapter 8: Disinfection of Endoscopes and Accessories.
Bacteremia is uncommon during EGD, with a reported frequency of 4 to 8%.153156(2708) This
proportion increases to 30% or more with esophageal dilation or sclerotherapy.153(2709) As the
organisms cultured from the blood are usually members of the oropharyngeal flora, antibiotics with
activity against these bacteria should be selected if prophylaxis is indicated. The possibility of
endocarditis is usually a concern in relation to EGD-related bacteremia, although there are very few
well-documented cases.157(2710) Other rare complications include meningitis158(2711) and

phlegmonous gastritis, the latter of which has occurred after injection of India ink159(2712) and
polypectomy.160(2713)
Guidelines for antibiotic prophylaxis have been developed by professional societies based on available
data and a consensus of experts.161164(2714) Zuckerman et al.165(2715) found that a decision with
regard to the advisability of antibiotic prophylaxis was necessary in 15% of patients undergoing EGD.
The ASGE guideline statement classifies procedures according to risk for bacteremia.163(2716) EGD
is regarded as having a low risk for bacteremia, whereas esophageal dilation and variceal
sclerotherapy have high risks. The ASGE guidelines also group clinical conditions according to risk for
the development of infective endocarditis into high-, intermediate-, and no-risk categories. Those at
high risk include prosthetic heart valves, a previous history of endocarditis, and surgically constructed
systemic pulmonary shunts or conduits. The risk of infection of a synthetic vascular graft decreases
over time with pseudointimal covering, which is usually complete in about 1 year. There are insufficient
data, according to the ASGE guideline, to recommend antibiotic prophylaxis for high-risk patients who
undergo procedures with low rates of bacteremia (e.g., diagnostic EGD including biopsy), and the need
for prophylaxis should therefore be considered on a case-by-case basis. For example, Bianco et
al.166(2717) demonstrated that immunosuppression is associated with an increased risk of clinically
significant bacteremia in a study of 151 bone marrow transplant patients. Clinical evidence of infection
along with positive blood cultures was noted in 19% of patients within 24 hr after endoscopy. Treatment
with prednisone was associated with a significant increase in the risk of bacteremia. Acute
graft-versus-host disease (GVHD) did not increase the risk of bacteremia, but for patients with GVHD
taking prednisone the risk was especially high. The prevention of infective endocarditis is the subject of
a review by Durack.167(2718)
Although the potential for infectious complications of endoscopy is usually placed within a framework of
risk to patients, patients who harbor infectious agents also represent a risk for endoscopists and
endoscopy nurses. Because infection with a transmissible agent may be unrecognized in many
patients, universal precautions against infection are therefore mandatory for all endoscopic
procedures, with particular attention to wearing gloves and gowns, the use of needles, avoidance of
contact with body fluids, and the handling of contaminated equipment. Chong et al.168(2719) found
that H. pylori was significantly more common when sera from 122 endoscopists and endoscopy nurses
were compared with sera from 510 blood donors using the enzyme-linked immunosorbent assay.
Perforation
EGD-related perforations occur most often in the pharynx or esophagus.107,109(2720) These can
result from poor technique, lack of patient cooperation, and the presence of a Zenker diverticulum. The
presence of cervical osteophytes has also been cited as a contributing factor.169(2721) The risk for a
perforation also increases with attempts to pass an endoscope through an obstructing lesion such as a
carcinoma. Gastric perforations may result from passage of the instrument through a lesion or when
biopsies are obtained from a deep ulcer. A perforation due to endoscopic treatment of a Dieulafoy
lesion has been reported.170(2722) Perforations related to use of an endoscopic biopsy forceps are
exceedingly rare. However, Scott and Holmes171(2723) reported two cases in which biopsies had
been obtained from the small bowel; the mucosa was normal in one case and a diagnosis of celiac
disease was made in the other. Fortunately, EGD-related perforations are rare, with a reported
incidence of 0.01 to 0.1%.107110(2724)
There are relatively few data concerning management of patients with esophageal perforations related
to instrumentation.172176(2725) The mortality rate appears to be higher in patients with serious
esophageal diseases such as carcinoma. Furthermore, perforation of the thoracic esophagus has a
higher associated mortality rate than perforation of the cervical esophagus.177(2726) Prompt
recognition has been emphasized as an important determinant of survival.173,177 In addition to

contrast radiography, a chest x-ray may be of assistance in making the diagnosis. Panzini et
al.178(2727) reviewed chest radiographs of 15 patients who sustained an esophageal perforation due
to instrumentation. There was evidence of the perforation in 12 cases, pneumomediastinum being the
most common finding (60%) followed by loss of contour of the descending aorta at the left diaphragm
(33%) owing to a density near the aorta in the left cardiophrenic angle.
There have been reports over many years of cases in which a perforation occurred during EGD in the
complete absence of any indication or suspicion of such a complication during the
procedure.179189(2728) In some cases, no site of the perforation was found.190(2729) Most of
these reports discount the possibility that the perforation was due to direct mechanical forces exerted
with an endoscope or an accessory such as a biopsy forceps. The mechanisms whereby this type of
perforation can occur are mostly speculative, although air insufflation and changes in air pressure
during EGD are usually proposed as contributing factors. In the majority of cases, a disease process
that altered the normal anatomy of the gastrointestinal tract was present. Frequently unrecognized
before EGD, these disorders include obstructing jejunal tumors,191(2730) small bowel
adhesions,192(2731) an inflammatory mass secondary to duodenal perforation by a foreign
body,192(2732) esophageal strictures, and duodenal or jejunal diverticula.193,194(2733) There are
two reports of cecal perforation in patients undergoing EGD. In that of Rex et al.,195(2734) no cecal
abnormality was found at surgery, but the patient had two obstructing carcinomas at the splenic
flexure. The patient described by Weiner196(2735) had a large, necrotic cecal cancer with multiple
fistulas to the small bowel including the proximal jejunum distal to the ligament of Treitz. There are also
a few reports in which no contributing disorders were present.197(2736) In that of Pope and
Adair,197(2737) the perforation was thought to have been due to an "uncoordinated" series of severe
pressure changes in the upper gastrointestinal tract. Esophageal rupture was attributed to pressure
changes in the esophagus in the case of Gubbins et al.,198(2738) although this patient also had an
esophageal stricture proximal to the site of perforation. Although many patients have developed
symptoms of the perforation at variable intervals after the procedure, other patients have remained
asymptomatic despite radiographic evidence of a pneumoperitoneum or pneumomediastinum that was
discovered only incidentally.194(2739) Management of patients with these "nonmechanical"
perforations appears to have been dictated by individual circumstances; conservative measures were
frequently employed if the patient remained asymptomatic,180,183,199(2740) whereas patients with
symptoms usually underwent operation. For example, the patient described by Fireman et al.200(2741)
developed a pneumoperitoneum, pneumomediastinum, and subcutaneous emphysema in the absence
of an obvious perforation but recovered with only conservative treatment.
Bleeding
Gastrointestinal bleeding is an uncommon complication of EGD, with a reported incidence of 0.03 to
0.1%.107110(2742) There is no evidence that patients taking aspirin are at increased risk for
significant bleeding when endoscopic biopsies are obtained at EGD.201,202(2743) Bleeding may be
due to treatment of a previously bleeding lesion such as esophageal varices, EGD-induced
gastroesophageal tears, or hemorrhage at a biopsy site.107(2744) However, even with endoscopic
evaluation of acute upper gastrointestinal bleeding, procedure-induced hemorrhage is rare
(0.1%).109(2745)
Miscellaneous Rare Complications

The use of topical anesthetic agents to induce oropharyngeal anesthesia has been associated with
several types of complications including acute methemoglobinemia203,204(2746) and systemic
allergic reactions (see also Chapter 38: Technique of Upper Gastrointestinal
Endoscopy).205,206(2747) An overtube may be used in conjunction with upper endoscopy to facilitate

esophageal variceal banding or removal of foreign bodies. The use of this device has been associated
with esophageal perforation207(2748) and injury to the mucosa, which may become "pinched"
between the overtube and the endoscope.208(2749)
Mucosal tears in the region of the cardia and esophagogastric mucosal junction without significant
bleeding have occurred during EGD.209,210(2750)
Serum amylase values may be elevated after EGD in about 10 to 20% of patients, but the increased
levels are due to increases in salivary (S-type) rather than pancreatic (P-type)
isoamylase.211,212(2751) The mechanism whereby this occurs and whether this finding has any
clinical significance are unknown. However, swelling of the parotid and submaxillary salivary glands
may occur in patients undergoing EGD.213,214(2752)
There are rare case reports of acute pancreatitis after EGD with215(2753) and without216(2754)
manipulation of the main duodenal papilla. In the case reported by Morales and Hixson,215(2755)
biopsies were obtained from the main duodenal papilla in a patient with the Gardner variant of FAP. A
duodenal hematoma and acute pancreatitis developed in a 14-year-old girl with neurofibromatosis after
duodenal biopsies were obtained at EGD.217(2756) The patient had a normal platelet count, and
parameters of coagulation were normal before and after development of the hematoma. A similar
occurrence was previously reported in a 23-year-old man.218(2757) A duodenal hematoma without
development of pancreatitis has also been reported in a child who underwent EGD to obtain biopsies
from the small bowel.219(2758)
There are two reports of serious complications (hypertensive crisis, ventricular tachycardia) when EGD
was performed in patients with pheochromocytomas.220,221(2759) Hypertensive crisis was induced
by the intravenous administration of an opiate (fentanyl) for conscious sedation, but the patient
recovered.220 The other patient died as a result of ventricular tachycardia, and the tumor was found in
the left adrenal gland at autopsy.221(2760)
Hyperglycemia has occurred in nondiabetic patients undergoing upper endoscopy.222(2761) Impaction
of an endoscope, usually as a result of a retroflexion ("turn around") maneuver within a relatively
narrow segment of the upper gastrointestinal tract, is extremely unusual but has been
described.223,224(2762)
Rare and unusual complications of upper endoscopy include precipitation of hemorrhage, sometimes
with exsanguination, from an aortoesophageal fistula;225(2763) rupture of an umbilical hernia in a
patient with ascites;226(2764) fatal cerebral air embolism in a patient with gastrointestinal bleeding, a
large duodenal ulcer, and a duodenocaval fistula;227(2765) acute herniation of the stomach into the
chest (paraesophageal hernia) with gastric volvulus in a 21-year-old man and an esophageal
diaphragmatic hiatus that measured 4 cm in diameter at surgery;228(2766) herniation of the stomach
and other viscera through a hiatus 7 cm in diameter, leading to acute respiratory distress in an
80-year-old woman;229(2767) rupture of the short gastric blood vessels of the proximal stomach (three
cases);230,231(2768) obstruction of the superior vena cava;232(2769) and prolonged ileus.233(2770)
REFERENCES
1.
Kahn KL, Kosecoff J, Chassin MR, et al. The use and misuse of upper gastrointestinal
endoscopy. Ann Intern Med 1988;109:66470.
2. Working Party of the Clinical Services Committee of the British Society of Gastroenterology.
Provision of gastrointestinal endoscopy and related services for a district general hospital. Gut
1991;32:95105.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
Bond JH. Control of the volume of gastrointestinal endoscopy. Gastrointest Endosc

1993;39:1023.
Brook RH, Park RE, Chassin MR, et al. Predicting the appropriate use of carotid
endarterectomy, upper gastrointestinal endoscopy, and coronary angiography. N Engl J Med
1990;323:11737.
American Society for Gastrointestinal Endoscopy. Appropriate Use of Gastrointestinal
Endoscopy. Manchester, MA: ASGE, 1992;68.
Colin-Jones DG. Management of dyspepsia: Report of a working party. Lancet 1988;1:5769.
Health and Public Policy Committee, American College of Physicians. Endoscopy in the
evaluation of dyspepsia. Ann Intern Med 1985;102:2669.
Kahn KL, Park RE, Vennes J, et al. Assigning appropriateness ratings for diagnostic upper
gastrointestinal endoscopy using two different approaches. Med Care 1992;30:101828.
Akdamar K, Ertan A, Agrawal NM, et al. Upper gastrointestinal endoscopy in normal
asymptomatic volunteers. Gastrointest Endosc 1986;32:7880.
Jones RH, Lydeard SE, Hobbs FDR, et al. Dyspepsia in England and Scotland. Gut
1990;31:4015.
Kahn KL, Greenfield S. The efficacy of endoscopy in the evaluation of dyspepsia. A review of
the literature and development of a sound strategy. J Clin Gastroenterol 1986;8:34658.
Morrissey JF, Reichelderfer M. Gastrointestinal endoscopy (Part 1). N Engl J Med
1991;325:11429.
Naji SA, Brunt PW, Hagen S, et al. Improving the selection of patients for upper
gastrointestinal endoscopy. Gut 1993;34:187- 91.
Talley NJ, Weaver AL, Tesmer DL, Zinsmeister AR. Lack of discriminant value of dyspepsia
subgroups in patients referred for upper endoscopy. Gastroenterology 1993;105:137886.
Mansi C, Savarino V, Mela GS, et al. Are clinical patterns of dyspepsia a valid guideline for
appropriate use of endoscopy? A report on 2253 dyspeptic patients. Am J Gastroenterol
1993;88:10115.
Boring CC, Squires TS, Tong T, et al. Cancer statistics, 1994. CA Cancer J Clin
1994;44:726.
Richter JE. Dyspepsia: Organic causes and differential characteristics from functional
dyspepsia. Scand J Gastroenterol 1991;26(Suppl 182):116.
Lockhart SP, Schofield PM, Gribble RJN, et al. Upper gastrointestinal endoscopy in the
elderly. BMJ 1985;290:283.
Forbat LN, Gribble RJN, Baron JH. Gastrointestinal endoscopy in the young. BMJ
1987;295:365.
Williams B, Luckas M, Ellingham JH, et al. Do young patients with dyspepsia need
investigation?. Lancet 1988;2:134951.
Hallissey MT, Allum WH, Jewkes AJ, et al. Early detection of gastric cancer. BMJ
1990;301:5135.
Peterson WL. Helicobacter pylori and peptic ulcer disease. N Engl J Med 1991;324:10438.
Graham DY, Lew GM, Klein PD, et al. Effect of treatment of Helicobacter pylori infection on
the long term recurrence of gastric or duodenal ulcer: A randomized, controlled study. Ann
Intern Med 1992;116:7058.
Hentschel E, Brandstatter G, Dragosics B, et al. Effect of ranitidine and amoxicillin plus
metronidazole on the eradication of Helicobacter pylori and the recurrence of duodenal ulcer.
N Engl J Med 1993;328:30812..
Kreuning J, Bosman FT, Kuiper G, et al. Gastric and duodenal mucosa in "healthy"
individuals. An endoscopic and histopathological study of 50 volunteers. J Clin Pathol
1978;31:6977.
DeLuca V, Winnan GG, Sheahan DG. Is gastroduodenitis part of the spectrum of peptic ulcer
disease?. J Clin Gastroenterol 1981;3:1722.
Johannessen T, Petersen H, Kleveland PM, et al. The predictive value of history in dyspepsia.
28. Sobala GM, Crabtree JE, Pentith JA, et al. Screening dyspepsia by serology to Helicobacter
pylori. Lancet 1991;338:946.
29. Castell DO, Knuff TE, Brown FC, et al. Dysphagia. Gastroenterology 1979;76:101524.
30. Ott DJ, Chen YM, Wu WC, Gelfand DW. Endoscopic sensitivity in the detection of
esophageal strictures. J Clin Gastroenterol 1985;7:1215.
31.
Ott DJ, Chen YM, Wu WC, et al. Radiographic and endoscopic sensitivity in detecting lower
esophageal mucosal ring. AJR 1986;147:2615.
32. McDonald GB, Sharma P, Hackman RC, et al. Esophageal infections in immunosuppressed
patients after marrow transplantation. Gastroenterology 1985;88:11117.
33. Wilcox CM. Esophageal disease in the acquired immunodeficiency syndrome: Etiology,
diagnosis and management. Am J Med 1992;92:41221.
34. Porro GB, Parente F, Cernuschi M. The diagnosis of esophageal candidiasis in patients with
acquired immune deficiency syndrome: Is endoscopy always necessary?. Am J Gastroenterol
1989;84:1436.
35. Parente F, Cernuschi M, Rizzardini G, et al. Opportunistic infections of the esophagus not
responding to oral systemic antifungals in patients with AIDS: Their frequency and treatment.
36. Howard PJ, Heading RC. Epidemiology of gastroesophageal reflux disease. World J Surg
1992;16:28893.
37. Heading RC. Epidemiology of oesophageal reflux disease. Scand J Gastroenterol
1989;24(Suppl 168):337.
38. Knuff TE, Benjamin SB, Worsham GF, et al. Histologic evaluation of chronic
gastroesophageal reflux. An evaluation of biopsy methods and diagnostic criteria. Dig Dis Sci
1984;29:194201.
39. Hetzel DJ, Dent J, Reed WD, et al. Healing and relapse of severe peptic esophagitis after
treatment with omeprazole. Gastroenterology 1988;95:90312.
40. Achem SR, Kolts BE, Wears R, et al. Chest pain associated with nutcracker esophagus: A
preliminary study of the role of gastroesophageal reflux. Am J Gastroenterol 1993;88:18792.
41. Hsia PC, Maher KA, Lewis JH, et al. Utility of upper endoscopy in the evaluation of noncardiac
chest pain. Gastrointest Endosc 1991;37:226.
42. Kahrilas PJ. Nutcracker esophagus: An idea whose time has gone?. Am J Gastroenterol
1993;88:1679.
43. Malagelada J, Camilleri M. Unexplained vomiting: A diagnostic challenge. Ann Intern Med
1984;101:2118.
44. Kozarek RA, Botoman VA, Patterson DJ. Long term follow-up in patients who have undergone
balloon dilation for gastric outlet obstruction. Gastrointest Endosc 1990;36:55861.
45. Kiss A, Wiesnagrotzki S, Abatzi TA, et al. Upper gastrointestinal endoscopy findings in
patients with longstanding bulimia nervosa. Gastrointest Endosc 1989;35:5168.
46. Eckardt VF, Giessler W, Kanzler G, Bernhard G. Does endoscopic follow-up improve the
outcome of patients with benign gastric ulcers and gastric cancer?. Cancer 1992;69:3015.
47. Llanos O, Guzman S, Duarte I. Accuracy of the first endoscopic procedure in the differential
diagnosis of gastric lesions. Ann Surg 1982;195:2246.
48. Tragardh B, Haglund U. Endoscopic diagnosis of gastric ulcer. Evaluation of the benefits of
endoscopic follow-up observation for malignancy. Acta Chir Scand 1985;151:3741.
49. Pruitt RE, Truss CD. Endoscopy, gastric ulcer, and gastric cancer. Follow-up endoscopy for
all gastric ulcers? Dig Dis Sci 1993;38:2848.
50. Church JM, McGannon E, Hull-Boiner S, et al. Gastroduodenal polyps in patients with familial
adenomatous polyposis. Dis Colon Rectum 1992;35:11703.
51. Noda Y, Watanabe H, Iida M, et al. Histologic follow-up of ampullary adenomas in patients
with familial adenomatosis coli. Cancer 1992;70:184756.
52. Offerhaus GJA, Giardiello FM, Krush AJ, et al. The risk of upper gastrointestinal cancer in
familial adenomatous polyposis. Gastroenterology 1992;102:19802..
53. Sarre RG, Frost AG, Jagelman DG, et al. Gastric and duodenal polyps in familial
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
adenomatous polyposis: A prospective study of the nature and prevalence of upper

gastrointestinal polyps. Gut 1987;28:30614.
Spigelman AD, Williams CB, Talbot IC, et al. Upper gastrointestinal cancer in patients with
familial adenomatous polyposis. Lancet 1989;2:7835.
. The Standards of Training and Practice Committee, American Society for Gastrointestinal
Endoscopy: The role of endoscopy in the surveillance of premalignant conditions of the upper
gastrointestinal tract. Gastrointest Endosc 1988;34:18s-20s.
ReMine SG, Hughes RW Jr, Weiland LH. Endoscopic gastric polypectomies. Mayo Clin Proc
1981;56:3715.
Spechler SJ, Goyal RK. Barrett's esophagus. N Engl J Med 1986;315:36271.
Dent J, Bremner CG, Collen MJ, et al. Working party report to the World Congresses of
Gastroenterology, Sydney 1990: Barrett's oesophagus. J Gastroenterol Hepatol 1991;6:122.
Geller AJ, Kolts BE, Achem SR, et al. The high frequency of upper gastrointestinal pathology
in patients with fecal occult blood and colon polyps. Am J Gastroenterol 1993;88:11847.
Zuckerman G, Benitez J. A prospective study of bidirectional endoscopy (colonoscopy and
upper endoscopy) in the evaluation of patients with occult gastrointestinal bleeding. Am J
Chen YK, Gladden DR, Kestenbaum DJ, Collen MJ. Is there a role for upper gastrointestinal
endoscopy in the evaluation of patients with occult blood-positive stool and negative
colonoscopy?. Am J Gastroenterol 1993;88:20269.
Calvey HD, Castleden CM. Gastrointestinal investigations for anaemia in the elderly: A
prospective study. Age Ageing 1987;16:399404.
Cook IJ, Pavli P, Riley JW, et al. Gastrointestinal investigation of iron deficiency anaemia.
BMJ 1986;292:13802.
Hsia PC, al-Kawas FH. Yield of upper endoscopy in the evaluation of asymptomatic patients
with Hemoccult-positive stool after a negative colonoscopy. Am J Gastroenterol
1992;87:15714.
Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in patients with iron-deficiency
anemia. N Engl J Med 1993;329:16915.
Gordon SR, Smith RE, Power GC. The role of endoscopy in the evaluation of iron deficiency
anemia in patients over the age of 50. Am J Gastroenterol 1994;89:19637.
Thomas WM, Hardcastle JD. Role of upper gastrointestinal investigations in a screening
study for colorectal neoplasia. Gut 1990;31:12947.
Achkar E, Carey WD, Petras R, et al. Comparison of suction capsule and endoscopic biopsy
of small bowel mucosa. Gastrointest Endosc 1986;32:27881.
Mee AS, Burke M, Vallon AG, et al. Small bowel biopsy for malabsorption: Comparison of the
diagnostic adequacy of endoscopic forceps and capsule biopsy specimens. BMJ
1985;291:76972.
Dandalides SM, Carey WD, Petras R, et al. Endoscopic small bowel biopsy: A controlled trial
evaluating forceps size and biopsy location in the diagnosis of normal and abnormal mucosal
architecture. Gastrointest Endosc 1989;35:197200.
Ladas SD, Tsamouri M, Kouvidou C, Raptis SA. Effect of forceps size and mode of
orientation on endoscopic small bowel biopsy evaluation. Gastrointest Endosc 1994;40:515.
The North Italian Endoscopic Club for the Study and Treatment of Esophageal Varices.
Pagliaro L, D'Amico G, Sorensen TIA, et al. Prevention of first bleeding in cirrhosis: A
meta-analysis of randomized trials of nonsurgical treatment. Ann Intern Med 1992;117:5970.
The Veterans Affairs Cooperative Variceal Sclerotherapy Group. Prophylactic sclerotherapy
for esophageal varices in men with alcoholic liver disease. N Engl J Med 1991;324:177984.
Poynard T, Cales P, Pasta L, et al. Beta-adrenergic-antagonist drugs in the prevention of
gastrointestinal bleeding in patients with cirrhosis and esophageal varices. N Engl J Med
1991;324:15328.
76. Gumaste VV, Dave PB. Ingestion of corrosive substances by adults. Am J Gastroenterol
1992;87:15.
77. Zargar SA, Kochhar R, Mehta S, et al. The role of fiberoptic endoscopy in the management of
corrosive ingestion and modified endoscopic classification of burns. Gastrointest Endosc
1991;37:1659.
78. Peterson WL, Barnett CC, Smith HJ, et al. Routine early endoscopy in
upper-gastrointestinal-tract bleeding: A randomized, controlled trial. N Engl J Med
1981;304:9259.
79.
Standards of Practice Committee, The American Society for Gastrointestinal Endoscopy. The
role of endoscopy in the management of non-variceal acute upper gastrointestinal bleeding.
Gastrointest Endosc 1992;102:13948..
81. Terblanche J, Burroughs AK, Hobbs KEF. Controversies in the management of bleeding
esophageal varices. N Engl J Med 1989;320:13938, 146975.
82. Stiegmann GV, Goff JS, Michaletz-Onody PA, et al. Endoscopic sclerotherapy as compared
with endoscopic ligation for bleeding esophageal varices. N Engl J Med 1992;326:152732.
83. Brady PG. Esophageal foreign bodies. Gastroenterol Clin North Am 1991;20:691701.
84. The American Society for Gastrointestinal Endoscopy. Standards of Training and Practice
Committee. Guidelines for clinical application: Esophageal dilatation. Gastrointest Endosc
1991;37:1224.
85. Lo AY, Shinya H. Endoscopic placement of long intestinal tubes. Am Surg 1993;59:6267.
86. Pasricha PJ, Ravich WJ, Hendrix TR, et al. Intrasphincteric botulinum toxin for the treatment
of achalasia. N Engl J Med 1995;332:7748.
87. Loizou LA, Grigg D, Atkinson M, et al. A prospective comparison of laser therapy and
intubation in endoscopic palliation for malignant dysphagia. Gastroenterology
1991;100:130310.
88. Dooley CP, Larson AW, Stace NH, et al. Double-contrast barium meal and upper
gastrointestinal endoscopy. A comparative study. Ann Intern Med 1984;101:53845.
89. Stevenson GW, Norman G, Frost R, Somers S. Barium meal or endoscopy? A prospective
randomized study of patient preference and physician decision making. Clin Radiol
1991;44:31721.
90. Longstreth GF. Long-term costs after gastroenterology consultation with endoscopy versus
radiography in dyspepsia. Gastrointest Endosc 1992;38:236.
91. Scott BB. Gastroenterology in the Trent region in 1992 and a review of changes since 1975.
Gut 1995;36:46872.
92. Schaap NP, Houben MH, Driessen WM, van Spreeuwel JP. Endoscopic studies of the
digestive tract as a service for family practitioners: Experience in the Eindhoven area. Ned
Tijdschr Geneeskd 1993;137:11426.
93. Jones R. Gastro-oesophageal reflux disease in general practice. Scand J Gastroenterol Suppl
1995;211:358.
94. Minoli G, Prada A, Gambetta G, et al. The ASGE guidelines for the appropriate use of upper
gastrointestinal endoscopy in an open access system. Gastrointest Endosc 1995;42:3879.
95.
Mahajan RJ, Barthel JS, Marshall JB. Appropriateness of referrals for open-access
endoscopy. How do physicians in different medical specialties dox Arch Intern Med
1996;156:20659.
96. Heatley RV. Open-access upper gastrointestinal endoscopy: Visions of the future or the
past?. Br J Hosp Med 1994;51:3669.
97. Saunders BP, Trewby PN. Open access endoscopy: Is the lost outpatient clinic of value?.
Postgrad Med J 1993;69:78790.
98. Quine MA, Bell GD, McCloy RF, et al. Appropriate use of upper gastrointestinal
endoscopyA prospective audit. Steering Group of the Upper Gastrointestinal Endoscopy
Audit Committee. Gut 1994;35:120914.
99. Adang RP, Vismans JF, Talmon JL, et al. Appropriateness of indications for diagnostic upper
gastrointestinal endoscopy: Association with relevant endoscopic disease. Gastrointest
Endosc 1995;42:3907.
100. Mansi C, Mela GS, Savarino V, et al. Open access endoscopy: A large-scale analysis of its
use in dyspeptic patients. J Clin Gastroenterol 1993;16:14953; discussion 1534.
101. Hungin AP, Thomas PR, Bramble MG, et al. What happens to patients following open access
gastroscopy? An outcome study from general practice. Br J Gen Pract 1994;44:51921.
102. Cappell MS. The safety and clinical utility of esophagogastroduodenoscopy for acute
gastrointestinal bleeding after myocardial infarction: A six-year study of 42 endoscopies in 34
consecutive patients at two university teaching hospitals. Am J Gastroenterol
1993;88:34450.
103. Cappell MS, Sidhom O. A multicenter, multiyear study of the safety and clinical utility of
esophagogastroduodenoscopy in 20 consecutive pregnant females with follow-up of fetal
outcome. Am J Gastroenterol 1993;88:19005.
104. Chardavoyne R, Ratner LE, Jaume JC, et al. Safety of endoscopy in the immediate
postoperative period following gastric anastomosis. Surg Endosc 1989;3:135.
105. Hart R, Classen M. Complications of diagnostic gastrointestinal endoscopy. Endoscopy
1990;22:22933.
106. Muhldorfer SM, Kekos G, Hahn EG, et al. Complications of therapeutic gastrointestinal
endoscopy. Endoscopy 1992;24:27683..
107. Mandelstam P, Sugawa C, Silvis SE, et al. Complications associated with
esophagogastroduodenoscopy and with esophageal dilation. Gastrointest Endosc
1976;23:169.
108. Davis RE, Graham DY. Endoscopic complications: The Texas experience. Gastrointest
Endosc 1979;25:1469.
109.
Gilbert DA, Silverstein FE, Tedesco FJ. National ASGE survey on upper gastrointestinal
bleeding: Complications of endoscopy. Dig Dis Sci 1981;26:55s-9s.
110. Schiller KFR, Cotton PB, Salmon PR. The hazards of digestive fibre-endoscopy: A survey of
British experience [Abstract]. Gut 1972;13:1027.
111. Baillie J. Complications of endoscopy. Endoscopy 1994;26:185203.
112. Gerstenberger PD, Plumeri PA. Malpractice claims in gastrointestinal endoscopy: Analysis of
an insurance industry data base. Gastrointest Endosc 1993;39:1328.
113. Narendranathan M, Kalam A. Respiratory distress during endoscopyReport of an unusual
case. Postgrad Med J 1987;63:8056.
114. Iber FL, Livak A, Kruss DM. Apnea and cardiopulmonary arrest during and after endoscopy. J
Clin Gastroenterol 1992;14:10913.
115. Bell GD, Reeve PA, Moshiri M, et al. Intravenous midazolam: A study of the degree of oxygen
desaturation occurring during upper gastrointestinal endoscopy. Br J Clin Pharmacol
1987;23:7038.
116. Dark DS, Campbell DR, Wesselius LJ. Arterial oxygen desaturation during gastrointestinal
endoscopy. Am J Gastroenterol 1990;85:131721.
117. Dhariwal A, Plevris JN, Lo NTC, et al. Age, anemia, and obesity-associated oxygen
desaturation during upper gastrointestinal endoscopy. Gastrointest Endosc 1992;38:6848.
118. Rostykus PS, McDonald GB, Albert RK. Upper intestinal endoscopy induces hypoxemia in
patients with obstructive pulmonary disease. Gastroenterology 1980;78:48891.
119. Freeman ML, Hennessy JT, Cass OW, Pheley AM. Carbon dioxide retention and oxygen
desaturation during gastrointestinal endoscopy. Gastroenterology 1993;105:3319.
120. Barkin JS, Krieger B, Blinder M, et al. Oxygen desaturation and changes in breathing pattern
in patients undergoing colonoscopy and gastroscopy. Gastrointest Endosc 1989;35:52630.
121. Arrowsmith JB, Gerstman BB, Fleischer DE, et al. Results from the American Society for
Gastrointestinal Endoscopy/US Food and Drug Administration collaborative study on the
1991;37:4217.
122. Gilger MA, Jeiven SD, Barrish JO, McCarroll LR. Oxygen desaturation and cardiac
arrhythmias in children during esophagogastroduodenoscopy using conscious sedation.
123. Block R, Jankowski J, Johnston D, et al. The administration of supplementary oxygen to
prevent hypoxia during upper alimentary endoscopy. Endoscopy 1993;25:26973.
124. Whorwell PJ, Smith CL, Foster KJ. Arterial blood gas tensions during upper gastrointestinal
endoscopy. Gut 1976;17:797800.
125. Bell GD, Bown S, Morden A, et al. Prevention of hypoxaemia during upper-gastrointestinal
endoscopy by means of oxygen via nasal cannulae. Lancet 1987;1:10224.
126. Solomon SA, Isaac T, Banerjee AK. Oxygen desaturation during endoscopy in the elderly. J R
Coll Physicians Lond 1993;27:168.
127. al-Qorain A, Adu-Gyamfi Y, Larbi EB, al-Shedokhi F. The effect of supplemental oxygen in
sedated and unsedated patients undergoing upper gastrointestinal endoscopy. J Int Med Res
1993;21:16570.
128.
Cotton PB. Pulmonary aspiration after fibre-endoscopy. BMJ 1972;4:4912.
129. Lipper B, Simon D, Cerrone F. Pulmonary aspiration during emergency endoscopy in patients
with upper gastrointestinal hemorrhage. Crit Care Med 1991;19:3303.
130. Bowling TE, Hadjiminas CL, Polson RJ, et al. Effects of supplemental oxygen on cardiac
rhythm during upper gastrointestinal endoscopy: A randomised controlled double blind trial.
Gut 1993;34:14927.
131. Fujita R, Kumura F. Arrhythmias and ischemic changes of the heart induced by gastric
endoscopic procedures. Am J Gastroenterol 1975;64:448.
132. Levy N, Abinader E. Continuous electrocardiographic monitoring with Holter
electrocardiocorder throughout all stages of gastroscopy. Am J Dig Dis 1977;22:10916..
133. Thompson AM, Park KGM, Kerr F, et al. Safety of fibreoptic endoscopy: Analysis of
cardiorespiratory events. Br J Surg 1992;79:10469.
134. Gupta SC, Gopalswamy N, Sarkar A, et al. Cardiac arrhythmias and electrocardiographic
changes during upper and lower gastrointestinal endoscopy. Mil Med 1990;155:911.
135. Mathew PK, Ona FV, Damevski K, Wallace WA. Arrhythmias during upper gastrointestinal
endoscopy. Angiology 1979;30:83440.
136. McAlpine JK, Martin BJ, Devine BL. Cardiac arrhythmias associated with upper
gastrointestinal endoscopy in elderly subjects. Scott Med J 1990;35:1024.
137. Segawa K, Nakazawa S, Yamao K, et al. Cardiac response to upper gastrointestinal
138. DiSario JA, Waring JP, Talbert G, et al. Monitoring of blood pressure and heart rate during
routine endoscopy: A prospective, randomized, controlled study. Am J Gastroenterol
1991;86:95660.
139. Wilcox CM, Faibicher M, Wenger NK, Shalek KA. Prevalence of silent myocardial ischemia
and arrhythmias in patients with coronary heart disease undergoing gastrointestinal tract
endoscopic procedures. Arch Intern Med 1993;153:232530.
140. Alvarado CJ, Stolz SM, Maki DG. Nosocomial infections from contaminated endoscopes: A
flawed automated endoscope washer. An investigation using molecular epidemiology. Am J
Med 1991;91:272S-280S.
141. Langenberg W, Rauws EA, Oudbier JH, Tytgat GN. Patient-to-patient transmission of
Campylobacter pylori infection by fiber-optic gastroduodenoscopy and biopsy. J Infect Dis
1990;161:50711.
142. Maloney S, Welbel S, Daves B, et al. Mycobacterium abscessus pseudoinfection traced to an
automated endoscope washer: Utility of epidemiologic and laboratory investigation. J Infect
Dis 1994;169:11669.
143. Schembre D, Bjorkman DJ. Review article: Endoscopy-related infections. Aliment Pharmacol
Ther 1993;7:34755.
144. Centers for Disease Control. Nosocomial infection and pseudoinfection from contaminated
endoscopes and bronchoscopes Wisconsin and Missouri. JAMA 1991;266:21978.
145. Spach DH, Silverstein FE, Stamm WE. Transmission of infection by gastrointestinal
endoscopy and bronchoscopy. Ann Intern Med 1993;118:11728.
146. Beecham HJ, Cohen ML, Parkin WE. Salmonella typhimurium. Transmission by fiberoptic
upper gastrointestinal endoscopy. JAMA 1979;241:10135.
147. Dean AG. Transmission of Salmonella typhi by fiberoptic endoscopy. Lancet 1977;2:134.
148. O'Connor BH, Bennett JR, Alexander JG, et al. Salmonellosis infection transmitted by
fibreoptic endoscopes. Lancet 1982;2:8646.
149. Greene WH, Moody M, Hartley R, et al. Esophagoscopy as a source of Pseudomonas
aeruginosa sepsis in patients with acute leukemia: The need for sterilization of endoscopes.
150. Birnie GG, Quigley EM, Clements GB, et al. Endoscopic transmission of hepatitis B virus. Gut
1983;24:1714.
151. Ayoola EA. The risk of type B hepatitis infection in flexible fiberoptic endoscopy. Gastrointest
Endosc 1981;27:602.
152. Chiaramonte M, Farini R, Truscia D, et al. Risk of hepatitis B virus infection following upper
gastrointestinal endoscopy: A prospective study in an endemic area. Hepatogastroenterology
1983;30:18991.
153. Botoman VA, Surawicz CM. Bacteremia with gastrointestinal endoscopic procedures.
154. Shull HJ Jr, Greene BM, Allen SD, et al. Bacteremia with upper gastrointestinal endoscopy.
155. Norfleet RG, Mitchell PD, Mulholland DD, Philo J. Does bacteremia follow upper
gastrointestinal endoscopy?. Am J Gastroenterol 1981;76:4202.
156. Baltch AL, Buhac I, Agrawal A, et al. Bacteremia after upper gastrointestinal endoscopy. Arch
Intern Med 1977;137:5947.
157. Pritchard TM, Foust RT, Cantey JR, Leman RB. Prosthetic valve endocarditis due to
Cardiobacterium hominis occurring after upper gastrointestinal endoscopy. Am J Med
1991;90:5168.
158. Carley NH. Streptococcus salivarius bacteremia and meningitis following upper
gastrointestinal endoscopy and cauterization for gastric bleeding. Clin Infect Dis
1992;14:9478.
159. Hornig D, Khn H, Stadelmann O, Btticher R. Phlegmonous gastritis after Indian ink
marking. Endoscopy 1983;15:2669.
160. Lifton LJ, Schlossberg D. Phlegmonous gastritis after endoscopic polypectomy. Ann Intern
Med 1982;97:3735.
161. The American Society of Colon and Rectal Surgeons. Practice parameters for antibiotic
prophylaxis to prevent infective endocarditis or infected prosthesis during colon and rectal
endoscopy. Dis Colon Rectum 1992;35:277..
162. The Standards Task Force. American Society of Colon and Rectal Surgeons. Practice
parameters for antibiotic prophylaxisSupporting documentation. Dis Colon Rectum
1992;35:27885.
163. The American Society for Gastrointestinal Endoscopy. Antibiotic prophylaxis for
gastrointestinal endoscopy. Gastrointest Endosc 1995;42:6305.
164. Dajani AS, Bisno AL, Chung KJ, et al. Prevention of bacterial endocarditis. Recommendations
by the American Heart Association. JAMA 1990;264:291922.
165. Zuckerman GR, O'Brien J, Halsted R. Antibiotic prophylaxis in patients with infectious risk
factors undergoing gastrointestinal endoscopic procedures. Gastrointest Endosc
1994;40:53843.
166. Bianco JA, Pepe MS, Higano C, et al. Prevalence of clinically relevant bacteremia after upper
gastrointestinal endoscopy in bone marrow transplant recipients. Am J Med 1990;89:1346.
167. Durack DT. Prevention of infective endocarditis. N Engl J Med 1995;332:3844.
168. Chong J, Marshall BJ, Barkin JS, et al. Occupational exposure to Helicobacter pylori for the
endoscopy professional: A sera epidemical study. Am J Gastroenterol 1994;89:198792.
169. Wright RA. Upper-esophageal perforation with a flexible endoscope secondary to cervical
osteophytes. Dig Dis Sci 1980;25:668.
170. Bedford RA, van Stolk R, Sivak MV Jr, et al. Gastric perforation after endoscopic treatment of
a Dieulafoy's lesion. Am J Gastroenterol 1992;87:2447.
171. Scott B, Holmes G. Perforation from endoscopic small bowel biopsy. Gut 1993;34:1345.
172. Pettersson G, Larsson S, Gatzinsky P, Sudow G. Differentiated treatment of intrathoracic
oesophageal perforations. Scand J Thorac Cardiovasc Surg 1981;15:3214.
173. Miller RE, Bossart PW, Tiszenkel HI. Surgical management of complications of upper
gastrointestinal endoscopy and esophageal dilation including laser therapy. Am Surg
1987;53:66771.
174. Griffin SC, Desai J, Townsend ER, Fountain SW. Oesophageal resection after instrumental
perforation. Eur J Cardiothorac Surg 1990;4:2113.
175. Nashef SA, Pagliero KM. Instrumental perforation of the esophagus in benign disease. Ann
Thorac Surg 1987;44:3602.
176. Prinsley PR, Murrant NJ. Cervical esophageal perforation caused by diagnostic flexible
esophagoscopy. J Otolaryngol 1989;18:3146.
177. Moghissi K, Pender D. Instrumental perforations of the oesophagus and their management.
Thorax 1988;43:6426.
178. Panzini L, Burrell MI, Traube M. Instrumental esophageal perforation: Chest film findings. Am
179.
Schiff L, Stevens RJ, Goodman S. Pneumoperitoneum following the use of the flexible
gastroscope. Ann Intern Med 1941;14:12837.
180.
Wilson JA, Hass WR. Pneumoperitoneum with spontaneous recovery following use of flexible
gastroscope. Gastroenterology 1948;10:731.
181.
Asher LM, Cohen S. Gastroscopic perforation of esophagus and stomach: Report of three
cases. Gastroenterology 1949;12:9669.
182.
Villarreal R. Pneumoperitoneum following gastroscopy with spontaneous recovery: Report of
a case. Gastroenterology 1950;15:3645.
183.
Nelson RS. Pneumoperitoneum following gastroscopy with spontaneous recovery on
conservative therapy. Gastroenterology 1953;24:2679.
184.
Chamberlin DT. Pneumoperitoneum following gastroscopy apparently without perforation. N
Engl J Med 1947;237:845.
185.
Katz D, Selesnick S. Massive pneumoperitoneum and pneumoretroperitoneum after
gastroscopy: Report of a case and review of the literature. Am J Dig Dis 1956;1:51220.
186.
Fierst SM, Robinson HM, Lasagne L. Interstitial gastric emphysema following gastroscopy: Its
relation to the syndrome of pneumoperitoneum and generalized emphysema with no evident
perforation. Ann Intern Med 1951;34:120212.
187. Taub SJ, Nagurka C, Raskin JB. Tension pneumoperitoneum as a nontraumatic complication
of upper gastrointestinal endoscopy. Gastrointest Endosc 1980;26:1534.
188. Goldstein MJ, Sherlock P. Silent esophageal perforation following esophagoscopy.
189. Katz D, Cano R, Antonelle M. Benign air dissection of the esophagus and stomach at
fiberesophagoscopy. Gastrointest Endosc 1972;19:724..
190. Minocha A, Richards RJ. Pneumomediastinum as a complication of upper gastrointestinal
endoscopy. J Emerg Med 1991;9:3259.
191. Aston NO, Hougton AD. Jejunal perforation after esophagogastroscopy. Gastrointest Endosc
1989;35:5867.
192. Jeffrey GP, Reed WD. Non-instrumental small bowel perforation following upper
193. Brown MW, Brown RC, Orr G. Pneumoperitoneum complicating endoscopy in a patient with
duodenal and jejunal diverticula. Gastrointest Endosc 1986;32:1201.
194. Martin WC, Nagai N, Diner W, Texter EC Jr. Benign retroperitoneal emphysema associated
with upper gastrointestinal endoscopy. Arch Intern Med 1978;138:75961.
195. Rex DK, Hawes RH, Goulet RJ. Cecal perforation after upper endoscopy and pyloric dilation.
196. Weiner BC. Complications of routine diagnostic upper endoscopy. Gastrointest Endosc
1987;33:53.
197. Pope AJ, Adair HM. Spontaneous perforation of the esophagus during gastroscopy.
198. Gubbins GP, Nensey YM, Schubert TT, Batra SK. Barogenic perforation of the esophagus
distal to a stricture after endoscopy. J Clin Gastroenterol 1990;12:3102.
199. Girardi A, Piazza I, Giunta G, Pappagallo G. Retroperitoneal, mediastinal and subcutaneous
emphysema as a complication of routine upper gastrointestinal endoscopy. Endoscopy
1990;22:834.
200. Fireman Z, Bar-Ziv J, Lurie B. Retroperitoneal, mediastinal, and subcutaneous emphysema
as a complication of routine upper gastrointestinal endoscopy. Endoscopy 1994;26:32930.
201. Shiffman ML, Farrel MT, Yee YS. Risk of bleeding after endoscopic biopsy or polypectomy in
patients taking aspirin or other NSAIDs. Gastrointest Endosc 1994;40:45862.
202. O'Laughlin JC, Hoftiezer JW, Mahoney JP, Ivey KJ. Does aspirin prolong bleeding from
gastric biopsies in man?. Gastrointest Endosc 1981;27:15.
203. Buckley AB, Newman A. Methemoglobinemia occurring after the use of a 20% benzocaine
topical anesthetic prior to gastroscopy. Gastrointest Endosc 1987;33:4667.
204. Dinneen SF, Mohr DN, Fairbanks VF. Methemoglobinemia from topically applied anesthetic
spray. Mayo Clin Proc 1994;69:8868.
205. Patel D, Chopra S, Berman MD. Serious systemic toxicity resulting from use of tetracaine for
pharyngeal anesthesia in upper endoscopic procedures. Dig Dis Sci 1989;34:8824.
206. Zimmerman J, Rachmilewitz D. Systemic anaphylactic reaction following lidocaine
administration. Gastrointest Endosc 1985;31:4045.
207. el-Newihi HM, Mihas AA. Esophageal perforation as a complication of endoscopic overtube
insertion. Am J Gastroenterol 1994;89:9534.
208. Berkelhammer C, Madhav G, Lyon S, Roberts J. Pinch injury during overtube placement in
upper endoscopy. Gastrointest Endosc 1993;39:1868.
209. Baker RW, Spiro AH, Trnka YM. Mallory-Weiss tear complicating upper endoscopy: Case
reports and review of the literature. Gastroenterology 1982;82:1402.
210. Watts HD. Mallory-Weiss syndrome occurring as a complication of endoscopy. Gastrointest
Endosc 1976;22:1712.
211. Pelletier G, Nee N, Brivet M, et al. Upper gastrointestinal endoscopy. An unrecognized cause
of hyperamylasemia. Dig Dis Sci 1987;32:2546.
212. Kobayashi T, Fukuchi S, Sawano S, et al. Changes in serum isoamylase activities after
fibergastroduodenoscopy and colonoscopy. Isoamylase after FGDS and FCS. Endoscopy
1979;11:1337.
213. Shields HM, Soloway RD, Long WB, Weiss JB. Bilateral recurrent parotid gland swelling after
endoscopy. Gastroenterology 1977;73:1645.
214. Slaughter RL, Boyce HW Jr. Submaxillary salivary gland swelling developing during peroral
endoscopy. Gastroenterology 1969;57:838.
215. Morales TG, Hixson LJ. Acute pancreatitis following endoscopic biopsy of the ampulla in a
patient with Gardner's syndrome. Gastrointest Endosc 1994;40:3679.
216. Deschamps JP, Allemand H, Janin Magnificat R, et al. Acute pancreatitis following
gastrointestinal endoscopy without ampullary cannulation. Endoscopy 1982;14:1056..
217. Karjoo M, Luisiri A, Silberstein M, Kane RE. Duodenal hematoma and acute pancreatitis after
upper gastrointestinal endoscopy. Gastrointest Endosc 1994;40:4935.
218. Zinelis SA, Lee M, Hershenson MD, et al. Intramural duodenal hematoma following upper
gastrointestinal endoscopic biopsy. Dig Dis Sci 1989;34:28991.
219. Ghishan FK, Werner M, Vieira P, et al. Intramural duodenal hematoma: An unusual
complication of endoscopic small bowel biopsy. Am J Gastroenterol 1987;82:36870.
220. Barancik M. Inadvertent diagnosis of pheochromocytoma after endoscopic premedication. Dig
Dis Sci 1989;34:1368.

221. Crowe J, Bottomley K, McCarthy CF. Death following fiberoptic gastroscopy caused by
pheochromocytoma. Endoscopy 1977;9:589.
222. Archimandritis A, Tjivras M, Tryphonos M, et al. Changes in blood glucose induced by upper
GI diagnostic endoscopy: The influence of various premedications. Ital J Gastroenterol
1994;26:6971.
223. Dobronte Z, Barta M. Impacted duodenoscope removed with the aid of a second endoscope.
224. Kavin H, Schneider J. Impaction of a fibre-optic gastroscope in the oesophagus: An unusual
complication of gastroscopy. S Afr Med J 1970;44:4789.
225. Benson MJ, Rouse D, van Someren N, et al. Fatal hemorrhage from an aorto-esophageal
fistula precipitated by flexible endoscopy. Gastrointest Endosc 1991;37:1936.
226. Bynum TE, Smalley TK. Rupture of ascitic umbilical hernia during esophagoscopy: An
unusual indirect complication. Gastrointest Endosc 1970;17:678.
227. Christl SU, Scheppach W, Peters U, Kirchner T. Cerebral air embolism after
gastroduodenoscopy: Complication of a duodenocaval fistula. Gastrointest Endosc
1994;40:3768.
228. Kumashiro R, Kamachi H, Kumamoto M, Inutsuka S. Iatrogenic paraesophageal hernia after
229. Momsen OH. Hiatal hernia causing an unusual complication of gastroscopy. Endoscopy
1990;22:601.
230. Dehn TC, Lee EC. Intraperitoneal hemorrhage following fiberoptic gastroscopy. Gastrointest
Endosc 1985;31:350.
231. Pricolo R, Cipolletta L. Intraperitoneal hemorrhage following upper gastrointestinal endoscopy.
232.
Dawson J. Superior vena caval obstruction complicating fibreoptic gastroscopy. BMJ
1977;2:368.
233. Hillman LC, Davies WA, Clarke AC. Prolonged ileus following gastroscopyA case report.
Endoscopy 1980;12:946. hapte.
Chapter 38 Technique of Upper Gastrointestinal Endoscopy

(2771)
(2772)
JACQUES VAN DAM, M.D., PH.D.
AMITABH CHAK, M.D.
Upper gastrointestinal (GI) endoscopy is an essential part of the evaluation of GI tract pathology. This
chapter is concerned with two major topics: the preparation of the patient for
esophagogastroduodenoscopy (EGD), with particular emphasis on conscious sedation, and the
technical points of maneuvering an endoscope through the pharynx, esophagus, stomach, and
duodenum. The discussion on conscious sedation is appropriate to all endoscopic procedures;
therefore, this topic is not developed in detail in later chapters on colonoscopy and endoscopic
retrograde cholangiopancreatography (ERCP).
The ability to manipulate the instrument while performing EGD within a patient is essential but second
in importance to observation and recognition. As a trainee learns to perform EGD, the ability to
maneuver the endoscope is almost always attained before the capacity for endoscopic diagnosis. The
thought processes of the novice are mainly concerned with moving the instrument from one point to
another. Thus preoccupied, the beginner frequently does not observe all that may be seen. In contrast,
the required maneuvers are second nature to the expert; they are performed in a natural, flowing
manner that requires minimum conscious attention. The expert thus becomes more cognizant of what
can be seen. The word see can have many connotations. It is never enough for an endoscopist simply
to collect observations in a mechanistic manner. Rather, observations must be synthesized with a
clinical problem and compared with prior experience. Because many disorders affect the upper GI
tract, there is much to be seen. Experience, therefore, has an important place in upper GI endoscopy.
Upper GI endoscopy extends the sense of sight. It is, in a very real sense, physical diagnosis for the
gastroenterologist and the GI surgeon.
The indications, contraindications, and complications of endoscopy are discussed elsewhere (see
Chapter 37: Indications, Contraindications, and Complications of Upper Gastrointestinal Endoscopy).
Endoscopic diagnosis, or the things to be "seen," are discussed throughout this book.
Preparation for EGD

Every endoscopic procedure includes preprocedure assessment, patient preparation, patient consent,
conscious sedation, performance of the procedure, patient recovery, and follow-up.
Assessment of the Patient

GI endoscopy has value in many clinical situations. It is impractical to require that every patient with an
indication for endoscopy be completely evaluated by the endoscopist who performs the procedure. In
fact, a large percentage of upper GI endoscopies are performed by an endoscopist on patients who
are directly referred by another physician specifically for the procedure.1,2(2773) Nevertheless, the
indications for the procedure must be well defined and carefully reviewed by the endoscopist, who
must also be satisfied that no contraindications to the procedure exist. These aspects of preparation
for EGD can be problematic if communication between the referring physician and endoscopist is less
than adequate.
To some degree, the endoscopist must rely on the thoroughness and ability of the referring physician
in determining the fitness of a patient for a procedure. This is less difficult when the endoscopist has an
established professional relationship with the referring physician, such as exists in a closed group
practice or health care system in which complete case records are available to the endoscopist. It is
more problematic when no relationship exists with a referring physician. In addition to establishing the
indication for the procedure, the endoscopist must verify that a patient has no associated illnesses that
may compromise the safety of the procedure. This often requires a more formal evaluation of the
patient prior to the procedure; the focus should be on disorders that have a direct bearing on the
procedure, such as a history or evidence of cardiovascular illness and respiratory disorders. A
complete medication history, including drug allergies, should be obtained to avoid drug interactions.
Furthermore, many drugs have adverse or beneficial effects on the GI tract and may establish or alter
the endoscopic findings.
Preprocedure Instruction
Patients should be advised in advance of the EGD that they will be unable to drive, operate machinery,
or perform intricate mental or psychomotor tasks for a number of hours afterward and that it is
necessary for another person to accompany them after the procedure. The patient must be told the
reason for the procedure, given a brief description of EGD, and apprised of the remote possibility of a
complication, including the most common untoward events. This process, usually referred to as
"informed consent," has ethical and legal implications.3(2774) Informed consent is the subject of an
excellent review by Plumeri.4(2775)
Although it is incumbent upon the endoscopist to provide patients with an adequate explanation of a
procedure, including indications and risks, Levy et al.5(2776) found that an excessively detailed
description of EGD did not reduce patient anxiety. In this prospective study, 243 consecutive patients
undergoing EGD were randomized to four groups and received either a brief description of the
procedure by the referring physician, a detailed description by the endoscopist, a comprehensive
explanation that included pictures showing each phase of the procedure, or a descriptive videotape
presentation. No significant differences were found in patient anxiety, as measured by the Spielberger
State and Trait Anxiety Scales, between any of the groups. No differences in anxiety level were found
with respect to age or ethnic background, although women scored significantly higher than men.
An informed, motivated, and relaxed patient generally allows the procedure to be performed in an
equable manner that assures safety and thoroughness.6(2777) Although a written description of the
procedure is often advantageous, it remains necessary to review the important points with the patient.
Because of the lingering effects of sedative drugs, patients undergoing EGD should also be given
postprocedure instructions before the procedure, including directions in the event of a delayed
complication. Written instructions are again advantageous.
Preparation
General
Patients must fast for at least 6 hr prior to EGD. When delayed or defective gastric emptying is a
problem, the required period of fasting may be longer, and in some cases aspiration of gastric contents
may be necessary before EGD is performed. Dentures and eyeglasses should be removed. It is
advisable to establish intravenous access for all endoscopic procedures, even those in which
conscious sedation is not used. In the prospective study of Smith et al.,7(2778) a 23-gauge Teflon
cannula provided more reliable access than a 23-gauge winged steel needle.
Other steps in the process of preparing the patient, the procedure room, and endoscopic equipment for
EGD are discussed in Section 1 (Chapter 4: The Endoscopy Unit; Chapter 5: The Gastrointestinal
Assistant; and Chapter 8: Disinfection of Endoscopes and Accessories).
Endoscopic Instruments
The Upper Gastrointestinal Panendoscope
Routine diagnostic EGD may be performed with fiberoptic or electronic (video) forward-viewing
endoscopes (see Chapter 2: Flexible Endoscope Technology: The Fiberoptic Endoscope; and Chapter
3: Flexible Endoscope Technology: The Video Image Endoscope). The working length of these
instrumentsthat is, the length of the insertion tubeis about 1 m. Four-way deflection of the distal tip
with at least 180 and 90 degrees of upward and downward deflection, respectively, and 90 degrees
each of leftward and rightward deflection is necessary for a thorough examination. Independently
locking mechanisms for the deflection controls maintain tip angulation during certain standard
endoscopic maneuvers. Diagnostic endoscopes are equipped with valves that enable the endoscopist
to insufflate the GI tract with air, clean the objective lens with water, or aspirate GI contents via built-in
channels. An accessory channel with a diameter of at least 2.0 mm (preferably larger) is essential for
obtaining mucosal biopsies and for passing therapeutic devices into the endoscopic field. Modern
endoscopes are fluid-tight and are therefore immersible for thorough cleaning and disinfection. All
endoscopes can also be coupled to cameras and video recorders for documentation of findings.
Electronic (video) endoscopy systems have replaced older fiberoptic instrument technology in many
endoscopy units. Because a fiberoptic instrument must be held up to the endoscopist's eye,
endoscopic procedures done with this type of instrument result in eye strain and musculoskeletal
stress on the arms, neck, and shoulders. An added risk is that the endoscopist may be splashed in the
eye or mouth by body fluids that back up in the accessory channel, as may occur with the passage of
accessories through the channel. Electronic endoscopes can be held in a more natural and
comfortable midbody position, thereby minimizing physical strain on the neck and shoulders. Viewing
an enlarged video image with both eyes is much easier than looking at the objective of a fiberoptic
endoscope with one eye. Display of endoscopic images on a video monitor also allows the endoscopy
assistant and others to participate in the procedure more easily. Because the accessory channel of the
electronic endoscope is held away from the endoscopist's face, the risk of being splashed by body
fluids is markedly reduced. Electronic systems are also better suited for endoscopic instruction and
have become the standard in most teaching institutions in the United States. The conversion to
electronic endoscopy has also been propelled by the fact that digitized images may be captured and
stored by computers (see Chapter 7: Electronic Image Management).
A general impression exists and is probably correct that smaller-diameter endoscopes are easier to
pass and are better tolerated by patients. However, the ideal diameter is difficult to determine.
Theoretically, reduction in insertion tube diameter could improve patient tolerance, even to the point
that no sedation would be needed. Thus, diagnostic EGD with ultrathin endoscopes (<8 mm diameter)
and without sedation has been proposed. Indeed, the feasibility of unsedated diagnostic panendoscopy
with an ultrathin endoscope using a transnasal route has been demonstrated.8(2779) However, patient
acceptance of endoscopy performed with an instrument 8 mm or less in diameter has not been
demonstrated to be any better than for procedures performed with an instrument 10 mm in diameter.
Furthermore, many features of the standard panendoscope, such as the presence of an accessory
channel, must be sacrificed or compromised in exchange for thinness. In terms of patient tolerance
and ease of the examination, the differences between an endoscope 10 mm in diameter and one of
smaller caliber are not great enough to justify the loss of standard features.
A wide variety of upper GI endoscope models, all with the basic features described earlier, are offered
by endoscope manufacturers. The general design of all of these instruments is the same. However,
subtle differences exist in the handling characteristics of the insertion tube. Some have greater torque
and longitudinal stability than others. Minor differences in viewing angle, deflection capability, and
accessory channel diameter are seen. The design of the control section also differs slightly among the
various manufacturers. Although these minor variations are not markedly important with respect to the
adequacy and safety of EGD, one tends to become accustomed to the design of a particular company.
Changing to an instrument made by another manufacturer usually entails a period of adjustment and
can be distracting. Minor differences in the stiffness characteristics of the insertion tube, the shape of
the deflection controls and the control section itself, the position of the accessory channel valve, the
weight, the location of the air/water and suction valves, the length of the bending section of the
insertion tube, the tightness of the deflection angle, and the like may be annoying at the least and may
even necessitate a change in examination technique. It is advisable, therefore, to use endoscopes
made by a single company.
The Side-Viewing Endoscope
A thorough examination of the entire upper GI tract may not be possible with a forward-viewing
endoscope. Because of anatomic features, certain regions cannot be viewed well or may not be seen
at all. These "blind areas" include the medial wall of the descending duodenum, the area of the
duodenal bulb immediately beyond the pylorus, and the region of the small bowel just beyond a
gastroenterostomy stoma.
Approximately 80 to 90% of the duodenal bulb can be seen with a forward-viewing endoscope. Even
though the recesses of the bulb beyond the pylorus may be difficult to view, a lesion in this area usually
produces other secondary findings that may give a clue to its presence. An ulcer in this hidden area,
for example, may produce edema, spasm, and deformity of the pylorus as well as erythema of the
bulbar mucosa, even though the crater itself cannot be seen.
The medial duodenal wall, including the duodenal papilla, is also a blind area. This region can be
viewed only tangentially with a forward-viewing endoscope, so that a lesion in this area, particularly one
involving the main duodenal papilla, may be easily overlooked. An instrument with side-viewing optics
provides a more satisfactory view of the medial aspect of the descending duodenum, and for this
reason it is sometimes necessary to use both types of instruments to accomplish a thorough and
complete examination of the upper GI tract. Techniques for use of this type of instrument are described
in Chapter 57: Technique of Endoscopic Retrograde Cholangiopancreatography.
Care and Handling
GI endoscopes are both useful and expensive. With proper care, an endoscope can be used to
perform hundreds, even thousands, of procedures. Just as there are correct methods for cleaning and
storage, there are also proper techniques for handling the instrument. The distal deflecting section of
the insertion tube should not be bent with the hand. Serious damage can result if the distal tip strikes a
hard surface. The insertion tube should never swing about freely when the instrument is being
transported. A straight, outward pull should be used to disconnect the endoscope from the light source.
A rocking or twisting motion during disconnection throws the connecting pins out of alignment and may
result in even more serious damage. The video processor should be turned off before the electronic
cable is disconnected to avoid damage to the electronics. Use of excessive force while connecting or
disconnecting the video processor cable from an electronic endoscope can bend the metal connecting
pins out of alignment. Water should be kept away from electronic equipment at all times.
An endoscope should always be checked to ensure that it is in working order before use. With the
instrument tip held 3 or 4 cm from an object, a quick view through the objective lens of a fiberscope or
at the video screen of the electronic system reveals any clouding or streaking in the image that may
indicate improper cleaning or leakage of water into the insertion tube as a result of damage or
mishandling.
Electronic endoscopes must be "white balanced" so that the system displays true color. The instrument
tip should be deflected through its full range of motion, using the deflection controls to determine not
only that the controls are functional but also that the degree of tip angulation in each direction is
reasonably close to that specified by the manufacturer. Over the course of time, the cables that control
tip deflection become stretched so that the degree of tip angulation decreases. This loss of function
may compromise the endoscopist's ability to maneuver the instrument in subtle but significant ways.
The distal end of the insertion tube may be held in a small container of clean water to check the
function of the air/water and suction systems. The accessory channel may be lubricated by passing a
biopsy forceps through the channel, placing a drop of silicone lubricant on the end of the forceps after it
exits the distal end of the instrument, and then withdrawing the forceps back through the channel. This
maneuver lubricates the channel within the distal bending section of the insertion tube, where an
accessory device passing through the channel encounters the greatest friction. Several kinds of
accessory channel valves are available, many of them diaphragm-type devices. The competence of
the valve should be checked because repeated passage of accessories wears out the diaphragm,
allowing fluid to leak. These valves should be replaced periodically.
Other steps preliminary to the actual EGD, such as positioning the patient on the procedure table in the
left lateral decubitus position, are discussed in Chapter 5: The Gastrointestinal Assistant.
Premedication
Wide variations are found throughout the world in the approach to patients who are about to undergo
GI endoscopy, particularly in the use of sedative and analgesic drugs. These differences reflect
availability of these agents, prevailing opinion among endoscopists concerning their use, and cultural
dissimilarities among nations.
Three basic types of medications are commonly used for EGD: topical anesthetics; sedatives,
especially the benzodiazepines; and narcotic analgesic agents. The drugs most widely used for
sedation are midazolam, alone or in combination with meperidine (also known as pethidine in some
countries) or fentanyl. Upper GI endoscopy can be performed without pharmacologic preparation in
highly motivated patients, especially those who are fearful of the effects of sedative drugs. Individual
factors such as personality, prior experience, and the presence of an escort influence levels of anxiety
and tolerance of endoscopic procedures.9,10(2780) In clinical situations of an emergency nature, it
may be advisable to avoid the use of drugs that depress respiration or to use such agents in small
doses.
Endoscopy without the use of sedative drugs has certain practical advantages: The risk of a
complication is reduced; the time and cost of administering the drug are eliminated; virtually no time is
required for postprocedure recovery; and postprocedure monitoring is minimized. Indeed, avoidance of
conscious sedation in favor of topical pharyngeal anesthesia alone may not only greatly reduce the
morbidity and mortality of GI endoscopy but may also protect the endoscope and result in less damage
to instruments.11(2781) The use of endoscopes with smaller-diameter insertion tubes facilitates upper
GI endoscopy without sedation; those with diameters of 1.0 cm or less are more easily passed and
better tolerated by patients.
The benefit of premedication for EGD remains controversial in many parts of the world. Although
benzodiazepines are generally preferred for sedation in Western countries, GI endoscopy is performed
routinely in many regions of Asia using only topical pharyngeal anesthesia.12(2782) Evaluation of
patient attitudes in Sweden demonstrated a desire to avoid conscious sedation prior to GI
endoscopy.13(2783) However, the results of a survey of more than 500 members of the American
Society for Gastrointestinal Endoscopy (ASGE) show that only 2.2% of endoscopists in the United
States do not routinely sedate patients for upper GI endoscopy.14(2784) Similarly, the results of a
postal questionnaire in the United Kingdom suggest that 2% of respondents do not use sedation for GI
endoscopy.15(2785)
Despite the advantages of eliminating premedication, GI endoscopy is nevertheless better tolerated by
patients if some form of sedation is administered.16,17(2786) The debate over the use of
premedication may be best suited to the academician because the reputation of the practicing
gastroenterologist depends in no small measure on the comfort and tolerance of patients during
procedures.18(2787)
The dose of any sedative drug(s) should be the minimum required to achieve successful endoscopy
while maintaining patient comfort and acceptance.19,20(2788) Drowsiness, slurred speech, and
physical relaxation have generally been assumed to correlate closely with a patient's perception of
comfort during an endoscopic procedure. However, the adequacy of sedation as estimated by the
endoscopist or GI assistant correlates poorly with patient satisfaction.21(2789) The best method of
avoiding complications is to carefully titrate the sedative dose according to the response of the
individual patient and to administer the drug(s) slowly to avoid hemodynamic and respiratory side
effects.22(2790)
Topical Pharyngeal Anesthesia
Although opinions diverge concerning the use of local anesthetic agents for EGD, these drugs are
widely used in practice. A nationwide survey in the United Kingdom showed that 63% of endoscopists
spray the patient's oropharynx with a local anesthetic prior to most upper GI endoscopic
procedures.15(2791)
A number of prospective studies of topical pharyngeal anesthesia for upper GI endoscopy have been
done. The results of most but not all studies indicate that topical pharyngeal anesthesia offers
advantages in terms of patient tolerance. Any assessment of tolerance for an endoscopic
procedurewhether determined by the patient, the endoscopist, an assistant, or an observer not
directly involved in the procedureis almost entirely subjective. Furthermore, tolerance is influenced
by a number of factors in addition to the use of medications.
Cantor and Baldridge23(2792) found no benefit for topical pharyngeal anesthesia in a randomized trial
in which patients received either viscous lidocaine gargle or placebo, or neither the drug nor the
placebo. The double-blind placebo-controlled trial of Chuah et al.24(2793) also found no benefit in
terms of ease of intubation and patient comfort. In a similar study of 150 patients, Lachter et
al.25(2794) found no differences with regard to coughing, gagging, or difficulty of intubation when use
of a topical anesthetic spray (Cetacaine) was compared with use of a placebo, except that intubation
was thought by endoscopists to be easier in patients who were undergoing endoscopy for the first time.
Gordon et al.26(2795) found that patients preferred topical anesthesia to placebo in a double-blind
randomized trial in 111 consecutive patients undergoing EGD; endoscopists also rated patient
tolerance as significantly better. In a randomized, double-blind, placebo-controlled trial in the United

Kingdom, the acceptability of upper GI endoscopy was significantly greater (p = .001) in those patients
who received lidocaine topical spray (50 mg) than in those who received a placebo.27(2796) In a
prospective study in Sweden, 200 ambulatory patients undergoing EGD without sedation were
randomized to receive either topical anesthesia or placebo.28(2797) Although no significant difference
in throat discomfort was found between the two groups, patients preferred, when given a choice, that
any subsequent endoscopic procedures be performed with topical anesthesia.28(2798) In a similar
study from Greece, 140 consecutive patients undergoing EGD were randomized to receive one of the
following regimens: (1) diazepam, 10 mg intramuscularly (IM) 30 min prior to endoscopy, and 10%
lidocaine spray for topical pharyngeal anesthesia, (2) 10 mg diazepam IM, (3) 10% lidocaine spray, or
(4) no premedication.29(2799) Upper GI endoscopy was well tolerated without sedation, and no
statistical difference was found between the groups. However, a tendency toward better tolerance was
seen in those patients who received topical pharyngeal anesthesia.29(2800) Both lowdose midazolam
and lidocaine spray had an additive beneficial effect on patient tolerance in another study of
comparable design from Switzerland of 200 patients undergoing diagnostic EGD on an ambulatory
basis.17(2801) A prospective evaluation of 2000 upper GI endoscopic examinations performed without
sedation in the Middle East concluded that topical pharyngeal anesthesia alone (10% lidocaine spray)
resulted in a safe, quick, and well-tolerated procedure.30(2802) Thus, a documented basis exists for
the common belief that patient tolerance of endoscopy is improved by the use of topical
analgesia.31(2803)
Topical pharyngeal anesthesia may be obtained by administration of one of several different agents by
spray, gargle, painting, or lozenge. Commonly used topical anesthetic agents include tetracaine
(Pontocaine), tetracaine plus benzocaine (Cetacaine), benzocaine (Hurricaine), and lidocaine
(Xylocaine). Smith et al.32(2804) compared preferences for three topical anesthetic sprays (tetracaine
plus benzocaine, benzocaine, and 10% lidocaine spray) and three gargles (2% lidocaine, and two
combinations of 2% lidocaine diluted 1:1 with mouthwash) in a randomized study of normal subjects
who had previously undergone upper GI endoscopy. Subjects underwent repeated procedures without
additional premedication and then ranked the agents in order of preference. Although individual
subjects had strong preferences, no consensus of opinion was reached as to a preferred agent with
regard to taste, degree of pharyngeal anesthesia, and tolerance for passage of the endoscope.
However, most subjects preferred administration by spray rather than gargling.
The optimum dose of topical pharyngeal anesthesia has been evaluated prospectively. Jameson et
al.33(2805) randomly evaluated three different doses of lignocaine spray (50 mg, 100 mg, or 200 mg)
in 60 consecutive patients undergoing EGD. Lignocaine spray in the 100-mg dosage improved patient
tolerance for upper GI endoscopy compared with the 50-mg dose; no added benefit was achieved with
the 200-mg dose.33(2806) Optimum dosing may be of clinical importance because topical anesthetic
agents are known to be absorbed in some degree into the systemic circulation, although their potential
for toxicity is rarely emphasized.34(2807) In 20 patents who received topical anesthesia prior to upper
GI endoscopy, serum concentrations of lidocaine and its metabolite monomethylglycinexylidide were
noted to be lower in those who received lidocaine as a 2% gel than in those who gargled with a 2%
lidocaine solution.35(2808) Although no untoward effects were observed in either group, the authors
recommended the gel form of lidocaine to minimize systemic absorption.
Complications related to use of topical anesthetic agents are rare. Patel et al.36(2809) reported several
serious systemic toxic effects, including one fatality, resulting from the use of tetracaine for pharyngeal
anesthesia in upper GI endoscopy. One case of systemic anaphylaxis is attributed to administration of
a lidocaine gargle prior to EGD.37(2810) The administration of topical benzocaine prior to endoscopy
has been reported to cause methemoglobinemia.3840(2811) This potentially life-threatening
complication has also occurred with the use of other topical anesthetic agents in sensitive
patients.40(2812) In one instance, topical benzocaine resulted in near-fatal methemoglobinemia in a

patient with preexisting NADH methemoglobin reductase deficiency.41(2813) Deficiency of this
enzyme is inherited and occurs with increased frequency in Inuit and Alaskan Native
Americans.42(2814) In other reported cases, a preexisting enzyme deficiency was not
noted.43,44(2815)
Cyanosis occurring immediately after the administration of topical anesthetics should be evaluated by
arterial blood gas measurements. If the measured PaO2 appears greater than clinically expected for
the degree of cyanosis, methemoglobinemia should be suspected. Methemoglobinemia is formed

when iron in the hemoglobin molecule becomes oxidized from the ferrous to the ferric state. In this
form, the molecule cannot bind oxygen. Treatment is infusion of methylene blue (1 mg/kg) over 10 min.
Although supplemental oxygen administration raises PaO2 in the arterial blood, it does not increase
oxygen binding or delivery to the tissue. Patients should be monitored in a critical care setting with
continuous pulse oximetry. Although it is uncommon, the endoscopist should be aware of the
possibility of systemic toxicity secondary to topical pharyngeal anesthesia.
Benzodiazepines
Diazepam (Valium; Roche Laboratories), a lipid-soluble benzodiazepine with active metabolites, is the
prototype sedative medication for GI endoscopy. The distribution half-life of diazepam is 30 to 60 min;
the plasma elimination half-life is 24 to 57 hr, considerably longer than the time required for upper GI
endoscopy.45(2816) Diazepam is metabolized either by demethylation to N-desmethyl diazepam or by
oxidation to 3-hydroxydiazepam, both of which may be further metabolized to the pharmacologically
active oxazepam.19(2817) Partial or complete amnesia for the endoscopic procedure is observed in 16
to 53% of patients premedicated with diazepam.46,47(2818) Longstreth et al.48(2819) found that a
significantly higher dosage of diazepam, mean 0.48 mg/kg, was required for adequate sedation in
patients who used two or more doses of a benzodiazepine per week than in patients who did not take
this type of drug on a regular basis; they required a mean dosage of 0.30 mg/kg.
Conventional preparations of diazepam contain solvents such as propylene glycol, phenylcarbinol,
ethanol, and sodium benzoate, which may cause pain at the site of injection and
thrombophlebitis.49(2820) Diazemuls, an alternate pharmacologic formulation of diazepam, is devoid
of solvents, containing instead a soybean oil-water emulsion designed to eliminate irritation at the
injection site.50(2821) When evaluated in a prospective, randomized, cross-over study, there were no
statistically significant differences between the two forms of diazepam with regard to degree of
sedation, patient compliance, or recovery of psychomotor function.50(2822) The short-term amnestic
effect of diazemuls mimics closely that of diazepam. However, the potency of diazemuls appears to be
less than that of diazepam.
Midazolam hydrochloride (Versed; Roche Laboratories) is a water-soluble, ultra-short-acting imidazo
derivative benzodiazepine first marketed in the United States in 1986. The metabolites of midazolam
are metabolized even faster than the parent compound and have no sedative properties.45(2823) The
drug undergoes oxidation to form hydroxymidazolam, which is also active but is rapidly conjugated to
the pharmacologically inactive glucuronide, which is excreted primarily in the urine.19,51(2824) The
distribution half-life is approximately one quarter that of diazepam; the plasma elimination half-life is
approximately one tenth that of diazepam.45(2825) Obesity, cirrhosis, and advanced age increase the
half-life of midazolam.52,53(2826)
Diazepam and midazolam have both been found to be satisfactory for conscious sedation for upper GI
endoscopy in comparative trials, although some important differences exist between the two
agents.5460(2827) The degree of amnesia for the procedure has been noted to be more substantial
for midazolam in prospective comparative trials.5662(2828) The degree of amnesia is apparently
dose related, being more profound with higher doses of the drug. Gilvarry et al.63(2829) evaluated
amnesia for upper GI endoscopy as evidence of drug efficacy and compared intravenously
administered diazepam (20 mg in patients less than 65 years of age; 10 mg in patients of greater age)
with intravenously administered midazolam (10 mg in patients less than 65 years of age; 5 mg in
patients more than 65 years old). Amnesia occurred significantly more frequently following
administration of midazolam, with 90% of patients having no recall for the procedure, compared with
63% of those who had received diazepam (p < .05).
Based on its pharmacologic properties, recovery of psychomotor functions after conscious sedation
with midazolam should be faster, as has been noted in several comparative trials.57,60(2830) In the
study of Sanders et al.,57(2831) side effects of diazepam (0.15 mg/kg) were noted up to 30 hr after
intravenous administration, although psychomotor functions were significantly impaired for up to 4 hr
after administration of midazolam (0.07 mg/kg). In other studies, no difference in recovery time was
seen for psychomotor functions;54,58(2832) in one trial, recovery time was found to be shorter with
diazepam despite its longer pharmacologic half-life.55(2833) Compared with diazepam, midazolam is
associated with substantially less pain and fewer venous complications at intravenous injection
sites.54,56,60(2834) In these respects, diazemuls appears to be equivalent to midazolam.58(2835)
Diazepam and midazolam were compared with respect to rates of serious cardiorespiratory
complications and death when used for GI endoscopy by evaluating data on more than 21,000
procedures in the computer-based management system data bank of the ASGE.64(2836) The rate for
serious cardiorespiratory complications was 5.4 per 1000 procedures; the mortality rate associated
with the use of these agents was 0.3 per 1000 procedures. No difference was found in the risk for a
cardiorespiratory complication when use of the two agents was compared, although the concomitant
use of a narcotic agent and the performance of a procedure on an urgent basis did increase this
risk.64(2837)
As noted earlier, the various benzodiazepine compounds may have substantially different
pharmacologic effects. Significant differences were observed in various physiologic and psychomotor
parameters in a randomized, double-blind trial in which patients were sedated for oral surgery with
intravenous lorazepam (0.05 mg/kg), diazepam (0.25 mg/kg), or midazolam (0.1 mg/kg).65(2838) A
number of effects were observed with sedation, including slurred speech, giddiness, dizziness,
diplopia, ptosis, and bradycardia. All three drugs produced a satisfactory degree of sedation, and all
initially lowered blood pressure. Lorazepam had a lesser effect on blood pressure than diazepam,
which in turn produced less hypotension than midazolam. However, the recovery period was longer
with lorazepam and was associated with more "giddiness and dizziness" than the other drugs.65(2839)
Such drug effects, if confirmed by additional large-scale studies, confirm that patient response to
sedative agents is variable and that the administration of these drugs should be individualized.
The ideal medication for conscious sedation should have predictable clinical effects in patients of all
ages and diverse medical infirmities, minimum cardiovascular side effects, rapid and reliable onset of
action, and rapid postprocedure recovery. Such a drug does not yet exist. Newer benzodiazepine
derivatives with potentially safer pharmacologic profiles are currently under investigation. When
flunitrazepam, a water-soluble benzodiazepine, was compared with midazolam as a premedication for
EGD, the former drug was found to produce significantly less increase in heart rate and decrease in
systolic blood pressure; however, its postprocedure recovery time was significantly longer than that for
midazolam.66(2840) Such drug effects further suggest that premedication eventually may be
individualized such that inpatients with cardiovascular disease may receive one drug (e.g.,
flunitrazepam) and outpatients without severe cardiopulmonary limitations may receive another (e.g.,
midazolam).
The speed with which benzodiazepines are intravenously administeredbolus administration (injection
over 5 sec) versus slow titration (injection over 2 to 3 min)also changes the pharmacodynamics of
the drug. The results of a nationwide survey in Great Britain disclosed that adverse results were
reported more frequently by endoscopists who "titrated" the sedative dose than by those who gave
intravenous sedation by bolus administration.15(2841) Specifically, episodes of severe hypoventilation
were reported by 47% of those who titrated the dosage of the sedative drug, compared with 37% of
those who gave a bolus injection. The number of prospective studies evaluating speed of
administration is small.
Bell et al.67(2842) noted that the theoretical advantage of rapid bolus injection of midazolam relates to
its more rapid onset, in which case the endoscopist can begin the procedure sooner than when the
drug is administered more slowly. A further advantage of bolus injection is that the patient generally
requires smaller doses of the drug, usually less than two thirds the dose of midazolam required when a
slowly titrated injection is given.67(2843) The theoretical disadvantage of bolus injection is the
increased risk of drug-induced respiratory depression, an effect that would be enhanced by giving the
benzodiazepine in conjunction with an opiate. Bell et al.67(2844) tested these theories prospectively in
more than 130 patients undergoing upper GI endoscopy using midazolam given as a bolus injection
over 5 sec. Patients were monitored for oxygen saturation using a pulse oximeter. One group of 54
patients was given supplemental oxygen via nasal cannula at a rate of 3 L/min, and the other group of
77 patients received oxygen only if their oxygen saturation dropped below 85%. Sedation and
antegrade amnesia were similar after both bolus injection and titration of midazolam, although the
titrated doses were greater. Despite the fact that the dose of mid-azolam given by intravenous bolus
injection was only two thirds that given by slow titration, the degree of oxygen desaturation during the
endoscopic procedure was greater and the ability of supplemental oxygen to prevent desaturation was
less in patients who received a bolus injection (p < .01).67(2845)
Swain et al.68(2846) compared bolus intravenous injection of 10 mg diazepam (given over 2 to 4 sec)
with titration of the dosage of diazepam (mean dosage, 15.9 mg given over 1 to 2 min) in a randomized
trial that included 100 patients. The level of patient cooperation during endoscopy was similar for both
groups. Respiratory depression occurred in one patient who received a bolus injection. At 30 min after
injection, psychomotor functions were significantly more impaired in patients who were sedated by
titration; the sedative effects of the drug also persisted for longer periods of time in patients sedated in
this manner. However, partial or complete amnesia for the procedure was significantly more frequent in
patients who were sedated according to the titration protocol. Although prevention of oxygen
desaturation, as observed by pulse oximetry, has not yet been shown to decrease morbidity and
mortality in patients undergoing EGD, the slow titration method of administering a benzodiazepine
appears to be safer than rapid intravenous injection of these drugs.
Oral administration of benzodiazepines has been studied as an alternative to intravenous injection.
This route of administration reduces cost and may shorten the recovery time after upper GI endoscopy.
Hedenbro et al.69(2847) evaluated 359 outpatients randomized to receive either triazolam, 0.125 mg
by mouth, or placebo. Triazolam reduced patient discomfort compared with placebo, but no
demonstrable difference in amnesia for the procedure was found. However, in view of the substantial
individual variation in response to benzodiazepines and other sedative drugs, it is unlikely that all
ambulatory patients are suitable candidates for oral premedication. Further evaluation of this route of
sedative administration is required before any definite conclusions can be drawn regarding efficacy and
safety.
Complications of benzodiazepines include the so-called paradoxical reaction. The expected results of
benzodiazepine administration are anxiolysis, amnesia, and mild sedation. If a patient responds to the
medication with agitation and anxiety, the response is termed paradoxical.70(2848) A patient was
reported to become delirious after intravenous midazolam (3 mg given in 1-mg increments), uttering
nonsensical words and requiring physical restraint by endoscopy personnel.70(2849) The response
was easily reversed with flumazenil, 0.2 mg given intravenously (see later). Potential venous
complications of benzodiazepine administration include phlebitis (pain), thrombophlebitis (pain with
induration), and thrombosis (induration without pain).71(2850) When compared with diazepam
prospectively, midazolam produced significantly less pain at the injection site in patients undergoing
endoscopic procedures.71(2851)
Cardiorespiratory complications associated with administration of benzodiazepines as premedication
for upper GI endoscopy are discussed in Chapter 37: Indications, Contraindications, and Complications
of Upper Gastrointestinal Endoscopy.
Specific benzodiazepine antagonists are widely available and should be on hand in every endoscopy
unit.20(2852) However, the availability of such antagonists should not encourage the misuse or
overdosage of the benzodiazepine agonist. Flumazenil (Mazicon; Roche Laboratories) was introduced
in 1981 as an imidazobenzodiazepine that blocks the central effects of benzodiazepine agonists by
competitive interaction at the benzodiazepine receptor site. The drug is bound more than 50% to
plasma protein, but its elimination half-life is relatively short, approximately 1 hr.72(2853) Flumazenil
antagonizes the anxiolytic, muscle relaxant, sedative, ataxic, anticonvulsant, amnestic, anesthetic, and
respiratory-depressive effects of benzodiazepines.72(2854) Flumazenil may require several minutes to
reverse fully an adverse reaction (e.g., respiratory depression) caused by the benzodiazepine
agonist.73(2855)
Flumazenil has been used to reverse conscious sedation in patients receiving midazolam or
diazepam.74,75(2856) Well-designed controlled clinical trials have shown that flumazenil significantly
reduces postprocedure recovery time after midazolam administration without reducing anterograde
amnesia.7679(2857) It has been suggested that patients can be safely discharged after reversal of
conscious sedation with flumazenil, provided that they are given detailed instructions.45,79,80(2858)
However, not all endoscopists agree that the benefit of prompt discharge is worth the cost of this
relatively expensive drug, and they suggest that benzodiazepine reversal be used only in
life-threatening situations.81(2859) Further studies of the long-term effects and duration of
benzodiazepine reversal are needed before any definitive conclusions can be drawn regarding the
efficacy and safety of this benzodiazepine antagonist.
In the study of Sanders et al.,82(2860) 120 ambulatory patients sedated with either diazepam (0.2
mg/kg) or midazolam (0.1 mg/kg) were randomized to receive either flumazenil (0.1 to 2 mg) or
placebo after EGD. Although flumazenil diminished the sedative effects of both benzodiazepines,
attenuation was not complete. Return of psychomotor function occurred significantly more rapidly in
patients who had been given midazolam. Sanders et al.82(2861) concluded that flumazenil could not
be used within acceptable margins of safety to expedite the discharge of patients.
Pearson et al.83(2862) assessed sedation, psychomotor skills, and amnesia after administration of
flumazenil or placebo in patients sedated for EGD with midazolam. Recovery from sedation and loss of
psychomotor functions was significantly more rapid in patients who received flumazenil, and no
difference was seen between the two groups of patients with respect to amnesia for the procedure. At
18 and 24 hr after administration of flumazenil, no recurrence of sedation was seen. Similar results
were noted in the double-blind, randomized trial of Jensen et al.80(2863) with respect to attenuation of
the sedative effects of midazolam and diazepam, but anterograde amnesia was also effectively
antagonized. These investigators concluded that flumazenil could be used to reduce patient recovery
time after sedation. In the study of Benjamin et al.,84(2864) most patients (85%) required only one or
two doses and remained alert during a 2-hr observation period. However, other studies suggest that
benzodiazepine reversal with flumazenil persists for only 2 hr, after which time some loss of effect may
occur.85,86(2865)
Narcotic Agents
Meperidine is frequently used for sedation for endoscopic procedures. It is generally administered
intravenously and is often used in combination with a benzodiazepine. Although elderly patients may be
more sensitive to the effects of meperidine, the half-life and clearance are unaffected by age.87(2866)
Fentanyl, a short-acting narcotic, also has been effective in producing sedation for upper GI
endoscopy.88(2867) When compared with a control group of patients receiving no narcotic analgesia,
patients receiving fentanyl reported a favorable sedative effect for EGD, and fentanyl administration
was not associated with a significant difference in the measured arterial oxygen saturation
(SaO2).89(2868)
Alfentanil, an opioid analgesic with a rapid onset of action, has been evaluated for its analgesic effects
when self-administered by patients undergoing endoscopic procedures, so-called patient-controlled
analgesia (PCA). In this study of Nguyen et al.,90(2869) PCA using alfentanil was found to be both
safe and readily accepted by patients and medical staff.
One advantage associated with the use of a narcotic is that the effects of the drug, particularly
respiratory depression, are reversible by intravenous administration of Narcan (naloxone).91(2870)
Most respondents to a survey of American endoscopists reported that they rarely (44%) or never
(11.3%) reverse the effects of narcotics at the completion of an endoscopic procedure, although a
smaller percentage of endoscopists either routinely (14.8%) or occasionally (29.9%) reverse
narcotics.14(2871) Because the duration of action of naloxone is shorter than that of meperidine, an
additional intramuscular dose of naloxone should be given in cases of serious respiratory depression.
Agents used in Combination
Benzodiazepines are often used in combination with opioids to induce conscious sedation for GI
endoscopy. The administration of a combination benzodiazepine/narcotic increases the risk of
hypoxemia92(2872) and an adverse cardiopulmonary event.20,93,94(2873) Bailey et al.95(2874)
investigated the respiratory effects of midazolam (0.05 mg/kg) and fentanyl (2.0 g/kg) in 12 healthy
volunteers. The incidence of hypoxemia (defined as an oxyhemoglobin saturation of less than 90%)
and apnea (no spontaneous respiratory effort for 15 sec) were evaluated. Midazolam alone produced
no significant respiratory effects; fentanyl alone resulted in hypoxemia without apnea in half of the
subjects. Midazolam and fentanyl in combination significantly increased the incidence of hypoxemia
(11 of 12 subjects) and apnea (6 of 12 subjects).95(2875) Jurell et al.92(2876) also demonstrated that
the combination of midazolam and meperidine significantly increased the risk of hypoxemia over that
with midazolam alone (44% vs. 8%). These investigators urged caution when a benzodiazepine is used
in combination with an opioid; precautions should include patient monitoring and the availability of
supplemental oxygen.
Hartke et al.94(2877) observed that patients premedicated with intramuscular meperidine and then
given midazolam for additional sedation exhibited frequent episodes of hypotension and respiratory
depression, as indicated by an increased number of apneic events and arterial oxygen desaturation.
These investigators also indicated that elderly patients and patients with underlying cardiopulmonary
disease may require noninvasive physiologic monitoring, including pulse oximetry and continuous
blood pressure measurements, and that supplemental oxygen therapy should be readily available.
Another study that emphasizes the synergistic properties of benzodiazepines and opioids is that of
Walton et al.96(2878) These investigators demonstrated that the degree of hypoxemia experienced by
patients receiving midazolam and the opioid agonist/antagonist nalbuphine was significantly more
profound with respect to incidence and severity than that observed in patients who received midazolam
alone.96(2879)
Reports of the effects of sedative drugs are flawed by the use of inappropriate doses, by comparing
nonequivalent doses of two drugs, by imprecise knowledge of potential side effects, and by the
absence of pharmacokinetic data.97(2880) Evaluation and comparison of many studies are therefore
difficult. For example, in one study of agents used in combination for conscious sedation, pethidine
was used in a dose of 25 to 50 mg and midazolam was "titrated to effect."98(2881) In this study,
oxygen saturation declined to a mean of 82%, 16 of 20 patients developed tachycardia during the
examination, and 10 patients developed supraventricular or ventricular extrasystoles. However, the
amount of sedative and degree of sedation were not specifically quantified and were therefore difficult
to compare with findings in other studies.98(2882) Al-Hadeedi and Leaper99(2883) evaluated 132
patients undergoing diagnostic or therapeutic ERCP using sedation (intramuscular
pethidine/meperidine and intravenous diazemuls/diazepam), and compared the effects of sedation in
these patients with the condition of 51 patients undergoing EGD without sedation. Comparison
between sedated and nonsedated patients in this study, however, is not easily done owing to the
significant inherent difference between these two procedures.
Other Agents
Various agents may be used to suppress GI motility, including atropine, glucagon, and
hyoscine-N-butyl bromide (Buscopan; Boehringer, Germany). The use of an anticholinergic agent such
as atropine or scopolamine may facilitate upper GI endoscopy by reducing secretions and thereby
aspiration, muscle spasm, and the risk of vasovagal responses. The results of a nationwide survey in
Great Britain show that 23% of those endoscopists responding to a questionnaire used an
anticholinergic agent routinely for upper GI endoscopy.15(2884) However, the side effects of atropine
include abdominal and bladder distention, and few controlled trials demonstrate any advantage of
anticholinergic premedication.100,101(2885)
Hedenbro et al.102(2886) studied the effects of anticholinergic medication (scopolamine) administered
intravenously or transdermally on the quality of endoscopic examination and patient discomfort in more
than 230 consecutively observed outpatients. The anticholinergic premedication did not improve the
quality of the diagnostic procedure when administered by either route, and patients noted dryness of
the mouth. The use of scopolamine is not recommended for routine upper GI endoscopy.102(2887)
Intravenous administration of glucagon (0.5 mg) or atropine (0.5 mg) given as premedication for EGD
decreased motility and reduced secretions in the study of Qvigstad et al.,103(2888) but neither drug
had an appreciable effect on the overall success of the procedure compared with a placebo. Schwartz
and Fazio104(2889) found no discernible benefit for premedication with an intramuscular injection of
atropine (0.6 mg) plus meperidine (1 mg/kg) given 30 min prior to EGD compared with premedication
with topical benzocaine and intravenous diazepam. Gerner et al.105(2890) found no significant benefit
when either glucagon or atropine was used in combination with diazepam or meperidine.
Propofol, a sterically hindered phenol, has a rapid onset, large volume of distribution, lack of active
metabolites, and short elimination half-life. These properties suggest that propofol may be suitable for
conscious sedation of patients who undergo GI endoscopy on an ambulatory basis. This drug may be
especially useful in patients who are difficult to sedate. Because of propofol's rapid onset and short
duration of action, additional doses may be required during an endoscopic procedure. Therefore,
greater peak levels of sedation to the point of anesthesia are more likely to occur with propofol than
with other sedative agents. Because of this narrow therapeutic range, assiduous patient monitoring is
imperative when this drug is used. For these reasons, it is recommended that propofol be used only by
physicians with specific training in anesthesiology.
Few reports have been made about the use of propofol for endoscopy. Patterson et al.106(2891)
evaluated 40 consecutive patients who were randomly assigned to receive propofol or midazolam.
Both agents were administered as single injections with similar endpoints of sedation. Amnesia (up to
the time of removal of the endoscope), degree of oxygen desaturation (mean decrease from 98 to
94%; considered clinically insignificant), and changes in heart rate and systemic arterial blood pressure
were similar for both groups of patients. Propofol provided a more rapid recovery and was associated
with less amnesia immediately after removal of the endoscope, two factors that may facilitate early
discharge. However, propofol was associated with pain on injection, and patient acceptance of the
drug was less than that for midazolam. In another comparative study from Sweden of 90 patients
undergoing EGD on an ambulatory basis, propofol (0.6 mg/kg) facilitated endoscopy to a greater extent
than midazolam (0.06 mg/kg) as evaluated by patient and physician questionnaires.107(2892)
However, propofol was considered a more demanding agent to administer because of its narrow
therapeutic range.107(2893) Thus, propofol may be less universally applicable for outpatient
endoscopy.
Droperidol is a short-acting butyrophenone tranquilizer and a powerful antiemetic that produces a
neuroleptic state characterized by reduced anxiety, decreased motor activity, mild sedation, and a
disassociation from surroundings. The onset of action is 3 to 10 min, and its effects may last 3 to 6 hr.
The effects of droperidol for upper GI endoscopy were studied in a prospective, randomized,
placebo-controlled trial of Barthel et al.108(2894) Droperidol (5 mg) significantly reduced the dose of
benzodiazepine and narcotic medication given prior to the procedure and reduced procedure time by
10%. No significant effects on mean arterial blood pressure were seen, but droperidol significantly
reduced the procedure-related increase in pulse rate. No consensus has been reached on the use of
droperidol as an adjunctive medication for upper GI endoscopy.
Alternative Methods of Preparation for EGD
Although serious complications related to pharmacologic preparation of patients for endoscopic
procedures are infrequent, it would nevertheless be desirable to avoid or minimize the use of drugs
during endoscopy. In addition to improved margins of safety, alternative methods of patient preparation
could reduce the cost and shorten recovery time. Unfortunately, relatively few attempts have been
made to develop ancillary or equivalent substitutes for conscious sedation.
Pound et al.109(2895) conducted a randomized, double-blind placebo-controlled study of orally
administered diphenhydramine (100 mg), acetaminophen (1000 mg), or both drugs in conjunction with
topical oropharyngeal anesthesia (Cetacaine) in patients undergoing EGD using a 7.9-mm diameter
"pediatric" endoscope. Administration of both drugs orally 30 to 60 min prior to the procedure was
found to improve patient tolerance for the procedure compared with topical anesthesia alone.
Hypnosis has been suggested as a method of preparing patients for endoscopy.110(2896) Local
glossopharyngeal and superior laryngeal nerve block anesthesia for upper GI endoscopy has been
described by DeMeester et al.111(2897) Studies have been done concerning the use of acupuncture
for EGD.112,113(2898) In general, data concerning the efficacy and safety of these various
approaches to preparation of patients for upper GI endoscopy are limited.
Escher et al.114(2899) measured plasma adrenocorticotropic hormone, cortisol, and catecholamine
levels in 32 patients before and after EGD to study the influence of music on procedure-related stress.
Patients were randomly assigned to one of two groups (unmatched for sex, age, or underlying
disease), one of which listened to music during the procedure. The type of music was selected by the
patient after a short discussion with a music therapist. Hormone levels were significantly lower
immediately after the procedure in patients who listened to music. Subjectively, patients also noted less
anxiety and fear concerning the procedure.
Monitoring
Cardiorespiratory complications are responsible for most deaths related to GI
endoscopy.115118(2900) Patient monitoring during endoscopic procedures includes observation of
the patient by the physician and GI assistant and measurements of blood pressure, pulse, respiratory
rate, cardiac rhythm, and oxygen saturation. The purpose of patient monitoring is to improve the overall
outcome for endoscopic procedures by decreasing the risks associated with parenteral sedation and
with the procedure itself.97(2901) It is often correctly noted that the use of special equipment to
monitor patients may be a useful adjunct to patient surveillance, but this must never substitute for the
conscientious clinical assessment of a patient's tolerance of an endoscopic procedure. However,
patient observation may be limited in a darkened room during a technically demanding procedure in
which the endoscopist's attention is focused on the task at hand.115(2902) Significant hypoxemia and
hypoventilation, as determined by patient monitoring devices, have been demonstrated to occur during
endoscopic procedures and yet go undetected despite rigorous clinical assessment.119,120(2903)
Such studies confirm the relative inaccuracy of clinical assessment in detecting changes in physiologic
parameters. Assignment of responsibility for patient monitoring to someone other than the endoscopist
has been shown in similar studies to be advisable.121(2904)
The primary goal of monitoring is the early detection of potentially harmful changes in a patient's
physiologic parameters. Although such changes are best detected by methods of continuous
monitoring,97(2905) optimum management of patients during GI endoscopy requires integration of
information from monitoring devices, clinical assessment of patient status, endoscopic procedural
activity, and knowledge of the patient's underlying medical condition.120,122(2906)
A great deal of attention has been directed toward the role of pulse oximetry in GI
endoscopy.97,123(2907) The pulse oximeter directs an infrared beam of light into the vascular bed of
the digit (finger oximeter) and uses the optically detected color changes in the reflected light to
estimate the percentage of hemoglobin bound to oxygen. By detecting variations at wavelengths of 660
and 940 nm, the light absorbance differences between oxyhemoglobin and deoxyhemoglobin can be
estimated and compared with a set of algorithms derived from data obtained from healthy
volunteers.124,125(2908) Estimations of oxygen saturation obtained in this manner have been shown
to be accurate within 6% for many commercially available pulse oximeters.125127(2909)
The use of pulse oximetry during GI endoscopy has several advantages: The device is simple to use, it
does not interfere with the procedure, it provides continuous monitoring of oxygen saturation, and it
sounds an alarm with low oxygen saturation and at the extremes of heart rate.97(2910) However, the
estimation of oxygen saturation by means of a pulse oximeter also has certain limitations. Because
these devices use only two discrete wavelengths of light, only two forms of hemoglobin may be
distinguished: oxyhemoglobin and deoxyhemoglobin. Thus, in patients with dyshemoglobinemias (e.g.,
sulfhemoglobin, carboxyhemoglobin, methemoglobin), substantial errors in the estimation of oxygen
saturation may be observed.124,128,129(2911) Ambient light from surgical lights and infrared heating
lamps also affects the accuracy of pulse oximeters, causing readings to be either falsely low or falsely
normal.130,131(2912) Covering or protecting the pulse oximeter sensor (i.e., shielding it from ambient
light) eliminates such interference.
Pulse oximeter readings are affected by nail polish applied to finger nails. Cote et al.132(2913)
demonstrated that finger sensors used on nail beds with black, blue, and green nail polish produced
statistically significant decreases in oxygen saturation readings compared with unpolished nails.
Although red and purple nail polishes do not introduce similar errors in oxygen saturation
measurements, all nail polish should routinely be removed prior to placement of the finger pulse
oximeter.124(2914) Movement of the sensor (motion artifact), compression of the sensor, and
diminished perfusion of the finger to which the sensor is attached may all be sources of error in the
assessment of oxygen saturation by means of a pulse oximeter.
GI endoscopy is a relatively safe procedure with a low complication rate. The appropriate level of
monitoring for individual patients undergoing endoscopy may differ according to clinical
circumstances.97(2915) Furthermore, the safety of GI endoscopy was established prior to the advent
of pulse oximetry. Which patients therefore should be monitored by a pulse oximeter?
Patients thought to be at increased risk for complications during GI endoscopy include those with
ischemic heart disease, cerebrovascular disease, morbid obesity, anemia, GI bleeding, severe
pulmonary disease, and renal or liver failure.115(2916) However, available data do not uniformly
support this contention. For example, the elderly and frail or infirm are often believed to require a
higher level of monitoring. Russell133(2917) performed continuous monitoring of arterial oxygenation

via pulse oximetry in 50 consecutive elderly patients (mean age 73.2 years, range 6589 years)
undergoing elective EGD. Patients were sedated with diazepam (0.1 mg/kg). All patients experienced a
fall in Sao2 (mean 5.7%, range 1 to 14%), but none had clinically significant arrhythmias. Therefore,
age alone may not be an appropriate risk factor for monitoring during GI endoscopic procedures.
Similarly, Visco et al.124(2918) studied 50 patients undergoing endoscopic procedures (EGD or
colonoscopy) and divided them into groups with (n = 11) or without (n = 39) a history of cardiac or
pulmonary disease or both. A significant decrease in oxygen saturation, as measured by finger pulse
oximetry, occurred in both groups, but no significant difference was seen in the degree of desaturation
between the two groups of patients. Therefore, a history of cardiac or pulmonary disease alone may be
inappropriate as an indicator for monitoring during endoscopic procedures.
Wilcox et al.134(2919) evaluated 25 hospitalized patients with well-defined coronary heart disease who
underwent endoscopic procedures. Although 24% of patients had one or more episodes of
electrocardiographic ischemia during the recording period (continuous recording during endoscopic
procedures and a prolonged baseline period), no patient had evidence of ischemia exclusively during
the endoscopic procedure. These investigators concluded that endoscopy in patients with stable but
severe coronary heart disease, when performed using standard medications and careful monitoring
techniques, rarely results in silent or symptomatic ischemia.134(2920) Despite the current lack of
evidence, however, it seems reasonable to assume that patients with severe pulmonary disease are at
greater risk for respiratory failure during endoscopy, even when minimum doses of sedative or narcotic
drugs are used.97(2921) In addition, patients with severe pulmonary disease may be more likely to
cough, gag, or aspirate gastric contents during endoscopic procedures.97(2922)
The issue of guidelines for patient monitoring remains controversial and largely unresolved.125(2923)
Some argue that guidelines should be established for "universal monitoring" of all patients undergoing
GI endoscopy.135137(2924) Indeed, some consider pulse oximetry to be mandatory during upper GI
endoscopy.138,139(2925) The opposite position states that the intensity of monitoring should be
proportional to the perceived risk factors for the individual patient, the level of sedation, and the type
and duration of the procedure. Unfortunately, no good data are available on which to base
recommendations, nor are such data likely to become available owing to the low incidence of severe
adverse outcomes for GI endoscopy.115(2926)
The results of a questionnaire sent to more than 500 members of the ASGE disclosed that pulse
oximeters and electrocardiographic monitoring were used by 65 and 55% of endoscopists,
respectively, and that electronic monitoring was used more frequently for patients undergoing
endoscopy in hospital-based units (99.5%) than in private offices (27%).14(2927) In an effort to
standardize the approach to patient monitoring, the ASGE has made a series of recommendations as
a practical guide for patient monitoring (Table 381).140,141(2928) In contrast, a nationwide survey in
Switzerland of all 173 practicing gastroenterologists was answered by 123 endoscopists (71.1%), who
reported data from more than 115,000 endoscopic procedures.142(2929) Sixty percent of procedures
were performed using conscious sedation, vital signs were monitored in less than 25%, and electronic
monitoring devices were virtually never used. The overall sedation-related complication rate was
0.10%, and no procedure-related deaths occurred.142(2930) Thus, the morbidity and mortality related
to conscious sedation in Switzerland are low, and no higher than in countries in which a high level of
monitoring is maintained.
Table 38-1 American Society for Gastrointestinal Endoscopy Guidelines for

Monitoring of Patients Receiving Conscious Sedation for Gastrointestinal Endoscopy
Table 38-1 American Society for Gastrointestinal Endoscopy Guidelines for

Monitoring of Patients Receiving Conscious Sedation for Gastrointestinal Endoscopy
Monitoring is one aspect of endoscopy unit policy. It should be part of the overall quality assurance
program for the endoscopy unit.
A well-trained gastrointestinal assistant, working closely with the endoscopist, is the most important
part of the monitoring process.
The use of extracorporeal equipment to monitor patients may be a useful adjunct to patient
surveillance, but is never a substitute for conscientious clinical assessment.
Although changes in blood pressure, pulse, cardiac rhythm, and oxygen saturation do occur during
endoscopy, no controlled studies address the question of whether noninvasive monitoring with
extracorporeal equipment decreases complications.
The amount of monitoring should be proportional to the perceived risk to the patient undergoing the
procedure. It may vary from one procedure to another.
The minimal clinical monitoring for all sedated patients should include the determination of heart rate,
blood pressure, and respiratory rate before sedation, during the procedure, immediately after the
procedure, and when the patient is released from the endoscopy area.
The proper role for pulse oximetry and continuous electrocardiographic monitoring during endoscopic
procedures is controversial and unsettled.
Given the cost of the equipment and the manpower to use it, the best decision about whether such
monitoring should be used would be based on data showing the effect on clinical outcome. Such data
do not exist.
However, when the individualized need of the patient indicates that measurement of cardiac rhythm or
oxygen saturation would complement the clinical assessment, the use of electrocardiographic
monitoring or pulse oximetry may be beneficial.
Adapted from Fleischer D. Monitoring the patient receiving conscious sedation for gastrointestinal
endoscopy: Issues and guidelines. Gastrointest Endosc 1989; 35:2626; and the American Society for
Gastrointestinal Endoscopy. Guidelines for clinical application: Monitoring of patients undergoing
gastrointestinal endoscopic procedures. ASGE Publications No. 1022. Gastrointest Endosc 1991;
37:1201.
In a position paper on monitoring of patients undergoing GI endoscopic procedures, the ASGE

correctly noted that, "There are no controlled trials addressing the question of whether non-invasive
monitoring of sedated patients undergoing endoscopic procedures decreases the frequency of
complication. Moreover, it is unlikely that a scientifically rigorous study will ever be conducted because
the very low complication rate of endoscopic procedures dictates that a prohibitively large number of
patients would need to be studied."141(2931) In an effort to address this problem, DiSario et
al.143(2932) conducted a prospective, randomized, controlled evaluation of automated cardiovascular
monitoring versus clinical observation in more than 600 patients undergoing routine endoscopy.
Hemodynamic parameters such as blood pressure and heart rate were monitored every 3 min in the
group randomized to automated monitoring. Hemodynamic aberrations occurred in 71% of monitored
patients, including hypotension (reduction in mean arterial pressure to less than 60 mm Hg) in 6%,
none of whom were symptomatic. No adverse outcomes occurred in either group. DiSario et
al.143(2933) concluded that automated monitoring of blood pressure and heart rate detects frequent
hemodynamic alterations in patients undergoing routine endoscopic procedures, but the alterations are
often clinically insignificant and such monitoring does not improve patient outcome.
Oxygen Saturation
Some endoscopists regard transient hypoxemia that responds to gentle stimulation and bradycardia
that resolves with cessation of endoscope advancement as physiologic responses, not
complications.144(2934) Indeed, arterial desaturation is frequently observed in patients who undergo
EGD without sedation.145(2935) However, given the volume of data concerning the role of oxygen
monitoring during endoscopy, the majority of clinicians consider arterial oxygen desaturation to be a
potential pulmonary complication related either to premedication or to the procedure itself.
Cardiopulmonary events, although uncommon, are a major source of complications during EGD,
accounting for 50% of the morbidity and 60% of deaths.116118(2936) Electrocardiographic changes
during endoscopy have been reported in patients with and without a history of cardiovascular disease,
and endoscopy-associated cardiac arrhythmias are most commonly associated with periods of
maximum oxygen desaturation.146,147(2937) A relationship between hypoxemia and
electrocardiographic ST-segment depression has been demonstrated during upper GI
endoscopy.148(2938)
Continuous monitoring of arterial oxygen saturation during upper GI endoscopy is becoming routine,
although it is not practiced universally. Therefore, the exact incidence and severity of
endoscopy-related arterial desaturation are not known. Oxygen desaturation may, however, occur
more commonly than recognized. In a prospective analysis, Dark et al.135(2939) found that 80 of 115
patients (70%) undergoing endoscopic procedures experienced oxygen desaturation, defined as a
greater than 4% decrease from baseline SaO2. Approximately one third of patients had severe
desaturation, which was defined as an Sao2 of less than 85%, reflecting hypoxemia (PaO2 <50 mm
Hg). Dark et al.135(2940) suggest that their data support the concept that continuous monitoring of
Sao2 should become standard practice during all GI endoscopic procedures. However,
Vasquez-Olivencia et al.149(2941) suggest that the data of Dark et al.135(2942) do not support this
conclusion. Although significant oxygen desaturation occurred, the desaturation did not predict or result
in cardiopulmonary complications during GI endoscopy.
This same argument is frequently cited for studies that demonstrate oxygen desaturation in the
absence of any clinically important sequelae. During colonoscopy, changes may be noted in SaO2,
blood pressure, and other physiologic parameters similar to those that may be encountered during
EGD. As in the study of Fennerty et al.,150(2943) no apparent morbidity was associated with such
changes, which these investigators regarded as physiologic. In another study designed to assess the
clinical utility of transcutaneous oxygen saturation monitoring during routine diagnostic and therapeutic
upper GI endoscopy and colonoscopy, 271 consecutive patients were sedated with meperidine and
diazepam or midazolam (mean doses, 50 mg and 6.4 or 3.3 mg, respectively) and monitored for
changes in oxygen saturation.151(2944) The observed decrease in mean Sao2 compared with
preprocedure levels (3.2%) was less than that considered normal during sleep, suggesting that
monitoring of oxygen saturation during routine endoscopic procedures was of no clinical
value.151(2945)
Factors that predispose to arterial oxygen desaturation are not clearly understood; however, studies
have implicated such variables as duration of the procedure, degree of premedication, patient age,
anemia, obesity, and endoscope diameter.123,135,146,152,153(2946) When the degree of oxygen
desaturation was prospectively evaluated in 120 patients undergoing upper GI endoscopy sedated with
either diazemuls, midazolam, or normal saline (control group), no significant difference was found in
the degree of oxygen desaturation between groups.154(2947) These data suggest that endoscopic
intubation (i.e., the endoscopic procedure itself) and not the sedation was primarily responsible for the
desaturation. However, equally convincing data indicate that sedation with benzodiazepines is the
primary cause of oxygen desaturation during EGD, and not merely endoscope intubation, although
hypoventilation and further arterial oxygen desaturation may be caused by either the mechanical effect
of the endoscope or a reflex stimulated by the endoscope.121,155(2948) Another potential contributing
factor for oxygen desaturation during upper GI endoscopy is the level of operator experience.
Significantly more oxygen desaturation has been found when trainees perform endoscopy.154(2949)
Although a number of studies have implicated various factors in the development of oxygen
desaturation during endoscopy, other prospective studies designed to detect statistical differences in
oxygen desaturation in patients with a variety of predisposing factors have not found such differences.
McKee et al.156(2950) used pulse oximetry (finger probe) to monitor oxygen saturation in 100
consecutive patients undergoing routine colonoscopy. A decrease in SaO2 to less than 90% was
demonstrated in 40 patients after intravenous sedation but prior to colonoscopy, 14 patients sustained
a decrease to less than 90% during colonoscopy, and 46 patients maintained Sao2 levels at greater
than 90% throughout the study period. No statistically significant differences were found when these
three groups were compared for age, body surface area, sedative dosage, smoking history, or history
of preexisting hypertension, diabetes, arrhythmias, angina, or myocardial infarction.156(2951) Data
such as these suggest that distinguishing those patients who are at increased or reduced risk for
oxygen desaturation may be difficult prior to endoscopy.
The inability to determine reliably which patients are at increased risk for oxygen desaturation during
endoscopic procedures has led to the recommendation that all patients receive supplemental oxygen
and that monitoring by pulse oximetry be initiated prior to all endoscopic procedures.124(2952) The
risk of hypoxia during upper GI endoscopy can be reduced by keeping the dose of sedative drugs to a
minimum, avoiding the use of combinations of drugs such as benzodiazepines and narcotics, using
smaller-diameter endoscopes to reduce the possibility of airway compression, and administering
supplemental oxygen.140,157,158(2953)
Supplemental Oxygen
Numerous studies verify that oxygen supplementation prevents or reduces episodes of arterial
desaturation during upper GI endoscopy.52,67,92,153,159164(2954) The use of supplemental
oxygen has been advocated as standard practice for all patients undergoing GI endoscopy, as well as
its continuation into the postendoscopy period.136(2955) The rate flow (2 vs. 3 L/min) and route of
administration (nasal vs. oral) appear to be of less importance than the timing of oxygen
administration.160(2956) Specifically, Bell et al.160(2957) found that preoxygenation (administration
prior to starting the procedure) significantly reduces the frequency of episodic oxygen desaturation
during upper GI endoscopy.
Jurell et al.92(2958) evaluated hypoxemia and electrocardiographic changes in a prospective study of
sedated patients with and without documented ischemic heart disease. Supplemental oxygen
significantly decreased but did not entirely prevent hypoxemia (oxygen saturation less than 90%) in
cardiac and control groups.92(2959) In addition, the incidence of electrocardiographic ST-segment
deviation in cardiac patients who were hypoxemic was significantly lower in the group receiving
supplemental oxygen (p = .0015). Jurell et al.92(2960) recommended that supplemental oxygen be
provided to patients with ischemic heart disease who undergo conscious sedation for endoscopic
procedures. However, Bowling et al.165(2961) found that supplemental oxygen during endoscopy did
not reduce the occurrence of cardiac rhythm disturbances in a randomized study of 103 patients over
age 60 years, with the exception of supraventricular ectopic beats, which did occur more commonly in
patients who did not receive supplemental oxygen.
Supplemental oxygen is most commonly administered nasally during upper GI endoscopy owing to the
obvious need to access the mouth for passage of the endoscope. However, Bell et al.52,67,166(2962)
demonstrated that the pattern of breathing is predominantly via the mouth rather than the nose after
intubation of the esophagus. Hebbard et al.167(2963) studied oxygen supplementation via nasal
cannula versus a catheter passed into the low oropharynx to eliminate the effect of mouth breathing.
Intranasal and intrapharyngeal methods of oxygen supplementation were found to be of similar
efficacy. Novel oxygenating mouth guards have been developed (Oxyguard, Tri-Med Specialties,
Overland Park, KS) to accommodate the alteration in breathing pattern that develops during
endoscpy.168(2964) This device is similar in design to the standard mouth guard, with the addition of a
side port through which tubing may be attached to deliver oxygen. The oxygenating mouth guard has
been tested in randomized studies and has been shown to better prevent severe oxygen desaturation
than endoscopy without supplemental oxygenation.168(2965) Some predict that use of oxygenating
mouth guards will become standard practice because they eliminate the need for nasal prongs and
therefore reduce costs related to the use of disposable nasal cannulas.157,169(2966)
Technique of Upper Gastrointestinal Endoscopy

Conventions
Certain conventions have been adopted with respect to deflection of the tip of the endoscope, use of
the deflection control knobs, and the endoscopic field. Deflection of the tip is usually referred to as
upward or downward and left or right. These terms have no meaning with respect to common
reference points (e.g., the sky is up). The easiest way to understand the terminology of tip deflection is
to relate the motion of the tip to the control section of the instrument when the insertion tube is perfectly
straight. In this configuration the air/water and suction valves face upward (toward the ceiling). Thus,
upward deflection of the tip bends it backward along the insertion tube in the direction of the valves;
downward deflection is in the opposite directionaway from the valves. Looking along the length of the
insertion tube from the control section, left deflection is toward the operator's left, right toward his or
her right. The actual motion of the tip within a patient with upward deflection may be something entirely
different from upward in relation to the physical surroundings.
Terms that describe the directions in which the control knobs are turned to deflect the tip are also
relative. For the purpose of description, the control section is viewed on the side that has the deflection
knobs. Each knob may then be considered to move in either a clockwise or a counterclockwise
direction. Thus, counterclockwise rotation of the innermost control knob deflects the tip upward. A
counterclockwise turn of the outer knob deflects the tip to the left. Clockwise rotation of the inner and
outer knobs results in downward and right tip deflection, respectively.
The endoscopic visual field may be divided in clockface fashion, the 12 o'clock position being at the top
of the field, 3 o'clock to the right, and so on. Most endoscopes have a small marker in the visual field at
the 12 o'clock position. As one looks through a fiberoptic instrument, upward tip deflection causes the
visual field to move toward the 12 o'clock position, right lateral deflection toward 3 o'clock, and so forth.
For video endoscopy, upward deflection moves the endoscopic field in the direction that corresponds
to the top of the video monitor.
Every beginning student of endoscopy is anxious to take hold of an endoscope, insert it into a patient,
and begin work. However, two preliminary and vitally important concepts must be understood to
acquire a long-range foundation for expertise in endoscopy: (1) the method of holding the instrument,
and (2) the endoscope-body position relationship.
Holding the Instrument

The first and most important lesson in endoscopy occurs when a novice picks up the instrument with
the left hand for the first time. Initial experience plays an important part in establishing a method of
holding the instrument. This method, once adopted, determines in large measure the "style" of
endoscopy. It should be clearly understood that the method of holding the instrument either facilitates
or hinders the development of expertise, particularly in other procedures such as ERCP and
colonoscopy.
From a human engineering standpoint, the control section of any endoscope is designed to fit the left
hand. In reality, the design of any instrument is a crude approximation because of great differences in
finger length and hand size. In practice, the hand compensates for the characteristics of the
instrument. Most fiberoptic endoscopes have two coaxial deflection control knobs placed near the
eyepiece objective. Similar coaxial control knobs are positioned below the electronic controls (e.g.,
image freeze, image capture) of electronic endoscopes. The control section nearest the insertion tube
is usually narrower than the upper part that houses the deflection controls and air/water and suction
valves. This allows for a more comfortable grip of the instrument. Future changes in the control section
will probably incorporate a more ergonomic design. The many methods for holding the endoscope can
be classified into two general categories: the two-finger and the three-finger grips.
Two-Finger Grip
Because endoscopes have two valves, it seems desirable to place a finger (index and middle) on each
valve. As the instrument is grasped, the thumb, index, and middle fingers come into relation to the
valves and control knobs, leaving the fourth and fifth fingers to hold the lower, narrower part of the
control section tightly (Figure 381). An endoscopist with short fingers finds that this grip places the
tips of the second and third fingers on the valves. For most people, however, the second and third
fingers come to lie across the valves. If so, there is no point in attempting to place the fingertips
squarely on the corresponding valves because this loosens the grip of the fourth and to some degree
the fifth finger.
(2967)Figure 381. Two-finger grip in frontal view showing index and middle fingers on
suction and air/water valves, respectively.
Exceptionally long fingers may reach the control knobs when the two-finger grip is used. This has an
advantage in that the knobs may be operated with the index and middle fingers as well as the thumb.
Lateral finger motion is limited, however, and extension of the fingers across the valves toward the
control knobs provides only the small benefit of holding a knob in place after it has been turned with the
thumb. In any case, the fingers of most individuals are not long enough to take advantage of this
without some compromise.
When the index and middle fingers are placed across the air/water and suction valves, the instrument
is held for the most part by the fourth and fifth fingers, thus the term two-finger grip. The top of the
control section of many endoscopes is contoured in such a way that it comes to rest against the lateral
aspect of the fourth finger, which thus supports some of the weight of the instrument. This reduces the
degree of grip strength needed to hold the instrument.
The thumb is a most important appendage. Although it is possible for a few endoscopists to reach
either of the two coaxial control knobs comfortably with the thumb, the left thumb of most endoscopists
is only long enough to reach the inner of the two knobs ("up/down"). Deflection of the insertion tube is
controlled mostly by turning the up/down deflection knob with the left thumb.
The main disadvantage of placing the second and third fingers across the valves (i.e., the two-finger
grip) is a partial loss of function of the thumb. This method of grasping the instrument tends to move
the thumb upward along the control section. The usual result is that the left thumb contacts the inner
control knob near the distal interphalangeal joint rather than at the tip of the thumb (Figure 382). The
knob is then moved mainly by opposition rather than flexion of the thumb. This provides a lesser range
of knob motion than is possible when the tip of the thumb is on the control knob and a combination of
flexion and opposition of the thumb is used to turn the control.
(2968)Figure 382. Two-finger grip. A, View from the side illustrating the relationship
between the deflection control knobs and the left thumb. B, View from the back of the hand
showing the umbilical cord of the endoscope running over the groove between the thumb and
the index finger.
It takes some time to become accustomed to holding an endoscope in the left hand for prolonged
periods. Until this becomes habitual, the student usually finds that the grip strength of the fourth and
fifth fingers gradually decreases through the course of several procedures. The control section tends to
drop downward in the hand so that eventually it may be supported almost entirely in the groove
between the thumb and index finger (see Figure 382). This greatly compromises the use of the left
thumb so that the novice increasingly uses the right hand to turn both control knobs. This slovenly use
of the left thumb as a "hanger" for the instrument is deplorable.
Three-Finger Grip
Another method of holding the instrument involves shifting the control section upward in the hand so
that the narrower lower part of the control section is grasped with the third, fourth, and fifth fingers, thus
the term three-finger grip (Figure 383). This has two effects: The middle finger is no longer available
to operate the air/water insufflation valve, but the tip of the thumb can now be placed on the inner
deflection knob (Figure 384). With this method, the left index finger must be used to operate both the
air/water and suction valves (see Figure 383). This is not especially difficult, although it poses a slight
disadvantage in that simultaneous operation of both the air/water and suction valves can be
accomplished only by shifting the instrument downward in the left hand or by using the fingers of the
right hand to operate one or both valves, the latter being the preferred method. In practice,
simultaneous depression of both valves is seldom necessary. The advantage of the three-finger grip is
that it permits greater use of the left thumb. The left thumb has more control over the inner (up/down)
deflection knob, and it can also be extended to the outer (left/right) deflection knob, allowing small
lateral deflections while the insertion tube of the endoscope is maneuvered with the right hand.
(2969)Figure 383. Three-finger grip shown in a frontal view. The index finger can be used
to operate both valves; the control section is gripped with remaining three fingers.
(2970)Figure 384. Three-finger grip. A, View from side illustrating the relationship between
the deflection control knobs and the left thumb. B, View from the back of the hand showing
the umbilical cord of the endoscope running over the groove between the thumb and the index
finger. The thumb can reach across to the outer deflection control knob.
Lateral Deflection
The deflection control knobs cannot be operated with any degree of skill with the left hand alone,
regardless of the method used for holding the instrument. It is especially difficult to rotate the two
control knobs in opposite directions. Manipulation of both knobs in addition to the suction and air/water
valves with one hand becomes extremely clumsy and ineffective for all but a few exceptionally
dexterous individuals. For practical purposes, the lateral deflection control knob, the outer of the two
coaxial systems, is turned with the right hand. The other function of the right hand is advancement and
withdrawal of the insertion tube.
Because the lateral deflection knob must also be operated with the right hand, it is not possible to
laterally deflect and advance the instrument at the same time. This seems to make a case for allowing
an assistant to advance the instrument while the endoscopist steers with both hands at the control
section. However, this two-person method is never as rapid or precise as when endoscopy is
performed by one individual and is unnecessary when performing upper GI endoscopy. As described
later, the technique of rotating or torquing the insertion tube eliminates or greatly decreases any
disadvantage inherent in the inability to laterally deflect and advance the instrument at the same time.
In fact, in most cases it is not necessary to remove the right hand from the insertion tube until after the
tip of the instrument has reached the apex of the duodenal bulb. As also described later, using both
deflection controls simultaneously while withdrawing the insertion tube is entirely possible.
The Endoscope-Body Position Relationship

The endoscope-body position relationship refers to the configuration of the insertion tube outside the
patient. This is determined entirely by the position of the endoscopist or, more precisely, by the position
of the control section of the endoscope relative to the point of entry of the endoscope into the patient.
The expert endoscopist determines the configuration and position of the insertion tube external to the
patient and takes advantage of this relation in maneuvering the instrument. Concentrating on what
seems to be occurring within the patient, the novice allows the insertion tube to assume any number of
configurations and is unaware that this greatly influences the ability to maneuver in both positive and
negative ways.
There are significant differences in the way fiberoptic and electronic (video) endoscopes are held in
relation to the endoscopist's body owing to the necessity with the former of looking through the
objective lens (Figure 385A and B). This produces slight but significant differences in the technique of
upper GI endoscopy.
(2971)Figure 385. Various stances for performing endoscopy. A, Stance with the fiberoptic
endoscope held up to the eye. B, Stance with the electronic (video) endoscope held near
endoscopist's epigastrium. C, Stance with the electronic (video) endoscope held on the left
side of the body.
By necessity, the control section of the fiberoptic endoscope is held up to the eye (see Figure 385A).
Thus, the position of a fiberoptic instrument depends on the height and stance of the endoscopist. The
control section of the electronic endoscope, on the other hand, is most comfortably positioned at about
the middle of the endoscopist's torso. Furthermore, changing the position of the control section of an
electronic instrument relative to the endoscopist and patient is much easier. Raising or lowering the
position of the control section relative to the surface of the examination table offers an additional and
highly advantageous method of rotating the insertion tube. Some endoscopists hold the control section
of electronic endoscopes upright near their epigastrium (see Figure 385B), whereas others hold it at
an angle from the left side of their body, with the left palm facing slightly upward (see Figure 385C).
The latter method is frequently adopted by endoscopists with relatively small hands who experience
difficulty in grasping the control section.
Although all modern endoscopes are considered flexible, they also have certain degrees of stiffness.
Resistance to deforming force is most evident in an instrument's torque stability. This means that a
twisting motion applied to the insertion tube at one end is transmitted along its long axis to the other
end with little or no loss of motion, provided that the insertion tube is as straight as possible. When the
instrument is straight, rotatory motion also has fidelity; that is, any degree of rotation at one end is
reproduced promptly and equally at the opposite end. When the insertion tube is in a looped
configuration, a twisting force applied at one end is absorbed to some degree by the loop, and the
rotatory motion at the opposite end is diminished. The degree of transmission of the rotatory motion is
also less predictable or may not occur at all, in which case the quick response at the distal end of the
instrument is lost.
The behavior of the insertion tube in response to rotatory motion gives rise to a guiding principle for all
endoscopic procedures: The straighter the instrument, the more precise the control. The GI tract
is obviously not straight, but neither is it a rigid tube. Some degree of straightening is therefore possible
during virtually any endoscopic procedure. Furthermore, only part of the insertion tube is within the
patient at any given time, and control of the configuration of the part that is outside the patient is
obviously possible.
All of the tip motions that are possible by combined use of the up/down and lateral deflection controls
can be reproduced by rotation of the insertion tube (torque) and movement of the up/down deflection
control with the left thumb. This almost completely removes the disadvantages associated with the
inability to advance and laterally deflect at the same time. As noted, this works only if the instrument is
relatively straight. Thus, coiling or looping of the section of the insertion tube external to the patient
should be avoided. In order to do this, it is usually necessary to stand back somewhat from the patient.
Afraid that they will lose their position, most beginning students of endoscopy tend to stand too close to
the patient, leftward from the patient's mouth (i.e., in the direction of the patient's feet), and to lean
forward (when using a fiberscope), sometimes almost to the point of hovering over the patient (Figure
386). This "novice crouch" communicates intense concentration and forewarns of lifelong lower back
pain, but it allows little or no effective rotation of the insertion tube. As discussed later, methods of
keeping the insertion tube relatively straight within the patient can be learned.
(2972)Figure 386. Endoscopist is standing too close to the procedure table using a
fiberscope (A) and an electronic (video) endoscope (B). In either case, the loop in the
instrument between the endoscopist and the table dampens rotation of the instrument except
for rotation performed with the right hand.
Three basic methods are available for rotating the insertion tube of the endoscope during EGD: (1)
turning the control section clockwise or counterclockwise by flexion or extension of the left wrist; (2)
turning the entire body left (Figure 387) or right (Figure 388) beginning from a position facing the
patient; and (3) raising or lowering the left shoulder (fiberoptic endoscopy) (Figure 389) or abducting
or adducting the left arm (electronic endoscopy) (Figure 3810).
(2973)Figure 387. Left rotation of the insertion tube is achieved by a leftward turn of the
torso with a fiberscope (A) and an electronic (video) instrument (B).
(2974)Figure 388. Right rotation of the insertion tube is achieved by a rightward turn of the
torso with a fiberscope (A) and an electronic (video) instrument (B).
(2975)Figure 389. Left rotation of the insertion tube of a fiberscope is achieved by leaning
forward and lowering the left shoulder. Note the position of the endoscopist slightly to the left
of the patient's mouth.
(2976)Figure 3810. A, Left rotation of the insertion tube of an electronic (video) endoscope
is achieved by abduction and external rotation of the left arm. B, Right rotation of the insertion
tube is achieved by adduction and internal rotation of the left arm.
Torquing the insertion tube with the right hand is ineffective if the control section is held in a fixed
position and may damage the instrument. Up to 90 degrees of rotation to the right occurs with flexion
of the left wrist, and a small degree of rotation to the left can be achieved by wrist extension. If the
endoscopist simply turns from a position facing the patient to the left or right, the result is a
corresponding rotation of the insertion tube in the same direction. With a fiberoptic instrument, a
significant degree of left rotation occurs if the endoscopist moves his or her left shoulder downward,
forward, and slightly to the left, and a small degree of right rotation also occurs by raising the left
shoulder. Leftward torque can be achieved with an electronic endoscope by abducting and externally
rotating the left arm, holding the control section down and away from the endoscopist (see Figure
3810A); rightward rotation can be achieved by adducting and internally rotating the left arm, bringing
the control section up and more toward the endoscopist's head (see Figure 3810B). The electronic
endoscope may be slightly more maneuverable than the fiberoptic endoscope because the control
section can be moved more freely. Rotation of the insertion tube can be accomplished simply by
raising or lowering the left hand in relation to the surface of the procedure table.
Rotation of the insertion tube is accomplished in practice by some combination of the various
maneuvers described earlier. Flexion of the left wrist and a turn by the endoscopist toward his or her
right is most effective when right (clockwise) torque is desired. Lowering the left shoulder or left arm is
usually satisfactory when left (counterclockwise) torque is called for.
When the insertion tube is straight, the position of the endoscopist relative to the point of entry into the
patient has added importance. When the endoscopist moves to the left of the patient's mouth, the
instrument rotates to the left as well (Figure 3811). Because the insertion tube is held with the right
hand, the endoscopist has a natural tendency to stand slightly to the left of the patient's mouth. A
similar but opposite change occurs if the endoscopist moves to the right relative to the patient's mouth.
(2977)Figure 3811. Endoscopist is standing to the left of the patient's mouth with a
fiberscope (A) and an electronic (video) endoscope (B). In these positions, the insertion tube is
necessarily rotated to the left.
Although EGD was at one time performed in either a sitting or standing position, most endoscopists
now prefer to stand. The earlier discussion should make it clear that a sitting endoscopist is less
mobile and therefore less able to rotate the insertion tube properly. Conversely, a standing endoscopist
is in a much better position to take advantage of the various methods of instrument rotation.
The behavior of the instrument in response to the maneuvers described can be demonstrated by
standing at an empty procedure table in the usual position for EGD, with the instrument laid out on the
table as if it were in a patient's stomach. It is helpful if the instrument is passed through a mouth guard,
and the guard is held by an assistant at a point that approximates its usual location between a patient's
teeth. The distal tip should be deflected upward to a slight degree so as to better appreciate the motion
imparted to the distal end of the instrument in response to the various maneuvers. The rotation
maneuvers are then performed while observing the behavior of the deflected tip of the instrument.
Passing the Endoscope

As the endoscope is passed, the patient should be in the left lateral position on the procedure table, left
hand under a pillow, right hand at the side, knees drawn up at a right angle to the torso, and the neck
slightly flexed. The mouthguard should be in place between the patient's teeth. The GI assistant has
certain specific and important functions during passage of the instrument (see Chapter 5: The
Gastrointestinal Assistant).
The endoscope may be passed by one of two methods: blind and direct vision. The most basic rule of
safe and effective endoscopic practice is that the instrument should not be advanced when one is
unable to see ahead. There is no reason to discard this rule for passage of the instrument, and
therefore direct vision is unarguably the better method of insertion.
Blind Passage
The blind method of passing the endoscope requires that the endoscopist place two fingers (generally
the index and middle fingers of the left hand) in the patient's mouth and over the back of the tongue.
Ideally the space between the fingers lies in the midline and forms a groove to guide the instrument
downward into the pharynx. Some physicians guide the instrument tip with the index finger as the
insertion tube is being advanced with the right hand.
Once within the space between the fingers, the instrument may be advanced into the posterior
pharynx. Note that the deflection controls should be in an unlocked position. They are not used during
this process. Resistance to further advancement of the instrument is encountered as the tip reaches
the level of the cricopharyngeus. This muscular sphincter is usually located about 15 to 18 cm from the
incisor teeth (or mandibular ridge). Thus, the right hand should be just behind the 20-cm mark (at
about 22 or 23 cm) on the insertion tube. If the insertion tube is placed distal to the 20-cm mark, it is
often necessary to shift its position backward so that an adequate length of instrument is available to
reach the esophagus. This may result in a momentary loss of control. When the instrument tip reaches
the cricopharyngeus, gentle forward pressure is maintained as the patient is asked to swallow. A
distinct sensation of relaxation is felt as the cricopharyngeus opens and the instrument moves forward
without resistance. As this last action is being performed, the endoscopist withdraws his or her fingers
from the patient's mouth.
Blind instrument passage has numerous disadvantages. The endoscopist, the endoscope, or both are
more likely to suffer a bite. Sedated patients have greater difficulty in controlling their reactions and are
slower in their response to commands. Placing two fingers and an endoscope in the posterior pharynx
is much more likely to cause gagging than is the instrument alone. The patient is less able to initiate a
swallow when he or she is barely able to move the tongue. The glottis, the pyriform sinuses, the larynx,
the cricopharyngeus, and even the proximal several centimeters of the esophagus are not examined,
and these structures are also more likely to be traumatized with this method. Trauma to these tissues
may create the appearance of an abnormality. Certain abnormalities such as an esophageal web may
be overlooked. Blind passage of the instrument is contraindicated in patients with dysphagia that
suggests a proximal esophageal lesion, especially if a Zenker's diverticulum is suspected or known to
be present.
Direct-Vision Passage
Observation begins and continues throughout the examination in the direct-vision method of insertion
as the instrument tip passes the back of the tongue. All of the structures of the posterior pharynx and
larynx are noted as the endoscope is advanced.
The endoscopist must actively manipulate the tip of the endoscope when using the direct-vision
method. It is impossible to use the lateral deflection knob during this maneuver because its use
requires that the right hand be shifted from the insertion tube to the control knob. This shift almost
always results in loss of control of the instrument tip. Steering the tip to the left or right is accomplished
by torquing the insertion tube with the right hand. Because the cricopharyngeus is located about 15 to
18 cm from the incisor teeth, the thumb of the right hand should be placed proximal to the 20-cm mark
on the insertion tube to begin with and kept in this place until the instrument tip reaches the upper
esophagus. Up and down deflection is done by the left thumb. Fortunately, most of the necessary
movements are relatively fine; marked degrees of twisting or turning are not required and indicate a
lack of expertise.
Before the instrument is passed, the lateral deflection control can be locked with the tip in a straight,
neutral position. The endoscopist then stands slightly to the left of the patient's mouth with both hands
at about the level of the mouth. The instrument is inserted to the base of the patient's tongue (Figure
3812), whereupon the tip is deflected upward with the left thumb. If this maneuver is performed
properly, it generally results in a view of the patient's epiglottis (Figure 3813). If this is not seen, the
instrument should be withdrawn somewhat while rotating it left and right slightly with the right hand.
This generally reveals an identifiable structure. Sometimes the instrument tip ends up in the left or right
vallecula of the epiglottis and may give the mistaken impression that the tip is in one of the pyriform
sinuses. However, its true location is easily recognized by the fact that the instrument has not been
advanced far enough to reach the pyriform sinus and by the paler yellow color of the mucosa in this
area.
(2978)Figure 3812. Endoscopic view of the base of the patient's tongue during direct vision
passage of endoscope. Note that the tongue is at the top of the image and the palate is at the
bottom.
(2979)Figure 3813. Endoscopic view as the instrument is advanced beyond the base of the
tongue. The epiglottis appears at the top of the image.
Once the epiglottis is visualized, the tip of the instrument is advanced. Some downward deflection and
slight rotation to the left or right may be necessary at this point to steer the tip around the epiglottis. As
the instrument is advanced, the next structures visualized are the larynx, vocal cords, and pyriform
sinuses (Figure 3814). If the patient swallows and the cricopharyngeus relaxes, the esophageal
lumen is seen between the pyriform sinuses posterior to the larynx. Ideally, the instrument should be
advanced in this direction, keeping to the midline by means of small degrees of rotation of the insertion
tube with the right hand and directed posteriorly away from the larynx. The latter is accomplished by
downward deflection with the left thumb. The patient is asked to swallow as the rosette of the closed
cricopharyngeus is encountered, and then the tip of the instrument is advanced into the esophagus a
few centimeters as the cricopharyngeus relaxes.
(2980)Figure 3814. Endoscopic view of the vocal cords. Slightly yellowish arytenoid
cartilages are seen on either side. The pyriform sinuses are at the left and right edges of the
field.
It is difficult in practice to perform the last part of the direct-vision method perfectly because the
instrument must be kept precisely in the midline as it is advanced toward the cricopharyngeus, and the
advancement must be perfectly timed to the patient's swallow. Therefore, the tip often ends up in one
of the pyriform sinuses, frequently the left because of its dependent position, and the endoscopist
encounters resistance to forward motion. The endoscopist has two options at this point. The instrument
tip can be withdrawn 1 or 2 cm and another attempt can be made to pass the cricopharyngeus.
Alternately, the endoscopist may rotate the instrument slightly toward the midline. If the tip is in the left
pyriform sinus, right rotation is needed. Small adjustments can also be made with the up/down control
as the patient is asked to swallow while the endoscopist maintains gentle forward pressure. The
required rotation and adjustment of the up/down deflection are very slight. Generally, the instrument
does not immediately enter the esophagus, but each swallow reveals some clue to the location of the
cricopharyngeus and thus allows for further and more directed adjustments of the tip position. Small
capillary vessels in the mucosa near the cricopharyngeus run longitudinally; if these are visible, they
also indicate the direction of the lumen.
Esophageal intubation may be unsuccessful by standard methods in the presence of anatomic
abnormalities (e.g., Zenker's diverticulum). Catheter- or wire-guided methods for intubating the
esophagus in such cases have been described.170,171(2981) The catheter or wire is passed under
either direct vision or fluoroscopy through the accessory channel of the endoscope and into the
esophageal inlet. The endoscope can then be passed into the esophagus using the wire or catheter as
a guide.
The endoscope should be withdrawn promptly if the patient has excessive coughing or stridor during
attempts at esophageal intubation. These are usually the first signs that the trachea has been
inadvertently intubated. A practiced endoscopist often recognizes cartilaginous tracheal rings and
immediately withdraws the endoscope, even before the patient responds to the presence of the
instrument in the airway. If the trachea is intubated inadvertently, esophageal intubation can be
reattempted after a few minutes unless laryngeal stridor persists. Persistence of stridor indicates
laryngeal edema, in which case endoscopy should be postponed and the patient observed carefully.
Esophagus
Once within the esophagus, the instrument can be advanced easily. Direct vision should be used at all
times. The surface mucosa is pale pink, smooth, flat, and somewhat glistening (Figure 3815). Small
vessels are usually visible in the mucosa, especially in the distal esophagus. These are oriented
longitudinally with the long axis of the esophagus.
(2982)Figure 3815. Normal esophageal mucosa.

The esophagus is essentially a straight tube. Although it possesses no landmarks, certain structures
external to the esophagus may be noted (see Chapter 44: Hiatus Hernia and Peptic Diseases of the
Esophagus). As the instrument is advanced through the esophagus, intermittent air insufflation may be
used to keep the lumen open. Peristaltic motion may or may not be evident. Some patients may retch
and vomit as the endoscope enters the esophagus. This may cause small amounts of gastric contents
to flow upward toward the pharynx. Quick use of suction prevents fluid from gaining access to the
pharynx and causing further coughing and gagging.
Normally, the gastroesophageal mucosal junction should be clearly evident as a color change
occurring at about 38 to 40 cm from the incisor teeth (Figure 3816). The pale pink mucosa of the
esophagus contrasts sharply with the red or deep orange color of the stomach. The mucosal junction
forms a line around the inner circumference of the lumen. This usually follows an undulating course so
that the junction is often referred to as the Z line. Normally, the diaphragmatic constriction of the gut
lumen should be noted within 2 cm of the Z line. The position of the diaphragm is more evident during
inspiration. Its constriction of the esophagus may be exaggerated by asking the patient to sniff.
(2983)Figure 3816. The gastroesophageal junction. The Z line where the pinkish-gray color
of the squamous esophageal mucosa changes to the redder color of the columnar gastric
mucosa can be seen just proximal to the diaphragmatic constriction.
Entering the Stomach

If only anterior and posterior directions are considered, the course of the endoscope through the upper
GI tract is from posterior to anterior to posterior. The esophagus is essentially a posterior organ,
whereas the stomach lies in a more anterior plane. Viewed from above, the stomach is largely anterior
in relation to the esophagus so that it can swing over the spine (Figure 3817). This means that a turn
in an anterior direction is often needed to advance the endoscope from the distal esophagus through
the cardia and into the proximal stomach. This may be accomplished by rotation to the left (anterior)
with upward deflection as the endoscope passes the Z line (Figure 3818). The easiest method of left
rotation is to lower the left shoulder in a slightly forward direction (fiberoptic endoscopy) or to abduct
and externally rotate the left arm (electronic endoscopy).
(2984)Figure 3817. Transverse view from above showing the relative positions of the
esophagus, stomach, and duodenum. Note that the stomach lies anterior to the other two
structures.
(2985)Figure 3818. Diagrammatic representation of the method of passing the endoscope

from the esophagus through the cardia and into the stomach proper. Left rotation and upward
deflection of the tip are required as the instrument is advanced.
An unsatisfactory alternative to direct-vision passage of the instrument tip through the cardia is to
advance the endoscope without rotation or deflection. It passes through this segment and usually stops
against the posterior aspect of the lesser curvature.
As the instrument tip enters the stomach, the rugal folds along the greater curvature of the body are
seen (Figure 3819). Often a small pool of fluid is found in this dependent location. This observation is
useful because it identifies the greater curvature of the stomach and therefore the opposite lesser
curvature and the anterior and posterior walls. If only a small amount of fluid is present, aspiration of
the gastric contents is probably unnecessary. If, however, the pool obscures the rugal folds beneath, it
is advisable to aspirate the gastric contents. This should be done as much as possible with the
instrument tip lying parallel to the mucosal surface and alternating suction and air insufflation. Applying
suction with the instrument tip perpendicular to the mucosa frequently draws some of the mucosa into
the accessory channel. This delays aspiration, and the mucosa, once freed, becomes hyperemic and
edematous.
(2986)Figure 3819. Endoscopic view of the stomach as the endoscope is advanced and
rotated toward the right. The gastric rugal folds along the greater curvature are seen coursing
toward the antrum, which is at the upper right.
If only a small volume of air was insufflated while traversing the esophagus, the stomach may still be
partially collapsed. When the initial view is unsatisfactory, it is necessary to insufflate the stomach with
air. The ideal degree of gastric distention is a matter of judgment. Generally, the distention should be to
the point that the rugal folds just begin to separate. Later in the procedure, it may be necessary to
distend the stomach to a greater degree to examine thoroughly all of the mucosa of the body,
especially if small mucosal abnormalities such as vascular malformations may be present. Patients
differ with respect to the degrees of gastric distention they tolerate. Some belch uncontrollably in
response to small volumes, and it may be necessary to accept less than ideal distention. Excessive air
insufflation almost always causes the patient some discomfort and should be used only briefly, if at all.
Furthermore, excessive insufflation markedly flattens the rugal folds, leading to the mistaken
impression of gastric atrophy.
Body of the Stomach

The endoscope is in a position of left rotation with upward tip deflection as a result of advancing from
the esophagus through the cardia and into the stomach. To bring the tip through the body and toward
the pylorus, it is necessary to rotate to the right as the insertion tube is advanced. After the endoscope
has passed through the cardia and traversed the proximal gastric body, the insertion tube external to
the patient is oriented mostly along the length of the patient's body; that is, it is rotated about 90
degrees to the left. To accomplish the necessary right rotation as the instrument is advanced through
the stomach, over the spine, and toward the pylorus, one need only bring the left shoulder (fiberoptic
endoscopy) or left arm (electronic endoscopy) back to the normal midline position so that the external
part of the insertion tube follows a straighter line of entry at the patient's mouth. This usually requires
that the endoscopist stand slightly back from the table. The advancement and right rotation of the
insertion tube through the stomach should be a smooth, coordinated maneuver. Small adjustments of
the up/down deflection are also needed, especially additional upward deflection along the lesser
curvature as the tip advances through the body of the stomach.
Paradoxical Motion
The esophagus and duodenum are relatively fixed, narrow, and inelastic structures. The stomach,
however, stretches considerably to accommodate intraluminal forces. As the tip of the instrument is
advanced toward the antrum and pylorus, the insertion tube invariably comes to lie along the greater
curvature. Because of the elasticity of the stomach, a considerable part of the forward force is
absorbed by the greater curvature, pushing the stomach toward the pelvis as a loop forms along the
curvature. Endoscopically it may appear that the instrument tip is not moving forward or even that the
tip is moving away from the pylorus during this maneuver, thus the term paradoxical motion. It is
necessary to continue to advance the instrument until the greater curvature loop is fully formed and the
instrument tip begins to move forward again (Figure 3820) Generally this also requires additional
upward deflection of the tip using the left thumb.
(2987)Figure 3820. Diagram showing the formation of the greater curvature loop. The point
of paradoxical motion has been passed, and the instrument is being advanced toward the
pylorus as the greater curvature loop continues to develop.
Antrum
Two differences between the antrum (Figure 3821) and the body (see Figure 3819) of the stomach
are obvious: Rugal folds are present in the body but not the antrum, and peristaltic contractions are not
noted in the body but are readily observed in the antrum. At this point in the examination, the tip of the
instrument is probably at the midpoint of the greater curvature or slightly beyond. Full upward deflection
of the tip often reveals a view of the angulus of the stomach, although this depends on the overall
configuration of the stomach itself. Some endoscopists perform the retroverted view of the fundus and
proximal stomach at this juncture.
(2988)Figure 3821. Endoscopic view of the antrum. A relaxed, open pylorus is visible.
If the endoscope has been advanced to the antrum quickly, smoothly, and with minimum air
insufflation, no contractions may be observable. Proceeding directly to the pylorus can be
advantageous at this point. The physiologic function of the pylorus is to close in response to gastric
distention so that repetitive antral peristaltic contractions grind the food against the pylorus and
progressively reduce the size of the intragastric food particles. When gastric distention has been kept
to a minimum and no antral contractions are occurring, the pylorus is in a relatively lax state.
Pylorus
The previous left rotation of the insertion tube is largely corrected as the tip of the instrument passes
along the greater curvature into the antrum and toward the pyloric ring, and the external section of the
insertion tube between the patient's mouth and the endoscopist's left hand is relatively straight. As the
tip advances through the antrum, the pylorus should be kept in the middle of the endoscopic field. This
requires small adjustments of the up/down deflection and minor degrees of rotation to the left or right.
Small left or right corrections can be made with the left hand and wrist alone, provided that the external
portion of the insertion tube is relatively straight.
The endoscope tip can be passed through the pyloric ring without difficulty in most patients. In other
cases, however, the pylorus may be closed and may offer resistance to advancement into the bulb.
This can also be an indication of disease in the region of the pylorus, which may itself be strictured. A
certain amount of judgment based on experience must be exercised about how hard to push the
instrument tip against the pyloric ring. If considerable resistance is encountered and the view of the
ring at a distance in the antrum suggests deformity, it may not be possible to enter the bulb.
Novice endoscopists usually have considerable difficulty in crossing the pylorus. One common mistake
is to persist in blindly searching for the pylorus when it has been lost from endoscopic view. Pressure
exerted against the antral wall during these futile efforts usually leads to antral contractions and spasm
of the pylorus. When the pylorus is lost from view while advancing the endoscope, withdrawing the tip
to the proximal antrum and beginning again are always preferable.
The cardinal rule in negotiating a "spastic" or persistently closed pylorus is to keep the ring in the
center of the endoscopic field as the instrument is advanced. This may require momentary
adjustments of the lateral deflection. Locking the lateral deflection control knob may help to keep the
tip steady, and the added rigidity may ease pyloric intubation. It may be difficult to maintain forward
pressure with the right hand while changing the lateral deflection. One way to maintain gentle forward
pressure is to rest the insertion tube on a pillow near the patient's mouth. It may be necessary for the
endoscopist to bend forward and press the insertion tube gently against the pillow. This maneuver
preserves forward pressure and prevents the instrument from falling back into the antrum as the right
hand moves to the lateral deflection control. A distinct sensation is often felt by the right hand on the
insertion tube as the pylorus relaxes and allows the endoscope to enter the bulb. Passage of the
endoscope through the pyloric ring produces a peculiar sensation, sometimes described as discomfort,
in some patients.
Duodenal Bulb
The bulb is a relatively small structure. Generally, the endoscopist leans forward when advancing the
instrument tip across the pylorus. The forward force exerted on the pylorus to gain entry to the bulb
usually carries the instrument tip to the apex. As a result, most of the bulb may not have been
examined. If the endoscopist leans back and places slight backward tension on the insertion tube, the
endoscope tip can be brought back into the pyloric channel, from which point the bulb can be
examined. This examination proceeds in a circumferential manner using both deflection controls. The
mucosa of the bulb has a slightly irregular, villous surface (Figure 3822), and the color is a pale tan, in
contrast to the pink color of the stomach. Accurately determining anatomic relationships within the
duodenal bulb is frequently difficult. In the study of Straker et al.,172(2989) the posterior wall was
located correctly by experienced endoscopists only 30% of the time.
(2990)Figure 3822. Endoscopic photograph of the duodenal bulb. The apex is seen toward
the center.
Once the bulb has been intubated, the novice is often anxious that the instrument not return to the
stomach, especially if difficulty was encountered in inserting it across the pylorus. This tendency to fall
back into the antrum can be countered to some degree by keeping the right hand on the insertion tube.
If the instrument does return to the stomach, the pyloric ring should be observed closely for any
evidence of ulcer or other disease process, especially if difficulty was encountered in passing the ring.
The endoscope almost always passes through the pyloric ring more easily after one successful
passage.
Descending Duodenum
The endoscopic view of the apex of the duodenal bulb usually discloses no obvious lumen because of
the acute angulation at the superior angle between the bulbar apex and the proximal descending
duodenum. Sometimes a valvular fold or two may be visible and a portion of the lumen may be noted,
but this is unusual.
Because the lumen ahead is not clearly defined, insertion of the tip of the instrument into the
descending duodenum is a partially blind maneuver. It is first necessary to place the tip at the apex of
the bulb. In most cases, the course of the lumen is directed superiorly and posteriorly. In the
endoscopic field, this is usually upward and to the right. Once in position, therefore, the tip is deflected
almost fully upward and fully or almost fully to the right and the insertion tube is rotated 90 degrees or
more to the right (Figure 3823). The latter rotation is best accomplished by a combination of flexing
the left wrist and turning toward the right (see Figure 388). This maneuver hooks the superior angle
and, if performed properly, provides a view of the descending duodenum (Figure 3824). Locking the
lateral deflection knob in the right position may help the tip to stay in the "hooked" position while the
instrument is withdrawn. Toward the end of this maneuver, minor adjustments in tip deflection may be
required to provide a tubular view of the duodenum.
(2991)Figure 3823. Diagram showing maneuvers to deflect the instrument tip from the apex
of the bulb into the proximal descending duodenum. Right rotation of the insertion tube is
required (see Figure 388) as well as full upward and right tip deflection.
(2992)Figure 3824. Endoscopic view of the proximal descending duodenum on completion

of the maneuver illustrated by Figure 3823.
Another method of insertion into the proximal descending duodenum also begins by positioning the
instrument tip at the apex of the bulb. The tip is then deflected to the right, and the insertion tube is
advanced while the tip is deflected sharply downward. In this maneuver the degree of downward
deflection required may be as much as 150 degrees. Some endoscopes, however, do not provide this
degree of tip angulation in the downward direction.
After the instrument tip is in the proximal descending duodenum, the next step is to withdraw the
instrument. Because further correction of tip deflection may be necessary, the withdrawal maneuver
can be performed with both hands at the control section. This is not difficult and can be done by simply
lifting the left wrist upward and stepping back a slight distance from the procedure table. An alternate
method of withdrawal is to use clockwise torque on the instrument tube with the right hand while
controlling the up/down control knob with the left thumb.
Withdrawal of the insertion tube at this point takes up the greater curvature loop in the stomach and
causes the tip to move forward in paradoxical fashion (Figure 3825). This straightening of the
insertion tube within the stomach and bulb generally results in advancement of the instrument to the
inferior angle of the descending duodenum. The amount of instrument that must be withdrawn in this
"straightening maneuver" varies but is often 30 cm or more. When the instrument is straight, only 55 to
60 cm remain within the patient. This can be noted by referring to the distance marks on the insertion
tube.
(2993)Figure 3825. Diagram showing a straightening maneuver. When performed in

conjunction with the maneuvers illustrated in Figure 3823, right rotation and withdrawal of
the instrument cause the tip of the insertion tube to move farther into the duodenum.
When the endoscope is deeply inserted and a large loop forms on the greater curvature of the
stomach, the degree of responsiveness of the insertion tube to rotation is greatly diminished, as
described earlier. Once the straightening maneuver is accomplished, however, the response to rotation
returns.
At the conclusion of the straightening maneuver, the instrument tip is near the inferior duodenal angle,
the insertion tube is relatively straight, and the endoscopist should have both hands at the control
section. The instrument should then be slowly withdrawn from the duodenum, making use of both
rotation and tip deflection to gain the best possible view. The difficulty this presents is that the tip may
suddenly fall back a considerable distance, even into the stomach, before an adequate examination
can be made. This occurs as the tip becomes hooked, especially at one of the duodenal angles or at
any point in the duodenal lumen that is more tortuous than usual.
The tendency for the insertion tube to slide suddenly back toward or into the stomach can be
anticipated by noting any resistance to withdrawal. This indicates that the tip has come to a turn in the
lumen and that by maintaining a view of the lumen with the deflection controls the endoscopist is also
maintaining a hook at the end of the endoscope. The tip should be deflected, usually with loss of the
view of the lumen, to release this hook. Further withdrawal is then possible without abruptly sliding
backward.
Once the instrument tip is in the descending duodenum, it is often possible to advance the instrument
by allowing a larger loop to form on the greater curvature. This often gives a somewhat different view
of the descending duodenum and may allow observation of areas that were not seen well during the
straightening/withdrawal maneuver. However, the formation of a large greater curvature loop may also
cause the instrument tip to move paradoxically in a backward direction as the insertion tube is
advanced.
In practice, it is often necessary to resort to both the withdrawn/straightened and greater curvature
looped positions, as well as to intermediate combinations of these two extremes, in order to piece
together a more complete picture of the descending duodenum. But even with maximum use of these
maneuvers, certain parts of the descending duodenum, particularly the medial wall, may not be seen
satisfactorily. The maneuver used in EGD for examination of the duodenum are very similar to those of
ERCP (see Chapter 57: Technique of Endoscopic Retrograde Cholangiopancreatography).
Distal Duodenum
Upper GI endoscopy is usually considered complete if the examination is carried out to the level of the
inferior duodenal angle. It can be difficult to insert the instrument tip into the third part of the duodenum
and extremely difficult or even impossible to reach the ligament of Treitz. Again, the reason for this is
the formation of the greater curvature loop in the stomach.
It is usually possible to obtain a view of the third portion from the inferior angle. When it is essential to
advance the tip into the third portion, it is usually necessary to form a loop of considerable size in the
stomach. This has several limitations. It usually causes discomfort to the patient, and the standard
panendoscope is not long enough to take great advantage of the forward motion achieved at the
expense of loop formation. In thin patients with lax abdominal wall muscles, the formation of the
greater curvature loop may be partially controlled by exerting external, upwardly directed pressure in
the left upper quadrant of the abdomen. Repeated straightening and loop formation may gradually
advance the instrument toward the ligament of Treitz in some cases. In a study by Brady et
al.,173(2994) fluoroscopy disclosed that endoscopists incorrectly estimated the position of the
endoscope within the duodenum in 47% of examinations. When incorrect, the distance the endoscope
had been advanced was overestimated about two thirds of the time. The endoscope was inserted to
the third part of the duodenum in 51% of cases and to the fourth portion or beyond in 38%. If an
examination of the distal duodenum or proximal jejunum is necessary, a longer endoscope such as a
pediatric colonoscope or a dedicated enteroscope should be used (see Chapter 50: Endoscopy of the
Small Intestine).
Retroversion (Retroflexion)
The endoscopist can withdraw the instrument from the level of the pylorus while keeping both hands at
the control section. This is done simply by lifting the left hand upward while stepping back from the
procedure table in slight increments. Keeping the right hand at the lateral deflection knob allows the
endoscopist to manipulate the distal tip in wide, circumferential arcs in order to view all areas of the
wall of the stomach. Electronic (video) endoscopes can be rotated by movement of the left hand in
wide circular arcs. It is not necessary to place the right hand back on the insertion tube until the
instrument has reached the cardia unless the retroversion maneuver is to be performed during this part
of the EGD.
The importance of the retroversion maneuver is that it provides an en face view of the angulus and the
fundus of the stomach. The region of the angulus is seen only tangentially on forward view. The fundus
is not seen at all.
To provide a view of the angulus (Figure 3826), the retroversion maneuver must be initiated at a point
opposite the angulus. This is facilitated by a partial loop on the greater curvature, which allows
maximum upward tip deflection without encountering the lesser curvature of the stomach with the tip of
the instrument.
(2995)Figure 3826. Retroflexed endoscopic view of the angulus. The antrum is toward the
bottom right, and the proximal stomach is toward the upper left.
Once in position opposite the angulus, the instrument tip is deflected upward as far as possible with the
left thumb. Generally, at least 180 degrees of tip deflection is required. Many instruments permit even
greater degrees of upward deflection. The instrument is then withdrawn while the tip is kept in this
attitude (Figure 3827). This provides a retroverted view of the instrument entering the stomach
through the cardia (Figure 3828). In addition to withdrawing the instrument, the endoscopist must add
a significant degree of rotation, as much as 180 degrees. This can be done to the left or right, but
rotation to the left is easier to accomplish. If left rotation is selected, the lateral deflection should be set
and locked full left (Figure 3829).
(2996)Figure 3827. Initial steps in the retroversion maneuver. The instrument tip has been
deflected fully in the upward direction. Note the position of the instrument tip at the midpoint
of the greater curvature as this maneuver is initiated.
(2997)Figure 3828. Retroverted endoscopic view of the fundus. The instrument can be seen
coming through the cardia, in the upper part of the image.
(2998)Figure 3829. Additional steps in the retroversion maneuver. The lateral deflection has
been set and locked full left, and the insertion tube has been rotated 180 degrees to the left.
Rotation and withdrawal should be accomplished simultaneously by degrees. The insertion tube must
be withdrawn with the right hand. In order to rotate to the left simultaneously, the left arm can be
abducted and externally rotated (see Figures 389 and 3810). If the left hand (holding the control
section) is brought toward the right hand at the same time and if the right hand simultaneously lifts the
insertion tube upward toward the oncoming left hand, a loop forms in the external configuration of the
instrument (Figure 3830). When this occurs, the insertion tube has in effect been rotated 180 degrees
to the left.
(2999)Figure 3830. Diagrams showing the formation of a full external left loop with a
fiberscope (A) and an electronic (video) endoscope (B). Formation of this loop results in
rotation of the portion of the insertion tube within the patient 180 degrees to the left.
The left rotation retroversion maneuver provides an excellent view of the fundus and cardia (see Figure
3828). This can be augmented by deflecting the tip to the left and right, using the right hand to
operate the lateral deflection knob. However, this view may still be incomplete, and it may also be
necessary to rotate the insertion tube to the right to complete the assessment of the area. This can be
done by simply undoing the 180-degree external left loop. To do so, it is necessary to raise the left
shoulder to the normal position, to step back from the procedure table, and to turn to the right (see
Figure 388). This must be done without further withdrawal (or advancement) of the insertion tube.
Because the right hand remains on the insertion tube throughout the retroversion maneuver, it is not
difficult to keep the insertion tube in place with respect to forward and backward motion even as it is
being rotated through 270 degrees or more.
The retroversion maneuver can be completed simply by moving the deflection controls to a neutral
position while the instrument tip is still in the proximal stomach. However, it is also advantageous to
return the tip to the proximal antrum while maintaining the retroverted view. This provides an additional
and somewhat different view of the proximal stomach and lesser curvature. To accomplish this, it is
only necessary to unrotate the instrument back to its usual position while advancing it. When the view
of the angulus is again obtained, the lateral deflection is returned to the neutral position with the right
hand and the tip is deflected downward with the left thumb to obtain a view of the pylorus. Withdrawal
to the level of the cardia then proceeds as described earlier.
Withdrawal of the Endoscope

When the tip of the instrument is returned to the cardia, it becomes more difficult to continue
withdrawing the instrument without placing the right hand on the insertion tube. The distance between
the endoscopist and the procedure table has become greater and the weight of the insertion tube
across this distance tends to cause it to slip backward out of the esophagus. Therefore, withdrawal of
the endoscope through the esophagus is performed with the right hand on the insertion tube. This still
allows a satisfactory range of tip motion by means of slight degrees of rotation with the right hand in
combination with up/down deflection by the left thumb. This provides very precise control of the tip and
allows reexamination of the structures of the posterior pharynx and larynx.
Air Insufflation and Suction

Deflation of the stomach as the instrument tip approaches the cardia during withdrawal diminishes the
patient's postprocedure discomfort. Air insufflation of the upper GI tract should be controlled precisely
during EGD. A common misconception of the beginning student is that further air insufflation is the
answer to all maneuvering problems. This may become so automatic and unconscious that the
instrument's air insufflation mechanism is activated throughout the procedure.
Overdistention of the bowel tends to make it difficult to view certain surfaces that are seen tangentially.
Suction and air insufflation can actually be used to manipulate the wall of the bowel. By collapsing the
bowel, areas that are seen poorly because of a tangential alignment with the instrument tip may be
seen more en face.
Gastrostomy
Endoscopy of the stomach, esophagus, and duodenum can be performed readily with a small-diameter
endoscope via a mature gastrostomy.174176(3000) This procedure has relatively few indications;
usually this access to the upper GI tract is used when some form of pathologic obstruction is present in
the esophagus or pharynx. The procedure should be performed under direct vision. Conscious
sedation is usually not required.
Postprocedure Recovery
Some aspects of the patient's recovery after a procedure are discussed in other chapters (see Chapter
5: The Gastrointestinal Assistant). The length of time required depends on the type and dosage of
sedation. Most patients rest or sleep for about 30 min after the intravenous administration of
meperidine, diazepam, or both. If a topical anesthetic was applied to the posterior pharynx, the patient
should not attempt to drink liquids until the effect of the agent dissipates. Generally, the patient may be
discharged after about 1 hr. He or she should be reminded not to drive or operate machinery for the
rest of the day.
REFERENCES
1.
Scott B, Atkinson M. Gastroenterology services: Regional review of changes over a five year
period (19811986). Gut 1989;30:695700.
2. Bramble MG. Open access endoscopyA nationwide survey of current practice. Gut
1992;33:2825.
3. Plumeri PA. Informed consent for gastrointestinal endoscopy in the '90s and beyond.
4. Plumeri PA. Informed consent for upper gastrointestinal endoscopy. Gastrointest Endosc Clin
North Am 1994;4:45561.
5. Levy N, Landmann L, Stermer E, et al. Does a detailed explanation prior to gastroscopy
reduce the patient's anxiety?. Endoscopy 1989;21:2635.
6. Wilson JF, Moore RW, Randolph S, Hanson BJ. Behavioral preparation of patients for
gastrointestinal endoscopy: Information, relaxation, and coping style. J Human Stress
1982;8:1323.
7. Smith MR, Bell GD, Fulton B, et al. A comparison of winged steel needles and Teflon
cannulas in maintaining intravenous access during gastrointestinal endoscopy. Gastrointest
Endosc 1993;39:336.
8. Shaker R. Unsedated trans-nasal pharyngoesophagogastroduodenoscopy (T-EGD):
Technique. Gastrointest Endosc 1994;40:3468.
9.
Webberley MJ, Cuschieri A. Response of patients to upper gastrointestinal endoscopy: Effect
of inherent personality traits and premedication with diazepam. BMJ [Clin Res]
1982;285:2512.
10.
Shapira M, Tamir A. The effect of accompanying persons on anxiety during gastroscopy
[Abstract]. Endoscopy 1992;24:612A.
11. Kingston RD. Sedation for upper gastrointestinal endoscopy. Gut 1991;32:8323.
12. Colin-Jones DG. Improving the standards of endoscopy. Gut 1991;32:7256.
13. Hedenbro JL, Lindblom A. Patient attitudes to sedation for diagnostic upper endoscopy.
14. Keeffe EB, O'Connor KW. 1989 ASGE survey of endoscopic sedation and monitoring
practices. Gastrointest Endosc 1990;36:S1318.
15. Daneshmend TK, Bell GD, Logan RFA. Sedation for upper gastrointestinal endoscopy:
Results of a nationwide survey. Gut 1991;32:125.
16. Thompson DG, Evans SJ, Murray RS, et al. Patients appreciate premedication for endoscopy.
Lancet 1980;2:46970.
17. Froehlich F, Schwizer W, Thorens J, et al. Conscious sedation for gastroscopy: Patient
tolerance and cardiorespiratory parameters. Gastroenterology 1995;108:697704.
18. Ross WA. Premedication for upper gastrointestinal endoscopy. Gastrointest Endosc
1989;35:1205.
19. Bell GD. Review article: Premedication and intravenous sedation for upper gastrointestinal
endoscopy. Aliment Pharmacol Ther 1990;4:10322.
20. Bell GD, McCloy RF, Charlton JE, et al. Recommendations for standards of sedation and
patient monitoring during gastrointestinal endoscopy. Gut 1991;32:8237.
21. Schutz SM, Lee JG, Schmitt CM, Baillie J. Patient satisfaction with conscious sedation for
22. Lauven PM. Pharmacology of drugs for conscious sedation. Scand J Gastroenterol Suppl
1990;179:16.
23. Cantor DS, Baldridge ET. Premedication with meperidine and diazepam for upper
gastrointestinal endoscopy precludes the need for topical anesthesia. Gastrointest Endosc
1986;32:33941.
24. Chuah SY, Crowson CP, Dronfield MW. Topical anaesthesia in upper gastrointestinal
endoscopy. BMJ 1991;303:695.
25. Lachter J, Jacobs R, Lavy A, et al. Topical pharyngeal anesthesia for easing endoscopy: A
double-blind, randomized, placebo-controlled study. Gastrointest Endosc 1990;36:1921.
26. Gordon MJ, Mayes GR, Meyer GW. Topical lidocaine in preendoscopic medication.
27. Leitch DG, Wicks J, el Beshir OA, et al. Topical anesthesia with 50 mg of lidocaine spray
facilitates upper gastrointestinal endoscopy. Gastrointest Endosc 1993;39:3847.
28. Hedenbro JL, Ekelund M, Jansson O, Lindblom A. A randomized, double-blind,
placebo-controlled study to evaluate topical anaesthesia of the pharynx in upper
gastrointestinal endoscopy. Endoscopy 1992;24:5857.
29. Archimandritis A, Tjivras M, Tryphonos M, et al. Is premedication necessary in diagnostic
endoscopy of the upper gastrointestinal tract?. Endoscopy 1991;23:240.
30. Al-Atrakchi HA. Upper gastrointestinal endoscopy without sedation: A prospective study of
2000 examinations. Gastrointest Endosc 1989;35:7981.
31. Keefe EB. Sedation and analgesia for endoscopy. Gastroenterology 1995;108:9323.
32. Smith JL, Opekun A, Graham DY. Controlled comparison of topical anesthetic agents in
flexible upper gastrointestinal endoscopy. Gastrointest Endosc 1985;31:2558.
33.
Jameson JS, Kapadia SA, Polson RJ, et al. Rational use of topical anesthesia in upper
gastrointestinal endoscopy [Abstract]. Gastroenterology 1992;102:14A.
34. le Lorier J, Larochelle P, Bolduc P, et al. Lidocaine plasma concentrations during and after
endoscopic procedures. Int J Clin Pharmacol Biopharm 1979;17:535.
35. Soda K, Shimanuki K, Yoshida Y, et al. Serum lidocaine and MEGX concentrations after
pharyngeal anesthesia for gastroscopy. Endoscopy 1994;26:34751.
36. Patel D, Chopra S, Berman MD. Serious systemic toxicity resulting from use of tetracaine for
pharyngeal anesthesia in upper endoscopic procedures. Dig Dis Sci 1989;34:8824.
37. Zimmerman J, Rachmilewitz D. Systemic anaphylactic reaction following lidocaine
administration. Gastrointest Endosc 1985;31:4045.
38. Grum DF, Rice TW. Methemoglobinemia from topical benzocaine. Cleve Clin J Med
1990;57:3579.
39. Buckley AB, Newman A. Methemoglobinemia occurring after the use of a 20% benzocaine
topical anesthetic prior to gastroscopy. Gastrointest Endosc 1987;33:4667.
40. O'Donohue WJ Jr, Moss LM, Angelillo VA. Acute methemoglobinemia induced by topical
benzocaine and lidocaine. Arch Intern Med 1980;140:15089.
41. Collins JF. Methemoglobinemia as a complication of 20% benzocaine spray for endoscopy.
42.
Scott EM, Hoskins DD. Hereditary methemoglobinemia in Alaskan Eskimos and Indians.
Blood 1958;13:795802.
43. Brown CM, Levy SA, Susann PW. Methemoglobinemia: Life-threatening complication of
endoscopy premedication. Am J Gastroenterol 1994;89:11089.
44. Dinneen Sf, Mohr DN, Fairbanks VF. Methhemoglobinemia from topically applied anesthetic
spray. Mayo Clin Proc 1994;69:8868.
45. McCloy RF, Pearson RC. Which agent and how to deliver it? A review of benzodiazepine
sedation and its reversal in endoscopy. Scand J Gastroenterol Suppl 1990;179:711.
46. Castiglioni LJ, Allen TS, Patterson M. Intravenous diazepam: An improvement in
pre-endoscopic medication. Gastrointest Endosc 1973;19:1346.
47. Stephens MJ, Jakobovits AW, Dudley FJ. A controlled trial of phenoperidine and diazepam in
48. Longstreth GFU, O'Brien PM, Youkeles LF. Determinants of the intravenous diazepam dose
required for gastroscopy. Gastrointest Endosc 1980;26:924.
49.
Hegarty JE, Dundee JW. Sequelae after the intravenous injection of three
benzodiazepinesDiazepam, lorazepam and flunitrazepam. BMJ 1977;2:13845.
50. Bianchi Porro G, Lazzaroni M, Parente F, et al. Diazemuls versus diazepam as premedication
in upper gastrointestinal endoscopy: A randomised crossover study. Endoscopy
1991;23:2035.
51.
Heizmann P, Eskert M, Ziegler WH. Pharmacokinetics and bioavailability of midazolam in
man. Br J Clin Pharmacol 1983;16:S4349.
52. Bell GD, Spickett GP, Reeve PA, et al. Intravenous midazolam for upper gastrointestinal
endoscopy: A study of 800 consecutive cases relating dose to age and sex of patient. Br J
Clin Pharmacol 1987;23:2413.
53. MacGilchrist AJ, Birnie GG, Cook A, et al. Pharmacokinetics and pharmacodynamics of
intravenous midazolam in patients with severe alcoholic cirrhosis. Gut 1986;27:1905.
54. Kawar P, Porter KG, Hunter EK, et al. Midazolam for upper gastrointestinal endoscopy. Ann R
Coll Surg Engl 1984;66:2835.
55. Bell GD, Morden A, Coady T, et al. A comparison of diazepam and midazolam as endoscopy
premedication: Assessing changes in ventilation and oxygen saturation. Br J Clin Pharmacol
1988;26:595600.
56. Whitwam JG, Al-Khudhairi D, McCloy RF. Comparison of midazolam and diazepam in doses
of comparable potency during gastroscopy. Br J Anaesth 1983;55:7737.
57. Sanders LD, Davies-Evans J, Rosen M, Robinson JO. Comparison of diazepam with
midazolam as I.V. sedation for outpatient gastroscopy. Br J Anaesth 1989;63:72631.
58. Berggren L, Eriksson I, Mollenholt P, Wickbom G. Sedation for fibreoptic gastroscopy: A
comparative study of midazolam and diazepam. Br J Anaesth 1983;55:28996.
59. Bianchi Porro G, Baroni S, Parente F, Lazzaroni M. Midazolam versus diazepam as
premedication for upper gastrointestinal endoscopy: A randomized, double-blind, crossover
60. Al-Khudhairi D, Whitwam JG, McCloy RF. Midazolam and diazepam for gastroscopy.
Anaesthesia 1982;37:10026.
61.
Ginsberg GG, Lewis JH, Gallager J, et al. Midazolam (MZ) vs diazepam (DZ) for colonoscopy:
An evaluation of predicted and actual dosing requirements [Abstract]. Gastroenterology
1992;102:11A.
62. Lee MG, Hanna W, Harding H. Sedation for upper gastrointestinal endoscopy: A comparative
study of midazolam and diazepam. Gastrointest Endosc 1989;35:824.
63. Gilvarry JM, Craig M, Fielding JF. Short report: Sedation for upper gastrointestinal
endoscopyDiazepam versus midazolam. Aliment Pharmacol Ther 1990;4:4235.
64. Arrowsmith JB, Gerstman BB, Fleischer DE, et al. Results from the American Society for
Gastrointestinal Endoscopy/US Food and Drug Administration collaborative study on the
1991;37:4217.
65. Van Der Bijl P, Roelofse JA, Joubert JJ De V, Van Zyl JF. Comparison of various physiologic
and psychomotor parameters in patients sedated with intravenous lorazepam, diazepam, or
midazolam during oral surgery. J Oral Maxillofac Surg 1991;49:6728.
66.
Krugliak P, Bar-Am I, Serov B, et al. Flunitrazepam versus midazolam as premedication for
upper gastrointestinal endoscopy: A double-blind randomized study [Abstract]. Gastrointest
Endosc 1992;38:228A.
67. Bell GD, Antrobus JHL, Lee J, et al. Bolus or slow titrated injection of midazolam prior to
upper gastrointestinal endoscopy? Relative effect on oxygen saturation and prophylactic value
of supplemental oxygen. Aliment Pharmacol Ther 1990;4:393401.
68. Swain DG, Ellis DJ, Bradby H. Rapid intravenous low-dose diazepam as sedation for upper
gastrointestinal endoscopy. Aliment Pharmacol Ther 1990;4:438.
69. Hedenbro JL, Ekelund M, Aberg T, Lindblom A. Oral sedation for diagnostic upper endoscopy
[see comments]. Endoscopy 1991;23:810.
70. Honan VJ. Paradoxical reaction to midazolam and control with flumazenil. Gastrointest
Endosc 1994;40:868.
71. Carrougher JG, Kadakia S, Shaffer RT, Barrilleaux C. Venous complications of midazolam
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
versus diazepam. Gastrointest Endosc 1993;39:3969.

Bonetti EP, Pieri L, Cumin R, et al. Benzodiazepine antagonist Ro 151788: Neurological and
behavioral effects. Psychopharmacology (Berlin) 1982;78:818.
Carter AS, Bell GD, Coady T, et al. Speed of reversal of midazolam-induced respiratory
depression by flumazenilA study in patients undergoing upper G.I. endoscopy. Acta
Anaesthesiol Scand Suppl 1990;92:5964; discussion 78.
al-Quorain A. The use of flumazenil in reversing the midazolam and diazepam sedation in
outpatients undergoing gastroscopy. Trop Gastroenterol 1993;14:514.
Lamy V, Finne R, Stenier P. Clinical evaluation of the short half-life injectable benzodiazepine
midazolam, and of the specific antagonist flumazenil in upper digestive tract endoscopy. Acta
Gastroenterol Belg 1991;54:31527.
Rosario MT, Costa NF. Combination of midazolam and flumazenil in upper gastrointestinal
endoscopy, a double-blind, randomized study. Gastrointest Endosc 1990;36:303.
Dunk AA, Norton AC, Hudson M, et al. The value of flumazenil in the reversal of
midazolam-induced sedation for upper gastrointestinal endoscopy. Aliment Pharmacol Ther
1990;4:3542.
Bartelsman JFWM, Sars PRA, Tytgat GNJ. Flumazenil used for reversal of midazolam
induced sedation in endoscopy outpatients. Gastrointest Endosc 1990;36:S912.
Birkenfeld S, Federico C, Dermansky-Avni Y, et al. Double-blind controlled trial of flumazenil
in patients who underwent upper gastrointestinal endoscopy. Gastrointest Endosc
1989;35:51922.
Jensen S, Knudsen L, Kirkegaard L, et al. Flumazenil used for antagonizing the central effects
of midazolam and diazepam in outpatients. Acta Anaesthesiol Scand 1989;33:268.
Ching CK, Holmes GKT. Flumazenil in routine upper gastrointestinal endoscopy. Gastrointest
Endosc 1990;36:634.
Sanders LD, Piggott SE, Isaac PA, et al. Reversal of benzodiazepine sedation with the
antagonist flumazenil. Br J Anaesth 1991;66:44553.
Pearson RC, McCloy RF, Morris P, Bardhan KD. Midazolam and flumazenil in
gastroenterology. Acta Anaesthesiol Scand Suppl 1990;92:214; discussion 47.
Benjamin SB, Ginsberg GG, Gallagher J, et al. Preliminary evaluation of flumazenil (FMZ) to
reverse post-endoscopic conscious sedation (CS) [Abstract]. Gastrointest Endosc
1992;38:226A.
Cooper SA, Quinn PD, MacAfee K, McKenna D. Reversing intravenous sedation with
flumazenil. Oral Surg Oral Med Oral Pathol 1991;72:29.
Young ER, Kestenberg SH, Toal CB. The effect of flumazenil on the recovery time of dental
patients sedated with diazepam. Anesth Prog 1989;36:4651.
Herman RJ, McAllister CB, Branch RA, Wilkinson GR. Effects of age on meperidine
disposition. Clin Pharmacol Ther 1985;37:1923.
Stephens MJ, Gibson PR, Jakobovits AW, et al. Fentanyl and diazepam in endoscopy of the
upper gastrointestinal tract. Med J Aust 1982;1:41920.
Ishido S, Knoshita Y, Kitajima N, et al. Fentanyl for sedation during upper gastrointestinal
Nguyen L, Chokhavatia S, Maunuksela EL, Heavner JE. Alfentanil analgesia during
colonoscopy [Abstract]. Endoscopy 1992;24:612A.
Handal KA, Schauben JL, Salamone FR. Naloxone. Ann Emerg Med 1983;12:43845.
Jurell KR, O'Connor W, Slack J, et al. Effect of supplemental oxygen on cardipulmonary
changes during gastrointestinal endoscopy. Gastrointest Endosc 1994;40:66570.
Murray AW, Morran CG, Kenny GNC, Anderson JR. Arterialoxygen saturation during upper
gastrointestinal endoscopy and the effects of midazolam/pethidine combination. Gut
1990;31:2703.
Hartke RH, Gonzalez-Rothi RJ, Abbey NC. Midazolam-associated alterations in
cardiorespiratory function during colonoscopy. Gastrointest Endosc 1989;35:2328.
Bailey PL, Pace NL, Ashburn MA, et al. Frequent hypoxemia and apnea after sedation with
midazolam and fentanyl. Anesthesiology 1990;73:82630.

96. Walton GM, Boyle CA, Thompson PJ. Changes in oxygen saturation using two different
sedation techniques. Br J Oral Maxillofac Surg 1991;29:879.
97. Cousins MJ. MonitoringThe anaesthetist's view. Scand J Gastroenterol Suppl
1990;179:127.
98.
Murray AW, Kenny GC. Monitoring during sedation and endoscopy. BMJ 1988;297:1337.
99. Al-Hadeedi S, Leaper DJ. Falls in hemoglobin saturation during ERCP and upper
gastrointestinal endoscopy. World J Surg 1991;15:8894.
100. Cattau EL Jr, Artnak EJ, Castell DO, Meyer GW. Efficacy of atropine as an endoscopic
premedication. Gastrointest Endosc 1983;29:2858.
101. Cook PJ, Bennett PN, Lennard-Jones JE, Warnes TW. Premedication for endoscopy: A trial
of atropine, pentazocine or pethidine as a supplement to diazepam. Scand J Gastroenterol
1978;13:339.
102. Hedenbro JL, Frederiksen SG, Lindbloom A. Anticholinergic medication in diagnostic
endoscopy of the upper gastrointestinal tract. Endoscopy 1991;23:199202.
103. Qvigstad T, Larsen S, Myren J. Comparison of glucagon, atropine, and placebo as
premedication for endoscopy of the upper gastrointestinal tract. Scand J Gastroenterol
1979;14:2315.
104. Schwartz SE, Fazio TL. Pre-endoscopic medication. A randomized double-blind trial of
atropine and meperidine as a supplement to diazepam. Scand J Gastroenterol
1979;14:74751.
105. Gerner T, Myren J, Larsen S. Premedication in upper gastrointestinal endoscopy. A
comparison of glucagon and atropine given in combination with diazepam and pethidine.
106. Patterson KW, Casey PB, Murray JP, et al. Propofol sedation for outpatient upper
gastrointestinal endoscopy: Comparison with midazolam. Br J Anaesth 1991;67:10811.
107. Carlsson U, Grattidge P. Sedation for upper gastrointestinal endoscopy: A comparative study
of propofol and midazolam. Endoscopy 1995;27:2405.
108. Barthel JS, Marshall JB, King PD, et al. The effect of droperidol on objective markers of
patient cooperation, and vital signs during esophagogastroduodenoscopy: A randomized,
double-blind, placebo-controlled prospective investigation. Gastrointest Endosc
1995;42:4550.
109. Pound DC, O'Connor KW, Brown ED, et al. Oral medications for upper gastrointestinal
endoscopy using a small diameter endoscope. Gastrointest Endosc 1988;34:32731.
110. Cavallo G, Cuomo R, Viscardi A, et al. Hypnosis for upper gastrointestinal endoscopy.
111. DeMeester TR, Skinner DB, Evans RH, Benson DW. Local nerve block anesthesia for peroral
endoscopy. Ann Thorac Surg 1977;24:27883.
112. Cahn AM, Carayon P, Hill C, Flamant R. Acupuncture in gastroscopy. Lancet 1978;1:1823.
113. Chu H, Zhao SZ, Huang YY. Application of acupuncture to gastroscopy using a fibreoptic
endoscope. J Tradit Chin Med 1987;7:279.
114. Escher J, Hohmann U, Anthenien L, et al. Music during gastroscopy. Schweiz Med
Wochenschr 1993;123:13548.
115. Bell GD. MonitoringThe gastroenterologist's view. Scand J Gastroenterol Suppl
1990;179:1823.
116. Schiller KFR, Cotton PB, Salmon PR. The hazards of digestive fibre-endoscopy: A survey of
British experience [Abstract]. Gut 1972;13:1027.
117. Silvis SE, Nebel O, Rogers G, et al. Endoscopic complications. Results of the 1974 American
Society for Gastrointestinal Endoscopy Survey. JAMA 1976;235:92830.
118. Davis RE, Graham DY. Endoscopic complications: The Texas experience. Gastrointest
Endosc 1979;25:1469.
119.
Nelson DB, Silvis SE, Yakshe PN, et al. Inaccuracy of clinical assessment in detecting
hypoxemia and hypoventilation during ERCP [Abstract]. Gastrointest Endosc 1992;38:227A.
120. Freeman ML, Hennessy JT, Cass OW, Pheley AM. Carbon dioxide retention and oxygen
desaturation during gastrointestinal endoscopy. Gastroenterology 1993;105:3319.
121.
Patterson KW, Noonan N, Keeling N, et al. Hypoxemia during upper gastrointestinal
endoscopy: Effects of sedation and esophageal intubation [Abstract]. Gastroenterology
1992;102:143A.
122.
Dusdieker N, Ruckdaschel J, Marshall V, Murph E. Clinical usefulness of automated monitors
during inpatient endoscopy [Abstract]. Gastrointest Endosc 1992;38:227A.
123. Iber FL, Sutberry M, Gupta R, Kruss D. Evaluation of complications during and after
conscious sedation for endoscopy using pulse oximetry. Gastrointest Endosc 1993;39:6205.
124. Visco DM, Tolpin E, Straughn JC, Fagraeus L. Arterial oxygen saturation in sedated patients
undergoing gastrointestinal endoscopy and a review of pulse oximetry. Del Med J
1989;61:53342.
125. Council on Scientific Affairs, American Medical Association. The use of pulse oximetry during
conscious sedation. JAMA 1993;270:14638.
126. Alexander CM, Tella LE, Gross JB. Principles of pulse oximetry: Theoretical and practical
considerations. Anesth Analg 1989;68:36876.
127. Tremper KK, Barker SJ. Pulse oximetry. Anesthesiology 1989;70:98108.
128.
Taylor MB, Whitman JG. The current status of pulse oximetry. Anesthesiology
1986;41:9439.
129. Eisenkraft JB. Carboxyhemoglobin and pulse oximetry. Anesthesiology 1988;68:3001.
130. Constarino AT, Davis SA, Keon TP. Falsely normal saturation reading with the pulse oximeter.
Anesthesiology 1987;67:8301.
131. Huch A, Huch R, Konig W, et al. Limitations of pulse oximetry. Lancet 1988;1:3578.
132. Cote CJ, Goldenstein EA, Fuchsman WH, Hoaglin DC. The effect of nail polish on pulse
oximetry. Anesth Analg 1988;67:6836.
133. Russell NJ. Monitoring during sedation for endoscopy. BMJ 1988;297:978.
134. Wilcox CM, Faibicher M, Wenger NK, Shalek KA. Prevalence of silent myocardial ischemia
and arrhythmias in patients with coronary heart disease undergoing gastrointestinal tract
endoscopic procedures. Arch Intern Med 1993;153:232530.
135. Dark DS, Campbell DR, Wesselius LJ. Arterial oxygen desaturation during gastrointestinal
136. Clark R, Goy J. Sedation for upper gastrointestinal endoscopy. Gut 1991;32:8323.
137. Fleischer D. Monitoring during gastrointestinal endoscopy: How much is necessary?.
Endoscopy 1991;23:3445.
138. Church JA, Kenny GNC, Anderson JR. Monitoring during endoscopy. Gastrointest Endosc
1990;36:5434.
139. Murray AW, Kenny GC. Sedation for endoscopy. Br J Anaesth 1989;63:2447.
140. Fleischer D. Monitoring the patient receiving conscious sedation for gastrointestinal
endoscopy: Issues and guidelines. Gastrointest Endosc 1989;35:2626.
141. The American Society for Gastrointestinal Endoscopy. Guidelines for clinical application:
Monitoring of patients undergoing gastrointestinal endoscopic procedures. ASGE Publication
No. 1022. Gastrointest Endosc 1991;37:1201.
142. Froehlich F, Gonvers JJ, Fried M. Conscious sedation, clinically relevant complications and
monitoring of endoscopy: Results of a nationwide survey in Switzerland. Endoscopy
1994;26:2314.
143. DiSario JA, Waring JP, Talbert G, et al. Monitoring of blood pressure and heart rate during
routine endoscopy: A prospective, randomized, controlled study. Am J Gastroenterol
1991;86:95660.
144. Simon IB, Lewis RJ, Satava RM. A safe method for sedating and monitoring patients for
upper and lower gastrointestinal endoscopy. Am Surg 1991;57:21921.
145. Iwao T, Toyonaga A, Harada H, et al. Arterial oxygen desaturation during non-sedated
diagnostic upper gastrointestinal endoscopy. Gastrointest Endosc 1994;40:27780.
146. Lieberman DA, Wuerker CK, Katon RM. Cardiopulmonary risk of
esophagogastroduodenoscopy. Role of endoscope diameter and systemic sedation.

147. Rostykus PS, McDonald GB, Albert RK. Upper intestinal endoscopy induces hypoxemia in
patients with obstructive pulmonary disease. Gastroenterology 1980;78:48891.
148. Murray AW, Morran CG, Kenny P, et al. Examination of cardiorespiratory changes during
upper gastrointestinal endoscopy. Comparison of monitoring arterial oxygen saturation,
arterial pressure, and the electrocardiogram. Anaesthesia 1991;46:1814.
149. Vasquez-Olivencia W, Badiga MS, Pitchumoni CS. Monitoring during gastrointestinal
150. Fennerty MB, Earnest DL, Hudson PB, Sampliner RE. Physiologic changes during
colonoscopy. Gastrointest Endosc 1990;36:225.
151. Berg JC, Miller R, Burkhalter E. Clinical value of pulse oximetry during routine diagnostic and
therapeutic endoscopic procedures. Endoscopy 1991;23:32830.
152. Dhariwal A, Plevris JN, Lo NTC, et al. Age, anemia, and obesity-associated oxygen
desaturation during upper gastrointestinal endoscopy. Gastrointest Endosc 1992;38:6848.
153. Block R, Jankowski J, Johnston D, et al. The administration of supplementary oxygen to
prevent hypoxia during upper alimentary endoscopy. Endoscopy 1993;25:26973.
154. Lavies NG, Creasy T, Harris K, Hanning CD. Arterial oxygen saturation during upper
gastrointestinal endoscopy: Influence of sedation and operator experience. Am J
155. Rimmer KP, Graham K, Whitelaw W, Field SK. Mechanisms of hypoxemia during
panendoscopy. J Clin Gastroenterol 1989;11:1722.
156.
McKee CC, Ragland JJ, Myers JO. An evaluation of multiple clinical variables for hypoxia
during colonoscopy. Surg Obstet Gynecol 1991;173:3740.
157. Bell GD. Who is for supplemental oxygen?. Gastrointest Endosc 1992;38:5146.
158. Hart R, Classen M. Complications of diagnostic gastrointestinal endoscopy. Endoscopy
1990;22:22933.
159. Bell GD, Bown S, Morden A, et al. Prevention of hypoxaemia during upper-gastrointestinal
endoscopy by means of oxygen via nasal cannulae. Lancet 1987;1:10224.
160. Bell GD, Quine A, Antrobus JHL, et al. Upper gastrointestinal endoscopy: A prospective
randomized study comparing continuous supplemental oxygen via the nasal or oral route.
161. Griffin SM, Chung SCS, Leung JWC, et al. Effect of intranasal oxygen on hypoxia and
tachycardia during endoscopic cholangiopancreatography. BMJ 1990;300:834.
162. Reed MW, O'Leary DP, Duncan JL, et al. Effects of sedation and supplemental oxygen during
upper alimentary tract endoscopy. Scand J Gastroenterol 1993;28:31922.
163. Hill DB, Stokes BD, Gilinsky NH. Arterial oxygen saturation during emergency
esophagogastroduodenoscopy. The effects of nasal oxygen. J Clin Gastroenterol
1994;18:2846.
164. al-Qorain A, Adu-Gyamfi Y, Larbi EB, al-Shedokhi F. The effect of supplemental oxygen in
sedated and unsedated patients undergoing upper gastrointestinal endoscopy. J Int Med Res
1993;21:16570.
165. Bowling TE, Hadjiminas CL, Polson RJ, et al. Effects of supplemental oxygen on cardiac
rhythm during upper gastrointestinal endoscopy: A randomised controlled double blind trial.
Gut 1993;34:14927.
166. Bell GD, Antrobus JHL, Lee J, et al. Pattern of breathing during upper gastrointestinal
endoscopy: Implications for administration of supplemental oxygen. Aliment Pharmacol Ther
1991;5:399404.
167. Hebbard GS, Royse CF, Bjorksten AR. Oxygen supplementation during upper gastrointestinal
endoscopy: A comparison of two methods. Endoscopy 1994;26:27882.
168. Brandl S, Borody TJ, Andrews P, et al. Oxygenating mouthguard alleviates hypoxia during
gastroscopy. Gastrointest Endosc 1992;38:4157.
169. Borody TJ. Methods of oxygen delivery during upper gastrointestinal endoscopy. Endoscopy
1994;26:3201.
170. Dabezies M. Esophageal intubation over a guide wire. Gastrointest Endosc 1993;39:597.
171. Tsang TK, Buto SK. Catheter-guided endoscopic intubation: A new technique for intubating a
difficult esophagus. Gastrointest Endosc 1992;38:4951.
172. Straker RJ, Bienvenu JC, Nord HJ. Endoscopic orientation within the duodenal bulb.
Endoscopy 1992;24:2667.
173. Brady CE III, Stewart DL, DiPalma JA, et al. Upper gastrointestinal endoscopyHow far does
the endoscope go?. Gastrointest Endosc 1985;31:3679.
174. Hentschel E, Dittrich H, Weiss W. Retrograde oesophagoscopy via gastrostomy. Endoscopy
1976;8:3840.
175. Gleason WA Jr, Danis RK. Gastroscopy per gastrostomy. Another indication for the use of a
small diameter flexible fiberoptic endoscope. Gastrointest Endosc 1976;23:767.
176. McDonald GB, Silverstein FE. Retrograde esophagoscopy with a fiberoptic endoscope.
Chapter 39 Esophageal Motility and Miscellaneous Disorders*(3001)

(3002)
JOSEPH R. MURPHY, M.D.
LAWRENCE F. JOHNSON, M.D.
This chapter deals with the endoscopic evaluation of esophageal motility disorders and other
miscellaneous conditions involving the esophagus. Although these disorders may be associated with
symptoms such as chest pain, dysphagia, and regurgitation, these symptoms are nonspecific. Thus,
most patients with these disorders undergo endoscopy to exclude benign and malignant obstructive
and mucosal diseases that cause similar symptoms. Endoscopy is very effective as a diagnostic tool to
exclude obstructive and mucosal abnormalities but is of limited value in assessing disorders of
esophageal peristalsis and emptying. Therefore, the evaluation of these conditions should include
other investigative techniques, such as a barium esophagogram, esophageal sonography,
scintigraphy, and esophageal manometry.
Esophageal Motility Disorders

Achalasia
Achalasia is a motility disorder characterized by loss of esophageal peristalsis and usually a
hypertensive lower esophageal sphincter (LES) that relaxes incompletely with swallowing, both of
which result in impaired esophageal emptying.1,2(3003) To some degree, the clinical presentation
varies according to the point in the course of the disease at which medical attention is sought.3(3004)
Dysphagia and chest pain occur early, whereas regurgitation and weight loss are more prominent with
longstanding disease. Our clinical observations of 34 patients with newly diagnosed achalasia who
were studied at Walter Reed Army Medical Center not only reflect the clinical experience of others but
also provide further insight.4(3005) All patients had dysphagia for solid food, and the majority also had
dysphagia for liquids. Regurgitation of food that tasted undigested, with neither the acidic taste of
gastric contents nor the bitter taste of bile, was noted by 66% of the patients. In addition, 42%
complained of pyrosis and substernal chest pain despite the absence of objective evidence of
gastroesophageal reflux (GER). Fermented intraesophageal contents have been shown to have a low
pH by monitoring, and perhaps this is the cause of pyrosis in these patients.5(3006) However, unlike
the pyrosis associated with reflux, that associated with achalasia tends to occur at times other than the
immediate postprandial period, to be unrelieved by antacid, and frequently to awaken patients at night.
In some patients with achalasia, pyrosis may occasionally be the predominant symptom. Ellis et
al.6(3007) reported four cases of achalasia in which surgery was performed a second time because of
a misdiagnosis of GER.
Weight loss, a finding previously emphasized in secondary achalasia,7,8(3008) was seen in 84% of
patients. We believe that weight loss was the best parameter for assessment of the chronicity and
severity of the disease. It directly correlated with the degree of abnormal esophageal emptying, as
assessed by radionuclide esophageal transit. Effective treatment, which resulted in decreased
symptoms and improved esophageal emptying, was always accompanied by subsequent weight gain.
Diagnosis
Barium swallow radiographs, radioisotope scintigraphy, esophageal motility, and esophagoscopy are
complementary procedures in the search for suspected achalasia. Characteristic roentgenographic
findings include symmetric dilation and tortuosity of the esophagus. The distal esophagus tends to be
smooth and symmetric, tapering to a "bird beak" configuration (Figure 391). At fluoroscopy, no
peristalsis is seen in the body of the esophagus, although repetitive nonperistaltic contractions may be
observed. The esophageal mucosa should appear smooth; if surface irregularity or nodularity is
present, malignancy should be considered.
(3009)Figure 391. Barium swallow x-ray demonstrates the tapering "bird beak"
configuration of the distal esophagus in achalasia.
Endoscopic ultrasonography has demonstrated widening or "thickening" of the fourth hyperechoic
sonographic layer (muscularis propria) in achalasia (Figure 392A), whereas thickening of the second
and third hypoechoic layers (mucosa and submucosa) is very suggestive of an infiltrating lesion (Figure
392B) (see Chapter 41: Endoscopic Ultrasonography of the Esophagus).9(3010)
(3011)Figure 392. Endoscopic ultrasonography. A, Enlargement of the fourth hyperechoic

area (arrow) representing a thickened muscularis propria at the level of the cardia in a patient
with achalasia. B, Pseudoachalasia due to gastric carcinoma, with thickening of the second and
third hypoechoic layers (arrows) representing submucosal tumoral infiltration arising from the
stomach. (A and B, Courtesy of Dr. J. Deviere.)
Radioscintigraphy performed with various isotope meals demonstrates a dilated, adynamic
esophagus.10(3012) In addition, esophageal transit time of the isotope is markedly prolonged,
indicative of the impaired esophageal emptying characteristic of this disease.11(3013)
Esophageal manometry should be performed to confirm the diagnosis. Absence of peristalsis is the
characteristic finding. Other manometric findings often present include (1) hypertensive LES pressure,
(2) incomplete relaxation of the sphincter with swallowing, and (3) resting pressure in the esophageal
body that is greater than gastric pressure.
Endoscopic Findings
Esophageal lavage using a large-bore tube should be performed prior to endoscopy in patients with
achalasia because retained solid and liquid material may be present in the esophagus despite
prolonged fasting (Figure 393). Lavage decreases the risk of pulmonary aspiration of this retained
material during the procedure.
(3014)Figure 393. Barium swallow x-ray showing a dilated tortuous esophagus containing
food and debris.
The dilated esophagus in the patient with longstanding achalasia may appear cavernous, simulating
the appearance of the colon at colonoscopy (Figure 394). In patients with a tortuous "sigmoid"
esophagus, endoscopy may be difficult because of the numerous bends and turns. Most patients with
achalasia have normal esophageal mucosa. However, as the esophagus dilates, mild diffuse erythema
may be noted distally. Whitish mucosal plaques, erosions, and discrete ulcers, presumably the result of
stasis of ingested material, may also be seen. Air insufflation may initiate distal esophageal
contractions.
(3015)Figure 394. A, Tortuous "sigmoid" endoscopic appearance. B, Dilated esophagus

containing food material. C, Impacted food bolus in the distal esophagus of a patient with
early achalasia. (B and C, From the collection of Dr. M. V. Sivak, Jr.)
As the endoscope is advanced from the distal esophagus to the stomach, mild resistance is often
encountered. However, with gentle forward pressure this can be overcome, and the instrument should
readily "pop" into the stomach. Undue difficulty in traversing the gastroesophageal junction suggests
the presence of underlying malignancy or a benign stricture. After esophageal dilation, the
gastroesophageal junction may occasionally become eccentrically positioned, preventing easy
intubation of the stomach.
The endoscopist should conduct a precise appraisal for a hiatus hernia, even though this is unusual in
achalasia, because esophageal perforation during pneumatic dilation is more common if a hiatus
hernia is present.12,13(3016) The gastric cardia must be evaluated using forward and retroflexed
views to exclude carcinoma of this area, a cause of secondary achalasia. Biopsies must be obtained of
any mucosal abnormalities noted in the cardia, especially in patients over the age of 50 and in those
with progressive symptoms of short duration.
Differential Diagnosis
Gastric adenocarcinoma with orad esophageal extension is the most common malignancy that mimics
achalasia, termed secondary achalasia or pseudoachalasia in such cases. Other tumors reported to
cause secondary achalasia, either by direct esophageal involvement or as a paraneoplastic
phenomenon, include oat cell and squamous cell carcinomas of the lung, reticulum cell sarcoma,
prostatic carcinoma, pancreatic cancer, Hodgkin's and non-Hodgkin's lymphoma, hepatocellular
carcinoma, mesothelioma, and esophageal leiomyomatosis.7,8,1420(3017) Secondary achalasia is
not always associated with tumor invasion of the LES or with malignancy. Sarcoidosis, amyloidosis,
eosinophilic esophagitis, and pancreatic pseudocysts have all been reported to imitate
achalasia.2124(3018) The esophageal manometric findings in primary and secondary achalasia are
virtually identical.25(3019)
Patients with secondary achalasia due to gastric adenocarcinoma tend to be older and to have
symptoms of relatively short duration along with weight loss.25,26(3020) Although relatively sensitive
and specific for secondary achalasia, these three clinical features are not completely reliable.8(3021)
Rozman and Achkar26(3022) compared the clinical presentations of 18 patients with secondary
achalasia due to cancer with those of 421 patients with idiopathic or primary achalasia. Cancer arose
at the gastroesophageal junction in 16 patients; other primary sites included duodenum and breast.
The mean age of patients in the former group was 57.1 years (range 15 to 78 years), versus 47.1
years (range 1 to 90 years) for patients with primary achalasia. However, 3 patients with secondary
achalasia were less than 40 years of age. The majority of patients with secondary achalasia had
symptoms for 6 months or less, and 88.2% had experienced weight loss, compared with 57.3% of
patients with primary achalasia. Endoscopy demonstrated cancer in two thirds of cases, whereas the
esophagogram suggested this diagnosis in only one quarter. Endoscopic findings that suggest
malignancy include mucosal nodularity or ulceration, decreased compliance of the gastroesophageal
junction, and resistance to passage of the endoscope into the stomach.

Chagas' disease, caused by infection with Trypanosoma cruzi, is generally confined to South and
Central America. Esophageal disease that is radiographically and manometrically indistinguishable
from achalasia can develop in patients with chronic infection.27(3023) Unlike achalasia, however,
megacolon, megaureter, and myocardial disease may also develop in patients with Chagas'
disease.28(3024)
Scleroderma and other connective tissue diseases may be associated with a moderately dilated
aperistaltic esophagus. Patients with scleroderma and an associated peptic stricture have been
misdiagnosed as having achalasia.29(3025) Although patients with progressive systemic sclerosis
have prolonged clearance as the primary parameter of abnormal reflux, they characteristically also
have low LES pressures.30(3026)
Treatment
Both medical and surgical methods of treatment are effective in achalasia.31,32(3027) Medical
treatment with pneumatic dilation is easy to perform, usually well tolerated by the patient, and rarely
associated with GER. However, this procedure may be complicated by esophageal
perforation.33(3028) Surgical esophageal myotomy generally results in longer and more complete
relief of symptoms,32(3029) but postoperative GER with stricture can be a major problem. Barrett's
esophagus has also been reported in patients with achalasia who have undergone
esophagomyotomy.34,35(3030) With surgical myotomy the mortality is low, but the cost and
preoperative morbidity are high. Parkman et al.36(3031) compared the cost of pneumatic dilation and
esophagomyotomy. On the basis of treatment of 123 patients by dilation, the cost of initial surgical
treatment was five times higher. Even when adjusted for the cost of treatment failures, the cost of
primary management by esophagomyotomy was 2.4 times greater than that for initial pneumatic
dilation. For these reasons, we initially recommend nonoperative therapy and refer patients for
operation only if they remain symptomatic after three attempts at pneumatic dilation.
The technique of esophageal pneumatic dilation described here is performed at our institution with the
use of a Hurst-Tucker pneumatic dilator. Prior to dilation, we inspect the dilator by inflating the bag to
check its symmetry and measure its circumference and by submerging it in water to discover leaks.
The entire procedure is performed under fluoroscopic control. Retained material is removed by
esophageal lavage, and the patient, who has fasted for 12 hr, is then premedicated with topical
oropharyngeal anesthetic spray and intravenous meperidine. A large-diameter bougie is passed under
fluoroscopic guidance to ensure that the patient does not have a stricture.
The pneumatic dilator is then passed into the esophagus, and the radiopaque inflatable bag is
centered lengthwise on the gastroesophageal junction. The bag is gradually inflated, taking care to
maintain the indentation or waist deformity created by the LES at the midpoint of the bag. Insufflation is
continued until the waist deformity is obliterated; pressure is then maintained for 30 to 60 sec (Figure
395). The level of pressure generated within the bag does not appear to be an important factor with
respect to either efficacy or perforation. The most important determinant in the success of the
procedure is full expansion of the bag. Reinflation of the bag, as a method of demonstrating that a
lower pressure is required for full expansion, has not been shown to be a good predictor of
success.37(3032)
(3033)Figure 395. Radiograph of a Hurst-Tucker dilator in the distal esophagus of a patient

with achalasia. Note that the indentation is positioned at the center of the pneumatic dilator.
(Courtesy of Dr. R. K. H. Wong.)
With the patient in the semiupright position after dilation, an esophagogram can be obtained with 1 to 2
ounces of Gastrografin given through the dilator or by mouth after its removal. If a small, walled-off
perforation is noted extending just beyond the lumen of the esophagus, the patient is not permitted any
oral intake and is closely observed. If Gastrografin flows freely into the mediastinum or left chest,
immediate surgery is indicated. The incidence of complications associated with pneumatic dilation
ranges from 0 to 6%, with intramuscular hematomas and small, walled-off perforations being the most
common.38,39(3034) These complications can usually be managed without surgical intervention using
conservative therapy such as intravenous antibiotics and nasogastric tube suction. When management
does not include immediate surgical intervention, the patient's response should be closely observed,
with particular attention to temperature profile and the development of leukocytosis.
Endoscopy occasionally plays a role in the nonoperative treatment of achalasia. If the esophagus is
tortuous and dilated, it may be difficult to position a pneumatic dilator properly at the gastroesophageal
junction. Various endoscopic methods for proper placement have been proposed. This can be
facilitated by passing the dilator over a guidewire previously placed in the stomach under endoscopic
control.40(3035) A long overtube of the type used for enteroscopy was used in one case to facilitate
placement of a pneumatic dilator.41(3036) A pneumatic dilator can be drawn into the stomach by
attaching it to the end of a small-diameter endoscope by means of a string that has been drawn
backward through the accessory channel of the endoscope.42,43(3037) Pneumatic dilation using a
modified Gruntzig dilator under endoscopic control without fluoroscopy has also been
described.44,45(3038) Witzel46(3039) described a type of pneumatic dilator that is mounted on the
insertion tube of a small-diameter forward-viewing endoscope. Since its description in 1981, use of this
device has been reported in only small numbers of patients.4749(3040) Endoscopic electrosurgical
myotomy has also been described.50(3041) Because information concerning safety and efficacy is
extremely limited, this procedure cannot be recommended.
A novel endoscopic treatment method has been proposed by Pasricha et al.51,52(3042) in which
botulinum toxin is injected into the LES. In a double-blind, placebo-controlled trial in 21 patients, toxin
injections resulted in a statistically significant improvement in follow-up symptoms and a decrease in
LES pressure at 1 week after treatment, although not to levels achieved by surgical myotomy or
pneumatic dilation.52(3043) At 6 months, 14 patients remained in remission. No adverse effects of
treatment or complications were observed. The mechanism of action of botulinum toxin in achalasia is
thought to be inhibition of the release of acetylcholine from nerve endings. Although this form of
therapy appears to be useful, currently available data must be considered preliminary in nature. In
particular, the long-term effectiveness and safety of this method of treatment have not been
established, and the effects of repeated injections are unknown. Whether botulinum toxin injections
can be considered an alternative to established methods of treatment in all or some patients depends
on the results of ongoing comparative trials. The successful use of botulinum toxin injection to treat
suspected secondary achalasia in a 74-year-old man with adenocarcinoma of the lung has been
reported.53(3044)
Cancer Risk
Patients with achalasia possibly have an increased risk for squamous cell carcinoma of the
esophagus.5457(3045) Characteristically, carcinoma occurs at the midesophageal level. Most are
squamous cell cancers, although one case of a verrucous carcinoma has been reported.58(3046)
Adenocarcinoma in Barrett's epithelium arising after surgical myotomy has also been
reported.59,60(3047)
The incidence and hence the importance of esophageal cancer in association with achalasia are
controversial.61(3048) In one review, estimates of the risk of cancer range from 0 to 33 times that of
the normal population.62(3049) In a series of 241 patients treated over 25 years by Streitz et
al.,62(3050) esophageal carcinoma developed in 9 patients, for a prevalence of 3.7%, or 1 cancer per
1138 patient-years of follow-up. Based on these data, these investigators estimated the risk for cancer
at 14.5 times that of the general population (age and sex adjusted). Periodic endoscopic surveillance
has been suggested as a method to diagnose such tumors at potentially curable stages. However, our
experience seems to indicate that the incidence of squamous cell carcinoma is quite low. Furthermore,
no convincing data indicate that endoscopic surveillance results in improved survival. Therefore, at this
time we do not recommend routine surveillance for cancer in patients with achalasia.
Diffuse Esophageal Spasm

Diffuse esophageal spasm (DES) is a rare esophageal motility disorder characterized clinically by
chest pain and dysphagia. Specific manometric abnormalities include repetitive simultaneous
esophageal contractions intermixed with normal peristalsis.6365(3051) These contractions may have
increased duration or amplitude as well as incomplete LES relaxation in response to
swallowing.66(3052)
Endoscopy is indicated in DES only to exclude other lesions that mimic this disorder; no characteristic
endoscopic abnormalities are seen. Because DES is infrequently associated with GER, endoscopic
evidence of esophagitis suggests mucosal injury (e.g., acid reflux, pill-induced, infectious, superficial
cancer), with esophageal spasm as a secondary phenomenon. Occasionally, air insufflation during
endoscopy can induce nonperistaltic esophageal contractions and initiate chest pain.
Various types of drugs, including nitrates, anticholinergics, sedatives, calcium-channel blockers, and
antacids, have been used to treat DES with varying degrees of success. Patients sometimes obtain
relief of symptoms after peroral bougienage.67(3053) Individuals with well-documented manometric
abnormalities characteristic of DES and with severe debilitating symptoms, including chest pain and
weight loss, have been treated with tricyclic antidepressants,68(3054) pneumatic dilation,69(3055) or
long surgical myotomy.70(3056)
Other Motility Disorders

Certain systemic disorders such as scleroderma and other collagen vascular diseases may have
associated abnormalities of esophageal peristalsis and LES function.30,66,71(3057) Although
symptoms of dysphagia and chest pain can be related in part to primary esophageal motor failure,
most symptoms in these patients result from esophagitis and strictures related to the GER that
complicates these conditions.
Endoscopic findings are generally related to underlying GER. For instance, severe erosive esophagitis
with hemorrhage and stricture is commonly encountered in patients with scleroderma. Barrett's
esophagus has also been described.72(3058) Occasionally, there may be endoscopic evidence of
Candida esophagitis with its characteristic "cheesy" white, plaquelike membrane in a seriously ill
patient with systemic symptoms. Severe circumoral involvement may also make it impossible to place
a standard bite guard in the patient's mouth or to pass a dilator; use of a small-diameter endoscope
and hydrostatic dilators may be required in such cases.
Management includes vigorous antireflux therapy, such as elevation of the head of the bed to promote
esophageal acid clearance and use of H2 receptor-blocking drugs or proton pump inhibitors to
decrease acid secretion. Benign esophageal strictures respond to peroral bougienage, whereas
antifungal therapy should effectively eliminate associated esophageal candidiasis. Although
fundoplication can significantly decrease reflux events in patients with erosive esophagitis due to
scleroderma, even one such event that is poorly cleared from the esophagus owing to aperistalsis and
fundoplication can still lead to prolonged acid mucosal contact73(3059) and its adverse consequences.
For this reason, as well as the effectiveness of medical therapy74(3060) and poor long-term results of
surgery,75(3061) antireflux surgery is not widely employed in patients with reflux due to scleroderma.
Esophageal Diverticulum
Esophageal diverticula are included in this section because this type of abnormality is generally
associated with an underlying motility disorder. Traditionally, esophageal diverticula are divided into
upper or pharyngeal esophageal, midesophageal (formerly referred to as traction diverticulum), and
lower esophageal or epiphrenic.
Pharyngeal Esophageal Diverticulum (Zenker's Diverticulum)
A pharyngeal esophageal diverticulum occurs when the hypopharyngeal mucosa forms a posteriorly
directed pouch by protruding between the inferior pharyngeal constrictor muscle and the
cricopharyngeal muscle in an area of junctional muscle weakness (Killian's dehiscence).76(3062) The
pathogenesis is thought to be disordered coordination between pharyngeal contraction and upper
esophageal sphincter relaxation. In one study, manometry demonstrated double pressure peaks in 8 of
10 patients with symptomatic hypopharyngeal diverticulum, indicating a split between the muscle
bundles in the sphincter.77(3063)
Symptoms from Zenker's diverticulum are usually related to the size of the pouch. A small diverticulum
is often asymptomatic or associated only with the sensation of a foreign body in the back of the throat.
As the diverticulum enlarges, it can cause unexplained regurgitation of undigested food or nocturnal
aspiration. An extremely large diverticular sac can compress and deform the esophageal lumen,
resulting in severe dysphagia and weight loss. Swallowed foreign objects may become lodged in
Zenker's diverticulum. These may be especially difficult to remove and can lead to perforation.78(3064)
Although the majority of patients with Zenker's diverticulum are more than 50 years of age, it has also
been described in younger patients, including children.79(3065)
Pharyngeal esophageal diverticulum is usually diagnosed by barium swallow roentgenography (Figure
396). Endoscopy is not required, and, because the diverticulum is located in the cricopharyngeal area,
it is usually not visualized during endoscopy. If endoscopy is undertaken in a patient with Zenker's
diverticulum to exclude another cause of symptoms or to evaluate another lesion, the procedure
should be performed cautiously. During blind intubation, the endoscope can inadvertently enter and
perforate the diverticulum, as the thin wall consists entirely of mucosa. Endoscopic intubation under
direct vision or over a previously placed guidewire can reduce this risk. Generally, one tends to enter
the diverticular sac preferentially because it lies proximal to the cricopharyngeus muscle and is easily
inflated during the procedure. The esophageal lumen is always found at or very close to the opening
into the diverticulum, and therefore the tip of the endoscope need not be passed into the diverticulum
(Figure 397).
(3066)Figure 396. Lateral (A) and anteroposterior (B) radiographs of Zenker's diverticulum.
(3067)Figure 397. Zenker's diverticulum (left side of field) containing food material. (From
Midesophageal Diverticulum
Esophageal diverticula may be single or multiple in the midesophagus (Figure 398). It was formerly
thought that a midesophageal diverticulum resulted from traction on the esophagus produced by
inflammatory conditions (e.g., tuberculosis, fungal infection) in the mediastinum. Recent evidence,
however, suggests that mid-esophageal diverticula result from esophageal motility disorders.80(3068)
Achalasia, DES, and scleroderma have been associated with these diverticula.81(3069) A diverticulum
of the esophageal body rarely produces symptoms, but occasionally it may bleed, perforate, or enlarge
to such a degree that it compresses the esophageal lumen.
(3070)Figure 398. Midesophageal diverticulum. (From the collection of Dr. M. V. Sivak,

Jr.)
An esophageal diverticulum is usually diagnosed on barium esophagogram. Endoscopy is not required,
but, in patients undergoing endoscopy for other reasons, the opening of the diverticulum is often noted.
If the diverticulum is extremely large and has a wide mouth, it may be difficult to determine which
opening represents the esophageal lumen and which the diverticular lumen (Figure 399). An
esophageal diverticulum does not preclude carefully performed endoscopy. However, during blind
intubation of the esophagus, such as occurs with peroral bougienage or passage of a nasogastric tube,
the diverticulum may be perforated. This risk can be reduced by performing dilation over an
endoscopically placed guidewire or under fluoroscopic control.
(3071)Figure 399. Large, wide-mouthed esophageal diverticulum (at right of endoscopic

field).
Because esophageal diverticula are usually asymptomatic, no therapy is required. However, in the rare
case of a diverticulum that bleeds, perforates, or causes esophageal obstruction, surgical treatment is
indicated. There are a few case reports of carcinoma arising in an esophageal diverticulum.82(3072)
Epiphrenic Diverticulum
Epiphrenic diverticula are usually solitary and located in the distal 3 cm of the esophagus (Figure
3910). As with the other types of esophageal diverticula, the pathogenesis appears to be abnormal
motility with a failure of the LES to relax fully, resulting in an increase in intraluminal pressure. Debas et
al.83(3073) found that 50 of 65 patients with an epiphrenic diverticulum had evidence of abnormal
motility, most often diffuse spasm or achalasia. Of the 15 patients with normal esophageal motility, 13
had a hiatus hernia and 5 of these had a high-grade distal esophageal stricture. In patients with an
epiphrenic diverticulum and achalasia, pneumatic dilation should be avoided because of an increased
risk of perforation. Combined extramucosal myotomy and diverticulectomy are the preferred
management.84(3074)
(3075)Figure 3910. Epiphrenic diverticulum, hiatus hernia, and severe esophagitis. (From
A large leiomyoma arising in the distal esophagus can occasionally present radiographically as a large
epiphrenic diverticulum (Figure 3911). This has been associated with massive bleeding.85(3076) The
tumor involves the lower third of the esophagus and grows outward in all directions, carrying normal
mucosal lining. The endoscopic finding is a pouchlike dilation of the lumen, an appearance that mimics
an epiphrenic diverticulum.
(3077)Figure 3911. A, Barium esophagogram of a leiomyoma mimicking an epiphrenic

diverticulum. B, View of exophytic leiomyoma (center of the surgical field) at thoracotomy.
Miscellaneous Esophageal Disorders

Mucosal Esophageal Rings
Distal esophageal ring (Schatzki ring, B ring) is a common finding on barium swallow roentgenograms
and at endoscopy (see also Chapter 44: Hiatus Hernia and Peptic Diseases of the
Esophagus).86(3078) These rings consist entirely of mucosa and occur at or near the
squamocolumnar mucosal junction. The pathogenesis of lower esophageal rings is unclear. Congenital
factors and chronic damage to the esophageal mucosa as a result of GER87(3079) or pill-induced
esophagitis88(3080) have been implicated (Figure 3912). Although most B rings are asymptomatic,
they may be associated with intermittent dysphagia or food impaction when they narrow the lumen to
less than 12 mm in diameter.89(3081) A thin, symmetric, circumferential membranous projection that
partially occludes the distal esophageal lumen at the squamocolumnar junction is demonstrated at
endoscopy. The ring has a fixed diameter, as demonstrated by failure to obliterate the membrane with
continuous air insufflation (Figure 3913).
(3082)Figure 3912. Pills above a Schatzki ring.
(3083)Figure 3913. A, Schatzki's ring without adequate insufflation. B, Schatzki's ring with
adequate insufflation. C, Retroflex view of Schatzki's ring without adequate insufflation. D,
Retroflex view of Schatzki's ring with adequate insufflation.
If the ring diameter is less than the critical 12 mm, slight resistance to passage of a standard diameter
endoscope may be perceived as it enters the stomach. On occasion the ring is ruptured during
intubation of the gastroesophageal junction, resulting in a mucosal tear and slight bleeding. This may
not be appreciated until the endoscope is withdrawn through the traumatized area. Small-caliber
endoscopes can pass through a Schatzki ring without difficulty; therefore, the lesion may not be
appreciated at endoscopy or its significance may be misjudged. If undue resistance is encountered
during intubation of the gastroesophageal junction or esophagitis is present, a peptic stricture rather
than a Schatzki ring should be suspected.
Conventional endoscopes are not designed to dilate esophageal lesions. Forceful attempts at
intubation of a fixed stenosis may result in instrument damage or, more importantly, injury to the
esophagus itself. Tight rings as well as benign and malignant strictures should be treated by
established techniques for dilation.
Peroral bougienage is effective therapy for most rings.86(3084) A single passage of a large-caliber
mercury-filled bougie ruptures the ring and usually alleviates symptoms permanently. Occasionally,
recurrent symptoms may necessitate repeat bougienage. Schatzki rings resistant to dilation have been
incised endoscopically using either laser90(3085) or monopolar electrocautery.91(3086) For unusually
intractable rings, surgery is also effective.92(3087)
Muscular Esophageal Rings

Muscular esophageal rings most commonly occur several centimeters proximal to the
squamocolumnar junction.93(3088) When observed during radiologic examination or at endoscopy,
this type of ring changes appearance as the esophageal musculature contracts and relaxes. This
dynamic feature, the characteristic location, and the lack of resistance to passage of the endoscope
help to differentiate a muscular ring from a mucosal ring or a peptic stricture. Muscular rings rarely
cause symptoms, but if dysphagia is present, peroral bougienage may be helpful.
Esophageal Webs
Esophageal webs are membrane-like structures, completely covered by squamous epithelium, that
may focally narrow the esophageal lumen. They may be solitary or multiple (Figure 3914).94,95(3089)
Webs usually occur in the cervical esophagus in the immediate postcricoid area but can be found at
any level of the esophagus.96(3090) They are usually diagnosed first by barium swallow radiographs
(Figure 3915). Symptoms are variable, although most often an esophageal web is asymptomatic and
represents only an incidental radiographic finding. However, some patients complain of severe
dysphagia, localized most often in the cervical area.97(3091) An association between webs and iron
deficiency anemia has been described (Plummer-Vinson or Paterson-Kelly syndrome), but webs
usually occur in the absence of any associated conditions.
(3092)Figure 3914. Multiple midesophageal webs.
(3093)Figure 3915. Lateral (A) and anteroposterior (B) barium swallow x-rays of an
esophageal web.
The presence of a cervical esophageal web is frequently overlooked at endoscopy because it is
inadvertently ruptured during blind intubation of the hypopharynx and cervical esophagus. Careful
endoscopic intubation under direct vision occasionally permits the demonstration of a thin, often
transparent, postcricoid membrane (Figure 3916). More often, the torn membrane is visualized during
withdrawal of the endoscope. Endoscopic evaluation should include thorough examination of the
cervical esophagus and hypopharynx because carcinoma in this region is more common in patients
with webs and associated iron deficiency anemia.98(3094) Because the web is usually ruptured by the
endoscope, further treatment is rarely required. In patients with persistent symptoms after endoscopy
or in those with multiple webs, peroral bougienage is effective therapy.99(3095) If this fails, endoscopic
lysis using a laser may be successful.90(3096)
(3097)Figure 3916. Proximal esophageal web. Note the traumatic area caused by an
instrument.
Extrinsic Compression of the Esophagus

Abnormal structures, aberrant vessels, and neoplasms adjacent to the esophagus may produce
symptoms due to compression or deviation of the esophageal lumen.100,101(3098) Examples include
large osteophytes of the cervical spine, cardiac chamber enlargement, tortuous or aneurysmal
deformity of the thoracic aorta (dysphagia aortica), aberrant right subclavian artery (dysphagia lusoria),
lung cancer, and enlarged mediastinal lymph nodes. The exact location of extrinsic esophageal
narrowing is best identified by barium swallow roentgenograms. Endoscopy generally demonstrates
deviation or tortuosity of the esophageal lumen or the smooth, rounded deformity of extrinsic
compression without disruption of the mucosa. On occasion, mucosal abnormalities indicative of
esophageal invasion by an extrinsic lesion may be noted. Extrinsic compression of the esophagus
does not usually cause symptoms except in malignancy or chronic infection, such as mediastinal
collagenosis from histoplasmosis. Endoscopy is contraindicated if a thoracic aortic aneurysm is
suspected. Endoscopy with a rigid instrument is hazardous in patients with posterior compression of
the cervical esophagus by osteophytes because of possible esophageal damage during intubation.
Peroral bougienage to treat dysphagia due to an extrinsic process is generally not successful because
the esophageal wall quickly collapses when the dilator is removed.
Esophageal Intramural Pseudodiverticula

Esophageal intramural pseudodiverticula (EPD) comprise a rare entity of flask-shaped openings limited
to the mucosa and submucosa; it is believed to be the result of chronic inflammation of the
submucosal glands.102(3099) The typical appearance on barium contrast radiography consists of
collar button-shaped outpouchings projecting in a perpendicular fashion from the luminal surface.
Although this condition is best demonstrated by radiography, the tiny openings, occasionally containing
inspissated whitish material, can be readily visualized at endoscopy.103(3100) EPD is most common in
patients over the age of 70 years, but it may occur at virtually any age, including childhood.104(3101)
Most often EPD is limited to a narrow segment around a stricture. GER disease is clearly an etiologic
factor with distal strictures.105,106(3102) Candidiasis,102(3103) caustic ingestion,106(3104) and
diabetes105(3105) have all been associated with EPD. Symptomatic treatment of the dysphagia can
be accomplished by dilation of the stricture, but the underlying pseudodiverticula do not resolve.
Esophageal Hematomas
Esophageal hematomas result from submucosal bleeding that can occur spontaneously in patients
with impaired hemostasis or after esophageal trauma such as that occurring, for example, with
vigorous retching or an aborted sneeze.107109(3106) Most hematomas due to trauma involve the
distal esophagus but may extend in an orad direction within the esophageal wall to involve the entire
organ. In patients with clotting abnormalities, hematomas may be multicentric. Symptoms include
dysphagia, odynophagia, and hematemesis.109(3107) The presence of an esophageal hematoma
may be suspected on barium swallow radiography. It appears as a submucosal mass indenting the
barium column, often with occlusion of the esophageal lumen. In fact, it may simulate esophageal
malignancy. However, radiographs several days later show rapid resolution of the process.
At endoscopy a hematoma appears as a reddish purple submucosal swelling that partially occludes the
esophageal lumen.102,110(3108) A multicentric lesion can have the appearance of multiple, discrete
submucosal ecchymoses. Mucosa overlying the hematoma is usually smooth, but occasionally it is
disrupted and irregular and simulates malignancy (Figure 3917). Patients with esophageal hematoma
rarely require transfusion or surgical therapy because most lesions resolve spontaneously over several
days.
(3109)Figure 3917. Esophageal hematoma with mucosal irregularity simulating a carcinoma.
Spontaneous Intramural Esophageal Rupture

Long linear lacerations of the esophageal mucosa, which appear to occur spontaneously, have been
described.111,112(3110) In most cases, the onset is sudden, with chest pain followed by
hematemesis, hematoma formation, and dysphagia. Frank perforation does not occur, and the lesions
usually resolve with conservative management.
Glycogenic Acanthosis
Glycogenic acanthosis is a benign mucosal lesion of the esophagus commonly seen at
endoscopy.113,114(3111) This lesion is easily recognized as many small (<1 cm diameter), pale or
whitish papules, usually in the distal esophagus (Figure 3918). Histologically, the lesion is
characterized by hyperplastic squamous acanthotic epithelium with increased glycogen. Its clinical
significance is unknown, although it has been suggested to be an indicator of GER. Because the entity
is common (3 to 15% of consecutive endoscopies),115,116(3112) it must be differentiated from other

more significant lesions that can have a similar appearance, such as moniliasis, leukoplakia, and early
esophageal carcinoma.
(3113)Figure 3918. Glycogenic acanthosis of the esophagus.
Heterotopic Gastric Mucosa (Inlet Patch)

Small, isolated islands of gastric mucosa, usually less than 1 cm in maximum diameter, may be found
within the esophageal squamous mucosa. These patches of reddish orange mucosa are easily
distinguished from the surrounding pinkish color of the normal squamous mucosa; staining the
squamous mucosa with Lugol's iodine solution sharply enhances this demarcation. They are usually
round or elliptical in shape but can have an elongated, a linear, or rarely a circumferential
configuration.117(3114) In about half of cases, multiple patches are present, usually located in close
proximity.118(3115) In consecutive endoscopic procedures, heterotopic gastric mucosa has been seen
in 2.8 to 10% of cases.117120(3116) Although it can be found at any level in the esophagus, most
patches are relatively close to the cricopharyngeal sphincter, hence the term inlet patch. Either fundic
or antral-type mucosa is found in biopsy specimens; parietal cells may also be present, indicating a
potential for acid secretion. In most cases, heterotopic gastric mucosa is an incidental finding with no
clinical relevance. However, Nakajima et al.119(3117) described 10 patients with heterotopic gastric
mucosa and complaints of pharyngeal discomfort, a globus sensation, or both that responded to
treatment with H2-receptor antagonist drugs. Isolated reports can be found of complications, including
adenocarcinoma, stricture formation, and tracheoesophageal fistula.121124(3118)
REFERENCES
1.
Cohen S, Lipshutz W. Lower esophageal sphincter dysfunction in achalasia. Gastroenterology

1971;61:81420.
2. Castell DO. Achalasia and diffuse esophageal spasm. Arch Intern Med 1976;136:5719.
3.
Barrett NR. Achalasia of the cardia: Reflections upon a clinical study of over 100 cases. Br
Med J 1964;1:113540.
4.
Wong RKH, Maydonovitch CL. Achalasia. In Castell DO (ed). The Esophagus. Boston: Little,
Brown and Co., 1992;233.
5. Smart HL, Foster PN, Evans DF, et al. Twenty-four hour esophageal acidity in achalasia
before and after pneumatic dilation. Gut 1987;28:8837.
6. Ellis FH, Crozier RE, Watkins E. Reoperative achalasia surgery. J Thorac Cardiovasc Surg
1986;92:85963.
7. Tucker HJ, Snape WJ Jr, Cohen S. Achalasia secondary to carcinoma: Manometric and
clinical features. Ann Intern Med 1978;89:3158.
8. Sandler RS, Bozymski EM, Orlando RC. Failure of clinical criteria to distinguish between
primary achalasia and achalasia secondary to tumor. Dig Dis Sci 1982;27:20913.
9. Deviere J, Dunham F, Rickaert F, et al. Endoscopic ultrasonography in achalasia.
10. Russell CO, Hill LD, Homes ER III, et al. Radionuclide transit: A sensitive screening test for
esophageal dysfunction. Gastroenterology 1981;80:88792.
11. Gross R, Johnson LF, Kaminski RJ. Esophageal emptying in achalasia quantitated by a
radioisotope technique. Dig Dis Sci 1979;24:9459.
12.
Vantrappen G, Hellemans J. Diseases of the Esophagus. New York: Springer-Verlag,
1974;341.
13. Laub W, Achkar E. Hiatal hernia in patients with achalasia. Am J Gastroenterol
1987;82:125660.
14. Davis JA, Kantrowitz PA, Chandler HL, Schatzki SC. Reversible achalasia due to
reticulum-cell sarcoma. N Engl J Med 1975;293:1302.
15. Benjamin SB, Castell DO. Achalasia and Hodgkin's disease: A chance association?. J Clin
16. Roark G, Shabot M, Patterson M. Achalasia secondary to hepatocellular carcinoma. J Clin
17. Goldschmiedt M, Peterson WL, Spielberger R, et al. Esophageal achalasia secondary to
mesothelioma. Dig Dis Sci 1989;34:128589.
18. Marshall JB, Diaz-Arias AA, Bochna GS, et al. Achalasia due to diffuse leiomyomatosis and
inherited as an autosomal dominant disorder: Report of a family study. Gastroenterology
1990;98:135865.
19. Idenburg FJ, Akkermans LM, Smout AJ, et al. Leiomyoma of the distal oesophagus mimicking
achalasia. Neth J Surg 1991;43:7981.
20. Makker HK, Chisholm R, Rate AJ, et al. Dysphagia due to secondary achalasia as an early
manifestation of squamous cell carcinoma. Postgrad Med J 1995;71:5024.
21. Dufresue CR, Jeyasingher K, Baker RR. Achalasia of the cardia associated with pulmonary
sarcoidosis. Surgery 1983;94:326.
22. Costigan DJ, Clouse RE. Achalasia like esophagus from amyloidosis: Successful treatment
with pneumatic balloon dilation. Dig Dis Sci 1983;28:7637.
23. Landres RT, Kuster GGR, Strum WB. Eosinophilic esophagitis in a patient with vigorous
achalasia. Gastroenterology 1978;74:12981301.
24. Woods CA, Foutch PG, Waring JP, et al. Pancreatic pseudocyst as a cause for secondary
25. Kahrilas PJ, Kishk SM, Helm JF, et al. Comparison of pseudoachalasia and achalasia. Am J
Med 1987;82:43946.
26. Rozman RW Jr, Achkar E. Features distinguishing secondary achalasia from primary
achalasia. Am J Gastroenterol 1990;85:132730.
27.
Bettarello A, Pinotti HW. Oesophageal involvement in Chagas' disease. Clin Gastroenterol
1976;5:1039.
28.
Koberle F. Enteromegaly and cardiomegaly in Chagas' disease. Gut 1963;4:399405.
29. Park RH, McKillop JH, Belch JJ, et al. Achalasia like syndrome in systemic sclerosis. Br J
Surg 1990;77:469.
30. Murphy JR, McNally P, Peller P, et al. Prolonged clearance is the predominant reflux
abnormality in patients with progressive systemic sclerosis with esophagitis. Dig Dis Sci
1992;37:83341.
31.
Cassella RR, Brown AL Jr, Sayre GP, et al. Achalasia of the esophagus: Pathologic and
etiologic considerations. Ann Surg 1964;160:47487.
32. Csendes A, Braghetto I, Henriquez A, Cortes C. Late results of a prospective randomised
study comparing forceful dilatation and oesophagomyotomy in patients with achalasia. Gut
1989;30:299304.
33. Nair LA, Reynolds JC, Parkman HP, et al. Complications during pneumatic dilation for
achalasia or diffuse esophageal spasm. Analysis of risk factors, early clinical characteristics,
and outcome. Dig Dis Sci 1993;38:1893904.
34. Sprung DJ, Gibb SP. Barrett's esophagus in a patient with achalasia. Am J Gastroenterol
1985;80:3303.
35. Kortan P, Warren RE, Gardner J, et al. Barrett's esophagus in a patient with surgically treated
achalasia. J Clin Gastroenterol 1981;3:35760.
36. Parkman HP, Reynolds JC, Ouyang A, et al. Pneumatic dilatation or esophagomyotomy
treatment for idiopathic achalasia: Clinical outcomes and cost analysis. Dig Dis Sci
1993;38:7585.
37.
Wong RKH, Maydonovitch CL. Parameters measured during and immediately post dilation do
not predict long term success in achalasia: A six year prospective study [Abstract]. Am J
38. Sanderson DR, Ellis FH Jr, Olsen AM. Achalasia of the esophagus: Results of therapy by
dilation, 19501967. Chest 1970;58:11621.
39. Vantrappen G, Hellemans J, Deloof W, et al. Treatment of achalasia with pneumatic dilations.
Gut 1971;12:26875.
40. Holloway RH, McCallum RW. Technique for pneumatic dilatation in achalasia complicated by
"sigmoid" esophagus. J Clin Gastroenterol 1982;4:1235.
41. Bernstein D, Barkin JS. Pneumatic dilation of a sigmoid esophagus in achalasia using an
overtube. Gastrointest Endosc 1993;39:54950.
42. Kerr RM, Ott DJ, Wu WC, Ward BW. Pneumatic dilatation of the achalasic esophagus
requiring the aid of an endoscope. Am J Gastroenterol 1987;82:747.
43. Neumann G, Samani U, Nas S. A new endoscopic technique for introducing pneumatic
dilators in patients with achalasia. Gastrointest Endosc 1992;38:598601.
44. Levine ML, Dorf BS, Moskowitz G, Bank S. Pneumatic dilatation in achalasia under
endoscopic guidance: Correlation pre- and postdilatation by radionuclide scintiscan. Am J
45. Levine ML, Moskowitz GW, Dorf BS, Bank S. Pneumatic dilation in patients with achalasia
with a modified Gruntzig dilator (Levine) under direct endoscopic control: Results after 5
years. Am J Gastroenterol 1991;86:15814.
46. Witzel L. Treatment of achalasia with a pneumatic dilator attached to a gastroscope.
Endoscopy 1981;13:1767.
47. Johnston BT, Collins BJ, Collins JS, Ferguson WR. Perendoscopic pneumatic dilatation in
achalasia: Assessment of outcome using esophageal scintigraphy. Dysphagia 1992;7:2014.
48. Frimberger E, Kuhner W, Kunert H, Ottenjann R. Results of treatment with the endoscope
dilator in 11 patients with achalasia of the esophagus. Endoscopy 1981;13:1735.
49. Di Matteo S, Cifani F, Ferrini G, et al. Perendoscopic pneumatic dilatation in achalasia: Short
and mid-term results. Ital J Surg Sci 1988;18:3337.
50. Ortega JA, Madureri V, Perez L. Endoscopic myotomy in the treatment of achalasia.
51. Pasricha PJ, Ravich WJ, Kalloo AN. Effects of intrasphincteric botulinum toxin on the lower
esophageal sphincter in piglets. Gastroenterology 1993;105:10459.
52. Pasricha PJ, Ravich WJ, Hendrix TR, et al. Intrasphincteric botulinum toxin for the treatment
of achalasia. N Engl J Med 1995;332:7748.
53. Vallera RA, Brazer SR. Botulinum toxin for suspected pseudoachalasia. Am J Gastroenterol
1995;90:131921.
54. Lortat-Jacob JL, Richard CA, Fekete F, et al. Cardiospasm and esophageal carcinoma:
Report of 24 cases. Surgery 1969;66:96975.
55. Wychulis AR, Woolam GL, Anderson HA, et al. Achalasia and carcinoma of the esophagus.
JAMA 1971;215:163841.
56.
Williams JL. Carcinoma of the oesophagus as a complication of achalasia of the cardia.
Thorax 1956;11:26874.
57. Just-Viera JO, Haight C. Achalasia and carcinoma of the esophagus. Surg Gynecol Obstet
1969;128:108195.
58. Roach E, Barr G. Verrucous carcinoma of the oesophagus and achalasia. J Gastroenterol
Hepatol 1993;8:1079.
59. Gallez JF, Berger F, Moulinier B, Partensky C. Esophageal adenocarcinoma following Heller
myotomy for achalasia. Endoscopy 1987;19:768.
60. Shah AN, Gunby TC. Adenocarcinoma and Barrett's esophagus following surgically treated
achalasia. Gastrointest Endosc 1984;30:2946.
61. Chuong JJH, DuBovik S, McCallum RW. Achalasia as a risk factor for esophageal carcinoma.
A reappraisal. Dig Dis Sci 1984;29:11058.
62. Streitz JM Jr, Ellis FH Jr, Gibb SP, Heatley GM. Achalasia and squamous cell carcinoma of
the esophagus: Analysis of 241 patients. Ann Thorac Surg 1995;59:16049.
63.
DiMarino AJ Jr, Cohen S. Characteristics of lower esophageal sphincter function in
symptomatic diffuse esophageal spasm. Gastroenterology 1974;666:16.

64. Mellow M. Symptomatic diffuse esophageal spasm: Manometric follow-up and response to
cholinergic stimulation and cholinesterase inhibition. Gastroenterology 1977;73:23740.
65. Richter JE, Castell DO. Diffuse esophageal spasm: A reappraisal. Ann Intern Med
1984;100:2427.
66. Cohen S. Motor disorders of the esophagus. N Engl J Med 1979;301:18492.
67.
Goldin NR, Burns TW, Herrington JP. Treatment of nonspecific esophageal motor disorders.
Beneficial effects of bougienage [Abstract]. Gastroenterology 1982;82:1069.
68. Clouse RE, Lustman PJ, Eckert TC, et al. Low dose trazadone for symptomatic patients with
esophageal contraction abnormalities. Gastroenterology 1987;92:102736.
69.
Vantrappen G, Hellemans J. Treatment of achalasia and related motor disorders.
70. Leonardi HK, Shea JA, Crozier RE, et al. Diffuse spasm of the esophagus. Clinical,
manometric, and surgical considerations. J Thorac Cardiovasc Surg 1977;74:73643.
71.
Cohen S, Laufer I, Snape WJ Jr, et al. The gastrointestinal manifestations of scleroderma:
Pathogenesis and management. Gastroenterology 198O;79:15566.
72. Sprung DJ, Gibb SP. Dysplastic Barrett's esophagus in scleroderma. Am J Gastroenterol
1985;80:51822.
73. Walters PF, DeMeester T. Foregut motor disorders and their surgical management. Med Clin
North Am 1981;65:1195222.
74. Petrokubi RJ, Jeffries GH. Cimetidine vs. antacid in scleroderma with reflux esophagitis.
75.
Mansour KA, Malone CE. Surgery for scleroderma of the esophagus: A 12 year experience.
Ann Thorac Surg 1988;5:5134.
76. Knuff TE, Benjamin SB, Castell DO. Pharyngoesophageal (Zenker's) diverticulum: A
reappraisal. Gastroenterology 1982;82:7346.
77.
Nilsson ME, Isberg A, Schiratzki H. The hypopharyngeal diverticulum: A simultaneous
cineradiographic and manometric examination. Acta Otolaryngol 1988;106:31420.
78. White IL. Severe complication of a Zenker's diverticulum with endoscopic diverticulotomy
rescue. Laryngoscope 1981;91:70819.
79. Fiser WP, Euler AR, Golladay ES. Pharyngoesophageal pulsion diverticulum (Zenker) in a
ten-year-old boy. J Pediatr Surg 1981;16:7279.
80. Kaye MD. Oesophageal motor dysfunction in patients with diverticula of the mid-thoracic
oesophagus. Thorax 1974;29:66672.
81. Penagini R, Bianch P. Primary achalasia associated with a midthoracic esophageal
diverticulum. J Clin Gastroenterol 1990;12:36567.
82. Fujita H, Kakegawa T, Shima S, Kumagaya Y. Carcinoma within a middle esophageal
(parabronchial) diverticulum: A case report and review of the literature. Jpn J Surg
1980;10:1428.
83. Debas HT, Payne WS, Cameron AJ, Carlson HC. Physiopathology of lower esophageal
diverticulum and its implications for treatment. Surg Gynecol Obstet 1980;151:593600.
84. Mulder D, Rosenkranz E, DenBesten L. Management of huge epiphrenic esophageal
diverticulum. Am J Surg 1989;157:3037.
85. Bozorgi S, Migliorelli FA, Cook WA. Leiomyoma of the esophagus presenting as a bleeding
epiphrenic diverticulum. Chest 1973;63:2814.
86. Goyal RK, Glancy JJ, Spiro HM. Lower esophageal ring. N Engl J Med
1970;282:1298305,135562.
87. Marshall JB, Kretschmar JM, Diaz-Arias AA. Gastroesophageal reflux as a pathogenic factor
in the development of symptomatic lower esophageal rings. Arch Intern Med
1990;150:166972.
88. Jamieson J, Hinder R, DeMeester T, et al. Analysis of 32 patients with Schatzki's ring. Am J
Surg 1989;158:5636.
89.
Schatzki R. The lower esophageal ring: Long term follow-up of symptomatic and
asymptomatic rings. Am J Roentgenol 1963;90:80510.

90. Krevsky B, Pusateu J. Laser lysis of an esophageal web. Gastrointest Endosc 1989;35:4513.
91. Raskin JB, Monteau T, Harary A, et al. Transendoscopic electrosurgical incision of lower
esophageal (Schatzki rings): A new treatment modality. Gastrointest Endosc 1985;31:3914.
92. Molins L, Payne S, Cameron A, et al. An unusually intractable Schatzki's ring. Ann Thorac
Surg 1988;45:3279.
93. Goyal RK, Bauer JL, Spiro HM. The nature and location of lower esophageal ring. N Engl J
Med 1971;284:117580.
94. Shiflett DW, Gilliam JH, We WC, et al. Multiple esophageal webs in adults. Gastroenterology
1979;77:5569.
95. Carlisle WR. A case of multiple esophageal webs and rings. Gastrointest Endosc
1984;30:1845.
96. Tedesco FJ, Morton WJ. Lower-esophageal webs. Am J Dig Dis 1975;20:3813.
97. Ekberg O. Cervical oesophageal webs in patients with dysphagia. Clin Radiol
1981;32:63341.
98. Chisholm M. The association between webs, iron and post-cricoid carcinoma. Postgrad Med J
1974;50:2159.
99. Gilat T, Rozen P. Fiberoptic endoscopic diagnosis and treatment of a congenital esophageal
diaphragm. Am J Dig Dis 1975;20:7815.
100.
Johnson LF, Moses FM. Endoscopic evaluation of esophageal disease. In Castell DO,
Johnson LF (eds). Esophageal Function in Health and Disease. New York: Elsevier,
1983;237.
101.
Cattau EL Jr, Castell DO. Symptoms of esophageal dysfunction. Adv Intern Med
1982;82:7346.
102. Choa SR, Sanders MM, Turner MA, et al. Esophageal intramural pseudodiverticulosis.
Gastrointest Radiol 1981;6:916.
103. Mahajan SK, Warshauer DM, Bozymski EM. Esophageal intramural pseudo-diverticulosis:
Endoscopic and radiologic correlation. Gastrointest Endosc 1993;39:5657.
104. Fromkes J, Thomas FB, Mekhjian H, et al. Esophageal intramural pseudodiverticulosis. Am J
Dig Dis 1977;22:690700.
105. Sabanathan S, Salama FD, Morgan WE. Oesophageal intramural pseudodiverticulosis.
Thorax 1985;40:84957.
106.
Muhletaler C, Lams P, Johnson A. Occurrence of oesophageal intramural
pseudodiverticulosis in patients with pre-existing benign oesophageal stricture. Radiology
1980;53:299303.
107. Shay SS, Berendson RA, Johnson LF. Esophageal hematoma. Four new cases, a review,
and proposed etiology. Dig Dis Sci 1981;26:101924.
108. Talley NA, Nicks R. Spontaneous submucosal haematoma of the oesophagus: Oesophageal
apoplexy. Med J Aust 1969;2:14650.
109. Spiller RC, Catto JV, Kane SP. Spontaneous dissecting intramural haematoma of the
oesophagus: A rare cause of haematemesis and dysphagia. Endoscopy 1981;13:12830.
110. Tim LO, Segal I, Mirwis J. Intramural haematoma of the oesophagus. The role of endoscopy.
S Afr Med J 1982;61:798800.
111. Steadman C, Kerlin P, Crimmins F, et al. Spontaneous intramural rupture of the oesophagus.
Gut 1990;31:8459.
112. Gutierrez del Olmo A, Loscos JM, Baki W, et al. Spontaneous intramural oesophageal
perforation. Endoscopy 1985;17:767.
113. Bender MD, Allison J, Cuartas F, et al. Glycogenic acanthosis of the esophagus: A form of
benign epithelial hyperplasia. Gastroenterology 1973;65:37380.
114.
Clemencon G, Gloor F. Benign epithelial hyperplasia of the esophagus: Glycogenic
acanthosis. Endoscopy 1974;6:214.
115. Stern Z, Sharon P, Ligumsky M, et al. Glycogenic acanthosis of the esophagus. A benign but
confusing endoscopic lesion. Am J Gastroenterol 1980;74:2613.
116. Vadva MD, Triadafilopoulos G. Glycogenic acanthosis of the esophagus and

gastroesophageal reflux. J Clin Gastroenterol 1993;17:7983.
117. Jabbari M, Goresky CA, Lough J, et al. The inlet patch: Heterotopic gastric mucosa in the
upper esophagus. Gastroenterology 1985;89:3526.
118. Borhan-Manesh F, Farnum JB. Incidence of heterotopic gastric mucosa in the upper
esophagus. Gut 1991;32:96872.
119. Nakajima H, Munakata A, Sasaki Y, Yoshida Y. pH profile of esophagus in patients with inlet
patch of heterotopic gastric mucosa after tetragastrin stimulation. An endoscopic approach.
Dig Dis Sci 1993;38:19159.
120. Feller SC, Weaver GA. Heterotopic gastric mucosa in the upper esophagus. Gastroenterology
1986;90:2578.
121. Kohler B, Kohler G, Riemann JF. Spontaneous esophagotracheal fistula resulting from ulcer
in heterotopic gastric mucosa. Gastroenterology 1988;95:82830.
122. Yarborough CS, McLane RC. Stricture related to an inlet patch of the esophagus. Am J
123. Powell RW, Luck SR. Cervical esophageal obstruction by ectopic gastric mucosa. J Pediatr
Surg 1988;23:6324.
124. Schmidt H, Riddell H, Walter B, et al. Adenokarzinome in heterotoper Magenschleimhaut des
proximalen sophagus [Adenocarcinoma of heterotopic gastric mucosa in the proximal
esophagus]. Leber Magen Darm 1985;15:1447.
Chapter 40 Benign and Malignant Tumors of the Esophagus

(3119)
(3120)
G. N. J. TYTGAT, M.D.
Clinical Features
The presenting symptoms of benign esophageal tumors are dysphagia and bleeding; the latter is the
result of central ulceration. The symptoms of esophageal malignancy include progressive dysphagia,
first for solids, then for liquids; weight loss; regurgitation; alteration of voice; pain localized to the
retrosternal, back, or cervical regions; bronchopulmonary-related complaints such as coughing spells
at night and symptoms of aspiration pneumonia and those resulting from a bronchopulmonary fistula;
and supraclavicular lymph node enlargement. These classic symptoms usually indicate an advanced
stage of the disease.
The most important symptom is dysphagia. The dysphagia that results from malignancy is usually of
gradual onset and steadily increasing severity. At first, the patient has difficulty with foodstuffs such as
meat and apples. Dysphagia worsens until, finally, it may be impossible to swallow saliva. In other
cases, there may be sudden impaction of a bolus of food. Difficulty in swallowing usually develops long
before the entire circumference of the esophageal lumen is compromised and often before there is
radiologic evidence of impediment in the movement of barium. Rather, it begins when about half the
intraluminal circumference has been invaded. Different systems for classifying the severity of
dysphagia are summarized in Table 401.13(3121)
TABLE 401
Scoring Systems for
Dysphagia
Stoller et al., 19771
0:
All foods
1:
All soft foods
2:
Blenderized foods
3:
Clear liquids only
TABLE 401
Scoring Systems for
Dysphagia
4:
Nothing (not even saliva)
Atkinson et al., 19792

0:
Taking normal diet
1:
Unable to swallow certain solids
2:
Limited to semisolid soft diet
3:
Liquids only
4:
Unable to swallow even liquids in adequate
amounts; intravenous fluid needed
O'Sullivan et al., 19813
None 0:
Minimal 1:
Moderate 2:
Severe 3:
No dysphagia
Occasional episodes
Requires liquids to clear
Episode of meat impaction, requiring
medical treatment
The current methods of assessing the severity of dysphagia and the response to dilating therapy are
subjective and highly variable. Usually, a weak correlation is found between the dysphagia score and
objective measurements of luminal diameter. Changes in the dietary habits of patients probably
account for the discrepancy. Patients may take longer to complete a meal, they might chew their food
for a longer period of time, or they may take smaller bites.4(3122) Luminal diameter should be
assessed at the narrowest part of a stricture with a biopsy forceps (diameter closed 2 mm, with opened
cups 6 mm) or by fluoroscopically monitoring the passage of barium pills of variable diameters.
Pain is not a prominent feature, except in the late stages when the tumor has spread locally. When
present, pain is usually located in the area of the sternum and often radiates to the back.
Patients with early esophageal malignancy are either asymptomatic or may have characteristically
vague complaints, such as an awareness of the passage of food through the esophagus; mild
odynophagia with localized or radiating pain; mild, paradoxical dysphagia due to esophageal spasm
proximal to the tumor; unilateral otalgia or pain in the jugular-carotid area; minimum dysphagia due to
dysfunction of the upper esophageal sphincter; and, rarely, a discrete weight loss.
Classification, Etiology, and Pathogenesis

Epithelial Tumors
Benign Epithelial Tumors
Squamous cell papilloma is a benign tumor composed of squamous cells covering finger-like
projections of lamina propria. These lesions have a fibrovascular core without inflammatory cells,
which is covered by an acanthoid epithelium of mature squamous cells. Squamous cell papillomas
usually occur in elderly men. They perhaps have an inflammation pathogenesis that may be related to
chronic irritation, chronic esophageal reflux, or prolonged nasogastric intubation, although infection
with human papilloma virus has also been suggested.5(3123) Despite the fact that multiple squamous
cell papillomas have been found in patients with tylosis and with acanthosis nigricans, available data do
not on the whole suggest progression to malignancy.
Malignant Epithelial Tumors
Squamous Cell Carcinoma
Squamous cell carcinoma is by far the most common malignant tumor of the esophagus. Its frequency
reaches epidemic proportions in Iran, China, and certain parts of Africa. The incidence in North
American black men is increasing rapidly.6(3124)
Squamous cell carcinoma is usually graded microscopically as poorly, moderately, or well
differentiated. Well-differentiated tumors are those with abundant keratin, easily demonstrated
intercellular bridges, and minimum nuclear and cellular pleomorphism. A well-differentiated papillary
growth is also termed a verrucous carcinoma. Poorly differentiated tumors have marked cellular atypia,
nuclear pleomorphism, and little or no keratin or intercellular bridges. Sometimes, the malignant cells
may appear spindle-shaped. Moderately differentiated tumors have a microscopic appearance
intermediate to that of well-differentiated and poorly differentiated cancers.
Squamous cell carcinoma of the middle and distal esophagus shows a marked predilection for
middle-aged and elderly men, especially those who use alcohol to excess and are heavy smokers.
Carcinomas of the upper third of the esophagus account for approximately 10% of all squamous cell
cancers. They develop almost exclusively in the region of the cricopharyngeus muscle, chiefly in young
women.
PREMALIGNANT CONDITIONS. A premalignant condition is an abnormal state or situation that

creates an environment favoring the development of cancer. Several entities of differing etiology have
been shown to be linked to malignant transformation of the esophageal epithelium. However, the
proportion of esophageal cancer cases related to these conditions is very small.7,8(3125)
Squamous cell carcinoma of the proximal esophagus, especially the postcricoid area, has been related
to longstanding iron deficiency and is reported in as many as 15% of women with the Plummer-Vinson
syndrome.9(3126) A number of other associated abnormal findings are reported with this syndrome.
The incidence of Plummer-Vinson syndrome has decreased considerably in recent decades, probably
as a result of improved nutrition, although the deficiency state theory of pathogenesis does not explain
all aspects of the syndrome. Esophageal webs may be found in the absence of any of the other
findings usually associated with the Plummer-Vinson syndrome. However, it has not been established
that isolated esophageal webs are associated with carcinoma (see Chapter 39: Esophageal Motility
and Miscellaneous Disorders).
Tylosis (hyperkeratosis palmaris et plantaris) is a rare autosomal dominant genetic disorder associated
with a striking susceptibility to esophageal cancer. The disease is characterized by thickening of the
skin of the palms and soles. In the largest series published on this disorder, esophageal cancer
developed in 37% of family members.10(3127)
The incidence of malignancy in patients with caustic burns of the esophagus, especially as a result of
lye ingestion, varies from 0.8 to 5%. Usually, there is a delay of over 20 years before the appearance of
carcinoma, although exceptions occur. The older the patient at the time of lye ingestion, the earlier the
carcinoma appears. Malignant transformation usually develops in areas of extensive scarring.11(3128)
It is uncertain whether achalasia predisposes to esophageal cancer. Some investigators believe that
patients probably have only a minimum, if any, increased risk of cancer. Others describe squamous
cell carcinoma in approximately 3% of patients, with variations of 1 to 10%.1214(3129) There is
speculation that chronic irritation resulting from stasis is the etiologic mechanism. Because patients are
accustomed to interference with the normal passage of food as a result of the underlying motility
disorder and because the enlarged esophagus will accommodate a rather bulky tumor, the carcinoma
is usually advanced when detected. Cancer may occur in both surgically and nonsurgically treated
patients (Figure 401) (see Chapter 39: Esophageal Motility and Miscellaneous Disorders).
(3130)Figure 401. Midesophageal cancer in a patient with longstanding achalasia.

Contrary to the report of Joske and Benedict,13(3131) few data support the concept that peptic
stricture predisposes to esophageal malignancy.
There are isolated case reports of esophageal squamous cell carcinoma in patients who have
undergone esophageal variceal sclerotherapy.1518(3132) However, no hard evidence exists that
previous sclerotherapy predisposes patients to esophageal cancer (see Chapter 35: Complications of
Sclerotherapy).
DYSPLASIA. Certain pathologic changes in the esophageal mucosa are associated with a greater
risk of cancer. Progressively worsening mucosal dysplasia is generally considered to be premalignant.

Dysplasia is characterized by disruption of the normal epithelial architecture, with proliferation of
atypical cells and partial loss of polarity. With severe dysplasia, the mucosa is totally replaced by
dysplastic cells that, however, remain bounded by an intact, uninterrupted basal lamina. Early in the
course of malignant lesions, when there is limited infiltration, it is often possible to discern evidence of
a transition from normal mucosa through various degrees of dysplasia into focally invasive malignancy
(Figure 402).
(3133)Figure 402. Microscopic appearance of early squamous esophageal cancer. Normal

mucosa merges via areas of severe dysplasia into early infiltrating malignancy. The finding of
normal and dysplastic epithelium merging gradually with early invasive cancer is suggestive of
a histologic sequence; that is, dysplasia developing into in situ carcinoma, which, in turn,
becomes superficial esophageal cancer.
Adenocarcinoma
Esophageal adenocarcinoma is a malignant tumor arising in glandular cell-type epithelium. The
incidence of adenocarcinoma of the esophagus and of the gastric cardia has been increasing in many
industrialized countries.19,20(3134) The cause of these absolute increases in incidence is largely
unknown and cannot be accounted for by improved methods of diagnosis. Primary adenocarcinoma
may develop either in heterotopic gastric mucosa or from submucosal mucous glands. At the
gastroesophageal junction, adenocarcinoma is more common than squamous cell carcinoma. For
adenocarcinoma to be accepted as esophageal in origin, it must lie entirely within the esophagus and
be bounded by squamous cell epithelium on all sides, except in the case of adenocarcinoma arising in
Barrett's epithelium. It may be very difficult, if not impossible, to distinguish adenocarcinoma found in
the distal esophagus as being the result of proximal spread from the stomach or as arising as a true
primary growth of the esophagus.
PREMALIGNANT CONDITIONS. Most adenocarcinomas that develop within the esophagus

are probably related to a transition to columnar cell-type esophageal mucosa known as Barrett's
esophagus or endobrachyesophagus. The exact prevalence of this entity is unknown, but frequencies
as high as 11.5% have been described in patients with symptoms and signs of esophageal reflux
disease.2123(3135) The reported incidence of adenocarcinoma in patients with Barrett's esophagus
varies from 2 to 9% but may be as high as 13%.2426(3136) Cancer usually occurs in the distal third
of the esophagus, but it may develop at more proximal levels, and it can be multifocal.
DYSPLASIA. Accumulating clinical experience indicates that severe dysplasia may be considered a
reliable marker of impending adenocarcinoma (Figures 403 and 404). In the absence of overt
carcinoma, the prevalence of dysplasia is usually around 5 to 10%. Several arguments support the
view that high-grade dysplasia is to be considered an unequivocal premalignant lesion. First,
retrospective studies of Barrett's adenocarcinoma have demonstrated dysplasia in approximately 90%

of the cases in which it was sought. Second, some patients who had undergone esophagectomy
because of the presence of high-grade dysplasia were found to have previously unrecognized
adenocarcinoma in the resected specimens, often with multiple foci of dysplasia. Third, in endoscopic
follow-up studies, detection of carcinoma has been shown to be preceded by increasing severity of
dysplasia.27,28(3137) Dysplasia develops almost exclusively in the specialized columnar cell-type and
intermediate-type metaplastic epithelium (Table 402).2935(3138) Fourth, a carcinoma incidence has
been calculated at 1 in 52 patient-years to 1 in 441 patient-years of observation, which is a 30- to
40-fold increased risk compared with the general population. Although the presence of high-grade
dysplasia is an ominous sign in a patient with columnar cell-lined esophagus, the time course for
development of invasive carcinoma from high-grade dysplasia seems to vary considerably.
TABLE 402
Screening for Adenocarcinoma in Patients With Columnar Cell-Lined Esophagus
SCREENING
METHOD
PATIENTS
(NO.)
FOLLOW-UP
(YEARS)
CARCINOMA
(NO.)
INCIDENCE
PER
PATIENT-YEAR
FOLLOW-UP
Spechler et al.,
198429
Sprung et al.,
198430
Cameron et al.,
198831
Endoscopy
105
3.3
1 out of 175
Endoscopy
41
4.0
1 out of 81
Postal
122
8.5
1 out of 441
Sampliner, 198632
Robertson et al.,
198827
Endoscopy
25
2.1
1 out of 92
Endoscopy
56
3.0
1 out of 56
62
2.6
Postal
155
4.4
1 out of 170
Endoscopy
50
5.2
1 out of 52
Endoscopy
220
4.1
1 out of 150
REFERENCE
Achkar and Carey,

198833
Van der Veen et
al., 198934
Hameeteman et
al., 198928
Polepalle and
McCallum,
199035
Total screened
836
27
(3139)Figure 403. Example of severe dysplasia in Barrett's columnar cell-lined esophagus,

characterized by architectural distortion, marked nuclear atypia, and cellular pleomorphism.
Mitoses are abundant, and there is focal loss of polarity. The dysplastic cells are still bound by
an intact basal lamina.
(3140)Figure 404. Early malignant lesion, detected in a patient with Barrett's esophagus on
routine endoscopic screening, developed in an area where multiple previous biopsies (see
Figure 403) had shown severe dysplasia. Biopsies in another area revealed focia of early
multifocal invasive malignancy.
Barrett's esophagus itself (without reference to the presence of dysplasia or carcinoma) has been
treated by neodymium:yttrium-aluminum-garnet (Nd:YAG) laser photoablation in a small number of
patients.36,37(3141) Preliminary results suggest that Barrett's epithelium regresses in response to this
form of treatment, but these data must be regarded as highly preliminary in nature. Because of the risk
of carcinoma associated with Barrett's esophagus, long-term prospective studies are mandatory, and
laser photoablation must therefore be regarded as an investigational procedure until its efficacy and
safety have been established.
There is an established consensus that endoscopic surveillance should be undertaken in patients with
Barrett's esophagus.38(3142) Because of a lack of prospective data, however, endoscopic surveillance
remains controversial.34,39(3143) In particular, there is a decided lack of prospective data concerning
the benefit of surveillance. In a study of patients who underwent esophagectomy for high-grade
dysplasia or carcinoma, Streitz et al.40(3144) found that adenocarcinoma was diagnosed at an earlier
stage in 19 patients who underwent endoscopic surveillance compared with 58 patients not under
surveillance. Furthermore, the 5-year actuarial survival of patients in the former group was 62%,
compared with 20% for the latter. In a similar study, Peters et al.41(3145) found earlier-stage
malignancies in patients who had undergone surveillance, and survival was also better in this group.
Despite the consensus in support of endoscopic surveillance, numerous aspects of this form of
management of patients with Barrett's esophagus are not standardized, including the interval between
procedures, the technical protocol for obtaining biopsies (number and pattern), and the histopathologic
classification of dysplasia in association with Barrett's esophagus. Because the presence of high-grade
dysplasia and carcinoma may not be evident at endoscopy, Levine et al.42(3146) recommended a
rigorous protocol for obtaining biopsies that includes the use of large ("jumbo") forceps to obtain
specimens in four quadrants at 2 cm from the level of the lower esophageal sphincter to the
squamocolumnar junction. Additional biopsies are obtained from any mucosal abnormality, no matter
how negli-gible its appearance at endoscopy. Biopsies obtained at each level are specifically identified
by placing them in separate bottles of fixative marked as to location. A brush cytology specimen of the
Barrett mucosa is also included.
Several expert pathologists have recommended that surveillance biopsies be classified as follows: (1)
negative for dysplasia; (2) low-grade or indefinite dysplasia; (3) high-grade dysplasia; (4) intramucosal
carcinoma; and (5) submucosal carcinoma.43,44(3147) There is no distinction in this five-level system
between carcinoma in situ and high-grade dysplasia.
Endoscopic surveillance with histopathologic evaluation of multiple biopsies for dysplasia is an
imperfect method for identification of patients at increased risk of adenocarcinoma. Accordingly, a
number of additional methods have been proposed in an effort to enhance the accuracy of endoscopic
surveillance including flow cytometry,45,46(3148) p53 gene expression,47(3149) ornithine
decarboxylase activity,48(3150) endoscopic ultrasonography (EUS),49(3151) and
chromoscopy.50(3152) Although some of these techniques appear promising, currently available data
are insufficient to support their use in clinical practice.
The majority of investigators recommend further surveillance at shortened intervals (3 to 6 months)
plus rigorous antireflux treatment for patients with low-grade or indefinite dysplasia. If dysplasia is not
confirmed at two subsequent surveillance endoscopies, the interval between procedures can be
lengthened. Low-grade dysplasia may persist for a number of years, and in some cases, it appears to
regress.51(3153)
The management of patients with Barrett's esophagus and confirmed high-grade dysplasia remains
controversial. Esophagectomy is favored by many but not all investigators.28,42,5255(3154) Although
ample evidence exists that high-grade dysplasia progresses to carcinoma, the length of time required
for this transformation is uncertainthis appears to be on the order of 6 months to 2 years. However,
there are well-documented cases in which overt carcinoma did not develop during periods of close
observation of up to 57 months.28,42,43(3155) Because of the morbidity and mortality of
esophagectomy, it can be argued that surgery should be deferred until there is definite evidence of
carcinoma. However, it is also established that unsuspected carcinoma will be found in perhaps as
much as a third of patients who undergo esophagectomy because of high-grade
dysplasia.53,56,57(3156)
Endoscopic methods such as Nd:YAG laser photoablation and photodynamic therapy have been
proposed for the treatment of patients with Barrett's esophagus and high-grade dysplasia or
early-stage carcinoma (see Chapter 19: Photodynamic Therapy).58(3157) An endoscopic alternative to
surgery would be of considerable value, especially in patients who are not candidates for
esophagectomy because of concurrent illnesses. However, data are limited and it is not yet established
that these forms of therapy are curative.
Rare Epithelial Malignancies
Depending on histologic characteristics, some rare forms of epithelial malignancy may be
distinguished, such as adenoid cystic carcinoma,59(3158) mucoepidermoid carcinoma,60(3159)
adenosquamous carcinoma (intermingled glandular and squamous components), undifferentiated
carcinoma without recognizable glandular or squamous elements, carcinosarcoma, and infiltrating
adenocarcinoma causing a pseudoachalasic image (Figure 405). A distinct but rare form of
undifferentiated carcinoma, referred to as oat cell carcinoma, sometimes associated with production of
hormones, has also been described.6163(3160)
(3161)Figure 405. A, Radiograph of carcinoma at the cardia, infiltrating under normal

squamous epithelium. B, Biopsy specimen obtained with large forceps in a patient with
pseudoachalasia. Photomicrograph demonstrates squamous cell mucosa (top) and
adenocarcinoma (bottom).
A carcinosarcoma consists of an intimate mixture of carcinomatous elements (often as epithelial pearls
and foci of highly anaplastic squamous cells) and sarcomatous cells growing together as a single
tumor. Carcinosarcoma of the esophagus is usually a polypoid nonulcerating lesion, only superficially
invasive, that metastasizes comparatively late in its course (Figure 406).
(3162)Figure 406. Polypoid exophytic nonulcerating carcinosarcoma obstructing the

esophageal lumen.
Nonepithelial Tumors
Benign Nonepithelial Tumors
Leiomyoma is the most common type of benign nonepithelial tumor encountered in the esophagus.
Usually found in the lower third of the esophagus, these lesions are mostly solitary, although multiple
tumors have been encountered. In some rare cases of diffuse leiomyomatosis, the esophagus is
studded throughout with these tumors.64(3163)

Other rare, solid intramural lesions are the fibroma, lipoma, neurofibroma, osteochondroma, and
glomus tumor.
Lipomas usually develop in the cervical segment of the esophagus, although they may occur at any
level. They are exceptionally rare in the esophagus and are frequently pedunculated.65,66(3164)
Granular cell myoblastoma, also called granular cell tumor (Abrikosov's tumor), is characterized
histologically by large polygonal or fusiform cells with eosinophilic cytoplasm and eccentric nuclei
arranged in compact nests. Usually the cells are S-100 positive on immunohistochemical examination,
a point favoring a neural origin. The overlying squamous mucosa may reveal evidence of hyperplasia
or acanthosis, which may be misleading on biopsy. The great majority occur in women; most are
located in the upper and lower third of the esophagus.67(3165) Multiplicity is possible.
Hemangiomas are composed of endothelium-lined spaces, varying in size from fine capillaries to larger
cavernous structures. Hemangioma of the esophagus is a rare entity, making up only 2 to 3% of
benign esophageal neoplasms.68(3166) Hemangiomas are evenly distributed along the length of the
esophagus. Size varies widely, the average diameter being 2 to 3 cm. Owing to the propulsive effect of
peristalsis, these lesions may develop long pedicles and become very mobile. Hemangiomas are
usually considered to be vascular malformations (hamartomas) rather than true neoplasms. Most are
probably congenital.
Malignant Nonepithelial Tumors
Leiomyosarcoma is the most frequently encountered malignant nonepithelial tumor. This lesion is
exceptionally rare, accounting for substantially less than 1% of malignant tumors.69(3167) In a review
of the world literature in 1986, Choh et al.70(3168) found less than 50 cases. Only a few case reports
include endoscopic findings.7174(3169) Consistent features include large size, ulceration, and rapid
growth.
Esophageal leiomyoblastoma is also an exceedingly rare lesion of low-grade malignant potential.
Endoscopic findings in the case report of Cantero et al.75(3170) were those of a large mass with
ulceration.
Lymphomatous and leukemic involvement of the esophagus appears to be rare. There are relatively
few reports of primary lymphoma of the esophagus.7679(3171) As with lymphomas arising elsewhere
in the gastrointestinal tract, the endoscopic appearance of primary lymphoma and secondary
lymphomatous involvement is highly variable. In several case reports, the appearance has included
extensive multiple ulcers, a solitary ulcerated mass, a mass lesion without ulceration, and polypoid
tumors.7680(3172)
Malignant melanoma of the esophagus is most commonly metastatic rather than primary (Figure
407). Metastasis of extraintestinal melanomas to the gastrointestinal tract usually produces numerous
black, umbilicated nodules.81(3173) The very rare primary lesion usually presents as a large, polypoid,
pedunculated lesion.8286(3174) Endoscopic Nd:YAG laser photoablation has been employed as a
method of palliation.87(3175)
(3176)Figure 407. A, Esophageal metastasis of melanosarcoma. B, Primary amelanotic

melanoma of the esophagus. (B, From the collection of Dr. M. V. Sivak, Jr.)
Esophageal involvement with Kaposi's sarcoma is regularly seen in patients with the acquired
immunodeficiency syndrome (AIDS) (Figure 408).
(3177)Figure 408. Esophageal localization of Kaposi's sarcoma.

Carcinoid tumors and choriocarcinoma have, on rare occasions, arisen in the esophagus.
Occasionally, biopsy specimens of an esophageal lesion will show large numbers of enterochromaffin
carcinoid cells that, after detailed study of the resected specimen, are shown to accompany a deeply
infiltrating, poorly differentiated adenocarcinoma.
Tumor-Like Lesions
Heterotopias
Islands of gastric heterotopia may be found in the proximal esophagus as small, sometimes slightly
elevated lesions. Such an inlet patch can be detected in up to 3.8% of consecutive
endoscopies88(3178) (see Chapter 39: Esophageal Motility and Miscellaneous Disorders). In the lower
esophagus, heterotopic tissue is often more extensive and usually continuous with the normal
glandular mucosa of the stomach. Gastric heterotopia in the lower esophagus, a subject of some
debate, is either congenital or associated with chronic reflux esophagitis.
Heterotopic Sebaceous Glands
Heterotopic sebaceous glands in the esophagus are rare and seen as yellow-gray plaquelike lesions.
They presumably develop through metaplasia of pluripotent cells in esophageal salivary-type mucous
glands.89,90(3179)
Congenital Cysts
Congenital cysts are malformations lined entirely by epithelium that is ciliated, squamous, or gastric, or
a combination of more than one of these types. They are usually the result of a developmental
abnormality occurring during formation and differentiation of the lower respiratory tract, esophagus,
and stomach.
Esophageal Duplications
Esophageal duplications are tubular or spherical structures, adherent to the esophagus. Esophageal
duplications rarely communicate with the esophageal lumen. They are usually lined with gastric-type
mucosa and a smooth muscle wall. They may be localized centrally but are more often on one side,
especially the right.
Fibrovascular (Fibrous) Polyps
A fibrovascular polyp is a polypoid lesion composed of a core of fibrous or adipose connective tissue
and blood vessels, covered by thickened but otherwise normal squamous epithelium.91(3180) This
lesion is almost invariably solitary. This lesion may be differentiated from a myoma by its elasticity and
tendency to develop a pedicle (Figure 409). When situated proximal to a sphincter, such polyps may
acquire the shape of a bell clapper. Most fibrovascular polyps are located in the proximal third of the
esophagus, and because most are pedunculated, regurgitation with subsequent asphyxia may
occur.92(3181)
(3182)Figure 409. Fibrovascular polyp with a long pedicle seen in a retroverted view from
the gastric cardia. (From the collection of Dr. M. V. Sivak Jr.)
The term glycogenic acanthosis has been applied to focal areas of excessive glycogen accumulation in
the superficial squamous cells of the mucosa. Mucosal biopsy specimens reveal hyperplasia of the
glycogen-rich intermediate squamous cells without cytologic atypia. This lesion, which has no known
role as a cancer precursor, is also discussed in Chapter 39: Esophageal Motility and Miscellaneous
Disorders.
Secondary Tumors
Tumors of adjacent organs, most commonly the bronchus or stomach, often invade the esophagus.
Occasionally, mediastinal lymph nodes as a site of metastatic cancer from another organ will provide
an avenue for esophageal involvement and obstruction. Metastatic esophageal involvement may also
occur in patients with breast cancer, sometimes many years after treatment of the primary lesion.
Diagnostic Techniques
Endoscopy
Benign Lesions
Papillomas
A papilloma is an exophytic excrescence of the esophageal epithelium. Papillomas appear as small
wartlike projections on the mucosal surface (Figure 4010). They are preferentially located in the distal
esophagus and are often multiple. Occasionally, large areas covered with plaques or verrucous
projections consisting of epithelial hyperplasia and sometimes hyperkeratosis may be observed in
association with chronic irritation such as occurs with longstanding achalasia. These areas have the
endoscopic appearance of whitish, slightly elevated mucosal nodules (Figure 4011). Although there
are numerous case reports of squamous papillomas, these lesions are encountered in less than 1% of
endoscopic examinations of the esophagus.9395(3183)
(3184)Figure 4010. Esophageal papilloma seen as an irregular wartlike projection.
(3185)Figure 4011. Nodular thickening of the esophageal squamous mucosa in longstanding

stasis.
Papillomas must be distinguished from verrucous squamous cell carcinoma and proliferating
granulation tissue.96(3186) Verrucous squamous cell carcinoma is also papillary and well
differentiated, but careful microscopic examination will reveal epithelial dysplasia and superficial
carcinomatous invasion (Figure 4012). Multiplicity, the usual case with papilloma, strongly favors a
benign lesion. When there is doubt or uncertainty regarding the true nature of a papillomatous lesion
based on endoscopic appearance and biopsy, endoscopic excision, if feasible, can be
undertaken.9799(3187) Because squamous papillomas are not associated with an increased risk of
malignancy, endoscopic excision is not necessary when the diagnosis has been established by
endoscopic biopsy.
(3188)Figure 4012. Verrucous squamous cell carcinoma of the distal esophagus. (From the
Leiomyomas
Esophageal leiomyomas exhibit great variation in size and shape. Most are spherical or ovoid,
although many (25% in one study100(3189)) may be crescentic or even annular in configuration. The
most common shape is roughly oval, with dimensions of approximately 6 to 8 cm in length and 4 to 6
cm in width. The endoscopic appearance is that of a bulging mass with intact or focally ulcerated but
freely movable mucosa stretched over the lesion (Figure 4013). The surface may be regular and
smooth or multilobulated. As leiomyomas are not infiltrative, the endoscope can usually pass these
lesions without difficulty, unless the tumor is completely or almost completely annular. The endoscopist
should be aware that the mobility of such intramural lesions may permit them to be displaced easily,
and hence, they can be overlooked at esophagoscopy. A leiomyoma arising in the distal esophagus
may rarely cause a clinical picture similar to that of achalasia.101(3190) By demonstrating that
submucosal tumor arises within the muscularis propria, endoscopic ultrasonography (EUS) can be
especially useful in establishing the diagnosis of leiomyoma.
(3191)Figure 4013. Esophageal leiomyoma covered with normal-appearing mucosa.

As a rule, biopsy is not advisable if the mucosa overlying a leiomyoma is normal in appearance.
Mucosal biopsies will not usually reveal the nature of the lesion, except in some cases when they are
taken from an ulcerated area. Biopsy is justified only if roughening, irregularity, or peculiar ulceration of
the mucosa overlying the tumor suggests malignancy. The decision to perform extramucosal
enucleation or partial resection is usually made at surgery, and there is a concern that multiple
endoscopic biopsies can interfere with enucleation. Thoracoscopic excision of esophageal leiomyomas
has been described.102(3192)
Endoscopic techniques for resection of esophageal submucosal tumors up to 2 cm in diameter have
been described.103,104(3193) Selection of tumors for resection relies heavily on EUS findings. These
methods should be regarded as investigational because experience with them is exceedingly limited
and their safety and efficacy have not been established.
Myoblastomas
Granular cell tumors (myoblastomas) appear endoscopically as tiny, whitish, sometimes slightly
umbilicated, submucosal lesions in the distal esophagus (Figure 4014).105(3194) About 20% of these
tumors are multiple, seen as slightly elevated plaques 0.5 to 1 cm in diameter.106(3195) In one case,
EUS demonstrated a hypoechoic mass in the submucosa without continuity to the muscularis
propria.107(3196) Granular cell tumors are not premalignant lesions, and excision is not required for
small lesions in asymptomatic patients.108,109(3197) If such tumors cause dysphagia or pain or are
larger than 1 cm, then local resection is the treatment of choice.108,110112(3198) Several such
lesions, up to 1 cm in diameter, have been removed endoscopically without untoward
effects.107,110,113,114(3199)
(3200)Figure 4014. Two small nodular elevations (one is umbilicated), covered by

normal-appearing squamous mucosa, that proved to be granular myoblastomas.
Hemangiomas
It may be possible to differentiate a hemangioma from other benign tumors at endoscopy if the lesion
has a pale, bluish coloration and if it empties when pressure is applied with the endoscope. However,
often the mucosa covering a hemangioma is indistinguishable from that of the surrounding esophagus
(Figure 4015). Usually, hemangiomas are not sufficiently small or pedunculated to a degree that
would make them suitable for removal by electrosurgical snare polypectomy. In principle, no biopsy
should be taken of any vascular lesion, because of the possibility of severe hemorrhage. This is
problematic when the nature of the lesion is not obvious endoscopically and biopsy is contemplated to
exclude malignancy. Data are limited on the consequences of endoscopic biopsy of such a lesion. The
endoscopist must be prepared to deal with profuse bleeding. There are no reports of extensive
experience in the endoscopic treatment of esophageal hemangiomas. If the lesion is large, attempted
endoscopic control of bleeding could result in more profuse bleeding. Theoretically, such lesions might
be treated by laser photocoagulation, but data are inadequate to support this point.
(3201)Figure 4015. Esophageal hemangioma. (From the collection of Dr. M. V. Sivak, Jr.)
Polyps
Fibrovascular polyps appear endoscopically as oval or elongated, lobulated, sausage-like masses with
a smooth or corrugated surface. It can be difficult to recognize the site of attachment. When the lesion
is bulky and highly vascular and has a broad insertion, surgical rather than endoscopic removal seems
indicated, although some large lesions have been removed successfully with endoscopic
polypectomy.115(3202) Laser photoablation has also been described.116(3203)
Cysts
Congenital cystic lesions of the esophagus (e.g., teratoma, epidermoid cyst, duplication cyst,
branchiogenic cyst, enterogenous cyst, or coelom cyst) usually have no characteristic endoscopic
appearance. Such lesions manifest as a nonspecific, extrinsic compression or an intramural tumor
(Figure 4016). Retention cysts are round or fusiform structures that have a yellowish transparency
and a smooth surface. They can be opened with a needle or biopsy forceps. This may result in the
discharge of a viscous orange-yellow fluid. Retention cysts are usually solitary and have no preferential
location. An isolated retention cyst is easily distinguished from polycystic chronic esophagitis
(esophagitis cystica) in which the mucosal surface is covered by innumerable retention cysts.
Microdiverticulosis and sometimes an inflammatory stricture may coexist with the latter entity.
(3204)Figure 4016. Esophageal cyst. (From the collection of Dr. M. V. Sivak, Jr.)
Glycogenic acanthosis (see Chapter 39: Esophageal Motility and Miscellaneous Disorders) is
commonly noted at endoscopy and is usually easy to differentiate from other lesions and tumors. The
plaques, usually less than 1 cm in diameter, sometimes tend to align along the longitudinal folds of the
esophagus and will dye blue with application of 2% Lugol's solution because of their increased
glycogen content.
Premalignant Lesions and Conditions
Endoscopic screening for severe dysplasia can be undertaken when conditions exist that predispose to
esophageal malignancy. Preferably, this should be done annually by endoscopy with multiple biopsies
with or without brush cytology. However, the most efficacious schedule of procedures and methods
has not been established. Unfortunately, data concerning the outcome of this approach are limited. It is
not known with certainty, for example, that such screening will result in curative treatment of significant
numbers of patients. Furthermore, the exact risk of cancer is not established for all conditions that are
considered to be premalignant. Screening should be decided on an individual case basis, because
data accumulated from large series of patients undergoing endoscopic screening are not available as
guidelines.
Systematic endoscopy of the upper digestive tract is indicated in patients with squamous malignancy of
the structures of the head and neck, and probably when mucosal dysplasia is discovered in this area,
because such tumors are known to be associated with carcinoma of the esophagus.117,118(3205)
Meticulous endoscopic examination of the esophagus has disclosed a second esophageal primary
cancer in from 1.8 to 7.3% of patients in reported series.119123(3206) A second primary malignancy
may also be present elsewhere in the head and neck or the respiratory tract.114(3207) Because
second primary cancers of the esophagus may be small and potentially early-stage lesions in these
patients, endoscopic examination using chromoscopy techniques with toluidine blue or Lugol's solution
is mandatory.122125(3208)
The endoscopic identification of dysplasia is problematic. Occasionally, dysplastic areas may be visible
endoscopically as whitish plaques, although most often such plaques consist of innocent-looking areas
of mucosal thickening. In some patients, dysplasia may be discovered in areas that appear inflamed or
manifest increased erythema or vascularity. In other patients, dysplasia may be present in areas that
appear as shallow depressions of the often slightly discolored mucosal surface. Despite these
observations, no peculiar endoscopic features allow unequivocal distinction between dysplastic and
nondysplastic mucosa.
Rarely, adenomatous polyps develop in the esophagus, especially in patients with columnar
metaplasia. The significance of this finding is uncertain, but the occurrence of a benign neoplasm in
Barrett's epithelium with its known premalignant potential should be regarded with circumspection.
Rare instances of multiple adenomatous lesions in Barrett's esophagus have been reported.126(3209)
Early Esophageal Cancer
Early esophageal cancer can be defined as malignancy that is confined to the mucosa and submucosa
and is without lymph node metastasis. Descriptions of the endoscopic appearance of early esophageal
malignancy come mainly from China127130(3210) and Japan,131,132(3211) although these lesions
are being recognized with greater frequency in other countries.133(3212)
Early esophageal malignancy may appear as a superficial erosion (Figure 4017). Usually, the
cancerous mucosa is quite fragile and bleeds easily on contact with the endoscope. The cancerous
area appears slightly depressed, with gray erosive spots against a reddish background. These
reddened areas usually have geographic shapes. Between the erosive defects, islands of whitish,
normal-looking mucosa may be seen.
(3213)Figure 4017. Discrete abnormalities of the esophageal lining due to early esophageal
carcinoma.
The endoscopic appearance may also be that of a plaquelike lesion in which the cancerous area
appears slightly elevated, with a granular or coarse knobby surface. Occasionally, small plaques are
scattered in the surrounding mucosa. In other areas these whitish plaques may be confluent, giving the
surface an orange-peel appearance. The cancerous epithelium of plaquelike lesions is usually
markedly thickened.
Sometimes, early esophageal malignancy presents as a flat patch of localized edema and congestion
(congestive type). Reddish spots may be found within and about the congested mucosal area. This
localized mucosal roughening and hyperemia has been compared to tree bark. The abnormal mucosa
tears easily, and friability (bleeding that develops with slight trauma) is a prominent feature.
Early esophageal cancer may also look like a circumscribed polypoid or protruded lesion (papillary
type) (Figure 4018). The irregular, thickened mucosa may have a granular, nodular, papillary, or
polypoid appearance. The diameter of this type of lesion is often less than 3 cm. The polypoid
protrusion sometimes has a relatively wide mucosal base, and occasionally, friability or superficial
erosive defects may be present.
(3214)Figure 4018. Polypoid projection in the distal esophagus due to early esophageal
carcinoma.
In some patients, the only visible abnormality is either a circumscribed area of altered pliability or a
circumscribed patch of mucosal discoloration (occult type).134(3215) These early cancers are
obviously difficult to detect. Only meticulous attention to detail, along with adequate biopsy sampling of
any suspicious area, will detect this type of early cancer.
In the Japanese system of classification, early esophageal cancers are categorized as protruding,
superficial, and ulcerating. This descriptive classification is in keeping with the accepted endoscopic
nomenclature for early gastric cancer.131(3216) However, the term erosive would seem more
appropriate than ulcerated.
The plaquelike and erosive types of early cancer in the esophagus may be difficult to differentiate
endoscopically from the mucosal changes resulting from reflux esophagitis. However, a zone of normal
mucosa is usually found between the distal edge of the neoplastic lesion and the gastroesophageal
junction. A normal mucosal segment is not a feature of reflux esophagitis, and this may be a useful
differential criterion.
Early malignancy may also be difficult to detect in patients with a columnar cell-lined esophagus,
especially with a Barrett-type ulcer (Figure 4019).134(3217) The columnar epithelium surrounding the
ulcer may be heaped up and polypoid, thus suggesting adenocarcinoma. However, the mucosal
pattern of this polypoid mucosa can sometimes be distinguished from the more disorganized pattern
with focal areas of discoloration seen in malignancy. Furthermore, fibroepithelial, inflammatory,
pseudopolypoid projections may sometimes be present at the margins of a Barrett-type ulcer.
However, the true nature of lesions that are easily misinterpreted at endoscopy can be determined only
by adequate biopsies.
(3218)Figure 4019. Irregular ulceration in a columnar cell-lined esophagus due to early

Barrett's adenocarcinoma.
There are reports of endoscopic treatment of early-stage esophageal cancers including
resection,135,136(3219) Nd:YAG and photodynamic therapy,137(3220) and injection
techniques.138(3221) Although experience with this approach to treatment is very limited, available
data appear to suggest that short-term results are favorable for the majority of patients. However, no
long-term prospective follow-up studies with large numbers of patients have been performed, so that
the frequency of recurrence or metastasis is essentially unknown. Moreover, differences in the several
proposed techniques of endoscopic resection make it difficult to know which method is the safest and
most efficacious. It is also uncertain whether current methods of tumor staging are adequate for the
diagnosis of early-stage esophageal cancer. Whether the selection of patients should depend on
accurate staging is debatable, since histopathologic staging can be performed on the resected
specimen. Until further data and information become available, endoscopic resection of early-stage
esophageal cancer should be regarded as an investigational procedure.
Advanced Cancer
The macroscopic appearance of advanced cancer is traditionally of three main types, although many
combinations and transitional forms exist. These types are (1) exophytic-polypoid, (2) ulcerative, and
(3) diffusely infiltrating. The exophytic-polypoid growth is usually characterized as a bulky,
cauliflower-like, wide-based lesion (Figure 4020). The coarsely nodular surface is often hemorrhagic
and may show additional erosive or ulcerative defects (Figure 4021). The ulcerative type is
characterized by central meniscoid-type necrosis (Figures 4022 and 4023) surrounded by
heaped-up edges (Figure 4024). Usually, a major part of the circumference of the esophageal lumen
is involved by the tumor. A diffusely infiltrating scirrhous cancer is manifest as obvious thickening and
rigidity of the esophageal wall of variable length, with fixation of the irregularly thickened, coarsely
nodular, rarely ulcerated mucosa to the deeper layers. Luminal narrowing is almost always present
(Figure 4025). The most common and classic endoscopic appearance of advanced esophageal
cancer is that of an eccentrically located ulcerated mass, with a projecting and overhanging margin
located in a rigid, aperistaltic segment.
(3222)Figure 4020. Exophytic adenocarcinoma growing intraluminally in a columnar

cell-lined esophagus. Note the sharply demarcated transition from columnar to squamous
mucosa opposite the carcinoma.
(3223)Figure 4021. Bulky exophytic mass with a coarsely nodular, hemorrhagic, focally
eroded, ulcerated surface.
(3224)Figure 4022. Central malignant ulceration in a tumor that has the shape of a meniscus.
(3225)Figure 4023. Large central ulceration creates the radiologic appearance of a

malignant meniscus sign.
(3226)Figure 4024. The proximal heaped-up edges of the lesion with a central ulceration
that is depicted in Figure 4025.
(3227)Figure 4025. Diffusely infiltrating malignancy causing severe luminal constriction,

surrounded by nodular elevations covered with discolored, partially necrotic mucosa.
A peculiar form of presentation, referred to as superficial spreading cancer, is defined as a lesion with
an intramucosal extension of malignancy of at least 2 cm from the main portion of the cancer. The
boundaries between involved and uninvolved mucosa may be indistinguishable.139(3228)
The endoscopist's impression may be inadequate for assessing the extent of disease in the
esophagus. From an endoscopic standpoint, it is important to realize that cancer has a tendency to
spread submucosally and to establish satellite lesions at some distance from the obvious primary
lesion (Figure 4026). Tiny mucosal elevations proximal to a main lesion should be given special
attention. These often have a somewhat yellowish hue. Biopsy usually proves these to be malignant
(Figure 4027). They probably are intramural metastases, although some authorities contend that such
apparent (sub)mucosal spread may actually be primary intramucosal carcinoma, with the intraepithelial
changes representing a form of "field cancerization."
(3229)Figure 4026. Satellite lesion several centimeters proximal to a malignant stenosis

caused by an infiltrating carcinoma.
(3230)Figure 4027. Small nodular elevations probably due to intramucosal metastases

proximal to a distal malignant obstruction seen at the top of the figure.
Careful endoscopic examination of the fundus and cardia of the stomach is mandatory. Detailed
inspection after retroflexion of the endoscope assists in the detection of cancer that has extended
proximally from the stomach into the esophagus.
According to Savary and Miller,140(3231) the following items compose the specified protocol for
endoscopic assessment of esophageal cancer:
1. Location of the proximal margin, usually measured from the incisor teeth, or the mandibular ridge
in the edentulous patient.
2. Extent and location longitudinally and circumferentially (tumor origin as from upper, middle, or
lower third influences therapeutic decisions).
3. Position of the distal edge in relation to the cardia.
4. Type of tumor (e.g., exophytic, ulcerating, or infiltrating).
5. Degree of mobility or fixation.
6. Diameter of the existing lumen.
7. Appearance of the mucosa proximal to the cancer.
8. Histologic type.
9. Status of the larynx, including recurrent laryngeal nerve function.
10. Tracheobronchoscopic examination, performed to detect not only tracheobronchial cancerous
involvement or a tracheobronchial fistula (Figure 4028) but also (early) concomitant lesions in the
respiratory tract.
(3232)Figure 4028. Large bronchoesophageal fistula (left). The original esophageal lumen
(right) is obstructed by an exophytic cancerous mass.
It must be stressed that it is almost impossible to predict the true depth of infiltration of an esophageal
malignancy from the luminal aspect alone. Frequently, lesions that appear to be small and limited
endoscopically ultimately prove to be deeply penetrating malignancies. Adequate staging should be
performed according to the tumor-node-metastasis (TNM) classification.141(3233) EUS plays a crucial
role in esophageal cancer staging, together with computed tomography and magnetic resonance
imaging (see Chapter 41: Endoscopic Ultrasonography of the Esophagus, for a discussion of the TNM
classification and EUS staging of esophageal cancer).
Vital Staining (Chromoscopy)

Chromoscopy (in vivo dye-staining or dye-scattering) is being used increasingly in some centers to
facilitate detection of early esophageal malignancy. Although its application is not worldwide, those who
have extensive experience with this method advocate its use for detection of small malignant
lesions142(3234) (specific techniques for this method are discussed in Chapter 12: Chromoscopy).
The stains used most often are 1 to 2% Lugol's solution, 1 to 2% toluidine blue, and 1 to 2% methylene
blue. Lugol's solution stains nonkeratinized squamous epithelium in proportion to its intracellular
glycogen content. Injured squamous epithelium or metaplastic columnar epithelium remains unstained
or weakly stained, depending on the severity of the mucosal injury. Early esophageal cancer
consistently shows negative staining. Lugol's solution is not useful for isolating dysplastic areas
because about 50% of such areas stain positively.
Nuclear DNA has a much stronger affinity than cytoplasmic RNA for toluidine blue.143(3235)
Dysplastic or neoplastic epithelium usually stains blue; the variable appearance includes tiny spots;
interlacing stripes; irregularly shaped, maplike areas; and extensive diffuse staining with sharp edges.
Methylene blue can be used to stain columnar epithelium. It may be taken up by intestinalized cells and
goblet cells but not by squamous epithelium. Methylene blue also accentuates mucosal relief, which
differs for columnar and squamous epithelium.
Chromoscopy may facilitate detection and determination of the extent of cancerous lesions, including
those not visible endoscopically. Multiple biopsies and cytologic specimens must be obtained from any
suspicious area.144,145(3236)
Biopsy and Cytology

A biopsy forceps with a central spike (bayonet) within the forceps' cups is probably the best instrument
for obtaining biopsy specimens from the esophagus. The spike is helpful in fixing the forceps on the
mucosa when the surface of the area of interest is tangential to the trajectory of the biopsy forceps as it
is advanced toward the lesion. A small forceps (7 French) does not permit full-thickness mucosal
sampling. This may be a problem when carcinoma is infiltrating submucosally. In this circumstance, a
larger forceps (9 French) may be useful. Repetitive biopsies in the same location can also be used to
sample deeper submucosal layers.
The histologic diagnosis of polypoid and meniscoid tumors is not difficult when the proximal tumor
margin is visible. Preferentially, biopsies should be taken from viable tumor tissue or from border areas
between obvious proliferative growth and the area of invasion. Central necrotic, ulcerated areas and
edematous adjoining walls should be avoided because the yield will be lower in these regions. Cancer
often extends proximally, and it may be useful to sample areas with mucosal pallor or tiny satellite
nodules. Multiple biopsies in longitudinal sequence along the esophageal wall will aid in defining the
proximal border of the cancer and in disclosing associated dysplasia. Contact between the endoscope
and the tumor is often sufficient to cause bleeding. This may lead to imprecise biopsy localization and
should be avoided if possible.
Esophageal cancer commonly presents as a stenotic lesion with apparently intact mucosa
characterized by a moderately irregular surface. Infiltration of the esophageal wall usually provokes a
fibrous reaction, this being perhaps the main cause of the stenosis. Moderately severe narrowing
should be investigated with a small-caliber endoscope. Usually, the endoscope may be passed through
such a lesion, so that biopsies can be obtained not only from the proximal edge but also from within the
tumor. Preferred biopsy sites include small nodular elevations, areas of whitish-yellowish discoloration,
and the edges of eroded areas.
A malignant stenosis may be so tight that it is not possible to pass any endoscope beyond the lesion.
Sometimes, biopsy from within the stenosed segment is impossible and tissue can be obtained only at
a point proximal to the stenosis. In this circumstance, it is not uncommon that biopsies, even in large
numbers, fail to establish a diagnosis. Inflammatory changes or a nonspecific fibrous reaction is a
common microscopic finding. Diagnostic yield can be increased by introducing the forceps into the
stricture to obtain tissue blindly. Unfortunately, it may be difficult to obtain adequate tissue by this
means, and the blind nature of the maneuver in itself lowers the incidence of positive results. In this
circumstance, it is especially important to obtain cytologic specimens from within the narrowed section.
An alternative approach is to gently dilate the stricture to allow guided biopsies from within and from
the distal border of the lesion, using a small-caliber endoscope. This maneuverthat is, dilation and
immediate biopsy or cytologyhas been shown to be practical and safe.146(3237) With any distal
esophageal lesion, it is essential that a retroflexed view of the gastroesophageal junction is performed
from the gastric side to allow biopsy of this area.
Most endoscopists prefer a brush for obtaining cytology specimens. Generally, the type used has a
plastic sheath into which the brush is withdrawn when not in use. This prevents contamination and loss
of specimen material as the brush is withdrawn through the accessory channel of the endoscope. If
both cytology and biopsy specimens are to be obtained, the cytology sample should be collected first.
Otherwise, blood in the area to be brushed will dilute the cytology sample. The brush should be swept
over the slough from ulcerated and necrotic areas, the rim and margins of the lesion, and the
surrounding tissue. Brush cytology is especially effective in the diagnosis of infiltrative lesions that
produce severe luminal narrowing, since the brush can be directed into the stricture. Positive samples
can be obtained in cases in which biopsies of the proximal rim are negative.
Salvage cytology can be performed by aspirating the contents of the accessory channel of the
endoscope into a mucous trap attached to the suction connector on the light guide section of the
universal cord. Fluid aspiration between biopsies yields a pool of cytology material by recovery of
material that clings to the biopsy forceps and is dislodged within the endoscope.147(3238)
Fine-needle aspiration is a new technique for obtaining cytologic samples.148150(3239) Usually, the
aspiration samples are taken with a sclerotherapy needle with a 4-mm, 23-gauge retractable needle.
Suction is applied while the needle is plunged in and out of the tissue. With the addition of fine-needle
aspiration, an accuracy of 100% may be achieved because aspiration cytology appears to have a high
diagnostic accuracy.
In several studies reported, a correct biopsy diagnosis of esophageal carcinoma was obtained in over
90% of patients, provided that six or more specimens were taken (Table
403).128,148,151165(3240) The higher the number of biopsies, the greater the chance of hitting
nonnecrotic tumor tissue. There is evidence that the yield of the first biopsy specimen is higher than
that of subsequent ones.158(3241) It is frequent, especially with distal infiltrating lesions, that more
than 12 specimens are needed to obtain a correct diagnosis; even with this number, a positive tissue
diagnosis occasionally may not be obtained. If, based on endoscopic assessment, esophageal cancer
is suspected but the diagnosis is not confirmed by cytologic and tissue sampling, then the procedure
must be repeated. The problematic nature of the false-negative biopsy becomes most significant when
the presence of carcinoma is not obvious at endoscopy. This pertains mostly to strictures that cannot
be distinguished from those due to peptic reflux. The distinction between malignancy and florid
proliferation of regenerating immature squamous cells in esophagitis may also be difficult for the
histopathologist.
Diagnostic Accuracy of Biopsy and Cytology in Malignancy of the

Esophagus and Cardia
TABLE 403
REFERENCE
PATIENTS
(NO.)
BX (NO.)
POS. BX
(%)
POS. CYTOL.
(%)
Nakamura et al.151
Kobayashi et al.152
>3
88
100
Prolla153
25
34
72
92
Bruni and Nelson154

Seifert and Atay155
103
68
94
87
Winawer et al.156
Hishon et al.157
30
Winawer et al.158
Witzel et al.159
POS. BX AND
CYTOL. (%)
85
64
81
4
20
66
97
50
90
100
100
47
610
77
89
Prolla et al.160
Eastman et al.161
73
68
78
89
69
71
76
Gutz and Wildner162
75
93
Diagnostic Accuracy of Biopsy and Cytology in Malignancy of the

Esophagus and Cardia
TABLE 403
REFERENCE
PATIENTS
(NO.)
Mortensen and
MacKenzie163
36
Graham et al.164
Yang et al.128
27
Wesdorp and Tytgat

Cusso et al.165
154
309
Kochhar et al.148
46
BX (NO.)
>7
POS. BX
(%)
POS. BX AND
CYTOL. (%)
81
86
92
96
88*
100
80
88
77
93.5
98
89
B 80
100
115
812
POS. CYTOL.
(%)
98
N 89
Bxbiopsies; Pospositive; Cytol. cytology; Bbrush; Nneedle aspiration.
* Salvage cytology.
Early esophageal carcinoma.

Not published.
The diagnostic accuracy of cytology is high. In many series, brush cytology yielded a positive rate of
diagnosis greater than 95% (see Table 403). When forceps biopsy and brush cytology are used in
combination, a positive tissue diagnosis can be achieved in almost 100% of esophageal cancers.
Therefore, brush cytology and biopsy are considered complementary rather than mutually exclusive.
The particular advantages of brush cytology when the luminal diameter through a tumor is only a few
millimeters have been emphasized. A disadvantage, albeit small, of cytology is the small percentage of
false-positive results that have occurred in virtually all published series.166(3242)
As mentioned before, aspiration cytology has also been shown to have a consistently high diagnostic
accuracy.148,149(3243)
Errors in Endoscopic Diagnosis

The endoscopic appearance of the esophagus may occasionally confuse the inexperienced examiner
with respect to recognition of cancer. A classic error is misinterpretation of a strongly adherent clot that
mimics an exophytic malignancy (Figure 4029). Also deceptive is the appearance of food residue
such as leguminous material impacted at a distal peptic stricture (Figure 4030).
(3244)Figure 4029. Tumor simulating a strongly adherent clot in a patient with an

esophageal laceration due to vomiting.
(3245)Figure 4030. Vegetable material or fruit skin tightly impacted in the esophagus.
Submucosal malignancy that mimics achalasia can be problematic. The clinical and radiologic
presentation of carcinoma of the gastric fundus and cardia encroaching on the gastroesophageal
junction can be indistinguishable from that of primary achalasia. In rare cases, the esophageal
manometric findings truly mimic primary achalasia.167,168(3246) Cancer should be suspected in

patients with so-called secondary achalasia when the duration of symptoms is less than 1 year, the
patient is older than 50 years, and excessive and rapid weight loss has occurred.169(3247) The ease
with which a small-caliber endoscope may be passed through such a lesion, because of the lack of
resistance to advancement of the instrument, may result in a serious oversight. Some authorities
recommend that a large-caliber dilator or large-diameter endoscope be passed through this segment
prior to pneumatic dilation to be certain that the sphincter area is not mechanically
constricted.170(3248) If a large bougie will not pass or if it meets more than minimum resistance, it is
likely that either a fibrotic stricture or an occult neoplasm is present and pneumatic dilation should not
be performed. Alternatively, one may perform detailed endosonographic study of the area of narrowing.
A carcinoma of the stomach with submucosal extension into the esophagus should be strongly
suspected when the narrowed segment appears to be in a position proximal to the diaphragm because
the narrow segment in true achalasia almost always straddles the diaphragm. Frequently, there are no
visible abnormalities at endoscopy and the usual small, superficial endoscopic biopsy specimens may
reveal only normal squamous or gastric mucosa. The use of a large biopsy forceps is one approach to
this problem; sampling is performed at the same site until tissue is obtained from the deeper layers of
the submucosa (see Figure 405).84(3249) A, Radiograph of carcinoma at the cardia, infiltrating under
normal squamous epithelium. B, Biopsy specimen obtained with large forceps in a patient with
pseudoachalasia. Photomicrograph demonstrates squamous cell mucosa (top) and adenocarcinoma
(bottom).
REFERENCES
1.
Stoller JL, Samer KJ, Toppin DI, et al. Carcinoma of the esophagus: A new proposal for the
evaluation of treatment. Can J Surg 1977;20:4549.
2. Atkinson M, Ferguson R, Ogilvie AL. Management of malignant dysphagia by intubation at
endoscopy. J R Soc Med 1979;72:8947.
3. O'Sullivan GC, DeMeester TR, Smith RB, et al. Twenty-four hour pH monitoring of
esophageal function. Arch Surg 1981;116:58190.
4. Goldschmid S, Boyce HW Jr, Brown JI, et al. A new objective measurement of esophageal
lumen patency. Am J Gastroenterol 1989;84:12558.
5. Syrjanen K, Pyrhonen S, Aukee S, et al. Squamous cell papilloma of the oesophagus: A
tumour probably caused by human papilloma virus (HPV). Diagn Histopathol 1982;5:2916.
6. Correa P. Precursors of gastric and esophageal cancer. Cancer 1982;50:255465.
7.
Tytgat GNJ, Mathus-Vliegen EMH. Cancer screening. In Hodgson HJF, Bloom SR (eds).
Gastrointestinal and Hepatobiliary Cancer. London: Chapman & Hall, 1983;15787.
8.
Tytgat GNJ, Mathus-Vliegen EMH, Offerhaus J. Value of endoscopy in the surveillance of
high-risk groups for gastrointestinal cancer. In Sherlock P, Morson BC, Barbara L, Veronesi V
(eds). Precancerous Lesions of the Gastrointestinal Tract. New York: Raven Press,
1983;30518.
9.
Wynder EL, Hultberg S, Jacobsson F, et al. Environmental factors in cancer of upper
alimentary tract. Swedish study with special reference to Plummer-Vinson (Patterson-Kelly)
syndrome. Cancer 1957;10:47087.
10.
Howel-Evans W, McConnell RB, Clarke CA, Sheppard PM. Carcinoma of the oesophagus
with keratosis palmaris et plantaris (tylosis). A study of two families. Q J Med 1958;27:41329.
11. Lansing PB, Ferrante WA, Ochsner JL. Carcinoma of the esophagus at the site of lye
stricture. Am J Surg 1969;118:10811.
12.
Williams JL. Carcinoma of the oesophagus as a complication of achalasia of the cardia.
Thorax 1956;11:26874.
13.
Joske RA, Benedict EB. The role of benign esophageal obstruction in the development of
carcinoma of the esophagus. Gastroenterology 1959;36:74955.
14. Just-Viera JO, Haight C. Achalasia and carcinoma of the esophagus. Surg Gynecol Obstet
1969;128:108195.
15. Macias Rodriguez MA, Soria de la Cruz MJ, Iglesias Arrabal M, Martin Herrera L. Esophageal
carcinoma after the endoscopic sclerotherapy of varices. Rev Esp Enferm Dig 1992;82:436.
16. Nakamura R, Watanabe M, Sugimura Y, et al. Early esophageal cancer discovered 6 months
after endoscopic injection sclerotherapy of esophageal varices. Gan No Rinsho
1988;34:11904.
17. Kokudo N, Sanjo K, Umekita N, et al. Squamous cell carcinoma after endoscopic injection
sclerotherapy for esophageal varices. Am J Gastroenterol 1990;85:8614.
18. Bochna GS, Harty RF, Harned RK, Markin RS. Development of squamous cell carcinoma of
the esophagus after endoscopic variceal sclerotherapy. Am J Gastroenterol 1988;83:5648.
19. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF Jr. Rising incidence of adenocarcinoma of the
esophagus and gastric cardia. JAMA 1991;265:12879.
20. Pera M, Cameron AJ, Trastek VF, et al. Increasing evidence of adenocarcinoma of the
esophagus and esophagogastric junction. Gastroenterology 1993;104:5103.
21. Naef AP, Savary M, Ozzello L. Columnar-lined lower esophagus: An acquired lesion with
malignant predisposition. Report on 140 cases of Barrett's esophagus with 12
adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:82635.
22. Bozymski EM, Herlihy KJ, Orlando RC. Barrett's esophagus. Ann Intern Med 1982;97:1037.
23. Spechler SJ. Barrett's esophagus: What's new and what to do?. Am J Gastroenterol
1989;84:2203.
24.
Wesdorp ICE, Bartelsman J, Schipper MEI, et al. Malignancy and premalignancy in Barrett's
esophagus: A clinical endoscopical and histological study. Acta Endosc 1981;11:31726.
25. Tytgat GNJ, Hameeteman W. The neoplastic potential of columnar-lined esophagus. World J
Surg 1992;16:30812.
26. Reid BJ. Barrett's esophagus and esophageal adenocarcinoma. Gastroenterol Clin North Am
1991;20:81734.
27. Robertson CS, Mayberry JF, Nicholson DA, et al. Value of endoscopic surveillance in the
detection of neoplastic change in Barrett's oesophagus. Br J Surg 1988;75:7603.
28. Hameeteman W, Tytgat GNJ, Houthaff HJ, van den Tweel JG. Barrett's esophagus:
Development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:124956.
29. Spechler SJ, Robbins AH, Rubins HB, et al. Adenocarcinoma and Barrett's esophagus. An
overrated risk?. Gastroenterology 1984;87:92733.
30.
Sprung DJ, Ellis FH, Gibb SP. Incidence of adenocarcinoma in Barrett's esophagus. Am J
Gastroenterol 1984;79:817A.
31.
Cameron AJ, Zinsmeister AR, Ballard DJ, et al. A population-based study of Barrett's
esophagus: Comparison of clinically diagnosed prevalence versus autopsy determined
prevalence. Gastroenterology 1988;94:A57.
32.
Sampliner RE. Does Barrett's esophagus need to be detected? [Editorial]. Arch Intern Med
1986;96:438A.
33. Achkar E, Carey W. The cost of surveillance for adenocarcinoma complicating Barrett's
esophagus. Am J Gastroenterol 1988;83:2914.
34. Van Der Veen AH, Dees J, Blankensteijn JD, Van Blankenstein M. Adenocarcinoma in
Barrett's oesophagus: An overrated risk. Gut 1989;30:148.
35. Polepalle SC, McCallum RW. Barrett's esophagus. Current assessment and future
perspectives. Gastroenterol Clin North Am 1990;19:73344.
36. Brandt LJ, Kauvar DR. Laser-induced transient regression of Barrett's epithelium. Gastrointest
Endosc 1992;38:61922.
37. Sampliner RE, Hixson LJ, Fennerty MB, Garewal HS. Regression of Barrett's esophagus by
laser ablation in an anacid environment. Dig Dis Sci 1993;38:3658.
38.
. The Standards of Training and Practice Committee, American Society for Gastrointestinal
Endoscopy: The role of endoscopy in the surveillance of premalignant conditions of the upper
gastrointestinal tract. Gastrointest Endosc 1988;34:18s-20s.
39. Spechler SJ. Endoscopic surveillance for patients with Barrett esophagus: Does the cancer
risk justify the practice?. Ann Intern Med 1987;106:9024.

40. Streitz JM Jr, Andrews CW Jr, Ellis FH Jr. Endoscopic surveillance of Barrett's esophagus.
Does it help?. J Thorac Cardiovasc Surg 1993;105:3837.
41. Peters JH, Clark GW, Ireland AP, et al. Outcome of adenocarcinoma arising in Barrett's
esophagus in endoscopically surveyed and nonsurveyed patients. J Thorac Cardiovasc Surg
1994;108:81321; discussion 8212.
42. Levine DS, Haggitt RC, Blount PL, et al. An endoscopic biopsy protocol can differentiate
high-grade dysplasia from early adenocarcinoma in Barrett's esophagus. Gastroenterology
1993;105:4050.
43. Reid BJ, Weinstein WM, Lewin KJ, et al. Endoscopic biopsy can detect high-grade dysplasia
or early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic
lesions. Gastroenterology 1988;94:8190.
44. Petras RE, Sivak MV Jr, Rice TW. Barrett's esophagus. A review of the pathologist's role in
diagnosis and management. Pathol Annu 1991;26:132.
45. Reid BJ, Blount PL, Rubin CE, et al. Flow-cytometric and histological progression to
malignancy in Barrett's esophagus: Prospective endoscopic surveillance of a cohort.
46. Fennerty MB, Sampliner RE, Way D, et al. Discordance between flow cytometric
abnormalities and dysplasia in Barrett's esophagus. Gastroenterology 1989;97:81520.
47. Ramel S, Reid BJ, Sanchez CA, et al. Evaluation of p53 protein expression in Barrett's
esophagus by two-parameter flow cytometry. Gastroenterology 1992;102:12208.
48. Garewal HS, Sampliner R, Gerner E, et al. Ornithine decarboxylase activity in Barrett's
esophagus: A potential marker for dysplasia. Gastroenterology 1988;94:81921.
49. Falk GW, Catalano MF, Sivak MV Jr, et al. Endosonography in the evaluation of patients with
Barrett's esophagus and high-grade dysplasia. Gastrointest Endosc 1994;40:20712.
50. Canto MIF, Setrakian S, Petras RE, et al. Methylene blue selectively stains intestinal
metaplasia in Barrett's esophagus. Gastrointest Endosc 1996;44:17.
51. Miros M, Kerlin P, Walker N. Only patients with dysplasia progress to adenocarcinoma in
Barrett's oesophagus. Gut 1991;32:14416.
52. Reid BJ, Haggitt RC, Rubin CE, Rabinovitch PS. Barrett's esophagus. Correlation between
flow cytometry and histology in detection of patients at risk for adenocarcinoma.
53. Rice TW, Falk GW, Achkar E, Petras RE. Surgical management of high-grade dysplasia in
Barrett's esophagus. Am J Gastroenterol 1993;88:18326.
54. Rusch VW, Levine DS, Haggitt R, Reid BJ. The management of high-grade dysplasia and
early cancer in Barrett's esophagus. A multidisciplinary problem. Cancer 1994;74:12259.
55. Williamson WA, Ellis FH Jr, Gibb SP, et al. Barrett's esophagus. Prevalence and incidence of
adenocarcinoma. Arch Intern Med 1991;151:22126.
56. Altorki NK, Sunagawa M, Little AG, Skinner DB. High-grade dysplasia in the columnar-lined
esophagus. Am J Surg 1991;161:979.
57. Pera M, Trastek VF, Carpenter HA, et al. Barrett's esophagus with high-grade dysplasia: An
indication for esophagectomy?. Ann Thorac Surg 1992;54:199204.
58. Overholt B, Panjepour M, Teftellar E, Rose M. Photodynamic therapy for treatment of early
59. Kim JH, Lee MS, Cho SW, Shim CS. Primary adenoid cystic carcinoma of the esophagus: A
case report. Endoscopy 1991;23:3841.
60. Sasajima K, Watanabe M, Takubo K, et al. Mucoepidermoid carcinoma of the esophagus:
Report of two cases and review of the literature. Endoscopy 1990;22:1403.
61. Briggs JC, Ibrahim NBN. Oat cell carcinoma of the oesophagus: A clinicopathological study of
23 cases. Histopathology 1983;7:26177.
62. Rosenthal SN, Lemkin JA. Multiple small cell carcinomas of the esophagus. Cancer
1983;51:19446.
63. Sasajima K, Hayashi N, Yamashita K, et al. Oat cell carcinoma of the esophagus with multiple
differentiation. J Clin Gastroenterol 1988;10:66771.

64. Seremetis MG, Lyons WS, deGuzman VC, et al. Leiomyomata of the esophagus: An analysis
of 838 cases. Cancer 1976;38:216677.
65. Peiser J, Ovnat A, Herz A, et al. Lipoma of the esophagus. Isr J Med Sci 1984;20:106870.
66. Zschiedrich M, Neuhaus P. Pedunculated giant lipoma of the esophagus. Am J Gastroenterol
1990;85:16146.
67. Patel RM, DeSota-LaPaix F, Sika JV, et al. Granular cell tumor of the esophagus. Am J
68. Hanel K, Talley NA, Hunt DR. Hemangioma of the esophagus: An unusual cause of upper
gastrointestinal bleeding. Dig Dis Sci 1981;26:25763.
69. Domergue J, Rouanet P, Joyeux H, et al. Muscular tumors of the esophagus. Apropos of 9
cases and review of the literature. J Chir (Paris) 1986;123:5558.
70. Choh JH, Khazei AH, Ihm HJ. Leiomyosarcoma of the esophagus: Report of a case and
review of the literature. J Surg Oncol 1986;32:2236.
71. Shiraishi M, Takahashi T, Yamashiro M, et al. A report of leiomyosarcoma of the esophagus.
Nippon Ronen Igakkai Zasshi 1995;32:28691.
72. Koga H, Iida M, Suekane H, et al. Rapidly growing esophageal leiomyosarcoma: Case report
and review of the literature. Abdom Imaging 1995;20:159.
73. Aimoto T, Sasajima K, Kyono S, et al. Leiomyosarcoma of the esophagus: Report of a case
and preoperative evaluation by CT scan, endoscopic ultrasonography and angiography.
Gastroenterol Jpn 1992;27:7739.
74. Partyka EK, Sanowski RA, Kozarek RA. Endoscopic diagnosis of a giant esophageal
leiomyosarcoma. Am J Gastroenterol 1981;75:1324.
75. Cantero D, Tada M, Okita K. Esophageal leiomyoblastoma: The role of endoscopic
procedures in the pre-operative diagnosis. Gastrointest Endosc 1993;39:56871.
76. Bolondi L, De Giorgio R, Santi V, et al. Primary non-Hodgkin's T-cell lymphoma of the
esophagus. A case with peculiar endoscopic ultrasonographic pattern. Dig Dis Sci
1990;35:142630.
77. Doki T, Hamada S, Murayama H, et al. Primary malignant lymphoma of the esophagus. A
case report. Endoscopy 1984;16:18992.
78. Mengoli M, Marchi M, Rota E, et al. Primary non-Hodgkin's lymphoma of the esophagus. Am
79. Nagrani M, Lavigne BC, Siskind BN, et al. Primary non-Hodgkin's lymphoma of the
esophagus. Arch Intern Med 1989;149:1935.
80. Bernal A, del Junco GW. Endoscopic and pathologic features of esophageal lymphoma: A
report of four cases in patients with acquired immune deficiency syndrome. Gastrointest
Endosc 1986;32:969.
81. Butler ML, Van Heertum RL, Teplick SK. Metastatic malignant melanoma of the esophagus: A
case report. Gastroenterology 1975;69:13347.
82. Aldovini D, Detassis C, Piscioli F. Primary malignant melanoma of the esophagus. Brush
cytology and histogenesis. Acta Cytol 1983;27:658.
83. Lautz HU, Schmidt FW, Cullen P. Primary malignant melanoma of the esophagus. Endoscopy
1986;18:2402.
84. Loftus BM, Broe P, Johnston O, Leader M. Primary amelanotic malignant melanoma of the
oesophagusA case report. Ir J Med Sci 1994;163:12831.
85. Milman PJ. Primary malignant melanoma of the esophagus. Gastrointest Endosc
1987;33:367.
86. Roesch W, Rohner HG. Primary malignant melanoma of the oesophagus. Endoscopy
1984;16:1868.
87. Woolery WA, Tripodis SP, Chacko DC. Endoscopic Nd:YAG laser and radiation therapy for
primary malignant melanoma of the esophagus. J Am Osteopath Assoc 1990;90:5436.
88. Feller SC, Weaver GA. Heterotopic gastric mucosa in the upper esophagus. Gastroenterology
1986;90:2578.
89. Auld RM, Lukas WM, Bordin GM. Heterotopic sebaceous glands in the esophagus.
90. Marcial MA, Villafana M. Esophageal ectopic sebaceous glands: Endoscopic and histologic
findings. Gastrointest Endosc 1994;40:6302.
91.
Stout AP. Tumors of the Soft Tissues. Sect II, fasc 5. Washington, DC: Armed Forces
Institute of Pathology. 1953.
92. Siddins M, Cade R. Fibrovascular polyp of the oesophagus. Aust N Z J Surg 1991;61:23740.
93. Franzin G, Musola R, Zamboni G, et al. Squamous papillomas of the esophagus. Gastrointest
Endosc 1983;29:1046.
94. Sablich R, Benedetti G, Bignucolo S, Serraino D. Squamous cell papilloma of the esophagus.
Report on 35 endoscopic cases. Endoscopy 1988;20:57.
95. Zeabart LE, Fabian J, Nord HJ. Squamous papilloma of the esophagus: A report of 3 cases.
96. Koerfgen HP, Husemann B, Giedl J, et al. Verrucous carcinoma of the esophagus.
Endoscopy 1988;20:3269.
97. Javdan P, Pitman ER. Squamous papilloma of esophagus. Dig Dis Sci 1984;29:31720.
98. Ravry MJ. Endoscopic resection of squamous papilloma of the esophagus. Am J
99. Shah AN. Endoscopic removal of squamous cell papilloma of esophagus: A case report and
review of literature. Surg Gastroenterol 1984;3:236.
100.
Totten RS, Stout AP, Humphreys GH, et al. Benign tumors and cysts of the esophagus. J
Thorac Surg 1953;25:60622.
101. Idenburg FJ, Akkermans LM, Smout AJ, et al. Leiomyoma of the distal oesophagus mimicking
achalasia. Neth J Surg 1991;43:7981.
102. Bardini R, Segalin A, Asolati M, et al. Thoracoscopic removal of benign tumours of the
oesophagus. Endosc Surg Allied Technol 1993;1:2779.
103. Eda Y, Asaki S, Yamagata L, et al. Endoscopic treatment for submucosal tumors of the
esophagus: Studies in 25 patients. Gastroenterol Jpn 1990;25:4116.
104. Benedetti G, Sablich R, Bonea M, Mariuz S. Fiberoptic endoscopic resection of symptomatic
leiomyoma of the upper esophagus. Case report. Acta Chir Scand 1990;156:8078.
105. Andrade J de S, Bambirra EA, de Oliveira CA, et al. Granular cell tumor of the esophagus: A
study of seven cases diagnosed by histologic examination of endoscopic biopsies. South Med
J 1987;80:8524.
106. Flood CA. Granular cell tumor of the esophagus. Gastrointest Endosc 1981;27:2257.
107. Tada S, Iida M, Yao T, et al. Granular cell tumor of the esophagus: Endoscopic
ultrasonographic demonstration and endoscopic removal. Am J Gastroenterol
1990;85:150711.
108. Brady PG, Nord HJ, Connar RG. Granular cell tumor of the esophagus: Natural history,
diagnosis, and therapy. Dig Dis Sci 1988;33:132933.
109. Orlowska J, Pachlewski J, Gugulski A, Butruk E. A conservative approach to granular cell
tumors of the esophagus: Four case reports and literature review. Am J Gastroenterol
1993;88:3115.
110. Giacobbe A, Facciorusso D, Conoscitore P, et al. Granular cell tumor of the esophagus. Am J
111. Gertsch P, Mosimann R. A rare tumor of the esophagus: The granular cell myoblastoma.
Report of a case and review of the literature. Endoscopy 1980;12:2459.
112. Sarma DP, Rodriguez FH Jr, Deiparine EM, Weilbaecher TG. Symptomatic granular cell
tumor of the esophagus. J Surg Oncol 1986;33:2469.
113. Cohle SD, McKechnie JC, Truong L, Jurco S III. Granular cell tumor of the esophagus. Report
of three cases. Am J Gastroenterol 1981;75:4315.
114. Farinati F, Sturniolo GC, Cecchetto A, et al. Granular cellmyoblastoma of the esophagus:
Endoscopic diagnosis and therapy. Gastrointest Endosc 1985;31:223.
115. Walters NA, Coral A. Fibrovascular polyp of the oesophagus: The appearances on computed
tomography. Br J Radiol 1988;61:6413.

116. Naveau S, Bedossa P, Mallet L, et al. Successful ablation of a large fibrovascular polyp of the
esophagus by endoscopic Nd:YAG laser therapy. Gastrointest Endosc 1989;35:2546.
117.
Goldstein HM, Zornoza J. Association of squamous cell carcinoma of the head and neck with
cancer of the esophagus. AJR 1978;131;7914.
118. Atkins JP Jr, Keane WM, Young KA, Rowe LD. Value of panendoscopy in determination of
second primary cancer. A study of 451 cases of head and neck cancer. Arch Otolaryngol
1984;110:5334.
119. Grossman TW, Toohill RJ, Lehman RH, et al. Role of esophagoscopy in the evaluation of
patients with head and neck carcinoma. Ann Otol Rhinol Laryngol 1983;92:36972.
120. Leipzig B, Zellmer JE, Klug D. The role of endoscopy in evaluating patients with head and
neck cancer. A multi-institutional prospective study. Arch Otolaryngol 1985;111:58994.
121. Okumura T, Aruga H, Inohara H, et al. Endoscopic examination of the upper gastrointestinal
tract for the presence of second primary cancers in head and neck cancer patients. Acta
Otolaryngol Suppl (Stockh) 1993;501:1036.
122. Shiozaki H, Tahara H, Kobayashi K, et al. Endoscopic screening of early esophageal cancer
with the Lugol dye method in patients with head and neck cancers. Cancer 1990;66:206871.
123. Ina H, Shibuya H, Ohashi I, Kitagawa M. The frequency of a concomitant early esophageal
cancer in male patients with oral and oropharyngeal cancer. Screening results using Lugol
dye endoscopy. Cancer 1994;73:203841.
124. Contini S, Consigli GF, Di Lecce F, et al. Vital staining of oesophagus in patients with head
and neck cancer: Still a worthwhile procedure. Ital J Gastroenterol 1991;23:58.
125. Hix WR, Wilson WR. Toluidine blue staining of the esophagus. A useful adjunct in the
panendoscopic evaluation of patients with squamous cell carcinoma of the head and neck.
Arch Otolaryngol Head Neck Surg 1987;113:8645.
126. McDonald GB, Brand DL, Thorning DR. Multiple adenomatous neoplasms arising in
columnar-lined (Barrett's) esophagus. Gastroenterology 1977;72:131721.
127. Wang GQ. Endoscopic diagnosis of early oesophageal carcinoma. J R Soc Med
1981;74:5023.
128. Yang G, Huang H, Qui S, et al. Endoscopic diagnosis of 115 cases of early esophageal
carcinoma. Endoscopy 1982;14:15761.
129. Guanrei Y, He H, Sungliang Q, Yuming C. Endoscopic diagnosis of 115 cases of early
esophageal carcinoma. Endoscopy 1982;14:15761.
130. Guanrei Y, Songliang Q, He H, Guizen F. Natural history of early esophageal squamous
carcinoma and early adenocarcinoma of the gastric cardia in the People's Republic of China.
131.
Ide H, Endo M, Kinoshika Y, et al. Clinicopathological aspects of superficial esophageal
cancer. Chir Gastroenterol 1976;10:916.
132. Haruma K, Tokutomi T, Tsuda T, et al. Superficial esophageal carcinoma: A report of 27
cases in Japan. Am J Gastroenterol 1991;86:17238.
133. Sabik JF, Rice TW, Goldblum JR, et al. Superficial esophageal carcinoma. Ann Thorac Surg
1995;60:896901.
134. Tytgat GN, Hameeteman W, Onstenk R, Schotborg R. The spectrum of columnar-lined
esophagusBarrett's esophagus. Endoscopy 1989;21:17785.
135. Endo M. Endoscopic resection as local treatment of mucosal cancer of the esophagus.
Endoscopy 1993;25:6724.
136. Inoue H, Endo M, Takeshita K, et al. Endoscopic resection of early-stage esophageal cancer.
137. Spinelli P, Mancini A, Dal Fante M. Endoscopic treatment of gastrointestinal tumors:
Indications and results of laser photocoagulation and photodynamic therapy. Semin Surg
Oncol 1995;11:30718.
138. Hagiwara A, Takahashi T, Kojima O, et al. Endoscopic local injection of a new drug-delivery
format of peplomycin for superficial esophageal cancer: A pilot study. Gastroenterology
1993;104:103743.
Soga J, Tanaka O, Sasaki K, et al. Superficial spreading carcinoma of the esophagus. Cancer
1981;50:16415.
140.
Savary M, Miller G. The Esophagus. Handbook and Atlas of Endoscopy. Solothurn, Switz:
Gassman. 1978.
141.
Hermanek P, Sabin LH. TNM Classification of Malignant Tumours. 4th ed. Heidelberg, Berlin,
New York: Springer-Verlag. 1987.
142. Kawai K, Takemoto T, Suzuki S, et al. Proposed nomenclature and classification of the
dye-spraying techniques in endoscopy. Endoscopy 1979;11:235.
143. Yamakawa T, von Hofe FC, Kagan R, et al. The use of tolonium in the diagnosis of malignant
gastric lesions. Arch Surg 1972;104:7737.
144. Misumi A, Harada K, Murakami A, et al. Role of Lugol dye endoscopy in the diagnosis of early
esophageal cancer. Endoscopy 1990;22:126.
145. Seitz JF, Monges G, Navarro P, et al. Endoscopic detection of dysplasia and subclinical
cancer of the esophagus. Results of a prospective study using toluidine blue vital staining in
100 patients with alcoholism and smoking. Gastroenterol Clin Biol 1990;14:1521.
146. Barkin JS, Taub S, Rogers AI. The safety of combined endoscopy, biopsy and dilation in
esophageal strictures. Am J Gastroenterol 1981;76:236.
147. Graham DY, Spjut HJ. Salvage cytology. A new alternative fiberoptic technique. Gastrointest
Endosc 1979;25:1379.
148. Kochhar R, Rajwanshi A, Malik AK, et al. Endoscopic fine needle aspiration biopsy of
gastroesophageal malignancies. Gastrointest Endosc 1988;34:3213.
149. Zarger SA, Khuroo MS, Mahajan R, et al. Endoscopic fine needle aspiration cytology in the
diagnosis of gastrooesophageal and colorectal malignancies. Gut 1991;32:7458.
150. Ingoldby CJ, Mason MK, Hall RI. Endoscopic needle aspiration cytology: A new method for
151.
Nakamura Y, Arimori M, Kumagai Y. Examination with esophagoscopes: Utilization of the
fiberscope. Stomach Intestine 1968;3:13617.
152. Kobayashi S, Prolla JC, Winans CS, et al. Improved endoscopic diagnosis of
gastroesophageal malignancy. Combined use of direct vision brushing cytology and biopsy.
JAMA 1970;212:20869.
153. Prolla JC. Cancer of the gastrointestinal tract. I. EsophagusHistopathology and cytology in
detection. JAMA 1973;226:15546.
154. Bruni HC, Nelson RS. Carcinoma of the esophagus and cardia: Diagnostic evaluation in 113
cases. J Thorac Cardiovasc Surg 1975;70:36770.
155.
Seifert E, Atay Z. Maligne Tumoren des Gastrointestinal-traktes. Fortschr Endoskop.
Stuttgart: Thieme. 1975.
156. Winawer SJ, Sherlock P, Belladonna JA, et al. Endoscopic brush cytology in esophageal
cancer. JAMA 1975;232:1358.
157. Hishon S, Lovell D, Gummer JW, et al. Cytology in the diagnosis of oesophageal cancer.
Lancet 1976;1:2967.
158. Winawer SJ, Melamed M, Sherlock P. Potential of endoscopy, biopsy and cytology in the
diagnosis and management of patients with cancer. Clin Gastroenterol 1976;5:57595.
159. Witzel L, Halter F, Gretillat PA, et al. Evaluation of specific value of endoscopic biopsies and
160. Prolla JC, Reilly W, Kirsner JB, Cockerham L. Direct-vision endoscopic cytology and biopsy in
the diagnosis of esophageal and gastric tumors: Current experience. Acta Cytol
1977;21:399402.
161. Eastman MC, Gear MW, Nicol A. An assessment of the accuracy of modern endoscopic
diagnosis of oesophageal stricture. Br J Surg 1978;65:1825.
162. Gutz HJ, Wildner GP. Die diagnostische Sicherheit von Endoskopie und Biopsie bei
stenosierenden Prozessen in Osophagus und Kardia. Dtsch Z Verdau Stoffwechselkr
1978;38:479.
139.
163. Mortensen NJ, MacKenzie EF. Accuracy of oesophageal brush cytology: Results of a
prospective study and multicentre slide exchange. Br J Surg 1981;68:5135.
164. Graham DY, Schwartz JT, Cain GD, Gyorkey F. Prospective evaluation of biopsy number in
the diagnosis of esophageal and gastric carcinoma. Gastroenterology 1982;82:22831.
165. Cusso X, Mones-Xiol J, Villardel E. Endoscopic cytology of cancer of the esophagus and
cardia: A long-term evaluation. Gastrointest Endosc 1989;35:3213.
166. McCallum RW. Esophageal achalasia secondary to gastric carcinoma. Am J Gastroenterol
1979;71:249.
167.
Tytgat GNJ, Bartelsman JFWM. Motility disorders of the oesophagus. In Misiewicz JJ,
Pounder RE, VeDables CW (eds). Diseases of the Gut and Pancreas. London: Grant
McIntyre. 1985.
168. Tulman AB, Boyce HW Jr. Complications of esophageal dilation and guidelines for their
prevention. Gastrointest Endosc 1981;27:22934.
169. Kahrilas PJ, Kishk SM, Helm JF, et al. Comparison of pseudoachalasia and achalasia. Am J
Med 1987;82:43946.
Chapter 41 Endoscopic Ultrasonography of the Esophagus

(3250)
(3251)
T. LOK TIO, M.D., PH.D.
The first attempt to adapt ultrasound (US) to medical purposes occurred in 1947 and was based on
experience gathered during World War II with the development of technology for the detection of
submarines.1(3252) In 1950, the American physician J. J. Wild2(3253) described US imaging of
gastric cancers in resection specimens. The use of a transrectal ultrasonic probe was first described
by Wild and Reid3,4(3254) in 1956, a report that proved to be seminal to the development of
endosonography over the ensuing 2 decades. Transrectal and transurethral ultrasonography for the
evaluation of the prostate gland was performed, respectively, by Watanabe et al.5(3255) in Japan
(1968) and Holm and Northeveld6(3256) in Denmark (1974). Bom et al.7(3257) in Holland described
an ultrasonic intracardiac scanner in 1970. Transesophageal ultrasonic cardiac imaging was performed
by Hisanaga and Hisanaga8(3258) in Japan in 1978.
Lutz and Rsch9(3259) in Germany in 1976 reported the use of an A-scan ultrasonography probe that
could be passed through the accessory channel of a therapeutic endoscope. DiMagno et al.10(3260)
and Strohm et al.11(3261) described transgastrointestinal endoscopic ultrasonography (EUS) using
prototype instruments with endoscopic and B-mode US capabilitiesthese instruments were the
forerunners of present-day endoscopic echoendoscopes. Thereafter, rapidly occurring improvements
in instruments as well as remarkable advances in knowledge and experience led to the widespread
clinical application of EUS in many countries throughout the world.1229(3262),3035(3263) The first
comprehensive atlas that correlated EUS images with other imaging modalities including histology was
published by myself and Tytgat13(3264) in 1986.
The esophagus is a tubelike organ lying behind the sternum and directly adjacent to the
tracheobronchial tree and lungs. Because of its relationship to these structures, which are either made
of bone or contain air, the esophagus cannot be assessed by transcorporeal US. However,
transesophageal US allows a direct approach to this target organ that avoids obstructing gas and bone
(see also Chapter 11: Principles of Endoscopic Ultrasonography). This chapter reviews the role of EUS
in the diagnosis or staging, or both, of esophageal and periesophageal diseases.
Instruments
Three different types of instruments are available for transesophageal EUS: echoendoscopes,
echoprobes, and catheter probes.
Echoendoscopes
An echoendoscope is an endoscopic instrument with a US device incorporated at its distal tip. In
general, the optical component of an echoendoscope provides a view field that is oriented lateral or
oblique to the long axis of the insertion tube; fields of view are relatively small compared with those
provided by standard diagnostic/therapeutic endoscopes. The technical specifications of currently
available echoendoscopes are summarized in Table 411.
TABLE 411
Characteristics of Currently Available Echoendoscopes and Catheter Echoprobes
TECHNICAL
DATA
OLYMPUS EU-M20
MECHANICAL G*
ECHOENDOSCOPE
Length (mm)
Shaft diameter (mm)
Tip diameter (mm)
Frequency (MHz)
Sector image
Focal distance (mm)
1055
11.7
13.2
7.5/12
360/180
30/25
Depth of
100
penetration (mm)
Direction
Forward
of view
oblique 45
Axial
0.2/0.12
resolution (mm)
Biopsy channel
Yes
Color Doppler
No
(duplex)
* GEcho-gastroscope.
J = Echo-duodenoscope.
OLYMPUS BU-M20
MECHANICAL J
ECHOENDOSCOPE
PENTAXHITACHI CURVE
ARRAY
ECHOENDOSCOPE
1250
11.2
12.5
7.5/12
360/180
30/25
1250
10.8
12
7.5
180
Variable
100
Lateral
0.2/0.12
Yes
No
OLYMPUS
MECHANICAL
CATHETER
ECHOPROBE
360
2030
100
Forward
oblique 80
?
2030
Endoscopic
guided view
Yes
Yes
30 mm with 7.5 MHz and 25 mm with 12 MHz.
The first prototype echoendoscope manufactured by Olympus Optical, Ltd.(Tokyo, Japan), had a rigid
tip, 45 mm in length, attached at the end of a side-viewing endoscope.11(3265) Other early prototypes,
never offered commercially, are the Pentax-Siemens echoendoscope, which had a rigid metal tip, and
the Toshiba-Pentax prototype and ACMI-SRI echoendoscopes, both of which employed linear array
transducer scanning systems.10,15(3266)
Currently available echoendoscopes may be classified into two types based on the orientation of the
scanning plane relative to the long axis of the insertion tube: radial sector and linear array. The former
category includes instruments in which the scanning plane is oriented perpendicular to the axis; the
latter comprises instruments with a scanning plane oriented parallel to the axis. There are fundamental
differences in the technology used to generate US images with these two types of instruments. Radial
sector scanning instruments employ a rotating acoustic mirror, whereas currently available linear
display echoendoscopes utilize an array of transducers. Color Doppler US for demonstration of blood
flow within vessels is possible only with the use of a transducer array. The sector of radial scanning
instruments can be switched from a full 360-degree image to 180 degrees or 90 degrees. The EU-M20
instrument offered by Olympus (Olympus America, Inc. Melville, NY) is an example of a radial sector
scanning instrument (Figure 411; see also Table 411). Use of US components of relatively small
diameter also allows for the incorporation into this instrument of an accessory channel with elevator. It
is possible, therefore, to use the EU-M20 echoendoscope for EUS-guided needle puncture for the
purpose of obtaining cytologic specimens. The Pentax-Hitachi instrument (FG-36-UX; Pentax Precision
Instrument Co. [Orangeburg, NY]) (see Table 411), an example of a curve phased array
echoendoscope, has a transducer array at the distal tip, forward-oblique viewing optics, and a US scan
field of 135 degrees; this instrument provides continuous visualization of a needle during EUS-guided
fine-needle aspiration (FNA) for cytologic specimens (see also Chapter 21: Endoscopic
Ultrasonography with Linear Array Instruments: Anatomy and EUS-Guided Fine-Needle Aspiration).
(3267)Figure 411. An Olympus echoendoscope EU-M20 with a small echoprobe (e)

attached at the tip of a lateral-viewing duodenoscope. A modified sclerosing needle (n) has
been passed through the accessory channel. Note the elevator (el) for maneuvering the needle.
Echoprobes
Echoprobes do not have optical components and are passed over a guidewire that has been placed
endoscopically or fluoroscopically. Elimination of the optical elements reduces the diameter of the
probe, thereby making this type of instrument especially suitable for imaging malignant tumors that
markedly narrow the esophageal lumen. The first prototype nonoptical ("blind") flexible echoprobe, built
by Aloka, was a radial scanning device with a diameter of 10 mm.22(3268) A more recently developed
Olympus prototype incorporates the Aloka scanning device into a small-diameter ("pediatric")
gastroscope (GIF-XP20).32(3269)
Catheter Echoprobes
Catheter echoprobes, available since 1988, are small-diameter US devices that can be introduced
through the accessory channel of a large-caliber endoscope (Figure 412).23,33,36(3270) As with
echoendoscopes, catheter echo-probes are available that offer scanning in radial sector and linear
planes. The Olympus catheter echoprobes (UM 2R/3R), outer diameter 2.5 mm, provide a radial sector
scan at frequencies of 12 MHz (UM-2R) and 20 MHz (UM-3R). A linear scanning catheter echoprobe
(Fujinon, Inc.) is also available.34(3271)
(3272)Figure 412. An Olympus catheter-like miniature echoprobe (e) passing through the
accessory channel of a large-caliber duodenoscope. ooptics of duodenoscope.
Technique
The technique for endosonography of the esophagus is similar to that for esophagogastroscopy and
includes topical oropharyngeal anesthesia and conscious sedation. Patients may experience
discomfort during passage of the rigid distal end of the echoendoscope, expansion of the balloon
attached to the transducer, or filling of the gastric lumen with water. The intravenous administration of
sedative drugs is therefore necessary. In general, endoscopic examination of the esophagus and
stomach with a standard forward-viewing endoscope is performed before EUS because the oblique- or
lateral-viewing optics of echoendoscopes do not permit adequate visualization of the esophagus. The
distance between the lesion and the incisor teeth should be accurately measured. When a stenotic
lesion is present, careful study at endoscopy of the route through the distorted lumen will reduce the
chance of a perforation.
The echoendoscope should be introduced into the stomach whenever possible. Thereafter, it is
gradually withdrawn into the esophagus to the target of interest, using information gathered at the
previous gastroscopy, mainly the distance to the target from the incisor teeth. Then, with further
withdrawal, the longitudinal extent (distance between the cranial and the caudal margins of the lesion)
and the maximum depth (distance between the mucosa and the deepest infiltration) should be carefully
measured. During withdrawal, the balloon over the transducer is filled with water so that its outer
surface makes light contact with the lesion. Excessive distention of the balloon may distort the
sonographic anatomy of the normal esophageal wall or the lesion, especially small, early-stage tumors.
Because the diameter of the lumen within a malignant stenosis is usually not constant, it may be
necessary to adjust the volume of water within the balloon in order to maintain satisfactory acoustic
coupling. Scanning should be performed in 1-cm contiguous sections, working from the distal to the
proximal edge of the lesion. Before pulling the instrument back through the cricopharyngeus and
pharynx, the water must be removed from the balloon. Photographic documentation of "frozen" images
may be obtained by means of a variety of "hard-copy" devices (e.g., thermal printer). However, EUS is
a real-time rather than a static imaging method and videotaping is therefore the most useful method of
documentation for later study and review. Display systems for EUS have the capability to label lesions,
which facilitates interpretation of images.
EUS Interpretation
The unique advantage of EUS as a technique for imaging the gastrointestinal tract is the ability to place
the imaging device in direct contact with a lesion. This makes possible the use of relatively high US
frequencies that provide high-quality, high-resolution images. In general, the higher the frequency of
US, the less the depth of penetration and vice versa. Also, the higher the frequency, the better the
resolution. Because the transducer is placed very close to target lesions during EUS, it is possible to
use relatively high frequencies. The most commonly used frequencies are 7.5 and 12 MHz; switching
between these two frequencies is possible with some echoendoscope systems (see Table 411).
Optimum acoustic coupling with the lesion is achieved by means of a water-filled balloon placed over
the transducer or by filling the stomach with water.
The gastrointestinal wall is generally visualized by EUS as five distinct layers. A relatively close
correlation exists between this sonographic structure or pattern and the histologic architecture of the
gut wallmucosa, submucosa, muscularis propria, and adventitia of the esophagus or the subserosa
and serosa of the stomach.1315(3273) However, the width of the various layers as measured by EUS
does not correlate with the real dimensions of the histologic components of the gut wall. This is
explained by the phenomenon of border echoes that occur as US passes from one medium to another
of different acoustic impendence.37(3274)
By convention, the sonographic layers of the gastrointestinal tract are numbered from innermost (first)
to outermost (fifth). Structures that are echogenic (echorich, hyperechoic) are displayed as white, and
sonolucent areas are displayed as black. The EUS definition of the gastrointestinal wall can be
summarized as follows:
The first echogenic structure and the second echopoor structure correspond to the border echo
created by the passage of US from water or balloon into the mucosa.
The third echogenic structure corresponds to the submucosa and the border echoes created by the
passage of the US energy from the submucosa into the muscularis propria.
The fourth echopoor structure corresponds to the muscularis propria minus the border echoes
between the submucosa and the muscularis propria and between the muscularis propria and the
submucosa.
The fifth echogenic structure corresponds to the adventitia of the esophagus or subserosa and serosa
of the stomach plus the border echo between the muscularis and the adventitia or the subserosa and
serosa.
For practical purposes, the border echo phenomenon may be discounted during EUS. Therefore, the
second (echopoor, hypoechoic) layer may be assumed to correspond to the mucosa, the third
(hyperechoic) layer to the submucosa, the fourth (hypoechoic) layer to the muscularis propria, and the
fifth (hyperechoic) layer to the adventitia of the esophagus or the serosa of the stomach.
Topographic Anatomy
Knowledge of topographic anatomy is essential to obtaining satisfactory cross-sectional images and to
defining the relationships between a target of interest and the surrounding structures (Figure 413).
Endosonographically, the esophagus can be divided into three segments: cervical, middle, and distal.
Cross-sectional imaging planes, similar to those of computed tomography (CT), should be used to
standardize the procedure.22(3275) As with CT, it is helpful to remember that EUS images are
displayed from the vantage point of the patient's feet.
(3276)Figure 413. Anatomic diagram of a posterior sagittal view of the chest showing the
topographic anatomic relationships among the esophagus (es), the descending aorta (dao), the
azygos vein (av), the right pulmonary vein (rpv), and the left pulmonary vein (lpv). The
azygos vein is running into the superior vena cava (svc). Note the relationship between the
esophagus and the thyroid gland (th) and between the esophagus and the right and left carotid
arteries (rca, lca). ivcinferior vena cava; didiaphragm.
At the level of the cervical esophagus, the most important landmarks are the carotid arteries, jugular
veins, and thyroid gland (Figure 414). The vertebral spine, seen as a concave hyperechoic structure
with dorsal shadowing, can be used as an additional reference point; the EUS image should be
manipulated so as to place vertebrae between the 5 and the 7 o'clock positions. The left and right
jugular veins should be positioned at 3 and 9 o'clock, respectively; the left and right carotid arteries
should be placed at 1 and 11 o'clock, respectively. The tracheavisualized as a hyperechoic line
structure with repeating echoes, the so-called reverberation phenomenacan be located between the
11 and the 12 o'clock positions. By maneuvering the EUS transducer and using the display functions to
manipulate the images, cross-sectional images similar to those of CT can be obtained.
(3277)Figure 414. Anatomic diagram of a cross-sectional view at the level of the cervical
esophagus showing the relationships between the esophagus and adjacent structures.
lxlarynx; ththyroid gland; cacarotid arteries; jvjugular veins, vvertebral body. The
vagus nerves (nv), the recurrent laryngeal nerves (nr), and the sympathetic nerve (ns) are not
visualized by endoscopic ultrasound (EUS).
The descending aorta is the primary landmark at the level of the middle esophagus, at approximately
26 to 36 cm from the incisor teeth; it should be positioned in the lower left of the EUS field. At
approximately 26 to 28 cm, the descending aorta connects via the aortic arch with the ascending aorta
(Figure 415). The azygos vein, found on the contralateral side from the descending aorta, terminates
in the superior vena cava. The trachea, located between the superior vena cava and the ascending
aorta, appears at EUS as multiple bright echoes in a linear, layer-like pattern, a manifestation of the
so-called reverberation phenomenon of the trachea. At the most distal part of this cross section,
proximally from the carina, the pulmonary artery can be found on the left upper quadrant ventrally from
the main bronchus. With optimum positioning of the echoendoscope, the pulmonary trunk and its
bifurcation can be imaged (Figure 416). With further insertion, the left atrium can be seen ventrally
and the descending aorta in the left dorsal quadrant at approximately the 5 o'clock position. The left
and right pulmonary veins entering the left atrium are present at approximately the 1 and the 11 o'clock
positions, respectively. With proper maneuvering of the echoprobe, the heart with its left atrium and left
ventricle separated by the mitral valve can be clearly visualized. On the contralateral side of the heart,
the right atrium and right ventricle separated by the triscupid valve are noted. The interventricular
septum separating the left and right ventricles can be clearly imaged as a muscle bundle (Figure
417).
(3278)Figure 415. Anatomic diagram of a supercarinal cross section of the chest at

approximately 26 to 28 cm from the incisor teeth. daodescending aorta; aaoascending
aorta; avazygos vein; svcsuperior vena cava. Note the close relationship between the
esophagus (es) and the trachea (tr) with its adjacent paratracheal lymph nodes (n). The
thoracic duct (td) can be seen between the aorta and the vertebral spine.
(3279)Figure 416. Anatomic diagram of a cross section of the chest at the level of carina
showing the right and left bronchus (rbr, lbr), descending aorta (dao), azygos vein (av), and
pulmonary trunk (ptr) with its left and right pulmonary artery (lpa, rpa). The ascending aorta
(aao) and superior vena cava (svc) are found beyond the left pulmonary artery. Note the
relationship between the carinal lymph nodes (n) and the esophagus (es). The descending aorta
(dao) and the azygos vein (av) are seen at the 5 o'clock and 7 o'clock positions, respectively.
(3280)Figure 417. Anatomic diagram of a cross section of the chest at approximately 30 cm

from the incisor teeth showing the left and right pulmonary veins (lpv, rpv) running into the
left atrium (la), left ventricle (lv), right atrium (ra), and right ventricle (rv). Note the mitral
valve and interventricular septum. The right perihilar lymph nodes (n) are found adjacent to
the right pulmonary vein. Note also the constant position of the descending aorta (dao) and the
azygos vein (av) in relation to the esophagus (es) and the vertebral body.
A small portion of the liver together with the left ventricle are imaged at the level of the distal
esophagus. Also present are the azygos vein and descending aorta, and occasionally, the hemiazygos
vein can be identified. On the contralateral side of the aorta, adjacent to the liver, the inferior vena
cavatogether with the hepatic veinscan be imaged. The left and right diaphragmatic crura are
found around the aorta; this anatomic feature may become important in staging esophageal cancers.
The left lobe of the liver and the spleen are readily imaged at the level of the gastroesophageal junction
just distal to the diaphragm (Figure 418). At the level of subcardial region, a small part of the tail of the
pancreas may be imaged. At the splenic hilum, the splenic vein and splenic arterypositioned
togethercan be used as landmarks to identify splenic lymph nodes. The splenic artery is located
directly adjacent to the gastric wall and has a tortuous configuration; the splenic vein, the major
landmark for imaging of the pancreas from the stomach, is positioned more dorsally than the splenic
artery. Using the splenic artery as a guide, the celiac axis, with its bifurcation into the hepatic and
splenic arteries, can be visualized. Identification of the celiac axis is very important for cancer staging
because metastasis of esophageal cancer to celiac lymph nodes constitutes distant metastases. The
left gastric artery, found along the lesser curvature of the stomach, is often accompanied by lymph
nodes.
(3281)Figure 418. Anatomic diagram of a cross section of the chest at the cardia
demonstrating the relationships between the fundus of stomach (fd) and the left lobe of the
liver (ll) and the spleen (sp). Note the constant position of the aorta (ao) in relation to the
vertebral body and the stomach. The inferior vena cava (vc) and the hepatic vein (hv) are
located within the liver parenchyma. The thoracic duct (td) is found to the right of the aorta.
Indications
The most important indications for EUS in the evaluation of diseases of the esophagus and chest are:
(1) cancer staging, (2) submucosal tumors, (3) gastroesophageal varices, (4) achalasia and
pseudoachalasia, and (5) mediastinal abnormalities.
Results
Cancer Staging
The American Joint Committee on Cancer (AJCC) in cooperation with the International Union Against
Cancer (UICC; Union Internationale Contre le Cancer) agreed in 1987 on a uniform classification
system of malignancy, including gastrointestinal cancers.38,39(3282) This new formulation, which has
become known as the TNM classification, refers to depth of penetration of the primary tumor into or
through the wall of the gastrointestinal tract (T), the status of regional lymph nodes with reference to
the presence or absence of malignant involvement (N), and the presence or absence of distant
metastases (M). In this system, involvement of nonregional lymph nodes is generally considered to be
a distant metastatic disease (M1).
According to the TNM system, tumors are classified from Stage 0 (best prognosis) to Stage IV (worst
prognosis). The recently (1987) modified TNM classification is widely accepted as the standard for
clinical staging of malignant tumors.2229(3283) This system has replaced other staging criteria,
including the length of the cancer, degree of obstruction, gross morphology, and resectability of the
tumor. Moreover, the definition of regional lymph nodes has been simplified; celiac lymph node
involvement is defined as a distant metastasis.38,39(3284) The TNM classification of esophageal
carcinoma is summarized in Table 412.
TNM Classification of
Esophageal Carcinoma Based on EUS
Findings
TABLE 412
PRIMARY TUMOR (T)

T1:
Tumor localized to lamina propria and submucosa
T2:
Tumor invading muscularis propria
T3:
Tumor with penetration into adventitia
T4:
Tumor with penetration into adjacent structures
(pericardium, aorta, tracheobronchial tree,
diaphragmatic crus, or liver)
Tx:
Primary tumor cannot be assessed
REGIONAL LYMPH NODE (N)
N0:
Hyperechoic, indistinctly demarcated lymph nodes
N1:
Hypoechoic lymph nodes with clearly defined
borders, or direct penetration of tumor into
adjacent lymph nodes
Esophageal Carcinoma Based on EUS
Findings
TABLE 412
Nx:
Regional lymph nodes cannot be assessed
DISTANT METASTASES (M)
M0:
No evidence of metastases in celiac lymph nodes
or direct penetration of tumor into adjacent lymph
nodes
M1:
Distant metastasis: hepatic metastasis and/or
lymph node metastasis at celiac trunk
Mx:
Distant metastasis cannot be assessed
STAGE GROUPING
Stage I
T1
N0
M0
Stage IIA
T2
N0
M0
T3
N0
M0
Stage IIB
T1
N1
M0
T2
N1
M0
Stage III
T3
N1
M0
T4
Any N
M0
Stage IV
Any T
Any N
M1
TNMtumor, node, metastases; EUSendoscopic
ultrasonography.
Since the introduction of EUS, many reports have been published concerning staging of esophageal
cancer.1429(3285) The overall reported accuracy of EUS for assessment of the depth of tumor
invasion and status of regional lymph nodes is summarized in Table 413. Lightdale,27(3286) in a
review of data from 406 patients gathered from seven centers, calculated the accuracy of EUS for
assessing depth of tumor invasion and regional lymph node status at 85% and 75%, respectively. In
general, the accuracy of EUS is not influenced by the cell type of the malignancy, whether squamous
cell or adenocarcinoma. However, Dittler and Siewert40(3287) found EUS to be less accurate in
predicting resectability of squamous cell cancer compared with adenocarcinoma. In this study, which
included 108 patients with squamous cell carcinoma and 59 patients with adenocarcinoma, EUS
correctly predicted resectability in 64% of the former group versus 82% of the latter patients. This
difference was attributed to the inability of EUS to detect the small submucosal metastases that are
common with squamous cell cancers.
Accuracy of EUS for T and N Staging of Esophageal

Carcinoma in Patients Who Underwent Surgery
TABLE 413
STUDY (YEAR)
T ACCURACY (%)
N ACCURACY (%)
Murata et al.43 (1988)

Takemoto et al.80 (1989)
173
88
88
18
72
79
Tio et al.22 (1989)

Sch??der et al.81 (1990)
74
89
81
22
77
86
Sugimachi et al.82 (1990)

Vilgrain et al.83 (1990)
33
90
51
73
50
Accuracy of EUS for T and N Staging of Esophageal

Carcinoma in Patients Who Underwent Surgery
TABLE 413
STUDY (YEAR)
T ACCURACY (%)
N ACCURACY (%)
Botet et al.44 (1991)

Rice et al.84 (1991)
50
92
88
22
59
69
Ziegler et al.24 (1991)*

Fok et al.85 (1992)
37
89
69
45
89
Heintz et al.86 (1992)

Kalantzis et al.87 (1992)
38
74
84
28
82
72
Nobre-Leito et al.88 (1992)

Rsch et al.89 (1992)
36
80
79
44
82
70
Souquet et al.90 (1992)

Dittler and Siewert40 (1993)
82
77
82
167
86
73
Grimm et al.91 (1993)

Hordijk et al.46 (1993)
63
86
90
41
76
Nattermann and
Dancygier92 (1993)
Napolitano et al.93 (1994)
44
80
55
21
71
80
34
76
82
Peters et al.51 (1994)
All studies T Stage
1126
84
N Stage
1007
77
Modified from Rsch T. Endosonographic staging of esophageal cancer: A review of
literature results. Gastrointest Endosc Clin North Am 1995; 5:53747.
* Examinations were performed with a linear array instrument. All other studies were
done with radial sector scanning instruments.
Although the overall accuracy of EUS in assessing depth of tumor invasion is high, there is some
variability with respect to T stage. Based on a review of 15 reported studies, Rsch41(3288) found
EUS accuracy in relation to T stage to be as follows: T180.5%, T276%, T392%, T486%.
Information concerning overstaging and understaging of esophageal cancer was available from 11
reports; overstaging occurred with 16.5%, 17%, and 6% of T1, T2, and T3 cancers, respectively,
whereas understaging was encountered in 10%, 5%, and 11% of T2, T3, and T4 malignant tumors. It
is evident, therefore, that T2 cancers present the greatest difficulty for EUS in terms of accurate
staging.
The evaluation of early-stage, superficial esophageal cancer is also problematic. Such a lesion may be
confined to the mucosa, or it may invade the submucosa. In either case, this tumor would be classified
as T1. However, there are marked differences in prognosis depending on whether a lesion has
invaded the submucosa. There is less possibility of lymph node metastasis for cancers confined to the
mucosa, whereas there is a significant chance that the malignancy has spread to lymph nodes when
there is invasion of the submucosa. This differentiation has implications for treatment in that tumors
confined to the mucosa are amenable to local endoscopic treatment whereas those that invade the
submucosa require surgical resection with lymph node dissection. The accuracy of EUS staging for 9
mucosal and 19 submucosal esophageal cancers in the study of Yoshikane et al.42(3289) was 67%
and 79%, respectively. In the series of Murata et al.,43(3290) EUS staging was correct for 10 of 12
esophageal cancers confined to the mucosa and 38 of 45 (84%) early-stage malignant tumors that had
invaded the submucosa.
EUS and CT were performed using the modified TNM classification in our initial study of 74 patients
with esophageal cancer.22(3291) The overall accuracy of EUS (89%) was significantly better than that
for CT (59%). EUS was found to be accurate in tumor staging regardless of the anatomic location of
the lesion (Figures 419, 4110 and 4111). CT was found to be substantially less accurate for staging
esophageal cancers; the difference in accuracy was more pronounced in early-stage tumors (T1 and
T2) at 82% for EUS compared with 12% for CT. EUS can differentiate early- and advanced-stage
cancers based on depth of tumor invasion (Figure 4112). In a study of 50 patients with esophageal
carcinoma, Botet et al.44(3292) found that EUS was significantly more accurate than CT in assessing
depth of tumor invasion (92% vs. 60%). The accuracy in several reported series for EUS compared
with CT for T stage of esophageal cancer is summarized in Table 414.
Comparison of EUS and CT in the Locoregional Staging of

Esophageal Cancer
TABLE 414
T STAGING
N STAGING
(YEAR)
EUS (%)
CT (%)
EUS (%)
CT (%)
Tio et al.22
1989
74
89
59
80
51
Vilgrain et al.83
Botet et al.44
1990
51
50
48
1991
50
92
60
88
74
Ziegler et al.24
Grimm et al.94
1991
37
89
51
69
51
1991
49
89
62
Heintz et al.95
Schder et al.81
1991
22
77
64
86
50
1991
22
86
57
81
48
Kalantzis et al.87
Hordijk et al.46
1992
28
82
50
72
46
1993
41
76
49
STUDY
Nattermann and
1993
44
80
55
84
57
92
Dancygier
All studies
T Stage
367
85
58
N Stage
328
75
54
Modified from Rsch T. Endosonographic staging of esophageal cancer: A review of literature
results. Gastrointest Endosc Clin North Am 1995; 5:53747.
(3293)Figure 419. A, Barium swallow showing circumferential esophageal cancer in the

upper third of the esophagus. Note the elongated filling defect (arrows) adjacent to the
primary bronchus (br) and the level of the carina (c). B, EUS reveals a circumferential
transmural esophageal cancer (t) with maximum penetration into the adjacent hyperechoic
line-like bronchial structure (br). Note the position of the descending aorta (ao) between the 3
o'clock and 5 o'clock positions and the azygos vein (av) at the contralateral side. The
hyperechoic line-like structurewith its adjacent reverberation phenomenanext to the
azygos vein represents the right bronchus (rbr).
(3294)Figure 4110. A, Barium swallow showing an obstructing esophageal cancer in the

upper third of the esophagus. Note the polypoid filling defect beyond the obstruction (arrows).
B, Computed tomography (CT) scan after intravenous injection of a contrast agent reveals a
circumferential tumor (t). aoaraortic arch; avazygos vein; svcsuperior vena cava. The
trachea is demarcated by the esophagus, the azygos vein, and the aorta. C, Corresponding EUS
reveals an almost circumferential hypoechoic tumor (t) with maximum penetration (arrows)
into the adjacent azygos vein (av). Note the aortic arch (aoar), superior vena cava (svc),
azygos vein (av) draining into the superior vena cava, and the trachea (tr), which is seen as a
hyperechoic white line.
(3295)Figure 4111. A, Barium swallow showing polypoid tumor (arrows) located at the
posterior wall of the midesophagus. vvertebra. B, EUS reveals a semicircular hypoechoic
tumor (t) with penetration into the adventitia. There is a clear transition between the tumor and
the aorta (ao). Note the small hypoechoic lymph nodes (n) adjacent to the aorta and the left
atrium (la). The sonographic appearance of these lymph nodes does suggest metastasis. The
EUS findings are consistent with a T3N0 tumor.
(3296)Figure 4112. A, EUS obtained at the level of the midesophagus reveals a hypoechoic
tumor (t) in the submucosa (sm) without penetration into the muscularis propria (mp)
compatible with a T1 cancer. wbwater-filled balloon. B, EUS shows a deep transmural
hypoechoic tumor (t) at the left side with penetration into the adjacent left bronchus (br).
These findings are consistent with a T4 esophageal cancer. aoaorta; laleft atrium. Note
the obvious difference in depth of tumor invasion between A and B.
Hypoechoic, sharply demarcated lymph nodes adjacent to the tumor or at the celiac axis are
considered strongly suspicious of metastasis (Figure 4113). In a study of 100 patients with
esophageal carcinoma, Catalano et al.45(3297) found that EUS had a sensitivity and specificity of
89.1% and 91.7%, respectively, with regard to the detection of lymph node metastases. Certain EUS
features were found to be predictive of malignancy. In increasing order of importance, these were
echopoor (hypoechoic) structure, sharply demarcated borders, rounded contour, and a diameter of
greater than 10 mm. If all four features were present, EUS was said to be 100% accurate in the
recognition of malignant lymph nodes. However, the distinction between inflammatory changes of
lymph nodes and metastatic lymph nodes is often difficult or impossible. Thus, in the near future,
EUS-guided FNA cytology will probably become more routine (see Chapter 21: Endoscopic
Ultrasonography With Linear Array Instruments: Anatomy and EUS-Guided Fine-Needle Aspiration).
(3298)Figure 4113. A, EUS reveals a hypoechoic transmural tumor (t) with penetration into
the adventitia. aoaorta. The sonographic appearances of two hypoechoic lymph nodes (n)
with clearly defined boundaries are strongly suggestive of metastasis. Note that the echo
pattern of lymph nodes is more hypoechoic than that of the primary tumor. These lymph nodes
are located adjacent to the left atrium (so-called retrocardial lymph nodes). B, EUS showing
two hypoechoic lymph nodes (n) with sharply defined borders located at the celiac axis. The
appearances are strongly suggestive of metastasis. A round, hyperechoic lesion (arrow) in the
left lobe (l) of the liver is most likely a hemangioma.
In the study of 50 patients with esophageal carcinoma by Botet et al.,44(3299) EUS was more accurate
than CT in assessing lymph node involvement (88% vs. 74%), although this difference was not
statistically significant. The accuracy of several reported series for EUS compared with CT for N stage
of esophageal cancers is summarized in Table 414.
In our initial series, it was not possible to pass the echoendoscope or even the (blind) echoprobe
through a malignant stenosis in 26% of cases. Similar difficulty has been encountered in most series,
as summarized in Table 415. In most, but not all reports, the inability to traverse a malignant stricture
adversely influenced the staging accuracy of EUS. For example, Hordijk et al.46(3300) found that EUS
had an overall T-stage accuracy of 76%. However, the accuracy for staging of easily traversable
malignant stenoses was 92% compared with 46% in cases where the obstructing cancer was
traversed with difficulty. Of interest, the accuracy of EUS for T stage of tumor stenoses that were
nontraversable was 87%. These investigators attributed this to the fact that most nontraversable
cancers were advanced-stage malignancies.
Frequency of Passage of
Echoendoscope Through Malignant
Esophageal Stenosis
TABLE 415
STUDY
PASSAGE (%)
Ziegler et al.24
Botet et al.44
37
81
50
74
Grimm et al.96
Grimm et al.91
49
74
51
94
Dittler and Siewert40

Van Dam et al.97
32
74
79
73
Nattermann and Dancygier92

Rsch et al.89
44
72
44
70
Heintz et al.86
63
67
Heyder and Lux98

Murata et al.43
35
64
269
64
Hordijk et al.46
Vilgrain et al.83
41
63
51
50
32
37
Dancygier and Classen99
Modified from Rsch T. Endosonographic staging of
esophageal cancer: A review of literature results.
A catheter echoprobe can often be passed through a tight malignant stricture that does not permit
passage of an echoendoscope (Figure 4114). Glover et al.47(3301) found that a nonoptical steerable
echoprobe could be passed through tight malignant strictures in 75 of 80 attempts in 50 patients.
Depth of tumor invasion and status of periesophageal lymph nodes were readily assessed with this
device at a frequency of 5 MHz. Assessment of the depth of invasion may prove to be more
problematic with catheter echoprobes that utilize higher frequencies.36,48,49(3302)
(3303)Figure 4114. A, Barium swallow radiograph reveals a filiform stenosis (arrows) in the
midesophagus due to a cancer. B, EUS obtained with a catheter echoprobe reveals a
circumferential hypoechoic tumor (t) with maximum penetration (arrows) directly adjacent to
the bronchus (br). The appearance is indicative of a T4 cancer. AOaorta.
Staging of distant metastases by EUS is usually not possible because of the limited penetration depth
of US, with some exceptions. In some cases, a focal metastatic lesion may be visualized in the left
lobe of the liver. Also, metastasis of esophageal carcinoma to celiac lymph nodes is classified as a
distant metastasis (M1) in the TNM system. The accuracy of EUS in staging distant metastases was
68% compared with 82% for CT in our initial series. CT had an accuracy of 90% compared with 70%
for EUS with respect to distant metastasis in the study by Botet et al.44(3304) of 50 patients with
esophageal carcinoma.
Despite the remarkable accuracy of EUS in assessing the depth of malignant tumor invasion and
status of lymph nodes, there is relatively little evidence that this precise information is being utilized in
making clinical decisions concerning the management of patients with esophageal carcinoma. In a
multicenter retrospective cohort study of 79 patients with locally invasive (T4) esophageal cancer,
Chak et al.50(3305) found no difference in survival for patients who underwent surgery (n = 42)
compared with patients who received only palliative therapy (n = 37). As a group, those who underwent
surgery were significantly younger and had a greater number of distal esophageal cancers. Age,
histolopathologic diagnosis, tumor location, and status of regional lymph nodes did not influence
survival in surgically treated patients. Peters et al.51(3306) used EUS findings as a guide to surgical
therapy. Patients with early-stage lesions underwent en bloc esophagogastrectomy; those with
advanced cancers underwent palliative transhiatus esophagectomy.
Follow-Up
EUS has been helpful after distal esophageal resection in detecting and staging submucosal recurrent
tumor in cases where endoscopy does not reveal any evidence of recurrence. In such cases, the tumor
may be buried under normal overlying mucosa and cannot be reached with a biopsy forceps. Lightdale
and Botet25(3307) reported that EUS had an accuracy of 95% in diagnosing recurrent cancer after
distal esophageal resection. We found that EUS-guided cytologic FNA cytology was also helpful in
obtaining a tissue diagnosis.22(3308)
EUS has also been used for restaging of esophageal carcinoma prior to surgery in patients who have
undergone radiotherapy and chemotherapy.52(3309) We performed EUS in 63 patients with
inoperable esophageal cancer treated with combined intraesophageal selectron therapy
(brachytherapy) and external irradiation.53(3310) EUS was shown to be useful as a guide to staging,
prognosis, and documentation of treatment response. Reductions in tumor thickness, number, and
size of malignant-appearing lymph nodes were found after successful treatment. The survival rate
correlated positively with the tumor thickness and negatively with the number and size of metastatic
lymph nodes.
Barrett's Esophagus
Data concerning EUS in Barrett's esophagus are conflicting; a definite role in this disease has not been
established. Falk et al.54(3311) prospectively evaluated nine patients with Barrett's esophagus and
high-grade dysplasia but not carcinoma as demonstrated by endoscopic biopsies. EUS findings,
obtained with a sector scanning echoendoscope at frequencies of 7.5 and 12 MHz, were compared
with the histopathologic assessment of surgically resected specimens. Intramucosal carcinoma was
found in three of nine specimens; EUS had identified a malignant tumor in only one of these patients,
although this lesion had been overstaged as an invasive carcinoma (T2N1). In two other patients with
no histologic evidence of cancer, EUS findings of mucosal nodularity were misinterpreted as indicating
the presence of an intramucosal carcinoma. These findings are in contrast to those of Srivastava et
al.,55(3312) who compared EUS findings in 15 patients with Barrett's esophagus to those in 13 normal
volunteers. Esophageal wall thickness as determined by EUS (12 MHz, rotating sector scanning
instrument) was measured at 2- to 3-cm intervals over the length of the esophagus. The mean
thickness of the wall in control subjects was 2.6 mm versus 3.3 mm in patients with Barrett's
esophagus without dysplasia and 4.0 mm in those with dysplastic Barrett's epithelium; the difference in
wall thickness for patients with nondysplastic and dysplastic Barrett's epithelium was statistically
significant. Focal areas of submucosal thickening that measured 7 and 10 mm were found in 2
patients. Evaluation of resection specimens from these 2 patients revealed submucosal carcinoma in
the areas of thickening as demonstrated by EUS.
Submucosal Tumors
Submucosal masses remain enigmatic for the endos-copist because the submucosal extent and the
probable origin of the lesion cannot be assessed from the endoscopic appearance. Biopsies obtained
with standard endoscopic forceps are usually not helpful. Specimens obtained with large ("jumbo,"
"giant") forceps may sometimes provide useful information, especially if several biopsies are obtained
at a single site, but this maneuver in the esophagus may be associated with increased risks of
bleeding, perforation, and mediastinitis.
EUS, in contrast to endoscopy alone, is very useful in the diagnosis of submucosal tumors (SMTs)
because it provides images of the gastrointestinal wall and surrounding structures. A leiomyoma is
visualized as a hypoechoic tumor originating from the muscularis propria with normal overlying mucosa
(Figure 4115).13(3313) A blood vessel or an ulcer may be present within the tumor. For tumors that
arise in the muscular components of the gut wall, the risk of malignancy is thought to increase with
increasing size. It is usually impossible to accurately determine the size of an SMT from the
endoscopic appearance alone because the bulk of many of these lesions projects beyond the
esophageal wall into the chest. In the case of a large SMT (diameter greater than 3 to 4 cm),
malignancy is suggested by the presence of indefinite, irregular boundaries and adjacent lymph nodes
with features of metastatic involvement. A leiomyosarcoma or leiomyoblastoma is imaged as a
hypoechoic tumor mass with an heterogeneous echo pattern; a central ulcer is often present. Other
findings that suggest malignancy include penetration into adjacent structures, fistula tract formation,
and distortion of the configuration of the involved organ. However, similar EUS features may be found
with metastatic carcinoid tumors, submucosal invasion by cancer arising in a bronchus, and
submucosal lymph node metastases from breast, esophageal, or gastric cancer. On occasion,
observation of a rapid increase in size of a smooth muscle tumor may be an indication of
malignancy.56(3314) Because differential diagnosis based on echo features alone is difficult,
EUS-guided FNA cytology is gaining importance in the evaluation of SMTs.35,57(3315)
(3316)Figure 4115. EUS reveals a circumscribed hypoechoic intramural tumor (t) with
sharply demarcated boundaries. The tumor is contiguous with the muscularis propria (mp).
Long-term follow-up by EUS and endoscopy confirmed the diagnosis of leiomyoma.
The EUS appearance was correlated with histopathologic findings in 24 of 42 patients in our series of
gastroesophageal SMTs.58(3317) Differentiation between leiomyoma and leiomyosarcoma was not
always possible unless there was infiltration into the adjacent structures or suspicious lymph nodes
were present. However, lipoma, ectopic pancreas, and paraesophageal varices could be differentiated
from leiomyoma based on echo pattern, configuration, and origin. Lipoma was seen as a hyper-echoic
submucosal mass with a density similar to that of fat or connective tissue (Figure 4116). Ectopic
pancreas appears as a hypoechoic submucosal rim with a central ductular structure. Paraesophageal
varices were imaged as serpent-like, anechoic structures located in the adventitia. A case of a granular
cell tumor of the esophagus has been described as a hypoechoic mass in the submucosa without
continuity to the muscularis propria.59(3318) In a series of 25 SMTs, Caletti et al.60(3319) found that
EUS identified all lesions except for one 5-mm diameter antral lesion. Compression by an adjacent
organ (liver, spleen, aorta) was found in 6 cases; compression by another pathologic structure was
found in 4 (pancreatic pseudocyst, endocrine tumor, liver metastases).
(3320)Figure 4116. EUS obtained at the subcardial region reveals a hyperechoic tumor (t) in
the submucosa. The echo pattern of the tumor is more hyperechoic compared with that of the
muscularis propria (mp). These finding are consistent with a lipoma. Note also the clear image
of the mucosa (m), submucosa (sm), and muscularis propria. The hyperechoic structure
adjacent to the muscularis propria corresponds to the subserosa and the serosa.
Boyce et al.61(3321) examined 83 patients with SMTs found at endoscopy. EUS findings were
compared in 40 cases to surgical pathology or angiography where there was a suspicion of esophageal
varices. EUS was correct in predicting the site of origin of the lesion in 13 of the 14 cases where the
esophagus was involved. Dancygier and Classen62(3322) described a "jumping sign" phenomenon
that is associated with leiomyoma. This refers to the fact that a leiomyoma will appear and disappear
rapidly ("jump") on the image monitor during the passage of the water-filled balloon over the target.
However, this phenomenon may also be observed with other types of SMTs. Yasuda et al.,63(3323)
based on the study of 308 SMTs, emphasized that there is no unique EUS feature that reliably
distinguishes a leiomyoma from a leiomyosarcoma. EUS correctly differentiated SMTs arising within
the wall of the esophagus and extraluminal sources of esophageal compression in a study of 55
patients with SMTs as demonstrated by radiography or endoscopy in the study of Murata et al.64(3324)
EUS also correctly identified the tissue of origin in 28 of 29 lesions that arose within the esophageal
wall.
Intramural congenital esophageal cyst is a relatively uncommon lesion found mainly in children and
occasionally in adult patients. At endoscopy, this lesion may be confused with more common SMTs
such as leiomyoma. EUS readily demonstrates the cystic nature of this congenital
anomaly.6567(3325)
Gastroesophageal Varices
Caletti et al.68(3326) described the EUS appearance of esophageal varices as echofree structures
beneath the mucosa or submucosa. In their series, EUS demonstrated only 14% of endoscopic Grade
I varices, 78% of Grade II, and 50% of Grade III varices. For the diagnosis of esophageal varices, EUS
is in effect less sensitive than endoscopy. Periesophageal collateral veins were demonstrated by EUS
in 57% of patients with endoscopic Grade I varices, 89% with Grade II, and 100% with Grade III
esophageal varices. Compression by the water-filled balloon usually accounts for the failure of EUS to
demonstrate esophageal varices, including those that are clearly visible at endoscopy. High-frequency
echoprobes may be more suitable for imaging intraepithelial and submucosal variceal channels,
although the limited depth of penetration of these instruments may be problematic with respect to
visualization of periesophageal collateral veins.69(3327)
For the diagnosis of gastric varices, Caletti et al.68(3328) found that EUS was more accurate than
endoscopy. These investigators also found the diameter of the azygos vein, as determined by EUS, to
be an important indicator of portal hypertension; a diameter between 5 and 13 mm was considered to
be abnormal.
In our initial experience, EUS demonstrated varices as anechoic structures in the submucosa.13(3329)
EUS demonstrated the presence of esophageal varices in 20 of 24 cases found by endoscopy. For
gastric varices, however, EUS made a correct diagnosis in 21 patients compared with 16 for
endoscopy alone. Wall thickening after sclerotherapy was found in 20 patients by EUS and in only 10
patients by endoscopy.70(3330) Collateral veins, however, could be seen only by EUS. Occasionally, a
perforating vein, located adjacent to the gastroesophageal junction, can be found tangential to a varix.
This observation could be important clinically because some investigators believe that obliteration of
this vein by sclerotherapy may prevent further variceal bleeding.
Intraoperative EUS has been proposed as an imaging method for evaluation of the collateral circulation
including the completeness of devascularization during nonshunting operations in patients with portal
hypertension.71(3331)
Achalasia and Pseudoachalasia

The differentiation between achalasia and pseudoachalasia is often impossible based solely on
endoscopic appearance. An extramural mass may compress the esophagus circumferentially and may
cause only minimal endoscopic findings such as a shallow ulceration.13(3332) EUS can readily image
the extent of an extraesophageal mass such as a bronchial carcinoma, which may invade the
submucosa of the esophagus. In one of our patients, EUS demonstrated that an aortic aneurysm at the
level of the distal esophagus was simulating the endoscopic appearance of achalasia.14(3333)
Deviere et al.72(3334) performed EUS in six patients with achalasia prior to dilation. Enlargement of
the muscularis propria (fourth hypoechoic layer) was noted in five patients. These investigators
suggested that this appearance is typical for achalasia. Histopathologic examination of an autopsy
specimen available from one of these five patients confirmed the EUS finding. In two cases with
pseudoachalasia due to tumor infiltration, the normal wall architecture was destroyed.
Ziegler et al.73(3335) obtained EUS images with a linear array echoendoscope in 16 patients with
achalasia; the EUS findings were considered to be within normal limits in 15 patients. There was no
evidence of hypertrophy of the muscularis propria in any patient. In the remaining one patient, an
intramural abnormality was imaged that proved to be an adenocarcinoma of the esophagus.
Van Dam et al.74(3336) performed EUS with a sector scan echoendoscope in 17 patients with
achalasia and measured wall thickness in four quadrants at 1-cm intervals in the esophagus beginning
at the gastroesophageal junction. Patients were divided into two groups: those with a tortuous
esophagus (n = 5) and those in whom the esophagus was relatively straight (n = 12). These
measurements were compared with similar measurements in 6 patients without esophageal
symptoms. At the gastroesophageal junction, there was no significant difference in wall thickness
between the two groups of patients and controls. Proximal to this level, however, a significant increase
in wall thickness was noted in patients with a tortuous esophagus compared with control subjects and
patients with achalasia and a straight esophagus. Van Dam et al.74(3337) pointed out that artifacts
frequently occur when EUS is performed in patients with achalasia and a tortuous esophagus because
of scanning tangential to the esophageal wall. These investigators concluded that EUS using current
techniques is not useful in the routine evaluation of patients with achalasia.
Mediastinal Abnormalities
EUS can distinguish intramural lesions from extramural tumor masses, two different lesions that may
have a similar endoscopic appearance. At EUS, a bronchial carcinoma appears as a hypoechoic tumor
mass adjacent to the bronchus; this lesion can extend into surrounding periesophageal structures
(Figure 4117). Occasionally, a bronchial cyst may be misdiagnosed at endoscopy as a leiomyoma.
The correct diagnosis is readily made at EUS, which visualizes the bronchial cyst as a hypoechoic
mass adjacent to the hyperechoic structure of the corresponding bronchus and adjacent to the
esophagus or aorta.13(3338)
(3339)Figure 4117. EUS image obtained at level of the midesophagus reveals an extensive
hypoechoic tumor mass (t) adjacent to the left bronchus (br). The tumor extends beyond the
left bronchus into the lateral side (arrows). Note that the distance between the esophagus and
the left atrium (la) is expanded owing to the tumor. aoaorta.
EUS may be helpful in the diagnosis of tracheobronchial carcinoma when bronchoscopy or CT does
not provide a diagnosis. In cases where there is a close relationship between the tumor and the
esophageal wall, EUS-guided FNA is very helpful in establishing a positive diagnosis.75(3340) We
have used EUS for the evaluation of lymph nodes adjacent to the carina (unpublished series). Such
lymph nodes are often difficult to detect by CT. When the FNA cytology of a carinal lymph node is
positive, some surgeons recommend that resection should not be attempted because of the difficulty of
total lymph node dissection in this area.
Wiersema et al.76(3341) reported three cases in which dysphagia was caused by enlarged mediastinal
lymph nodes due to histoplasmosis. EUS demonstrated large, hypo-echoic lymph nodes adjacent to
the esophagus. Regional thickening of the muscularis propria was present in all cases, and focal
anechoic areas were noted within the adjacent lymph nodes that were causing esophageal
compression. EUS-guided FNA demonstrated reactive changes.
Miscellaneous Conditions
Miller et al.77(3342) performed endoluminal sonography using a high-resolution US probe in patients
with systemic sclerosis and in normal volunteers. Autopsy specimens of a normal esophagus and from
patients with systemic sclerosis were also imaged. These studies demonstrated a hyperechoic
abnormality within the hypoechoic muscularis propria that appeared to correspond to the presence of
fibrosis on histologic sections from postmortem specimens. The degree of the hyperechoic abnormality
corresponded to the severity of esophageal manometric abnormalities and acid reflux.
Taniguchi et al.78(3343) developed a miniature high-resolution US device that can be attached to the
gastrointestinal wall by means of suction. In an animal study, these investigators observed changes in
the width of the inner circular muscular layer during esophageal contractions as well as a decrease in
width after esophageal dilations. Traumatic injury to the mucosal surface and esophageal wall was not
encountered.
Primary lymphoma of the esophagus is a rare condition. The primary EUS feature in a case reported
by Bolondi et al.79(3344) was diffuse hypoechogenic thickening, frequently of a marked degree, of the
esophageal wall.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
Dussik KT, Dussik F, Wyt L. Auf dem Weg zur Hyperphonographie des Gehirnes. Wien Med
Wild JJ. The use of ultrasonic pulses for the measurement of biological tissue changes.
Surgery 1950;27:1837.
Wild JJ, Reid JM. Echographic tissue diagnosis. Proceedings of the 4th Annual Conference of
Ultrasonic Therapy, Detroit, August 27. 1955.
Wild JJ, Reid JM. Diagnostic use of ultrasound. Br J Phys Med 1956;19:24857.
Watanabe H, Kato H, Tanaka M, Terasawa Y. Diagnostic application of ultrasonotomography
for the prostate. Jpn J Urol 1968;59:2737.
Holm HH, Northeveld AA. Transuretral ultrasonic scanner. J Urol 1974;111:23848.
Bom N, Lancee CT, van Egmond FC. An ultrasonic intracardiac scanner. Ultrasonics
1972;10:726.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Hisanaga K, Hisanaga A. A new real-time sector scanning system of ultra-wide angle and
real-time recording to entire cardiac images: Transesophagus and transchest methods.
Ultrasound Med Biol 1978;4:391401.
Lutz H, Rsch W. Transgastroscopic ultrasonography. Endoscopy 1976;8:2035.
DiMagno EP, Buxton GL, Regan PT, et al. Ultrasonic endoscopy. Lancet 1980;1:62931.
Strohm WD, Phillip J, Hagenmller F, Classen M. Ultrasonic tomography by means of an
ultrasonic fiberendoscope. Endoscopy 1980;12:2414.
Lux G, Heyder N, Demling L. Endoscopic ultrasonographyTechnique, orientation and
diagnostic possibilities. Endoscopy 1982;4:2205.
Tio TL, Tytgat GNJ. Atlas of Transintestinal Ultrasonography. Aalsmeer, The Netherlands:
Mur-Kostverloren. 1986.
Tio TL, Tytgat GNJ. Endoscopic ultrasonography in the assessment of intra- and transmural
infiltration of tumors in the oesophagus, stomach and papilla of Vater and in the detection of
extra-esophageal lesions. Endoscopy 1984;16:20310.
Aibe T, Fuji T, Okita T, Takemoto T. A fundamental study of normal layer structure of the
gastrointestinal wall visualized by endoscopic ultrasonography. Scand J Gastroenterol Suppl
1986;123:615.
Tio TL, den Hartog Jager FC, Tytgat GN. Endoscopic ultrasonography of non-Hodgkin
lymphoma of the stomach. Gastroenterology 1986;91:4018.
Tio TL, den Hartog Jager FC, Tytgat GN. The role of endoscopic ultrasonography in
assessing local resectability of oesophagogastric malignancies. Accuracy, pitfalls, and
predictability. Scand J Gastroenterol Suppl 1986;123:7886.
Takemoto T, Ito T, Aibe T, Okita K. Endoscopic ultrasonography in the diagnosis of
esophageal carcinoma, with particular regard to staging it for operability. Endoscopy
1986;18(suppl 3):225.
Murata Y, Muroi M, Yoshida M, et al. Endoscopic ultrasonography in the diagnosis of
esophageal carcinoma. Surg Endosc 1987;1:116.
Yasuda K, Kiyota K, Nakajima M, Kawai K. Fundamentals of endoscopic laser therapy (ELT)
for G.I. tumorsNew aspects with endoscopic ultrasonography (EUS). Endoscopy
1987;19(suppl 1):26.
Heyder N. Endoscopic ultrasonography of tumours of the oesophagus and the stomach. Surg
Endosc 1987;1:1723.
Tio TL, Cohen P, Coene PP, et al. Endosonography and computed tomography of
esophageal carcinoma. Preoperative classification compared to the new (1987) TNM system.
Tio TL, Coene PP, Schouwink MH, Tytgat GNJ. Esophagogastric carcinoma: Preoperative
TNM staging with endosonography. Radiology 1989;173:4117.
Ziegler K, Sanft C, Zeitz M, et al. Evaluation of endosonography in TN staging of oesophageal
cancer. Gut 1991;32:1620.
Lightdale CJ, Botet JF. Esophageal carcinoma: Pre-operative staging and evaluation of
anastomotic recurrence. Gastrointest Endosc 1990;36(suppl 2):116.
Lightdale CJ, Botet JF, Zauber AG. Endoscopic ultrasonography in the staging of esophageal
cancer: Comparison with dynamic CT and surgical pathology [Abstract]. Gastrointest Endosc
1990;36:191.
Lightdale CJ. Endoscopic ultrasonography in the diagnosis, staging and follow-up of
esophageal and gastric cancer. Endoscopy 1992;24:297303.
Heintz A, Hohne U, Schweden T, Junginger T. Preoperative detection of intrathoracic tumor
spread of esophageal cancer: Endosonography versus computed tomography. Surg Endosc
1991;5:758.
Dittler HJ, Holscher AH, Siewert JR. Staging des sophaguskarzinoms mit endoluminalem
Ultraschall. Dtsch Med Wochenschr 1991;116:5616.
Vilmann P, Jacobsen GK, Henriksen FW, Hancke S. Endoscopic ultrasonography with guided
fine needle aspiration biopsy in pancreatic disease. Gastrointest Endosc 1992;38:1723.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Tio TL, Sie LH, Tytgat GNJ. Endosonography and cytology in diagnosis and staging of
pancreatic body and tail carcinoma. Dig Dis Sci 1993;1:5964.
Grimm H, Binmoeller K, Soehendra N. Ultrasonic esophagoscope. Gastrointest Endosc
1992;38:4903.
Tio TL, Coene PP, Luiken GJ, Tytgat GN. Endosonography in the clinical staging of
esophagogastric carcinoma. Gastrointest Endosc 1990;36(suppl 2):S210.
Silverstein FE, Martin RW, Kimmey MB, et al. Experimental evaluation of an endoscopic
ultrasound probe: In vitro and in vivo canine studies. Gastroenterology 1989;96:105862.
Wiersema MJ, Hawes RH, Tao LC, et al. Endoscopic ultrasonography as an adjunct to fine
needle cytology of the upper and lower gastrointestinal tract. Gastrointest Endosc
1992;38:359.
Saisho H, Sai K, Tsuyuguchi T, et al. A new small probe for ultrasound imaging via
conventional endoscope. Gastrointest Endosc 1995;41:1415.
Kimmey MB, Martin RW, Hagitt RC, et al. Histologic correlates of gastrointestinal ultrasound
images. Gastroenterology 1989;96:43341.
Sobin LH, Hermanek P, Hutter RVP. TNM classification of malignant tumors: A comparison
between the new (1987) and the old editions. Cancer 1988;6:23104.
Hermanek P, Sobin LH. TNM Classification of Malignant Tumors. 4th ed. New York,
Heidelberg: Springer-Verlag. 1987.
Dittler HJ, Siewert JR. Role of endoscopic ultrasonography in esophageal carcinoma.
Endoscopy 1993;25:15661.
Rsch T. Endosonographic staging of esophageal cancer: A review of literature results.
Yoshikane H, Tsukamoto Y, Niwa Y, et al. Superficial esophageal carcinoma: Evaluation by
endoscopic ultrasonography. Am J Gastroenterol 1994;89:7027.
Murata Y, Suzuki S, Hashimoto H. Endoscopic ultrasonography of the upper gastrointestinal
tract. Surg Endosc 1988;2:1803.
Botet JF, Lightdale CJ, Zauber AG, et al. Preoperative staging of esophageal cancer:
Comparison of endoscopic US and dynamic CT. Radiology 1991;181:41925.
Catalano MF, Sivak MV Jr, Rice T, et al. Endosonographic features predictive of lymph node
metastasis. Gastrointest Endosc 1994;40:4426.
Hordijk ML, Zander H, van Blankenstein M, Tilanus HW. Influence of tumor stenosis on the
accuracy of endosonography in preoperative T staging of esophageal cancer. Endoscopy
1993;25:1715.
Glover JR, Sargeant IR, Bown SG, Lees WR. Non-optic endosonography in advanced
carcinoma of the esophagus. Gastrointest Endosc 1994;40:1948.
Maruta S, Tsukamoto Y, Niwa Y, et al. Evaluation of upper gastrointestinal tumors with a new
endoscopic ultrasound probe. Gastrointest Endosc 1994;40:6038.
Frank N, Grieshammer B, Zimmermann W. A new miniature ultrasonic probe for
gastrointestinal scanning: Feasibility and preliminary results. Endoscopy 1994;26:6038.
Chak A, Canto M, Gerdes H, et al. Prognosis of esophageal cancers preoperatively staged to
be locally invasive (T4) by endoscopic ultrasound (EUS): A multicenter retrospective cohort
Peters JH, Hoeft SF, Heimbucher J, et al. Selection of patients for curative or palliative
resection of esophageal cancer based on preoperative endoscopic ultrasonography. Arch
Surg 1994;129:5349.
Rice TW, Boyce GA, Sivak MV Jr, et al. Esophageal carcinoma: Esophageal ultrasound
assessment of preoperative chemotherapy. Ann Thorac Surg 1992;53:9727.
Tio TL, Blank LE, Wijers OB, et al. Staging and prognosis using endosonography in patients
with inoperable esophageal carcinoma treated with combined intraluminal and external
irradiation. Gastrointest Endosc 1994;40:30410.
Falk GW, Catalano MF, Sivak MV Jr, et al. Endosonography in the evaluation of patients with
Barrett's esophagus and high-grade dysplasia. Gastrointest Endosc 1994;40:20712.
55. Srivastava AK, Vanagunas A, Kamel P, Cooper R. Endoscopic ultrasound in the evaluation of
Barrett's esophagus: A preliminary report. Am J Gastroenterol 1994;89:21925.
56. Koga H, Iida M, Suekane H, et al. Rapidly growing esophageal leiomyosarcoma: Case report
and review of the literature. Abdom Imaging 1995;20:159.
57. Wegener M, Adamek RJ, Wedmann B, Pfaffenbach B. Endosonographically guided
fine-needle aspiration puncture of paraesophagogastric mass lesions: Preliminary results.
Endoscopy 1994;26:58691.
58. Tio TL, den Hartog Jager FC, Tytgat GNJ. Endoscopic ultrasonography for the evaluation of
smooth muscle tumors in the upper gastrointestinal tract. Gastrointest Endosc
1990;36:34250.
59. Tada S, Iida M, Yao T, et al. Granular cell tumor of the esophagus: Endoscopic
ultrasonographic demonstration and endoscopic removal. Am J Gastroenterol
1990;85:150711.
60. Caletti GC, Brocchi E, Ferrari A, et al. Guillotine needle biopsy as a supplement of
61.
Boyce GA, Sivak MV Jr, Rsch TR, et al. Evaluation of submucosal upper gastrointestinal
tract lesions [Abstract]. Gastrointest Endosc 1990;36:195.
62. Dancygier H, Classen M. How can we diagnose the depth of cancer invasion in the
esophagus?. Endoscopy 1986;18:1921.
64. Murata Y, Yoshida M, Akimoto S, et al. Evaluation of endoscopic ultrasonography for the
diagnosis of submucosal tumors of the esophagus. Surg Endosc 1988;2:518.
65. Ott DJ, Wolfman NT, Wu WC, et al. Endoscopic ultrasonography of benign esophageal cyst
simulating leiomyoma. J Clin Gastroenterol 1992;15:857.
66. Rommens J, Deviere J, Gevenois PA, et al. Intramural esophageal cyst: The role of
endoscopic ultrasonography. J Belge Radiol 1994;77:156.
67. Geller A, Wang KK, DiMagno EP. Diagnosis of foregut duplication cysts by endoscopic
ultrasonography. Gastroenterology 1995;109:83842.
68. Caletti GC, Brochii E, Baraldin M, et al. Assessment of portal hypertension by endoscopic
ultrasonography. Gastrointest Endosc 1990;36:S27.
69. Nishizono M, Haraguchi Y, Eto T, et al. Endoscopic ultrasonography using a 15/20 MHz probe
in a direct contact technique: Evaluation and application in esophageal and gastric varices.
Fukuoka Igaku Zasshi 1994;85:2515.
70. Tio TL, Kimmings N, Rauws E, et al. Endosonography of gastro-esophageal varices:
Evaluation and follow-up of 76 cases. Gastrointest Endosc 1995;42:14550.
71. Nakamura H, Endo M, Shimojuu K, et al. Esophageal varices evaluated by endoscopic
ultrasonography: Observation of collateral circulation during non-shunting operations. Surg
Endosc 1990;4:6974; discussion 75.
72. Deviere J, Dunham F, Rickaert F, et al. Endoscopic ultrasonography in achalasia.
73. Ziegler K, Sanft C, Friedrich M, et al. Endosonographic appearance of the esophagus in
achalasia. Endoscopy 1990;22:14.
74. Van Dam J, Falk GW, Sivak MV Jr, et al. Endosonographic evaluation of the patient with
achalasia: Appearance of the esophagus using the echoendoscope. Endoscopy
1995;27:18590.
75. Schder G, Isrighaus H, Kubale B, et al. Endoscopic ultrasonography of the mediastinum in
the diagnosis of bronchial carcinoma. Thorac Cardiovasc Surg 1991;39:299303.
76. Wiersema MJ, Chak A, Wiersema LM. Mediastinal histoplasmosis: Evaluation with
endosonography and endoscopic fine-needle aspiration biopsy. Gastrointest Endosc
1994;40:7881.
77. Miller LS, Liu JB, Klenn PJ, et al. Endoluminal ultrasonography of the distal esophagus in
systemic sclerosis. Gastroenterology 1993;105:319.
78. Taniguchi DK, Martin RW, Trowers EA, et al. Changes in esophageal wall layers during
motility: Measurements with a new miniature ultrasound suction device. Gastrointest Endosc
1993;39:14652.
79. Bolondi L, De Giorgio R, Santi V, et al. Primary non-Hodgkin's T-cell lymphoma of the
esophagus. A case with peculiar endoscopic ultrasonographic pattern. Dig Dis Sci
1990;35:142630.
80.
Takemoto T, Itoh T, Fukumoto Y, et al. Endoscopic ultrasonography in preoperative staging of
esophageal cancer. In Dancygier H, Classen M (eds). 5th International Symposium on
Endoscopic Ultrasonogaphy. Munich: Demeter Verlag (Z Gastroenterol Suppl) 1989;348.
81.
Schder G, Koch B, Seitz G, et al. Endosonographisches Staging beim
sophaguskarzinomEin prospektiver Vergleich mit herkmmlichen bildgebenden
Verfahren. In Henning H, Soehendra N (eds). Fortschritte der gastroenterologischen
Endoskopie. Munich: Demeter Verlag, 1990;801.
82. Sugimachi K, Ohno S, Fujishima H, et al. Endoscopic ultrasonographic detection of
carcinomatous invasion and of lymph nodes in the thoracic esophagus. Surgery
1990;107:36671.
83.
Vilgrain V, Mompoint D, Palazzo L, et al. Staging of esophageal carcinoma: Comparison of
results with endoscopic sonography and CT. AJR 1990;155:27781.
84. Rice TW, Boyce GA, Sivak MV Jr. Esophageal ultrasound and the preoperative staging of
carcinoma of the esophagus. J Thorac Cardiovasc Surg 1991;101:53644.
85. Fok M, Cheng SW, Wong J. Endosonography in patient selection for surgical treatment of
esophageal carcinoma. World J Surg 1992;16:1098103.
86. Heintz A, Wahl W, Mildenberger P, et al. Endosonographie beim sophaguskarcinom.
Ergebnisse einer klinischen Untersuchung and einer In-vitro-Analyse. Chirurg
1992;63:62933.
87.
Kalantzis N, Kallimanis G, Laoudi F, et al. Endoscopic ultrasonography and computed
tomography in preoperative (TNM) classification of esophageal carcinoma [Abstract].
88.
Nobre-Leito C, Santos AA, Miodes Correira J, et al. Esophageal carcinoma: Preoperative
staging with endosonography [Abstract}. Endoscopy 1992;24(suppl 1):379.
89. Rsch T, Lorenz R, Zenker K, et al. Local staging and assessment of resectability in
carcinoma of esophagus, stomach and duodenum by endoscopic ultrasonography.
90. Souquet JC, Napoleon B, Pujol B, et al. Endosonography-guided treatment of esophageal
carcinoma. Endoscopy 1992;24(suppl 1):3248.
91. Grimm H, Binmoeller KF, Hamper K, et al. Endosonography for preoperative locoregional
staging of esophageal and gastric cancer. Endoscopy 1993;25:22430.
92. Nattermann C, Dancygier H. Endosckopischer Ultraschall im properativen TN-staging des
sophaguskarzinoms. Eine vergleichends Studie zwischen Endosonographie and
Computertomographie. Ultraschall Med 1993;14:1005.
93.
Napolitano V, Allaria A, Amato G. Accuracy of endoscopic ultrasonography in the preoperative
staging of esophageal cancer [Abstract]. Endoscopy 1994;26:823.
94.
Grimm H, Maydeo A, Hamper K, et al. Results of endoscopic ultrasound and computed
tomography in preoperative staging of esophageal cancer: A prospective controlled study
[Abstract]. Gastrointest Endosc 1991;37:279.
95. Heintz A, Hhne U, Schweden F, Junginger T. Endosonographie versus Computer
tomographie bei der properativen Stadienbeurteiling von sophaguskarzinomen. Z
96.
Grimm H, Hamper K, Maas R, Soehendra N. Properatives Staging von sophagusUnd
Magenkarzinoma mit der Endosonographie. Vergleich mit der Computertomographie. In
Henning H, Soehendra N (eds). Forschritte der gastroenterologischen Endos-copie. Munich:
Demeter Verlag, 1990;749.
97. Van Dam J, Rice TW, Catalano MF, et al. High-grade malignant stricture is predictive of
esophageal tumor stage. Risks of endosonographic evaluation. Cancer 1993;71:29107.

98. Heyder N, Lux G. Malignant lesions of the upper gastrointestinal tract. Scand J Gastroenterol
Suppl 1986;123:4751.
99. Dancygier H, Classen M. Endoscopic ultrasonography in esophageal diseases. Gastrointest
Endosc 1989;35:2205.
Chapter 42 Palliation of Malignant Obstruction: Dilation and Peroral

Prosthesis
(3345)
(3346)
G. N. J. TYTGAT, M.D.
The prognosis of patients with esophageal cancer is generally poor, with an overall 5-year survival rate
of only 5%. By the time the diagnosis is made, most patients are beyond cure. Dysphagia,
regurgitation, and hypersalivation due to local obstruction or incessant coughing on the basis of a
tracheobronchoesophageal fistula become the most important distressing factors in the final stage of
the disease. As such patients usually have a short life expectancy, it is often desirable to avoid the
morbidity associated with surgery, radiotherapy, or chemotherapy. At the present time, management of
the vast majority of patients with esophageal cancer is strictly palliative. The principal goal is to
maintain a patent esophageal lumen. Other aims include avoidance of complications, pain relief, and
restoration of nutrition, as well as avoidance of prolonged hospitalization and high cost of therapy.
Because there is no chance of cure with endoscopic therapy, patients eligible for endoscopic
techniques should include those with unresectable, inoperable, or recurrent tumors. An ideal palliative
technique should be fast, painless, long-lasting, safe, and not expensive and should require only brief
hospitalization. Thus far, no single treatment fulfills all these characteristics. This overview covers the
current possibilities with dilation therapy and insertion of an esophageal prosthesis.
Esophageal Dilation
Many patients with dysphagia due to carcinoma can be managed entirely by esophageal dilation.
Initially, it is advisable to introduce the dilating device over a guidewire, especially when the remaining
lumen is narrow, tortuous, or eccentric or when there is a bronchoesophageal fistula. Several devices
are available for guided dilation.
Equipment and Technique

The dilating equipment most commonly used at pres-ent consists of the following types: polyvinyl
dilators (Savary-Gilliard type or American Endoscopy type [American Endoscopy, Mentor, OH]), metal
olives (Eder-Puestow type), hydrostatic balloons, KeyMed advanced dilators (KeyMed, Ltd., Southend,
Essex, England), and mercury-filled rubber bougies (Maloney or Hurst types).1,2(3347) The main
advantages of the Savary-type dilators are the easy passage through the pharynx and the gradual
dilation produced.
In principle, all malignancies should be dilated over a guidewire under combined
endoscopic/fluoroscopic control, especially if difficulty is anticipated.1(3348) If the malignancy can be
passed with a small-caliber endoscope, then the guidewire may be positioned under endoscopic
control alone. Very tight strictures may require passage of an angiographic catheter or an ultrathin
lumen finder of the type used for endoscopic biliary work.
For tight or tortuous malignant strictures, many clinicians prefer to use polyvinyl tapered dilators
passed over a guidewire under fluoroscopic control for initial dilation of the tumor. The marked flexibility
of the smaller-diameter Savary-type or American Endoscopy-type dilators limits their usefulness in
dilating tight and complex malignant strictures. Therefore, initial dilation of long, tight, and tortuous
malignant strictures may require dilation with the Eder-Puestow-type metal olives. The tapered tip of
the thermoplastic dilators (17 cm Savary; 12 cm American Endoscopy) makes passage through tight
strictures relatively easy. Care should be taken not to force the dilator tip onto the proximal end of the
spring tip of the guidewire; this may lock them together, necessitating removal of the entire apparatus.
The lubricated dilator is passed over the taut guidewire until it is well beyond the stricture. Back traction
must be maintained on the guidewire while its tip is kept stabilized in the stomach during passage of
the dilator. It is crucial that the tip of the guidewire remain in constant position within the stomach
during and between dilations. The position of the spring tip can be monitored fluoroscopically or by
checking that the length of the wire projecting from the patient's mouth has not changed.
Although there has been some enthusiasm for endoscopic balloon dilators, many clinicians have found
them less effective for dilating tight, irregular cancers. Indeed, success with through the scope (TTS)
hydrostatic balloon dilators in difficult strictures is often unpredictable. After lubrication, the
decompressed TTS hydrostatic balloon is passed via the endoscope into the stricture and positioned
with its midpoint in the narrowed area, under fluoroscopy or by guess. The Rigiflex balloon
(Microvasive Endoscopy, Boston Scientific Corp., Natick, MA) is passed over a guidewire under
fluoroscopic observation. If the malignant stricture is longer than the balloon, then dilation is performed
by successively repositioning and inflating the balloon along the full length of the stricture. Proper
balloon positioning may also be obtained by inflating the balloon distal to the stricture, pulling the
proximal end of the balloon against the distal end of the stricture to the point that the stricture resists
further withdrawal, and after deflating, pulling back the appropriate length to center the balloon in the
stricture. This maneuver may have to be repeated several times. Without fluoroscopic control, it is
almost impossible to be certain of proper balloon position and complete balloon distention in complex
strictures.
The Keymed advanced dilator (KAD) system consists of spindle-shaped bougie dilators made of
medical-grade silicone rubber mounted by a simple screw junction on a flexible steel shaft. They are
barium-impregnated for fluoroscopic visualization. Because of their narrow tip and tapering contour,
these bougies usually pass smoothly through strictures, and effective dilation is thereby
guaranteed.3(3349)
Mercury-filled rubber bougies that are inserted blindly can be used only to maintain a rather straight
lumen once the latter has been initially dilated over a guidewire. Bougienage with mercury-weighted
dilators is satisfactory, provided the dilation path is not too eccentric or tortuous.
Repeated dilations are usually required to maintain luminal patency.4,5(3350) The interval between
dilations varies considerably depending on the procedure being used. A luminal diameter of 12 to 13
mm will usually eliminate dysphagia if the narrowed segment is short. For tumors of great longitudinal
extension, dilation to a much larger diameter is usually necessary. For many patients, considerable
symptomatic improvement is obtainable by dilation.6,7(3351)
Dilation can be continued as needed during radiation therapy. No evidence demonstrates that properly
performed dilation carries additional risk when performed in patients who have received or are
receiving radiotherapy.6,8(3352) For example, Ng et al.8(3353) compared complication rates for
dilation of 61 patients with esophageal malignancies who had undergone prior radiotherapy and 126
patients who had similar tumors but no prior radiotherapy. There were no significant differences in
perforation rates per patient (6.5% for those previously treated by radiotherapy vs. 8% for the control
group) or procedure performed (3% radiotherapy compared with 4.7% in the control population).
Dilation can also be accomplished safely in patients with a bronchoesophageal fistula. In some patients
with bulky exophytic tumors, dilation may be ineffective because the tumor mass recollapses the lumen
after the procedure. In others, dilation may be excessively painful or may be required too often to justify
its continued use. This usually happens in patients with a markedly scirrhous cancer and those with a
recurrence after prior radiotherapy. Marked angulation may also render dilation excessively difficult. In
such circumstances, placement of an endoprosthesis may be more effective and acceptable. However,
the success of peroral prosthesis insertion depends to a great extent on adequate dilation of the
malignant stenosis.
The rate at which dilators of increasing size should be passed is debatable. Some authors consider it
safest to perform gradual dilations, using only a few successive dilators of increasing diameter in
several sessions extending over 5 to 10 days.9(3354) Others prefer more rapid dilation, the rate of
insertion for successive dilators being determined by the responsiveness of the malignant stenosis and
the degree of resistance to passage of each dilator. It is fair to state that the characteristics of the
dilation of obstructing malignancies often depend on trial and error. Just how much force may be used
remains utterly subjective and depends in large part on prior experience. It should always be realized
that dilating a carcinomatous growth consists in essence in "tearing" abnormal tissue apart. The extent
and depth of this tearing process determine the major risks of the procedure, bleeding and perforation.
Advantages of dilation include relative simplicity, low cost, widespread availability of equipment, short
procedure time, and relative safety. Furthermore, dilation is a useful if not essential adjunct to other
treatments such as prosthesis insertion and laser photoablation.
The main disadvantage of dilation therapy is that the relief of dysphagia is often of short duration.
Frequent dilations, up to once per week, may be needed to maintain reasonable patency in the
esophagus.
Despite many reports suggesting acceptable relief of dysphagia, there often is reluctance to offer this
palliative modality to patients with esophageal cancer, mainly because of fear of complications.
However, when performed by an experienced operator, the overall complication rate for dilation is
acceptable; this varies between 3 and 10% and compares favorably with other palliative modalities.
Peroral Endoprostheses
Indications and Contraindications
Insertion of an endoprosthesis is the ultimate palliative measure for a malignant stricture when the
possibilities for radiotherapy and surgery have been exhausted. Endoprosthesis insertion is indicated
when dilation becomes ineffective, too many procedures are required, dilation becomes too difficult for
the patient or physician, or the required frequency of dilations becomes unacceptable. Another
well-accepted indication is a malignant bronchoesophageal fistula that results in incessant coughing
and in aspiration pneumonia. Life expectancy should be at least 6 weeks, except for patients with
fistulas. Contraindications include location of the cancer within 2 cm of the pharyngoesophageal
sphincter; short life expectancy, excepting cases with a fistula; an uncooperative or unmotivated
patient; and circumstances in which there is no prospect of improving the quality of life, for example,
certain patients with cerebral metastases.
Relative contraindications are total luminal obstruction prohibiting the passage of a guidewire,
noncircumferential tumor growth prohibiting proper anchoring of the prosthesis, excessively soft and
necrotic lesions with poor anchoring qualities, profusely bleeding lesions, almost horizontal orientation
of the malignant lumen interfering with proper liquid and food passage through the prosthesis, and
multiangulated lesions that render dilation impossible.5,1012(3355)
Types of Endoprostheses
Custom-Made Endoprostheses
It may be preferable to construct a prosthesis in order to meet special requirements with regard to
length, stiffness, caliber, and other features, as determined by the characteristics of the malignant
lesion.1315(3356) Inexpensive polyvinyl tubing is useful for this purpose. Preferably this should have
an outside diameter of 15.7 mm, an inside diameter of 12.5 mm, and a wall thickness of 1.6 mm. The
length of the malignant segment is calculated in centimeters from endoscopic measurements. The
prosthesis should be 5 to 6 cm longer than the stenotic segment, that is, 2.5 to 3.0 cm longer than the
stenosis at each end (Figure 421). To make the tube less slippery, a spiral indentation may be
created by softening the prosthesis in hot mineral oil and threading a wire spiral along its length. Next,
the proximal end is widened into a funnel shape by heating the last 2.5 cm of the tube in mineral oil at
100 C for 30 to 60 sec. This heated end is then forced into a funneling device (e.g., a large glass
centrifuge tube or an obturator) to stretch this end of the tube to a diameter of about 25 mm. A
radiopaque line may be inserted in the everted ends of the flange and also added along the length of
the tube or a metal band may be incorporated in the phalanged portion as a radiopaque
marker.16(3357) A proximal funnel-shaped phalange is essential to minimize leakage of liquids around
the upper edge of the prosthesis and into fistulas. The opposite end of the tube is cut on a bevel, and
the rough edges are filed and smoothed. Retainer rings 5 to 8 mm in length are glued to the prosthesis
and filed in the desired direction to prevent migration of the device. To prevent displacement, a ring
may be placed at positions corresponding to the proximal and distal margins of the stenotic segment.
An additional retainer ring may be positioned at a point corresponding to a fistulous tract into the lung.
To prevent kinking or collapse of the prosthesis in the case of sharp angulation or excessive scirrhous
compression by the tumor, the wall of the prosthesis may be strengthened by adding a
plastic-embedded metal coil to the outside of the prosthesis.1012(3358)
(3359)Figure 421. Various steps during preparation of a custom made endoprosthesis. Left
to right, Original Tygon tubing, lateral spiral indentation, creation of a proximal funnel,
addition of proximal and distal radiopaque lines, addition of proximal and distal shoulders to
prevent migration, and addition of a plastic-embedded metal coil.
A section of polyvinyl tubing about 50 cm in length is prepared as a pusher (ramming) tube by
funneling one end as described previously. This will be used to drive the prosthesis into place. To
prevent invagination of the pusher tube funnel into the prosthesis during insertion, the margin of the
pusher funnel may be thickened by partially filling it in with rubber.
Commercially Available Endoprostheses
Several types of nonexpandable endoprostheses are also available commercially (Figure 422). The
following are most commonly used in clinical practice.
(3360)Figure 422. Commercially available prostheses. Left to right, Celestin, KeyMed,

Procter-Livingston, and Wilson Cook.
The Wilson Cook prosthesis (Wilson Cook Medical Inc., Winston-Salem, NC) is made of silicone,
reinforced with a metal spiral. It has an outer diameter of 16 mm and an inner diameter of 12 mm.
They are available in lengths from 4.4 to 16.4 cm.
The KeyMed-Atkinson prosthesis (KeyMed, Southend-on-Sea, Essex, UK) is a radiopaque silicone
rubber tube of 11.7-mm bore and 16-mm outside diameter with a nylon spiral incorporated in its wall to
prevent kinking. A shoulder is attached to the distal end of the tube to prevent migration. Prostheses
are available in lengths ranging from 14 to 19 cm.
The ESKA-Buess esophageal tube (ESKA, Lubeck, Germany) is made of silicone and has an
oval-shaped proximal funnel to avoid compression of the trachea. The distal funnel completely
collapses when passed through a stenosis. The tube is given stability by a metal spiral embedded
within the wall. Metal hooks integrated in the proximal funnel can be hooked by a snare.17(3361)
The Medoc-Celestin tube (Medoc, Ltd., Tetbury, England) is made of latex and contains a nylon spiral.
The outer diameter is 15 mm, and the inner diameter is 12 mm. A collapsible flange is added distally to
maintain position. This tube fits snugly against the esophageal wall because of its tulip-shaped end.
Tubes are available in lengths of 12.5, 15, and 21 cm. Latex may undergo deterioration when exposed
to hydrochloric acid, bile, and radiation.18(3362)
Self-expandable prostheses have been adapted for palliation of patients with esophageal cancer.
Made of stainless steel mesh and originally intended for biliary and vascular applications, this type of
stent is designed to overcome some of the disadvantages of the solid-type ("plastic") prosthesis, such
as perforation and dislodgment. These stents may expand to a maximum diameter of about 20 mm
when released in the stricture. Woven in the form of a tubular mesh from surgical-grade stainless alloy
filaments, these stents are pliable, self-expanding, flexible in the longitudinal axis, and when fully
extended, are from 6 to 10 cm long.19(3363) The stent is maintained in a compressed form, about 6
mm in diameter, by a doubled-over plastic membrane, the outer layer of which can be progressively
retracted with hydraulic assistance, thus permitting the prosthesis to expand once it has been placed in
proper position within the malignant stricture (Figure 423).
(3364)Figure 423. Malignant distal esophageal obstruction before (left) and after (right)
bypassing with an expandable metal stent.
A number of self-expanding stents of differing design are now available commercially (e.g., Z Stent,
Wilson Cook Medical Inc., Winston-Salem, NC; Wallstent, Schneider, Bulach, Switzerland; Ultraflex,
Microinvasive Endoscopy, Boston Scientific Corp., Natick, MA; Esophacoil, InStent, Eden Prairie, MN).
Although designs may vary, all are based on the same mechanical principles. It has not been
established that any one design is superior in terms of ease of introduction, palliation of dysphagia, or
stent-related complications. The most important modification of the self-expanding metallic stent
design has been the addition of a silicone membrane.20(3365) Fully expanded, metal stents have
internal diameters up to 18 mm and an implanted length of 10 to 12 cm. With the exception of the Z
Stent, self-expanding metal stents shorten in length during deployment.
Fistula Prostheses
It is often necessary to utilize a modified prosthesis to obtain complete occlusion of a large malignant
fistula. The Wilson Cook fistula prosthesis (Wilson Cook Medical Inc., Winston-Salem, NC)
incorporates a circumferential cuff composed of a foam rubber sponge within a silicone sheath. On
release of a vacuum in the covering sheath, the foam rubber within the ring expands to close the
fistula.21(3366) The vacuum is created using a small catheter; when the vacuum is induced, the cuff
shrinks to an outer diameter of 2.6 cm. After insertion, the vacuum is released and the cuff fills with air,
allowing the foam rubber-filled cuff to expand to a diameter of 4 cm. The addition of such a soft,
self-inflating foam cuff to the esophageal prosthesis provides the ability to palliate some patients with
respiratory fistulas that are technically difficult to manage (Figure 424).
(3367)Figure 424. Wilson Cook fistula prosthesis before (left) and after (right) expansion.
The ESKA-Buess fistula funnel (ESKA, Lubeck, Germany, supplied as a separate part, is wider in
diameter, and the point of luminal sealing lies higher above the stenosis, allowing fistula closure that
would not have been possible with the standard tube.
Introducing Devices
A prosthesis may be inserted over a small-caliber endoscope,14,22,23(3368) over an Eder-Puestow
dilator shaft,24(3369) over a Savary-Gilliard-type bougie, or with various introducing devices including
those offered by Wilson Cook, Inc., and KeyMed, Ltd.,25,26(3370) the Celestin EndoProsthesis
Introducer (Medoc, Ltd., Tetbury, England),27(3371) the Dumon-Gilliard EndoProsthesis Delivery
System (Wilson Cook Medical Inc., Winston-Salem, NC), and the Am-sterdam Metal Introducer
(Fujinon, Inc., The Netherlands) (Figure 425).28(3372)
(3373)Figure 425. Amsterdam metal prosthesis introducer. The distal end of the introducer
can be rendered progressively stiffer and straighter by turning the knob at the end of the
introducer. This facilitates bypassing tortuous and angulated lesions.
The Amsterdam Metal Introducer is inserted over a guidewire, and the stiffness and straightness of its
distal end can be gradually increased. This facilitates the insertion of a prosthesis through lesions that
are angulated owing to factors such as tortuous tumor growth, previous surgery, or severe
kyphoscoliosis.
The Nottingham-KeyMed introducing device (KeyMed, Ltd., Southend-on-Sea, Essex, UK) consists of
two stainless steel tubes; the outer tube terminates in a metal olive, and the inner tube has a
detachable flexible tip that allows a plastic expansion cup to be inserted. When the outer tube is
pushed forward on the inner, the metal olive is pushed into the plastic cup, causing it to expand and
grip the prosthesis at the inside of its distal end. The outer and inner tubes are then locked in this
position. This assembly is completed by a plastic pusher (rammer) tube which fits over the two steel
tubes, so that its distal end fits into the proximal funnel of the prosthesis. The whole assembly is
passed over a previously placed guidewire. When the funnel of the prosthesis rests on the proximal
edge of the malignant stricture, the steel tubes are unlocked and removed with the guidewire while the
rammer temporarily holds the endoprosthesis in place.
There has been an increasing interest in guiding a prosthesis over a 10.5-mm-diameter Savary-bougie
and inserting the prosthesis with the help of a Dumon-Gilliard prosthesis pushing tube. Alternatively,
one may use the Wilson Cook prosthesis introducer, which consists essentially of a compression
balloon catheter, that fixes the inner side of a prosthesis during insertion.
The self-expanding metal stent is loaded on a small-diameter delivery catheter or a metal
guidewire.19(3374)
Placement of the Endoprosthesis

It is essential that a thorough explanation of the procedure be given to the patient and the patient's
family before an endoprosthesis is placed.
The location of the proximal margin of the tumor, its length, and its luminal diameter must be
accurately determined. Roentgenograms are of assistance in demonstrating the length and
configuration of the tumor as well as the presence of a fistula, but endoscopic measurement most
reliably determines the distance from the incisor teeth (or mandibular ridge) to the proximal edge of the
tumor and the length of the cancerous segment.
The most essential step is dilation of the stricture to the appropriate diameter. Because the stenosis
initially may be rigid and very narrow, this usually requires passage of a guidewire through the lumen of
the tumor under endoscopic and, if necessary, fluoroscopic control (Figure 426). The number of
dilators of increasing diameter passed over the guidewire depends entirely on the degree of narrowing,
extent, and tortuosity of the lesion (see earlier). A luminal diameter of about 15 mm (46 to 52 French)
is usually required to accommodate a prosthesis. In general, soft, necrotic lesions require less dilation
in comparison with scirrhous, tortuous malignancies; the latter sometimes require dilation to 18 to 20
mm. For easily dilated lesions, one dilation procedure may suffice. For scirrhous and tortuous lesions,
usually not more than three or four dilators of sequentially increasing size are passed at each
procedure. As the process of dilation proceeds, a small-diameter endoscope should be passed to
examine the regions within and distal to the tumor to determine the proximity of the distal margin to the
cardia as well as exact distances from the teeth to the proximal and distal margins. The endoscope
itself serves as a convenient measuring device.
(3375)Figure 426. Left, Extensive esophageal carcinoma. Right, The guidewire is inserted
carefully in the remaining original lumen without provoking bleeding.
A measurement equal to that from the incisor teeth to the location point for the proximal funnel of the
prosthesis is marked on the pusher tube, measuring from its forward end backward. The pusher tube
and the prosthesis are lubricated well and positioned on an introducing device. As noted previously,
this could be a small-caliber endoscope, a bougie, a special introducing device, or an Eder-Puestow
dilator shaft.
Intravenous drugs should be administered for sedation. Meperidine (pethidine) plus diazepam, or
thalamonal (Innovara combination of 2.5 mg droperidol and 0.05 mg fentanyl) plus diazepam or
midazolam are satisfactory. A relatively deep level of sedation is usually required. General anesthesia
may be necessary and appropriate in some patients.
The introducing device carrying the endoprosthesis is passed over a guidewire under fluoroscopic
control. After the tip of the endoscope (or other introducing device) has been advanced well past the
distal margin of the tumor, the endoscope or the introducing device is held in a steady position while
the pusher tube is advanced (using the inserting device as a guide) to the predetermined point marked
on the pusher. The prosthesis should now be seated properly, with its proximal funnel located above
the proximal margin of the tumor. With the prosthesis in the proper position, the operator holds the
pusher tube steady as the introducing device is quickly removed. Rotation of the pusher tube then
disengages it from the prosthesis, whereupon the pusher is withdrawn. When using the
KeyMed-Atkinson system, it is particularly important to be certain that the endoprosthesis is not
displaced during withdrawal of the introducer.
If the lesion has been dilated adequately, placement of an endoprosthesis should require only a few
minutes. The endoscope is then inserted to check the position of the prosthesis. If it is seated too
deeply, the prosthesis can be grasped with a foreign body retrieval forceps, grasping forceps, or
polypectomy snare to gently pull it upward. A small-caliber endoscope that will pass through the
prosthesis should be used. Thereby, the position of the prosthesis, particularly the distal end relative to
the cardia, can be determined, and the presence of a patent channel to the stomach confirmed. Care
must be exercised to avoid dislodgment of the prosthesis on withdrawal of the endoscope. To avoid
this, the pusher tube may be left in place to stabilize the prosthesis during this maneuver.
It is mandatory to carefully suction all blood-tinged fluid from the oral cavity of the heavily sedated
patient during the procedure in order to avoid aspiration. If midazolam has been administered for
sedation, one may wish to antagonize its sedative effects by intravenous administration of flumazenil.
With pharmacologic reversal of conscious sedation, the patient is sufficiently awake shortly after the
procedure and is then capable of informing the operator of any untoward symptoms.
The patient should be examined after the procedure for the presence of air within the subcutaneous
tissues of the neck and presternal area, and a chest x-ray should be obtained within 1 to 2 hr to search
for evidence of a perforation. Initially, the patient is not permitted any oral intake of food or fluids. When
the patient is fully awake, luminal patency and function of the prosthesis are verified by obtaining a
roentgenographic contrast study; a water-soluble contrast material should be used initially. The
prosthesis must be viewed from every possible angle. If no leakage is evident, further views may be
obtained using barium as a contrast agent, as this will be more satisfactory for documentation of the
position and function of the prosthesis. When it is certain, by meticulous roentgenographic
examination, that leakage has been excluded, the patient is permitted to eat and drink.
An endoprosthesis can be placed anywhere in the esophagus except near the proximal esophageal
sphincter (Figure 427). When the funnel edge is located 2 cm or less from the cricopharyngeus, it is
usually not tolerated by the patient. However, with certain modifications, a prosthesis may be tolerated
by selected patients with tumors that involve the proximal sphincter (see later). An endoprosthesis can
be inserted in elderly and frail patients and those with marked kyphoscoliosis as well as patients with
limited respiratory function, provided the proximal funnel will not compress the upper respiratory tract. It
can be introduced in patients with tumor recurrence after surgical resection or radiotherapy and in
patients with obstruction owing to mediastinal spread of bronchial or other metastatic malignancy
(Figure 428). An endoprosthesis can also be used to prevent pulmonary complications in many
patients with esophagopulmonary fistulas (Figure 429).
(3376)Figure 427. Proximal funnel of an endoprosthesis (left) bypassing a huge esophageal

cancer that has been compressed by the tube (right).
(3377)Figure 428. Typical example of extensive exophytic malignancy in the midesophagus

(left), bypassed by an endoprosthesis (right).
(3378)Figure 429. Endoscopic appearance of a malignant bronchoesophageal fistula (at right

side of field).
It is usually not difficult to insert an endoprosthesis when a malignant stenosis is short and straight.
Because the proximal tumor shelf must be adequate to anchor the device, the preferred tumor
configuration is concentric rather than a longitudinal growth that occupies only a portion of the
circumference of the esophageal wall. If the cancer involves less than half the circumference, the
stenosis is usually not adequate to hold the prosthesis in place (Figure 4210). Great care must be
exercised in the presence of a malignant intramural sinus, because such tumors are usually very
necrotic and easily allow the formation of a false passage into the lesion during manipulation. In these
circumstances, the guidewire, once correctly positioned, should not be removed until the prosthesis is
finally inserted into the correct luminal axis (Figure 4211).
(3379)Figure 4210. Noncircumferential polypoid tumor that is unsuitable for palliation with
a prosthesis because of poor anchoring capabilities.
(3380)Figure 4211. Left, Extensive adenocarcinoma of the cardia, fixed to the descending
aorta with sinus tract formation. Right, Bypass of the lesion by a prosthesis.
A large hiatus hernia is often present in patients with esophageal cancer, but it is possible to properly
insert a prosthesis in most cases (Figure 4212). Depending on the extent of the tumor and the size of
the hernia, either a short endoprosthesis can be introduced, so that its distal end just enters the hernia,
or a long tube can be used, so that the hernia is bypassed and the distal end of the tube lies within the
stomach and below the diaphragm (Figure 4213). The latter is preferred in the majority of patients
because a short prosthesis may become occluded at its distal end by prolapsing mucosal folds.
(3381)Figure 4212. Malignancy of the cardia (left) involving the distal esophagus (arrow) in
a patient with a hiatus hernia, making insertion of the guidewire into the stomach difficult but
not impossible. The prosthesis bypasses the hernia (middle and right) and enters the stomach.
The upper rim of the endoprosthesis is clearly visible over the air pocket in the hernia pouch
(arrows, middle and right).
(3382)Figure 4213. Left, Massive esophageal cancer proximal to a large hiatus hernia. Right,
Both cancer and hiatus hernia are bypassed with a prosthesis.
When it has been established that displacement has not occurred and that the lumen is adequate, the
patient is ready to eat and can be discharged. Detailed instructions should be given concerning
management of the prosthesis and diet. In general, a regular diet is tolerated, provided dentition is
adequate. The patient must be instructed to eat only in a bolt-upright sitting position, to chew food
carefully, and to take copious drafts of fluid during and after meals. Obstruction of the prosthesis to a
minor degree may be relieved by drinking carbonated beverages, which seem by their gaseous action
to disrupt and eliminate accumulated debris.
When a prosthetic tube straddles the cardia, measures to control gastroesophageal reflux are
required: sleeping with the head of the bed elevated, antacids after meals, and H2-receptor or proton
pump blocking drug therapy at night if necessary. When the prosthesis is very long and broaches the
cardia, the patient should be advised that all meat should be finely chopped, crusts should be removed
from bread, and all vegetables should be strained in order to avoid episodes of blockage. These steps
should be taken in addition to the antireflux measures. Ingestion of large pills or capsules intact should
also be avoided.
Technical Problems and Limitations

It is usually not difficult to bypass a relatively straight tumor that is not exceptionally long with virtually
any type of prosthesis. No major problems are to be expected in the absence of severe angulation or
surgical deformity. Tumors in the midesophagus tend to have a relatively straight configuration,
whereas marked angulation is more likely to be present with those that involve the cardia. Insertion of a
prosthesis is the only possible method of palliation of a patient with a malignant broncho-esophageal
fistula (Figure 4214).5,28,29(3383) Satisfactory occlusion of fistulas can be achieved in the majority of
patients with this type of lesion.
(3384)Figure 4214. Left, Large bronchoesophageal fistula (arrow) is visible in the left aspect
of this x-ray film. Right, Occlusion of the fistula was achieved only after adding a wide
shoulder (arrow) to the prosthesis. The segment with the shoulder is situated just proximal to
the origin of the fistulous tract.
Difficulties with the introduction of an endoprosthesis can be expected when any of the following
conditions exist:
1. There is complete luminal obstruction (Figure 4215).
(3385)Figure 4215. A patient had total aphagia due to complete malignant obstruction of the
esophagus.
2. Tumor growth causes sharp angulation of the esophageal lumen (which may occur in cancer of
the distal esophagus and cardia or after surgical resection).
3. A concomitant sliding or paraesophageal hernia is present.
4. The tumor is unusually necrotic or excessively hard.
5. A fistula is present in the absence of appreciable luminal narrowing.
6. The origin of a fistula is located at the proximal or distal ends of a malignant stricture.
7. The malignancy approaches the proximal esoph-ageal sphincter.
Several of these problems can be resolved, provided appropriate adaptations of both prosthesis and
introducing device are effected. However, considerable levels of expertise and experience are required
before attempting to deal with these technical problems. In addition, the capability for construction of
an endoprosthesis to unique specifications is essential because a custom-made, specifically designed
device is frequently crucial to the solution of a particular technical problem (see "Custom-Made
Endoprostheses," earlier in this chapter). As a general rule, fluoroscopic in addition to endoscopic
guidance is usually essential in dealing with the technical problems outlined previously.
When obstruction is total, it may be possible to pass one of several types of tiny, flexible, atraumatic
lumen finders through the lesion. When this fails, a portion of the proximal aspect of the tumor may be
removed by laser photodestruction. Ideally, a radiologic contrast study will have been obtained prior to
total obstruction; if so, this provides valuable information on the location, longitudinal extent, tortuosity,
angulation, and eccentricity of the mass. On reopening the lumen, an atraumatic lumen finder is
passed, and then catheters of increasing diameter are inserted in sequence over this wire until a point
is reached at which a standard guidewire can be placed through the lumen. Metal olives or other
dilation devices can be introduced along this guide to achieve progressive dilation.
Sharp angulation of the lumen may occur as a result of tortuous tumor growth, previous surgery, prior
radiotherapy, severe kyphoscoliosis, or any combination of these. This may cause difficulty during
dilation and insertion of the prosthesis. To transmit the dilating force in the desired direction toward the
tip of the dilating device, and to facilitate insertion of the prosthesis, it is advisable to make use of
devices that can gradually stiffen and straighten the path for dilation and insertion and thereby reduce
the acuteness of the angulation. Once a prosthesis has been inserted, compression and kinking may
easily occur with severely angulated lesions. This hinders the passage of food and predisposes to food
impaction soon after insertion. Incorporation of a metal coil in the endoprosthesis prohibits kinking and
may eliminate this problem. This type of strengthened prosthesis is also preferable in the case of very
scirrhous tumors (Figure 4216).
(3386)Figure 4216. Left, Linitis plastica extending into the esophagus (arrows). Middle,
Compression and kinking (arrows) of a regular prosthesis due to angulation and the scirrhous
character of the lesion. Right, Correction of this problem using a prosthesis strengthened with
a metal coil.
When a fistulous tract begins at the proximal or distal end of a malignant stricture, it is not unusual that
a radiologic contrast study will demonstrate persistent leakage despite apparent satisfactory positioning
of a prosthesis. In the case of a proximal fistula, the insertion of a device with an extra-wide funnel (up
to 30 mm in diameter) and a wide proximal shoulder may arrest leakage in some patients. It is also
possible to use a tube that has been wrapped round with several layers of a polyvinyl alcohol sponge
that swells once it is in place. Insertion of a cuffed prosthesis (Wilson Cook) is another possible
alternative for sealing of a fistulous tract (see description earlier) (Figure 4217). An additional
alternative for management of a malignant tumor with a fistula tract is the use of a Buess fistula tube.
(3387)Figure 4217. Malignant bronchoesophageal fistula bypassed by a Wilson Cook fistula

tube. Note the air in the foam cuff surrounding the Wilson Cook prosthesis.
Occasionally, a fistula will not be occluded despite correct deployment of a cuffed prosthesis. It may
happen that radiographic examination does not visualize the air-filled cuff surrounding the prosthesis.
Presumably, fluid and exudate block the air-inlet channel after removing the insufflation wick. If at all
possible, the level of the air channel should be above the fluid level in the esophagus when removing
the connecting catheter. Full insufflation of the cuff should be confirmed fluoroscopically prior to
removing the insufflation wick.
Other attempted techniques for closure of malignant fistula tracts include injection of rapidly
polymerizing amino acid solutions to solidify fibrin in the fistulous tract, injection of soluble collagen
particles at the edges of the prosthesis funnel to obtain proper sealing of the wall around the proximal
end of the prosthesis, and anchoring the tube with an adhesive.30(3388)
An extra-large prosthesis with distal widening may be used to arrest distal leakage. Fistulas with origins
very close to the proximal sphincter are difficult, albeit not impossible, to treat with an endoprosthesis.
It is exceedingly difficult to manage a large fistula arising in a tumor that does not cause narrowing
sufficient to anchor a prosthesis. Possible solutions include the use of a prosthesis with a wider
proximal funnel or to glue an extra retainer ring to a custom-made tube in an attempt to anchor the
device and prevent leakage (Figure 4218).
(3389)Figure 4218. Left, Failure of a commercial prosthesis to seal a large

bronchoesophageal fistula. Right, Adequate occlusion of the fistula with an adapted prosthesis
with a widened funnel.
Migration of an endoprosthesis may be expected, especially with a concomitant hiatus hernia, with
tumors approaching or bridging the gastroesophageal junction, and in the case of extrinsic but
eccentric compression. To obviate migration, anchoring devices are usually required at variable points
along the prosthesis, depending on tumor configuration. Proximal dislodgment because of
diaphragmatic movement may be prevented by adding a distal rim to the prosthesis. Distal migration,
which can occur with a soft tumor or with eccentric compression due to invading pulmonary cancer,
may be prevented by adding an extra flange at the proximal end just below the funnel. To prevent distal
dislodgment, a 7-French, 60-cm-long polyethylene catheter can also be attached to the flange of a
prosthesis, rerouted through the patient's nose, and attached to the ear in the form of a loop.31(3390)
These anchoring catheters are usually well tolerated.
The position of a prosthesis in cases of cancer of the cardia is occasionally so oblique that the funnel
opening is partially occluded, thereby preventing proper emptying. Gluing an extra shoulder some
distance beneath the funnel may align the latter with the esophageal axis.
Esophageal involvement by lung cancer often creates major problems.32(3391) This condition
represents somewhat less than 15% of all indications for endoscopic palliation. The overall prognosis
is even worse than that of primary esophageal cancer. Furthermore, the risk of perforation associated
with palliative manipulations is higher by comparison. The mechanisms responsible for this increase
are poorly understood but might be explained partially by the fact that extrinsic strictures due to primary
pulmonary cancer are often asymmetric and noncircumferential. Another possible explanation may be
the firm consistency of such extrinsic tumors, which makes dilation more hazardous.
The success rate for closure of a malignant fistula is lower if the primary tumor is pulmonary rather
than esophageal. When a fistula develops secondary to invasion of the esophageal wall by lung
cancer, there is usually little or no tumor-related esophageal stenosis. Therefore, the prosthesis cannot
be adequately anchored and may fail to seal the space between the stent and the fistulous orifice.
Another complication that is fairly specific to extrinsic esophageal involvement is tracheobronchial
obstruction owing to pressure exerted on the tumor by the prosthesis. A careful bronchoscopy to
assess the risk for upper airway obstruction is therefore mandatory prior to esophageal intubation.
If a cancer is located 2 cm or less from the cricopharyngeus, a foreign body sensation or stridor due to
laryngeal compression, or both, can be expected if a standard prosthesis is employed. This problem
can be circumvented only by inserting a prosthesis with a small, short funnel, or one without a
funneling section, provided that this does not cause airway compression.31(3392) The operator must
be prepared to remove the prosthetic device immediately if acute stridor occurs after insertion. Astute
inspection of lateral x-ray films may alert the endoscopist to the possibility of compression if the tumor
and airway are in close proximity, and this complication may thereby be avoided.
Goldschmid et al.33(3393) treated 10 patients with stenosis of the cervical esophagus close to or
involving the upper sphincter. A custom-made prosthesis was successfully placed in 8 patients, 5 of
whom had a tracheoesophageal fistula. A custom-designed prosthesis could not be placed in 2
patients; 2 others complained of a foreign body sensation, but this was mild and tolerable. Spinelli et
al.34(3394) attempted endoscopic intubation in 8 patients with cancer of the cervical esophagus. All
patients tolerated the prosthesis well, experienced no respiratory impairment, and reestablished
nutrition via the oral intake of food. Loizou et al.35(3395) inserted modified Celestin endoprostheses in
8 patients with malignant tumors located within 2 cm of the cricopharyngeus, including 3 with
tracheoesophageal fistulas. Intubation was successful in all cases, with long-term palliation of
dysphagia being successful in 6 patients. A foreign body or other sensation was noted by 2 patients,
but this was characterized as being of mild to moderate severity. Two patients experienced tracheal
aspiration; distal migration of the tube occurred in 2 patients.
A tumor that is within 2 cm of or even proximal to the cricopharyngeal sphincter need not be
considered an absolute contraindication to placement of an endoprosthesis. Although the sensation of
the presence of a foreign body may be absent or minimum, stent migration is common. Frequent
instrumentation, surgical manipulation, and marked involvement of the proximal esophagus by tumor
decrease the likelihood that the foreign body sensation will be severe. In addition, the use of a
tailor-made prosthesis of appropriate diameter (usually less than 12 mm in diameter) with a pliable
proximal funnel may decrease the patient's awareness of the presence of a foreign body. When a
patient is unable to tolerate even a small-caliber tube, it may be possible to ablate tumor tissue by laser
photoablation, or alternatively, the patient may be treated with intermittent dilation or radiotherapy.
Tumors with multiple stenotic segments that are sharply angulated one to another usually cannot be
dilated properly with existing equipment. An attempt may be made to guide metal wires over an
introducer that can be straightened and strengthened.
Results
Nonsurgical insertion of a prosthesis over a flexible mercury-weighted bougie was reestablished by
Boyce and Palmer in 1975. Also at that time, Tytgat et al.14(3396) stressed the importance of guided
insertion over an endoscope. Atkinson et al.26(3397) and Clestin and Campbell27(3398) developed
methods for guided insertion that are similar.
There have been numerous studies of the usefulness of an esophageal
prosthesis.5,17,23,24,29,32(3399),3659(3400) Technical success rates for insertion of various
"plastic-type" endoprostheses are given in Table 421.
Selected Published Series of Patients With Obstructing Malignant

Tumors of the Esophagus and Gastric Cardia Treated by Insertion of an
Endoprosthesis
TABLE 421
AUTHORS
YEAR
PATIENTS (NO.)*
SUCCESS (%)
PROSTHESIS
Weisel et al.44
1959
103
O'Conner et al.45
Palmer46
1963
388
97
Polyethylene
1973
75
100
Polyvinyl
Polyethylene
1977
49
94
Procter-Livingstone
Hegarty et al.47
Selected Published Series of Patients With Obstructing Malignant

Tumors of the Esophagus and Gastric Cardia Treated by Insertion of an
Endoprosthesis
TABLE 421
AUTHORS
YEAR
PATIENTS (NO.)*
1977
108
1979
200
1980
35
Clestin et al.23
Jones et al.50
1980
91
1981
55
87
Celestin
Soehendra51
Ogilvie et al.52
1981
60
100
Celestin
1981
118
85
KeyMed
Balmes et al.53
1982
78
91
Haring
van Blankenstein et al.54

Boyce24
1982
119
92
KeyMed
1982
41
100
Polyvinyl
Watson55
Lux et al.56
1982
32
91
1983
60
82
KeyMed/polyvinyl
Rose and Smith57

Fugger et al.58
1983
100
93
Atkinson/Celestin
1990
110
91
KeyMed
Spinelli et al.59
1994
76
100
Atkinson/Celestin/Wilson
Cook
Kairaluoma et al.48
Den Hartog Jager
al.29
Bergerault et al.49
et
SUCCESS (%)
PROSTHESIS
Celestin
96.5
Polyvinyl
Celestin
* Number of patients with esophageal and gastric cardia malignancy.

Percentage of patients successfully intubated.
Includes only tumors of distal esophagus and gastric cardia.
Loizou et al.60(3401) attempted to measure the quality of life for a small number of patients who
underwent either laser photoablation or prosthesis placement because of dysphagia due to malignant
tumors. Initially, a significant improvement correlated with relief of dysphagia. However, quality of life
decreased as the disease progressed. In a retrospective review of cases, Oliver et al.61(3402) found
no difference in survival of patients with squamous cell carcinoma of the esophagus not suitable for
surgery who were treated by radiotherapy compared with those who underwent only insertion of a
prosthesis.
Comparison with Laser Photoablation
Several studies have compared laser photoablation and prosthesis insertion in the palliative treatment
of esophageal cancer.41,6266(3403) Unfortunately, most of the currently available data have not
been derived from prospective randomized trials.
Barr et al.62(3404) compared laser therapy performed on a random basis with laser therapy followed
by intubation and concluded that both approaches were equally effective in relieving dysphagia and
maintaining quality of life. Because 21% of patients in this study had complete esophageal occlusion,
endoscopic laser therapy (ELT) was performed prior to stent placement to allow passage of a
guidewire and to shape the tumor to accept a stent. This usually required a single ELT session. If a
patient in the laser-treated group required ELT more often than monthly, treatment was regarded as
having failed. Laser treatment was also regarded as failed if dysphagia developed due to extrinsic
compression by the tumor or the occurrence of a fibrous stricture. Patients in whom laser treatment
was deemed a failure underwent stent placement. The ability to swallow, before therapy and during the
remainder of life, was graded using a 0 to 4 scale in which 0 indicated normal swallowing and 4
signified total dysphagia. There was no significant difference in dysphagia grade before or after
treatment, although there was greater fluctuation in dysphagia grade in patients treated by laser alone.
Overall, there was no significant difference in the quality of life between the two groups. The
complication rate was significantly higher in patients who underwent insertion of a prosthesis (40%)
compared with those treated by laser alone (10%), but most prosthesis-related complications were
readily managed by conservative measures.
Richter et al.41(3405) compared the efficacy of prosthesis insertion with laser photoablation in a
retrospective study of 53 patients with obstructing gastroesophageal cancers. Only 12 patients
underwent initial laser therapy. Both approaches provided similar relief of dysphagia, although that
afforded by insertion of a prosthesis was more lasting. Patients in whom a prosthesis had been placed
required less intervention for the management of recurrent dysphagia and consumed fewer resources.
Alderson and Wright63(3406) also compared endoscopic intubation and laser in a prospective
randomized trial that included 40 patients with malignant obstruction involving the middle or distal third
of the esophagus. Technical success was equal for both groups. Functional success was uninfluenced
by degree of dysphagia prior to treatment or location of the tumor. However, the functional result
achieved by laser photoablation was significantly better if the circumferential length of the tumor was 4
cm or less. The internal diameter of the stents used in this study was 11 mm; better results might have
been achieved with stents of larger inner diameter. For malignant stenoses more than 4 cm in length,
Alderson and Wright concluded that insertion of a prosthesis at a single procedure was the better
approach of management.
Buset et al.64(3407) compared outcomes for endoscopic management of obstructive
gastroesophageal cancer in a nonrandomized study that included 28 patients treated by laser
photoablation and 116 patients who underwent prosthesis placement. Technical success rates were
comparable at 100% and 95%, respectively. The morbidity associated with laser therapy was 3.6%,
whereas that related to prosthesis placement was 13.8%. The mortality rate for prosthesis insertion
was 4.3%; no deaths resulted from laser treatment. These investigators regarded intubation as
preferable in patients with fistulas, extensive strictures, and very long tumor stenoses; laser
photoablation was considered more satisfactory for tumors that involve the proximal esophagus and
those that have an asymmetric or polypoid configuration that does not lend itself to anchoring of a
prosthesis.
Loizou et al.65(3408) compared laser therapy with intubation for palliation of malignant dysphagia in a
prospective, nonrandomized, two-center trial. The results in 43 patients treated by
neodymium:yttrium-aluminum-garnet (Nd:YAG) laser photoablation in one center were compared with
those for 30 individuals who underwent endoscopic intubation in another group. The two treatment
groups were said to be comparable with regard to mean age and severity of dysphagia pretreatment as
well as the position, length, and histologic type of the tumor. Functional success was defined as
improvement in dysphagia by at least one grade. Early (2-week) and late functional success was
significantly better in the stent group (100% vs. 77% for early success; 90% vs. 63% for late success).
Initial and long-term improvement in dysphagia grade were essentially the same for patients with
tumors in the midesophagus. However, for tumors that crossed the gastric cardia, the immediate and
long-term results were significantly better in the prosthesis group (100% vs. 59% for early success;
92% vs. 52% for late success). In patients with longer survival, palliation of dysphagia was significantly
better in those who had undergone laser therapy, but these patients required significantly more
treatment sessions and more days spent in hospital. The perforation rate was significantly higher
(13%) for patients who underwent placement of a prosthesis compared with those who had laser
therapy (2%). Mean survival was slightly longer in the laser group (6 months vs. 5 months) than in the
intubation group, but this difference was not statistically significant.
Hahl et al.66(3409) retrospectively compared the results of endoscopic laser therapy and insertion of a
prosthesis. The technical success rate was 87% in 69 patients who underwent laser therapy versus
89% for prosthesis insertion in 27 patients. Both methods were effective in palliating dysphagia, but
mortality and complication rates were lower for patients who underwent laser therapy. In this group, a
complication occurred in 8.7% and none was fatal. By contrast, complications occurred in 48% of
patients who underwent prosthesis placement; the mortality for complications related to prosthesis
insertion was 11%. None of the prosthesis-treated patients survived 1 year; 1-year survival was 12%
for laser-treated patients.
In a prospective, randomized study involving 40 patients, Carter et al.67(3410) found that improvement
in dysphagia grade was significantly better with laser therapy compared with stent insertion. Although
quality of life was not assessed in this study, median weight loss and time spent in hospital were
significantly less in the laser-treated group. No perforations occurred among patients who underwent
stent insertion; 3 perforations were noted in the laser-treated group (1 fatal). Carter et al.67(3411)
concluded that the two modalities should be considered complementary.
Patients were randomized in the study of Reed et al.68(3412) to one of three treatment groups: stent
insertion, stent plus external beam radiotherapy, and laser therapy plus radiation. Among the three
treatment groups, there were no significant differences in major outcome variables except for a higher
complication rate for patients who underwent stent placement. However, numbers of patients in each
treatment group were small.
Self-Expandable Metal Stents
There has been a remarkable increase in reported series of patients with obstructing malignancies of
the esophagus or cardia who have undergone placement of a self-expanding metal
prosthesis20,6983(3413) (Table 422).
TABLE 42-2 Selected Published Series of Patients With Dysphagia Due to Obstructing Malignant
Tumors Who Underwent Placement of an Expandable Metal Prosthesis (Part i of ii )
YEAR
PATIENTS
(NO.)
PERFORATION
NECROSIS/
BLEEDING
TUMOR
IN/OVERGROWTH
Schaer et al.20
1992
69
1992
Neuhaus et al.
1992
10
84
1993
40
86
1993
21
AUTHORS
Bethge et al.
71
Cwikiel et al.
Knyrim et al.
1994
32
79
1994
19
81
1994
14
1994
119
13
1994
23
1995
28
1995
114
75/11
88
Wu et al.
Wagner et al.
Raijman et al.
82
Song et al.
83
Ell et al.
73
Ell et al.
74
Grund et al.
Tumors Who Underwent Placement of an Expandable Metal Prosthesis (Part i of ii )
AUTHORS
YEAR
PATIENTS
(NO.)
De Palma et al.76
1995
32
PERFORATION
NECROSIS/
BLEEDING
TUMOR
IN/OVERGROWTH
5
Tumors Who Underwent Placement of an Expandable Metal Prosthesis (Part ii of ii )
MIGRATION
FOOD
IMPACTION
PAIN
MORTALITY (%)
Schaer et al.20
Z-stent
69
Wallstent
Neuhaus et al.
Wallstent
84
Nitinol
86
Wallstent
AUTHORS
Bethge et al.
71
Cwikiel et al.
Knyrim et al.
STENT TYPE
6.3
79
81
7.14
Gianturco
88
Wu et al.
Wagner et al.
Raijman et al.
82
Song et al.
12
Gianturco-Rosch-Z
Nitinol
Ultraflex Nitinol
83
Wallstent
73
11
Gianturco-Z
Grund et al.
Ultraflex
De Palma et al.76
Ultraflex
Ell et al.
Ell et al.
74
Expandable metal stents have several advantages and some limitations compared with the solid
("plastic") prosthesis. From a technical standpoint, they are easier to insert and have been implanted
under endoscopic as well as fluoroscopic guidance.72,82,84(3414) High success rates for deployment
have been reported in all published studies (Figure 4219). Generally, a lesser degree of dilation of the
tumor is required prior to placement, although at least partial dilation is necessary in the majority of
cases.74(3415) If the tumor is especially rigid, expansion of the metal prosthesis may be insufficient to
provide adequate palliation of dysphagia. In such cases, some investigators have resorted to using
hydrostatic balloon dilators to increase the inner diameter of the prosthesis.74,81(3416) Bethge et
al.85(3417) found in autopsy specimens that uncovered (see later) Wallstents became incorporated
into normal tissue proximal and distal to a tumor, the luminal aspect of the stent being covered by a
"collagenous reactive layer" that did not encroach on the lumen. Stents were also found incorporated
into malignant tissue.
(3418)Figure 4219. Endoscopic appearance of an expandable metal stent bypassing a large

esophageal cancer.
The major problem with the metal stent has been obstruction owing to inward growth of tumor through
the openings in the wire mesh. This complication, which occurs in from 0 to 66% of patients or about
27% of cases on average (see Table 422), can usually be managed by various endoscopic methods
including laser photoablation and electrocoagulation. The reported frequency of obstruction owing to
ingrowth of tumor is in part related to patient survival. The frequency is usually relatively low in studies
in which follow-up is of short duration, whereas higher rates are noted in long-term follow-up studies
that include larger numbers of surviving pa-tients. Membrane-coated metal stents have been
developed in an attempt to prevent tumor ingrowth. Presumably, this modification will reduce the
incidence of tumor ingrowth, although this has not been demonstrated in a controlled trial. It is evident
that some tumors can still disrupt the covering membrane and encroach on the esophageal
lumen.20,70(3419) Although membrane-coated devices appear to reduce ingrowth of tumor, the
frequency of stent migration appears to be increased compared with that associated with noncoated
devices.74,83(3420)
Complications, especially those associated with device implantation, appear to be less frequent in
patients undergoing placement of a metal stent. Knyrim et al.86(3421) reported the results of a trial in
which patients with obstructing tumors were randomized to undergo implantation of a metal
expandable prosthesis or a rigid "plastic" prosthesis (21 patients in each group). There was a
statistically significant difference in complications; none occurred in association with implantation of the
metal prosthesis as opposed to 9 complications (43%) among the patients who underwent insertion of
a plastic prosthesis including 3 that were fatal (14% procedure-related mortality). Fatal complications
included 2 perforations and 1 episode of aspiration pneumonia. In this study, general anesthesia was
employed during insertion of plastic prostheses. It has also been suggested that dilation of tumors in
patients randomized to receive a plastic prosthesis (to 20 mm in diameter) was unnecessarily
aggressive.87(3422) Some patients experience severe retrosternal or epigastric pain after implantation
of a metal stent; this may become persistent in some cases. As with the plastic prosthesis,
complications related to gastroesophageal reflux may occur in patients with implanted metal stents.
Food impaction may also occur, but usually this may be corrected by various endoscopic maneuvers.
Other untoward events include a foreign body sensation and formation of a pouch at the proximal edge
of the prosthesis.
A significant improvement in the ability to swallow has been noted in most but not all patients who
undergo placement of a metal expandable prosthesis. However, no evidence indicates that the
expandable metal prosthesis provides better palliation of dysphagia or greater improvement in the
quality of life compared with the plastic prosthesis. Restoration of the ability to swallow was equal in the
prospective randomized trial of Knyrim et al.86(3423) Dysphagia recurred in equal numbers of patients
in this study; prosthesis migration was more commonly the cause with a rigid plastic device, whereas
tumor ingrowth or overgrowth was the usual cause with a metal stent. Although the number of required
interventions was slightly greater in patients with plastic stents, this difference was not statistically
significant and the rate of reintervention was about equal for both groups. Grund et al.74(3424) in a
study of 114 patients encountered tumor ingrowth in 66% of cases; this occurred at a mean of 7 weeks
following stent placement.
The cost of a metal prosthesis in terms of the actual purchase price of the device itself is substantially
higher than that of a plastic prosthesis. In the trial of Knyrim et al.,86(3425) the overall cost of
treatment was less for patients with metal stents because of costs related to management of
complications in patients who undergo implantation of a plastic stent and fewer patient days in hospital
after placement of a metal stent. If further data confirm a low rate of complications, it may be possible
to insert metal expandable stents in selected patients without hospitalization and thereby reduce the
cost of management.
The initial prototypes of expandable metal stents were not designed for occlusion of malignant fistulas.
Later designs incorporated covering membranes consisting of several types of material including
silicone and polyure-thane. These "covered" metal stents have been implanted in patients with
fistulas.73,75,78,8891(3426) Although the number of patients with malignant fistulas who have been
treated by implantation of these devices is small, the success rates reported to date for occlusion of
fistulas have been uniformly high. In some cases, tracheobronchial stents have been placed prior to
insertion of the esophageal prosthesis.89(3427)
Complications
The overall estimated complication for insertion of a plastic-type esophageal prosthesis based on
reported series is between 10 and 30% (Table 423).23,24,29,4459(3428) Comparisons of
procedure-related mortalities in populations with esophageal cancer that have a high spontaneous
mortality are of uncertain relevance because procedure-related mortality depends to a large extent on
the underlying condition, timing of the introduction of the prosthesis, the degree of cachexia, and
various other factors. When the percentage of patients with broncho-esophageal fistulas or extensive
previous surgery or radiotherapy is high, a further increase in the complication rate is to be expected.
In the study of Ng et al.,8(3429) the complication rate associated with insertion of a prosthesis in 29
patients with esophageal malignancies who had undergone previous radiotherapy was compared with
that for prosthesis placement in 99 similar patients who had no prior x-ray treatment. The frequency of
perforation was 4% in both groups, although migration of the prosthesis after placement was
significantly more frequent in the prior radiotherapy group (21% vs. 3%). However, the data presented
by Ng et al.8(3430) were not collected in a prospective, randomized trial. Previous surgery and
radiotherapy usually create altered anatomic relationships, rendering dilation and intubation more
difficult and therefore more hazardous. If, on the other hand, only patients with rather straight, not
excessively long, nonirradiated esophageal cancers are selected for intubation, the complication rate
will be low. Moreover, it is sometimes difficult to decide the extent to which the mortality is
procedure-related or attributable to the underlying condition. The effort expended in searching for and
treating complications is related to the general condition of the patient, in addition to the size and
spread of the tumor, and these factors probably bias the collection of data. Unless all of these factors
are standardized, comparisons between studies and with corresponding surgical series are probably
irrelevant, the existing data merely giving an impression of the magnitude of the problem.
TABLE 423
Complications in Selected Published Series of Patients With Obstructing Malignan
Gastric Cardia
YEAR
PATIENTS
(NO.)*
PERFORATION
(%)
Weisel et al.44
1959
103
16
O'Conner et al.45
Palmer46
1963
388
1.2
1973
75
1.3
Hegarty et al.47
1977
49
Kairaluoma et al.48
Den Hartog Jager et
al.29
1977
108
1979
200
Bergerault et al.49
Clestin et al.23
1980
35
1980
91
AUTHORS
BLEEDING
(%)
NECROSIS/
FISTULA (%)
1
0
8
1.5
8.5
5.5
OBSTRUCT
TABLE 423
Complications in Selected Published Series of Patients With Obstructing Malignan
Gastric Cardia
YEAR
PATIENTS
(NO.)*
PERFORATION
(%)
Jones et al.50
Soehendra51
1981
55
8.3
1981
60
10
Ogilvie et al.52
Balmes et al.53
1981
118
11.8
1982
78
van Blankenstein et al.54

Boyce24
1982
119
10.4
7.3
1982
41
Watson55
Lux et al.56
1982
32
9.3
1983
60
11
Rose and Smith57

Fugger et al.58
1983
100
1990
110
AUTHORS
BLEEDING
(%)
NECROSIS/
FISTULA (%)
5.1
6.4
2.4
2.2
5
1994
2.3
Spinelli et al.59
76
* Number of patients with esophageal and gastric cardia malignancy.
Various causes including tumor overgrowth, peptic stricture, and food bolus.
Includes only tumors of distal esophagus and gastric cardia.
The insertion of a plastic endoprosthesis is usually well tolerated and has an acceptable immediate
mortality in comparison with surgical alternatives. The most common complications are perforation,
dislocation, tumor overgrowth, stricturing due to reflux esophagitis, pressure necrosis, and blockage by
food.
Perforation
Perforation is the most important life-threatening complication. Predisposing factors are previous
radiotherapy, acute angulation (especially in case of extensive tumor involvement of stomach or
cardia), and previous surgery. After any insertion procedure, the patient should be examined
endoscopically to confirm that a satisfactory lumen has been established and to reinspect the
esophagus and pharynx for signs of perforation. Thereafter, the patient should be examined for the
presence of subcutaneous air in the tissues of the neck and shoulders, and plain radiographs should
be obtained of the chest and neck. Most patients with a perforation will have radio-graphically
demonstrable air in the neck or mediastinum because air is insufflated under positive pressure during
esophagoscopy. If there is no evidence of perforation by endoscopic and physical examinations,
contrast radiography is performed once the patient has fully recovered from the effects of sedative and
analgesic drugs and is no longer at risk for aspiration of the contrast material.28,9294(3431) Initially,
a water-soluble medium is used; this may be followed by the use of barium if no leak is demonstrated.
Occasionally, a small tear or leak will be detected only by barium swallow.
The vast majority of the perforations should be suspected and recognized immediately, either
endoscopically or clinically. Subcutaneous emphysema develops rapidly. A pneumomediastinum
should be readily evident roentgenographically as well as free air below the diaphragm, especially if the
tear is large and the unwary endoscopist has insufflated large amounts of air during the procedure.
Occasionally, the first evidence of perforation is the appearance of a small amount of pleural fluid on
chest x-ray.
OBSTRUCT
The most appropriate therapy for perforation depends on the interval of time elapsed between the
procedure and the detection of the complication. If a perforation occurs during the dilation or insertion
phase of the procedure, and if the narrowed segment has been sufficiently dilated, then one may wish
to complete the insertion procedure, aiming to seal the perforation with the prosthesis, hoping thereby
to prevent mediastinal contamination. Alternatively, one may prefer to institute conservative treatment
for 7 to 10 days and then resume the insertion procedure after the tear has sealed. Conservative
treatment consists of elimination of oral food intake; adequate aspiration below, at, and proximal to the
perforation site; and systemic administration of antibiotics to cover the broad spectrum of potential
bacterial patho-gens. Some clinicians use nasopharyngeal suction only for those patients with
pharyngeal tears and not for those with perforations at other sites.
Blind insertion of a sump tube into the esophagus for aspiration is not advisable because it may enter
the perforation. Rather, a guidewire should be positioned in the correct lumen under endoscopic
control. In doing this, air insufflation via the endoscope must be avoided. Using the wire as a guide, a
sump tube can be placed in correct position distal to the site of perforation. After removal of the
guidewire, the proximal end of the sump tube may be rerouted from the throat to the nasopharynx and
out one nostril. Nutrition may be maintained either by nasogastric feeding using a fine-bore tube
passed through the stenosis or through the prosthesis into the antrum or jejunum or by intravenous
alimentation. For the purpose of feeding, it is sometimes desirable to glue a second tube to the
aspirating tube. The distal tip of the second tube should be positioned at the ligament of Treitz (Figure
4220).
(3432)Figure 4220. In a patient with iatrogenic perforation (left), one may elect first to insert
a guidewire under endoscopic control up to the ligament of Treitz (middle). After removing
the endoscope and rerouting the wire through the nose, a double catheter is insertedthe
distal end of the first one being placed at the ligament of Treitz for feeding purposes, with the
proximal tube being positioned at the perforation so that its aspiration holes are below, at, and
above the previously determined level of perforation (right).
A perforation rate of around 6 to 8% seems to be an unavoidable attribute of nonsurgical intubation as
practiced at the present time. With more precise selection of patients, a perforation rate of less than
5% may be achieved. Tumor location, length, and histologic type do not seem to affect the risk of
perforation in any significant way. Whether some perforations can be avoided through more gradual
dilation is uncertain at present, but this is a distinct possibility. Whenever feasible, surgical intervention
should be avoided in frail preterminally ill patients.93(3433) It should not be assumed automatically that
perforation is synonymous with total disaster, especially when it is recognized immediately and before
mediastinal contamination has developed. With conservative measures, most patients will survive this
complication.93(3434) A perforation does, however, prolong hospitalization. Instrumental esophageal
tears differ from spontaneous or foreign body perforations in that they usually are recognized early, and
because the esophagus is empty, there is little extravasation into the mediastinum. Immediate surgical
intervention appears therefore inappropriate in the vast majority of patients with malignant disease.
Adequate conservative management results in a more than 80%, if not 90%, survival from the event.
The key to successful conservative management is early diagnosis and avoidance of mediastinal
contamination.2,93(3435) Massive perforation into the pleural cavity with rapid accumulation of fluid
and air in the pleural space carries a poor prognosis, and surgical repair is seldom feasible when the
growth itself is lacerated.
Dislocation
Migration or dislocation may occur with any type of prosthesis. Recurring migration can usually be
prevented by using a prosthesis with distal or proximal retainer rings. However, such adaption may not
always prevent migration in the patient's course. Migration often occurs when there is insufficient
anchoring of the tube. Removal of a dislodged prosthesis from the stomach requires that the distal end
be grasped and brought up first through the dilated cardia and tumor. When removal of a dislodged
tube from the stomach appears excessively difficult, it may be left in the stomach, provided it does not
lodge in the antropyloric channel and cause gastric outlet obstruction. Usually, the patient is not aware
of such a foreign body. Dislodged tubes may also migrate unnoticed through the intestinal tract.
Tumor Overgrowth
Tumor overgrowth proximal or distal to the prosthesis may result in obstruction (Figure 4221). This
may be corrected by insertion of a longer prosthesis after photoablation of the overgrowing cancerous
tumor using an Nd:Yag laser or after dilation of the narrowed segment. With these maneuvers, it is
almost always possible to restore luminal patency.
(3436)Figure 4221. Two examples of the endoscopic appearance of cancer overgrowth

obscuring or partially occluding a prosthesis.
Stricture Due to Reflux Esophagitis
Fibrous stricturing related to severe reflux esophagitis may occur despite antireflux measures, although
this is rare. This may be managed either by dilation with the original prosthesis in place or by removal
of the prosthesis and introduction of a longer one.
Pressure Necrosis
Pressure necrosis caused by the funnel edge of the prosthesis usually occurs in an area of the wall
that has been invaded by tumor or previously irradiated, or both. This causes pain and leads to
formation of a mediastinal fistula. Deep pressure necrosis may extend to the aorta, with resultant
exsanguination. The likelihood of pressure necrosis is higher when marked angulation between the
esophagus and the prosthesis results in greater pressure being exerted along one portion of the
circumference of the prosthesis. This may occur in patients with kyphoscoliosis or hiatus hernia, or
when the axis of the prosthesis is in a transverse position relative to the long axis of the body, as in the
case of tumors of the distal esophagus and cardia (Figures 4222 and 4223).
(3437)Figure 4222. Left, Esophageal cancer in a patient with kyphoscoliosis and a hiatus
hernia. Middle, A barium swallow shortly after insertion of an endoprosthesis demonstrates an
obvious eccentric indentation and compression of the esophageal wall (arrow). Right, A
barium swallow a few months later demonstrates striking pressure necrosis of the esophageal
wall and intramural migration of the funnel (arrow). This was corrected by insertion of a
longer prosthesis.
(3438)Figure 4223. Left to right, The primary presenting esophageal cancer is 3 cm long,
but distant metastasis is present; recurrence of the cancer (arrowhead) after radiotherapy; after
insertion of the first endoprosthesis; pressure necrosis and leakage (arrows) due to asymmetric
compression; correction of this problem with a longer prosthesis (arrowhead).
Food Blockage
A common complication that occurs in the follow-up period after prosthesis insertion is blockage of the
tube by food; this is usually the result of inadequate mastication or of dietary indiscretion (Figure
4224). In the majority of patients, the blockage can be cleared easily either by carefully passing a
cytology brush through the endoprosthesis to push the food into the stomach or by gently moving a
small-diameter endoscope up and down through the prosthesis to displace the impacted food bolus.
Some physicians have the patient drink a papain solution first ( oz of 1-to-1 solution every hr for 2
hr). The use of this enzyme is intended to loosen meat fibers. However, complications have been
reported with its use, including esophageal necrosis95(3439) and hypernatremia.96(3440) Laboratory
studies have also cast doubt on whether a papain solution will actually reduce the size of cubes of
meat.97(3441) If all other measures fail, the prosthesis should be removed and exchanged with a new
one.
(3442)Figure 4224. Left, Adenocarcinoma of the cardia extending into the esophagus.
Middle, The cancer is bypassed by an endoprosthesis. Right, Characteristic appearance of
clogging (arrow) due to poorly masticated meat.
Prosthesis Removal
Removal of an endoprosthesis may be necessary in the management of several of the complications
outlined previously. This can be carried out using a strong grasping forceps, polypectomy snare, the
KeyMed-Nottingham introducer, or a small-caliber endoscope. The latter functions as a hook after the
tip of the instrument is deflected at a point below the distal edge of the prosthesis.
Alcohol Injection Techniques

Several reports exist concerning the palliative treatment of patients with obstructing malignant tumors
of the esophagus and gastric cardia by endoscopic injection of absolute alcohol.98100(3443) This
leads to tissue necrosis and thereby an improvement in the ability of patients to swallow. One patient
with a granular cell tumor of the esophagus has been treated by this method.101(3444) Existing
information is based on prospective, albeit noncontrolled, series of relatively small numbers of cases.
Some patients treated by alcohol injection have also undergone esophageal dilation, and it is likely that
this contributed to the relief of dysphagia. Although alcohol injection appears to be a potentially useful
technique, efficacy and safety are not well established.
Payne-James et al.98(3445) performed endoscopic injections of absolute alcohol in 11 patients without
a complication. Chung et al.99(3446) treated 36 patients with inoperable cancers including lesions
involving the cardia. Dysphagia recurred at a mean duration of 35 days after treatment. Complications
occurred in 3 patients (8.3%), including mediastinitis in 1 and tracheoesophageal fistulas in 2. Nine
patients underwent alcohol injections in the series of Moreira et al.100(3447) No complications
occurred, although treatment had to be halted in 1 patient with a broncho-esophageal fistula that was
present prior to initiation of injections. All three groups of investigators noted an improvement in
dysphagia grade in response to treatment. In two series, however, esophageal dilation was also
performed.98,99(3448) Pretreatment dilation was integral to the treatment protocol in the series of
Chung et al.99(3449) In all three series, the mean duration of palliation before the recurrence of
dysphagia was approximately 30 days.
Conclusions
The insertion of an esophageal endoprosthesis through an inoperable carcinoma of the esophagus or
cardia provides acceptable relief of dysphagia and usually allows a patient to remain at home during
the terminal stages of disease. Starvation is at least a common contributing cause of death in
untreated cancer of the esophagus and cardia. The insertion of an endoprosthesis combats this
deterioration and often results in an immediate weight gain during the first few weeks after insertion.
However, patients with a large tumor load, anorexia, and nausea will not gain weight.
any clinicians feel that intubation provides poor palliation because swallowing is far from normal and
blockage of the tube is not infrequent. Others feel that with a properly positioned prosthesis and good
mastication, swallowing function is acceptable and certainly helps prevent aspiration and dehydration.
Outpatient supervision enables early detection of tube dysfunction; this can usually be dealt with by
endoscopic means. The essential value of an esophageal prosthesis lies in the improvement in the
quality of life that occurs with the elimination of dysphagia, thus allowing terminally ill patients to remain
at home.
Esophageal intubation with a plastic prosthesis is a well-established palliative treatment for esophageal
cancer. Although endoscopic insertion techniques offer greater safety than the surgical methods
utilized mainly in the past, there is still a significant complication rate and mortality.
The advantage of prosthesis insertion is rapid and sustained relief of dysphagia in most patients. In
contrast to laser therapy, repeated procedures to establish luminal patency are usually not required.
Placement of a prosthesis is the treatment of choice for patients with bronchoesophageal fistulas. The
cost of prosthesis placement is significantly less than that for laser therapy. Self-expanding metal
stents are thought to have distinct advantages over the more conventional plastic prosthesis because
they may be inserted with less trauma and fewer side effects.
The disadvantages of prosthesis insertion include a high complication rate and procedure-related
mortality. Occasionally, it is necessary to limit a patient's diet to soft or blenderized foods to prevent
stent occlusion. A disadvantage of self-expanding metal stents may be a short duration of palliation
because of ingrowth of tumor or granulation tissue through the meshes of the stent. The efficacy,
safety, and particular benefits of self-expanding metal stents must be established by careful evaluation
of a large number of patients.
It is pointless and ill-advised to think in terms of a standard approach to the palliative care of patients
with malignant obstruction of the esophagus or cardia. At pres-ent, no standard therapy exists, and the
physician taking care of such patients still requires much ingenuity to solve each individual's problems.
The several available techniques are in fact complementary. The particular characteristics of the tumor
or of the clinical situation may favor one or several therapies over the other(s). Ideally, several
possibilities should be at hand, but beyond doubt, stent insertion should be considered as an important
and essential modality.
Endoscopic intubation has many attractive features especially for the frail and elderly. The single
procedure, usually without general anesthesia, the short hospital stay, and an immediate improvement
in swallowing are considerable gains. Indeed, for the majority of patients with a malignancy of the
esophagus or cardia, the insertion of a prosthesis appears to be the fastest and cheapest way of
obtaining longstanding palliation of dysphagia. With particular types of lesions, however, an increase in
the technical difficulty of prosthesis insertion, complications, and procedure-related mortality can be
anticipated. In such cases, initial palliation by laser photoablation should be considered. This
individualized, selective approach will presumably reduce the overall number of complications.
REFERENCES
1.
2.
3.
4.
5.
Tytgat GNJ. Dilatation therapy of benign esophageal stenoses. World J Surg 1989;13:1428.
Tytgat GNJ. Endoscopic therapy of esophageal cancer: Possibilities and limitations.
Endoscopy 1990;22:2637.
Cox JG, Dakkak M, Buckton GK, Bennett JR. Dilators for esophageal strictureA description
of a new bougie and a comparison of current instruments. Gastrointest Endosc
1989;35:5514.
Boyce HW Jr, Palmer ED. Techniques of Clinical Gastroenterology. Springfield, IL: Charles C
Thomas. 1975.
Parker CH, Peura DA. Palliative treatment of esophageal carcinoma using esophageal
dilation and prosthesis. Gastroenterol Clin North Am 1991;20:71729.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Heit HA, Johnson LF, Siegel SR. Palliative dilation for dysphagia in esophageal carcinoma.
Lundell L, Leth R, Lind T, et al. Palliative endoscopic dilatation in carcinoma of the esophagus
and esophagogastric junction. Acta Chir Scand 1989;155:17984.
Ng TM, Spencer GM, Sargeant IR, et al. Management of strictures after radiotherapy for
esophageal cancer. Gastrointest Endosc 1996;43:58490.
Tulman AB, Boyce HW Jr. Complications of esophageal dilatation and guidelines for their
Tytgat GNJ, Huibregtse K, Bartelsman JFWM, Den Hartog Jager FCA. Endoscopic palliative
therapy of gastrointestinal and biliary tumours with prostheses. Clin Gastroenterol
1986;15:24971.
Tytgat GNJ. Upper intestinal and biliary tract endoprosthesis. Dig Dis Sci 1986;31(suppl):57S.
Tytgat GNJ, Den Hartog Jager FCA, Bartelsman JFWM. Endoscopic prosthesis for advanced
esophageal cancer. Endoscopy 1986;18(suppl 3):329.
Boyce HW, Palmer ES. Techniques of Clinical Gastroenterology. Springfield, IL: Charles C
Thomas. 1975.
Tytgat GNJ, Den Hartog Jager FCA, Haverkamp HJ. Positioning of a plastic prosthesis under
fiberendoscopic control in the palliative treatment of cardioesophageal cancer. Endoscopy
1976;8:1805.
Earlam R, Cunha-Melo JR. Malignant oesophageal strictures: A review of techniques for
palliative intubation. Br J Surg 1982;69:618.
Sivak MV Jr. Therapeutic endoscopy of the esophagus. Surg Clin North Am 1982;62:80720.
Buess G, Schellong H, Kometz B, et al. A modified prosthesis for the treatment of malignant
esophagotracheal fistula. Cancer 1988;61:167984.
Branicki FJ, Ogilvie AL, Willis MR, Atkinson M. Structural deterioration of prosthetic
oesophageal tubesAn in vitro comparison of latex rubber and silicone rubber tubes. Br J
Surg 1981;68:8614.
Domschke W, Foerster ECh, Matek W. Self-expanding mesh stent for esophageal cancer
stenosis. Endoscopy 1990;22:1346.
Schaer J, Katon RM, Ivancev K, et al. Treatment of malignant esophageal obstruction with
silicone-coated metallic self-expanding stents. Gastrointest Endosc 1992;38:711.
Lux G, Wilson D, Wilson J, Demling L. A cuffed tube for the treatment of oesophago-bronchial
fistulae. Endoscopy 1987;19:2830.
Etienne J, Clestin LR. Oesophageal intubation: Past and present. Acta Endoscopica
1979;9:2359.
Clestin LR, Etienne J, Raimbert P, et al. Traitement endoscopique des stnoses
oesophagiennes par prothse de Clestin. Nouv Presse Med 1980;9:21557.
Boyce HW Jr. Medical management of esophageal obstruction and esophageal-pulmonary
fistula. Cancer 1982;50:2597600.
Atkinson M, Ferguson R, Ogilvie AL. Management of malignant dysphagia by intubation at
endoscopy. J R Soc Med 1979;72:8947.
Atkinson M, Ferguson R, Parker GC. Tube introducer and modified Celestin tube for use in
palliative intubation of oesophagogastric neoplasms at fiberoptic endoscopy. Gut
1978;19:66971.
Clestin LR, Campbell WB. A new and safe system for oesophageal dilatation. Lancet
1981;1:745.
Van den Brandt-Grdel V, Den Hartog Jager FCA, Tytgat GNJ. Palliative intubation of
malignant esophagogastric obstruction. J Clin Gastroenterol 1987;9:290.
Den Hartog Jager FCA, Bartelsman JFWM, Tytgat GNJ. Palliative treatment of obstructing
esophagogastric malignancy by endoscopic positioning of a plastic prosthesis.
Straumann A, Gyr KI, Stalder GA. Verschluss maligner oesophago-bronchialer Fistel mit
Klebstoff-fixierter Endoprothese. Schweiz Rundsch Med Prax 1984;73:10867.
31. Brckner H, Grimm H, Nam VC, et al. Transnasal fixation of the esophageal prosthesis.
32. Buset M, Des Marez B, Cremer M. Endoscopic palliative intubation of the esophagus invaded
by lung cancer. Gastrointest Endosc 1990;36:3579.
33. Goldschmid S, Boyer HW, Nord HJ, et al. Treatment of pharyngoesophageal stenosis by
polyvinyl prosthesis. Am J Gastroenterol 1988;83:5138.
34. Spinelli P, Cerrai FG, Meroni E. Pharyngo-esophageal prostheses in malignancies of the
cervical esophagus. Endoscopy 1991;23:2134.
35. Loizou LA, Rampton D, Bown SG. Treatment of malignant strictures of the cervical
esophagus by endoscopic intubation using modified endoprostheses. Gastrointest Endosc
1992;38:15864.
36. Atkinson M, Ferguson R. Fibreoptic endoscopic palliative intubation of inoperable
oesophagogastric neoplasms. Br Med J 1977;1:2667.
37. Ell C, Riemann JF, Demling L. Endoscopic occlusion of a neoplastic esophagomediastinal
fistula by a fast-hardening amino acid solution. Gastrointest Endosc 1986;32:2878.
38. Hordijk ML, Dees J, Van Blankenstein M. The management of malignant
esophago-respiratory fistulas with a cuffed prosthesis. Endoscopy 1990;22:2414.
39.
Tytgat GNJ, Mathus-Vliegen EMH, Den Hartog Jager FCA. Laser versus endoprosthesis. In
Riemann JF, Ell C (eds). Lasers in Gastroenterology. Stuttgart: Georg Thieme Verlag,
1989;109.
40. Ogilvie AL, Dronfield MW, Ferguson R, Atkinson M. Palliative intubation of oesophagogastric
neoplasms at fibreoptic endoscopy. Gut 1982;23:10607.
41. Richter JM, Hilgenberg AD, Christensen MR, et al. Endoscopic palliation of obstructive
esophagogastric malignancy. Gastrointest Endosc 1988;34:4548.
42. Gasparri G, Casalegno PA, Camandona M, et al. Endoscopic insertion of 248 prostheses in
inoperable carcinoma of the esophagus and cardia: Short-term and long-term results.
43. Cotton MH, Sammon AM. Carcinoma of the oesophagus in Transkei: Treatment by intubation.
Thorax 1989;44:42.
44.
Weisel W, Raine F, Watson RR, Frederick JJ. Palliative treatment of esophageal carcinoma:
A method and its evaluation. Ann Surg 1959;149:20716.
45.
O'Conner T, Watson R, Lepley D, Weisel W. Esophageal prosthesis for palliative intubation.
Further evaluation of 378 patients. Arch Surg 1963;87:2759.
46. Palmer ED. Peroral prosthesis for the management of incurable esophageal carcinoma. Am J
47. Hegarty MM, Angorn IB, Bryer JV, et al. Pulsion intubation for palliation of carcinoma of the
oesophagus. Br J Surg 1977;64:1605.
48.
Kairaluoma MI, Jokinen K, Karkola P, Larmi TKI. Celestin tube palliation of unresectable
oesophageal carcinoma. J Thorac Cardiovasc Surg 1977;73:36.
49.
Bergerault P, Denez B, Mahe J, et al. Protheses endooesophagiennes de Clestin posees par
voie endoscopique. Experience de 35 cas de cancer de l'oesophage. Ann Gastroenterol
Hepatol 1980;16:3740.
50. Jones DB, Davies PS, Smith PM. Endoscopic insertion of palliative oesophageal tubes in
oesophagogastric neoplasms. Br J Surg 1981;68:1978.
51. Soehendra N. Endoskopisches Tubuseinfuhren bei malignen osophagus und Kardiastenosen.
52.
Ogilvie AL, Dronfield MW, Ferguson R, Atkinson M. Outcome of endoscopic intubation in 100
patients with oesophagogastric carcinoma [Abstract]. Gut 1981;22:A414.
53. Balmes JL, Baghdadi H, Michel H. Palliative endoscopic treatment of malignant esophageal
stenosis using a Haring prosthesis. Digestion 1982;23:318.
54.
van Blankenstein M, Tan TG, Dees J. Fibreoptic intubation of malignant oesophago-gastric
strictures by the Atkinson method. Neth J Surg 1982;34:112.
55. Watson A. A study of the quality and duration of survival following resection, endoscopic
intubation and surgical intubation in oesophageal carcinoma. Br J Surg 1982;69:5858.

56. Lux G, Groitl H, Reimann JF, Demling L. Tumor stenosis of the upper gastrointestinal
tractNonsurgical therapy by bridging tubes. Endoscopy 1983;15:20712.
57. Rose JD, Smith PM. Fibre endoscopic insertion of palliative oesophageal tubes with the
Nottingham introducer. J R Soc Med 1983;76:2668.
58. Fugger R, Niederle B, Jantsch H, et al. Endoscopic tube implantation for the palliation of
malignant esophageal stenosis. Endoscopy 1990;22:1014.
59. Spinelli P, Cerrai FG, Ciuffi M, et al. Endoscopic stent placement for cancer of the lower
esophagus and gastric cardia. Gastrointest Endosc 1994;40:4557.
60. Loizou LA, Rampton D, Atkinson M, et al. A prospective assessment of quality of life after
endoscopic intubation and laser therapy for malignant dysphagia. Cancer 1992;70:38691.
61. Oliver SE, Robertson CS, Logan RF, et al. What does radiotherapy add to survival over
endoscopic intubation alone in inoperable squamous cell oesophageal cancer?. Gut
1990;31:7502.
62. Barr H, Krasner N, Raouf A, Walker RJ. Prospective randomised trial of laser therapy only
and laser therapy followed by endoscopic intubation for the palliation of malignant dysphagia.
Gut 1990;31:2528.
63. Alderson D, Wright PD. Laser recanalization versus endoscopic intubation in the palliation of
malignant dysphagia. Br J Surg 1990;77:11513.
64. Buset M, des Marez B, Baize M, et al. Palliative endoscopic management of obstructive
esophagogastric cancer: Laser or prosthesis?. Gastrointest Endosc 1987;33:35761.
1991;100:130310.
66. Hahl J, Salo J, Ovaska J, et al. Comparison of endoscopic Nd:YAG laser therapy and
oesophageal tube in palliation of oesophagogastric malignancy. Scand J Gastroenterol
1991;26:1038.
67. Carter R, Smith JS, Anderson JR. Laser recanalization versus endoscopic intubation in the
palliation of malignant dysphagia: A randomized prospective study. Br J Surg
1992;79:116770.
68. Reed CE, Marsh WM, Carlson LS, et al. Palliation of advanced esophageal cancer. Ann
69. Bethge N, Knyrim K, Wagner HJ, et al. Self-expanding metal stents for palliation of malignant
esophageal obstructionA pilot study of eight patients. Endoscopy 1992;24:4115.
70. Fleischer DE, Bull-Henry K. A new coated self-expanding metal stent for malignant
esophageal strictures. Gastrointest Endosc 1992;38:4946.
71. Neuhaus H, Hoffmann W, Dittler HJ, et al. Implantation of self-expanding esophageal metal
stents for palliation of malignant dysphagia. Endoscopy 1992;24:40510.
72.
Winkelbauer FW, Schofl R, Niederle B, et al. Palliative treatment of obstructing esophageal
cancer with nitinol stents: Value, safety, and long-term results. AJR 1996;166:7984.
73. Ell C, May A, Hahn EG. Gianturco-Z stents in the palliative treatment of malignant esophageal
obstruction and esophagotracheal fistulas. Endoscopy 1995;27:495500.
74. Grund KE, Storek D, Becker HD. Highly flexible self-expanding meshed metal stents for
palliation of malignant esophagogastric obstruction. Endoscopy 1995;27:48694.
75. Bethge N, Sommer A, Vakil N. Treatment of esophageal fistulas with a new
polyurethane-covered, self-expanding mesh stent: A prospective study. Am J Gastroenterol
1995;90:21436.
76. De Palma GD, Galloro G, Sivero L, et al. Self-expanding metal stents for palliation of
inoperable carcinoma of the esophagus and gastroesophageal junction. Am J Gastroenterol
1995;90:21402.
77. Tranberg KG, Stael von Holstein C, Ivancev K, et al. The YAG laser and Wallstent
endoprosthesis for palliation of cancer in the esophagus or gastric cardia.
78. Watkinson AF, Ellul J, Entwisle K, et al. Esophageal carcinoma: Initial results of palliative
treatment with covered self-expanding endoprostheses. Radiology 1995;195:8217.
79. Wagner HJ, Stinner B, Schwerk WB, et al. Nitinol prostheses for the treatment of inoperable
malignant esophageal obstruction. J Vasc Interv Radiol 1994;5:899904.
80. Segalin A, Granelli P, Bonavina L, et al. Self-expanding esophageal prosthesis. Effective
palliation for inoperable carcinoma of the cervical esophagus. Surg Endosc 1994;8:13435.
81. Raijman I, Walden D, Kortan P, et al. Expandable esophageal stents: Initial experience with a
new nitinol stent. Gastrointest Endosc 1994;40:61421.
82. Song HY, Do YS, Han YM, et al. Covered, expandable esophageal metallic stent tubes:
Experiences in 119 patients. Radiology 1994;193:68995.
83. Ell C, Hochberger J, May A, et al. Coated and uncoated self-expanding metal stents for
malignant stenosis in the upper GI tract: Preliminary clinical experiences with Wallstents. Am
84. Cwikiel W, Stridbeck H, Tranberg KG, et al. Malignant esophageal strictures: Treatment with
a self-expanding nitinol stent. Radiology 1993;187:6615.
85. Bethge N, Sommer A, Gross U, et al. Human tissue responses to metal stents implanted in
vivo for the palliation of malignant stenoses. Gastrointest Endosc 1996;43:596602.
86. Knyrim K, Wagner HJ, Bethge N, et al. A controlled trial of an expansile metal stent for
palliation of esophageal obstruction due to inoperable cancer. N Engl J Med
1993;329:13027.
87. Alderson D, Blazeby JM. Expanding metal stents in the gastrointestinal tract. Br J Surg
1995;82:14413.
88. Wu WC, Katon RM, Saxon RR, et al. Silicone-covered self-expanding metallic stents for the
palliation of malignant esophageal obstruction and esophagorespiratory fistulas: Experience in
32 patients and a review of the literature. Gastrointest Endosc 1994;40:2233.
89. Weigert N, Neuhaus H, Rsch T, et al. Treatment of esophagorespiratory fistulas with
silicone-coated self-expanding metal stents. Gastrointest Endosc 1995;41:4906.
90. Do YS, Song HY, Lee BH, et al. Esophagorespiratory fistula associated with esophageal
cancer: Treatment with a Gianturco stent tube. Radiology 1993;187:6737.
91. Nelson DB, Silvis SE, Ansel HJ. Management of a tracheo-esophageal fistula with a
silicone-covered self-expanding metal stent. Gastrointest Endosc 1994;40:4979.
92. Hine KR, Atkinson M. The diagnosis and management of perforations of esophagus and
pharynx sustained during intubation of neoplastic esophageal strictures. Dig Dis Sci
1986;31:5713.
93.
Tytgat GNJ. Esophageal perforation: Diagnosis, prevention and therapy. Eur J Gastroenterol
Hepatol 1990;2:193202.
94. Wesdorp ICE, Bartelsman JFWM, Huibregtse K, et al. Treatment of instrumental esophageal
perforation. Gut 1984;25:398404.
95.
Holsinger JW Jr, Fuson RL, Sealy WC. Esophageal perforation following meat impaction and
papain ingestion. JAMA 1968;204:1889.
96.
Zarling EJ, Moeller DD. Bezoar therapy: Complications using Adolph's Meat Tenderizer and
alternatives from literature review. Arch Intern Med 1980;141:169970.
97. Goldner F, Danley D. Enzymatic digestion of esophageal meat impaction. A study of Adolph's
Meat Tenderizer. Dig Dis Sci 1985;30:4569.
98. Payne-James JJ, Spiller RC, Misiewicz JJ, Silk DB. Use of ethanol-induced tumor necrosis to
palliate dysphagia in patients with esophagogastric cancer. Gastrointest Endosc
1990;36:436.
99. Chung SC, Leong HT, Choi CY, et al. Palliation of malignant oesophageal obstruction by
endoscopic alcohol injection. Endoscopy 1994;26:2757.
100. Moreira LS, Coelho RC, Sadala RU, Dani R. The use of ethanol injection under endoscopic
control to palliate dysphagia caused by esophagogastric cancer. Endoscopy 1994;26:3114.
101. Moreira LS, Dani R. Treatment of granular cell tumor of the esophagus by endoscopic
injection of dehydrated alcohol. Am J Gastroenterol 1992;87:65961.
Chapter 43 Palliation of Malignant Obstruction: Lasers and Tumor

Probes
(3450)
(3451)
DAVID E. FLEISCHER, M.D.
The birth of modern endoscopy can be traced to the late 1950s and early 1960s, but it was during the
decade of the 1970s when endoscopy came into widespread use and the range of applications
blossomed. Gastrointestinal (GI) endoscopy was born as a diagnostic modality, but gradually a handful
of therapeutic applications emerged, notably the treatment of GI bleeding. Initially, monopolar
electrocoagulation was the only modality available for hemostasis. By the mid 1970s, however,
something very different had emerged for this purpose: endoscopic laser therapy (ELT). Not only did
the laser appear to be safer and more effective than monopolar electrocoagulation, but the word laser
created new expectations. It connoted treatment that was precise, high-tech, futuristic, perhaps even
revolutionary, and certainly on the "cutting edge."
Initial enthusiasm for the use of lasers at endoscopy was based entirely on anecdotal
reports.13(3452) Laboratory investigations established the potential efficacy of this new
modality,413(3453) but uncertainty existed as to which wavelength and therefore which type of laser
was best suited to the goal of endoscopic hemostasis. This question led to heated debate. Proponents
of the argon laser argued that it was safer and equally effective. Advocates of the
neodymium:yttrium-aluminum-garnet (Nd:YAG) laser argued that it was more effective, also safe, and
clearly more versatile. Because it seemed to be effective in the treatment of bleeding GI lesions, and
because monopolar electrocoagulation was the only, albeit less-than-satisfactory alternative, the laser
became very popular. Despite its high cost and a lack of randomized controlled trials, work with lasers
burgeoned, and data were gradually accumulated that established the safety and efficacy of
endoscopic laser photocoagulation for control of GI bleeding (see Chapter 29: Laser Therapy).
The use of lasers in endoscopy to resolve clinical problems widened in the early 1980s with the
development of techniques for treatment of GI neoplasms. Even more creative applications were
proposed and developed, and interest in the use of lasers reached truly remarkable levels. Since that
time, however, there has been a steady decline in enthusiasm for the use of lasers in conjunction with
GI endoscopy. Although it is well established that bleeding from many types of GI lesions can be
controlled by laser photocoagulation, treatment of GI neoplasms is now the main indication for the use
of lasers at endoscopy. The initial wave of enthusiasm generated by the introduction of lasers into GI
endoscopy has waned. The important question: What happened?
Laser light has many unique properties. Because it is monochromatic, it can be focused very precisely;
it is this property that led to its use in the GI tract. Despite the laser's many singular properties,
however, it is fundamentally the generation of thermal energy within tissue that has been used to
advantage by GI endoscopists. But if the ability to focus heat is the only desired attribute, there is then
the possibility that other less expensive, equally effective, and more portable devices might be
developed for the same purpose. In fact, this is what happened.
Expense has been another factor in the declining use of lasers in GI endoscopy. The first machines
available commercially cost approximately $100,000 and required special water and electric
connections. Newer machines do not require special power sources or a water supply for cooling and
are therefore more portable. The cost of purchase of a laser has also decreased, but it is still
substantial in relation to that for other thermal hemostatic devices.
Currently, no company deals exclusively in the development and commercialization of lasers for GI
applications. This overall loss of focus and the declining interest in the use of lasers in GI endoscopy
have led to less research into new applications for lasers in the management of digestive diseases.
Every new technology passes through phases of emergence, expanding interest, and decline.
Although lasers have now a limited, albeit important and well-established place in GI endoscopy, there
is nevertheless the potential for an expanded role in the future in both diagnosis and therapy. If this is
to occur, however, it will be necessary for researchers and clinical practitioners to take advantage of
the unique feature of laser lightits monochromaticity. This singular property, which separates lasers
from all other forms of light and energy, is the key to future developments.
Principles of Laser Physics

The word laser is an acronym for light amplification by stimulated emission of radiation. This
phenomenon occurs when matter is excited to a higher energy state and then emits light as it returns
to a lower energy state (see Chapter 10: Laser Physics and Laser-Tissue Interactions). Laser light is
intense, monochromatic, and easily focused. The heart of any laser is the medium used to generate
the beam. This determines the wavelength produced, which, in turn, gives each type of laser its
specific properties. If the lasing medium is a gas (e.g., argon), it can be excited with an electric
discharge. If the medium is solid (e.g., Nd:YAG), a different energy source must be used, usually an
intense lamp that is energized by an electric power supply. Several types of lasers have been used
therapeutically in medicine. Those most frequently employed in GI endoscopy are the Nd:YAG, argon,
and pulsed-dye lasers; other lasers used include the potassium titanyl phosphate (KTP) and holmium
lasers. Technologic advances may allow the CO2 laser to be used with flexible endoscopes, although
the current lack of a flexible waveguide remains problematic.
Laser-Tissue Interaction
Cummins14(3454) emphasized that the fundamental events that occur with laser-tissue interaction are
absorption and scattering. Differences in laser-tissue interactions among the various lasers are mainly
due to the presence of water and hemoglobin in biologic tissues.
For each type of laser, energy absorption varies according to wavelength (Figure 431). Hemoglobin is
highly absorptive of light at short wavelengths, but its absorptive capacity is low within and beyond red
wavelengths. Water is highly absorptive at long wavelengths, but its absorption is low in the visible
range. The energy produced by the argon laser, therefore, is highly absorbed by hemoglobin, but
absorption by water is low. Both hemoglobin and water absorb, to some extent, the energy produced at
the wavelength of the Nd:YAG laser.
(3455)Figure
431.
Wavelength
range
and
absorption
of
argon,
neodymium:yttrium-aluminum-garnet (Nd:YAG), and CO2 lasers. Argon has high absorption
in hemoglobin, and carbon dioxide has high absorption in water. (From Enderby CE. Medical
laser fundamentals. In Fleischer D, Jensen D, Bright-Asare P [eds]. Therapeutic Laser
Endoscopy in Gastrointestinal Disease. Boston: Martinus Nijhoff, 1983; 18, with kind
permission from Kluwer Academic Publishers.)
Scattering is also an important factor in the interaction of laser light and tissue.15(3456) Some energy
is scattered as the laser beam enters tissue. This occurs initially as the beam strikes a tissue surface,
although some further scattering occurs within the tissue substance. On entering the tissue, scattering
may occur in many directions, whereupon the energy is either absorbed or rescattered. Tissue
interaction is therefore more complex than simple straight-line penetration of the beam (Figure 432).
Penetration is defined and measured as the depth at which the energy level is only 10% of the initial
energy level. The tissue penetration of the Nd:YAG laser is about 4 mm; that is, at a tissue depth of 4
mm, the amount of energy remaining is about 10% of the incident energy. Argon laser penetration is
approximately 1 mm.
(3457)Figure 432. Tissue penetration. Laser beam penetration is not uniform and depends
on the interrelationship of absorption and scattering. (From Enderby CE. Medical laser
fundamentals. In Fleischer D, Jensen D, Bright-Asare P [eds]. Therapeutic Laser Endoscopy
in Gastrointestinal Disease. Boston: Martinus Nijhoff, 1983; 18, with kind permission from
Kluwer Academic Publishers.)
The phenomena of absorption and scattering are interdependent and determine the pattern of light
intensities at all points within the tissue. Wavelength, tissue constituents, and therapeutic spot size are
the major determinants of the intensity pattern. Spot size refers to the surface area of tissue struck by
the laser energy. Spot size for a flexible fiber waveguide of given diameter and fixed angle of beam
divergence is proportionate to the distance between the fiber and the tissue surface (Figure 433). The
actual effect on tissue is further dependent on the power output at which the machine is set and the
pulse duration (time during which the laser beam is applied to the tissue). When tissue is heated, there
are critical temperature levels at which endothelial damage, cell death, protein coagulation, blood
vessel constriction, and tissue vaporization occur. These can be correlated with certain endoscopic
observations and histologic changes that reflect the actual events occurring within the tissue (Table
431).
TABLE 431 Relationship Between Temperature and Tissue Event

CRITICAL
TEMPERATURE (C)
45
HISTOLOGIC
EVENT
Cell death, edema,
endothelial
damage,
vasodilation
Protein coagulates
60
80
100
210+
Denatured collagen
contracts, blood
vessels constrict
Tissue water boils
Dehydrated tissue
ENDOSCOPIC
MANIFESTATION
Erythema, edema cuff
Tissue turns gray-brown,

blood turns black
Tissue "puckers"
Vaporization causes divot

Blackened tissue burns,
disappears, and leaves
glowing embers
From Fleischer D. Lasers in gastroenterology. Am J Gastroenterol 1984;
79(5):40615.
(3458)Figure 433. Relationship of beam size on tissue and distance from tip of fiber to
tissue. With a divergent beam, spot size will be related to distance. (From Enderby CE.
Medical laser fundamentals. In Fleischer D, Jensen D, Bright-Asare P [eds]. Therapeutic
Laser Endoscopy in Gastrointestinal Disease. Boston: Martinus Nijhoff, 1983; 18, with kind
permission from Kluwer Academic Publishers.)
From a clinical standpoint, tissue events can be classified as either coagulation or vaporization. The
coagulative effect results in hemostasis when laser energy is employed to stop acute GI bleeding and
to eliminate lesions that have the potential to bleed (e.g., vascular ectasias, visible vessels in ulcer
craters). It is the vaporizing effect by which tissue is ablated when laser energy is used to destroy
neoplastic tissue and by which normal tissue is incised for therapeutic purposes (e.g.,
infundibulectomy, drainage of cystic lesions).
It is possible that new techniques can be developed that allow tissue interactions to take place by
selective absorption of a given wavelength rather than by coagulation or vaporization. Selective
absorptivity may produce a desired tissue effect without generating high temperatures within tissue.
For example, a nonbleeding, red, vascular malformation might be obliterated using a KTP laser that
produces a wavelength that is selectively absorbed by the color red. Because this effect does not
depend on thermal energy, therapy is more efficient and less destructive of normal tissue, as this does
not absorb the KTP wavelength.
Equipment
Of the several types of lasers available, the argon and Nd:YAG have been used most extensively in GI
endoscopy. Photodynamic therapy makes use of an argon laser; the solid-state Nd:YAG laser, which
emits invisible light at a wavelength of 1.06 nm, has been employed widely for photoablation of
malignant tissue. Of the several types available, the Nd:YAG laser offers the greatest depth of
penetration and is therefore best suited for tumor ablation. Commercially available medical Nd:YAG
lasers have maximum power outputs in the range of 60 to 110 watts (W) at the fiber tip. Most
machines permit the operator to select an output between 1 W and the maximum output.16,17(3459)
Although medical lasers are mounted on wheels, most machines are large and not truly
portable.16(3460)
The laser beam is transmitted through a fiber (waveguide) that is made of either quartz or silica (glass).
There are two types of fiber waveguides: noncontact and contact. Each has advantages and
disadvantages. Both types can be used with an Nd:YAG laser.
The noncontact fibers usually have a coaxial design with an inner fiber for transmission of the laser
beam and a surrounding sheath that allows for insufflation of a gas, usually carbon dioxide. This
coaxial gas flow is intended to cool the fiber tip, prevent the accumulation of debris on the tip, and dry
any fluid deposits. Accumulated debris will result in rapid heating, so that the tip may melt when the
laser is fired. The inner fiber usually has a diameter of about 1 mm; the outer diameter ranges from 2
to 2.4 mm.16(3461) Because most photoablation procedures are time consuming, high rates of coaxial
gas flow can result in excessive abdominal distention.
Contact fibers are similar to noncontact waveguides except that a sapphire tip is placed at the end of
the fiber. This markedly increases the energy density at the end of the fiber. Therefore, the power
output of the laser must be decreased substantially; high power will in fact damage the sapphire tip.
Usually, the desired tissue effect can be achieved with much lower energy levels compared with those
employed with noncontact fibers. However, only a small area of tissue is treated with each pulse,
whereas much larger volumes can be treated with a noncontact fiber. In contrast to the latter type of
waveguide, damage or destruction of the sapphire tip during firing of the laser as a result of
accumulated debris is much less problematic with the contact fiber.
The laser beam exits the noncontact fiber waveguide at a fixed divergence angle, typically about 10
degrees. Therefore, the beam that strikes a tissue surface has the geometry of a cone with its apex at
the fiber tip. As the fiber is moved closer to a tissue surface, the diameter of the circular base of the
cone (spot size) decreases; conversely, spot size increases as the tip is moved away from the surface.
If the power output of the laser is constant, total laser energy delivered to the tissue is confined to a
progressively smaller area as spot size decreases. The degree of tissue heating per unit of time
increases as spot size decreases, although the heat energy is confined within a progressively smaller
volume of tissue. Essentially, this means that the range of tissue changes (see Table 431) that occur
in response to laser energy can be controlled to a degree by changing the distance between the tip of
the waveguide and the tissue surface. An aiming beam is incorporated into the design of most medical
lasers and waveguides (Nd:YAG laser light at a wavelength of 1.06 nm is invisible). This second laser
projects a circle of visible light that corresponds to the spot size on the tissue surface. Waveguides are
available that project the laser beam at angles of 60 to 90 degrees to the long axis of the fiber to allow
lateral aiming.18(3462) Disposable waveguides are available commercially, but these are relatively
expensive. Nondisposable fibers can be cut and polished if damage has occurred during a procedure.
Contact fibers do not have an aiming beam because these devices are placed directly in contact with
the tissue. Furthermore, the concepts and techniques that relate to manipulation of spot size are
irrelevant with use of a contact fiber. Other suggested advantages of the contact waveguide in contrast
to the noncontact fiber include:
1. Repetitious cleaning of the fiber tip during a treatment session is not necessary.
2. Less smoke is produced.
3. Sapphire tips having different geometric designs produce different tissue effects including cutting,
coagulation, and vaporization.
4. The device functions with a laser of relatively low power (10 to 15 W).
A major disadvantage of the contact waveguide is that relatively long treatment sessions are required
to achieve a desired result in comparison to the time required for treatment with a noncontact fiber.
A variety of other accessories are also needed for laser therapy. Protective filters are not needed when
using an electronic endoscope.17(3463) If the laser is to be used with a fiberoptic instrument, however,
an optical filter must be placed over the ocular of the endoscope to avoid inadvertent eye
injuries.19(3464) For the same reason, special goggles must always be worn during use or calibration
of the laser.19(3465) A wide variety of different types of equipment may be needed during an
endoscopic laser ther-apy (ELT) session. Some items that should be readily available include:
over-the-wire dilators, polypectomy snares, foreign body retrieval devices, and cleaning brushes (for
dbridement of necrotic tissue from the tumor surface).
Plumes of smoke are usually generated during laser treatment sessions, especially when higher power
settings are used to photoablate tissue. Aside from obscuring the endoscopic field of view, smoke
frequently exits via the accessory channel port of the endoscope and the patient's mouth. There is a
slight concern that these plumes of smoke could contain infectious agents (see later in this chapter).
The smoke is also irritating to the eyes and mucous membranes. It is therefore advisable that some
mechanism be used to evacuate the smoke. Machines (smoke evacuators) are available commercially
for this purpose. When these are connected to lengths of flexible plastic piping, the ends of which are
placed near the patient's mouth and the accessory port of the endoscope, most of the smoke can be
eliminated from the procedure room. One advantage of an electronic endoscope is that the
endoscopist's face is further away from the accessory channel of the endoscope.
Clinical Applications
Gastrointestinal Neoplasms
Tumors of the esophagus, stomach, duodenum, main duodenal papilla, colon, and rectum have been
treated by ELT. For the most part, palliation has been the objective of therapy, but in some cases,
curative treatment has been intended and achieved.
An endoscopic approach to the palliative management of GI cancer has several appealing aspects:
1. It can preclude the need for palliative surgery and general anesthesia, with their attendant
morbidity and mortality.
2. It diminishes considerably the likelihood of systemic side effects.
3. It can be performed under direct vision.
4. Unlike radiation therapy, there is no maximum dose, so that if a tumor recurs in the same area
after treatment, therapy can be repeated.
ELT is limited in that it does not affect pathologic tissue outside of the GI lumen, and also the duration
of benefit is usually not more than a few months.
The goals of palliative therapy are usually relief of obstruction or prevention of blood loss. Generally,
attempts to eliminate or reduce pain have been ineffective, presumably because pain is related to
neoplastic spread outside the GI lumen. Bleeding may be treated either by coagulating bleeding sites
on the tumor surface or by destroying the tumor itself and thereby its vasculature. Treatment for
bleeding is usually ineffective if the tumor is excessively large (i.e., measuring more than 8 cm in any
dimension).
Esophageal Cancer
Satisfactory palliation may be achieved by surgery in about 40% of patients who have resectable
esophageal cancers.20,21(3466) However, surgery in these patients has an appreciable morbidity as
well as a mortality rate that ranges from 7 to 29%.21,22(3467) This has led to the development of less
invasive and less expensive methods, largely endoscopic, for palliation (Table 432).
TABLE 432
Endoscopic Treatment of Esophageal
Cancer
THERMAL
Nd:YAG laser (other lasers: KTP [holmium])
Bipolar electrocoagulation
Monopolar electrocoagulation (snare polypectomy)
Photodynamic therapy
Intraluminal radiation ("afterloading")
MECHANICAL
Esophageal prosthesis
Percutaneous gastrostomy
INJECTION
Nd:YAGneodymium:yttrium-aluminum-garnet;
KTPpotassium-titanylphosphate.
Treatment Strategy
The usual goals of treatment are expeditious relief of dysphagia to the point that the patient can both
meet his or her caloric requirements by the oral intake of food and feel comfortable eating in public.
The luminal diameter necessary to achieve these goals varies, but is generally in the range of 11 to 13
mm. Whether the lumen can be considered adequate depends entirely on subjective relief of
dysphagia. If the patient experiences minimum or no improvement in the ability to swallow, despite
apparently adequate luminal diameter, further ELT may not be beneficial and alternative treatment
methods should be considered. The length of the dysphagia-free interval after ELT depends on a
number of variables in addition to the laser treatment per se (e.g., rate of tumor growth, response to
ancillary therapy). If obstruction recurs, further therapy can be performed. This absence of restriction
on the number of treatment sessions is an important difference between ELT and radiotherapy and is
one of the appealing aspects of ELT. Therapy may become incrementally less effective over time, in
which case the dysphagia-free interval between sessions becomes progressively shorter. As the tumor
becomes longer and more circumferential, it may also take on a firm, almost fibrotic character. In such
cases, an alternative form of therapy may be needed (e.g., endoprosthesis). The timing of this decision
is important because persistence with minimally effective ELT may make it more difficult to insert an
endoprosthesis.
Patient Evaluation
When esophageal cancer is newly diagnosed, it is highly desirable at the outset that the patient
undergo evaluation by a team of physicians with expertise in medical oncology, radiation therapy, and
cancer surgery to determine the best strategy for treatment. Several types of imaging studies are
usually needed to determine whether surgery, radiotherapy, chemotherapy, ELT, some other form of
endoscopic palliation, or a combination of these modalities should have a role in the management of
the individual patient.
An upper GI series will provide information on the location, length, shape, and nature of the
esophageal cancer (Figure 434). Often, this is the best method for defining acute angulations and
luminal distortion (Figure 435). Endoscopic evaluation with biopsy provides a tissue diagnosis and
allows the physician to evaluate the geometry of the tumor. Is it primarily exophytic? How much of the
tumor is submucosal? Is there a polypoid, exophytic component? For many years, the standard for
assessing the spread of the tumor has been computed tomography (CT) of the thorax, ideally with oral
contrast. If the malignancy is located proximally, CT views of the neck should be included; if distal in
location, views of the upper abdomen are needed. With respect to involvement of lymph nodes and the
possibility of curative surgical resection, the information provided by CT is relatively imprecise. CT is,
however, a standard form of imaging that is generally available. Endoscopic ultrasonography provides
more precise information as to depth of tumor invasion and the presence or absence of nodal
involvement (see Chapter 41: Endoscopic Ultrasonography of the Esophagus).
(3468)Figure 434. Barium contrast x-ray of a tumor involving the distal esophagus. This
provides information about the location, length, and shape of the esophageal cancer.
(3469)Figure 435. Barium contrast x-ray demonstrates distortion with angulation caused by
an esophageal cancer.
Treatment Technique
The laser should be calibrated according to the manufacturer's recommendations to ensure that it
delivers the power output selected by the operator. The waveguide should be inspected for structural
integrity; the aiming light should provide a perfectly circular spot in the case of noncontact fibers.
Conscious sedation is almost always adequate for ELT. The procedure has been performed without
sedation.23(3470) However, the degree of discomfort experienced by patients will usually be greater
than that associated with other diagnostic and therapeutic upper GI endoscopic procedures.
Depending on the length of the tumor and degree of stenosis, ELT usually requires somewhat longer
periods of time (usually 20 to 40 min). Dilation may be required, larger-diameter (therapeutic)
endoscopes may be used, abdominal distention may arise from insufflation of gas via the laser fiber,
and pain may result from the transfer of heat generated by the treatment to nearby normal tissues and
structures. Therefore, deeper levels of sedation and analgesia are often needed for ELT compared
with a standard diagnostic procedure.
The type of endoscope used for ELT depends on a number of factors. A two-channel instrument allows
for evacuation of smoke and gas without removal of the waveguide, but it will be more difficult or
impossible to reach the distal margin of the tumor. Moving the waveguide between channels facilitates
aiming of the laser beam toward different surfaces of the tumor with less need to torque the endoscope
for targeting. If the malignant stenosis is extremely tight or the lumen is circuitous, use of a
smaller-diameter, single-channel instrument may be more advantageous. The basic technique for
treating esophageal cancer involves the delivery of laser energy to the tumor via a fiber waveguide
passed through an endoscope. In simple terms, the laser beam is focused on the target tissue and the
tissue is destroyed. Within this framework, however, there are important variations in technique that
may be categorized: 1. antegrade versus retrograde; 2. noncontact versus contact; 3. high-power
versus low-power.
As first described, ELT was performed using the antegrade (prograde) approach (Figures 436 and
437).24(3471) The endoscope is passed to the proximal margin of the tumor, so that the obstructed
lumen is seen directly and from above. The beam is focused centrally, working from the constricted
lumen toward the periphery. Therapy is carried out by drilling, usually at sequential sessions, through
the obstruction. With the retrograde technique, the endoscope is advanced through the tumor to or
beyond the distal margin. Treatment is carried out proceeding in a cephalad direction, so that the
proximal margin is treated last. Dilation may be required in order to reach the distal margin of the
tumor.
(3472)Figure 436. Antegrade treatment technique using an endoscopic Nd:YAG laser with a
quartz waveguide delivery system. A, Endoscope advanced to the superior margin of the
tumor, with the waveguide protruding from the accessory channel. Treatment (day 1) begins
centrally around the residual lumen proceeding toward but not to the wall. Cross sections at
left: Initial changes are coagulative. With continued thermal damage, vaporization occurs. B,
On day 3 (48 hr later), the laser-treated tissue is necrotic. C, After the necrotic tissue is
removed, laser treatment is performed at same endoscopic session, this time a few centimeters
distal to the original site on day 1. D, The same process is repeated on day 5. Treatment
progresses until the lumen is opened through the entire length of the neoplastic tissue. (A-D,
From Fleischer D. Lasers in gastroenterology. Am J Gastroenterol 1984; 79[5]:40615.)
(3473)Figure 437. Endoscopic views of laser-treated esophageal carcinoma. A, Squamous

cell carcinoma prior to treatment. The lesion occupies an eccentric position, with the lumen
visible in the lower left of the field. B, Result of first treatment. C-E, Appearance of tumor 1
week after first treatment. C and D, Proximal view. E, Distal view. (A-E, From the collection
of Dr. M. V. Sivak, Jr.)
Both the antegrade and retrograde approaches have advantages and disadvantages. The antegrade
method is technically easier because the laser energy can be delivered perpendicular to the target.
Additionally, a larger-diameter endoscope can be used with either a single large accessory channel or
two channels. Larger-channel instruments make it easier to evacuate smoke. With the antegrade
method, however, edema produced by thermal injury at the proximal treatment site may impede
advancement of the endoscope toward the more distal areas of the tumor. Charred and necrotic debris
may also block access to the distal aspect of the tumor, which therefore remains untreated. In such
cases, two or more sessions may be required to reestablish the lumen. Prior dilation of the tumor,
which entails a finite risk, is not required with the antegrade approach.
The retrograde approach is conceptually appealing because more of the tumor can be treated.
Treatment can often be completed in one session. Pietrafitta and Dwyer25(3474) emphasized the
value of a single session in which the occluded lumen is dilated using a guidewire and tapered
polyvinyl dilators, the endoscope is advanced to the distal margin of the tumor, and treatment is
performed in a retrograde manner. The approach is technically more difficult, and the possibility of
perforation is increased when the beam is aimed away from the axis of the lumen. This is often
necessary when the endoscope is advanced through a narrow and tortuous segment. When the
endoscope is wedged tightly within the tumor, it may difficult to deflect the tip, in which case aiming of
the laser beam may be less precise. Firing the laser when the tip of the waveguide is within the
accessory channel will damage the endoscope.
When there is complete obstruction of the esophagus or cardia, it is often difficult or impossible to
establish a lumen by dilation. Even if dilation can be performed, it may not be possible to establish a
diameter that will permit passage of a small-diameter endoscope. More than one dilation session may
be required. Bleeding from the surface of the tumor may impede tissue vaporization because the blood
absorbs some of the laser energy.26(3475) It is, in fact, sometimes difficult or impossible to even
identify the lumen at the proximal aspect of the tumor. In such cases, there are several options
including antegrade ELT. When the tumor is especially long and stenotic, a sound may be used to
identify the lumen. A biopsy forceps has been employed for this purpose.27(3476) A flexible,
2.5-cm-diameter ceramic probe that withstands the heat of the laser has been developed, although
experience with this device is limited.28(3477)
On balance, both methods (antegrade and retrograde) are acceptable, and each endoscopist must
decide which method is preferable.
Only noncontact fibers were used during the early developmental phases of ELT. As described
previously, the use of this type of waveguide made it possible for the endoscopist to achieve a range of
tissue effects simply by changing the distance between the tissue and the tip of the fiber. Moving the tip
of the fiber nearer and further from the tissue results, respectively, in a smaller or larger spot size.
When the tip is moved toward the tissue and the spot size becomes smaller, the energy density
increases; if the tip is moved away, the spot size increases and density decreases. An increase in
energy density results in a higher temperature and vice versa; tissue effect is determined by
temperature (see Table 431). If, however, a noncontact probe is placed in contact with the tissue, the
energy density will be high enough to destroy the tip of the fiber. Furthermore, soot, charred tissue, and
debris will accumulate on the tip of a noncontact fiber during the course of treatment. It is, therefore,
necessary to intermittently inspect the distal end of the noncontact fiber during a treatment session and
to clean it as necessary with a small brush. Distortion of the spot produced by the aiming beam usually
indicates that the tip should be cleaned. Because of the possibility of damage, it is advisable to have at
least three fibers available during a procedure.
Tissue effect is also governed by the power output of the laser. The power output of most commercially
available Nd:YAG is selectable, usually between 1 and 100 W. For a given spot size with a noncontact
fiber, a higher power setting results in a higher energy density (temperature) within the tissue. When
the temperature within the tissue is raised to 60 C, coagulation occurs; at temperatures of 100 C,
intracellular water boils and tissue vaporization takes place. Selection of a high- or low-power setting
offers benefits and disadvantages in either case.
Rapid vaporization is the objective with high-power settings. This approach results in large amounts of
smoke and soot within the lumen. In addition to obscuring visualization of the tumor, the use of rapid
vaporization techniques makes it necessary to clean the objective lens of the endoscope and the tip of
the waveguide. During a treatment session, the endoscope may have to be removed several times for
this purpose. A nasogastric tube can be fixed to the endoscope just proximal to the tip to remove
smoke and gas.29(3478) This is cumbersome and probably unnecessary. The maximum extent of
tissue coagulation and necrosis produced by ELT will not be immediately evident during a procedure.
Further necrosis will become manifest for several days after an ELT session, the net effect being
greater than that which was evident at the time of the procedure. Therefore, it is unnecessary and
theoretically hazardous to attempt to vaporize all visible malignant tissue during a given ELT session.
The desired tissue effect with low power is tissue coagulation. With this more gradual approach,
repeated tissue dbridement is usually required. The use of lower energies results in less smoke.
However, it generally takes longer to reestablish an adequate lumen, and in many cases, this delay in
palliation is undesirable. Although a matter of considerable debate, no data indicate that either high- or
low-power settings offer greater safety or provide a higher level of patient comfort.
The duration of a laser pulse is another factor that determines tissue effect. Virtually all lasers allow
selection of a pulse duration. For a given spot size, a longer pulse produces higher tissue temperatures
and vice versa. With a longer pulse, it is also possible to move the noncontact waveguide over the
surface of the tumor or to move the contact fiber within the tissue. With high power output and a long
pulse, movement of the waveguide may be used to cut away portions of the tumor. However, this type
of maneuver produces more smoke, which, in turn, restricts some of the effects that can be achieved,
owing to loss of visualization of the tumor. Lasers are in almost all cases fired by means of a foot
pedal. Even though a given pulse duration has been selected, a pulse of shorter duration may be
generated according to the length of time the laser has been fired by means of the foot pedal.
Therefore, the pulse duration set on the laser usually becomes the upper limit that the endoscopist
wishes to use.
In actual practice, an experienced endoscopist manipulates three variables (power output, spot size,
pulse duration) to achieve a range of effects.
There is no absolute limit on the amount of energy delivered to a tumor in a given ELT session. In
practice, however, this is rarely more than 15,000 joules (J), even with techniques that employ
high-power settings, with 3000 to 8000 J being more typical. The mean energy delivered per treatment
session in published reports has ranged from 2500 to 8500 J, with a range of about 500 to 12,000 J in
the individual patient.30,31(3479)
Soft, exophytic projections of a tumor can frequently be removed with a polypectomy snare with or
without electrosurgical cutting current. If this results in bleeding, hemostasis can be reestablished by
laser photocoagulation (lower power, large spot size).
Esophageal cancer has a penchant for extension beneath normal mucosa. Because the lateral depth
of a submucosal extension is not evident at endoscopy, ELT in such regions of a tumor may be
associated with an increased risk of perforation.32,33(3480) It has also been suggested that extensive
treatment of submucosal tumor may increase the risk of stricture formation.33(3481) The efficacy of
treating these areas has also been questioned.34(3482) In contrast to visible malignant tissue,
treatment of areas of submucosal spread may result in patient discomfort or pain because normal
mucosa is very sensitive to heat. Some investigators found that discomfort and pain are reduced by
injection of these submucosal areas with lidocaine.33,35(3483) The efficacy and safety of this
maneuver are not well established. Although submucosal tumor extension is problematic in many
respects, the options for treatment are limited and ELT may be beneficial.
When the technique of ELT for esophageal cancer was first described, three or more separate
sessions were often required to achieve luminal patency. Now, luminal patency can usually be
reestablished in two sessions or fewer. If more than one session is required, procedures are usually
spaced at 48-hr intervals. This allows tissue necrosis to develop fully and gives the patient some
respite between treatments. Treatment intervals as long as 1 week have been employed, but such
prolonged intervals during the initial phase of treatment are unnecessary and can delay relief of
dysphagia by several weeks. If ELT has been performed within the preceding few days, it is usually
necessary and advisable to dbride necrotic tumor before further treatment is performed. Otherwise, a
substantial portion of the laser energy will not reach viable tumor tissue. A sterilized cleaning brush is
suitable for abrading the tumor to remove necrotic surface debris. With increasing experience, the
number of ELT sessions required per patient probably decreases29(3484) and success rate
increases.36(3485)
ELT can be performed safely on an ambulatory basis in many patients.34(3486) Depending on the
clinical status of the patient and the technical difficulty of ELT, hospitalization may be necessary.
Results of Therapy
The first report of ELT for esophageal cancer appeared in 1982 when Fleischer et al.24(3487)
demonstrated that an Nd:YAG laser could be used to relieve obstruction. In addition to efficacy, this
study established the technical feasibility and safety of Nd:YAG laser photoablation. In all patients, the
goal of treatment was palliation. Since this seminal report, more than 30 other studies have reached
similar conclusions.25,26,30,32,33,3549(3488),5063(3489)
The results of laser therapy for obstructing esophageal cancer must be assessed in terms of technical
and functional outcomes. Technical success may be defined as the luminal diameter achieved by
treatment. The relative ease of passage of a standard endoscope (diameter 9 to 12 mm) through the
tumor is one method for assessing technical outcome.41,6466(3490) An improvement in the ability to
swallow is the simplest definition of functional success. Improvement by at least one grade on a 4- or
6-point scale (Table 433) is usually accepted as evidence of functional improvement. More relevant,
albeit less quantifiable, definitions include the ability to ingest solid food, to achieve adequate nutrition,
and to remain outside the hospital.16,31,41,43,66(3491) Functional success should also be evaluated
in terms of early and late phases because the initial improvement in the ability to swallow may not be
sustained throughout the course of a patient's illness. Technical and early functional success is defined
in most studies, but many lack data on long-term maintenance of the ability to swallow. Some
investigators have found that a satisfactory response to initial therapy is predictive of long-term
functional success.27,30(3492) The benefit of ELT may be greatest in patients in whom dysphagia is
severe and in those whose pretreatment performance status is good.43,64,67(3493)
TABLE 433 Typical Scale for Grading
Dysphagia
DYSPHAGIA GRADE
0
1
2
3
4
5
SYMPTOM
No dysphagia
Dysphagia for some solid food
Dysphagia for all solid food
Dysphagia for soft food
Dysphagia for liquids
Dysphagia for saliva
Technical success rates reported in various retrospective and prospective studies have ranged from 93
to 100%, with a average of around 95%. Early functional success rates vary from 70 to 100%, with an
average of approximately 75%. Obviously, reestablishment of a lumen with a diameter that should
allow normal swallowing does not always achieve the desired clinical result. Differences between
technical and functional success after ELT have been stressed by Mellow and Pinkas.43(3494) In their
study, luminal patency was achieved in 97% of patients, but this did not always equate with functional
success, the latter having been achieved in only 70% of the patients. The possible reasons for
discrepancies between technical and functional success include anorexia, painful tumor load, general
debility, pharyngeal dysphagia, treatment complications, and esophageal dysmotility as a result of
radiotherapy, ELT, or the effects of the tumor itself. In the study by Mellow and Pinkas,43(3495) poor
functional outcome correlated best with poor performance status at accession.
Late functional success was generally not equivalent to the initial relief of dysphagia. Depending on the
length of follow-up, dysphagia recurred in from 40 to 60% of patients. Symptoms usually reappeared
around 4 to 10 weeks after initial ELT. Once an adequate lumen was established, patients were
retreated in early report series only if dysphagia recurred or became more severe. Because recurrent
dysphagia might have been expected in relatively large numbers of patients, some endoscopists
preferred to reassess patients by endoscopy at regular intervals. The majority of patients responded to
further ELT if new tumor growth occurred.
Restoration of the ability to swallow, even if dysphagia is only partially relieved, can be expected to
improve the nutritional status of many patients with esophageal cancer. Increased survival of patients
undergoing ELT has been suggested by a few studies that utilized historical controls.41,68,69(3496) In
most studies, however, no effect on survival has been found and mean survival after ELT has ranged
from 14 to 20 weeks.29,30,35,53,66,70(3497)
Many studies concerned with ELT for esophageal carcinoma have focused on specific aspects of
therapy, including technique, selection of patients, and clinical parameters that predict outcome.
Tumors arising at the gastroesophageal junction can be technically difficult to treat if they cause an
acute angle to develop between the esophagus and the stomach.33,34(3498) Fleischer and
Sivak35(3499) performed palliative ELT in 15 patients with advanced adenocarcinoma of the gastric
cardia. Improvement in the ability to swallow was achieved in all cases. If obstruction had occurred
after a previous resection, it was generally possible to reestablish luminal patency in one treatment
session.
Fleischer and Sivak33(3500) determined that certain parameters affect initial outcome of ELT in a
study of 15 patients with gastroesophageal cancer. There was no difference in response to treatment
in relation to histopathologic tumor type, whether squamous cell or adenocarcinoma. It was technically
easier to treat mucosal rather than submucosal tumors. Results were least satisfactory and difficulty
was technically greatest with tumors involving the cervical esophagus. Tumors at the
gastroesophageal junction were problematic if they distorted the remaining natural lumen from a
vertical to a horizontal orientation. The best responses were achieved in patients with tumors in a
straight segment of the midesophagus or distal esophagus, particularly if the tumors were less than 5
cm in length. Some other studies found the response to ELT to be better in patients with relatively short
tumors.27,33,34(3501) But other investigators found no relation between tumor length and patient
response.30,31,43,59,71(3502) Rather than total length of the tumor, circumferential involvement,
especially over a long segment of the esophagus, may correlate more closely with response to
treatment.72,73(3503)
In a prospective study of short- and long-term efficacy of ELT that included 25 patients with advanced
esophageal cancer, Ahlquist et al.30(3504) achieved technical success in all patients and found that
the ability to swallow was improved in 80%. A single laser treatment provided adequate palliation in
more than half of the patients until the time of death. In those patients in whom restenosis occurred,
retreatment was necessary at a median of 3.2 months after initial treatment. Results of treatment were
best in patients with adenocarcinoma and those with a good appetite. Other studies found that patients
with adenocarcinoma have better early and later functional responses to ELT. Adenocarcinoma
appeared to be associated with a better response in the study of Naveau et al.;27(3505) but in the
study of Maunoury et al.,23(3506) the response was worse. No relationship has been found between
patient response and tumor histopathology in the majority of studies.29,31,33,34,43,72,73(3507)
Experience with sapphire-tipped contact waveguides is considerably less than that with noncontact
fibers.29,7478(3508) Some investigators suggested that the contact fiber is especially useful and
safer when there is complete obstruction of the esophagus and the initial goal of therapy is to establish
at least a minimum lumen.75(3509) Ell et al.46(3510) and Radford et al.52(3511) compared the use of
noncontact and contact waveguides with an Nd:YAG laser for treatment of malignant esophageal
obstruction. Ell et al.46(3512) concluded that the noncontact method was more effective in the average
case. However, in situations where there is tight stenosis, total obstruction, or overgrowth at the end of
an endoprosthesis, these investigators concluded that the contact method could be very helpful.
Radford et al.,52(3513) in a prospective randomized trial comparing the two techniques, concluded that
the contact method offered no advantage with respect to the number of treatment sessions required,
relief of dysphagia, or the occurrence of complications.
Ell and Demling58(3514) conducted an international inquiry concerning ELT for tumor stenoses of the
upper GI tract. They received responses from 47 centers that provided data on the treatment of
approximately 14,000 patients. Data provided by individuals who had treated more than 50 patients
were analyzed. Initial treatment was successful in 83%.
Tumors that arise in the cervical esophagus are the most difficult to treat, and patient response tends
to be less satisfactory than for tumors in other locations.33,34,73(3515) Even if a lumen of adequate
diameter is achieved, the clinical result may be unsatisfactory. Most investigators found recurrent
anastomotic carcinoma following surgical resection to be technically easy to treat. These tumors
respond especially well to laser therapy, perhaps because the cancer is relatively short,16,53(3516)
although some investigators dispute this point.23(3517)
Complications
In the international survey of Ell and Demling,58(3518) which included information on more than
14,000 patients who underwent ELT, the calculated complication rate was 4.1%. There were 49
reported complications including 25 perforations, 9 fistulas, 8 hemorrhages, and 7 sepses. The
mortality rate was 1%. In reported series, the procedure-related mortality for ELT is approximately
1%.16,27,29,58(3519)
Major complications of ELT include esophageal perforation, esophagopulmonary fistula, severe
bleeding, and sepsis. Complication rates vary from study to study. Esophageal perforation has been
reported in from 2 to 10% of ELT sessions.16,27,31,79(3520) Frequently, a combination of factors,
including dilation prior to ELT, is responsible for a perforation. Taken together, the frequency of other
major complications is about equal to that for perforation.16,58(3521) An esophagopulmonary fistula
occurs after 1 to 7% of ELT sessions.27(3522) However, fistulas develop spontaneously in from 6 to
15% of patients with esophageal cancer.80(3523) Patients undergoing ELT may also undergo
concomitant chemotherapy or radiotherapy. Therefore, it is often difficult or impossible to know
whether a fistula is the direct result of ELT.
Bacteremia is relatively common and occurs in from 34 to 40% of patients during and after
ELT.81,82(3524) This is thought to be due to passage of the endoscope through the tumor and not to
the application of laser energy.82(3525) Kohler and Riemann81(3526) found an incidence of
bacteremia of 34% in a series of 32 ELT procedures in 20 patients. Sepsis developed in 2 patients, 1
of whom died as a result of the infection. Guidelines for the prophylactic administration of antibiotics
published by the American Society for Gastrointestinal Endoscopy (ASGE) categorize procedures
according to risk for bacteremia.83(3527) Although not specifically mentioned in the ASGE guideline
statement, existing data indicate that the rate of bacteremia associated with ELT is comparable with
that for esophageal dilation. It would, therefore, be prudent to administer antibiotics to patients at
increased risk for infective endocarditis. The ASGE guidelines also classify clinical conditions into high,
intermediate, and no risk for endocarditis. Those conditions with a high risk include prosthetic heart
valves, a previous history of endocarditis, and surgically constructed systemic-pulmonary shunts or
conduits.
There have been no cases of laser-induced fires or explosions within the GI tract. For pulmonary
applications, however, there have been cases in which an endotracheal tube has caught fire, with
devastating injury to the patient. This type of complication is thought to be less likely in the
management of tumors of the upper GI tract because there is less oxygen in this environment and no
foreign body such as an endotracheal tube.
"Minor complications" are relatively common after ELT. These include low-grade fever, leukocytosis,
pain during treatment, mild chest discomfort for 1 to 2 days, pain or discomfort due to abdominal
distention, minor bleeding, and worsening dysphagia.33,35,36,66(3528) Laser photocoagulation
usually controls bleeding that arises during treatment. A decrease in the patient's ability to swallow
owing to acute self-limited edema usually persists for 1 to 2 days at most. Fibrous strictures may also
arise after repeated ELT, but these can usually be managed by dilation or placement of an
endoprosthesis.
Combined Therapy
Brachytherapy (afterloading) is a form of radiotherapy in which a container of radioactive material is
placed within an esophageal cancer. Nd:YAG laser therapy and afterloading with iridium-192 have
been used in combination in several studies.50,55,8486(3529) Forty-eight patients were treated with
endocavitary loading after prior recanalization with the laser in an uncontrolled study by Bader et
al.50(3530) Dysphagia was permanently relieved in 77%; restenosis occurred in 23%. There were 2
fatal complications: death resulted from acute myocardial infarction 4 days after laser therapy in 1
patient, and massive tracheal hemorrhage occurred after laser therapy and dilation in another patient
with malignant invasion of the trachea. In another study of 21 patients, dysphagia improved in all
cases, with no procedure-related complications.85(3531)
Shmueli et al.84(3532) treated 32 patients with intraluminal radiotherapy following initial palliation by
ELT. Patients with squamous cell carcinoma (n = 14) also received external beam radiotherapy. In this
uncontrolled trial, 9 patients who survived a median of 22 weeks did not require further treatment.
However, a median of three additional ELT sessions was required in the remaining 23 patients in
whom the median survival was 40 weeks (4 to 102 weeks). Further dilation became necessary in 11
patients who developed fibrous strictures. Combination fibrous/malignant strictures developed in an
additional 12 patients who were treated by dilation and ELT. Placement of an esophageal prosthesis
was eventually necessary in 6 patients.
A favorable effect on malignant dysphagia was not found in the trial of Hagenmller et al.55(3533)
Their preliminary results revealed that, although the dysphagia-free intervals increased, restenosis
occurred in virtually all patients. In a later prospective, randomized trial from this same group of
investigators,86(3534) the interval between establishment of luminal patency by ELT and recurrence of
dysphagia was almost twice as long in patients who also underwent brachytherapy, although only for
those patients with epidermoid carcinoma. No benefit of afterloading therapy was found in the
subgroup of patients with adenocarcinoma, and these patients required a significantly greater number
of endoscopic procedures per month of life. Furthermore, 21% of patients treated by ELT and
brachytherapy developed esophagitis, compared with none of those treated by ELT alone. No other
benefit was evident in either group. Sander et al.86(3535) concluded that afterloading therapy was
beneficial only in patients with epidermoid carcinomas who had a high performance status.
Preliminary data from a study by Sargeant et al.87(3536) suggested that combined ELT and external
beam radiation increases the interval between maintenance laser treatments and that it may prolong
life compared with historical controls treated by ELT alone.
Comparison with Other Methods of Palliation
ELT has been compared with other forms of palliative therapy for esophageal and gastroesophageal
cancer.
In the prospective trial of Low and Pagliero,88(3537) 23 patients were randomized to undergo ELT or
brachytherapy. Early and late improvement in dysphagia and performance scores were approximately
the same in each group. Complication rates were also not substantially different. ELT has been
compared with insertion of an endoprosthesis in the palliation of patients with obstructing esophageal
cancer in several studies.26,51,89(3538) In general, these two methods have been found to be equally
effective for relief of dysphagia. However, the complication rate for ELT is substantially lower than that
for placement of a prosthesis in several studies.26,59,89(3539) In the study by Carter et al.,90(3540)
however, three perforations occurred in the laser-treated group of patients (one fatal) versus none in
the group that underwent stent placement. Studies of ELT versus endoprosthesis placement often
exclude certain patients who may be better served by one or another form of therapy based on
technical considerations alone. Placement of an endoprosthesis is usually preferable when there is a

tracheopulmonary fistula, extrinsic tumor compression, rapid growth of the malignancy that would
necessitate frequent laser sessions, and a lack of experience with ELT. Laser therapy is generally
preferable if the tumor is eccentric, noncircumferential, soft, and highly exophytic; if the malignant
stenosis is extremely tight and impassable; if the tumor is located close to the cricopharyngeus; and if
there is significant bleeding from the tumor surface. Comparative studies of ELT versus stent
placement are discussed in Chapter 42: Palliation of Malignant Obstruction: Dilation and Peroral
Prosthesis.
No prospective comparative trials have studied ELT versus implantation of a self-expandable metal
prosthesis. Tumor overgrowth at either end of an endoprosthesis is often managed by laser
photoablation.91,92(3541)
Mellow and Pinkas41(3542) compared survival for a group of patients with squamous cell carcinoma of
the esophagus treated by ELT with that for a group of patients treated by conventional radiation
therapy. Survival for the group treated by ELT was significantly longer whether measured from the
onset of radiation therapy or from the time of recurrent symptoms after radiotherapy. These results
remain to be verified by other trials.
From an initial group of 88 patients with squamous cell carcinoma of the esophagus, Reed et
al.93(3543) prospectively randomized 27 patients with nonresectable and obstructing tumors to one of
three treatment groups: (1) insertion of a prosthesis alone; (2) insertion of a prosthesis followed by
radiation therapy; and (3) ELT followed by radiation therapy. There were no significant differences
between treatment groups with respect to survival or relief of dysphagia. However, 8 complications
occurred among patients who underwent prosthesis placement compared with none in the
laser-treated group, a statistically significant difference. Treatment time was significantly greater for
patients who underwent combined radiotherapy and ELT compared with that for patients who received
a prosthesis, although only 1 patient in the combined therapy group required repeat photoablation.
From this study, Reed et al.93(3544) concluded that: (1) ELT should be used first to establish a lumen
because of its lower complication rate and efficacy; (2) this should be followed by radiation therapy,
preferably with a reduced total dosage or change in fractionation schedule to reduce treatment time;
and (3) stent placement should be reserved for patients with fistulas or those with extrinsic
compression caused by the tumor.
ELT using an Nd:YAG laser was compared with injection of polidocanol in 37 consecutive patients with
obstructing tumors in the prospective, randomized trial of Angelini et al.31(3545) Injections (1 to 2 ml at
0.5 cm along the length of the tumor) were performed weekly with a 10- to 12-mm-long sclerotherapy
needle until a satisfactory result was achieved. Thereafter, further injection sessions were performed
as needed. There was no difference in functional result between the two groups, although further
treatment after the initial series was performed significantly less often in patients who had undergone
injection therapy. Polidocanol injection therapy is inexpensive, readily available, technically easy, and
well tolerated by patients. However, experience with this technique is limited and its safety and efficacy
are not well established.
Gastric Cancer
ELT for gastric carcinoma may be performed for palliation or cure in the case of early gastric cancer
(EGC). However, ELT has also been performed as definitive treatment for EGC.
Advanced Gastric Cancer
Palliative treatment of advanced gastric cancer is, in many ways, similar to that for esophageal cancer.
Tumors that arise in the cardia of the stomach, although gastric in origin, result in obstruction and
dysphagia. Endoscopic palliative treatment is therefore comparable with that for esophageal cancers.
Tumors that arise at or near the pylorus can also cause obstructive symptoms.
Spinelli et al.94(3546) reported a series of 308 patients with upper GI cancers treated with ELT that
included 72 patients with gastric obstruction and 54 patients with cancers at an anastomotic site.
Luminal patency was restored in 94% of the patients, and ELT improved swallowing in 74% of the
patients. By life table analysis, 1-year survival was 23% when luminal patency was achieved, and 7%
when treatment failed.
Early Gastric Cancer
One of the more exciting applications of laser therapy is the definitive treatment of EGC. Most reports
come from Japan, and the results are very encouraging. Definitive treatment depends on early and
accurate diagnosis. EGC has also been treated by endoscopic mucosectomy. Whereas ELT does not
provide a tissue specimen for histopathologic diagnosis and staging, the entire malignant tumor is
usually recovered after mucosectomy. EUS is therefore frequently used for tumor staging prior to ELT.
A study by Wyman et al.95(3547) evaluated contact Nd:YAG laser radiation in vitro in nine specimens
of advanced gastric cancer obtained immediately after resection. The purpose of this study was to
define treatment parameters and to correlate energy settings with the degree of tissue injury. Tissue
vaporization extended into the submucosa in 91% of the lesions treated, an observation relevant to the
definitive treatment of EGC by ELT.
There is an increasing number of reports of ELT as definitive treatment for EGC; these come mainly
from Japan where gastric cancer is often found in an early and potentially curable stage.96104(3548)
Watanabe et al.103(3549) reported complete eradication in 37 of 42 patients (88%) with EGC treated
by ELT. Hiki et al.105(3550) performed ELT in 34 patients with EGC. Complete eradication was
achieved in 22, but residual malignancy was detected in 11 patients, and the status of 1 patient was
undetermined. In the series of 111 patients with EGC treated by Yasuda et al.104(3551) with ELT, 12
patients underwent subsequent laparotomy at which residual tumor was found in 25% of patients.
Seventy-three of the 99 patients who did not undergo surgery were followed for a mean of 2.7 years
(range 1 to 7 years), during which time 81% remained free of tumor. Yasuda et al.104(3552) indicated
that ELT has been supplanted by endoscopic mucosal resection (mucosectomy), as the latter
technique provides a tissue specimen for staging. Depending on the location and size of the cancer,
however, mucosectomy was not always complete and was therefore supplemented by ELT. Of the 13
patients in whom both procedures were performed, 11 had complete removal of the tumor.
Criteria for the selection of patients for ELT have been suggested by Hiki et al.106(3553) based on a
review of a large number of cases of EGC. Because of the technical limitations of endoscopic therapy,
the cancer should be of a size that can be eliminated completely by the initial treatment. Lesions less
than 1 cm in diameter that are either protruding or depressed without ulceration are suitable. Ideally,
there should be no evidence of deep tissue invasion or lymph node metastasis. However, ELT may be
suitable in patients with more advanced lesions if they have severe coexisting disease that increases
the risk of surgery, are of advanced age, or refuse surgery.
Bleeding Neoplasms
Little published information exists concerning ELT for bleeding tumors in the upper GI tract. The
experience of the author is that laser therapy is seldom of benefit if the neoplastic lesion is diffuse and
larger than 6 to 8 cm in diameter. However, for selected focal lesions, such as those caused by
metastatic melanoma, carcinoid, and renal or breast neoplasms, ELT is often beneficial.
Other Upper GI Neoplasms
Tumors of the duodenum and main duodenal papilla are rare, but there are instances of laser
treatment where such tumors caused obstruction or blood loss. Both Bowers and Sivak [personal
communications] have treated such patients. The patient treated by Sivak had recurring episodes of
pancreatitis due to a villous tumor of the main duodenal papilla. The tumor remained after local
surgical resection was attempted, and duodenoscopic laser therapy was undertaken with the prospect
of decreasing episodes of pancreatitis and destroying premalignant tissue. The patient responded over
the short term; however, bouts of pancreatitis returned, and the benign lesion was extirpated by means
of a modified Whipple procedure. Lasers have been used to treat villous tumors of the duodenum that
do not involve the main papilla.107(3554)
Hoi and Schneider108(3555) used ELT to treat a patient with a granular cell tumor of the esophagus.
Naveau et al.109(3556) ablated a large fibrovascular polyp in the esophagus. ELT has been used to
treat Barrett's epithelium, and transient regression has been reported.110(3557) However, this
application should be considered investigational, as data on efficacy and safety are extremely limited.
Miscellaneous Applications
ELT has been utilized for a variety of purposes other than the palliation of patients with obstruction or
bleeding due to malignant tumors. These applications have included incision of various benign
strictures, endoscopic papillotomy (see Chapter 60: Endoscopic Papillotomy),111(3558) lithotripsy of
bile duct stones (see Chapter 62: Calculus Disease of the Bile Ducts),112114(3559) and drainage of
pseudocysts (see Chapter 76: Endoscopic Management of Pancreatic Disease).115(3560)
Sander et al.116(3561) describe the use of the laser to treat two peptic strictures of the esophagus and
two postoperative strictures in the rectosigmoid colon. There are case reports of ELT for a variety of
other obstructing lesions, including an antral web,117(3562) pyloric stenosis,118(3563) congenital
duodenal diaphragm in an adult,119(3564) and anastomotic strictures of the esophagus and colon
following surgical stapling procedures.120(3565)
Laser Safety
There are two considerations with regard to laser treatmentsafety of the patient and safety of the
health care personnel utilizing the laser. Complications of ELT and patient safety are discussed
previously. All considerations with respect to the safety of health care personnel also apply to patients.
Injury will occur if an errant laser beam strikes the eye of a health care worker or patient. This type of
injury may also occur if the beam is reflected back via the image guide bundle of a fiberoptic
endoscope. The former case, which is extremely unlikely provided reasonable precautions are
observed, has not been reported. The latter situation is of much less concern with the use of electronic
endoscopy systems.
Despite the lack of reported complications, the importance of ophthalmic safety must not be minimized.
A laser beam striking the eye could cause considerable damage, depending on its intensity. The CO2
laser can damage the corneal surface by desiccation. Exposure of the retina to argon or Nd:YAG laser
energy may cause damage, with resultant blind spots in the visual field. Ophthalmic safety levels for
laser exposure have been determined by the American National Standards. For typical endoscopic
treatment, levels of Nd:YAG laser reflected light from the mucosa through the endoscope could be in
the range of 100 to 200 mW/cm2.
If ELT is performed with a fiberoptic instrument, a filter must be placed over the ocular of the
instrument. In the case of the argon laser, the filter material is orange, which color-blocks the
blue-green light produced by the argon laser. Because this filter will color-distort the endoscopic view, it
is put in place by mechanical means during the actual firing of the laser. With the Nd:YAG laser, a
transparent filter is used. This fits over the ocular at the proximal end of the fiberoptic endoscope or
may be inserted into the laser itself. With video endoscopes, filters are also required, albeit not for eye
protection. Without a filter, the laser beam, on firing, would overwhelm the charge-coupled device
within the electronic endoscope and cause a "white-out," thereby eliminating visual observation when
the laser is fired.
There has always been a concern that infectious agents may be released in the plume of smoke
generated when a laser is fired at tissue. Although there are no instances of definite complications
occurring as a result of this for lesions of the GI tract, it has been the topic of much concern. In 1967,
Hoye et al.121(3566) described viable tumor cells in the plume after treatment with an Nd:YAG laser.
Oosterhuis122(3567) found nonviable cells after treatment of melanotic lesions with a CO2 laser.
Bellina et al.123(3568) used a CO2 laser to treat patients with condylomata acuminata. The presence
of intact but dehydrated cells suggested the possibility that an infectious agent might remain viable
within the cytoplasm or nucleus. However, metabolic studies failed to demonstrate biologic activity.
Voorhies et al.124(3569) treated brain tumors in rats with a CO2 laser and found no evidence of cell
viability in the plume of laser smoke. Fleischer and Reeves125 analyzed smoke produced by Nd:YAG
laser treatment of esophageal cancer in humans. They concluded that there were no identifiable
hazards from airborne cancer cells. Particles were extremely small (0.1 m). The presence of viral
particles could not be excluded. A study by Garden et al.126(3570) raised the issue of infectious vapor
release during CO2 laser treatment. Two models were used, and in both instances, intact papilloma
virus was present in plumes of smoke generated with laser therapy.
Although no evidence exists of transmission of disease from the plume of smoke that is generated with
laser treatment, currently available data are not adequate to state categorically that it is not possible
that transmission of disease could occur. On the other hand, attention to suctioning the smoke that
occurs after laser treatment and the use of appropriate exhaust systems make this author feel that the
treatment of any endoscopic laser treatment of GI diseases should not be impeded by this concern.
Lasers in Perspective: Present and Future

Although the aforementioned applications of lasers in GI endoscopy have been significant, some would
say that the role of lasers in the management of GI diseases has fallen short of initial expectations.
Lasers are capable of generating light of a single wavelength. Of the thousands of wavelengths that
make up the electromagnetic spectrum, only a few have been available for use in GI endoscopy. There
are several reasons for this. Aside from cost considerations, the inability to deliver laser energy via a
flexible endoscope is a major limitation. For example, the CO2 laser could have many applications in
the treatment of digestive diseases, but a practical method of delivering the laser energy via a flexible
waveguide has not been developed. For various reasons, therefore, the endoscopist has had only a
limited selection of wavelengths. Current ELT takes advantage of the fact that the laser can precisely
focus relatively large amounts of energy. All major applications are based on the thesis that focused
thermal energy will be beneficial. Little attention has been given in a practical sense to the primary
characteristic of a laserthat is, that it produces monochromatic light. It remains to be seen whether
this characteristic can be used to achieve highly selective nonthermal effects in contrast to the current
use of lasers to produce indiscriminate thermal effects in tissue.
What is the future of lasers for GI endoscopy? It is likely that more work will be undertaken with
diagnostic applications. It has been established that laser-induced fluorescence (LIF) spectroscopy can
be used to distinguish between normal and abnormal tissue (see Chapter 17: Biomedical Tissue
Spectroscopy and Chapter 18: Spectroscopic Diagnosis and Treatment of Gastric
Cancer).127130(3571) LIF, which is based on excitation of tissue with highly specific wavelengths of
light, could be the basis of imaging systems that screen for GI malignancy in patients at high risk. The
use of tissue-sensitizing agents, as in photodynamic therapy, is another area of development with
many potential applications for endoscopic diagnosis and therapy (see Chapter 19: Photodynamic
Therapy). The development of free-electron lasers will allow selection of specific wavelengths. This
opens a number of possibilities; for example, it may be that a certain wavelength will be preferentially
absorbed by a specific tumor or pathologic tissue. This principle has already been utilized for the
selective treatment of vascular abnormalities with dye lasers.131(3572) Instead of using high powers
for generalized thermal destruction, selective absorption with small percentages of the energy leads to
nonthermal destruction of the vascular lesion. It may be possible to treat selectively below the mucosal
surface, without damaging the surface tissue, for therapy of submucosal lesions. "Smart-laser"
technology is another intriguing combination of diagnostic and therapeutic applications of
lasers.132(3573) A diagnostic beam is aimed at the area where the pathologic process exists. A tissue
"signature" is created by various interactions between the laser light and the tissue. When this
information is fed through a multichannel optical analyzer computer, different patterns can be
distinguished. These data provided by the computer are used to target a therapeutic laser beam, via
the same fiber used for analysis, at specific pathologic tissue (Figure 438).
(3574)Figure 438. Smart laser technology. After a pattern is generated, the laser can be
programmed to fire a therapeutic beam if an appropriate signal is given.
ELT, to date, has been an important addition to therapeutic endoscopy, and has provided the
groundwork for many other novel treatment modalities. All of the initial expectations have not been
realized, but it is likely that many more applications of laser physics in GI endoscopy await
development.
Tumor Probes
Investigators have queried whether alternatives to ELT could be utilized to destroy obstructing
neoplastic tissue. The most promising and commonly employed alternative to date has been the
bipolar coagulating (BICAP) tumor probe developed by Johnston et al.133(3575) This device,
conceived along the lines of the Eder-Peustow dilator system, is less expensive and more portable
than laser systems and it may offer some technical advantages. The tumor probe is similar in many
respects to multipolar electrocoagulating probes used to treat bleeding lesions (see Chapter 28:
Thermal Contact Methods for Endoscopic Hemostasis). The effect on tissue achieved with the tumor
probe is primarily electrosurgical and is relatively localized near the tissue surface.
Equipment
The probe comprises three sections: a spring tip connected to an olive that is attached to a long,
flexible shaft with an electric connection at its proximal end (Figure 439A). The spring tip, which
serves as the leading point for the device, can be unscrewed from the olive, but the shaft does not
detach from the olive. The olive contains the electrocoagulating unit, which operates on the principle of
bipolar electrocoagulation (see Chapter 28: Thermal Contact Methods for Endoscopic Hemostasis).
Olives (with shaft attached) are available in various diameters; the smallest is 6 mm, and the largest is
15 mm (Figure 439B). The flexible shaft, 60 cm in length, has circular marks that indicate distance
between a mark and the olive. At the proximal end of the shaft is a Y connector, one branch of which is
aligned with the shaft. A central lumen runs through this section of the Y connector, the shaft, the olive,
and the spring tip, so that the entire device can be passed over a guidewire. The other arm of the Y
connector attaches to an electric cord that runs to a 50-W electrosurgical generator (Figure 439C).
(3576)Figure 439. A, Spring tip of a bipolar coagulating tumor probe (BICAP). B, Active
electrode of the tumor probe is seated in the olive or head of the probe. The olives are
available in different sizes. C, At the opposite end of the shaft is an electric adaptor that
connects to an electrosurgical generator.
Technique
The selection and evaluation of patients for treatment with the tumor probe are the same as that for
ELT (see previously). The tumor probe has been used for treatment of obstructing tumors of the
esophagus and rectum. There are two general approaches to treatment: retrograde and antegrade.
Retrograde Technique
Endoscopy is performed with a small-caliber instrument to define the length and topography of the
tumor. It is ideal if the endoscope can be passed to the distal margin of the esophageal tumor and
beyond. If this cannot be accomplished, dilation is carried out to allow subsequent passage of the
tumor probe. A guidewire is placed under endoscopic or fluoroscopic guidance (see Chapter 42:
Palliation of Malignant Obstruction: Dilation and Peroral Prosthesis). Dilation should be carried out to
the maximum diameter possible within the margins of safety; then, the tumor probe is passed over the
guidewire. Using measurements obtained endoscopically and with fluoroscopic confirmation, one can
determine when the probe has reached the proximal margin of the tumor. There will usually be some
resistance as the probe is advanced through the length of the tumor and beyond its distal margin. With
fluoroscopic guidance, the probe is pulled back to the distal margin of the tumor. This position is
confirmed by reference to endoscopic measurements, fluoroscopic observations, and the palpable
"feel" of the probe as it impacts against the distal margin of the tumor. Depending on the length of the
active portion of the olive electrode, sequential burns are carried out along the length of the tumor.
Therefore, if the active electrode is 1 cm in length and the tumor is 5 cm in length, approximately five
station burns will be required to deliver thermal energy along the entire length of the tumor.
Fluoroscopic observations are made at each station. The length of time that energy is applied is
predetermined for each probe and is based on experiments in animals. The approximate time at each
station is 15 sec.
Treatment of an esophageal cancer with a tumor probe is demonstrated in Figure 4310. Treatment at
the distal margin and over most of the length of a tumor cannot be observed directly at endoscopy;
however, when the probe is pulled to the proximal margin, a small-caliber endoscope can be passed
alongside the probe to observe the effects of treatment at the proximal margin. After treatment, the
tumor probe is removed and an endoscope can be passed into the treated area to observe the effects
of therapy. Often, a circumferential white coagulum is evident over the entire length of the tumor. On
some occasions, erythema as well as coagulation results in a zebra-like striping pattern (Figure
4311). Generally, one or two sessions will be required to treat the entire tumor.
(3577)Figure 4310. Endoscopic view of tumor probe therapy at the proximal level of an
esophageal cancer.
(3578)Figure 4311. Endoscopic view after tumor probe treatment. Therapy has produced a striped
effect.
Antegrade Technique
Treatment begins at the proximal margin of the tumor. Once the proximal margin of the tumor has
been defined endoscopically and a guidewire has been placed, the tumor probe is passed alongside
the endoscope and the first treatment is delivered at the proximal margin under endoscopic
visualization. After this first burn, the tumor probe is advanced distally to the next station, using the
markings on the shaft and endoscopic observation to position the olive. As this proceeds, fluoroscopy
is used to confirm the position of the probe. The tumor probe is gradually advanced through the
neoplasm at 1-cm station intervals. After treatment is completed at the distalmost station, the probe is
withdrawn and the endoscope is used to visualize the treated tumor.
Results of Treatment
Fleischer et al.134(3579) described the tissue effects of the tumor probe in a canine model designed to
investigate the mechanism of perforation and stricture formation, the two major untoward effects. With
the standard 360-degree probe, endoscopic coagulation and stricture formation were observed acutely
but strictures resolved with time. Histologically, tissue injury was evident at a depth of 1.5 mm, that is,
approximately one third of the esophageal wall thickness. A prototype probe with electrocoagulating
elements arranged over a 270-degree radius of the olive, developed to avoid circumferential
coagulation and thereby prevent stricture formation, was also evaluated. Stenosis was less likely with
this probe, but a fistula developed in one dog. Although the 270-degree probe appears useful in
concept, its actual application was limited by a lack of precision in placing the probe in the desired
location/orientation.
A prototype electrocoagulation probe was used for palliation of esophageal cancer in a preliminary
multicenter pilot study that included 20 patients.135(3580) The mean number of initial treatment
sessions was 1.7. Dysphagia and tumor lumen diameter improved after treatment. The mean
treatment interval between the initial and a subsequent treatment session was approximately 2
months. Major complications included esophageal pulmonary fistulas in 2 patients and delayed
hemorrhage in 2 patients.
McIntyre et al.,136(3581) in a prospective, randomized study, compared bipolar tumor probe therapy
with endoprosthesis insertion in the palliative treatment of advanced esophageal malignancy. Of the 17
patients treated with tumor probe, 1 suffered an esophageal perforation following dilation and 1
developed a fistula 4 weeks after treatment. Patients required retreatment at a median of 28 days. Of
the 13 patients who underwent insertion of a prosthesis, 2 suffered perforation. Dysphagia improved in
both treatment groups, this being somewhat better in the tumor probe group. The probe was felt to be
especially advantageous for treatment of the most proximal lesions. The prosthetic tube provided
longer-lasting relief of dysphagia.
Jensen et al.137(3582) compared low-power Nd:YAG laser treatment with the tumor probe for
palliation of esophageal malignant strictures. Of the 28 patients treated, the first 14 underwent ELT and
the last 14 were treated with the tumor probe. In the latter group, all patients had coagulation of
malignant strictures in one session. Prior to treatment, patients could ingest only liquids; 86% of
patients in both groups could eat a soft or solid diet after initial treatment. Treatment results were not
statistically different during a median follow-up and survival of 16 weeks. Treatment-related
esophageal strictures developed in 21% of patients treated by ELT. The only fistula developed in a
patient with a noncircumferential tumor that was treated with a tumor probe.
The comparative advantages and disadvantages of ELT and palliation with tumor probes were
reviewed by Reilly and Fleischer.138(3583) The tumor probe was considered to have these
advantages: it is portable and inexpensive, and palliation should be facilitated because a large area of
the tumor can be treated in one session. The disadvantages include a need for fluoroscopy in most
cases and the fact that the probe delivers a burn over the 360-degree circumference of the olive.
Because of the latter characteristic, Reilly and Fleischer138(3584) recommended that the probe not be
used in patients with tumors that are not circumferential.
Therapeutic Clogology
The laser and tumor probe are examples of devices and techniques that are applied in the new
interdisciplinary subspecialty dedicated to the unclogging of biologic cylinders. The tasks are similar for
the GI endoscopist as they are for the cardiologist working on coronary arteries, for the pulmonologist
or thoracic surgeon trying to unplug an airway endoscopically, for the urologist relieving a ureteral
obstruction, for the peripheral vascular physician recanalizing the iliac artery, and for the interventional
radiologist attempting to restore flow in the biliary tree. In addition, their tools (stents, balloons, dilators,
pusher tubes, lasers, tumor probes, baskets) are more similar than different. Particularly in the area of
laser technology, the issues in all of the subspecialties in which clogologists work are similar.139(3585)
The essence is that the pathologic process must, in some way, be separated from the normal tissue
and that the treatment must be selectively delivered to the pathologic process. Already, photodynamic
therapy, endoscopic ultrasonography, video endoscopy with computer analysis, spectral fluoroscopy,
and antibody labeling are being used by different subspecialties. It is logical and efficient that more
cross-fertilization should occur between subspecialties.
REFERENCES
1.
Dwyer RM, Yellin AE, Craig J, et al. Gastric hemostasis by laser phototherapy in man. A
preliminary report. JAMA 1976;236:13834.
2.
Nath G, Gorisch W, Kiefhaber P. First laser endoscopy via a fiberoptic transmission system.
3.
Frhmorgen P, Reidenbach J, Bodem R, et al. Experimental examinations on laser
4. Bown SG, Salmon PR, Kelly DF, et al. Argon laser photocoagulation in the dog stomach. Gut
1979;20:6807.
5. Dwyer RM, Haverback BJ, Bass M, Cherlow J. Laser-induced hemostasis in the canine
stomach. JAMA 1975;231:4869.
6. Silverstein FE, Auth DC, Rubin CE, Protell RL. High power argon laser treatment via standard
endoscopes. I. A preliminary study of efficacy in control of experimental erosive bleeding.
7. Silverstein FE, Protell RL, Piercey J, et al. Endoscopic laser treatment. II. Comparison of the
efficacy of high and low power photocoagulation in the control of severely bleeding
experimental ulcers in dogs. Gastroenterology 1977;73:4816.
8. Silverstein FE, Protell RL, Gulacsik C, et al. Endoscopic laser treatment. III. Development and
testing of gas-jet-assisted argon laser waveguide in control of bleeding experimental ulcers.
9. Silverstein FE, Protell RL, Gilbert DA, et al. Argon vs. neodymium YAG laser photocoagulation
of experimental canine gastric ulcers. Gastroenterology 1979;77:4916.
10.
Waitman AM, Grant DZ, DeBeer R, Chryssanthou C. Endoscopic laser photocoagulation:
Comparison of argon and neodymium-YAG [Abstract]. Gastrointest Endosc 1979;25:52.
11.
Nath G, Gorisch W, Kreitmair A, Kiefhaber P. Transmission of a powerful argon laser beam
through a fiberoptic flexible gastroscope for operative gastroscopy. Endoscopy 1973;5:213.
12. Butler ML, Morris W. Use of argon laser in the treatment of experimentally induced upper
gastrointestinal bleeding in primates. Gastrointest Endosc 1978;24:1178.
13.
Jensen DM, Machicado GA, Tapia JI, et al. Bipolar electrocoagulation and argon laser
photocoagulation of colonic lesions [Abstract]. Gastrointest Endosc 1980;26:69.
14.
Cummins L. Laser tissue interaction. In Fleischer D, Jensen D, Bright-Asare P (eds).
Therapeutic Laser Endoscopy in Gastrointestinal Disease. Boston: Martinus Nijhoff,
1983;938.
15.
Enderby CE. Medical laser fundamentals. In Fleischer D, Jensen D, Bright-Asare P (eds).
Therapeutic Laser Endoscopy in Gastrointestinal Disease. Boston: Martinus Nijhoff,
1983;18.
16. Fleischer D. Laser therapy in gastrointestinal diseases. Adv Intern Med 1987;32:40728.
17. Fleischer D. Endoscopic laser therapy for esophageal cancer: Pres-ent status with emphasis
on past and future. Lasers Surg Med 1989;9:616.
18.
Hashimoto D. Development and early clinical applications of a lateral-aiming laser probe.
Lasers Med Sci 1987;2:258.
19. Frank F. Technical prerequisites and safety considerations for the use of the Nd:YAG laser in
gastroenterology. Endoscopy 1986;18(suppl 1):69.
20. Earlam R, Cunha-Melo JR. Oesophageal squamous cell carcinoma: I. A critical review of
surgery. Br J Surg 1980;67:38190.
21. Orringer MB. Palliative procedures for esophageal cancer. Surg Clin North Am
1983;63:94150.
22. Cox J, Bennett JR. Light at the end of the tunnel? Palliation for oesophageal carcinoma. Gut
1987;28:7815.
23. Maunoury V, Brunetaud JM, Cochelard D, et al. Endoscopic palliation for inoperable
malignant dysphagia: Long term follow-up. Gut 1992;33:16027.
24. Fleischer D, Kessler F, Haye O. Endoscopic Nd:YAG laser therapy for carcinoma of the
esophagus: A new palliative approach. Am J Surg 1982;143:2803.
25. Pietrafitta JJ, Dwyer RM. Endoscopic laser therapy of malignant esophageal obstruction. Arch
Surg 1986;121:395400.
26. Buset M, des Marez B, Baize M, et al. Palliative endoscopic management of obstructive
esophagogastric cancer: Laser or prosthesis?. Gastrointest Endosc 1987;33:35761.
27. Naveau S, Chiesa A, Poynard T, Chaput JC. Endoscopic Nd-YAG laser therapy as palliative
treatment for esophageal and cardial cancer. Parameters affecting long-term outcome. Dig
Dis Sci 1990;35:294301.
28. Ell C, Hochberger J, Lux G, Riemann JF. Laser-resistant guide probe for laser treatment of
endoscopically impassable tumour stenoses. Endoscopy 1986;18:278.
29. Bown SG, Hawes R, Matthewson K, et al. Endoscopic laser palliation for advanced malignant
dysphagia. Gut 1987;28:799807.
30. Ahlquist DA, Gostout CJ, Viggiano TR, et al. Endoscopic laser palliation of malignant
dysphagia: A prospective study. Mayo Clin Proc 1987;62:86774.
31. Angelini G, Pasini AF, Ederle A, et al. Nd:YAG laser versus polidocanol injection for palliation
of esophageal malignancy: A prospective, randomized study. Gastrointest Endosc
1991;37:60710.
32. Bown SG. Endoscopic laser therapy for oesophageal cancer. Endoscopy 1986;18(suppl
3):2631.
33. Fleischer D, Sivak MV Jr. Endoscopic Nd:YAG laser therapy as palliation for esophagogastric
cancer: Parameters affecting initial outcome. Gastroenterology 1985;89:82731.
34. Lightdale CJ, Zimbalist E, Winawer SJ. Outpatient management of esophageal cancer with
endoscopic Nd:YAG laser. Am J Gastroenterol 1987;82:4650.
35. Fleischer D, Sivak MV Jr. Endoscopic Nd:YAG laser therapy as palliative treatment for
advanced adenocarcinoma of the gastric cardia. Gastroenterology 1984;87:81520.
36. Ell Ch, Riemann JF, Lux G, Demling L. Palliative laser treatment of malignant stenoses in the
upper gastrointestinal tract. Endoscopy 1986;18:216.
37. Buset M, Dunham F, Baize M, et al. Nd:YAG laser, a new palliative alternative in the
management of esophageal cancer. Endoscopy 1983;15:3536.
38. Swain CP, Bown SG, Edwards DAW, et al. Laser recanalization of obstructing foregut cancer.
Br J Surg 1984;71:1125.
39.
Krasner N, Beard J. Laser irradiation of tumours of the oesophagus and gastric cardia. Br
Med J Clin Res 1984;288:82930.
40. Groisser VS. YAG laser therapy of gastrointestinal tumors. Gastrointest Endosc
1984;30:3112.
41. Mellow MH, Pinkas H. Endoscopic therapy for esophageal carcinoma with Nd:YAG laser:
Prospective evaluation of efficacy, complications, and survival. Gastrointest Endosc
1984;30:3349.
42. Cello JP, Gerstenberger PD, Wright T, et al. Endoscopic neodymium-YAG laser palliation of
nonresectable esophageal malignancy. Ann Intern Med 1985;102:6102.
43. Mellow MH, Pinkas H. Endoscopic laser therapy for malignancies affecting the esophagus
and gastroesophageal junction. Arch Intern Med 1985;145:14436.
44. Mathus-Vliegen EMH, Tytgat GNJ. Laser photocoagulation in the palliative treatment of upper
digestive tract tumors. Cancer 1986;57:3969.
45. Goldberg SJ, King KH. Endoscopic Nd:YAG laser coagulation as palliative therapy for
obstructing esophageal carcinoma. Am J Gastroenterol 1986;81:62933.
46.
Ell Ch, Hochberger J, Lux G. Clinical experience of noncontact and contact Nd:YAG laser
therapy for inoperable malignant stenoses of the oesophagus and stomach. Lasers Med Sci
1986;1:1436.
47.
Fleischer D. Endoscopic laser therapy for esophagogastric cancer: Clinical experience as a
background for new approaches [Abstract]. Gastroenterology 1987;92:1391.

48. Pietrafitta JJ, Carstens MH, Dwyer RM. Endoscopic laser therapy for the treatment of
malignant esophageal obstruction. Lasers Surg Med 1987;7:48790.
49. Takemoto T, Okazaki Y, Otani T, et al. Palliative laser therapy. Endoscopy 1987;19(suppl
1):125.
50. Bader M, Dittler HJ, Ultsch B, et al. Palliative treatment of malignant stenoses of the upper
gastrointestinal tract using a combination of laser and afterloading therapy. Endoscopy
1986;18(suppl 1):2731.
51.
Lux G, Groitl H, Ell C. Tumor stenoses of the upper gastrointestinal tractTherapeutic
alternatives to laser therapy. Endoscopy 1986;1(suppl 1):3743.
52. Radford CM, Ahlquist DA, Gostout CJ, et al. Prospective comparison of contact with
noncontact Nd:YAG laser therapy for palliation of esophageal carcinoma. Gastrointest Endosc
1989;35:3947.
53. Rutgeerts P, Vantrappen G, Broeckaert L, et al. Palliative Nd:YAG laser therapy for cancer of
the esophagus and gastroesophageal junction: Impact on the quality of remaining life.
54.
Mosquet L, Bruentaud JM, Mudry J, et al. Are lasers useful in the palliative treatment of
digestive cancers?. Lasers Surg Med 1984;3:340.
55. Hagenmller F, Sander C, Sander R, et al. Laser and endoluminal 192-iridium radiation.
56. Fleischer D. The Washington Symposium on Endoscopic Laser Therapy, April 1819, 1985.
57.
Fleischer D. The laser in gastroenterology with emphasis on upper gastrointestinal
malignancies. Symposium held at the World Congress of Gastroenterology, Sao Paolo, Brazil,
September 12, 1986. Gastrointest Endosc 1986;33:11920.
58. Ell C, Demling L. Laser therapy of tumor stenoses in the upper gastrointestinal tract: An
international inquiry. Lasers Surg Med 1987;7:4914.
1991;100:130310.
60. Riemann JF, Ell C, Lux G, Demling L. Combined therapy of malignant stenoses of the upper
gastrointestinal tract by means of laser beam and bougienage. Endoscopy 1985;17:438.
61. Schulze S, Fischerman K. Palliation of oesophagogastric neoplasms with Nd:YAG laser
treatment. Scand J Gastroenterol 1990;25:10247.
62. Zittel TT, Allgaier D, Grund KE. Laser therapy for esophageal cancer. Results and additional
endoscopic treatments. Surg Endosc 1994;8:1096100.
63. Holting T, Friedl P, Schraube N, et al. Palliation of esophageal cancerOperative resection
versus laser and afterloading therapy. Surg Endosc 1991;5:48.
64. Brennan FN, McCarthy JH, Laurence BH. Endoscopic Nd-YAG laser therapy for palliation of
upper gastrointestinal malignancy. Med J Aust 1990;153:2731.
65. Fleischer D. A comparison of endoscopic laser therapy and BICAP tumor probe therapy for
esophageal cancer. Am J Gastroenterol 1987;82:60812.
66. Fleischer D, Kessler F. Endoscopic Nd:YAG laser therapy for carcinoma of the esophagus: A
new form of palliative treatment. Gastroenterology 1983;85:6006.
67. Alexander GL, Wang KK, Ahlquist DA, et al. Does performance status influence the outcome
of Nd:YAG laser therapy of proximal esophageal tumors?. Gastrointest Endosc
1994;40:4514.
68. Karlin DA, Fisher RS, Krevsky B. Prolonged survival and effective palliation in patients with
squamous cell carcinoma of the esophagus following endoscopic laser therapy. Cancer
1987;59:196972.
69. Siegel HI, Laskin KJ, Dabezies MA, et al. The effect of endoscopic laser therapy on survival in
patients with squamous-cell carcinoma of the esophagus. Further experience. J Clin
70. Barr H, Krasner N. Prospective quality-of-life analysis after palliative photoablation for the
treatment of malignant dysphagia. Cancer 1991;68:16604.
71. Mason RC, Bright N, McColl I. Palliation of malignant dysphagia with laser therapy:
Predictability of results. Br J Surg 1991;78:13467.
72. Alderson D, Wright PD. Laser recanalization versus endoscopic intubation in the palliation of
malignant dysphagia. Br J Surg 1990;77:11513.
73. Brunetaud JM, Maujoury V, Cochelard D, et al. Parameters affecting laser palliation in
patients with advanced digestive cancers. Lasers Surg Med 1989;9:16973.
74.
Daikuzono N, Suzuli S, Tajiri H, et al. Laserthermia: A new computer controlled contact
Nd:YAG system for interstitial local hyperthermia. Lasers Surg Med 1088;8:2548.
75. Houghton A, Mason R, Allen A, et al. Nd:YAG laser treatment in the palliation of advanced
oesophageal malignancy. Br J Surg 1989;76:9123.
76. Sankar MY, Joffe SN. Endoscopic contact Nd:YAG laser resectional vaporization (ECLRV)
and esophageal dilatation (ED) in advanced malignant obstruction of the esophagus. Am Surg
1991;57:25968.
77. Steger AC, Hira N. The palliative endoscopic treatment of inoperable oesophagogastric and
rectal cancers: A low power direct contact laser technique. Ann R Coll Surg Engl
1987;69:1668.
78. Suzuki S, Aoki J, Shiina Y, et al. New ceramic endoprobes for endoscopic contact irradiation
with Nd:YAG laser: Experimental and clinical applications. Gastrointest Endosc
1986;32:2826.
79. Murray FE, Bowers GJ, Birkett DH, et al. Palliative therapy of advanced esophageal
carcinoma: An alternative perspective. Am J Gastroenterol 1988;83:8169.
80. Boyce HW Jr. Palliation of advanced esophageal cancer. Semin Oncol 1984;11:18695.
81. Kohler B, Riemann JF. Incidence of bacteremia after endoscopic laser treatment of stenosing
processes in the upper gastrointestinal tract. Am J Gastroenterol 1987;82:10268.
82. Wolf D, Fleischer D, Sivak MV Jr. Incidence of bacteremia with elective upper gastrointestinal
endoscopic laser therapy. Gastrointest Endosc 1985;31:24750.
83. The American Society for Gastrointestinal Endoscopy. Antibiotic prophylaxis for
84. Shmueli E, Srivastava E, Dawes PJDK, et al. Combination of laser treatment and intraluminal
radiotherapy for malignant dysphagia. Gut 1996;38:8035.
85. Renwick P, Whitton V, Moghissi K. Combined endoscopic laser therapy and brachytherapy for
palliation of oesophageal carcinoma: A pilot study. Gut 1992;33:4358.
86. Sander R, Hagenmller F, Sander C, et al. Laser versus laser plus afterloading with
iridium-192 in the palliative treatment of malignant stenosis of the esophagus: A prospective,
randomized, and controlled study. Gastrointest Endosc 1991;37:43340.
87. Sargeant IR, Loizou LA, Tobias JS, et al. Radiation enhancement of laser palliation for
malignant dysphagia: A pilot study. Gut 1992;33:1597601.
88. Low DE, Pagliero KM. Prospective randomized clinical trial comparing brachytherapy and
laser photoablation of palliation of esophageal cancer. J Thorac Cardiovasc Surg
1992;104:1739.
89. Hahl J, Salo J, Ovaska J, et al. Comparison of endoscopic Nd:YAG laser therapy and
oesophageal tube in palliation of oesophagogastric malignancy. Scand J Gastroenterol
1991;26:1038.
90. Carter R, Smith JS, Anderson JR. Laser recanalization versus endoscopic intubation in the
palliation of malignant dysphagia: A randomized prospective study. Br J Surg
1992;79:116770.
91. Barr H, Krasner N, Raouf A, Waker RJ. Prospective randomised trial of laser therapy only and
laser therapy followed by endoscopic intubation for the palliation of malignant dysphagia. Gut
1990;31:2528.
92. Sargeant IR, Loizou LA, Tulloch M, et al. Recanalization of tube overgrowth: A useful new
indication for laser in palliation of malignant dysphagia. Gastrointest Endosc 1992;38:1659.
93. Reed CE, Marsh WM, Carlson LS, et al. Palliation of advanced esophageal cancer. Ann
94. Spinelli P, Dal Fante M, Mancini A. Endoscopic palliation of malignancies of the upper
gastrointestinal tract using Nd:YAG laser: Results and survival in 308 treated patients. Lasers
Surg Med 1991;11:5505.
95. Wyman A, Sweetland HM, Rogers K. Parameters for and safety of contact Nd:YAG laser
irradiation of early gastric cancer: An in vitro study. Gastrointest Endosc 1991;37:1704.
96. Imaoka W, Ida K, Katoh T, et al. Is curative endoscopic treatment of early gastric cancer
possible?. Endoscopy 1987;19(suppl 1):711.
97. Korenaga D, Watanabe A, Saito A, et al. Laser treatment for poor-risk patients with early
gastric cancer: Post treatment pathology. Eur J Surg Oncol 1991;17:3168.
98. Sakita T, Koyama S, Ishii M, et al. Early cancer of the stomach treated successfully with an
endoscopic neodymium-YAG laser: Report of a case. Am J Gastroenterol 1981;76:4415.
99. Suguro M, Hasegawa T, Suzuki S, Hanyu F. Clinical evaluation of laser endoscopy for the
treatment of gastric tumors. Surg Endosc 1987;1:1318.
100. Hiki Y, Sakakibara Y, Mieno H, et al. Endoscopic treatment of gastric cancer. Surg Endosc
1991;5:113.
101. Takemoto T. Laser therapy of early gastric carcinoma. Endoscopy 1986;18(suppl 1):326.
102.
Wang JT, Lin JT, Shun CT, Wang TH. Non-surgical radical treatment of early gastric cancer
by endoscopic contact Nd-YAG laser: Report of a case. Taiwan I Hsueh Hui Tsa Chih
1990;89:3147.
103. Watanabe M, Sugino Y, Imai Y, et al. Revaluation of endoscopic laser therapy for treatment of
early cancer in the stomach. Keio J Med 1993;42:2068.
105. Hiki Y, Shimao J, Yamao Y, et al. The concepts, procedures, and problems related in
endoscopic laser therapy of early gastric cancer. A retrospective study on early gastric
cancer. Surg Endosc 1989;3:16.
106. Hiki Y, Shimao H, Mieno H, Sakakibara Y. Laser therapy for early upper gastrointestinal
carcinoma. Surg Clin North Am 1992;72:57180.
107.
Hunter JM. Endoscopic laser applications in the gastrointestinal tract. Surg Endosc
1989;69:114766.
108.
Hoi M, Schneider S. Endoscopic laser therapy of a granular cell tumor of the esophagus.
Lasers Surg Med 1988;8:169.
109. Naveau S, Bedossa P, Mallet L, et al. Successful ablation of a large fibrovascular polyp of the
esophagus by endoscopic Nd:YAG laser therapy. Gastrointest Endosc 1989;35:2546.
110. Brandt LJ, Kauvar DR. Laser-induced transient regression of Barrett's epithelium. Gastrointest
Endosc 1992;38:61922.
111.
Sander R, Poesl M. Endoscopic papillotomy with Nd:YAG laser. Lasers Surg Med
1986;6:3658.
112. Sander R, Poesl M, Frank F. An Nd:YAG laser with a water guided laser beam. Gastrointest
Endosc 1988;34:3369.
113. Ell Ch, Lux G, Hochberger J, et al. Laser lithotripsy of common bile duct stones. Gut
1988;29:74651.
114.
Orii K, Nakahara A, Takase Y, et al. Choledocholithotomy by YAG laser. Surgery
1981;94:1202.
115.
Remine S, Setzer S, Shapshay S. Gastric mucosal dissection using the Nd:YAG laser contact
probe. Treatment of pancreatic pseudocysts. Lasers in general surgery. Surg Clin North Am
1992;72:789.
116. Sander R, Psl M, Spuhler A. Management of non-neoplastic stenoses of the GI-tract.
Endoscopy 1984;16:14951.
117. Al-Kawas FH. Endoscopic laser treatment of an obstructing antral web. Gastrointest Endosc
1988;34:34951.
118. Lambert M, Faintuch J. Laser recanalization of pyloric stenosis. Lasers Surg Med
1989;9:2825.
119.
Gertsch P, Mossmann R. Endoscopic treatments of a congenital duodenal diaphragm.
120. Chen PC, Wu CS, Chang-Chien CS, Liaw YF. YAG laser endoscopic treatment of an
esophageal and sigmoid stricture after end-to-end anastomosis stapling. Gastrointest Endosc
1984;30:25860.
121. Hoye RC, Ketcham AS, Riggle GC. The airborne dissemination of viable tumor by high energy
neodymium laser. Life Sci 1967;6:11925.
122.
Oosterhuis JW. Tumor Surgery with the CO2 Laser: Studies with the Cloudman S91 Mouse
Melanoma. Grningen: Veenstra-Visser Offset. 1969.
123. Bellina JH, Stjernholm RL, Kurpel JE. Analysis of plume emissions after papovavirus
irradiation with the carbon dioxide laser. J Reprod Med 1982;27:26870.
124.
Voorhies RM, Lavyne MH, Strait TA, Shapiro WR. Does the CO2 Fleischer DE, Reeves W.
Analysis of smoke produced by endoscopic Nd:YAG laser treatment of esophageal cancer: Is
airborne spread of cancer cells a concern? [Abstract]. Gastrointest Endosc 1987;33:180.
126. Garden JM, O'Banion MK, Shalnitz LS, et al. Papillomavirus in the vapor of carbon dioxide
laser-treated verrucae. JAMA 1988;259:1199202.
127. Kapadia CR, Cutruzzola FW, O'Brian K, Stetz ML. Laser-induced fluorescence spectroscopy
on human colon mucosa. Gastroenterology 1990;99:1507.
128. Schomacker KT, Frisoli JK, Compton CC, et al. Ultraviolet laser-induced fluorescence of
colonic polyps. Gastroenterology 1992;102:115560.
129. Cothren RM, Richards-Kortum R, Sivak MV Jr, et al. Gastrointestinal tissue diagnosis by
1990;36:10511.
130.
Yakshe PN, Bonner RF, Cohen P, et al. Laser-induced fluorescence spectroscopy may
distinguish colon cancer from normal human colon [Abstract no. 142]. Gastrointest Endosc
1989;35:184.
131. Nishioka N, Tan O, Bronskin B, et al. Selective vascular coagulation of the rabbit colon using
flashlamp-excited dye laser. Gastroenterology 1988;95:125864.
132.
Leon MB, Smith PD, Bonner RT. Laser angioplasty delivery systems: Design consideration. In
White RW, Spundfest WS (eds). Lasers in Cardiovascular Disease. Chicago: Yearbook
Medical. 1987.
133.
Johnston J, Quint R, Petruzzi C, Namihiri Y. Development and experimental testing of a large
BICAP probe for palliative treatment of obstructing esophageal and rectal malignancy
[Abstract no. 127]. Gastrointest Endosc 1985;31:156.
134.
Fleischer DE, Randard R, Kamath R, et al. Endoscopic and histologic observations following
BICAPt probe application in a canine model. Gullet 1991;1:12630.
135. Johnston JH, Fleischer D, Petrini J, Nord HJ. Palliative bipolar electrocoagulation therapy of
obstructing esophageal cancer. Gastrointest Endosc 1987;33:34953.
136. McIntyre AS, Morris DL, Sloan RL, et al. Palliative therapy of malignant esophageal stricture
with the bipolar tumor probe and prosthetic tube. Gastrointest Endosc 1989;35:5315.
137. Jensen DM, Machicado G, Randall G, et al. Comparison of low-power YAG laser and BICAP
tumor probe for palliation of esophageal cancer strictures. Gastroenterology
1988;94:126370.
138. Reilly HF III, Fleischer DE. Palliative treatment of esophageal carcinoma using laser and
tumor probe therapy. Gastroenterol Clin North Am 1991;20:73142.
139. Fleischer D. Lasers in gastroenterology. Am J Gastroenterol 1984;79:40615.
Chapter 44 Hiatus Hernia and Peptic Diseases of the Esophagus

(3586)
H. WORTH BOYCE, JR., M.S., M.D.
Anatomy of the Distal Esophagus and Gastric Cardia

Knowledge of the internal anatomy of the esophagus and the influences on it by extrinsic, contiguous
structures is essential for accurate interpretation of endoscopic findings. The importance of precise
measurements of distance from a fixed point of reference, such as the incisor teeth or alveolar ridge,
cannot be overemphasized. Evaluation of the therapeutic response by repeated endoscopy depends
on precision in recording the findings during the prior endoscopic procedure. Measurement of lesion
size and its accurate location within quadrants of the luminal circumference also are essential.
Nowhere is this practice more important than in treating the patient from whom esophageal biopsy
specimens have been procured, one of which reveals malignant change. Careful observation and
recording of the biopsy or lesion site reduce the risk of being unable to return to the same location to
obtain biopsies or to assess the response to therapy, or both. In this chapter, distance measurements
use the incisor teeth or alveolar ridge as the point of reference.
The esophagus begins at the cricopharyngeus muscle in the neck at the level of the C5-C6 vertebral
interspace, which is about 16 cm distal to the incisor teeth (Figure 441). In the adult, the esophagus
extends for approximately 25 cm, ending at its junction with the stomach 2 to 3 cm below the
diaphragmatic hiatus. The potential diameter of the lumen is approximately 25 mm throughout its
length. The first 3 to 4 cm in the cervical region are relatively collapsed because of pressure from
surrounding structures. The esophagus enters the superior mediastinum at the 20-cm level and opens
readily with air insufflation, this response being enhanced by the negative intrathoracic pressure. At 23
cm from the incisor teeth is a visible arterial pulsation through the left anterior wall of the esophagus in
the area of contact with the aortic arch. Slight compression of the esophageal wall may be seen at this
level. At 26 to 27 cm, it comes into contact with the posterior wall of the left main stem bronchus.
Rarely is there any intraluminal evidence of this relationship.
(3587)Figure 441. Esophageal anatomy and relationships with adjacent organs. Distances in
centimeters are the average for an adult as measured from the incisor teeth with a flexible
endoscope. LESlower esophageal sphincter.
At about 30 cm, a and v waves transmitted by the left atrium through the left anterior esophageal wall
are easily detected. Contact with the left atrium extends from about 30 to 35 cm. At this level, the distal
esophagus curves gently leftward and anteriorly to the proximal end of the lower esophageal sphincter,
which usually begins between 37 and 40 cm.
The esophagus in the average patient passes through the diaphragmatic hiatus at approximately 38 to
40 cm and ends distal to the squamocolumnar junction (SCJ) at about 41 to 42 cm. The normal level of
the SCJ may vary 1 to 2 cm, depending on the type of endoscope used and the care with which
measurements are made. More accurate distance measurements are made during endoscope
withdrawal than during insertion, because bowing of the insertion tube is minimized. Distances
measured with a rigid endoscope are 1 to 2 cm less than measurements made with flexible
endoscopes.
If minimal air insufflation of the proximal and mid-esophagus is used during endoscopy, the closed
lower esophageal sphincter region may be readily demonstrated. At the point of closure of the proximal
end of the sphincter, several (usually four to six) longitudinal, symmetric mucosal folds can be seen to
disappear in the center of the lumen (Figure 442). This closure produces a rosette appearance, with
the lumen precisely centered where the longitudinal folds converge. The tone of the lower esophageal
sphincter relaxes with primary or secondary peristalsis and opens in response to gentle insufflation. As
the closed, normal esophageal sphincter is approached with the endoscope, it relaxes with gentle
pressure, and passage through the sphincter encounters little or no resistance. As the high-pressure
sphincter zone relaxes, the SCJ can be identified 1 to 2 cm beyond, and it is possible to see into the
tubular cavity of the proximal stomach, the cardia. The proximal end of the lower esophageal sphincter
is most easily demonstrated in patients with achalasia because of dilation of the body of the esophagus
and typical hypertonicity of the closed sphincter.
(3588)Figure 442. Endoscopic view of the upper limits of the closed lower esophageal
sphincter, the so-called rosette.
The esophagus passes through the diaphragmatic hiatus in the adult at 38 to 40 cm from the incisor
teeth. The level of the hiatus margin is not as readily demonstrated in patients with normal anatomy as
in those who have a hiatus hernia; however, it is usually possible to determine the level with relative
precision. As the lumen is gently insufflated with air, the patient is asked to sniff or inhale rapidly, at
which time the diaphragmatic hiatus margin moves inferiorly quickly or gradually, depending on the
breathing maneuver used to demonstrate its location.
The squamous mucosal lining of the esophagus is pearly pink or pinkish gray and contrasts sharply
with the orange-red color of the gastric columnar epithelium (Figure 443A). The esophageal mucosa
is only slightly transparent and reflects light moderately. The gastric mucosa has a glistening surface
because of the presence of mucus but is more transparent and consequently absorbs light well.
Gastric mucosa requires more light than other anatomic structures for adequate photography.
(3589)Figure 443. Endoscopic view of the squamocolumnar junction (SCJ) just proximal to
the level of the diaphragmatic hiatus. A, Note the dramatic color difference between the
squamous and the columnar epithelium. B, Retroverted view of SCJ in normal position
(arrow) just proximal to the angle of His in the gastric cardia. C, Linear vessels disappear at
the SCJ. The level of disappearance of the distal ends of these vessels correlates closely with
the normally positioned junction of columnar and squamous epithelium. Note that the gastric
folds extend to the SCJ.
The junction of the squamous and columnar epithelium appears after minimum inflation as a slightly
irregular or undulating line, called the Z line. This line of demarcation between the two types of mucosa
is readily identifiable in the absence of pathologic changes. If there is uncertainty about the location of
this junction, it can be dramatically demonstrated by application of several milliliters of 2.5% Lugol's
solution through an endoscopic catheter (Figure 443B).1(3590) This stains the esophageal mucosa in
about 30 sec. Staining often clears within 10 min (see Chapter 12: Chromoscopy). In addition to
surface characteristics and color, the distal extent of the esophageal squamous epithelium is clearly
demarcated by multiple, linear, and frequently branching small blood vessels that abruptly disappear
just proximal to the upper end of the gastric folds of the cardia at the junction of squamous and
columnar epithelium (Figure 443C).
A crescentric protrusion may be detected in some patients just distal to the SCJ in the left lateral or
greater curvature aspect of the cardia. This elevation has been called the semilunar-shaped fold or the
crescent of the cardia, and it is thought to correspond to the horseshoe-shaped grouping of oblique
muscle fibers, called the sling muscle fibers, that are draped around the anatomic muscular union of
esophagus and stomach, forming the angle of His (see Figure 441). This structure is usually 39 to 41
cm from the incisor teeth, depending on the phase of respiration and the degree of bowing in a flexible
endoscope.
After the endoscope is passed into the proximal stomach, a retroversion maneuver should be
performed to view the cardia and fundus from below (Figure 444). In the normal setting, the insertion
tube of the endoscope can be seen coming through a snugly fitting intraabdominal segment of
esophagus. The snug fit in this region is sustained throughout respiration and during moderate
insufflation of the stomach, except that transient relaxation in response to primary or secondary
peristalsis can be detected. The classic snug appearance of the region is almost always demonstrated
in patients with achalasia because of the increased tone in the lower esophageal sphincter segment
(Figure 445). During retroversion with insufflation, the SCJ can be seen in some patients from several
millimeters to 1 cm above the distal margin of this intraabdominal segment. In a patient with a hiatus
hernia, the area is patulous, and the diameter depends on the diameter of the esophageal
diaphragmatic hiatus.
(3591)Figure 444. Correlative drawings explaining the anatomy in a retroverted view of the
gastroesophageal junction (GEJ). A, Drawing from an endoscopic photograph (Figure 445)
of the normal cardia region. B, Cross-sectional diagram showing endoscope position and
sagittal view of the GEJ.
(3592)Figure 445. Endoscopic retroverted view of the GEJ and the fundus of the stomach.
The distal margin of squamous epithelium is seen just below the endoscope.
Normal Endoscopic and Radiologic Correlations

The distal esophageal lumen fills evenly and symmetrically during barium contrast
radiography.2,3(3593) The column of barium begins to taper just proximal to the lower esophageal
sphincter. With complete barium filling in the region of the gastroesophageal junction, the proximal
margin of the lower esophageal sphincter is demonstrated as a smooth narrowing of the barium
column beginning 1 to 2 cm above the diaphragmatic hiatus. After passage of barium through this
region and complete closure of the lumen, a barium-free segment that is 3 to 4 cm long straddles the
diaphragmatic hiatus; this segment corresponds to the closed lower esophageal sphincter. The angle
of His on the greater curvature aspect demarcates the distal end of the sphincter region, also called
the submerged or abdominal segment of the esophagus (Figure 446A; see also Figures 444 and
445). The normally located SCJ cannot be identified radiographically. However, in patients with
herniation of the proximal stomach through the hiatus, a lower esophageal ring can be demonstrated,
the location of which corresponds with the level of the intrathoracically displaced SCJ (Figure 446B).
(3594)Figure 446. A, Profile of the GEJ showing normal relationships and the position of
the lower esophageal sphincter relative to the diaphragmatic hiatus. B, Profile of the GEJ
herniated into the thorax via the esophageal hiatus of the diaphragm. The contour of the lower
esophageal sphincter is altered readily by air or barium to produce this contour when it is
displaced into the thorax above a hiatus hernia. (A and B, Modified from Goyal RK: The
lower esophageal sphincter. Viewpoints Dig Dis 1976; 8:1.)
Earlier radiographic and anatomic studies caused much confusion about this area relative to the
structures that are seen in normal individuals and in patients with hiatus hernias. In a normal individual
there are no rings, no asymmetric bulges, and no bulbous contour of the distal esophagus. These
alterations are seen only in a patient with a hiatus hernia. Wolf3(3595) redefined the radio-graphic
anatomy of the gastroesophageal region and clearly demonstrated the significance of the various
radiographic contours that may be observed in this area.
One of the reasons there has been so much confusion concerning the normal anatomy is that the
sliding hiatus hernia is common in the population. Many radiologists have insisted on the classic
definition of a hiatus hernia, the criteria for which do not apply to patients with small hernias. It is
imperative for the physician who performs endoscopy to understand the proper radiographic criteria for
hiatus hernia as defined by Wolf so that accurate endoscopic-radiographic correlations may be made.
Hiatus hernias, deemed by some in the past as unrelated to acid reflux disease, may be found in
virtually every patient with reflux esophagitis, stricture, columnar-lined esophagus, and esophageal
ulcers when proper radio-graphic and endoscopic criteria are used. The high incidence of hiatus
hernias in patients with complicated gastroesophageal reflux disease is well documented.4(3596)
Endoscopic Diagnosis of Hiatus Hernia

The first step in endoscopic diagnosis of hiatus hernia is recognition of the important intraluminal
landmarks used in defining this entity. Despite argument over the criteria for endoscopic diagnosis, a
consensus exists concerning the major diagnostic points.510(3597)
Under normal circumstances, the SCJ has been observed to migrate during swallowing and with
respiration by as much as 2 cm above the diaphragmatic hiatus. Dagradi et al.,5,10(3598) Trujillo et
al.,6,7(3599) and others agree that displacement of the SCJ more than 2 cm proximal to the
diaphragmatic hiatus is abnormal. This opinion correlates well with the modern radiographic criteria for
hiatus hernia as reported by Wolf.3(3600) In patients with hiatus hernia, the diaphragmatic hiatus and
cardia are often rather patulous, and the lumen opens with minimal insufflation. In many instances, this
area is so widely patent that it can be seen from the proximal esophagus. The anatomic and radiologic
correlation of findings at the gastroesophageal junction are demonstrated diagrammatically in Figure
443A and B, respectively.
After the SCJ is identified, the maneuvers previously described to localize the diaphragmatic hiatus
should be used. With minimal degrees of herniation, displacement of the SCJ proximal to the
diaphragmatic hiatus by more than 2 cm is the primary endoscopic criterion for diagnosis of a hiatus
hernia (Figure 447). A hernia pouch is not identifiable when this minimum criterion is applied, but in a
moderately sized or large hiatus hernia, the gastric mucosal folds can be seen running proximally over
the hiatus margin into the bulbous cavity of the distended hernia pouch. When the patient inspires, the
diaphragmatic margin moves downward to give the appearance of gastric mucosa gliding upward over
this margin into the chest. If the patient sniffs, there is an abrupt, short, downward motion of the
diaphragm, producing the appearance of gastric mucosa gliding over the hiatus margin into the chest.
(3601)Figure 447. Endoscopic view of the SCJ displaced above the hiatus, representing a
small hiatus hernia. The junction has straightened and a dynamic lower esophageal B ring has
been formed under the influence of luminal distention with air.
After observing the diaphragmatic hiatus from above, the endoscope should be passed into the
proximal stomach and retroverted. In most instances of hiatus hernia, the initial view from this vantage
reveals a widened diaphragmatic hiatus with gastric mucosa lying loosely over the hiatus margin and
gastric folds running upward into the hiatus hernia pouch (Figure 448; see also Figure 447). If the
patient is asked to sniff or take a deep breath, the diaphragmatic margin descends as the stomach
glides in a proximal direction over its edge. Several gastric folds usually can be seen running along the
herniated pouch's greater curvature or posterior wall aspects to terminate just distal to the SCJ. These
folds normally terminate within 5 to 10 mm of the normal location of the SCJ. This termination point can
be utilized endoscopically and radiographically as a marker for the approximate location of the normal
SCJ. The cephalad margin of the longitudinal gastric folds in the hernia pouch correspond to the level
of the gastroesophageal muscular junction as well.
(3602)Figure 448. A, Drawing of an endoscopic retroverted view of a hiatus hernia. Note the
location of a lower esophageal B ring at the level of the SCJ. B, Drawing of a sagittal view of
a hiatus hernia showing endoscope position and correlative anatomy for the retroverted view
shown in A.
In some patients with a hiatus hernia and normal lower esophageal sphincter tone, the esophageal wall
in the region of the esophageal sphincter is closed snugly around the endoscope (Figures 449 and
4410). Some relaxation may occur with primary and secondary peristalsis or after greater degrees of
air insufflation (Figure 4411). In patients with reflux esophagitis, especially those with a
columnar-lined esophagus who tend to have the lowest sphincter pressures, considerable free space
exists around the endoscope in the region of the lower esophageal sphincter, just proximal to the
hernia pouch.
(3603)Figure 449. Retroverted endoscopic view into a hiatus hernia showing the SCJ and
the closed lower esophageal sphincter segment around the endoscope. The fold or ridge
appearance in the right foreground is due to gastric mucosa lying over the left margin of the
diaphragmatic hiatus.
(3604)Figure 4410. A, Antegrade endoscopic view of a classic, static lower esophageal B

(Schatzki) ring with hiatus hernia beyond. B, Retrograde view in same patient of the B ring
from inside the hiatus hernia. The lower esophageal sphincter is closed snugly around the
endoscope above the ring.
(3605)Figure 4411. Retroverted endoscopic view into a distended hiatus hernia pouch that
clearly demonstrates a lower esophageal B ring from below. The lighter color of the gastric
mucosa is due to the closer proximity of the illuminating endoscope to the gastric mucosa than
to the esophageal mucosa. The lower esophageal sphincter, located about 1 to 2 cm proximal
to the ring, is relaxed around the endoscope.
When viewing the region of the SCJ from below in a patient with normal sphincter pressure, closure of
the proximal end of the lower esophageal sphincter can be observed. The point of maximum closure is
1 to 2 cm above the SCJ (see Figure 449). This level of closure corresponds with the so-called
esophageal A ring or sphincter contraction ring in location and contour, as seen during radiography and
endoscopy (see Figure 446B).
With a hiatus hernia of moderate or larger size, the retroverted endoscope can be pulled back to the
level of the diaphragmatic hiatus or even a short distance into the hernia pouch, affording a close-up
view of the hernia pouch and the SCJ from below.
It is important to observe carefully and record the characteristics of the distal esophagus and proximal
stomach in patients with hiatus hernia who have no gastroesophageal reflux sequelae. The location of
the diaphragmatic hiatus in relation to the proximal stomach, the level of the SCJ, and the proximal
extent of the gastric mucosal folds in the hernia pouch are characteristics used in the precise
endoscopic diagnosis of hiatus hernia and reflux sequelae, including the earliest stages of a
columnar-lined esophagus(Barrett's esophagus). The levels of these landmarks should be recorded on
every esophagoscopy report.
In evaluating patients with a known or suspected hiatus hernia, it is appropriate to use more than
minimal air insufflation. The radiologist makes use of changes in patient position and increased
amounts of barium to demonstrate the same anatomy. Because patients with hiatus hernia tend to
belch frequently, a considerable amount of insufflated air may be required to adequately demonstrate
the landmarks discussed. Sliding esophageal hiatus hernias are common, particularly in older patients.
Nevertheless, it is important to attempt to demonstrate this entity by radiography or endoscopy or both
in all patients, especially those with upper gastrointestinal or pulmonary symptoms. If the endoscopic
criteria are used with the radiographic criteria of Wolf,3(3606) it is not difficult to recognize a sliding
hiatus hernia. It is the clinician's responsibility to determine whether this finding in the individual patient
is significant. If neither the radiologist nor the endoscopist diligently reports this entity, the patient's
physician may not suspect a reflux-related etiology for atypical or obscure complaints. The important
relationship between the anatomic defect of a hiatus hernia and gastroesophageal junction
incompetence is being documented and better understood by sophisticated physiologic
studies.11(3607)
The endoscopic appearance of the esophagus and proximal stomach after antireflux surgery is
discussed in Chapter 52: Endoscopy in the Postoperative Upper Gastrointestinal Tract.
Lower Esophageal B Ring or Schatzki Ring

With inflation of the distal esophagus, the SCJ gradually changes from its usual serrated appearance
to a straight line in many patients (Figure 4212A-D). Adding a bit more air or a Mueller maneuver (i.e.,
inspiration against a closed glottis) by the patient to enhance negative intrathoracic pressure and
thereby increase the effect of positive intraluminal pressure maximally distends the hiatus hernia pouch
distally and the esophageal lumen proximal to the SCJ (Figure 4412E). The line of junction between
the two types of mucosa at the normal location has limited distensibility, and with optimum distention, it
often protrudes into the lumen as a perfectly straight, weblike elevation around the entire
circumference of the lumen. The lower esophageal B ring may be clearly demonstrated during
antegrade and retrograde endoscopy using the same breathing maneuvers as previously mentioned,
illustrating the precise anatomic relationship between the SCJ and the lower esophageal ring (see
Figures 4410 and 4411). This mucosal junction corresponds with the location of the lower
esophageal or B ring. The radiologist can readily demonstrate this in a patient with a hiatus hernia by
distending this same region with an adequate quantity of barium. The B ring indicates a hiatus hernia,
whether demonstrated by radiography or endoscopy.3(3608)
(3609)Figure 4412. Schematic representation of the sequence of anatomic changes leading

to demonstration of a hiatus hernia and a lower esophageal B ring by gradual inflation and
luminal distention during endoscopy. A, View as the sphincter region is first opened. B,
Further inflation opens the sphincter region and brings the SCJ (S-C JX) to the level of the
diaphragmatic hiatus. C, Further inflation brings the SCJ above the hiatus margin. Negative
intrathoracic pressure enhances luminal distention. D, With further inflation, the SCJ changes
from an irregular, serrated, or undulated contour to a straight line. A hiatus hernia pouch is
now clearly visible. E, With further inflation or by having the patient sniff or perform a
Mueller maneuver (inspire against a closed glottis), the combination of intraluminal positive
pressure and intrathoracic negative pressure causes the SCJ to protrude into the lumen as a
smooth, symmetric lower esophageal B ring.
Close inspection reveals that the B ring forms precisely at or within 2 to 3 mm proximal to the SCJ (see
Figure 4411). When the B ring appears and disappears with various degrees of lumen distention, it is
considered a dynamic phenomenon. If the ring manifests with minimal air distention, especially if it
significantly compromises the lumen diameter, it is considered a static structure and called a Schatzki
ring. Both dynamic and static rings are 3 mm or less in thickness, and are symmetric around the entire
circumference. Close inspection of the B ring from the distal side by retroversion may reveal varying
degrees of squamous mucosa extending circumferentially in an irregular line 2 to 3 mm caudad to the
apex of the ring. When the hiatus hernia is examined from below using optimum inflation, the lower
esophageal B ring is readily demonstrable (see Figures 448B, 4410B, and 4411). The fact that this
ring occurs at the junction of the two types of mucosa suggests that its formation depends on some
normal anatomic feature of this contact point. This theory is supported by the finding that B rings do not
occur in patients with a columnar-lined esophagus. Apparently, its absence is related to the proximal
displacement of the SCJ. It is not demonstrable in such cases because the more proximal esophagus
cannot be adequately distended or because the unique anatomic characteristic that permits its
development at the normally located mucosal junction is absent at the level of the proximally displaced
junction in the columnar-lined esophagus.
Patients with true B rings appear to have a decreased frequency of reflux esophagitis. Whether a B
ring protects against esophagitis or esophagitis prevents development of a B ring is unknown.
The term Schatzki ring often is incorrectly used in radiology and endoscopy reports. Unfortunately, any
ringlike stenosis often is called a Schatzki ring whether it conforms to the correct anatomic definition or
not. When there is asymmetry of the ring, evidence of scarring, ring thickness greater than 3 mm in
any portion, prestenotic diverticula, or severe inflammation, the terms Schatzki or B ring should not be
used. Such aberrant changes are more precisely referred to as ringlike stenosis or stricture. Dilation
therapy may be performed differently for true rings and inflammatory ringlike stenoses.
Acid-Pepsin Reflux Esophagitis

Endoscopic documentation of a hiatus hernia correlates much better with the clinical syndromes of
gastro-esophageal reflux than does any other finding, including manometric measurements of the
lower esophageal sphincter.4,12(3610) Endoscopic criteria plus the radiographic criteria of
Wolf3(3611) should be used in any study that purports to correlate reflux sequelae with hiatus hernia.
The pathophysiology of gastroesophageal reflux, although investigated extensively by radiography,
endoscopy, and manometry over the past 2 decades, is still poorly understood. It is becoming more
apparent that reflux and reflux-related esophageal disorders result from a combination of anatomic and
functional defects, of which sliding hiatus hernia and reduced lower esophageal sphincter pressure are
only two.11,13(3612) The composition and quantity of refluxed material, its pH, and its duration of
contact with the mucosa are important factors in the pathogenesis of the endoscopic and histologic
changes of erosive reflux esophagitis.
Injury to the squamous epithelium of the distal esophagus occurs when there is a sufficient frequency
and duration of exposure of this tissue to highly acid or alkaline material. The reason for this
susceptibility to injury in certain patients and resistance in others is not understood. In general, it is
accepted that the duration of acid or alkaline contact with the distal esophagus is a major factor in the
pathogenesis of the inflammatory changes recognized as reflux esophagitis. It has been suggested
that the most injurious refluxant is a combination of acid-pepsin and bile acids.
The radiographic findings in esophagitis are usually minimal or nonexistent, unless the examination is
done by an expert radiologist using double-contrast technique. Accurate initial diagnosis and
assessment of response to therapy depend on esophagoscopy and procurement of biopsies.
The inflammatory changes due to reflux, even that of a severe degree, may be so mild as to produce
neither symptoms nor sequelae in some patients.1416(3613) In others, esophagitis may proceed
inexorably to produce mucosal friability, erosion, ulceration, stricture, or a columnar-lined esophagus.
Such progression rarely occurs in a patient who is receiving appropriate treatment for reflux disease. It
is virtually impossible for any endoscopist with special interest and vast experience with the endoscopic
study of reflux sequelae to recall a single case in which a patient developed any sequelae of acid reflux
after initial diagnosis if the patient was compliant with basic treatment measures. Reflux sequelae such
as stricture, esophageal ulcer, and columnar-lined esophagus virtually always develop before the initial
endoscopic diagnosis and rarely if ever afterward. In an occasional patient, the changes are of
sufficient chronicity and severity to cause the development of a columnar-lined
esophagus.1719(3614)
Esophagitis secondary to reflux of gastric contents (i.e., hydrochloric acid or bile acids) or possibly of
other substances always involves the SCJ. The SCJ becomes less distinct as hyperemia, erythema,
and erosion progress. In some cases, it is impossible to locate the junction with certainty because of
the inflammation. The presence of tenacious exudate surrounded by a margin of hyperemia is typical
of erosive esophagitis. Evidence of erosive esophagitis proximal to the SCJ may appear as proximally
directed finger-like extensions or as isolated patches of eroded squamous mucosa (i.e., stepping-stone
erosions). In all cases, however, the SCJ will be involved at some point (Figure 4413). Esophagitis
proximal to the SCJ with no abnormality at the level of the junction should lead to a suspicion of
another cause such as infection with Candida or herpesvirus, drug-induced injury, or malignancy (see
Chapter 45: Special Varieties of Esophagitis).
(3615)Figure 4413. Endoscopic view showing erosions of reflux esophagitis. A large

erosion extending proximally from the SCJ is shown at right center. Isolated "islands" of
erosion with central white exudate are present proximally. A hiatus hernia pouch is seen in the
distance.
Reflux esophagitis should be diagnosed endoscopically only if mucosal erosion or frank ulceration are
pres-ent.10,16(3616) Hyperemia and friability are suggestive but not sufficiently reliable unless
accompanied by histologic confirmation. Erosions may be linear (i.e., parallel to the long axis of the
esophagus) with the distal end at the SCJ or oval to round and surrounded by squamous epithelium
(see Figure 4413). The margins of erosions often are intensely hyperemic. Mucosal friability is
common.
Sonnenberg et al.16(3617) devised a simple classification of erosive esophagitis: grade 1 (mild),
isolated round or linear erosions; grade 2 (severe), confluence of erosions involving the total luminal
circumference; and grade 3 (complicated), erosions as described plus deep ulcers, stenosis, or
columnar-lined esophagus. More detailed grading systems have also been described.2022(3618)
The most widely used grading system for reflux esophagitis since 1977 has been that of Savary and
Miller.20(3619) This system is based on the extent of mucosal erosion as the primary lesion. There are
four levels of severity: grade 1, single or isolated erosions, oval or linear, affecting only one longitudinal
fold; grade 2, multiple erosions affecting more than one longitudinal fold, with or without confluence
and noncircumferential; grade 3, circumferential erosion; grade 4, complications of ulcers, strictures, or
a columnar-lined esophagus, alone or associated with lesions of grades 1 to 3.
In 1991, Armstrong et al.21,22(3620) proposed a new grading system that classifies the specific
lesions associated with reflux esophagitis. It uses the mnemonic MUSE (i.e., metaplasia [columnar],
ulcers, strictures, and erosions) Each type of lesion is classified in severity by four grades: 0, absent, 1,
mild, 2, moderate, and 3, severe. This system may provide the necessary detail for a standardized
esophagitis grading system that is acceptable for clinical and research purposes.
Studies of biopsies by Ismail-Beigi et al.14(3621) showed that specific histologic changes occurring
proximal to the lower esophageal sphincter may reflect acid injury in the absence of clinically significant
disease. These histologic features, such as lengthening of the rete pegs, thickening of the basal cell
layer, and the increase in number of rete pegs as described by Kobayashi and Kasugai,15(3622) if
found more than 2 cm proximal to the lower esophageal sphincter, indicate clinically significant
reflux-related inflammation in symptomatic patients. The presence of polymorphonuclear leukocytes

and edema of the esophageal mucosa are the unequivocal signs of esophagitis and correlate with the
typical endoscopic findings.16(3623) The finding of intraepithelial eosinophils has been proposed as a
new diagnostic criterion for esophagitis.23(3624)
Because the histologic criteria of Ismail-Beigi et al.14(3625) were established using manometric
localization without direct endoscopic correlation, they must be modified during endoscopy to use
standard anatomic landmarks for accurate diagnosis. The normal SCJ lies within the region of the
lower esophageal sphincter, about 1 to 2 cm distal to the proximal margin of the sphincter as
determined by manometry (see Figure 441). A mucosal biopsy taken 1 to 2 cm above the proximal
margin of the lower sphincter is about 3 to 4 cm proximal to the SCJ. This location for obtaining
biopsies in cases without visible diagnostic surface changes should meet the criteria for proper site
selection.
When esophagitis is suspected in the absence of visible diagnostic changes in the mucosal surface,
an adequate mucosal biopsy must be obtained 3 cm or more above the SCJ. Adequate implies full
mucosal thickness, ideally to the level of the muscularis mucosa or at a minimum to the full epithelial
layer. With proper technique, adequate mucosal biopsies can be obtained with the standard flexible
forceps.12(3626) Use of the larger, so-called jumbo biopsy forceps provides the best specimens for
diagnosis of esophagitis and columnar-lined esophagus. Histologic changes diagnostic of acid-induced
esophagitis appear before symptoms and endoscopically observable findings; histologic changes also
persist after endoscopic indicators have disappeared in response to therapy.
Endoscopically demonstrable, erosive reflux esophagitis is nearly always associated with a hiatus
hernia (see Figure 4413). In some instances, exudate on the luminal circumference may be detected
around the distal esophagus. This usually indicates chronic reflux esophagitis associated with some
degree of stenosis (Figure 4414). Stenosis of moderate degree may not be obvious to the casual
observer and may go unrecognized when endoscopy is performed with relatively small-diameter
instruments (e.g., 1.0 cm), because no resistance to their passage is encountered. Dysphagia often
occurs owing to reflux in the absence of a fibrous stricture. Inflammatory stenosis due to mucosal
inflammation, vascular congestion, and edema should be recognized as such and not referred to and
treated as a fibrous stricture. Diagnosis of a stricture requires evidence of fibrotic or neoplastic
stenosis. Inflammatory stenosis typically responds to antireflux measures and acid suppression by
H2-receptor blocking drugs or protein pump inhibitors. Dilation accomplishes little in such cases and
should be reserved for true strictures caused by fibrosis.
(3627)Figure 4414. Endoscopic view of severe distal reflux esophagitis with luminal
stenosis and circumferential exudate within the stenosis.
More than 90% of patients with symptoms of reflux esophagitis respond well to antireflux measures,
ant-acids, and/or the use of H2-receptor blocking drugs or proton pump inhibitory agents. Antireflux
surgery is reserved for those who remain symptomatic despite optimum medical therapy.
Alkaline Reflux Esophagitis

Alkaline reflux esophagitis is most likely to occur in two situations that permit pure duodenal contents to
enter the esophagus: after total gastrectomy with esophagoduodenostomy or
esophagojejunostomy24,25(3628) and in patients with hypochlorhydria or achlorhydria.26(3629)
Undoubtedly, some patients with a portion of the stomach remaining after surgery have a combined
acid-alkaline reflux. Most patients with alkaline reflux esophagitis after partial distal gastrectomy have a
hiatus hernia, and this probably plays a role in the pathogenesis.
The term alkaline reflux esophagitis implies that the alkalinity of the refluxed material is the causative
factor. This is misleading, because it is more likely that bile salts, proteolytic pancreatic enzymes, a
combination of the two, the presence of hydrochloric acid, or some unidentified substances are
responsible for injury to the esophageal squamous mucosa. The usual clinical manifestations of
pyrosis, retrosternal pain, or both, as well as a bitter- or sour-tasting regurgitant, are identical to the
symptoms of acid reflux esophagitis. Dysphagia, odynophagia, iron deficiency anemia, or less often,
acute and severe hemorrhage may be presenting problems.
Barium contrast radiography is primarily helpful in defining postoperative anatomic alterations.
Strictures may be detected with proper technique or a solid bolus challenge, but as with acid-pepsin
reflux esophagitis, the erosive changes of alkaline reflux esophagitis rarely can be detected
radiographically.
Esophagoscopy usually reveals mucosal erosions, hyperemic and friable mucosa, exudate, and in
some patients, stenosis and stricture. These erosions may be single or multiple, but they always arise
at the SCJ in the intact esophagus or at the esophagoduodenal or esophagojejunal anastomosis after
total gastrectomy. The extent of mucosal injury varies, but the endoscopic findings often are
impressive. Because a columnar-lined esophagus may develop in some patients, the mucosal
alterations of this condition should be searched for whether the reflux is predominantly alkaline or acid.
The endoscopic findings and histologic changes in biopsies of the squamous mucosa are
indistinguishable from those attributed to acid reflux.24,25(3630)
Antireflux positional measures and bile acid binding with aluminum hydroxide or cholestyramine plus
an H2-receptor blocking agent have been used in therapy, but the response often is unsatisfactory.
The use of a mucosal barrier-enhancing agent such as sucrose sulfate has been suggested and often
employed, but no therapeutic benefit has been proved. Use of a proton pump inhibitor has produced
improvement in some patients. Surgical management can be undertaken when patients fail to respond
to medical therapy. The most effective procedure is diversion of the duodenal contents far enough
distally into the small intestine to prevent esophageal reflux. This is usually accomplished with an
isoperistaltic Roux-en-Y limb of jejunum at least 40 cm long.25(3631)
Sentinel Fold
The sentinel fold or polyp was first described in 1973.27(3632) Some patients with a hiatus hernia
demonstrated by radiography, endoscopy, or both have a polypoid fold in the hiatus hernia pouch just
distal to the mucosal junction, most often on the left or greater curvature aspect. This finding by
radiography is a reliable indicator of significant inflammation in the region of the SCJ just proximal to
this fold, but it must be differentiated at endoscopy from varices by observation and from neoplasm by
obtaining biopsies. The usual finding is an area of focal, severe erosion or ulceration or signs of prior
ulceration between the proximal margin of the sentinel fold and the SCJ only a few millimeters away
(Figure 4415). This fold has been variously described as a pseudotumor or inflammatory polyp, and
the lesion has been removed unnecessarily by electrosurgical snare polypectomy during endoscopy.
Forceps biopsies from the fold reveal normal columnar epithelium with underlying acute and chronic
inflammation. The focal ulceration or erosion located between the SCJ and the sentinel fold is usually
healed after 12 weeks of adequate antireflux and acid-suppression therapy. The fold disappears or is
significantly reduced in size.
(3633)Figure 4415. Typical radiographic and endoscopic features of a sentinel fold. A,

Barium contrast esophagram reveals a typical sentinel fold with polypoid enlargement at the
proximal end of this gastric mucosal fold, which is located just distal to the SCJ. B, Active
focal reflux-related shallow ulceration with a white base (9 o'clock position) is bordered below
by a sentinel polyp located at the proximal end of the enlarged fold shown in the radiograph.
A tag of squamous mucosa, another manifestation of inflammatory response, is present just

proximal to the ulceration.
In rare instances, hypertrophy of the columnar epithelium below the ulceration and the squamous
mucosa above produce columnar and squamous inflammatory polypoid lesions as shown in Figure
4415B. Biopsies must be obtained to differentiate the columnar mucosa of the sentinel fold from
adenomas that may occur as a manifestation of macroscopic dysplasia in patients with Barrett's
esophagus. These focal adenomas appear to carry the same potential for malignant degeneration as
the typical Barrett's dysplasia, because some have been reported to contain foci of invasive
adenocarcinoma.28(3634)
Distal Esophageal Stenosis and Stricture

Causes and Clinical Features
Stenosis is the appropriate term to describe evidence of narrowing of the esophagus before
endoscopic examination. Complete evaluation may reveal that this narrowing is caused by
inflammation, spasm without evidence of fibrosis, or both. Peroral dilation is of little value in treating an
inflammatory stenosis. The dysphagia due to this abnormality should respond to antireflux medical
therapy. Actual fibrosis or cicatrix formation with narrowing should be designated as a stricture. This
condition requires peroral dilation in addition to antireflux medical therapy. Dysphagia for solid foods is
always present when the lumen diameter is reduced to 13 mm (39 French) or less. However, large
swallowed boluses of solid food may elicit dysphagia with lesser degrees of narrowing. Dysphagia is
always a significant symptom and requires evaluation by radiography with use of a bolus challenge, as
necessary, to demonstrate the obstruction and by endoscopy in all cases.
For practical purposes, a benign stricture at the SCJ is related to reflux of gastric contents and is
associated with a hiatus hernia. Reflux-related strictures always occur at the SCJ, regardless of its
location. If the mucosal junction has migrated proximally, as in a patient with Barrett's esophagus, the
reflux-related benign stricture occurs at the SCJ, whatever its location within the esophagus. Strictures
related to reflux are less than 1 cm long unless there has been some additional contributing factor,
such as an indwelling nasogastric tube or secondary esophagitis due to impaction of a pill or tablet.
The contour of the esophageal wall proximal to the stricture may be smooth and tapered (Figure
4416) or irregular and fibrotic (Figure 4417). As the degree of luminal stenosis increases, there will
be less erosive esophagitis proximal to the stricture because the stricture reduces the degree of reflux.
If submucosal injury of a more chronic nature has developed, fibrous bands or pseudodiverticula may
be present proximal to the stricture (see Figure 4417). These structural alterations create a hazard for
peroral dilation unless fluoroscopic control and a guidewire dilation system are used for initial therapy.
(3635)Figure 4416. Endoscopic view of typical chronic reflux-related distal esophageal

stricture. Submucosal fibrosis is apparent. Length of stricture is less than 1 cm.
(3636)Figure 4417. Endoscopic view of a severe reflux-related stricture with residual

esophagitis and gross mural deformity and diverticulum formation due to fibrosis.
Treatment of Esophageal Stricture

An esophageal stricture has a significantly adverse effect on a patient's quality of life. Social activities
may be restricted, nutrition and hydration are impaired, meals represent agony not pleasure, salivary
gland secretions become problematic, and bronchopulmonary aspiration is common. These problems
are intensified by the fact that most patients with esophageal obstruction produce large amounts of
saliva (i.e., sialorrhea). Restoration of an adequate lumen relieves these symptoms and signs of
obstruction. The main goal of therapy is to establish and maintain a patent esophagus with the lowest
risk and cost to the patient.29(3637)
Even after the diagnosis of an esophageal stricture is firmly established and a plan for dilation outlined,
close follow-up and periodic reevaluation of the stricture are necessary. The frequency of dilation is
determined by the etiology and characteristics of the stricture and the degree of remaining
inflammation. The luminal diameter to be achieved and the interval between evaluations and dilations
must be determined individually for each patient.30(3638)
Instruments for Peroral Esophageal Dilation
Two basic types of esophageal dilators are adequate for treating more than 90% of esophageal
strictures. The rubber, mercury-filled, tapered-tip dilators (e.g., Hurst, Maloney) (Figure 4418A and B)
are preferred for most reflux-related strictures. The tapered thermoplastic dilators (e.g., Savary-Gilliard
dilators, Wilson Cook Medical Inc., Winston-Salem, NC) (Figure 4418C) and the metal olive-tip
dilators (e.g., Eder-Puestow) (Figure 4418D) are passed over a spring-tipped guidewire. These types
are ideal for initial dilation of very narrow, fibrotic strictures or strictures with associated diverticula, an
eccentric or angulated lumen position, or an esophagopulmonary fistula.
(3639)Figure 4418. Hurst (bullet tip) (A) and Maloney (tapered tip) (B) rubber,
mercury-filled dilators. Savary thermoplastic tapered dilator (C) and Eder-Puestow metal olive
dilator (D) are both used with spring-tipped guidewires, which are seen at left exiting from the
dilators.
The interchangeable metal olives of the Eder-Puestow system were originally available in
odd-numbered French sizes from 21 through 45 French, but they are being manufactured in even
sizes up to 60 French. Extraordinary care should be taken when using Eder-Puestow olives of
diameters greater than 50 French. Because of their metal construction and the mechanical advantage
provided by their introduction on a flexible metal rod, the hazards of overstretching and perforation at
the stricture are increased. Mural injuries during dilation are more often located above or below rather
than within the strictured segment, suggesting that improper technique may be the responsible factor
rather than stretching alone. Perforation of the esophageal or gastric wall using the Eder-Puestow wire
is usually related to retroflexion and penetration by the spring tip of the guidewire rather than the
stretching produced by the olives; thus, the importance of fluoroscopic control.30(3640) The design of
the guidewire for the Savary system is less likely to produce mural injury because of the differential
flexibility of its spring tip.
Savary-type dilators (see Figure 4418C) combine features of the Maloney dilator (tapered tip) and the
Eder-Puestow system (guidewire). The original model of the Savary dilator has a 1-cm-long
radiopaque band in the shaft just proximal to the tapered tip and a metal cylinder within the tip. The
dilators are constructed of a flexible thermoplastic material and have a central channel for the
guidewire. Another design similar to that of the Savary system has a shorter tapered tip and is
constructed of a flexible material impregnated with barium so that the entire dilator is radiopaque.
There are several important differences in the design of the guidewires provided with the several
over-the-wire systems for mechanical esophageal dilation. The spring tip on the leading end of the
Savary guidewire has been modified to provide differential flexibility, which is maximum at tip and least
at junction point with the wire. This modification reduces the tendency for the spring to become flexed
and redirected 180 degrees in response to only minimal tip pressure, as may happen with the standard
Eder-Puestow guidewire spring tip. Despite this modification, the safest method of dilation over a
guidewire requires proper positioning with the aid of fluoroscopy and intermittent fluoroscopic
observation to ensure that the guidewire remains in proper position throughout the procedure. The
Savary guidewire has been modified by the addition of distance markers along the wire at 20-cm
intervals, starting 40 cm from the tip to 120 cm.31(3641) In selected cases, dilation can be performed
safely by placing this wire through the endoscope without fluoroscopic control.32(3642) This technique
requires levels of skill and judgment that can only be acquired through supervised training. The
decision to use this technique must be based on careful assessment of several factors, including the
type and severity of the stricture and the anatomic features distal to the stricture.31(3643)
Balloon dilators patterned after the Grundzig arterial balloon catheter are available (e.g., Rigiflex
balloon dilators, Microvasive Endoscopy, Boston Scientific Corp., Natick, MA) (Figure 4419). It has
been claimed that hydrostatic balloon dilation of esophageal strictures is safer than other methods, but
there is no proof for this claim, nor is there satisfactory evidence that this method of dilation is more
effective than any of the others.3336(3644)
(3645)Figure 4419. Hydrostatic dilators. These are available from several manufacturers in
several sizes. When they are collapsed and well lubricated, they can be passed into the
stricture via the accessory channel of an endoscope. (Courtesy of Microvasive Endoscopy,
Division of Boston Scientific Corp., Natick, MA.)
Balloon dilators are 4 to 8 cm long and have radiopaque markers at either end of the balloon, which is
positioned at the distal end of a relatively stiff catheter that may or may not have a central channel.
Several models are available. The only way during dilation to determine if the balloon has been
expanded to maximum diameter without formation of a waist is by distending the balloon with a dilute
water-soluble contrast material under fluoroscopic observation. The newest, high-compliance models
become extremely rigid at maximum distention. This prevents formation of the waist that previously
resulted in incomplete dilation of very tight strictures. A special apparatus with pressure gauge for
distention is recommended for optimum dilation.
Balloon dilators are available with diameters from 8 to 60 French. However, in contrast to other
esophageal dilators, sets with a complete range of sequentially increasing diameters are not available
commercially. Consequently, there may be a tendency to use available balloon sizes and thereby dilate
at a more rapid pace than is ordinarily considered safe. The smaller balloons are advantageous for
initial dilation of the most severe strictures. After the lumen diameter has been increased to 16 to 20
French with small balloon dilators, the Savary dilators can be used over a guidewire to complete the
therapy in several sessions.
The spindle-shaped Tucker dilators, which have loops of silk thread at each end, are designed for
retrograde passage through a gastrostomy using a pull-through technique by means of an indwelling
heavy silk or nylon thread. These dilators are available in sizes 12 through 40 French. Several dilators
of successively larger size are connected at the thread loops and are pulled in tandem through the
gastrostomy and then retrograde through the stricture.29(3646) The required gastrostomy is a major
disadvantage, and this system is seldom used for peptic stricture. The use of Tucker dilators is
sometimes necessary to treat long, severe stricturing of the esophagus, as may occur with ingestion of
caustic substances such as lye, for which repeated dilation may be required.
Preparing the Patient and the Esophagus
Patients with esophageal obstruction severe enough to cause dysphagia for all solid foods should
undergo dilation only after a 24-hour period on clear liquid intake to ensure that the esophagus is clean
and empty.29(3647) This is important when endoscopy is planned for diagnosis or in conjunction with
dilation. A complete clinical evaluation, including review of the patient's medical history, is requisite
before dilation. The endoscopist must be aware of clinical situations that increase the risk of a
complication, such as the use of anticoagulation and chemotherapy, and should consider the
indications for antibiotic prophylaxis. Informed consent should be obtained from the patient before the
procedure and should include a discussion of inherent risks and alternative therapies.
The use of premedication depends on several factors, including the patient's psychologic state, the
cause and physical characteristics of the stricture, the customary routine for dilation, presence of
associated diseases, and the anticipation of unusual degrees of discomfort during dilation.
Unfortunately, premedication is sometimes used for the benefit of the physician or surgeon who
performs the dilation rather than for the patient.
Most patients undergoing dilation of simple reflux-related strictures or rings tolerate remarkably well the
10 sec or so of mild to moderate discomfort associated with passage of each dilator. The ability to
depart after only a brief period of observation and to resume normal functions at home and work,
including the ability to operate a motor vehicle shortly after the procedure, is appreciated by most
patients. However, patients with narrow, long, or angulated strictures that require wire-guided dilation
are best premedicated with sedative and analgesia-producing drugs in a manner similar to that for
endoscopy. Most prefer some form of oropharyngeal topical anesthesia, induced by gargle or spray, to
reduce the gag reflex and discomfort of swallowing the dilator. A 1% lidocaine (Xylocaine) gargle is
effective and safe for this purpose.
The entire dilation procedure is infinitely more acceptable and tolerable if the physician who will
perform the procedure explains the expected sensations to the patient, provides assurances that
breathing will be unimpaired, and continues to offer comforting reassurance throughout the dilation.
The equanimity, compassion, and honesty of the operator are reflections of his training and
competence. Heavy handedness and the lack of skill and concern for the patient's comfort are readily
sensed by the patient, and the end result is a less than satisfactory and sometimes dangerous
procedure.
Technique of Peroral Dilation
Little has been written regarding specific recommendations for techniques of dilating benign
esophageal strictures. Consequently, there is no consensus on the principles of stricture management.
Every esophageal stricture seems to have a "personality" of its own, and variations in technique often
are necessary for successful therapy. However, several general principles of peroral dilation technique
merit consideration.30(3648)
The first principle of stricture management is to avoid a hurried approach to the restoration of lumen
patency. Many reflux-related esophageal strictures require months to years to develop to the point of
causing dysphagia, and there is little need to attempt to return the stenotic lumen to a normal diameter
in a single sitting or within a few days. Second and subsequent dilations may be scheduled at intervals
of 1 day to several weeks, depending on the etiology and duration of the stricture and on the response
of the patient and the stricture to the initial dilation. The ultimate success of therapy should not be a
cause for concern, because essentially all benign esophageal strictures are dilatable over some period
on an ambulatory or inpatient basis.
A second principle is to pass no more than three dilators that meet with moderate or greater resistance
at any one session. The proper size of a dilator often is determined by passing smaller ones first to
assess the degree of stricturing and the pathway to the stomach. The point at which resistance is
encountered and, in some cases, the characteristics of the stricture can he sensed by a practiced
operator; this is impossible when dilators are habitually passed with excessive force and abruptness.
Dilators that meet minimal or no resistance and obviously produce no stretching of the stenotic
segment are not counted in the "rule of three." The principle is to limit the dilation session to passage
of three dilators that encounter moderate or greater resistance. Only two or even one dilator may be
the appropriate number when the stricture is very narrow, pain is severe, and bleeding occurs or when
the degree of narrowing is extreme and symptoms have been present for years, as for strictures due to
corrosive injury.
A third important principle is the use of fluoroscopy to monitor dilator position and path, especially
during initial treatment sessions. The lumen axis for most typical benign esophageal strictures is
directly in line with the axis of the proximal esophagus and proximal stomach. The rubber,
mercury-filled bougies (e.g., Maloney) ordinarily used for simple strictures can be expected in most
patients to pass safely into the strictured lumen and through it without appreciable lateral pressure on
the wall except within the stricture. However, there are enough variations in these usual anatomic
relationships that the physician is well advised to keep the dilator tip under visual control at all times.
Fluoroscopy may be omitted with simple strictures that have been dilated under fluoroscopy initially and
are known to have no associated anatomic factors that increase risk.
The nonavailability of fluoroscopy in many endoscopy suites has prompted several studies to evaluate
the necessity of fluoroscopic control during esophageal dilation.31,34,37(3649) These concern the
Savary wire-guided method using a marked guidewire and the use of mercury-filled, rubber Maloney
dilators. Fleischer et al.31(3650) found dilation to be safe in 100 patients when a marked guidewire had
been placed endoscopically. However, when multiple dilations are required, the patient must accept the
risk and cost of repeated endoscopy procedures associated with this approach.
The Savary guidewire is easily and safely passed under fluoroscopic control. After the initial diagnostic
endoscopy, the need for guidewire passage through the endoscope is uncommon when fluoroscopy is
readily available, unless the lumen takes an angular path and "shelving," diverticula, or other anatomic
distortions are present, as seen after some surgical resections. Precise fluoroscopic localization of the
guidewire tip during dilation is believed to increase the safety of this technique. When the guidewire
cannot be passed using fluoroscopic control alone, it can be inserted through the accessory channel of
a forward-viewing panendoscope. Some physicians who perform endoscopy prefer a combination of
fluoroscopy and endoscopy for positioning the guidewire. Even though the guidewire is being passed
through the endoscope into the proximal end of the stricture under direct vision, its spring tip may
become markedly angulated within or distal to the stricture. This may occur completely beyond the
endoscopic view, and fluoroscopy should be used when attempting to pass the guidewire through the
endoscope with the instrument tip situated proximal to a stricture.
It is easier to prevent the development of a far-advanced stricture than to treat this condition. The value
of early treatment when the lumen is not markedly narrowed nor fibrosis very advanced cannot be
overemphasized. Neglected esophageal narrowing is unfortunately still a dilemma and will remain so
until the need for continuous efforts at prophylaxis is better appreciated. All patients who have had
dysphagia as a result of an esophageal stricture, even though it is adequately dilated initially, should
undergo clinical reevaluation at least annually. Subtle modifications of diet consistency, chewing habit,
and fluid ingestion can mask the slow progression of a chronic stricture.
Columnar-Lined or Barrett's Esophagus

Barrett's or columnar-lined esophagus is more common than previously appreciated.18,19,21,38
Because metaplastic epithelium in a columnar-lined esophagus is considered premalignant, its early
recognition is imperative. It is possible to recognize minimal degrees of a columnar-lined esophagus
based on an understanding of the normal luminal topographic relationships. Short segments of
columnar-lined mucosa appear to have a premalignant potential that is similar to that for longer
segments that extend to the proximal esophagus.
Development of Barrett's esophagus requires a hypotensive lower esophageal sphincter and a hiatus
hernia. Recognition of this relationship is important for accurate diagnosis. Mucosal biopsies taken
from a hernia pouch have resulted in overdiagnosis and confusion in the analysis of data from
surveillance studies. These problems result from the failure of the endoscopist to understand normal
endoscopic anatomy, especially the recognition of small sliding hiatus hernias. When the SCJ is
displaced proximally above the esophageal hiatus, the gastric folds that extend through the hiatus are
recognizable. These normal folds stop at a point several millimeters to 1 cm distal to the normal
position of the SCJ (see Figures 443 and 447). Because a measurement of 1 cm or less above the
proximal end of the gastric folds indicates the expected site of a normally positioned SCJ in a patient
with a hiatus hernia, it is easy to determine whether the junction is displaced proximally (Fig 4420).
(3651)Figure 4420. A, View of the gastric folds in a hiatus hernia associated with a
columnar-lined esophagus. Note that the faint linear vascular pattern ends at the proximal
margin of these folds, indicating the level where the SCJ should normally be located. The SCJ
in this patient was displaced 4 cm proximally into the esophagus. B, Close-up view showing
the abrupt change in mural depth of the subepithelial vessels as they pass the level that
corresponds to that of the SCJ as well as that of the GEJ.
The endoscopic features that aid in diagnosis of columnar-lined esophagus include the proximal
margin of the gastric folds in the hernia pouch, the patulous lower esophageal sphincter region, the
dislocated SCJ (>1 cm cephalad of the proximal margin of the gastric folds), and absence of a lower
esophageal mucosal B or Schatzki ring. By reference to these endoscopic landmarks, minimal degrees
of metaplastic epithelial change in the distal esophagus can be suspected and confirmed by obtaining
biopsy specimens. It is unnecessary to await development of the classic later stages with SCJ
dislocation into midesophagus to diagnose this condition. A short columnar-lined esophagus often
exists in patients with minimal symptoms, but it is not recognized in such cases if this diagnosis is
considered only in relation to severe reflux symptoms, proximal esophageal stricture, or a long (i.e., 2
or 3 cm or more) segment of columnar-lined esophagus.
The SCJ is normally located at or within 1 cm of the proximal end of the gastric mucosal folds in a
hiatus hernia (Figure 4421; see also Figure 447). This level corresponds closely to the true
gastroesophageal muscular junction. Another landmark, albeit difficult to detect, that can be used to
assess mucosal junction dislocation is the transverse mucosal fold, the luminal marker for the level of
the angle of His, which also corresponds to the gastric sling muscle and the esophagogastric muscular
junction. This region is always markedly patulous because of the hypotensive lower esophageal
sphincter and hiatus hernia and is easy to examine in a patient with a columnar-lined esophagus. The
true anatomic location of the gastric folds in the hiatus hernia cannot be appreciated if the lumen is
overinflated. The hiatus hernia pouch associated with a columnar-lined esophagus must be deflated
adequately to allow these folds to return to their most cephalad location. Only then is the fold level
properly of assistance in measuring the SCJ displacement. Careful observation reveals the
columnar-lined segment to be relatively dilated, although peristalsis or tertiary contractions are present.
The displaced SCJ may be found at any level in the esophagus. By retroverting the tip of the
endoscope below the diaphragmatic hiatus and withdrawing into the hiatus hernia, the columnar-lined
segment and the displaced SCJ may be examined from a different perspective (see Figure 4421).
(3652)Figure 4421. Drawing of a retroverted endoscopic view of a hiatus hernia and

displaced SCJ with an interposed segment of metaplastic columnar mucosa lining the distal
esophagus. The endoscope was omitted from this drawing to permit a full view of the surface
anatomy.
Identification of the SCJ in normal individuals and patients with the typical columnar-lined esophagus
usually is not difficult because of differences in mucosal color and texture. The mucosal color in the
columnar-lined segment is orange-red to red; the color usually is more reddish than normal gastric
mucosa and is distinctly different from the pinkish gray of normal esophageal squamous mucosa (see
Figure 443A). It may be impossible to precisely locate the SCJ in the presence of severe esophagitis
or a stricture. Close-up observation of the texture of columnar mucosa reveals a pitted or villoid
pattern. Several biopsy specimens from areas of suspected columnar metaplasia can provide
histologic confirmation of the diagnosis.1719(3653) Metaplastic columnar (intestinalized) epithelium
with alcian blue-positive staining goblet cells is considered diagnostic of a columnar-lined esophagus.
Only this cell type is known to become dysplastic. The exact location of biopsies and the relation of
these sites to surrounding landmarks should be recorded when samples are obtained.
Normal squamous esophageal mucosa is characterized by a pink-gray, smooth surface and many
underlying small vessels oriented parallel to the long axis of the lumen (see Figure 443C). These
small vessels disappear in the area where the normally located SCJ approximates the proximal margin
of the gastric mucosal folds. The level at which these vessels disappear, typically at the most proximal
margin of the gastric folds, marks the level of the true gastroesophageal muscular junction (see Figure
4420B). Several milliliters of 2.5% Lugol's solution can be instilled to stain squamous epithelium a
brownish-black when the location of the junction remains uncertain (Figure 4422).1(3654) Islands of
residual squamous epithelium in the columnar-lined segment also are readily identified by this
technique (Figure 4422B and C; see also Chapter 12: Chromoscopy).
(3655)Figure 4422. A, Irregular SCJ in the midesophagus. B, Lugol's iodine staining of

squamous epithelium in the same patient. Note the small islands of stained squamous
epithelium below the SCJ. C, Close-up view of stained islands of squamous epithelium shown
in the 5 o'clock and 9 o'clock positions.
The SCJ in Barrett's esophagus can vary widely in location and appearance. It often has an undulating
appearance with orange-red, finger-like cephalad extensions of metaplastic epithelium between distally
oriented projections of squamous mucosa, and the whole junction appears asymmetric and interrupted
(Figure 4423). There is a common tendency to attribute this appearance to esophagitis with erosion
alone, although in some cases the orange-red, proximally directed streaks result from metaplastic
columnar epithelium at the site of previous linear erosions. In some patients, islands of metaplastic
columnar epithelium surrounded by squamous epithelium may be found proximal to the level of the
main SCJ.
(3656)Figure 4423. Extensions of peninsula-like areas of metaplastic columnar epithelium in

Barrett's esophagus that may be interpreted as "esophagitis" unless biopsies are obtained. Such
areas of metaplastic epithelium are never covered by exudate as is often seen over erosions.
The endoscopic landmarks and topography previously described should be used to avoid missing the
diagnosis of columnar-lined esophagus. Overdiagnosis occurs when the mucosa in a hiatus hernia
pouch is interpreted as being in the tubular esophagus. An adequate number of biopsy specimens
(preferably with large forceps) and precise histologic interpretation usually prevents misdiagnosis.
Underdiagnosis occurs with failure to recognize short segments of columnar mucosa lining the distal
esophagus or failure to recognize that most of the esophagus is lined by columnar epithelium when the
dislocated SCJ is in the cervical esophagus. This error is more likely to occur when proximal
esophagitis or stricture is absent and when the endoscopist is careless and fails to examine every
millimeter of the esophageal mucosa with care.
Endoscopists who are unaware of the frequency and appearance of heterotopic gastric mucosa, the
so-called inlet patches, in the cervical esophagus are prone to make an incorrect diagnosis of a
columnar-lined or Barrett's esophagus. Such misdiagnosis is unfortunate and leads to unnecessary
concern and surveillance for a condition that is not prone to dysplasia or carcinoma. The inlet patch is
discussed in Chapter 39: Esophageal Motility and Miscellaneous Disorders.
Esophageal Ulcer in Columnar-Lined Esophagus

Deep, chronic ulcers of the esophagus in most instances indicate a columnar-lined esophagus. Such
ulcers always occur at the SCJ or distal to it in the columnar-lined segment and primarily involve the
columnar epithelium. Instances of acid reflux-related, benign, solitary ulceration in an esophageal
segment totally lined by squamous epithelium have not been confirmed. Ulcers of this type in a
columnar-lined esophagus may have small islands of squamous mucosa located about their margins,
but they are never surrounded entirely by squamous epithelium. Esophageal ulcers have the typical
endoscopic features of a gastric ulcer, and are prone to bleed or penetrate; these events often occur
with minimal prior symptoms (Figure 4424A and B). Because adenocarcinoma occurs with increased
frequency in columnar-lined esophagus, all ulcers and strictures should be evaluated by obtaining
multiple biopsies or cytology specimens (Figure 4424C).21,3841(3657) (Adenocarcinoma in relation
to Barrett's esophagus is discussed in Chapter 40: Benign and Malignant Tumors of the Esophagus.)
(3658)Figure 4424. A, Solitary, benign, bleeding ulcer in Barrett's esophagus. B, Close-up of

a solitary benign ulcer. C, Ulcer in Barrett's esophagus. Biopsies revealed adenocarcinoma;
this lesion was staged by endoscopic ultrasonography as T3N1.
Esophageal Stricture in Columnar-Lined Esophagus

Stenosis due to acid reflux esophagitis is reported to develop in about 40% of patients with
columnar-lined esophagus, and it typically occurs at the SCJ.17,18(3659) Stenosis at the SCJ is
usually caused by mucosal inflammation without submucosal injury and responds readily to medical
therapy. Because of its mucosal origin, a reflux-related stenosis with columnar-lined esophagus may
be observed to relocate proximally over the years as the SCJ migrates in a patient with inadequately
treated reflux. When submucosal injury occurs and heals with fibrosis, a stricture is formed that
requires dilation plus medical therapy for acid reflux.
A stricture within the columnar-lined segment of esophagus is uncommon but may be caused by the
fibrosis that results from an active ulcer or ulcer healing or by an invading carcinoma. It can occur at
any level in the columnar-lined segment. The lumen of the segment typically has a normal or slightly
larger diameter. Careful endoscopic observation permits detection of even minimal degrees of
stenosis, mural rigidity, or asymmetry in lumen contour. Because reflux esophagitis-associated
stenosis or stricture only occurs at the SCJ, any stenosis distal to the SCJ in the columnar-lined
segment only can be explained as secondary to submucosal fibrotic changes from a healed or active,
benign esophageal ulcer or by adenocarcinoma.
Because patients with columnar-lined esophagus have an increased risk for the development of
adenocarcinoma that is 30 to 40 times that of the general population, endoscopic screening
examinations, including procurement of multiple biopsies, should be performed at 1- to 2-year
intervals.18,19,21,38,40,42(3660) Biopsies should be taken in four quadrants at 2-cm intervals from
the proximal end of the gastric folds in the hiatus hernia pouch to the SCJ. In some cases, the typical
intestinal metaplasia of Barrett's esophagus involves less than 1 cm2 of surface area and can be focal
or circumferential. In addition to the landmarks previously discussed, close inspection can reveal a
distinctly villoid surface topography characteristic of dysplastic epithelium. Biopsies should be taken
from any areas that have these surface features. Chromoscopy using toluidine blue or methylene blue
can also assist in the identification of metaplastic, dysplastic, and neoplastic changes.43(3661)
The increasing incidence of adenocarcinoma of the esophagus in association with a columnar-lined
esophagus is alarming.41(3662) The survival rate for patients after the development of symptoms of
adenocarcinoma is poor, and early detection of dysplasia by screening biopsies offers the only hope for
decreasing the mortality from this disease.3842(3663) Various aspects of this disorder are also
discussed in Chapter 40: Benign and Malignant Tumors of the Esophagus.
No proven method exists to reverse the process of metaplasia in the columnar-lined esophagus and to
restore squamous epithelium. Such efforts have included antireflux surgery, acid suppression by
proton pump inhibition, and the use of localized argon laser thermal ablation combined with 40 mg of
omeprazole daily.44(3664) The technique of photodynamic laser therapy is being evaluated for
treatment of high-grade dysplasia and early adenocarcinoma associated with the columnar-lined
esophagus (see Chapter 40: Benign and Malignant Tumors of the Esophagus).45(3665)
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Nothmann BT, Wright JR, Schuster MM. In vivo vital staining as an aid to identification of
esophagogastric mucosal junction in man. Am J Dig Dis 1972;17:91924.
Gelfand DW, Ott DJ. Anatomy and technique in evaluating the esophagus. Semin Roentgenol
1981;16:16882.
Wolf BS. Sliding hiatal hernia: The need for redefinition. Am J Roentgenol Radium Ther Nucl
Med 1973;117:23147.
Jonsell G. The incidence of sliding hiatal hernias in patients with gastroesophageal reflux
requiring operation. Acta Chir Scand 1983;149:637.
Dagradi AE, Killeen RN, Schindler R. Esophageal hiatus sliding hernia: An endoscopic study.
Trujillo NP, Boyce HW Jr. Gastroscopy: Aid to the detection of small sliding-type hiatal
hernias. South Med J 1968;61:14.
Trujillo NP, Slaughter RL, Boyce HW Jr. Endoscopic diagnosis of sliding-type diaphragmatic
hiatal hernias. Am J Dig Dis 1968;13:85567.
Ortega JA. New criterion in the esophagoscopic diagnosis of sliding type hiatal hernia. Am J
Rsch W. Gastro-oesophageal reflux and hiatus hernia endoscopy. Postgrad Med J
1974;50:199201.
Dagradi AE. Endoscopic examination of the gastro-esophageal area. Gastrointest Endosc
1969;15:1757.
Sloan S, Rademaker AE, Kahrilas PJ. Determinants of gastro-esophageal junction
incompetence: Hiatal hernia, lower esoph-ageal sphincter, or both?. Ann Intern Med
1992;117:97782.
Johnson LF, DeMeester TR, Haggitt RC. Endoscopic signs for gastroesophageal reflux. J Clin
Edwards DAW, Thompson H, Shaw DG, et al. Symposium on gastroesophageal reflux and its
complications. Gut 1973;14:23353.
Ismail-Beigi F, Horton PF, Pope EC 2nd. Histological consequences of gastroesophageal
reflux in men. Gastroenterology 1970;58:16374.
Kobayashi S, Kasugai T. Endoscopic and biopsy criteria for the diagnosis of esophagitis with
a fiberoptic esophagoscope. Am J Dig Dis 1974;19:34552.
Sonnenberg A, Lepsien G, Muller-Lissner SA, et al. When is esophagitis healed? Esophageal
endoscopy, histology and function before and after cimetidine treatment. Dig Dis Sci
1982;27:297302.
Paull A, Trier JS, Dalton MD, et al. The histologic spectrum of Barrett's esophagus. N Engl J
Med 1976;295:47680.
Sjgren RW Jr, Johnson LF. Barrett's esophagus: A review. Am J Med 1983;74:31321.
Herlihy KJ, Orlando RC, Bryson JC, et al. Barrett's esophagus: Clinical, endoscopic,
histologic, manometric, and electrical potential difference characteristics. Gastroenterology
1984;86:43643.
Savary M, Miller G. The Esophagus: Handbook and Atlas of Endoscopy. Solothurn, Schweiz:
Gassman. 1978.
Armstrong D, Blum AL, Savary M. Reflux disease and Barrett's esophagus. Endoscopy
1992;24:917.
Armstrong D, Monnier PH, Nicolet M, et al. Endoscopic assessment of esophagitis. Gullet
1991;1:637.
Winter HS, Madara JL, Stafford RF, et al. Intraepithelial eosin-ophils: A new diagnostic
criterion for reflux esophagitis. Gastroenterology 1982;83:81823.

24. Wickbom G, Bushkin FL, Woodward ER. Alkaline reflux esophagitis. Surg Gynecol Obstet
1974;139:26771.
25. Pellegrini CA, DeMeester TR, Wernly JA, et al. Alkaline gastro-esophageal reflux. Am J Surg
1978;135:17784.
26.
Palmer ED. Subacute erosive ("peptic") esophagitis associated with achlorhydria. N Engl J
Med 1960;262:9279.
27.
Boyce HW Jr. Endoscopic Diagnosis of Hiatal Hernia and Related Disorders. Dallas: ASGE
Postgraduate Course. 1973.
28. Lee RG. Adenomas arising in Barrett's esophagus. Am J Clin Pathol 1986;85:62932.
29.
Boyce HW Jr, Palmer ED. Techniques of Clinical Gastroenterology. Springfield: Charles C
Thomas, 1975;22182.
31. Fleischer DE, Benjamin SB, Cattau EL, et al. A marked guide wire facilitates esophageal
dilation. Am J Gastroenterol 1989;84:35961.
32. Kadakia SC, Cohan CF, Starnes EC. Esophageal dilation with polyvinyl bougies using a
guidewire with markings without the aid of fluoroscopy. Gastrointest Endosc 1991;37:1837.
33. Shemesh E, Czerniak A. Comparison between Savary-Gilliard and balloon dilatation of benign
esophageal strictures. World J Surg 1990;14:51821 [discussion 5212].
34.
Kozarek RA, Gelfand MD, Patterson DJ, Ball TJ. A prospective trial of American Savary (A-S)
dilators versus hydrostatic balloons (HB) in complex esophageal strictures [Abstract].
35. Cox JGC, Winter RK, Maslin SC, et al. Balloon or bougie for dilatation of benign esophageal
stricture?. Dig Dis Sci 1994;39:77681.
36. Saeed ZA, Winchester CB, Pieretta S, et al. Prospective randomized comparison of polyvinyl
bougies and through-the-scope balloons for dilation of peptic strictures of the esophagus.
37. McClave SA, Wright RA, Brady PG. Prospective randomized study of Maloney esophageal
dilationBlinded versus fluoroscopic guidance. Gastrointest Endosc 1990;36:2725.
38. Naef AP, Savary M, Ozzello L. Columnar-lined lower esophagus: An acquired lesion with
malignant predisposition. Report on 140 cases of Barrett's esophagus with 12
adenocarcinomas. J Thorac Cardiovasc Surg 1975;70:82635.
39. Haggitt RC, Tryzelaar J, Ellis FH, Colcher H. Adenocarcinoma complicating columnar
epithelium-lined (Barrett's) esophagus. Am J Clin Pathol 1978;70:15.
40. Hameeteman W, Tytgat GNJ, Houthaff HJ, van den Tweel JG. Barrett's esophagus:
Development of dysplasia and adenocarcinoma. Gastroenterology 1989;96:124956.
41. Pera M, Cameron AJ, Trastek VF, et al. Increasing evidence of adenocarcinoma of the
esophagus and esophagogastric junction. Gastroenterology 1993;104:5103.
or early adeno-carcinoma in Barrett's esophagus without grossly recognizable neoplastic
43. Canto MIF, Setrakian S, Petras RE, et al. Methylene blue selectively stains intestinal
metaplasia in Barrett's esophagus. Gastrointest Endosc 1996;44:17.
44. Berenson MM, Johnson JD, Markowitz NR, et al. Restoration of squamous mucosa after
ablation of Barrett's esophageal epithelium. Gastroenterology 1993;104:168691.
45. Overholt B, Panjepour M, Teftellar E, Rose M. Photodynamic therapy for treatment of early
Chapter 45 Special Varieties of Esophagitis

(3666)
H. WORTH BOYCE, JR., M.S., M.D.

Esophagitis of varying degrees and clinical significance is caused by infections, chemical agents
including drugs, ionizing radiation, Behet's disease, and epidermolysis bullosa, a rare hereditary
disease. Idiopathic disorders such as eosinophilic gastroenteritis and Crohn's disease rarely have been
associated with esophagitis. The more common of these relatively uncommon disorders are
discussed, with emphasis on the role of endoscopy in their management.
Atypical cases of reflux esophagitis may cause confusion in diagnosis, especially when infections,
chemical-induced esophagitis, or malignancy are superimposed on changes caused by reflux.
Reflux-related esophagitis is by far the most common inflammation of the esophagus. Its clinical,
radiographic, and especially endoscopic-histologic features confirm the diagnosis without difficulty. A
complete understanding and awareness of the spectrum of characteristic endoscopic changes caused
by reflux as well as familiarity with important clinical clues are prerequisites for confirming the diagnosis
of nonpeptic esophagitis. The endoscopic criteria for reflux-related acid-peptic and alkaline esophagitis
are reviewed in Chapter 44: Hiatus Hernia and Peptic Diseases of the Esophagus.
Esophageal Infections
Candida (Monilial) Esophagitis
The esophagus is the visceral organ most frequently infected by Candida albicans, by far the most
common infection of the esophagus. Kodsi et al.1(3667) reported a prospective study of the incidence
of monilial esophagitis among 370 consecutive patients who underwent endoscopy for upper
gastrointestinal symptoms. There were 27 cases (7%) of proven Candida esophagitis in this large
series; 14 of these patients had symptoms of esophagitis. Other authors also have shown that this
disease may appear in the absence of esophageal symptoms or any of the various predisposing
conditions considered in the past as necessary for the establishment of Candida infection.2(3668)
The classic predisposing conditions for Candida infection include diabetes mellitus, malignancy and
other chronic debilitating diseases, esophageal obstruction, long-term corticosteroid or antibiotic
therapy, and especially compromise of the immune system.1,3(3669) In the absence of predisposing
illness, Candida esophagitis is mainly a disease of the elderly. Kodsi et al.'s review suggests that any
of three physiologic processes may play a role in the pathogenesis: impaired immunity, impaired
esophageal motility, and impaired carbohydrate metabolism associated with aging.1(3670) Less than
half the patients with esophagitis will have thrush, the oropharyngeal form of this infection.4(3671)
Dysphagia and odynophagia are the most common symptoms. Some patients complain only of an
awareness of food passing down the esophagus; others have retrosternal pain with radiation to the
scapular region or along the left of the lower thoracic spine. Malnutrition may result when these
symptoms are severe or prolonged. The clinical syndrome plus findings on barium esophagogram in
the typical patient provides strong evidence for diagnosis. Although the barium esophagogram is
abnormal in many cases, the radiographic findings are variable and, by themselves, nondiagnostic.
One group of investigators found esophageal symptoms plus oral candidiasis in patients with acquired
immunodeficiency syndrome (AIDS) to be sufficient for the diagnosis of Candida esophagitis and
suggested that esophagoscopy be performed in only selected cases.5(3672) Another study found that
oral candidiasis and esophageal symptoms do not reliably predict the presence of Candida esophagitis
in patients with AIDS and that esophagoscopy is necessary for accurate diagnosis.6(3673)
It is surprising that the first report of the radiographic findings in Candida esophagitis was not published
until 1956.7(3674) Radiographic abnormalities include motility disturbances manifested by diminished
peristalsis, irritability, or spasm. The early mucosal changes of irregularity or granularity may progress
to a shaggy appearance with edema and ulcerations. Focal accumulations of fungal debris may
simulate varices or neoplasm. With progression of mural infection into the submucosa, edema
becomes more apparent, ulceration may progress, and multiple sinus-like recesses or intramural
pseudodiverticula may be demonstrated. This appearance has been likened to that of the colon wall in
some severe cases of ulcerative colitis. Intramural infection may progress to stenosis owing to edema
and spasm or to frank fibrotic stricture.
The endoscopic appearance of Candida esophagitis ranges from an erythematous friable mucosa to
complete covering of the mucosal surface by a heavy, shaggy cream-to-white pseudomembrane
throughout the esophagus (Figures 451 and 452). The most common finding is white-to-cream
plaques scattered throughout the esophagus, usually with greater density in the lower two thirds. The
mucosa just proximal to the squamocolumnar junction often is less involved. This plus the
characteristic exudative plaques usually makes it easy to differentiate Candida esophagitis from reflux
and other forms of esophagitis. However, when Candida esophagitis is combined with other
esophageal diseases, the endoscopic appearance may not be typical.
(3675)Figure 451. Endoscopic view of monilial esophagitis showing typical focal

yellow-white plaques scattered over the mucosa. The degree of hyperemia around these
plaques is highly variable.
(3676)Figure 452. Endoscopic view of severe monilial esophagitis with thick exudate
producing partial occlusion of the lumen.
A grading scale for Candida esophagitis used by Kodsi et al.1(3677) describes the stages of this
infection: Grade I, a few raised white plaques up to 2 mm in diameter with hyperemia but no edema or
ulceration; Grade II, multiple raised white plaques greater than 2 mm in size with hyperemia but no
edema or ulceration (see Figure 451); Grade III, confluent linear and nodular elevated plaques with
hyperemia and frank ulceration; Grade IV, findings of Grade III plus increased friability of the mucus
membranes and occasional narrowing of the lumen (see Figure 452).
Atypical surface changes often occur, so that direct smears of exudate from plaques or the mucosal
surface as well as mucosal biopsies should be obtained for diagnosis. When ulcerations are present,
mycelia can usually be demonstrated in biopsies of the ulcer base. Because routine H & E stains
demonstrate mycelia poorly, silver methenamine staining should be requested.
Esophagoscopy with collection of material using a brush-type endoscopic accessory for direct
potassium hydroxide smear is the procedure of choice for diagnosis.1,4(3678) This rapid method for
confirmation involves microscopic demonstration of mycelial forms in the exudate, which is mixed with
a drop of 10% potassium hydroxide solution under a coverslip. The pseudomycelia of Candida species
are also readily identified on routinely stained cytology smears. Cultures do not differentiate between a
commensal or an infectious status for Candida, and hence are of value only in species determination
or for drug sensitivity testing. A serum agglutinin titer greater than 1:160 is considered confirmatory in
patients with other evidence of Candida esophagitis.1(3679) This test is rarely positive at this titer in
the absence of other evidence of candidiasis.
Symptoms usually disappear with therapy, but clinical improvement does not necessarily indicate
disappearance of the mucosal lesions. Endoscopy is necessary to confirm complete healing, although
not essential in patients at average risk who have no residual symptoms after treatment.
Nystatin, 200,000 units, given as a combined gargle and swallow every 1 to 2 hr while the patient is
awake, has been the standard therapy with a good cure rate and low risk of side effects.8(3680)
More recently, newer antifungal agents have been used successfully, especially in patients with severe
nystatin-resistant infections or immunodeficiency states, especially in AIDS. Until recently, the
preparation favored for this therapy was an imidazole compound, ketoconazole, which was said to be
less toxic than amphotericin B, miconazole, and flucytosine.9(3681) Ketoconazole is well absorbed
when taken orally and is given as a single daily dose of 200 mg for 2 weeks. Absorption is reduced
when there is gastric acid hyposecretion or elevation of the gastric pH by medications. Ketoconazole
should not be used for initial therapy of monilial esophagitis and is best reserved for cases associated
with severe systemic disease that have not responded to nystatin.
A new oral triazole antifungal agent, fluconazole, has been shown to have better absorption, no
dependency on gastric pH as does ketoconazole, and a longer half-life (16 hr compared with 8 for
ketoconazole). Reduced absorption of ketoconazole in patients with gastric hyposecretion, such as
those with AIDS, probably accounts for the fact that the non-pH-dependent fluconazole has greater in
vivo activity against C. albicans. In another report, 90% of patients with AIDS-related Candida
esophagitis had resolution of the esophagitis when treated with fluconazole compared with 40%
resolution in similar patients in the other treatment group who received ketoconazole.10(3682) Over
40% of the ketoconazole-treated patients had resolution of symptoms without healing by endoscopy,
whereas this occurred in only 4% of patients treated with fluconazole. At this time, fluconazole, 100 to
200 mg daily for 2 weeks after resolution of symptoms for a minimum of 3 weeks and a maximum of 8
weeks, appears to offer the best healing rates for severe cases of Candida esophagitis, especially in
immunosuppressed individuals.10(3683)
Herpetic Esophagitis
Until recent years, most reports of proven herpesvirus (simplex or zoster) and cytomegalovirus (CMV)
infections of the esophagus have been postmortem studies.1115(3684) In that of Nash and
Ross13(3685) on 3000 consecutive autopsies, herpes esophagitis was found in 0.4%. When only
patients with malignancy were considered, the overall incidence increased to almost 4%.13(3686)
Flexible endoscopy is the most reliable method for the diagnosis of herpetic esophagitis, and the
number of reports of antemortem diagnosis is increasing.1522(3687) When it occurs in an otherwise
healthy person, the clinical course is so short that response to symptomatic therapy may occur before
referral for endoscopic evaluation. Relatively few cases were documented in persons without
predis-posing conditions until the advent of flexible endoscopy.16,18,23(3688) The most common
occurrence is in immunosuppressed or preterminal patients with or without malignancy, especially
those with AIDS.11,15,24(3689)
The classic presentation of herpetic esophagitis was that of a patient, gravely ill as a result of a
malignancy, with severe dysphagia, odynophagia, and retrosternal pain. Usually Candida esophagitis is
first suspected, but this diagnosis should be questioned when there is no response to antifungal
therapy. Such patients should have further endoscopy and biopsy to evaluate for a viral etiology before
proceeding to drug therapy for Candida infection.
Physical examination may reveal either oral mucosal or vesicular skin lesions, or both, of herpes
simplex or herpes zoster. Herpes simplex is much more commonly the infectious agent. When oral or
cutaneous lesions are present, confirmation of their viral etiology should be attempted. Some patients
may present with a mild, febrile, viral illness syndrome. Rosen and Hajdu12(3690) reported that the
esophagus was the only organ affected in 72% of patients.
Radiographic examination by barium esophagogram is not adequate for specific diagnosis, although
distinct abnormalities may be seen. Findings include multiple, diffuse, small, nodular filling defects,
mucosal edema, numerous small ulcers, and irregular, vertically oriented plaquelike
projections.25,26(3691) These changes stop abruptly at the gastroesophageal junction. Some
radiologists regard air contrast esophagography as the preferred examination.25(3692) The
radiographic findings may be identical to those of monilial esophagitis. Further examination by
esophagoscopy with cytologic study and biopsy is usually necessary for diagnosis.
Esophagoscopy may reveal diffuse, segmental, or focal involvement.16,18,19,27(3693) The classic
early lesion is a small (less than 5 mm) papule or vesicle that ultimately ulcerates. The mucosa
surrounding this lesion is erythematous but that between lesions may be normal in the early stages.
These small vesicles will be seen only early in herpetic esophagitis, at a time when both biopsy and
culture may be negative for the herpes simplex virus.28(3694) As the vesicles evolve into erosions and
later distinct ulcers, the likelihood of biopsy or culture confirmation increases significantly.29(3695) The
ulcers are round with margins that are distinct and elevated above the plane of the mucosa (Figure
453). Some have referred to these as volcano ulcers because of the overhanging inflamed
margin.20(3696) The base of the ulcer is covered with white-to-yellow exudate. As ulcers enlarge, they
may coalesce, with the resulting larger ulceration tending to be linear and oriented in the long axis of
the esophagus. With severe involvement, part or all of the esophagus may be covered with a thick
exudate with no discrete ulcerations to provide a clue to the etiology.30(3697) There is a tendency for
ulcerations in the distal esophagus to be larger or to coalesce more readily, even becoming
circumferential, suggesting a possible potentiation of the virus-induced ulceration by material refluxed
from the stomach (Figure 454).
(3698)Figure 453. Multiple small ulcers typical for herpetic esophagitis are shown in this
endoscopic view from the midesophagus. Two classic herpetic ulcers are present to the left of
center, and multiple smaller ulcers in a linear orientation are shown on the opposite side.
Distal to this level, small lesions have coalesced to form larger ulcers.
(3699)Figure 454. Herpetic esophagitis with complete circumferential mucosal ulceration in

the distal esophagus. The margin between intact and ulcerated areas is shown in the
foreground between the 5 and the 9 o'clock positions.
There is one report of severe bleeding from the distal esophagus in a patient with herpetic esophagitis
proved by viral culture and cytology.19(3700) Esophageal manometry 4 weeks later revealed a
hypotensive lower esophageal sphincter and absent peristalsis in the distal 7 cm of the esophagus.
Whether the low sphincter pressure and abnormal motility permitted enhanced mucosal injury via
reflux or whether viral esophagitis led to the reduced sphincter pressure and a motility disorder was not
determined. In a more recent retrospective study, gastrointestinal bleeding was attributable to herpetic
esophagitis in 8 of 23 patients.30(3701)
A tracheoesophageal fistula complicating herpetic esophagitis in a patient with AIDS has been
reported.31(3702)
Active herpetic infection is best demonstrated histologically in the relatively intact epithelium adjacent
to the ulcer margin, and thus this area is the preferred site for obtaining biopsies.13,14,18(3703) The
ulcers are typically shallow and do not penetrate the muscularis mucosa. The ulcer bed contains acute
and chronic inflammatory cells, fibrin, and necrotic debris devoid of the microscopic findings
characteristic of herpes infections. In some cases, concomitant herpes and monilial invasion of the
tissue may be demonstrated.20,27(3704)
Confirmation of viral esophagitis requires the use of accessory techniques. Either a positive viral
culture of biopsy material, a biopsy that reveals the diagnostic eosinophilic intranuclear inclusions
(Cowdry type A bodies), or cytologic smears showing multinucleated epithelial cells with ground-glass
changes in the nuclei will confirm the diagnosis of herpes simplex infection. The discrete intranuclear
inclusions are often surrounded by a clear halo, intranuclear vacuoles, and a ground-glass appearance
of the nucleus with deposition of chromatin beneath the nuclear membrane.13,14(3705) Ultrastructural
study by electron microscopy may demonstrate the virus within epithelial cells. CMV esophagitis is
reported to have distinctive histologic features that permit differentiation from herpetic
esophagitis.32(3706)
Immunohistochemical staining of infected cells with monoclonal antibodies to herpes simplex virus is
both sensitive and specific. Culture of infected tissue is the most accurate method of diagnosis, having
the least false negativity.31(3707)
Serologic confirmation of acute herpetic infection may also be obtained by observing a fourfold rise in
complement-fixing antibody to herpes simplex virus type I in the patient's serum. The opportunity for
serologic, cytologic, or biopsy confirmation decreases rapidly 48 hr after the onset of the
lesion.18(3708) Endoscopic observations may provide the only clue to the diagnosis in at least half of
the patients.
After confirmation of the diagnosis, appropriate therapy consists of supportive measures such as liquid
to soft diet, topical anesthesia with viscous lidocaine (Xylocaine) or an antacid-anesthetic combination,
and analgesics as needed. The symptoms abate and the lesions heal within 7 to 21 days, usually in
less than 14 days.18(3709) In severe cases of herpes simplex and herpes zoster infections diagnosed
antemortem, especially in high-risk, immunocompromised patients, the use of agents such as acyclovir
and adenine arabinoside have proved effective.25,28,33(3710)
Cytomegalovirus Esophagitis
CMV, another opportunistic infectious agent in patients with AIDS, should be understood by
endoscopists. CMV is one of the most common infectious agents found in these immunocompromised
patients and most often causes ulceration in the esophagus and colon. Patients usually present with
dysphagia and odynophagia, but diagnosis may be delayed because of concomitant esophagitis owing
to multiple infectious agents. Upper gastrointestinal bleeding was the presenting manifestation in 15%
of patients in the series of Wilcox et al.34(3711)
CMV infection may produce large ulcers in the distal esophagus, with discrete margins showing little or
no inflammatory changes. The large ulcer is usually shallow and may be solitary or multiple. In a series
of 33 patients with CMV esophagitis identified endoscopically by Wilcox et al.,34(3712) 28% had large
lesions and 43% had lesions that were less than 1 cm in diameter. Multiple ulcers were present in 58%
and diffuse erosive disease in 6% of patients. In the majority of patients, ulcers were located in the
mid- or distal esophagus. Although most ulcers were characterized as shallow, deep lesions were
found in 8% of patients. In 12% of patients, the ulcers were described as having a "heaped-up"
appearance. Based on these findings, Wilcox et al.34(3713) emphasized that CMV esophagitis has
variable endoscopic appearances, with multiple, well-circumscribed ulcerations being the most
common.
Endoscopic biopsy is essential for diagnosis, and specimens must be obtained from both ulcer base
and margin. It is most helpful if deep biopsies are obtained from the ulcer bed and margin because the
virus typically infects endothelial cells and fibroblasts. Cytologic brushings also may reveal endothelial
or fibroblast cytoplasmic inclusions. Bonacini et al.35(3714) reported that viral cultures of brushings or
biopsy specimens are more sensitive (67%) than histologic examination (35%) for viral esophagitis.
The squamous cells are not infected and therefore do not provide any assistance on histologic
examination.
CMV esophagitis is defined as grossly abnormal esophageal mucosa with histologic or culture
evidence of CMV. The histologic criteria used to identify CMV-infected cells include enlargement of
infected cells (three to four times the size of normal cells); a homogeneous eosinophilic intranuclear
inclusion body; and granular intracytoplasmic inclusion bodies. The diagnosis is considered accurate
when two of the three criteria are pres-ent. Immunoperoxidase produces nuclear or cytoplasmic brown
staining and has been reported as more sensitive than H & E staining by some36(3715) and less
sensitive by others.37(3716)
Some evidence indicates that polymerase chain reaction in esophageal biopsies offers promise as a
rapid test to rule out CMV infection in gastrointestinal tissue.37(3717)
Currently available therapeutic agents approved for use in the United States for CMV esophagitis
include ganciclovir and foscarnet. Variable therapeutic results have been reported with ganciclovir.
Foscarnet has been reported to produce complete symptom relief and ulcer healing in 15 of 17
patients with only 3 relapses during 7 months after treatment.38(3718)
Laine et al.39(3719) reported that patients with AIDS and Candida esophagitis often have evidence of
CMV infection, but the symptoms usually are due to Candida and resolve with antifungal therapy.
Consequently, ganciclovir should be reserved for patients with evidence of CMV esophagitis without
Candida on post-antifungal treatment or follow-up endoscopy.
Human Immunodeficiency Virus Esophagitis

Evidence exists that acute esophageal ulcers are caused by the human immunodeficiency virus (HIV)
at the time of seroconversion, but this can rarely be confirmed.40(3720) Other evidence indicates that
the giant chronic ulcers that occur in patients with AIDS are probably the result of HIV infection. These
lesions are negative on biopsy for the known viral and fungal agents. Examination of biopsies of the
base of these ulcers by in situ hybridization revealed HIV RNA in lymphocytes and mononuclear cells.
Patients usually present with severe odynophagia.41(3721)
At endoscopy, there may be single or multiple, small in most cases, or large, deep ulcerations with
transverse ridges visible in the base representing circular muscle bundles of the esophageal
muscularis propria. The margins show variable degrees of inflammation, are often irregular, and
overhang the ulcer margin. These ulcers are similar in appearance to CMV ulcers, and every effort
should be made to distinguish the two because of differences in therapy. HIV-related ulcers have been
shown to either resolve spontaneously or respond to sucralfate and corticosteroid therapy, with either
marked symptom improvement or healing.42,43(3722) CMV-related ulcers require antiviral therapy.
Wilcox and Schwartz43(3723) provided an excellent review of the endoscopic features of idiopathic
esophageal ulceration associated with HIV infection. The report was based on the findings of 68
idiopathic ulcers in 23 patients. Multiple ulcers were present in 57% of patients, usually in the distal
esophagus, and were considered of giant diameter in about 33%. The ulcers were characterized as
either shallow or intermediate in 53% and deep in 7%. The ulcer or margin, or both, were noted to be
heaped up in 40% of the patients. Mucosal bridging was seen in 2 patients. Since these lesions were
not uniform in number, size, or appearance, biopsy with histologic examination is imperative for
accurate diagnosis.
The only effective therapy currently available is corticosteroids, to which most patients can be expected
to respond. A course of prednisone, 40 mg daily by mouth, tapering 10 mg per week for a 1-month
course, has been recommended as safe and effective with a low relapse rate.43(3724)
Other Causes of Large Esophageal Ulcers

Large esophageal ulcers have also been reported as caused by the papovavirus,44(3725) and some
were cited as being induced by direct drug injury to the esophagus.45(3726) Large, solitary esophageal
ulcers were reported in two patients with AIDS and multiple circumferential ulcers in one AIDS patient;
in each case, the ulcers were located at the level of the aortic arch. This location, a history of taking
zidovudine in the recumbent position without fluid, and a dramatic disappearance of symptoms in all
with confirmed healing in two patients after stopping zidovudine treatment suggested that these lesions
were due to direct esophageal injury by this drug.46(3727) No endoscopic feature will allow
differentiation of these large ulcers from other causes. A good history plus biopsy with histology and
accessory procedures will be necessary to define their etiology.
Bacterial Esophagitis
Infectious esophagitis may be caused by a number of organisms and is an uncommonly suspected
diagnosis.47(3728)
Bacterial esophagitis has been defined as histopathologically demonstrable bacterial invasion of
esophageal mucosa or deeper layers with no concomitant fungal, viral, or neoplastic involvement or
previous surgery of the esophagus.48(3729) Organisms reported as responsible for acute esophagitis,
usually in immunocompromised patients with other severe diseases, include: group A -hemolytic
streptococcus, which produces a diffuse phlegmonous esophagitis when submucosa and muscularis
propria are involved (also called esophagitis dissecans superficialis and esophagitis exfoliativa in older
reports);47(3730) mixed bacterial infections;49(3731) and Lactobacillus acidophilus.50(3732) Cultures
should be taken in any case of esophagitis that does not have the clinical and endoscopic criteria for
the diagnosis of reflux or chemical-induced injury. This practice is especially important in
immunocompromised patients.48(3733)
In the case of acute streptococcal esophagitis reported by Howlett,51(3734) barium radiography
showed only some mucosal irregularity in the distal esophagus. Endoscopy revealed marked exudative
inflammation throughout but more severe in the proximal esophagus. The response to penicillin
therapy was dramatic and complete. From endoscopic observations reported in these rare cases of
bacterial esophagitis, no distinctive morphologic features suggested the diagnosis. The endoscopic
appearance may include exudate in plaques or more complete pseudomembranes, bleeding erosions,
early erythema, or in some, normal-appearing mucosa.48(3735)
Tuberculous Esophagitis
The recent increase in systemic tuberculosis plus the rising number of immunocompromised patients
with AIDS is leading to a rising incidence of the once rare esophageal tuberculosis. The esophagus
has usually been involved by direct spread from contiguous mediastinal lymph nodes, the
tracheobronchial tree, or regional lymphatic channels. Consequently, most lesions are found in the
midesophagus about the level of the carina. Mediastinal fibrosis due to tuberculosis can also cause
esophageal compression and dysphagia without direct evidence at endoscopy of esophageal
involvement.52(3736)
Tuberculous esophagitis may present with dysphagia, odynophagia, bleeding (ulcers or
arterioesophageal fistula), or cough related to development of a tracheo-esophageal
fistula.5357(3737) Diagnosis should be suspected based on barium esophagogram, especially in
patients with AIDS, in whom tuberculosis must always be considered.58(3738) A few cases have been
described in which the esophagus was considered the primary site of tuberculous
infection.59,60(3739)
Tuberculosis of the esophagus may have several endoscopic appearances: an irregular, gray-based
ulcer with rolled nodular margin; a hypertrophic or a granular-appearing lesion; or simply a protruding
subepithelial mass. The ulcerative form is most common; ulcers may occur as solitary or multiple in
various sizes, with shallow smooth edges and a gray-white base. The surrounding mucosa may
contain small nodules or ulcerations. The hyperplastic form produces a granulomatous fibrosis in the
esophageal wall with stricture of the lumen. The least common presentation is the granular form with
multiple miliary granulomas in the mucosa.61(3740) All of the forms of the lesion may mimic
carcinoma, and diagnosis is often difficult.59(3741)

Biopsies infrequently reveal caseating granulomas and acid-fast bacilli, but specimens should be taken
from the edge of the lesions. Diagnosis is often dependent on clinical factors, the radiologic findings,
and response to antituberculosis therapy. The differential diagnosis must include other uncommon
granulomatous infections.
Esophageal tuberculosis is usually cured with anti-tuberculosis chemotherapy. In some cases, fistulas
may heal on medical therapy.62(3742) Successful treatment is less likely in patients with AIDS, who
may develop disseminated tuberculosis as an early manifestation. The risk of death from the disease is
significantly higher in these patients.
Histoplasmosis
A case report of disseminated histoplasmosis in a 49-year-old Hispanic man with AIDS has been
reported in which the fungus also involved the esophagus.63(3743) The endoscopic findings included
numerous round, umbilicated submucosal nodules, 3 to 8 mm in diameter, in the distal half of the
esophagus. Silver stains of endoscopic biopsies demonstrated the presence of the organism within the
squamous mucosa.
Chemical-Induced Esophagitis
Corrosive Esophagitis
Esophageal injuries due to corrosive ingestion, accidental or suicidal, have been recognized for at least
200 years. Although once fashionable among those with suicidal intent, corrosive ingestion has
become recognized as a poor method to end one's life. Often the consequence is either no esophageal
damage at all or significant nonfatal esophageal injury that leads to chronic, painful, costly, and
regrettable suffering for the patient.
The Federal Caustic Act, which became law in the United States in 1927, helped to reduce accidental
esophageal injury. Potent corrosive chemical agents, although modified and better labeled and
packaged, are still readily available.64(3744) The stage is set for this horrible injury to occur if suicide
is the goal, the patient is in an alcoholic stupor, or careless adults allow children access to toxic
substances. Caustic ingestion in children is also discussed in Chapter 53:
Esophagogastroduodenoscopy in the Pediatric Patient.
There has been much debate over the pathogenesis and early therapy for corrosive
injuries.6571(3745) The site and extent of injury and the tissue effect produced by caustic agents vary
primarily with the type of ingestant (liquid, paste, powder, or solid), the quantity ingested, and the
duration of tissue contact. Alkalis such as lye (concentrated sodium or potassium hydroxide) dissolve
tissues by liquefaction necrosis and may diffuse rapidly through several layers. In contrast, acids
produce coagulation necrosis, a tissue response that tends to limit depth of penetration and possibly
the risk of perforation. As a rule, alkalis tend to injure the esophagus more than the stomach and vice
versa for acid substances. However, ingestion of acid can cause esophageal injury with subsequent
stricture formation.72(3746) Because of the everchanging compositions of commonly used corrosive
agents such as drain cleaners, it is recommended that physicians contact a poison control center for
information before making major decisions on diagnosis and therapy for ingestion of a suspected
corrosive substance.
Patients who claim or are suspected to have ingested a corrosive agent should be evaluated carefully.
The clinical presentation varies from asymptomatic to profound shock with either severe chest or
abdominal pain, or both, and signs of infection. Chemical injury to the lips or oropharyngeal mucosa is
not reliable, by either its presence or its absence, as an indicator of the existence or extent of
esophageal injury. In several studies, these signs and symptoms of corrosive ingestion did not reliably
identify all patients with serious esophageal injury.7376(3747) Early endoscopy, within 24 hr, provides
the most precise answer to these questions. Older reports warn of perforation when rigid endoscopes
are used, but recent experience with small-diameter, forward-viewing, flexible endoscopes does not
indicate an increased risk when precautions are taken.
After initial clinical assessment and supportive care, an esophagogram should be obtained in patients
suspected by history and physical examination of having esoph-ageal or gastric injuries. A
water-soluble contrast medium is generally recommended. Radiography alone is inadequate to assess
such injury. When there is no clinical or radiographic evidence of perforation, mediastinitis, or
peritonitis, it is reasonable to proceed with endoscopy unless contraindicated. In addition to signs of
perforation, contraindications include severe respiratory distress with pharyngeal edema and necrosis,
shock, or the nonavailability of a physician experienced in endoscopy.
Concern over proper therapy and potential complications can be resolved in many cases of
questionable ingestion or minor injury by a complete endoscopic examination of the esophagus,
stomach, and proximal duodenum. The procedure can be performed in nearly all adult patients and
most older children under the standard sedative-analgesic premedication for routine endoscopy. The
endoscope should be passed under direct vision with careful examination of all visible mucosa from
mouth to proximal duodenum, using as little insufflated air as possible to achieve an adequate
examination. The location and extent of all injuries should be recorded with precision. On completion of
the gastroscopy, as much air as possible should be aspirated via the endoscope.
The grading of degree of injury is difficult, especially when endoscopy is delayed beyond the first 24 hr.
However, some criteria have proved helpful over the years.
Mucosal hyperemia and edema without exudate or ulcerations that exhibits minimum to no friability on
contact with the endoscope is classified as first-degree injury (Figure 455). Although a few small
erosions may be pres-ent, there is no significant bleeding. This degree of injury is confined to mucosa
and consequently not associated with stricture formation or other major complications.
(3748)Figure 455. Endoscopic view of corrosive injury of the esophageal mucosa due to lye.
Hyperemia without exudate, erosion, or ulceration (shown on the right side) is compatible
with first-degree injury. On the left side are exudate, patches of hyperemia, and some
petechiae, indicative of second-degree injury.
The main difficulty in relating diagnosis to prognosis occurs with the forms of second-degree injury
(see Figure 455). More inflammation with erosions, exudate, friability, bleeding, and esophageal
spasm is seen. More advanced stages of second-degree injury show ulcers, more bleeding and
exudate, and areas suggestive of necrotic mucosa with impending mucosal slough. Differentiation
between advanced second-degree and third-degree injuries is difficult. Prognosis is indefinite for this
stage, but therapy probably should be the same as for third-degree injury because endoscopic
estimates of the depth of injury are imprecise.
The patient who has signs of mediastinitis, peritonitis, pneumomediastinum, pleural effusion, or
tracheo-esophageal fistula obviously has a third-degree injury. Other procedures, namely
surgical-therapeutic maneuvers, may be more appropriate than endoscopy for acute management.
Endoscopic findings include large areas of mucosal slough with hemorrhage and areas of
gray-to-black membrane formation with the appearance of an eschar. The lumen may be practically
closed by edema, or it can be dilated without evidence of tone or peristalsis, a sign of deep mural
injury. There may be normal areas, or the entire esophagus may be involved. When these ominous
findings are first detected, the examiner may proceed cautiously, with minimum insufflation of air as
long as visualization is adequate. Retroversion maneuvers should be avoided. It is helpful to know the
precise extent of injury to stomach and duodenum as well, when complete endoscopic examination is
possible. The object is to obtain a rapid but thorough assessment with minimum risk of
endoscope-related aggravation of the underlying pathologic process.

Further radiographic contrast study may be desirable after the patient's condition has stabilized during
the first several days postinjury. Radiographic findings are not reliable with milder injury, but are usually
abnormal to some extent with second- and third-degree injuries. The detailed radiographic findings
have been reviewed.77,78(3749)
Accurate diagnosis of the degree and extent of corrosive injury permits confident conclusions with
regard to prognosis, natural history, and methods of therapy. Only supportive care and follow-up
recommendations are needed for cases of accidental ingestion without any detectable mucosal injury.
Patients with no mucosal injury who have attempted suicide should be hospitalized for observation and
psychiatric evaluation.
Patients with identifiable mucosal injury of the pharynx, respiratory tract, esophagus, stomach, or
duodenum should be hospitalized. Supportive care should be consonant with the degree of injury, and
early surgical consultation should be included as appropriate. A nasogastric tube, used for aspiration
early in the course and later for enteric feeding, should be placed promptly after endoscopy. The
duration of time the tube should remain in place and the appropriate point to begin feeding vary and
should be determined according to the type of injury and the patient's general status. An indwelling
nasogastric tube preserves access to the stomach for future dilation.
The role of corticosteroid and antibiotic therapy in corrosive injury has not been established by
controlled studies in adult humans.67,68(3750) In a controlled, randomized trial involving 60 children
who had ingested a caustic substance, Anderson et al.79(3751) found that cortico-steroids provided no
benefit. In particular, no significant difference was found in the development of strictures or need for
surgery between the treatment and the nontreatment groups. Corticosteroids would probably be helpful
if given before the damaging agent was ingested, as suggested by animal studies.80,81(3752)
Because cortico-steroids cannot be given before the injury, are of unproven value, and may enhance
the progression of septic complications, we avoid their use.
The use of antibiotics appears rational in patients with second- and third-degree burns to minimize the
risks of local and systemic sequelae of the bacterial infections that develop soon after these types of
injuries. Various antibiotics have been recommended, but no regimen has proved superior in humans.
The goal is to provide antibiotic coverage initially for those organisms usually found in the oropharynx
that will most likely contaminate the area of injury.
A prospective study of the role of flexible endoscopy in 81 patients led to a recommendation for use of
a more detailed endoscopic classification system.82(3753) The standard Grades 0 to 3 classification
was modified by subdividing Grade 2 burns into Grades 2a and 2b, and Grade 3 burns into 3a and 3b.
Such subdivision allowed relatively accurate prediction of sequelae and prognosis. No complications
occurred in patients with Grade 0, 1, or 2a injury. The risk of subsequent stricture formation was 71.4%
with Grade 2b. Patients with Grade 3a burns did not require emergency surgery, whereas among those
with Grade 3b the mortality rate was 64.7% and 66.7% of the survivors required surgery for sequelae.
Early surgery may reduce mortality and morbidity in patients with Grade 3b injuries. All major
complications and deaths occurred in patients with Grade 3b injuries. This grading system also served
as a reliable guide for therapy and hospitalization time. No complications of endoscopy occurred in the
81 patients having a total of 381 endoscopic examinations performed within hours of corrosive
ingestion or between the 3rd and the 10th weeks. Strictures developed in 71% of those with Grade 2b
injury and in all survivors with Grade 3 injury.
Once the extent and course of the injury are determined, early peroral dilation should be
considered.47,65,83(3754) At the latest, this should begin within 7 to 10 days postinjury in those
patients without extraesophageal complications. Dilation should be performed under fluoroscopic
control without force. The starting time for dilation in those patients with complicated third-degree
injuries must be determined on an individual basis. However, these patients are very likely to develop
severe stricture, usually by the 3rd week, so that earlier rather than delayed dilation is desirable.
Dilation can be started when the patient's condition is stable, usually within the first week, depending
on the grade of injury. Dilation using Savary dilators over a guidewire under fluoroscopic control is a
preferred method. The details of peroral dilation technique and frequency are reported
elsewhere.83,84(3755) In cases of severe or neglected esophageal injury, retrograde dilation with
Tucker dilators is the safest and most effective therapy.70,83(3756) Technical aspects of esophageal
dilation are also discussed in Chapter 44: Hiatus Hernia and Peptic Diseases of the Esophagus. The
use of esophageal stents has been evaluated for prevention of strictures in infants and children and
appears to be effective in the small number of cases reported thus far.69,71(3757)
Long-term follow-up must take into consideration the tendency for severe strictures to recur and their
increased risk, many years after ingestion of a chemical agent, of squamous cell
carcinoma.85,86(3758) In the study of Broor et al.,87(3759) the initial success rate for dilation of
corrosive-induced strictures was equal to that for peptic strictures. Long-term, however, the number of
symptomatic recurrences per patient-month of follow-up was significantly higher for patients with
corrosive-induced strictures, although the recurrence rate decreased over the course of time. At 12
months, it was significantly less than that during the first 6 months of follow-up, and after 3 years, there
was no significant difference in recurrence rate. A perforation occurred in 9 of 1373 dilations, although
8 of these occurred in patients with chemical-induced strictures.
Drug-Induced ("Pill") Esophagitis

Although more than 50 drugs are recognized as causes for over 450 reported cases of drug-induced
esophagitis, this potentially lethal disorder is little known among primary care physicians and many
specialists. Esophageal injury due to "pills" was first reported in 1970 by Pemberton.88(3760) Several
comprehensive reviews have been published.8993(3761)
Ordinarily, little thought is given to instructing patients in the proper way to swallow medications that
are in solid form, either tablet or capsule. The delay in passage of tablets through the esophagus under
the influence of recumbency and a small volume of fluid was amply demonstrated by Evans and
Roberts.94(3762) They found a delay in passage of an aspirin-sized barium tablet through the
esophagus beyond 5 min in over 50% of subjects when in the recumbent position. The message is
clearavoid low fluid volume and recumbency when taking any form of medication.95(3763)
The more commonly used drugs that have caused esophageal injury in the United States include
potassium chloride tablets (especially the wax matrix form such as Slow-K), tetracycline, doxycycline,
clindamycin, alendronate (Fosamax), quinidine gluconate (Quinaglute), ferrous sulfate, nonsteroidal
anti-inflammatory drugs, and ascorbic acid.88,89,96105(3764) The drug formulation appears to be an
important factor in the potential for delayed passage and mucosal injury. For instance, capsules of
doxycycline have been shown to remain in the esophagus three times as frequently as tablets of the
same compound.99(3765) The size of the tablet or capsule seems important as well. Three of our last
five cases of "pill" esophagitis have been severe esophageal strictures at the level of the aortic arch in
patients taking quinidine gluconate. This tablet is 13 mm in diameter.
Common factors in the genesis of this form of esophagitis are esophageal "physiologic" narrowing at
the level of the aortic arch, obstruction, disordered esophageal motility, or any combination of these. In
the absence of obstruction, the patient usually gives a history of taking the medicine while in a
recumbent position or without sufficient liquid just before retiring at night. The obstructive problem that
seems to enhance the development of drug-induced injury most often is esophageal stenosis owing to
inflammation. Drug-induced esophagitis may also be associated with esophageal motor abnormalities,
such as achalasia, esophageal spasm, or with left atrial enlargement, which may cause compression
and partial occlusion of the esophageal lumen. The most common contributory factors are ingestion of
the drug without adequate fluid, recumbency immediately after ingestion, tablets or capsules of large
size, and the chemical composition of the drug.

The pill-taking habits of the patient are a vital aspect of the history in suspected cases. Certain
diseases and their related therapy suggest the diagnosis and site of injury. The patient with
cardiomegaly and left atrial enlargement typically sustains esophageal injury from potassium chloride
or quinidine at a level 30 to 35 cm from the incisor teeth, that is, at the level of esophageal
compression by the enlarged left atrium (Figure 456). The person without intrinsic esophageal
disease or cardiomegaly will more often develop pill esophagitis from transient impaction of drugs at
the level of the aortic arch (22 to 24 cm) or proximally (Figure 457). After discontinuation of the drug,
the acute mucosal injury heals within several weeks without other therapy and severe mural fibrosis,
and stricture is the end result (Figure 458). The teenager with acne or the traveler (likely a physician
or nurse) trying to prevent traveler's diarrhea typically presents with a lesion in the proximal esophagus
caused by doxycycline. The patient with arthritis may develop esophagitis from aspirin, naproxen
(Naprosyn), indomethacin, or other anti-inflammatory drugs.106(3766) In Great Britain, a major cause
is emepronium bromide, an agent used to treat urinary frequency.107,108(3767) Another important
point is that drug injury may aggravate, or render atypical, the lesions of reflux esophagitis.91(3768)
(3769)Figure 456. Endoscopic view of severe inflammatory stenosis at the level of the left
atrium (30 to 35 cm) caused by potassium chloride formulated in a wax matrix (Slow-K). The
patient had cardiomegaly and left atrial enlargement that compressed the esophagus.
(3770)Figure 457. Endoscopic view of severe inflammation, ulceration, and fibrosis that
resulted in a very tight stricture at the level of the aortic arch (23 cm). The squamous mucosa
below the stricture was entirely normal to the squamocolumnar mucosal junction at 38 cm.
Quinidine gluconate was the responsible drug.
(3771)Figure 458. Severe fibrotic stricture due to quinidine gluconate examined several
weeks after discontinuation of the drug. Intramural fibrosis has produced the linear elevations
radiating from the stricture at the 12, 5, and 7 o'clock positions. Because of the severe mural
injury, gradual dilation over several months is required to treat this type of stricture.
Over 50 drugs have been reported to cause esophageal injury. Injury is probably asymptomatic in most
cases. The classic symptoms are odynophagia (74%), continuous retrosternal pain (72%), dysphagia
(20%), or combinations of these symptoms. Elderly patients may have dysphagia alone, and often
drug-induced esophagitis and stenosis are never considered in the differential diagnosis. Our most
recent five patients complained of severe dysphagia without odynophagia or chest pain and were
considered to be suffering from occult malignancy because of their advanced ages and the location
and radio-graphic appearance of the stenotic lesions. A proper history suggested the correct diagnosis
in all five. Three were taking quinidine gluconate and had strictures at the aortic level, the fourth was
taking potassium chloride and quinine sulfate in tablet form, and the fifth potassium chloride (Slow-K).
The latter had a stricture at 30 cm, the level of his enlarged left atrium that was compressing his
esophagus (see Figure 456). With continued ingestion of the offending drug, the esophageal injury
may extend proximally from the original stenosis, thereby making diagnosis and therapy more difficult
and increasing the risk of severe injury. Potassium chloride injury has been the most lethal, with at
least nine deaths resulting from either ulceration with bleeding, perforation, or mediastinal abscess and
from the long-term effects of severe esophageal stricture.96(3772)
It has become apparent that certain drugsfor example, doxycycline, ferrous sulfate, and emepronium
bromideproduce only acute mucosal injury that heals completely when the drug is stopped and does
not lead to stricture formation. Others, such as quinine sulfate and quinidine gluconate, will cause
severe esophageal stricture (see Figures 457 and 458).
Barium contrast radiography may provide clues by demonstrating esophageal stenosis or ulceration,
but these findings are misleading unless drug-induced injury is suspected from the history. The atypical
location of such injury in the region of the aortic arch or opposite an enlarged left atrium in patients
without known prior esophageal disease should suggest drug-induced injury or malignancy.
Reflux-related lesions do not occur at either of these levels unless a columnar-lined esophagus is
present. Double-contrast radiography may provide more detail, but without a proper history, accurate
diagnosis is unlikely.
Documentation of drug-induced esophagitis requires esophagoscopy. The findings vary but may
include one or more of the following: a focal area of marked edema and erythema, usually with some
degree of mucosal erosion covered by a pseudomembrane or exudative plaque; "punched-out"
ulceration, possibly "kissing" type ulcers on opposite walls; and either inflammatory stenosis or stricture
(see Figure 456). Instances of prolonged and repeated injury manifest longer segments of
inflammation, ulceration, and fibrosis, findings most often diagnosed as probably malignant in nature
but with negative findings by biopsy (see Figure 457). No endoscopic characteristics are helpful for
diagnosis other than the focal nature of the lesion, its atypical location, or its occurrence just proximal
to an area of esophageal stenosis. Residual particles of the medication at a stricture may provide a
clue to diagnosis.97(3773) When drug injury is superimposed on a preexisting reflux stricture or occurs
in a patient with achalasia, the diagnosis can be confusing and long delayed. Drug injury should be
considered as a possible etiology in all cases of esophagitis and stricture until proved otherwise.
The chemical properties of the medication as well as the duration of mucosal contact are important
factors in the degree of esophageal injury. In most cases, therapy is simple. Ingestion of the possible
offending drug (tablet or capsule) should be stopped, at least until the evaluation is completed. The
medication may be resumed if neither stenosis nor motility disorder is present, provided the patient
uses ample fluid and remains upright after swallowing, uses a liquid form of the drug, crushes the
tablet, or empties the contents of capsules into a suitable liquid for ingestion. Inflammatory erosion and
stenosis may resolve with such changes in dosing procedure, but strictures usually require peroral
dilation over variable periods of time. A complete evaluation of strictures by endoscopy and biopsy to
eliminate an occult neoplasm or other lesion as the reason for the pill impaction is absolutely essential.
All patients, especially those bedridden and those with esophageal motor disorders or esophageal
stenosis of any type, must be warned regarding the danger of ingesting any medication in the
recumbent position or with insufficient amounts of liquid immediately before going to bed. All patients
should be advised to take all tablets or capsules with at least 2 oz of fluid and then to remain in the
upright position for at least 10 min. For patients who are bedridden, larger volumes of fluid should be
given or liquid forms of drugs prescribed.
Radiation-Induced Esophagitis
Radiation-induced esophagitis is usually the result of mediastinal irradiation for bronchogenic
carcinoma.109,110(3774) Less common causes include irradiation for metastatic breast and testicular
carcinoma, lymphoma, and primary squamous cell carcinoma. The dosage and rate of exposure that
result in symptomatic esophagitis vary, but most patients have received between 2500 and 6000 rads.
The natural history of this condition is not well understood.
The clinical syndrome associated with radiation injury is manifested primarily by dysphagia,
odynophagia, and spontaneous chest pain. The barium contrast esophagogram demonstrates no
morphologic abnormality in most cases, but altered esophageal motility is the most frequently noted
abnormality.111,112(3775) Early in the course, the dysphagia seems to be due primarily to edema and
motor disturbances, but from 3 to 18 months after therapy, a stricture will usually be found. The
proximal end of the stricture typically is smooth and tapers gradually. Angulation of the lumen at the
level of the stricture may be due to radiation effects, adherence of the esophagus to adjacent
mediastinal structures, or recurrent neoplasm.
In some patients, esophagoscopy may help to explain symptoms or uncertain findings demonstrated
by radiographic study. The nature of the mucosal changes found by endoscopy depends on the
dosage, rate, and duration of therapy, and perhaps other factors such as prior or concomitant
administration of chemotherapeutic agents that may enhance irradiation effects.113(3776) Infection by
C. albicans or the herpesviruses may be superimposed. The endoscopic appearance of the mucosa
may include hyperemia, friability, erosions or ulceration, and exudate. If the patient survives several
years, telan-giectatic vessels are usually seen in the mucosa in the field of irradiation. Biopsy reveals
nonspecific acute and chronic inflammation. Special stains for monilial pseudohyphae and a search for
cellular changes of viral esophagitis should be included in the evaluation. The possibility of recurrent
malignancy must always be considered when interpreting endoscopic findings.
Ideally, prevention of radiation injury would be the best alternative, with further modifications in the
mechanics of radiation delivery and the use of arachidonic acid metabolism pathway inhibitors and
radioprotectants.109(3777)
Therapy when the condition is acute consists first of modifying the irradiation dosage and rate of
subsequent treatments. Supportive care by dietary modifications and administration of topical
anesthetics and analgesics before meals and at bedtime is usually adequate for the limited course of
this disorder.
The best therapy for postirradiation stricture is peroral dilation.83(3778) A more gradual dilation
program may be required for this type of stricture because the esophageal stenosis is very firm and
more elongated than strictures due to reflux. Tapered polyvinyl (e.g., Savary) dilators passed over a
guidewire are the best instruments for the initial dilation program in such cases. Later, mercury-filled
bougies (e.g., Maloney) dilators may be adequate for maintenance of patency. In a series of 103
consecutive patients with radiation-induced esophageal strictures reported by Swaroop et
al.,114(3779) dilation provided effective relief of dysphagia in 66% of patients. The dysphagia-free
interval ranged from 3 to 84 weeks, with a median of 16 weeks. Among the 75 patients followed
long-term, 32 required repeated dilation. Procedure-related perforations occurred in 2 patients and 1
patient developed a delayed tracheoesophageal fistula; 2 patients expired as a result of
treatment-related complications. Nelson et al.115(3780) found intramural hydrocortisone injected via a
transendoscopic needle to be helpful as a supplement to peroral dilation in more resistant cases, but
others found this to be a difficult exercise that is not definitely helpful.
Stasis Esophagitis
In some patients with chronic esophageal obstruction, a significant amount of residual food material
and medications may remain in contact with esophageal mucosa for days. The classic circumstance is
achalasia. The mucosa becomes irritated to a point that visible changes are evoked that appear as
patchy areas of hyperemia, rarely with any erosion or ulceration unless pill esophagitis is
superimposed. These findings may be widely scattered but tend to be prominent in the distalmost
esophagus, probably because retained material remains longest in contact with the mucosa in this
region. The epithelium appears thickened with creasing, an appearance that suggests elephant skin.
Histologically, the alterations are nonspecific, and biopsy is not helpful unless a more significant
abnormality such as early malignancy is suspected. A fibrotic stricture can be the end result in some
cases.
Nonspecific Esophagitis
Eosinophilic Esophagitis
Eosinophilic esophagitis has been reported as a component of the eosinophilic gastroenteritis

syndrome.116118(3781) Eosinophilic gastroenteritis is an uncommon disorder characterized by
eosinophilia, infiltration of the stomach or small intestine by eosinophils, and symptoms related to the
site and severity of involvement. As many as half the patients have either a history of allergy or a
specific food intolerance.119(3782) The clinical presentation correlates with the layer of the gut wall
that is predominantly involved. Bleeding and malabsorption are prominent with mucosal disease,
obstruction occurs with involvement of muscularis propria, and eosinophilic ascites, the rarest form,
occurs with serosal disease. Dysphagia is the primary symptom associated with esophageal
involvement and may be caused by disturbed motility or stenosis.116,118(3783)
The diagnosis is suspected on the basis of symptoms plus eosinophilia. Radiography with barium
reveals evidence of gut wall infiltration. Thickening and rigidity, mucosal nodules, and stenosis may be
seen.118(3784)
A severe stricture appeared at 24 cm from the incisors in a case reported by Picus and
Frank.118(3785) A small-diameter pediatric endoscope would not pass the stricture. The mucosa at
the proximal margin of the stricture was ulcerated, and a 3-mm polyp was removed at that level.
Microscopic study of this and the biopsy specimen revealed edema and eosinophilic infiltration of the
mucosa and submucosa.
Corticosteroid therapy usually provides objective and symptomatic improvement but may be required
on a chronic basis. Residual stricture of the esophagus is managed by peroral dilation.
Behet's Disease
Behet's disease was described first in 1937 in patients with recurrent oral (Figure 459A) and genital
ulcerations and iridocyclitis.120,121(3786) Later, skin, joint, vascular, neurologic, and intestinal (jejunal,
ileal, and colonic) lesions were reported. The intestinal lesions are easily confused with Crohn's
disease.121(3787) Esophagoscopy is not necessary unless a patient has symptoms of esophageal
involvement.122(3788)
(3789)Figure 459. A, Small "aphthoid" erosions of the buccal mucosa in a patient with
Behet's disease. B, Fibrotic stricture and polypoid nodules due to chronic inflammation and
ulceration of Behet's disease. These changes are located just proximal to the ulcerations
shown in C. C, Two deep ulcerations of Behet's disease with little surrounding inflammation
and small amount of exudate in the ulcer bases. Ulcers are shown in the 9 and 12 o'clock
positions. The patient was on moderate-dose corticosteroid therapy.
Esophageal involvement is rare, and up until 1986, only 34 cases had been reported.120(3790) The
esophagus may be involved at any level, but the midesophagus is reportedly more prone to the
development of erosions and ulceration. In some patients, hypertrophic changes and a polypoid
appearance owing to chronic inflammation develop in the mucosa (see Figure 459B). Large ulcers
may develop, and these respond unpredictably to cortico-steroid therapy. Strictures arise in the area of
ulcers and require periodic dilation (see Figure 459C). Perforation and tracheoesophageal fistula
have rarely been reported secondary to Behet's disease.120,121,123(3791)
Dermatologic Disorders
Esophageal involvement may occur in several disorders of the skin, including pemphigus vulgaris,
bullous pemphigoid, benign mucous membrane pemphigoid, and epidermolysis bullosa dystrophica.
Bullae may be noted in the esophagus in all of these conditions, although they are rarely observed in
patients with benign mucous membrane pemphigoid. Esophageal intubation in certain of these
disorders may be hazardous (see later in this chapter).

Pemphigus Vulgaris
Pemphigus vulgaris usually appears during the middle decades of life and is somewhat more common
in women. About half of patients will have esophageal involvement, but this is usually
asymptomatic.124(3792) However, patients may complain of odynophagia, dysphagia, and
occasionally gastrointestinal bleeding. In addition to hemorrhagic bullae, endoscopic findings include
patchy erythema, erosions, white plaques, and occasionally, sheets of desquamated
mucosa.124126(3793) Esophageal involvement may become evident in the absence of cutaneous
manifestations, when the cutaneous findings are in remission, or during treatment for cutaneous
disease.127129(3794) Secondary monilial or herpetic infection may occur in patients treated with
high-dose corticosteroids. In addition to inflammatory cells in the submucosa, acantholysis is the
typical finding in endoscopic biopsies.
Bullous Pemphigoid
Bullous pemphigoid is a relatively benign disease that mainly afflicts older patients. Oropharyngeal
involvement occurs in about 20% of patients; the incidence of esophageal involvement is uncertain
because this is probably asymptomatic in many patients.130(3795) When symptoms occur, they
include dysphagia, odynophagia, and bleeding. Rarely, in symptomatic patients, the esophageal
mucosa may slough as a cast, an occurrence termed esophagitis dissecans superficialis.131(3796)
Stricture formation is relatively uncommon, even after slough of the esophageal mucosa.
Benign Mucous Membrane Pemphigoid
Benign mucous membrane pemphigoid is, despite its name, a severe chronic disease that arises
during the middle decades of life. It is characterized by formation of bullae on various mucosal
surfaces including the esophagus, mouth, and pharynx in addition to the skin. Esophageal involvement
is relatively rare (probably less than 15% of patients) and may be asymptomatic. There are no typical
endoscopic findings, with bullae being noted only infrequently. Furthermore, there is no characteristic
histopathologic finding in endoscopic biopsies. However, passage of the endoscope may be
associated with formation of hemorrhagic bullae.132(3797) Because endoscopy is not likely to provide
a definitive diagnosis and is potentially hazardous, there are relatively few indications for the
procedure. Dilation of esophageal strictures has resulted in at least temporary relief of dysphagia in
reported cases.133(3798) There is relatively little information as to the best technique for dilation, given
the small number of case reports. Obviously, dilation should be as atraumatic and as gradual as
possible. Attempts to dilate the esophagus to a relatively normal luminal diameter may be hazardous if
the stricture is especially narrow.
Epidermolysis Bullosa
Epidermolysis bullosa is uncommon and inherited and appears clinically in three forms. The simple
type is transmitted as autosomal dominant, is characterized by subepithelial bullae after minor trauma,
rarely involves mucous membranes, subsides at puberty, and makes up about 45% of cases. The
hyperplastic dystrophic form, transmitted by a single autosomal gene of high penetrance, accounts for
about 30% of cases, is characterized by deep and superficial cutaneous bullae, and involves mucous
membranes in 20%.134(3799) Most of the serious consequences, such as esophageal strictures and
microstomia, occur in the recessive dystrophic form of epidermolysis bullosa, but are also seen
infrequently in the junctional and simplex forms.135(3800) The most serious and least common form of
this disease, polydysplastic epidermolysis bullosa, appears at birth and worsens with age as a
mutilating and potentially fatal process. Any part of the body, including the esophagus, may be
involved.
Esophageal involvement may occur at any time, is more common in men than in women, and may
result in aspiration, hemorrhage, malnutrition, perforation rarely, and possibly, an increased risk of
carcinoma with prolonged disease. Dysphagia and odynophagia, alone or together, are the prominent
symptoms of esophageal involvement. Dysphagia does not necessarily indicate stricture and may be
due to spasm and pain in the presence of vesicles or ulceration, or both.136138(3801) Dysphagia has
been reported in 20% of patients with the dominant dystrophic form of the disease, with dilation being
needed in about half.139(3802) Dysphagia is most common with the recessive dystrophic form, being
found in 72% of patients in one series, half of whom required dilation.139(3803) Strictures can occur at
any level, but are most commonly located in the cervical esophagus.135(3804)
Barium contrast radiography can reveal disordered motility, mucosal nodules, ulcers, or stenosis.
Stenosis, initially due to inflammation and later to true strictures, is found in the proximal third of the
esophagus in about 50% of the patients, in the distal third in about 25%, and at multiple sites in 25%.
Diffuse stricturing is distinctly uncommon. When multiple, the strictured segments usually are from 2 to
5 cm in length.
The indication for esophagoscopy should be carefully evaluated in each patient. Because there
appears to be, based on older reports, an increased risk of perforation, it has been suggested that
esophagoscopy is best performed when the disease is relatively quiescent or under control with
corticosteroid therapy.134(3805) Such decisions depend to a great extent on the patient's need for
relief of dysphagia.
Esophagoscopy should be performed with a flexible, small-caliber, fiberoptic, forward-viewing
endoscope using direct observation throughout introduction and removal to minimize mucosal trauma.
Multiple mucosal bullae of variable size are seen with active disease and are easily ruptured. The
mucosa is friable, and granulation tissue may be seen at sites of earlier disease activity. With
continued activity, the mucosa may become atrophic and scarred, such changes leading eventually to
strictures that require dilation.
Dilation of strictures with epidermolysis bullosa should not be considered as routine. Special care must
be taken with anesthesia, and all manipulations that may cause trauma to facial skin and
oropharyngeal mucous membranes. Some have advised that strictures be dilated in these patients
when possible during anesthesia that is induced for other purposes.139(3806) The use of
through-the-scope hydrostatic balloon dilators has been suggested as less traumatic than a Maloney or
Savary dilator that exerts "shearing" force; this is an unproven approach.139(3807)
In addition to peroral dilation for strictures and topical anesthetics for relief of pain, dietary
modifications minimize the inevitable malnutrition that will develop if these measures are not instituted.
Corticosteroid therapy is helpful in some cases. Severe strictures may be unresponsive to peroral
dilation. Esophagectomy with colon interposition has been reported as dramatically effective in two
brothers with severe strictures.140(3808) Such treatment should be advised, if ever, only when proper
peroral dilation has failed and expert surgical help is available.139(3809)
Crohn's Disease
Clinically evident involvement of the esophagus is rare in patients with Crohn's disease, although a
number of cases have been reported.141144(3810) In the series of Geboes et al.145(3811) of 500
patients with Crohn's disease encountered over a 4-year period, 9 were found to have esophageal
involvement. In the majority of reported cases, Crohn's disease was known to be present in the more
commonly involved sites such as the colon and ileum. However, a few cases have been reported in
which the presenting manifestations were esophageal in nature despite the presence of asymptomatic
involvement of the distal bowel,146(3812) or in which esophageal Crohn's disease was found in the
apparent absence of involvement of other segments of the gastrointestinal tract.147(3813)
Symptoms and other clinical manifestations of esophageal Crohn's disease depend on the stage of
evolution of the disease. Dysphagia with odynophagia, which may be severe, is a common presenting
symptom.146(3814) In addition to evidence of Crohn's disease elsewhere in the gastrointestinal tract,
patients with esophageal involvement are more likely to have a high Crohn's disease activity index and
extraintestinal systemic manifestations.145,148(3815) As with Crohn's disease elsewhere in the
gastrointestinal tract, fistula formation and the development of dense, fibrotic strictures may occur. In
the five cases reported by Huchzermeyer et al.,149(3816) the principal site of involvement was the
distal esophagus, although a "tendency" to extend proximally was also noted. It has not been
established, however, that there is a preferential site. Huchzermeyer et al.149(3817) divided the clinical
and endoscopic manifestations of esophageal Crohn's disease into two stages: an inflammatory stage,
in which the endoscopic findings range from mild erosive changes to more severe ulceration, and a
stage in which the predominant finding is stenosis and stricture formation. No clinical, endoscopic, or
radiographic findings are absolutely specific for esophageal Crohn's disease. The "cobblestone"
appearance that is typical of Crohn's disease elsewhere in the gastrointestinal tract may be present
rarely. A careful search should be conducted for noncaseating granulomas in endoscopic biopsies, but
discovery of this confirmatory finding is relatively rare.145,149(3818) In many cases, therefore, the
diagnosis becomes one of exclusionin a patient with known Crohn's diseaseof other possible
causes for the clinical, radiographic, and endoscopic findings.
The erosive and ulcerative changes of Crohn's disease are usually reversible with standard treatment
methods, although esophageal lesions may persist in the absence of esophageal
symptoms.146,148(3819) In the study of D'Haens et al.,148(3820) 8 of the 14 patients did not have
symptoms or evidence of esophageal involvement during exacerbations of the disease; in 3 patients,
esophageal manifestations developed in parallel with clinical flare-ups of the disease.
Unlike most other areas of the gastrointestinal tract involved by Crohn's disease, the esophagus is
readily accessible to endoscopic and endoscope-guided treatment measures such as dilation.
Short-term success has been reported with esophageal dilation, using mainly wire-guided techniques,
in a small number of case reports.150152(3821) Mathis et al.150(3822) reported fibrin sealing of a
fistula in one patient, although this patient eventually underwent operation for an esophagobronchial
fistula.
Esophageal Melanocytosis
Focal and diffuse black pigmentation of the esophagus has been found at endoscopy in a small
number of patients.153,154(3823) In a study of 1000 consecutive endoscopic examinations of the
esophagus from India, Sharma et al.155(3824) found evidence of esophageal melanosis in 21 cases.
The findings in the index case consisted of a pigmented mucosal patch; the results of histochemical
staining of endoscopic biopsy specimens were consistent with the presence of melanin. In the
identified cases, the pigmentation was found in the mid- and distal esophagus. Endoscopic evidence of
esophagitis was found adjacent to the pigmented areas in about a third of subjects. Sharma et
al.155(3825) speculated that the prevalence of melanocytes in normal esophageal mucosa may vary
by race. Yamazaki et al.156(3826) performed a detailed study of four samples of melanotic lesions of
the esophagus detected at esophagoscopy that included histochemical staining and electron
microscopy. By comparison with 13 control samples of normal esophageal epithelium, there was an
increase in numbers of melanocytes in the basal mucosal layer, an increase in the quantity of melanin
in these cells, and a transfer of this pigment to keratinocytes, macrophages, and fibroblasts in the
lamina propria. Based on these observations, these investigators suggested that the term
melanocytosis be used rather than melanosis, as the former more accurately describes the histologic
findings.
Primary melanoma is known to arise in the esophagus, although the number of reported cases is
small. In a few cases, the primary tumor has been associated with more diffuse pigmentation in the
esophagus.157,158(3827) This has led to speculation that melanosis (melanocytosis) of the
esophagus may be premalignant. This hypothesis has not been proved, and the significance of
melanocytosis is therefore uncertain at present.
REFERENCES
1.
Kodsi BE, Wickremesinghe C, Kozinn PJ, et al. Candida esophagitis: A prospective study of
27 cases. Gastroenterology 1976;71:7159.
2. Andersen LI, Frederiksen HJ, Appleyard M. Prevalence of esophageal Candida colonization in
a Danish population: Special reference to esophageal symptoms, benign esophageal
disorders, and pulmonary disease. J Infect Dis 1992;165:38992.
3. Eras P, Goldstein MJ, Sherlock P. Candida infection of the gastrointestinal tract. Medicine
(Baltimore) 1972;51:36779.
4. Sheft DJ, Shrago G. Esophageal moniliasis. The spectrum of the disease. JAMA
1970;213:185962.
5. Porro GB, Parente F, Cernuschi M. The diagnosis of esophageal candidiasis in patients with
acquired immune deficiency syndrome: Is endoscopy always necessary?. Am J Gastroenterol
1989;84:1436.
6. Pennazio M, Arrigoni A, Spandre M, et al. Endoscopy to detect oral and oesophageal
candidiasis in acquired immune deficiency syndrome. Ital J Gastroenterol 1992;24:3247.
7.
Andren L, Theander G. Roentgenographic appearances of esophageal moniliasis. Acta
Radiol 1956;46:5714.
8. Kantrowitz PA, Fleischli DJ, Butler WT. Successful treatment of chronic esophageal
moniliasis with a viscous suspension of nystatin. Gastroenterology 1969;57:42430.
9. Gregory DW, Blackburn WD, Madhavan SV, et al. Esophageal candidiasis: Treatment with
ketoconazole. South Med J 1983;76:13078.
10. Laine L, Dretler RH, Conteas CN, et al. Fluconazole compared with ketoconazole for the
treatment of Candida esophagitis in AIDS. A randomized trial. Ann Intern Med
1992;117:65560.
11.
Pearce JA, Dagradi A. Acute ulceration of the esophagus with associated intranuclear
inclusion bodies. Arch Pathol 1943;35:88997.
12. Rosen P, Hajdu SI. Visceral herpesvirus infections in patients with cancer. Am J Clin Pathol
1971;56:45965.
13. Nash G, Ross JS. Herpetic esophagitis. A common cause of esophageal ulceration. Hum
Pathol 1974;5:33945.
14. Lasser A. Herpes simplex virus esophagitis. Acta Cytol 1977;21:3012.
15.
Berg JW. Esophageal herpes: Complication of cancer therapy. Cancer 1955;8:73140.
16. Depew WT, Prentice RSA, Beck IT, et al. Herpes simplex ulceration esophagitis in a healthy
subject. Am J Gastroenterol 1977;68:3815.
17. Owensby LC, Stammer JL. Esophagitis associated with herpes simplex infection in an
immunocompromised host. Gastroenterology 1978;74:13056.
18. Springer DJ, DaCosta LR, Beck IT. A syndrome of acute self-limiting ulcerative esophagitis in
young adults probably due to herpes simplex virus. Dig Dis Sci 1979;24:5359.
19. Fishbein PG, Tuthill R, Kressel H, et al. Herpes simplex esophagitis: A cause of
upper-gastrointestinal bleeding. Dig Dis Sci 1979;24:5404.
20. Agha FP, Lee HH, Nostrant TT. Herpetic esophagitis: A diagnostic challenge in
immunocompromised patients. Am J Gastroenterol 1986;81:24653.
21. Byard RW, Champion MC, Orizaga M. Variability in the clinical presentation and endoscopic
findings of herpetic esophagitis. Endoscopy 1987;19:1535.
22. Mosimann F, Cuenoud PF, Steinhauslin F, Wauters JP. Herpes simplex esophagitis after
renal transplantation. Transpl Int 1994;7:7982.
23. Pazin GJ. Herpes simplex esophagitis after trigeminal nerve surgery. Gastroenterology
1978;74:7413.
24. Whitley R, Barton N, Collins E, et al. Mucocutaneous herpes simplex virus infections in
immunocompromised patients. A model for evaluation of topical antiviral agents. Am J Med
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
1982;73(suppl):23640.
Skucas J, Schrank WW, Meyers PC, Lee CS. Herpes esophagitis: A case studied by
air-contrast esophagography. AJR 1977;128:4979.
Meyers C, Durkin MG, Love L. Radiographic findings in herpetic esophagitis. Radiology
1976;119:212.
Brayko CM, Kozarek RA, Sanowski RA, Lanard BJ. Type I herpes simplex esophagitis with
concomitant esophageal moniliasis. J Clin Gastroenterol 1982;4:3515.
Kadakia SC, Oliver GA, Peura DA. Acyclovir in endoscopically presumed viral esophagitis.
McDonald GB, Sharma P, Hackman RC, et al. Esophageal infections in immunosuppressed
patients after marrow transplantation. Gastroenterology 1985;88:11117.
McBane RD, Gross JB Jr. Herpes esophagitis: Clinical syndrome, endoscopic appearance,
and diagnosis in 23 patients. Gastrointest Endosc 1991;37:6003.
Cirillo NW, Lyon DT, Schuller AM. Tracheoesophageal fistula complicating herpes esophagitis
in AIDS. Am J Gastroenterol 1993;88:5879.
Toghill PJ, McGaughey M. Cytomegalovirus esophagitis. Br Med J 1972;2:294.
Whitley RJ, Soong SJ, Dolin R, et al. Adenine arabinoside therapy of biopsy-proved herpes
simplex encephalitis. National Institute of Allergy and Infectious Diseases collaborative
antiviral study. N Engl J Med 1977;297:28994.
Wilcox CM, Straub RF, Schwartz DA. Prospective endoscopic characterization of
cytomegalovirus esophagitis in AIDS. Gastrointest Endosc 1994;40:4814.
Bonacini M, Young T, Laine L. The causes of esophageal symptoms in human
immunodeficiency virus infection. A prospective study of 110 patients. Arch Intern Med
1991;151:156772.
Theise ND, Rotterdam H, Dieterich D. Cytomegalovirus esophagitis in AIDS: Diagnosis by
endoscopic biopsy. Am J Gastroenterol 1991;86:11236.
Goodgame RW, Genta RM, Estrada R, et al. Frequency of positive tests for cytomegalovirus
in AIDS patients: Endoscopic lesions compared with normal mucosa. Am J Gastroenterol
1993;88:33843.
Nelson MR, Connolly GM, Hawkins DA, Gazzard BG. Foscarnet in the treatment of
cytomegalovirus infection of the esophagus and colon in patients with acquired immune
deficiency syndrome. Am J Gastroenterol 1991;86:87681.
Laine L, Bonacini M, Sattler F, et al. Cytomegalovirus and Candida esophagitis in patients
with AIDS. J Acquir Immune Defic Syndr 1992;5:6059.
Rabenech L, Popovic M, Gartner S, et al. Acute HIV infection presenting with painful
swallowing and esophageal ulcers. JAMA 1990;263:231822.
Levine MS, Loercher G, Katzka DA, et al. Giant, human immunodeficiency virus-related ulcers
in the esophagus. Radiology 1991;180:3236.
Sohol-Anderson ML, Prelutsky DJ, Westblom TU. Giant aphthous ulcers in AIDS patients:
Treatment with low-dose corticosteroids. AIDS 1991;5:15378.
Wilcox CM, Schwartz DA. Endoscopic characterization of idiopathic esophageal ulceration
associated with human immunodeficiency virus infection. J Clin Gastroenterol 1993;16:2516.
Schechter M, Pannain VL, De Oliveira AV. Papovavirus-associated esophageal ulceration in a
patient with AIDS. AIDS 1991;5:238.
Kimmel ME, Boylan JJ. Fistulous degeneration of a giant esophageal ulcer in a patient with
acquired immune deficiency syndrome. Am J Gastroenterol 1991;86:898900.
Edwards P, Turner J, Gold J, Cooper DA. Esophageal ulceration induced by zidovudine. Ann
Intern Med 1990;112:656.
Palmer ED (ed). The Esophagus and Its Diseases. New York: Hoeber, 1952;242.
Walsh TJ, Belitsos NJ, Hamilton SR. Bacterial esophagitis in immunocompromised patients.
Givler RL. Esophageal lesions in leukemia and lymphoma. Am J Dig Dis 1970;15:316.
McManus JPA, Webb JN. A yeast-like infection of the esophagus caused by Lactobacillus
acidophilus. Gastroenterology. 68:58361975.

51. Howlett SA. Acute streptococcal esophagitis. Gastrointest Endosc 1979;25:1501.
52.
Ramakantan R, Shah P. Dysphagia due to mediastinal fibrosis in advanced pulmonary
tuberculosis. AJR 1990;154:613.
53. Kaufman JP. Odynophagia/dysphagia. Gastroenterol Clin North Am 1988;17:599614.
54. Newman RM, Fleshner PR, Lajam FE, Kim U. Esophageal tuberculosis: A rare presentation
with hematemesis. Am J Gastroenterol 1991;86:7515.
55. De Silva R, Stoopach PM, Raufman JP. Esophageal fistulas associated with mycobacterial
infection in patients at risk for AIDS. Radiology 1990;175:44953.
56. Mokoena T, Shama DM, Ngakane H, Bryer JV. Oesophageal tuberculosis: A review of eleven
cases. Postgrad Med J 1992;68:1105.
57. Gupta SP, Arora A, Bhargava DK. An unusual presentation of oesophageal tuberculosis.
Tuber Lung Dis 1992;73:1746.
58. Goodman P, Pinero SS, Rance RM, et al. Mycobacterial esophagitis in AIDS. Gastrointest
Radiol 1989;14:1035.
59. Laajam MA. Primary tuberculosis of the esophagus: Pseudotumoral presentation. Am J
60. Seivewright N, Feehally J, Wicks AC. Primary tuberculosis of the esophagus. Am J
61. Gordon AH, Marshall JB. Esophageal tuberculosis: Definitive diagnosis by endoscopy. Am J
62. Eng J, Sabanathan S. Tuberculosis of the esophagus. Dig Dis Sci 1991;36:53640.
63. Forsmark CE, Wilcox CM, Darragh TM, Cello JP. Disseminated histoplasmosis in AIDS: An
unusual case of esophageal involvement and gastrointestinal bleeding. Gastrointest Endosc
1990;36:6045.
64. Leape LL. Liquid lyeStill a hazard. N Engl J Med 1971;284:14434.
65.
Salzer H. Fruhbehandlung der Speiserohrenveratzung. Wien Klin Wochenschr 1920;33:307.
66.
Palmer ED. Esophagitis due to corrosive agents. In Palmer ED (ed). The Esophagus and Its
Diseases. New York: Hoeber, 1952;288.
67. Webb WR, Koutras P, Ecker RR, Sugg WL. An evaluation of steroids and antibiotics in
caustic burns of the esophagus. Ann Thorac Surg 1970;9:95102.
68. Haller JA Jr, Andrews HG, White JJ, et al. Pathophysiology and management of acute
corrosive burns of the esophagus: Results of treatment in 285 children. J Pediatr Surg
1971;6:57884.
69. Reyes HM, Lin CY, Schlunk FF, et al. Experimental treatment of corrosive esophageal burns.
J Pediatr Surg 1974;9:31727.
70. Tucker JA, Yarington CT Jr. The treatment of caustic ingestion. Otolaryngol Clin North Am
1979;12:34350.
71. Hill JL, Norberg HP, Smith MD, et al. Clinical technique and success of the esophageal stent
to prevent corrosive strictures. J Pediatr Surg 1976;11:44350.
72. Broor SL, Kumar A, Chari ST, et al. Corrosive oesophageal strictures following acid ingestion:
Clinical profile and results of endoscopic dilatation. J Gastroenterol Hepatol 1989;4:5561.
73. Gorman RL, Khin-Maung-Gyi MT, Klein-Schwartz W, et al. Initial symptoms as predictors of
esophageal injury in alkaline corrosive ingestions. Am J Emerg Med 1992;10:18994.
74. Symbas PN, Vlasis SE, Hatcher CR Jr. Esophagitis secondary to ingestion of caustic
material. Ann Thorac Surg 1983;36:737.
75. Zargar SA, Kochhar R, Nagi B, et al. Ingestion of strong corrosive alkalis: Spectrum of injury
to upper gastrointestinal tract and natural history. Am J Gastroenterol 1992;87:33741.
76. Stenson K, Gruber B. Ingestion of caustic cosmetic products. Otolaryngol Head Neck Surg
1993;109:8215.
77. Franken EA Jr. Caustic damage of the gastrointestinal tract: Roentgen features. Am J
Roentgenol Radium Ther Nucl Med 1973;118:7785.
78.
Muhletaler CA, Gerlock AJ Jr, deSoto L, et al. Acid corrosive esophagitis: Radiographic
findings. AJR 1980;134:113740.

79. Anderson KD, Rouse TM, Randolph JG. A controlled trial of corticosteroids in children with
corrosive injury of the esophagus. N Engl J Med 1990;323:63740.
80.
Weisskopf A. Effects of cortisone on experimental lye burn of the esophagus. Ann Otol Rhinol
Laryngol 1952;61:68191.
81.
Rosenberg N, Kunderman PJ, Vroman L, Moolten SE. Prevention of experimental esophageal
stricture by cortisone. II. Control of suppurative complications of penicillin. Arch Surg
1953;66:5938.
82. Zargor SA, Kochhar R, Mehta S, Mehta SK. The role of fiberoptic endoscopy in the
management of corrosive ingestion and modified endoscopic classification of burns.
83.
Boyce HW Jr, Palmer ED. Techniques of Clinical Gastroenterology. Springfield, IL: Charles C
Thomas, 1975;23751.
85.
Kiviranta UK. Corrosive carcinoma of the esophagus: 381 cases of corrosion and 9 cases of
corrosion carcinoma. Acta Otolaryngol 1952;42:8995.
86. Appelqvist P, Salmo M. Lye corrosion carcinoma of the esophagus: A review of 63 cases.
Cancer 1980;45:26558.
87. Broor SL, Raju GS, Bose PP, et al. Long term results of endoscopic dilatation for corrosive
oesophageal strictures. Gut 1993;34:1498501.
88. Pemberton J. Oesophageal obstruction and ulceration caused by oral potassium therapy. Br
Heart J 1970;32:2678.
89. Kikendall JW, Friedman AC, Oyewole MA, et al. Pill-induced esophageal injury. Case reports
and review of the medical literature. Dig Dis Sci 1983;28:17482.
90. Bott S, Prakash C, McCallum RW. Medication-induced esophageal injury: Survey of the
literature. Am J Gastroenterol 1987;82:75863.
91.
Bonavina L, DeMeester TR, McChesney L, et al. Drug-induced esophageal strictures. Am
Surg 1987;206:17383.
92. McCord GS, Clouse RE. Pill-induced esophageal strictures: Clinical features and risk factors
for development. Am J Med 1990;88:5128.
93.
Boyce HW. Drug-induced esophageal damage. In Tytgat GNJ, van Blankenstein M (eds).
Current Topics in Gastroenterology and Hepatology. Stuttgart, New York: Georg Thieme,
1990;17095.
94. Evans KT, Roberts GM. Where do all the tablets go?. Lancet 1976;2:12379.
95.
Applegate GR, Malmud LS, Rock E, et al. "It's a hard pill to swallow" or "Don't take it lying
down" [Abstract]. Gastroenterology 1980;78:1132.
96. Collins FJ, Matthews HR, Baker SE, Stakova JM. Pill-induced oesophageal injury. Br Med J
1979;1:16736.
97. O'Meara TF. A new endoscopic finding of tetracycline-induced esophageal ulcers.
98. Mason SJ, O'Meara TF. Drug-induced esophagitis. J Clin Gastroenterol 1981;3:11520.
99. Carlborg B, Densert O. Esophageal lesions caused by orally administered drugs. An
experimental study in the cat. Eur Surg Res 1980;12:27082.
100. Walta DC, Giddens JD, Johnson LF, et al. Localized proximal esophagitis secondary to
ascorbic acid ingestion and esophageal motor disorder. Gastroenterology 1976;70:7669.
101. Amendola MA, Spera TD. Doxycycline-induced esophagitis. JAMA 1985;253:100911.
102. Kato S, Kobayashi M, Sato H, et al. Doxycycline-induced hemorrhagic esophagitis: A pediatric
case. J Pediatr Gastroenterol Nutr 1988;7:7625.
103. Ovartlarnporn B, Kulwichit W, Hiranniramol S. Medication-induced esophageal injury: Report
of 17 cases with endoscopic documentation. Am J Gastroenterol 1991;86:74850.
104. Puhakka HJ. Esophageal ulceration due to oral medications. ORL J Otorhinolaryngol Relat
Spec 1987;49:3216.
105. Winckler K. Tetracycline ulcers of the oesophagus: Endoscopy, histology and roentgenology
in two cases, and review of the literature. Endoscopy 1981;13:2258.
106. Rodino S, Sacca N, De Medici A, Giglio A. Multiple esophageal ulcerations caused by a
granular formulation of aspirin. Endoscopy 1994;26:50910.
107. Barrison IG, Trewby PN, Kane SP. Oesophageal ulceration due to emepronium bromide.
Endoscopy 1980;12:1979.
108. Hillman LC, Scobie BA, Pomare EW, Austad WI. Acute oesophagitis due to emepronium
bromide. N Z Med J 1981;93:46.
109. Chowan NM. Injurious effects of radiation on the esophagus. Am J Gastroenterol
1990;85:11520.
110. Mascarenhas F, Silvestri ME, Sa da Costa M, et al. Acute secondary effects in the esophagus
in patients undergoing radiotherapy for carcinoma of the lung. Am J Clin Oncol
1989;12:3440.
111.
Seaman WB, Ackerrnan LV. The effect of radiation on the esophagus: A clinical and
histologic study of the effects produced by the betatron. Radiology 1957;68:53441.
112. Goldstein HM, Rogers LP, Fletcher GH, Dodd GD. Radiological manifestations of
radiation-induced injury to the normal upper gastrointestinal tract. Radiology
1975;117:13540.
113. Phillips TL, Fu KK. Quantification of combined radiation therapy and chemotherapy effects on
critical normal tissues. Cancer 1976;37:1186200.
114. Swaroop VS, Desai DC, Mohandas KM, et al. Dilation of esophageal strictures induced by
radiation therapy for cancer of the esophagus. Gastrointest Endosc 1994;40:3115.
115. Nelson RS, Hernandez AJ, Goldstein HM, Saca A. Treatment of irradiation esophagitis. Am J
116. Dobbins JW, Sheahan DG, Behar J. Eosinophilic gastroenteritis with esophageal involvement.
117. Landres RT, Kuster GGR, Strum WB. Eosinophilic esophagitis in a patient with vigorous
118.
Picus D, Frank PH. Eosinophilic esophagitis. AJR 1981;136:10013.
119.
Goldberg HI, O'Kieffe D, Jenis EH, Boyce HW Jr. Diffuse eosin-ophilic gastroenteritis. AJR
1973;119:34251.
120. Anti M, Marra G, Rapaccini GL, et al. Esophageal involvement in Behet's syndrome. J Clin
121. Yashiro K, Nagasako K, Hasegawa K, et al. Esophageal lesions in intestinal Behet's disease.
Endoscopy 1986;18:5760.
122. Bottomley WW, Dakkak M, Walton S, Bennett JR. Esophageal involvement in Behet's
disease. Is endoscopy necessary?. Dig Dis Sci 1992;37:5947.
123. Mori S, Yoshihira A, Kawamura H, et al. Esophageal involvement in Behet's disease. Am J
124. Eliakim R, Goldin E, Livshin R, Okon E. Esophageal involvement in pemphigus vulgaris. Am J
125. Wood DR, Patterson JB, Orlando RC. Pemphigus vulgaris of the esophagus. Ann Intern Med
1982;96:18991.
126. Barnes LM, Clark ML, Estes SA, Bongiovanni GL. Pemphigus vulgaris involving the
esophagus. A case report and review of the literature. Dig Dis Sci 1987;32:6559.
127. Goldberg NS, Weiss SS. Pemphigus vulgaris of the esophagus in women. J Am Acad
Dermatol 1989;21:11158.
128. Goldin E, Lijovetzky G. Esophageal involvement by pemphigus vulgaris. Am J Gastroenterol
1985;80:82830.
129. Mobacken H, Al Karawi M, Mohamed A, Coode P. Oesophageal pemphigus vulgaris.
Dermatologica 1988;176:2669.
130. Sharon P, Greene ML, Rachmilewitz D. Esophageal involvement in bullous pemphigoid.
131.
Stevens AE, Dove GAW. Oesophageal cast: Oesophagitis dissecans superficialis. Lancet
1960;2:1279.
132. Witte JT, Icken JN, Lloyd ML. Induction of esophageal bullae by endoscopy in benign mucous
membrane pemphigoid. Gastrointest Endosc 1989;35:5668.
133. Isolauri J, Airo I. Benign mucous membrane pemphigoid involving the esophagus: A report of
two cases treated with dilation. Gastrointest Endosc 1989;35:56971.
134. Manier JW, Kaplan AP. Polydysplastic epidermolysis bullosa with esophageal stricture.
135. Ergun GA, Kin AN, Dannenberg AJ, Carter DM. Gastrointestinal manifestations of
epidermolysis bullosa. A study of 101 patients. Medicine 1992;71:1217.
136.
Nix TE Jr, Christianson HB. Epidermolysis bullosa of the esophagus: Report of two cases and
review of the literature. South Med J 1965;58:61220.
137. Katz J, Gryboski JD, Rosenbaum HM, Spiro HM. Dysphagia in children with epidermolysis
bullosa. Gastroenterology 1967;52:25962.
138. Warren RB, Warner TF, Gilbert EF, Pellet JR. Acquired double-barrel oesophagus in
epidermolysis bullosa dystrophica. Thorax 1980;35:4726.
139. Travis SPL, Eady RAJ, Thompson RPH. Esophageal complications in epidermolysis bullosa.
Gut 1991;32:156970.
140. Absolon KB, Finney LA, Waddill GM Jr, Hatchett C. Esophageal reconstructioncolon
transplantin two brothers with epidermolysis bullosa. Surgery 1969;65:8326.
141. Fernandez CM, Ledo L, Cunqueiro R, et al. Early esophageal Crohn's disease: An endoscopic
view. Gastrointest Endosc 1990;36:3034.
142. Kuboi H, Yashiro K, Shindou H, et al. Crohn's disease in the esophagusReport of a case.
Endoscopy 1988;20:11821.
143. Klein GL, Ament ME, Sparkes RS. Monozygotic twins with Crohn's disease: A case report.
144. Gelfand MD, Krone CL. Dysphagia and esophageal ulceration in Crohn's disease.
145. Geboes K, Janssens J, Rutgeerts P, Vantrappen G. Crohn's disease of the esophagus. J Clin
146. Freson M, Kottler RE, Wright JP. Crohn's disease of the oesophagus. A case report. S Afr
Med J 1984;66:4178.
147. LiVolsi VA, Jaretzki A III. Granulomatous esophagitis. A case of Crohn's disease limited to the
esophagus. Gastroenterology 1973;64:3139.
148. D'Haens G, Rutgeerts P, Geboes K, Vantrappen G. The natural history of esophageal Crohn's
disease: Three patterns of evolution. Gastrointest Endosc 1994;40:296300.
149. Huchzermeyer H, Paul F, Seifert E, et al. Endoscopic results in five patients with Crohn's
disease of the esophagus. Endoscopy 1977;8:7581.
150. Mathis G, Sutterlutti G, Dirschmid K, et al. Crohn's disease of the esophagus: Dilation of
stricture and fibrin sealing of fistulas. Endoscopy 1994;26:508.
151. Yanaga K, Iwamatsu M, Yukizane T, Shimizu T. Esophageal stricture in Crohn's diseaseA
case report. Jpn J Surg 1986;16:6872.
152. Bonavina L, Bettineschi F, Cecchetto A, Pomerri F. Crohn's disease of the esophagus.
Dysphagia 1989;4:1769.
153. Kimball MW. Pseudomelanosis of the esophagus. Gastrointest Endosc 1978;24:1212.
154. Dumas O, Barthelemy C, Billard F, et al. Isolated melanosis of the esophagus: Systematic
endoscopic diagnosis. Endoscopy 1990;22:945.
155. Sharma SS, Venkateswaran S, Chacko A, Mathan M. Melanosis of the esophagus. An
endoscopic, histochemical, and ultrastructural study. Gastroenterology 1991;100:136.
156.
Yamazaki K, Ohmori T, Kumagai Y, et al. Ultrastructure of oesophageal melanocytosis.
Virchows Arch A Pathol Anat Histol 1991;418:51522.
157. Piccone VA, Klopstock R, LeVeen HH, Sika J. Primary malignant melanoma of the
esophagus associated with melanosis of the entire esophagus. First case report. J Thorac
Cardiovasc Surg 1970;59:86470.

158. Takubo K, Kanda Y, Ishii M, et al. Primary malignant melanoma of the esophagus. Hum
Pathol 1983;14:72730.
Chapter 46 Peptic Diseases of the Stomach and Duodenum

(3828)
(3829)
DAVID Y. GRAHAM, M.D.
The simplicity, convenience, and safety of modern endoscopy has resulted in major advances in
diagnosis, treatment, and knowledge of the prevalence and natural history of peptic diseases. The
indications and uses for endoscopy continue to evolve and its optimum, cost-effective role for
diagnosis or management of the dyspeptic patient remains unsettled. Primarily, endoscopy is used as
a diagnostic test. Because diagnostic testing accounts for a substantial portion of health care
expenditures,1,2(3830) it is important to ask: "What is the value of endoscopy as a test?"
Currently, the value of a diagnostic test is often viewed in a global context and equated with cost
effectiveness. The physician often considers value in terms of accuracy, sensitivity, specificity, and
positive and negative predictive values. Because most tests are performed on patients who do not
have the disease that is the object of the test, reported figures for accuracy may be misleading.
Accuracy is often reported as the percentage of correct diagnoses. If one evaluated 100 patients with
gastric ulcer, 5 of which were malignant, and 4 of the cancers were detected, the accuracy may be
reported as 99% (99 of 100 correct diagnoses). In reality, 20% of the cancers were overlooked. Better
measures of a test are its sensitivity for a given disease (i.e., the number of cases detected divided by
the total number of cases present in the population tested) and specificity. Physicians are more
interested in the positive predictive value and the negative predictive value, measures of how likely the
result is either correct or wrong. Video endoscopy for detection of peptic diseases of the upper
gastrointestinal tract meets modern criteria for high sensitivity and specificity as well as high positive
and negative predictive values.
Clinical usefulness is a subjective concept that measures the translation of a test result into a change
in the patient's well-being. A finding may be clinically important (specifies or explains symptoms, is
diagnostic, justifies therapy, or indicates further evaluation) or trivial. A negative result may have
clinical utility by redirecting the evaluation and reassuring the patient. With functional disorders, the
principal value of diagnostic evaluation is the exclusion of disease or the reassurance of the patient, or
both. One consideration with respect to clinical utility, therefore, is the ability to exclude disease when
there is a possibility that some disorder is present.
Evaluation of the Dyspeptic Patient

For years, single-contrast barium radiography was the standard method for evaluation of dyspeptic
patients. Since the mid-1980s, the trend has been to replace this radiologic study with endoscopy. The
modern, recommended radiologic approach to the evaluation of the upper gastrointestinal tract is a
biphasic study (i.e., double contrast followed by single contrast during the same session) using
state-of-the-art equipment. However, radiology is still inferior to endoscopy because the latter allows
detailed examination of the esophagus, stomach, and duodenum and can thus certify the absence of
disease and identify superficial lesions (e.g., small erosions in the distal esophagus and the small
gastric or duodenal ulcer that is poorly defined or overlooked radiographically).
Many authors regard endoscopy and the upper gastrointestinal x-ray as complementary tests; I cannot
support that view. The x-ray series may provide information not obtainable by
endoscopyobservations pertaining to motility, for examplebut it is best described as a safe albeit
inaccurate test. In comparison with endoscopy, approximately 20% of patients with a variety of
gastrointestinal complaints will be misdiagnosed by upper gastrointestinal radiography. The reported
false-negative rate is as high as 32%, and the false-positive rate ranges from 8 to 21%.38(3831) Note
that half of radiologic "missed" gastric ulcers or cancers are clearly visible during retrospective
evaluation of the x-ray films;9(3832) the abnormalities were overlooked in the context of the "routine
examination." Double-contrast studies tend to have fewer false-negative results, but this is balanced by
an increase in false-positive tests. These figures apply not only to superficial inflammatory disease but
also to advanced disease, including gastric ulcer and advanced gastric cancer.
Three prospective studies comparing diagnostic radi-ology and endoscopy have been
published.8,10,11(3833) Two used modern biphasic radiology techniques and confirmed that
endoscopy has definite advantages over state-of-the-art radiography. Despite the fact that radiographic
evaluation of the upper gastrointestinal tract is inherently inaccurate, expensive, time-consuming, and
overutilized,12,13(3834) it remains the second most frequently ordered diagnostic x-ray study in the
United States.12(3835) The reason for the majority of these studies is unclear, especially since the
examination may not effect a change in therapy, regardless of the findings.14(3836) Marton et
al.14(3837) documented that many upper gastrointestinal series are destined to be normal because
most of the patients undergoing the procedure do not have symptoms, signs, or other indications of
upper gastrointestinal disease. Presumably, a major reason for requesting the procedure is the
reassurance provided to patient or physician, or both.
Stevenson6(3838) suggested that every hospital should internally audit the performances of its
radiology and endoscopy services, and if a significant false-negative rate is demonstrated for the upper
gastrointestinal x-ray series, it should be abandoned. This would likely be the result in most hospitals,
possibly because radiologists seem to have lost interest in performing or teaching barium contrast
studies in conjunction with the growing availability of newer, more sophisticated tests such as
computed tomography. There are still a few truly superb gastrointestinal radiologists who achieve
outstanding results. Most physicians do not have access to such individuals but fortunately do have
access to highly skilled endoscopists.
Endoscopy, more sensitive than standard or double-contrast radiography, is now the diagnostic test of
choice in patients with symptoms that may indicate upper gastrointestinal disease. Op den
Orth15(3839) summarized the current status of barium examination of the upper gastrointestinal tract
as: "The recommended diagnostic pathway will be beneficial only if the upper GI [gastrointestinal]
examinations are performed by skilled radiologists with a high degree of expertise who are willing to
devote a relatively large part of their time to upper GI examinations and to communicate with
endoscopists. In any other circumstance, primary endoscopywith its inherent disadvantageswill
often be necessary." I agree. If the provisional diagnosis is functional disease, the primary goal of a
diagnostic test is reassurance. Any reassurance derived by physicians from upper gastrointestinal
radiography should be negligible, given the magnitudes of false-negative and false-positive results.
Although patients derive assurance that they do not have a serious disease, the upper gastrointestinal
series has little or no effect on the management of disease,14,16(3840) suggesting either that
physicians will embark on a course of therapy irrespective of test results or, more likely, that patients
believed to be suffering from a serious disease are not referred for upper gastrointestinal x-rays.
Endoscopy is superior to upper gastrointestinal x-rays, but can it be recommended as the sole
diagnostic technique for evaluation of patients with complaints referable to the upper gastrointestinal
tract? Potential problems with this approach are the low but definite complication rate, the cost in terms
of time, human resources, and expense, and the potential for overutilization. Currently, the latter does
not appear to be a problem. For example, Beavis et al.17(3841) reported results from a one-visit
endoscopy clinic for evaluation of patients with dyspepsia. Patients were referred by their practicing
physicians and were examined by a gastroenterologist who determined whether endoscopy was
necessary (98%). Endoscopy was successfully performed in 186 patients (93%); there were abnormal
findings in 70%. The frequency of significant peptic disease was over 50% (26.7% pyloric channel
ulcer, duodenal ulcer, or scar; 17.7% esophagitis; 5.9% gastric ulcer; and 1% gastric cancer).
However, as more patients with functional complaints are referred for endoscopy, the percentage of
false diagnoses may well increase. If the endoscopist were forced to examine a high volume of
patients without organic disease, as is the case for the radiologist, the yield of true findings by
endoscopy might also diminish.
The question remains whether endoscopy, as the only available diagnostic test, will be overused.
Although x-ray procedures are expensive, the cost pales in comparison with that of endoscopy. It is
estimated that dyspeptic symptoms will develop in approximately 1% of a given population in the
course of a year.18,19(3842) There remains the principal issue: how best to evaluate these patients.
The best current approach is to choose the most sensitive and accurate test, currently video
endoscopy, for those cases where signs and symptoms point to actual disease in the upper
gastrointestinal tract.20(3843) This may change as we integrate new knowledge concerning
Helicobacter pylori into our practices (see later).21(3844)
Endoscopy in Peptic Disease of the Stomach and Duodenum

Although the pathogenesis of peptic ulcer disease is not yet fully understood, the recent recognition
that there are several categories of causation has changed our concepts of pathogenesis as well as
our approach to the diagnosis and management of patients. The three most common causes are (1)
infection with the bacterium H. pylori, (2) use of nonsteroidal anti-inflammatory drugs (NSAIDs), and (3)
the presence of a pathologic acid hypersecretory state such as the Zollinger-Ellison
syndrome.22(3845) Patients undergoing hepatic artery infusion chemotherapy may develop gastric and
duodenal ulcers; serious complications such as perforation may also occur.23,24(3846)
In a broad sense, ulcers are still thought to be caused by an imbalance between aggressive and
defensive factors. The former may be endogenous (acid and proteolytic digestion by pepsin) or
exogenous (corrosive drugs or H. pylori infection), or both. The recognized defensive factors include
gastric mucus, bicarbonate secreted by cells of the gastric mucosal surface, epithelial cell
regeneration, and gastric mucosal blood flow. The fact that gastric mucosa is not digested by acid and
pepsin has fascinated and intrigued scientists since the 17th century, but the mechanism underlying
this phenomenon is still not understood. Both acid and pepsin are required for rapid digestion to occur,
and pepsin is always found whenever gastric acid is present. Early investigators were impressed with
the concept of a "living principle," as it was discovered that living tissue is digested less rapidly than
damaged or dead tissue. Subsequently, it was found that living tissue could be digested.25(3847)
Gastric mucosa is digested whenever it is damaged. For example, topical application of concentrated
ethanol or intramucosal injection of silver nitrate is followed by digestion of the damaged mucosa.
The sequence of events leading to acute ulcer formation is thought to be mucosal hemorrhage,
followed by erosion, an increase in the depth of injury, and formation of an acute ulcer. The transition
from acute to chronic ulcer is not understood and must be extremely rare. The natural tendency of
erosions and acute ulcers is to heal. This has been documented by experimental studies in which
acute ulcers were produced by cutting or injecting corrosive chemicals into the submucosa; these
lesions healed rapidly.25(3848) A chronic ulcer represents failed or abnormal wound healing.
Definition of Peptic Disease

Peptic ulcer is a generic term that signifies a break in any part of the gastrointestinal mucosa exposed
to acid and pepsin. The name implies that the defect is caused by digestion of the gastric or duodenal
mucosa by acid and pepsin. An ulcer can also occur in an achlorhydric stomach, suggesting that
another mechanism must be operative in such cases. Chronic ulcers in the presence of achlorhydria
are usually the result of drug-induced injury. Gastric and duodenal mucosal disruptions are defined
histologically as erosions, acute ulcers, or chronic ulcers.
Petechiae
Gastric petechiae (mucosal hemorrhages) are common and most often disappear without
incident.26,27(3849) Petechiae are thought to be precursors of erosions and ulcers, but they are also
frequently observed in normal individuals who experience no known sequelae. Petechiae range in size
from pinpoint to several millimeters in diameter and appear to be located in the mucosa (Figures 461
and 462). The intactness of the mucosa overlying petechiae can be ascertained from the appearance
of mucosal highlights. Showers of petechiae, especially in the fundus, may occur after retching. The
endoscopic terminology used to describe petechiae is confusing and overlaps with erosions. For
example, mucosal hemorrhages associated with NSAID ingestion often appear black or brown,
suggesting that gastric acid has had contact with the blood. It is not known whether this is a result of an
actual small breach in the mucosa (an erosion) or whether the permeability to acid in the cells overlying
the hemorrhage is increased. This distinction is largely of scholarly interest because such lesions heal
and disappear rapidly regardless of whether they are red or black.
(3850)Figure 461. Tiny gastric petechiae on the lesser curvature of the antrum.
(3851)Figure 462. Large areas of petechiae seen in the antrum after aspirin ingestion.
Erosions
An erosion is an epithelial defect that does not penetrate the muscularis mucosae and heals
spontaneously without scar formation, usually within 2 to 8 days. The conversion of an erosion to a
chronic gastric ulcer or gastric cancer must be extraordinarily uncommon. Erosions may be situated in
either the acid- or the non-acid-secreting mucosa and commonly occur at multiple sites. Two general
types of erosions are distinguished: flat (aphthous or typical) and complete (varioliform).
Schindler28(3852) first described endoscopically observed flat erosions in the fundus of the stomach
and noted that they resembled aphthous ulcers, having a grayish base and a red halo. Endoscopically,
the fully developed erosion is characterized by a central depression with or without a necrotic floor, a
red rim, and prominent reaction in the surrounding mucosa (Figures 463 and 464). The
double-contrast radiologic appearance of a flat erosion is a small (0.5 to 1.5 cm) filling defect
surrounded by a halo.29(3853) High-density barium must be used to visualize these lesions.
Double-contrast radiologic studies using the biphasic technique are greatly inferior to endoscopy for
detection of typical erosions.
(3854)Figure 463. Aphthoid erosion on a gastric fold. Note the lack of deformity of the
mucosa, signifying that it is very superficial and does not involve the muscularis mucosae.
(3855)Figure 464. Gastric petechiae and erosions seen after approximately 24 hr of aspirin
administration (two tablets four times per day) to a normal, healthy volunteer. The picture is
that of acute injury.
Complete erosions are seen endoscopically as small nodules with a central necrotic base. They are
usually located on gastric folds, especially along the greater curvature of the antrum and distal body
(Figure 465). Radiographically, they appear as filling defects on gastric folds with a central collection
of barium. Both endoscopic and radiologic approaches are effective in identification of complete
erosions. Flat erosions heal rapidly and typically reflect acute injury, whereas complete erosions may
remain essentially unchanged for years.30,31(3856) However, in our experience, complete erosions
associated with H. pylori infection often heal and disappear with eradication of the bacterium.
(3857)Figure 465. Complete or varioliform erosions along the greater curvature of the
antrum. Complete erosions may remain for many years without any evident change. If
associated with Helicobacter pylori infection, they resolve, in our experience, when the
infection is successfully treated.
Erosions are seen in 2 to 7% of routine autopsies, most commonly in the fundus, and are reported to
occur in 10 to 80% of gastric specimens resected because of chronic gastric or duodenal
ulcer.32(3858) Gastric erosions are observed in approximately 10% of patients undergoing endoscopy,
about half of whom will have other demonstrable lesions, usually of peptic origin.33,34(3859) The
prevalence of erosion (determined as the percentage of individuals with erosion among those
undergoing endoscopy) has been reported as 3 to 4%,35,36(3860) 11%,33,37(3861) and as high as
15%.32(3862) As with frank ulceration,38,39(3863) there seems to be a seasonal incidence, with
erosions being most commonly found from October through November and May through
June.33(3864) The actual incidence of erosions among Western populations is not known because
there have been few studies of the gastric mucosa in normal healthy individuals. Gastroscopy is
performed on large segments of the general population of Japan to detect early gastric cancer;
mucosal erosions are found in 3 to 4% of individuals who undergo screening endoscopy.35(3865) In
the United States, gastroscopy is most often performed on normal individuals during clinical
investigations of the effects of drugs on the gastric mucosa. We have observed mucosal abnormalities,
most commonly mucosal hemorrhages, in approximately 10% of such procedures. Mucosal
hemorrhages (gastric petechiae) are usually located in the fundus and vary from few to many, and as
noted previously, they may also occur in a shower-like fashion immediately after retching or after the
endoscope has been advanced over the mucosal surface. Gastric erosions are encountered
occasionally, but less frequently, and are usually in the antrum or sometimes in the distal body of the
stomach. When the location of gastric and duodenal erosions was studied in 574 patients with acute or
chronic lesions, these results were obtained: In approximately 54%, erosions were confined to the
antrum; in 11%, to the fundus; in 10%, to the entire stomach; in 14%, to the duodenal bulb; and in
8.4%, to the antrum and duodenal bulb.32(3866)
Erosions are found infrequently in volunteers for studies of pharmacologic drugs who refrain from
ingestion of alcohol, spicy foods, and all medications for at least 2 weeks prior to gastroscopy.
Identification of the typical lesions of NSAID (usually aspirin) injury (Figures 466, 467, 468 and
469), a frequent finding at endoscopy, should prompt questioning of the patient about recent use of
over-the-counter medications, many of which (e.g., "cold" remedies) contain aspirin.
(3867)Figure 466. Typical picture of a trivial nonsteroidal anti-inflammatory drug

(NSAID)-induced injury to the gastric mucosa. There are multiple small erosions with
brown-black staining of the center as a result of local bleeding and petechiae.
(3868)Figure 467. Typical NSAID-induced injury to the gastric mucosa. The damage is
more severe, with a large erosion or a small acute ulcer on the lesser curvature in the
prepyloric area and a large, very superficial erosion on the greater curvature.
(3869)Figure 468. Acute superficial NSAID-induced injury to the duodenum seen after 1
day of aspirin injection (650 mg four times per day). Biopsy confirmed that the damage did
not extend to the muscularis mucosae.
(3870)Figure 469. Typical gastric erosion in the prepyloric area of the stomach.
Typical erosions are usually less than 0.5 cm in diameter; many, if not most, investigators designate
lesions whose diameter is greater than 0.5 cm as acute ulcers. Endoscopically, it is impossible to
determine whether a small lesion extends through the muscularis mucosae. The estimation of depth is
not reliable because swelling and edema at the edges of acute erosions may exaggerate the
impression of depth. It is possible to determine whether a lesion involves the deeper muscle layer
(muscularis propria) by noting its effect on the normal muscular contraction waves that move through
the antrum. If the deeper muscles are affected, the smooth peristaltic progression of these waves is
altered. Rsch32(3871) suggested that the biopsy forceps may be used to determine whether the
lesion (and the submucosa) can be pulled toward the lumen, apart from the gastric wall. This is a
useful technique that will reveal involvement of the muscularis propria but will not indicate a breach of
the muscularis mucosa. Consequently, it cannot differentiate between a shallow acute ulcer and an
erosion, but this distinction probably has little clinical significance.
It is difficult to discern whether erosions actually cause symptoms, as they are usually associated with
another disease that is responsible for the patient's discomfort. In addition, most patients undergoing
examination have some symptoms, and it is easy to incriminate gastric petechiae or erosions as
causative. The superficial mucosal lesions induced with aspirin during clinical investigations are usually
asymptomatic; conversely, when symptoms are present, visible mucosal lesions are frequently absent.
Cameron and Higgins40(3872) found that linear erosions on the summit of mucosal folds at the level of
the diaphragmatic constriction in patients with large hiatus hernias are more likely to be associated with
chronic iron deficiency anemia.
Ulcers
Definition
Ulcers are distinguished from erosions by the fact that ulcers breach the muscularis mucosa and
erosions do not. Typical erosions are shallow, circular or oval lesions, usually less than 0.5 cm in
diameter, with sharply defined edges (see Figure 463). The defect does not penetrate the muscularis
mucosa, so that glandular elements remain in the base and healing is rapid. Acute ulcers are
commonly multiple, have the same general appearance and distribution as erosions, but are larger and
penetrate through the muscularis mucosa. There is little or no fibrosis in the base of an acute ulcer,
whereas chronic ulcers have underlying fibrous tissue (Figure 4610). Chronic gastric ulcers are
usually solitary, although they are multiple in 6 to 10% of cases. Ulcers are recurrent clinically and
histologically and are usually less than 2 cm in diameter. Large (giant) ulcers are usually benign
chronic gastric lesions on the lesser curvature of the stomach. There is a report of a patient with a
refractory giant gastric ulcer that was due to mural eosinophilic gastritis.41(3873)
(3874)Figure 4610. Photomicrograph of a benign gastric ulcer. Note disruption of the

muscularis propria and intense fibrotic reaction in the ulcer base.
Ulcer Healing
If a piece of gastric mucosa is removed surgically, the mucosa rapidly regenerates to close the defect,
so that the site is not identifiable macroscopically. Regeneration can occur with remarkable rapidity, as
demonstrated by healing of the mucosal surface within hours after acute superficial injury. Much of the
available information concerning ulcer healing comes from carefully performed experimental studies in
animals.25,42,43(3875) After acute gastric mucosal injury, or resection of a section of mucosa, the
muscularis mucosa adjacent to the defect contracts and thereby invaginates the intact mucosal edge
toward the base of the defect. In small defects such as erosions, adjoining mucosal cells rapidly
migrate into the defect and cover the area. This is followed by invagination of the mucosa and
restoration of normal glandular anatomy. In defects extending through the muscularis mucosa (acute
ulcers), the ulcer base lacks a residuum of normal glandular elements, and healing is entirely
dependent on the migration of adjacent cells to cover the defect. Although the normal glandular
anatomy may be restored, the muscularis mucosa is not regenerated, and new glands lie on fibrous
tissue instead of on the muscularis mucosa. This is not evident by gross observation, and usually it can
be detected only by histologic examination. Because the endoscopist is largely restricted to visual
information, it may be difficult to accurately distinguish whether an ulcer is acute or chronic when
visualized initially.
In chronic ulceration, the reparative processes are similar but the end result is different. The slow
healing of a chronic ulcer implies either a defective or inhibited healing process or an imbalance
whereby the factors leading to ulcer formation and perpetuation are not overcome by reparative
processes. A chronic ulcer is the summation of the simultaneous processes of mucosal destruction
and regeneration. Fibroblastic proliferation, leading to a dense scar, occurs when the reparative
processes are unable to restore epithelial continuity rapidly (see Figure 4610). The defect in healing is
focal, as evidenced by the fact that the mucosa surrounding the ulcer shows rapid healing if large
mucosal biopsies are taken from the normal-appearing mucosa adjacent to a chronic ulcer.
Importantly, such biopsies include muscularis mucosa and are therefore technically acute ulcers.
The ulcer scar is initially covered by a single, fragile layer of gastric mucosal cells that is easily
removed. Later, the mucosal architecture may be largely restored, but the scar remains and is
detectable both microscopically and macroscopically. An ulcer scar, and hence the fact that an ulcer is
or was present, can be identified by simple palpation of the surface of the stomach and by endoscopic
detection of folds radiating from a central point (Figure 4611). The ulcer scar that remains after
healing can often be defined endoscopically or by double-contrast x-rays. It has been shown to change
over time as the fibrous tissue contracts and becomes less apparent.44(3876) The events of ulcer
formation and healing are similar in the duodenum. As in the stomach, the muscularis mucosa does
not regenerate and Brunner's glands, normally beneath the muscularis mucosa, become situated
directly beneath the villous architecture after ulcer healing. However, this is not definite evidence of a
previous ulcer, since the mucosa can lie directly over Brunner's glands in some normal
individuals.45(3877)
(3878)Figure 4611. Large scar at the site of a healed gastric ulcer. The stomach is deformed
such that the radiating folds are likely to be demonstrated on a barium study.
Gastric Ulcer
Location of Gastric Ulcer
Endoscopy is superior to x-ray in the identification of gastric and duodenal ulcers, but it is not
perfect.8,46(3879) Some ulcers may not be seen, most often because of poor endoscopic technique.
An ulcer should be suspected when there is gastric irritability, particularly as evidenced by general or
localized muscle spasm. The majority of benign ulcers are located on the lesser curvature of the
stomach near the angularis, an area that requires rigorous evaluation (Figures 4612 and
4613).25,47,48(3880) An archlike deformity of the angularis (Henning's sign), the result of contraction
of a healing ulcer, should arouse suspicion and occasion careful inspection.
(3881)Figure 4612. Location of benign gastric ulcers in relationship to the distance from the
pylorus. The majority of benign ulcers will be found on the lesser curvature within 3 cm of the
angulus. (Adapted from Ivy AC, Grossman MI, Bachrach WH. Peptic Ulcer. Philadelphia:
Blakiston Company, 1950.)
(3882)Figure 4613. Gastric ulcer on the lesser curvature is shown "peeking out" from behind
the insertion tube of the endoscope. Ulcers in this location will frequently be missed when the
endoscopist performs a single retroflexion in the antrum and then withdraws the instrument
into the fundus.
There are several approaches to ensuring complete visualization of the gastric mucosa. One method is
to take advantage of gastric peristalsis to enhance examination of the lesser curvature of the antrum
distal to the angularis. As peristaltic contractions move toward the pylorus, successive segments of the
mucosal surface can be observed on the contraction wave. This offers an en face albeit brief and
constantly changing view of this part of the stomach, which is normally seen at a tangent with
forward-viewing instruments. The second approach is to administer glucagon intravenously to
temporarily paralyze the stomach, which will allow for examination in detail. These methods are
complementary.
Because most gastric ulcers are on the lesser curvature, the proximal and distal sides of the angularis
must be examined carefully. One common error is retroflexion of the endoscope in the antrum followed
immediately by withdrawal to examine the body, fundus, cardia, and gastroesophageal area. We
recommend that two retroflexions be performed: one in the antrum to examine the angularis, the distal
lesser curvature, and a limited portion of the body. This is followed by un-retroflexion and examination
of the antrum and body directly as the endoscope is withdrawn; finally, a second retroflexion maneuver
is performed in the midbody to examine the fundus, cardia, and gastroesophageal area. This results in
excellent and complementary views of the antrum and body and discloses lesions that hide behind the
insertion tube of the endoscope during the retroflexion maneuver (see Figure 4613).
On occasion, an ulcer is diagnosed when none is pres-ent. These phantom ulcers are caused by local
muscular contractions that produce partially hidden or closed-over pockets. To avoid this error, it is
important to wash such suspected lesions to remove trapped mucus; when in doubt, glucagon should
be administered (0.5 to 1 mg intravenously) to relax the spasm and allow visualization. If the view is
still impaired, the mucosa within the fold can be grasped with a biopsy forceps and inverted for
inspection.
Mucosal Histology
Histologic examination of the gastric mucosa is one extremely useful method of separating ulcers as to
possible causative associations. It is our opinion that whenever a gastric or duodenal ulcer is
discovered by endoscopy, mucosal biopsy of normal-appearing antral mucosa should be performed.
Because H. pylori infection is common4953(3883) and, by definition, is present in every case with an
H. pylori ulcer, the goal is not to identify the presence of an H. pylori infection but rather to exclude it.
Histologically normal mucosa (preferably two separate biopsies) virtually excludes H. pylori infection
and makes NSAID use (overt or covert), Zollinger-Ellison syndrome, or a rarer cause of ulcer the most
likely possibilities (see "Mucosal Biopsies for H. pylori Detection," later in this chapter).
H. Pylori Ulcers
Oi et al.48(3884) found that benign gastric ulcers occur in mucosa that has the histologic features of
the pyloric gland-type mucosa. However, these histologic characteristics were found proximal to the
normal junction between the pyloric and the fundic (acid-secreting) types of mucosa. The proximal
anatomic extent of the pyloric mucosa was found to be variable. When ulcers were found in the
proximal stomach, there was an associated cephalad extension of the pyloric gland mucosa,
sometimes almost to the cardia. This hypothesis was consistent with previous data regarding ulcer
location and explained the presence of ulcers in the most proximal aspect of the lesser curvature of the
stomach. It is now recognized that the mucosa surrounding an ulcer is fundic mucosa altered by
gastritis and not simply displaced pyloric gland mucosa. Endoscopic studies, using the Congo red dye
test54(3885) or indigo carmine,55(3886) are being carried out to investigate the nature of the gastric
mucosa in which ulcers occur. With the Congo red test, mucosal areas that secrete acid (fundic
mucosa) turn blue-black with application of the dye, whereas non-acid-producing mucosa (pyloric
gland) does not change color (see Chapter 12: Chromoscopy). Tatsuta and Okuda54(3887) found a
close correlation between the extent of gastritis and the location of gastric ulcers; that is, fundal
gastritis was extensive when the ulcer was located in the proximal stomach and less when the ulcer
was located in the distal stomach. To determine the true ancestry of the non-acid-secreting mucosa,
these authors measured mucosal gastrin content and found that the gastrin content of the pyloric
mucosa remained high even in severe antral gastritis (2096 ng/g vs. 7557 ng/g in normal mucosa).
Both normal fundic mucosa and mucosa adjacent to gastric ulcers occurring in the proximal stomach
contained almost no gastrin (5.1 + 4 ng/g) showing that gastric ulcers were located in fundal gastritis
instead of antral (pyloric gland-type) mucosa.
We now recognize that progressive gastritis is a feature of H. pylori infection and that eradication of the
infection results in improvement in the histology and clinical cure of the ulcer
disease.21,22,5659(3888) The recognition that H. pylori infection may cause ulcers does not explain
why ulcers are focal defects; the pathophysiologic reason for this is unknown. We now know that large
areas of gastric mucosa are abnormal, which may help explain the previous observations that
recurrent ulcers (relapses) frequently develop apart from, and often proximal to, the site of the original
ulcer.
Nsaid Ulcers
Large benign gastric ulcers in histologically normal mucosa are also common in countries where
NSAIDs such as aspirin are widely used.60(3889) One frequent site is the greater curvature in the
antrum, possibly because this area is the most dependent portion of the stomach of the normal, upright
adult, and it has been suggested that this type of ulcer (sump ulcer) is directly related to ingestion of
corrosive drugs.61(3890) Although the surrounding normal gastric mucosa allows these ulcers to be
differentiated histologically, they cannot be distinguished endoscopically from ulcers with other causes
because typical benign gastric ulcers also occur in this location. Location alone is insufficient for
determining cause, and even histology may be misleading; both NSAID use and H. pylori infection are
common, so that it is likely that NSAID ulcers will be found in H. pylori-infected individuals and that H.
pylori ulcers will be present in patients who use NSAIDs chronically.60,6265(3891) Some patients
may have both forms of ulcer. One method to distinguish whether an ulcer is an NSAID ulcer in an H.
pylori-infected individual is to eradicate the H. pylori infection. If the ulcer does not recur with continued
NSAID use, it was likely an H. pylori ulcer. We recommend that eradication of H. pylori be attempted in
all infected patients who have a peptic ulcer.
Ulcers Associated with Pathologic Hypersecretion of Acid
The best endoscopic clues to the presence of the Zollinger-Ellison syndrome are multiple or single
ulcers distal to the duodenal bulb (postbulbar ulcers) and severe erosive gastroesophageal reflux
disease plus a duodenal ulcer. The value of the latter criterion is negated by the presence of delayed
gastric emptying. The observant endoscopist will often note that the stomach contains a considerable
quantity of clear secretions and that the secretions rapidly reaccumulate after suctioning. It is useful to
aspirate several milliliters of gastric secretion for measurement of acidity. A hydrogen ion concentration
of 100 mmol/L or more in the fasting state suggests continuing maximum secretion and a pathologic
hypersecretory state. A serum gastrin level should be obtained on any patient with atypical ulcer
disease. H. pylori infection is uncommon in patients with ulcers associated with the Zollinger-Ellison
Syndrome.66(3892) Meticulous observation, especially in the duodenum, may occasionally disclose
the presence of a small submucosal nodule, particularly in patients with multiple endocrine neoplasia
type I; biopsies or fine-needle aspiration cytology may disclose the presence of a neuroendocrine
tumor.67(3893)
Endoscopic Appearance of Benign Ulcers
Thorough evaluation of an ulcer includes analysis of the site, the shape, and the stage of a lesion. To
identify and characterize ulcers, the stomach must be adequately distended, so that the entire mucosal
surface can be inspected. Ulcers on the lesser curvature of the proximal antrum and linear ulcers
between folds on the greater curvature of the body of the stomach are particularly difficult to identify.
Sometimes marked contraction of the gastric musculature causes an ulcer to be hidden within a deep
crevice. Careful examination is essential when muscular contraction causes the stomach to have an
abnormal or unusual shape. If it is impossible to sufficiently distend the stomach to efface large folds
and visualize the mucosa, the intravenous administration of glucagon (0.5 to 1 mg) will usually relax
the muscular contractions, so that inspection of the entire mucosal surface is possible.
The basic shape of a gastric ulcer is round and discrete, although oval, oblong, or linear ulcers may
occur. Most linear ulcers are positioned with their long axis parallel to the long axis of the stomach,
although in the antrum, the long axis may be transverse and have a saddle appearance (Figures
4614, 4615, 4616, 4617, 4618 and 4619). Linear ulcers are probably transformations of round
or oval types that arise as part of the healing process. Air insufflation within the stomach during
gastroscopy separates the folds and spreads out the ulcer crater to permit detailed inspection of its
edges, base, and surrounding mucosa. The edge of a typical ulcer is sharp and appears as if it were
punched out; undermining of the border may also be discerned. The bottom or base of the ulcer is
usually white, gray-white, or yellow-white, although it may become greenish because of bile staining or
brown because of hemorrhage (Figures 4615 and 4619). Particles of food or barium may adhere to
the ulcer, and occasionally pinkish streaks are seen within the whitish floor. The ulcer base or floor has
a smooth surface except in large ulcers where an uneven surface with nodules and ridges may be
present (Figure 4620).
(3894)Figure 4614. Large saddle-shaped gastric ulcer.
(3895)Figure 4615. Typical round gastric ulcer at the angulus (incisura) of the stomach.
(3896)Figure 4616. Very deep ulcer was hidden within a fold and unapparent until glucagon
was administered intravenously.
(3897)Figure 4617. A, Lesser curvature gastric ulcer crater filled with foreign material. B,
Close-up of the base of the ulcer.
(3898)Figure 4618. Healing gastric ulcer. The edges have become flattened and are almost
level with the plane of the surrounding gastric mucosa.
(3899)Figure 4619. Sharp, bile-stained edge of a large benign gastric ulcer on the greater
curvature of the antrum.
(3900)Figure 4620. Close-up of the edge of a typical benign gastric ulcer. Note that the base
is not completely smooth. The presence of small protuberances in the base is not unusual with
large ulcers.
Skilled endoscopists can usually estimate the diameter of an ulcer with reasonable accuracy. In
addition, it is possible to use a reference scale, such as a measuring device or open biopsy forceps, to
estimate ulcer size (Figure 4621). It is difficult to accurately assess depth because endoscopes
provide only monocular vision. However, depth can be estimated by viewing the ulcer from different
angles with attention to variations in illumination and shadow. Newer imaging techniques with
computer-aided video endoscopy and endoscopic ultrasonography allow precise measurement of size
and depth but currently have limited clinical utility in the management of typical peptic ulcer disease
(see Chapter 20: Stereoscopic Endoscopy).
(3901)Figure 4621. A, Measuring device (American Cystoscope Makers, Inc.) used to

determine the size of a lesion. Each black or white bead is 2 mm in length. B, Use of open
biopsy forceps to estimate the size of an ulcer.
A deep gastric or duodenal ulcer may occasionally penetrate through the gastric wall and into an
adjacent organ such as the liver.6871(3902) This complication is frequently not recognized at
endoscopy and is diagnosed by discovery of hepatic tissue in biopsy specimens.68,69(3903) Benign
peptic ulcer is a common cause of gastrocolic fistula.72,73(3904) The presence of this type of fistula
may not be recognized at upper endoscopy; barium enema is the most reliable method of diagnosis.
It is difficult to distinguish a chronic ulcer from an acute ulcer. When the ulcer is large, the distinction is
easier. If restriction of gastric motility is evident, the presence of fibrosis may be inferred, and this is,
therefore, evidence of chronicity. When mucosal biopsies are taken, the hard, gritty character of a
fibrotic ulcer base can be appreciated, and the ulcer, as noted previously, cannot be separated from
the underlying musculature. In contrast, acute ulcers are soft and pliable, and shallow ones can be
pulled away from the gastric musculature (muscularis propria) with the biopsy forceps.
In the acute stage, the mucosa surrounding an ulcer may be edematous with a mild slope upward
toward the margin (Figure 4622A). But as the ulcer heals, this swelling disappears and the margin
becomes flat (Figures 4623 and 4624; see also Figure 4622B and C). Most ulcers disappear
without visible scar formation (Figure 4625), although the mucosal surface at the site of healing may
appear finely granular (Figure 4626). Healing may be either symmetric or asymmetric, the latter
leading to development of a linear scar. When asymmetric healing occurs, the ulcer may pass through
a dumbbell-shape stage, with eventual division into two ulcers. The initial size of an ulcer is important
as a determinant of whether a scar or a gross deformity of the stomach will be produced by the healing
process. Fortunately, scars are typically produced only by large chronic gastric ulcers. The scar of a
healed gastric ulcer is identifiable as an arched deformity of the gastric lumen, as a collection of gently
sloping and converging folds accompanied at times by a central pit or depression (Figure 4627), or as
a localized regular mucosal depression without discoloration or hyperemia. Folds converging toward
the edge of the ulcer are seen less frequently at endoscopy than with x-ray, although lessening gastric
distention by removal of insufflated air may allow endoscopic visualization.
(3905)Figure 4622. A, Large antral NSAID-induced gastric ulcer. Serial endoscopic

examinations were performed over a period of 6 months as the ulcer healed. B, The edges
have become flat during the healing stage. C, The residual defect is approximately 1.5 mm in
diameter. Deformity of the underlying gastric wall is clearly visible.
(3906)Figure 4623. Changes in endoscopic appearance of a gastric ulcer. (From Tsuneoka

K, Takemoto T, Fukuchi S, eds. Fiberoscopy of Gastric Disease. Tokyo: Igaku-Shoin,
1973:139.)
(3907)Figure 4624. Gastric ulcer, nearly healed.
(3908)Figure 4625. Linear scar at the site of a recently healed gastric ulcer. Over time such a
scar will become increasingly difficult to identify, in contrast to the large scar seen after
healing of large deep ulcers.
(3909)Figure 4626. Granular base in the center of a well-healed gastric ulcer scar.
(3910)Figure 4627. Large scar with granular mucosa at the site of a previously large, healed
gastric ulcer.
The time required for complete ulcer healing is related to the initial size of the ulcer. Large ulcers take
longer to heal than small ulcers, a relationship that holds for both stomach and duodenal ulcers. With
treatment, most gastric ulcers heal within 12 weeks; large ulcers, that is, those with a diameter over 2.5
cm, are allowed 16 weeks of therapy before they are designated as resistant.74,75(3911) Duodenal
ulcers are generally smaller than gastric ulcers and most heal within 4 to 6 weeks after initiation of
treatment. Healing of an H. pylori ulcer is not associated with healing of the surrounding gastritis; acute
and chronic inflammation persist at the ulcer site.57(3912)
Distinction Between Benign and Malignant Ulcers
The final diagnosis with regard to whether an ulcer is benign or malignant is composed of a number of
elements: clinical impression, radiologic assessment, endoscopic assessment, and finally, the
interpretation of biopsy and cytology specimens. It is inappropriate to consider any of these elements
exclusively.
The first goal of endoscopy is to determine the cause of symptoms, for example, the presence of a
gastric ulcer. Once an ulcer has been identified, an additional goal is to prove that the ulcer is
benign.76,77(3913) Radiographically, the most reliable criteria for a benign ulcer are radiating folds,
regular shape, a smooth base, and location of the base on or outside the stomach wall. The most
significant radiographic criteria of malignancy are an adjacent mass, gastric wall rigidity, and failure of
the ulcer base to project beyond the contour of the stomach.
The endoscopist has considerable advantage over the radiologist because of the ability to carefully
examine a magnified image of the ulcer, including the margin, base, and surrounding mucosa. Further,
observations can be made repeatedly from different vantage points. Gastric carcinomas usually
present as mass lesions and, as such, are easily diagnosed. An early gastric cancer may mimic a
benign ulcer, although stigmata of cancer can usually be found by careful inspection (see Chapter 48:
Benign and Malignant Tumors of the Stomach). In the absence of a mass, endoscopic features that
suggest carcinoma include a stepwise depression of the ulcer edge; small extensions of the crater that
blur the ulcer's edge (the edge or margin of a benign ulcer is usually sharply demarcated); bleeding
from the edge of the crater; a necrotic, "dirty" appearance of the crater itself; the presence of abnormal
folds (clubbing or fusion of the tips) at the margin; disruption of the mucosal folds before they reach the
crater; and an irregular or moth-eaten appearance of the surrounding mucosa (Figure 4628).
(3914)Figure 4628. Disruption of the gastric folds is seen well before the edge of the ulcer
crater, signifying that this is a malignant ulcer. The marginal wall rises steeply from the
surrounding mucosa, also a strong indicator of malignancy. Interestingly, the presentation was
acute upper gastrointestinal hemorrhage from the ulcer.
Gabrielsson78(3915) compared gastroscopic and radiographic evaluations of 20 benign and 20
malignant gastric ulcers that were carefully selected as diagnostic problems. Each ulcer had been
diagnosed by both x-ray examination and gastroscopy, and obvious carcinomas were excluded. The
benign ulcers were large and scarred; in no case was carcinoma restricted to the mucosa or
submucosa. The most distinctive x-ray signs were an incisura opposite a benign ulcer crater and
abnormal and disrupted folds with malignant ulcers. The strongest endoscopic criterion for malignancy
was a disruption (flattening) of the folds before they reached the ulcer margin or marginal wall,
whereas blurring of the ulcer edge was the next most reliable endoscopic marker (found in 50% of
those patients with malignancy). Obviously, absence of this sign is meaningless unless the entire
circumference of the ulcer margin can be inspected. A marginal wall that rose steeply from the
surrounding mucosa was a strong indication for malignancy, but this also occurred in a benign case.
No conclusions could be drawn from a regular ulcer shape, whereas an asymmetric shape strongly
indicated malignancy. In spite of these guidelines, this study conclusively demonstrated that neither
x-ray examination nor endoscopic inspection alone was adequate to differentiate between benign and
malignant gastric ulcers.
Biopsy of Gastric Ulcer
The purpose of endoscopic biopsy is to confirm the clinical impression of the nature of a lesion and to
exclude other diseases that have a similar endoscopic appearance. The diagnosis is rarely in doubt
with a large polypoid or ulcerating tumor. A sufficient number of biopsies will ensure that the cell type of
the tumor is identified correctly and the diagnosis confirmed. In the event that all biopsies are negative,
the procedure will be repeated or the patient will be referred for laparotomy. The real problem is the
cancer that appears to be a benign gastric ulcer endoscopically. Although endoscopic appearance
alone is not reliable in the differentiation of benign from malignant ulcers,79,80(3916) some
endoscopists nevertheless use clinical judgment exclusively and do not obtain biopsies. An accuracy of
greater than 95% is ensured with this approach simply because only about 5% of benign-appearing
gastric ulcers are actually malignant (diagnostic accuracy, 95%; sensitivity, 0%). Whether all patients
with radiographically "benign" ulcers should undergo endoscopy for the purpose of biopsy remains
controversial. If the ulcer is diagnosed first by endoscopy, hopefully the usual case in the future, the
question is moot. The ulcer first diagnosed by radiography is the problem. If the ulcer is judged
radiographically to be benign in this circumstance, endoscopy can be delayed until the first follow-up
evaluation, for example, at 8 weeks. This allows time for healing, a factor that helps distinguish benign
from malignant lesions. If there is any degree of uncertainty as to whether a lesion is radiographically
benign, it is prudent to proceed with endoscopy and biopsy. The primary indications for endoscopy in a
patient with a radiologically identified ulcer are to exclude cancer and to categorize the ulcer as to
cause. As a general rule, patients with gastric ulcers should be followed until healing has been
confirmed endoscopically. Serial endoscopic observations with procurement of biopsies and cytology
specimens is mandatory whenever there is the slightest suspicion of malignancy.81(3917) Endoscopic
follow-up will identify relatively few gastric cancers, but some of these cancers will be early-stage,
curable lesions.79,81(3918)
The endoscopist must develop a routine to ensure that a sufficient number of biopsies are obtained to
exclude a small cancer. The prognosis of these "early" cancers is much better than that for advanced
gastric carcinoma; thus their detection is imperative. Unfortunately, few studies have addressed the
question as to where to direct the biopsy forceps.
We believe that endoscopists can specifically recognize abnormal anatomic features and direct the
biopsy to the area most likely to contain a carcinoma. In fact, the experienced endoscopist proves to be
quite skillful at correct identification of the area most likely to contain a cancer. For example, Graham
et al.82(3919) obtained a positive diagnosis with the first biopsy specimen in 70% of cases ultimately
proved to have cancer. Hatfield et al.83(3920) provided the best guidelines for placement of biopsies.
They performed an in vitro study in which they obtained biopsies from freshly resected, ulcerating
carcinomas and then compared the biopsy site with the frequency of positive biopsy specimens (Figure
4629). Fifty percent of the biopsies obtained from either the margin or the base of the ulcer yielded a
diagnosis of cancer. There was a marked reduction in the percentage of positive specimens when the
biopsies were obtained beyond the inner edge of the ulcer (Figure 4629). This study also emphasized
the value to the endoscopist in personally reviewing the biopsy slides, as this provides a continuous
critique of biopsy technique. If a large number of slides contain normal gastric epithelium, the
ulcerating lesion was missed too often, and the evaluation was inadequate. Depending on the clinical
situation, repetition of the procedure may be required.
(3921)Figure 4629. Percentage of positive biopsy specimens obtained in relation to the site
of biopsy of a gastric carcinoma that simulates a benign ulcer. The 100% yield is related to the
fact that a skilled endoscopist can accurately identify the site that has the greatest likelihood of
a positive biopsy.
How many biopsy specimens are sufficient to ensure that a seemingly benign lesion is truly benign?
When an obvious cancer is present, the endoscopist is likely to take many biopsies. "More is better,"
and at least seven or eight biopsy specimens should be obtained. The endoscopist may be less
inclined to obtain an adequate number from benign-appearing ulcers. At a minimum, four biopsies
(three from the margin and one from the base) should be obtained along with directed cytology
specimens. If cytology specimens are not collected, it is prudent to increase the number of biopsies
from an apparently benign ulcer to seven or eight.
It is a good rule not to place more than four or five biopsy specimens in a single bottle of fixative. When
a large number of specimens are pooled, they are often embedded at different levels within the paraffin
block, and consequently, some may be lost during the initial sectioning process or remain unsectioned
deeply within the block. Another important point to stress is that the physician should personally review
the biopsies to ensure that all specimens obtained were examined by the pathologist.
A variety of biopsy forceps in various sizes with biopsy cups of different shapes and configurations
(solid vs. fenestrated) are available. The forceps may also contain a spear or bayonet, which was
designed to permit procurement of several specimens without withdrawing the forceps. Although there
are few controlled studies comparing biopsy forceps, there is little difference in terms of positive yield
with respect to the type of forceps used. We have obtained a high yield of positive biopsies with the
small forceps designed for pediatric instruments, as well as with large instruments designed for
large-channel gastroscopes. Large biopsy forceps are particularly useful when detailed histologic
evaluation of the duodenal, gastric, or esophageal mucosa is desired. Forceps with fenestrated (open)
cups possibly reduce the problem of "crush artifact," but there is no evidence for a significant increase
in the yield for cancer. I personally find little use for forceps with a bayonet because I withdraw the
forceps after each biopsy and the bayonet often bisects or damages the mucosa. We recommend that
biopsies be obtained with the forceps with the largest cup that can be accommodated by the particular
endoscope in use. The best mucosal biopsies are obtained with large cup forceps such as the
Olympus FB-13K (Olympus America, Inc., Melville, NY). We have the bayonet in these forceps
removed by our dental service. This device is almost indestructible and therefore becomes dull with
regular use, and we therefore advise that the cups be sharpened at 6- to 8-month intervals. Again, our
dental service performs this operation for us. These large cup forceps may not obtain large samples
from some very firm ulcers or cancers. Other excellent choices of forceps are the Radial JawTM Max
Capacity and Radial JawTM Large Capacity II series from Microvasive (Microvasive Endoscopy,
Boston Scientific Corp., Natick, MA). These are available without needles and with large (jumbo) cups
(3.3 mm diameter closed) or standard cups (2.2 mm diameter closed) (Figure 4630). They are sold
as single-use items but are sturdy and can be reused multiple times, provided they are subjected to
proper mechanical cleaning and sterilization. A box of five is approximately the same cost as one
multiple-use forceps. Unfortunately, they do not provide good specimens from hard fibrotic lesions. A
bayonet-type forceps is particularly useful for acquisition of biopsies from lesions that are difficult to
reach; the bayonet impales the lesion, thus ensuring that a biopsy is obtained from the correct site. We
recommend that both bayonet and nonbayonet (e.g., bayonet removed) forceps be available in each
unit.
(3922)Figure 4630. Three large-cup biopsy forceps.

It should never be assumed that all endoscopic biopsy forceps provide equivalent specimens. In
particular, the forceps supplied with an endoscope by a manufacturer may not necessarily be
satisfactory. Many forceps in fact provide poor specimens; there are no substitutes for trial, experience,
and interest.
For any particular biopsy forceps, the largest specimens will be obtained when the forceps is oriented
perpendicular to the surface of the target lesion. Gentle pressure may then be exerted downward on
the target, and the entire circumference of each cup is uniformly in contact with the tissue prior to
closure of the forceps. When orientation to a lesion is tangential, additional shearing forces come into
play and the tissue to be sampled is not distributed evenly within the open grasp of the forceps. This
makes it more difficult to maintain the position of the forceps as it is closed. If the forceps are closed
with a rapid, snapping motion, these problems are frequently amplified. Rather, closure should be with
a somewhat slower, steady motion (the experienced assistant can frequently sense whether the
specimen will be adequate and whether of soft or hard consistency). Whenever possible, the
endoscope should be maneuvered to place the target tissue in the best possible orientation before
biopsies are obtained. When this proves to be difficult, one of the most frequently overlooked problems
is the excessively distended stomach. Simply applying suction for several seconds before closing the
forceps will collapse the wall and provide a larger and more representative biopsy. Where it is not
possible to adequately orient the endoscope to the target, a bayoneted forceps may be useful. In some
cases, it may even be necessary to resort to use of a side-viewing endoscope.
The lift-and-cut technique that utilizes a twin-channel endoscope and electrosurgical snare excision of
a piece of mucosa held in the grasp of a biopsy forceps (Figure 4631) has been suggested for
submucosal lesions,84(3923) but this is not used widely. More recently, endoscopists have begun
using needles (similar to sclerotherapy injection needles, but of a larger internal diameter) for needle
biopsies of infiltrating or submucosal lesions (Figure 4632).85(3924) Needle aspiration biopsy is a

useful technique, especially for submucosal lesions.
(3925)Figure 4631. Technique of lift-and-cut biopsy of submucosal lesions. (Adapted from

Martin TR, Onstad GR, Silvis SE, Vennes JA. Lift and cut biopsy technique for submucosal
sampling. Gastrointest Endosc 1976; 23:2930.)
(3926)Figure 4632. Metal and plastic biopsy needles and a tube of preservative for
collection by aspiration of cytology specimens from submucosal or infiltrating lesions.
The collection of cytology specimens in addition to biopsies generates more controversy than the
question deserves. Cytology specimens improve the diagnostic accuracy of endoscopy.86(3927)
Cytology specimens are frequently not collected, probably because this lengthens the procedure,
breaks the endoscopic routine, and is inconvenient. There are two efficient ways to obtain specimens:
brushing and salvage cytology technique. We rarely perform brush cytology, except when anatomic
features of a lesionsuch as a very narrow stricture of the esophagus, stomach, or colonpreclude
collection of adequate biopsy specimens. It has been suggested that the yield of positive results from
brush cytology is better if the samples are obtained before biopsy.87(3928) Although it has not been
demonstrated directly that brushing influences the diagnostic yield of subsequent biopsies, brushing
often causes bleeding that obscures the lesion and interferes with targeting of the forceps. It is our
intuitive opinion that positive biopsy yield is less after brushing because of a reduced ability to carefully
pinpoint the cancer.
We prefer the salvage cytology technique because it is rapid, does not require an additional
instrument, and can be performed as part of the routine biopsy technique.88(3929) The technique is
simple; the only thing required is a mucus trap that is attached between the endoscope and the suction
line. The contents of the accessory (biopsy) channel are aspirated between the collection of each
biopsy specimen. The result is about 5 to 20 ml of fluid. This is diluted with alcohol or another suitable
fixative and submitted for cytologic examination. The simplest fixative is "sputum fixative," readily
available from any hospital's laboratory. This technique salvages malignant cells remaining within the
accessory channel and those that fall from the biopsy forceps. Salvage cytology has a high positive
diagnostic yield; its advantage is that it does not interrupt the endoscopic routine. We reviewed the
results of gastric salvage cytology and biopsy in 671 patients; salvage cytology had a sensitivity of 91%
and a specificity of 100% for detection of adenocarcinoma.89(3930) The combination of forceps biopsy
and cytology had a sensitivity of 100%.
Duodenal Ulcer
The normal duodenal bulb is large with smooth walls (Figure 4633); the bulb is almost never normal
when duodenal ulcer disease is present (Figure 4634). The typical duodenal bulb in patients with
duodenal ulcer disease is smaller than normal, with obvious deformity and mucosal abnormalities.
These abnormalities, such as the presence of gastric metaplasia (Figure 4635), are less apparent
with fiberscopes than with video endoscopes. Introduction of video endoscopy has been a major
advantage for recognizing H. pylori-related and ulcer-related pathology. Duodenal and gastric ulcers
are similar endoscopically. The average duodenal ulcer is round and located in the duodenal bulb
(Figure 4636). Ulcers appear with equal frequency on the anterior and posterior aspects of the bulb.
Because they are found uncommonly in the second or third portion of the duodenum, postbulbar
occurrence suggests a high level of acid secretion and the Zollinger-Ellison syndrome.
(3931)Figure 4633. Normal, nondeformed duodenal bulb.
(3932)Figure 4634. Markedly deformed duodenal bulb with multiple pseudodiverticula.

Because pseudodiverticula trap barium, they make it difficult to interpret radiographic contrast
studies with regard to whether an active ulcer is present. If contrast radiography is used to
follow healing, pseudodiverticula also make it difficult to determine whether an ulcer has
healed.
(3933)Figure 4635. Duodenal bulb with islands of gastric metaplasia extending into the bulb
from the gastroduodenal junction.
(3934)Figure 4636. Typical duodenal ulcers. A, Ulcer on the inferior wall of the duodenal
bulb. B, Ulcer on the superior or apical wall of the bulb. C, Ulcer on the anterior wall of the
bulb.
When the duodenum is irritable or markedly scarred, it may be difficult to exclude the presence of an
ulcer. The entire duodenal bulb must be inspected to ensure that an ulcer is not overlooked. As in the
stomach, administration of glucagon will abolish duodenal contractile activity and facilitate examination
of the mucosal surface. We routinely give glucagon when examining patients in clinical trials and are
regularly impressed that ulcers would have been overlooked without administration of this drug (Figure
4637). Marked scarring is a more difficult problem, particularly as food particles and other material
may be trapped within mucosal folds (Figure 4638). The small-diameter (pediatric) endoscope is
ideally suited for examination of the scarred duodenal bulb because its small diameter and flexibility
permit examination of each recess; video endoscopes with their wide angle of view are equally good
(Figures 4639 and 4640). Postbulbar ulcers, particularly those on the medial wall just beyond the
junction of the first and second portions, may be particularly difficult to identify; adequate examination
of this area may require a pediatric or side-viewing instrument.
(3935)Figure 4637. Deep duodenal ulcer that was not seen until after administration of
glucagon to relax the marked spasm of the musculature.
(3936)Figure 4638. Chronic duodenal ulcer with foreign material trapped in the base.
Duodenal ulcers are usually located in a bed of metaplastic epithelium. A large deep ulcer is
seen inferiorly. Superiorly, the mucosa is gastric metaplasia, whereas to the left and
superiorly, normal-appearing villus duodenum is seen.
(3937)Figure 4639. A, The base of a typical duodenal ulcer covered with white exudate. B,
After the exudate has been removed by suction.
(3938)Figure 4640. The sharp margins of a typical duodenal ulcer. The mucosa at the
margin is typically that of gastric metaplasia.
Ulcers within the pyloric channel or within 1 cm of the pyloric channel (prepyloric ulcers) (Figures
4641, 4642, 4643 and 4644) behave clinically as duodenal ulcers, with the exception that
symptoms of gastric outlet obstruction due to edema, deformity, and narrowing of the pylorus are more
common. Ulcers within, or just beyond, the pyloric channel may be difficult to visualize. A useful
technique to avoid missing an ulcer in these locations is to carefully examine the area while
withdrawing the endoscope from duodenum to stomach. The endoscope is passed into the duodenum,
the tip deflected into one quadrant with reference to the pyloric ring, and then the instrument is slowly
withdrawn while maintaining the tip deflection in order to examine the bulbar mucosa immediately distal
to the pylorus (Figure 4544). This will also flatten out spastic folds in the pyloric valve itself. The
endoscope is then reintroduced, and the technique is repeated for each of the other quadrants.
(3939)Figure 4641. Typical pyloric channel ulcer on the lesser curvature side of the channel.
The lumen is not obstructed.
(3940)Figure 4642. Large ulcer extending from the pyloric channel into the duodenal bulb.
(3941)Figure 4643. Circumferential ulceration at the pylorus with a foreign body (food) in
the actual channel.
(3942)Figure 4644. Deep pyloric channel ulcer that markedly deformed the channel and
could be seen only when the endoscope was withdrawn and the tip deflected toward the ulcer.
When duodenal ulcers are very large, they are designated giant ulcers. Large ulcers have an
associated increased frequency of complications.90(3943) Ulceration of the entire duodenal bulb may
not be appreciated by upper gastrointestinal radiography. The radiographic clue is that the bulb
remains large and its outline does not change through multiple films, that is, no muscular contractions
are evident. With H2-receptor antagonist or proton pump therapy, giant ulcers will heal, but because of
their large size, healing requires a longer period of time in comparison with that for typical small ulcers.
Considerable duodenal deformity usually accompanies healing.
An interesting complication of pyloric or prepyloric ulceration is the development of a double pylorus
(Figure 4645).91,92(3944) In this instance, the ulcer deepens, penetrates the wall, and eventually
perforates into the duodenum.93,94(3945) The remaining column of tissue between the two channels
may be a thick muscular band or a thin band that is easily torn and removed as the endoscope enters
the duodenum. In a few cases, the bridge between the fistulous tract and the normal pylorus may be
lost, leaving a large opening between the stomach and the duodenum.95(3946) There is little
information on the long-term consequences of double pylorus, although this complication appears to
have no serious sequelae and endoscopic incision or disruption of the bridge is generally not
necessary.94,95(3947) There is, however, one report of a patient with symptoms of gastric outlet
obstruction who responded to endoscopic electrosurgical incision of a "thick" bridge separating the
pyloric and the fistulous openings.96(3948) Previous hydrostatic balloon dilation had been ineffective in
this patient.
(3949)Figure 4645. Double pylorus. The smaller opening into the bulb is the fistula. (From
Healing of duodenal and channel ulcers is usually associated with a visible scar that becomes less
apparent over time (Figures 4646, 4647, 4648, 4649, 4650 and 4651). Gastric outlet
obstruction as a result of peptic ulcer disease may respond to endoscopic balloon dilation.97(3950)
(3951)Figure 4646. Healing duodenal ulcer that has become linear.
(3952)Figure 4647. Duodenal ulcer that has broken up into erosions and scars as it heals.
(3953)Figure 4648. Healing duodenal ulcer at the stage of a red scar with active
inflammation.
(3954)Figure 4649. The healing ulcer site is becoming more organized. This appearance
arises rapidly when H. pylori infection has been successfully treated.
(3955)Figure 4650. The ulcer site becomes unrecognizable as the bulb returns toward
normal after successful treatment of H. pylori infection.
(3956)Figure 4651. A and B, Residual deformity and scars in two patients with inactive
duodenal ulcer disease.
Role of Endoscopy
Duodenoscopy is more accurate than radiography in recognizing the presence of a duodenal ulcer and
in distinguishing various anatomic abnormalities that simulate an ulcer (see Figure 4634). The role of
endoscopy in diagnosis and management of the patient with dyspepsia has changed and will continue
to change as new diagnostic approaches become widely available. It is still common practice for the
young patient with typical symptoms to be managed without a precise endoscopic diagnosis and for
endoscopy to be performed in the older patient who is more likely to have another cause for symptoms
such as gastric ulcer or cancer and in any other patient with complicated disease or disease significant
enough to require consultation with a subspecialist. Endoscopy not only confirms the clinical diagnosis
but also excludes other disorders in the stomach or esophagus. Another indication for duodenoscopy
in patients with suspected duodenal ulcer disease is the concern that a complication of the disease
may be present. These include gastric outlet obstruction (regular nausea and vomiting), intractability,
or bleeding. Furthermore, peptic ulcer disease as a basic clinical problem has changed somewhat.
Formerly, diagnosis was the initial issue. Although this remains a fundamental consideration, the cause
of the ulcer is now an additional, clinically relevant question. Endoscopy will reveal evidence of
diseasesuch as small linear esophageal erosions or the presence of gastric or duodenal ulcer
scarsthat is not available by other means. It provides the opportunity to obtain gastric mucosal
biopsies to diagnose H. pylori infection. Based on current knowledge, it seems prudent to offer
endoscopy to all patients with undiagnosed chronic dyspepsia in whom no definite contraindication for
endoscopy is present. As noted previously, endoscopy should include antral mucosal biopsies for
histologic examination. A rapid urease test can also be performed if desired but cannot take the place
of histologic examination of biopsies properly obtained and examined. Culture for H. pylori is expensive
and not usually helpful except in the setting of research.
Mucosal Biopsies for H. Pylori Detection
For diagnosis of H. pylori infection, tissue biopsies can be processed according to standard methods
for preparation of paraffin sections. Formalin fixation, standard in most laboratories, is
satisfactory.98(3957) A number of factors must be considered for optimum specimen evaluation. The
tissue should be oriented so that both the surface and the gastric pits can be examined. Because H.
pylori predominantly inhabits the mucous layer overlying the surface mucosa and within the gastric pits,
biopsies should be handled so as to minimize tissue distortion and so that as little mucus as possible is
dislodged. Thus, methods that involve placing the fresh biopsy on a supporting structure such as a
piece of filter paper should be avoided. Although many biopsies tend to curl owing to contraction of the
smooth muscle of the muscularis mucosa, optimum orientation of the tissue at the time of paraffin
embedding is easily achieved by a trained technician. We recommend that the specimen be "shaken
off" the forceps within the formalin bottle and sent to the pathologist for examination. Biopsies from the
edge of an ulcer or the area immediately surrounding the ulcer may show acute and chronic
inflammation related to the presence of the ulcer and not to H. pylori infection. It is thus essential to
obtain biopsies from normal-appearing mucosa away from the ulcer.
A mucosal mapping study in 89 patients with H. pylori infection found that the organism was present in
all biopsies obtained from the lesser curvature of the antrum in the vicinity of the angulus.99(3958)
Although biopsies obtained from the body of the stomach were less likely to demonstrate the organism,
the differences were not significant.
A positive histology is one that shows acute-on-chronic inflammation and H. pylori. Normal mucosa
shows neither. It may be necessary to advise pathologists as to the expectations implicit in obtaining
the biopsy; formerly the only question posed was "Is it cancer?" and they may be unaware (by practice
and education) that now the biopsies are intended to answer an additional question concerning cause.
Duodenal Ulcer Follow-Up
The natural history of the duodenal ulcer is to heal (see Figures 4646, 4647, 4648, 4649, 4650
and 4651), so that follow-up endoscopy to simply assess healing is not needed after a course of
adequate therapy and control of symptoms. Likewise, repetition of endoscopy is not necessary in every
patient to confirm that recurrence of typical symptoms is indeed due to an active ulcer. The current
question is, "Has the H. pylori infection been eradicated?" By the current definition, eradication means
no evidence of infection 1 or more months after completion of anti-H. pylori therapy.100(3959)
Inclusion of the 1-month interval as part of the definition is necessary, so that the organism has time to
repopulate the stomach after suppression and provide evidence of recrudescence of the infection.
There is now evidence that the endoscopy unit is a site of H. pylori transmission.22,101(3960) We now
recognize that the stomach is not sterile and that endoscopic equipment and accessories require
rigorous disinfection between uses (see Chapter 8: Disinfection of Endoscopes and Accessories).
The Future of Endoscopy in Peptic Ulcer Disease

A number of noninvasive tests that allow accurate diagnosis of the H. pylori status of a patient are
becoming widely available.98(3961) Accurate first- and second-generation serologic tests are already
available in large commercial laboratories; rapid "in-the-office" serologic tests are being introduced.
The 13C-urea breath test will soon be widely available. In contrast to serologic tests, the 13C-urea
breath test determines the presence or absence of active H. pylori infection and should become the
method of choice for follow-up of patients after attempts at eradication of the organism.
One can predict that there will be a move toward noninvasive testing for H. pylori infection in patients
with dyspepsia, many of whom will be treated by their primary care physician without endoscopic
evaluation. This change in approach may also lead to a resurgence of upper gastrointestinal barium
radiology. Patients with a barium study considered diagnostic for the presence of a benign ulcer and
serologic or urea breath test evidence of H. pylori infection will surely be treated for H. pylori
eradication. I anticipate that dyspeptic patients will be referred for endoscopy for the following reasons:
(1) dyspepsia not suggestive of gastroesophageal reflux (possibly only those resistant to therapy with
proton pump inhibitors), (2) dyspepsia but without H. pylori infection, and (3) symptoms that fail to
respond to anti-H. pylori therapy. Such patients often fall into what is now a vague diagnostic group of
nonulcer or essential dyspepsia. There is much to be learned about the cause(s), approach, and
treatment of these conditions.
Drug-Induced Mucosal Damage

The sequence of events leading to NSAID ulcer formation is thought to be mucosal hemorrhage,
erosion, increasing depth of damage, and finally, formation of an acute ulcer. This sequence is
illustrated best by drug-induced gastric mucosal damage, and aspirin-induced injury is probably the
most thoroughly understood.102,103(3962) The initial event in humans is unknown, although recent
studies suggest that polymorphonuclear cells play an important role, an early event being occlusion of
small vessels by "white thrombi."104(3963) This lesion probably precedes the formation of mucosal
hemorrhages.
Acetylsalicylic acid (aspirin) was introduced by Bayer in 1899. Currently, more than 300 proprietary
preparations contain aspirin, and the yearly consumption is in excess of 20 billion tablets. Endoscopic
evidence of the adverse effects of aspirin on the gastric mucosa was first reported in 1938 by
Douthwaite and Lintott.105(3964) The degree of gastric mucosal damage caused by aspirin is
dependent on gastric pH, dosing schedule, and topical aspirin concentration.106(3965) Ultrastructural
changes occur within 3 min of oral aspirin administration, and the potential difference across the
gastric mucosa changes immediately. Administration of a single 650-mg dose of aspirin is associated
with endoscopically visible acute gastric mucosal injury in nearly 100% of volunteer
subjects.107,108(3966) Endoscopically, the first evidence of damage is pinpoint red, brown, or black
dots on and within the mucosa, often associated with streaking of blood on the surface (Figure 4652).
This change may be visible within 30 min after aspirin ingestion. Within 1 to 2 hours, mucosal
hemorrhages ranging from 2 to several millimeters in diameter appear, and by 8 hours, small erosions
may be evident, occurring typically in areas previously having the heaviest involvement with petechial
hemorrhage. By 24 hours, erosions ranging in size from 3 to 5 mm may be seen, as well as larger
acute lesions that are either erosions or small acute ulcers.107,108(3967) If the ulcerative process
breaches the muscularis mucosae, larger erosions or acute gastric ulcers may occur. No area of the
stomach is resistant to the mucosal injury induced by aspirin; visible damage occurs wherever aspirin
contacts the gastric mucosa.106(3968) Even in the same patient, the location of acute superficial injury
is not constant; it can occur in any portion of the stomach. Crossover studies have shown that the
location of injury is not predictable.107(3969) When aspirin is administered in tablet form, frequent and
severe injury usually occurs in the antrum. The reason for this predilection is unknown. It may be due
to the difference between antral and fundal mucosa (the former, for example, lacks parietal cells) or
the antral mucosa may be less tolerant to aspirin. The fact that aspirin particles presumably fall into the
antrum in a person standing or sitting upright is probably also important.85(3970)
(3971)Figure 4652. Gastric petechiae with streaking of blood on the gastric surface signifies
that injury extends to the lumen.
Dose and frequency of administration influence the degree and extent of acute visible mucosal injury.
For example, 650 mg of aspirin administered as a single dose is associated with endoscopically visible
petechial hemorrhage, but erosions are rarely seen and, when present, are small.107,108(3972) When
650 mg of aspirin is administered twice a day, a larger percentage of patients develop erosions; the
percentage rises again with further increases in dose frequency or amount.
With continual aspirin administration, the gastric mucosa adapts, previous lesions heal, and the
evidence of damage resolves.107,109,110(3973) We compared the extent of gastric mucosal damage
associated with 7 days versus 1 day of therapy.107(3974) The degree of mucosal injury reached a
maximum early in the course of therapy, and then decreased despite continued aspirin administration.
Gastric mucosal healing took longer after a single day of aspirin therapy (median, 10 days) than it did
following 7 days of aspirin therapy (median, 4 days). Thus, the gastric response to continuous aspirin
injury is to heal and to do so at an accelerated rate. The mechanism behind this compensation is as
yet unclear. It is also probable that this compensatory adaptation is variable, delayed healing occurring
in some patients, because there is a definite association between NSAID dosage and the occurrence
of gastrointestinal side effects such as ulcer formation.
Endoscopic comparative studies have found marked differences among NSAIDs in their propensity to
cause acute mucosal damage. Projections based on the results of such short-term studies were
predicated on the hypothesis that the results of studies of acute NSAID administration would identify
NSAIDs with little or no propensity to damage the gastric mucosa. The underlying assumption inherent
in this approach is that there would be a relationship between the likelihood of acute mucosal injury
and major gastrointestinal side effects associated with chronic use. Unfortunately, studies of mucosal
damage after acute NSAID administration have failed to provide information useful for predicting the
relative frequency of clinically important gastroduodenal injury after chronic administration of any given
NSAID. Studies showing that parenterally administered NSAIDs and enteric-coated NSAIDs have the
same propensity to cause ulcers as does oral administration have led to the seemingly inescapable
conclusion that the severe untoward effects of these agents represent a systemic reaction and not
simply a topical effect. We now believe that all NSAIDs have the tendency to cause ulceration and the
attendant complications of bleeding and perforation. Higher doses of these anti-inflammatory agents
are associated with increased risk.
Investigation of NSAID-Induced Gastroduodenal Injury

One of the earliest clinical observations demonstrated that ingestion of aspirin, the prototypical NSAID,
is associated with gastric petechiae and small mucosal erosions. Serial endoscopies showed that the
petechiae appeared soon after aspirin administration, whereas erosions appeared after a latent period
of many hours.106,107(3975) These observations were remarkably reproducible, and it became
standard practice to compare each new NSAID with aspirin in endoscopic studies. These studies were
typically of short duration (1 or 2 weeks) and were undertaken in healthy volunteers. Invariably, each
new NSAID has proved to be "less damaging" to the stomach than aspirin, and the results of these
studies have been used to help market new NSAIDs with the implication that less damage in a short
endoscopic study in normal volunteers can be equated with "safer." Pharmaceutical companies who
market antiulcer drugs employed similar logic and experimental protocols to investigate whether
antiulcer agents would prevent, or ameliorate, mucosal damage after acute administration of NSAIDs.
The tacit assumption was that if a drug could prevent endoscopically visible acute injury with aspirin, it
could also prevent the large and more serious ulcers that are increasingly recognized with NSAID use.
We now have sufficient experience to realize that many of the interpretations from endoscopic studies
after acute NSAID administration were wrong. Until recently, there was considerable confusion as to
the best method for presentation of the results of such studies. Rating scales for mucosal damage can
be grouped into two types: those that count lesions and those that assess both the severity and the
distribution of lesions. Gastric damage is not linear, that is, a grade 4 lesion is not simply twice as
severe as a grade 2. Therefore, each scale results in a generalized classification of damage from mild
to severe. It is assumed that acute ulcers are a more severe injury than erosions, and erosions are
more severe than mucosal hemorrhages.
Perhaps the most potentially misleading method of data presentation involves division of the upper
gastrointestinal tract into areas, scoring each for the presence or absence of specific lesions, and then
adding the scores for comparison. For example, separate numerical values were obtained for five
areas (esophagus, fundus, body and antrum of the stomach, and the duodenal bulb). Each area was
evaluated for the presence of erythema, mucosal hemorrhages, erosions, or ulcers, and each type of
lesion was given a numerical value. The scores for each type of lesion were then added to achieve a
final "score" for an individual patient. Results obtained with different drugs were compared using
standard parametric statistical analyses. A serious drawback of this method is that the results are
treated as numerical when they really represent ordered categories. For example, a patient with a
2-cm-diameter gastric ulcer might be assigned the score of 10, which would be identical to the score of
a patient with gastric erythema, 15 petechiae, and 2 tiny erosions.
Another serious problem in data analyses was that of multiple comparisons (pairwise comparison of
each area, each lesion type, and each drug), which almost ensures the investigator of at least one
"statistically significant" comparison. The greatly increased likelihood of a type I error was ignored.
Consequently, a test drug might fail to provide overall protection and yet the report would conclude that
"drug A significantly protected a specific area (e.g., the duodenum) against aspirin-induced injury."
There are examples in which the number of erosions actually increased and yet the overall result was
that a study drug successfully "protected" against NSAID-induced damage.
Some investigators have used a "global" assessment based on the original model of Lanza et
al.111(3976) Such rating scales have been progressively modified with use, and our most recent
version (Table 461) minimizes petechial hemorrhage, so that, no matter how many are present, the
score cannot increase above what is usually considered trivial damage. Current thinking among
investigators is that mucosal erythema is meaningless because this finding has no known clinical
sequelae and there is considerable interobserver variability inherent in its assessment. Petechial
hemorrhage represents at best trivial damage. Erosions are of unknown significance, but most
investigators agree that more erosions are "worse" than few; large erosions or acute ulcers are
considered "worse" than a few erosions. None of these findings has been shown to be of value for
predicting the propensity of a particular drug to cause chronic ulceration, bleeding, or perforation, nor
has prevention of such lesions in acute studies been shown to have predictive value for prevention of
either large ulcerations or their complications.12(3977)
Gastroscopic Scoring System

for NSAID-Induced Mucosal Injury
TABLE 461
GRADE
0
1
2
3
4
DESCRIPTION (CIRCLE SCORES)

Mucosal Hemorrhages
No evidence of mucosal hemorrhages
One area of mucosal hemorrhage
More than one area of mucosal hemorrhage but
neither numerous nor widespread
Numerous (3 or more) areas of mucosal
hemorrhage
Widespread involvement of gastric mucosa with
mucosal hemorrhage
Gastroscopic Scoring System

for NSAID-Induced Mucosal Injury
TABLE 461
GRADE
0
A
B
C
D
Score:
Video?
DESCRIPTION (CIRCLE SCORES)

Final Score
Normal stomach (includes grades 1, Mucosal
Hemorrhages)
Mucosal hemorrhage (grade 2 or greater, Mucosal
Hemorrhages)
Gastric erosions (1 or 2)
3 or more areas of gastric erosion
Large areas of erosion with widespread
involvement of stomach
Stomach [ ] Duodenum [ ]
Yes [ ] No [ ] Tape #
Spices and Gastric Mucosal Damage

The ability of spicy foods to cause gastric mucosal damage has not been fully investigated. Spices that
appear to be associated with symptoms of gastric distress when taken with food are black pepper, chili
pepper, mustard seeds, and cloves.112(3978) In humans, mustard and paprika have been shown to
affect gastric acid secretion; mustard (1 g) has a depressive action, whereas paprika (0.5 to 1 g) may
cause an increase in gastric acidity.113(3979) Celery salt, nutmeg, sage, cinnamon, cloves, and
pepper have no apparent effect on gastric secretion.113 Chili pepper has been shown to increase
gastric mucosal epithelial cell turnover, as evidenced by an increase in DNA content of gastric
aspirates.114(3980) Schneider et al.112(3981) studied the effects of several spices. Endoscopy was
performed after each spice (2 times the amount used in highly spiced foods) was allowed to remain
in the stomach for 10 min. No significant change in the gastric mucosa was observed in patients who
received cinnamon, nutmeg, allspice, thyme, black pepper, chili pepper, cloves, or paprika. Thyme and
mustard seed produced mild erythema, but no symptoms. Chili pepper produced moderate mucosal
hyperemia, and black pepper produced severe hyperemia, although neither elicited symptoms.
Capsaicin (8-methyl-N-vannillyl-6-neneamide) is the pungent irritating compound that is the active
ingredient in Capsicum fruits such as jalapeo peppers. There are at least 200 commercially available
varieties of Capsicum from around the world, ranging from capsaicin-free "sweet" varieties to the
Mombasa variety, which is said to be too hot to be eaten. The consumption of peppers is in the range
of 6 million metric tons annually and rising.
We used video endoscopy to investigate the effect of eating capsaicin-containing spicy foods on the
gastric mucosa.115(3982) Four meals were employed: a bland meal of unpeppered steak and French
fries (negative control), a bland meal accompanied by 1950 mg of aspirin (positive control), a spicy
meal of Mexican food (30 g of jalapeo peppers), and a pepperoni pizza. Twelve subjects were studied
in a randomized, crossover trial, with the test meal given at noon and in the evening; each subject
received all four test meals. Each study consisted of a baseline endoscopy that was repeated
approximately 12 hr after the last test meal. Eleven of 12 individuals ingesting the bland meal plus
aspirin developed multiple gastric erosions. In contrast, the median endoscopic score for the other
three meals was 0. Single cases of a single erosion were present after the Mexican and the pizza
meals. In another experiment, the effect of spices applied directly to the gastric mucosa was studied;
approximately 30 g of fresh jalapeo peppers was ground in a food processor and then placed directly
into the stomach. Endoscopy after 24 hr revealed no visible mucosal damage. The results of this study
indicate that ingestion of highly spiced meals by normal individuals is not associated with
endoscopically demonstrable gastroduodenal mucosal damage. We concluded that although
capsaicin-containing spices increase gastric acid secretion and their use may be associated with
gastric distress, they add to the flavor and enjoyment of eating and do not appear to cause mucosal
damage.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Conn RB. Clinical laboratories. Profit center, production industry, or patient care resource?. N
Engl J Med 1978;298:4227.
Griner PF, Liptzin BJ. Use of the laboratory in a teaching hospital: Implications for patient
care, education, and hospital costs. Ann Intern Med 1971;75:15763.
Tedesco FJ. Endoscopy in the evaluation of patients with upper gastrointestinal symptoms:
Indications, expectations, and interpretation. J Clin Gastroenterol 1981;3:6771.
Cotton PB. Upper gastrointestinal endoscopy. Br J Hosp Med 1976;16:715.
Cotton PB. Fibreoptic endoscopy and the barium meal: Results and implications. BMJ
1973;2:1619.
Stevenson GW. Who needs radiology?. Gastrointest Endosc 1980;26:11925.
Wormsley KG. Diagnostic procedures. In Horrobin DE (ed). Annual Research Reviews. Vol 2.
Duodenal Ulcer. Edinburgh: Churchill Livingstone, 1979;88.
Dooley CP, Larson AW, Stace NH, et al. Double-contrast barium meal and upper
gastrointestinal endoscopy. A comparative study. Ann Intern Med 1984;101:53845.
Gelfand DW, Ott DJ, Tritico R. Causes of error in gastrointestinal radiology. I. Upper
gastrointestinal examination. Gastrointest Radiol 1980;5:917.
Chandie Shaw MP, van Romunde LKJ, Janssens AR, et al. Peptic ulcer and gastric
carcinoma: Diagnosis with biphasic radiography compared with fiberoptic endoscopy.
Radiology 1987;163:3942.
Gelfand DW, Ott DJ, Munitz HA, Yu MC. Radiology and endoscopy: A radiologic viewpoint.
Population exposure to x-rays. U.S. 1970. Publication (FDA) 738047. Washington, DC: U.S.
Department of Health, Education, and Welfare. 1973.
Revesz G, Shea FJ, Ziskim MC. Patient flow and utilization of resources in a diagnostic
radiology department. Radiology 1972;104:216.
Marton KI, Sox HC Jr, Wasson J, Duisenberg CE. The clinical value of the upper
gastrointestinal tract roentgenogram series. Arch Intern Med 1980;140:1915.
Op den Orth JO. Use of barium in evaluation of disorders of the upper gastrointestinal tract:
Current status. Radiology 1989;173:6018.
Bulpitt CJ, Rowntree RK, Semmence A. A randomized controlled trial of the effects of
screening for ulcer-type dyspepsia. J Epidemiol Community Health 1982;36:1725.
Beavis AK, La Brooy S, Misiewicz JJ. Evaluation of one-visit endoscopic clinic for patients with
dyspepsia. BMJ 1979;1:13879.
Gear MWL, Ormiston MC, Barnes RJ, et al. Endoscopic studies of dyspepsia in the
community: An "open-access" service. BMJ 1980;280:1135.
Gear MWL, Barnes RJ. Endoscopic studies of dyspepsia in a general practice. BMJ
1980;280:11367.
Bytzer P, Hansen JM, Schaffalitzky de Muckadell OB. Empirical H2-blocker therapy or prompt
endoscopy in management of dyspepsia?. Lancet 1994;343:8116.
Walsh JH, Peterson WL. Treatment of Helicobacter pylori infection in the management of
peptic ulcer disease. N Engl J Med 1995;333:98491.
Graham DY. Helicobacter pylori: Its epidemiology and its role in duodenal ulcer disease. J
Wells JJ, Nostrant TT, Wilson JA, Gyves JW. Gastroduodenal ulcerations in patients
receiving selective hepatic artery infusion chemotherapy. Am J Gastroenterol 1985;80:4259.
24. Doria MI Jr, Doria LK, Faintuch J, Levin B. Gastric mucosal injury after hepatic arterial
infusion chemotherapy with floxuridine. A clinical and pathologic study. Cancer
1994;73:20427.
25.
Ivy AC, Grossman MI, Bachrach WH. Peptic Ulcer. Philadelphia: Blakiston Company. 1950.
26. Palmer ED. Editorial: Whatever happened to gastric petechiae?. Gastrointest Endosc
1974;21:801.
27.
Dagradi AE, Stempien SJ, Lee ER, Juler G. Hemorrhagic-erosive gastritis. Gastrointest
Endosc 1968;14:14750.
28.
Schindler R. Lehrbuch und Atlas der Gastroskopie. Munchen, Lehmann. 1923.
29. Catalano D, Pagliari U. Gastroduodenal erosions: Radiological findings. Gastrointest Radiol
1982;7:23540.
30. Franzin G, Manfrini C, Musola R, et al. Chronic erosions of the stomachA clinical,
endoscopic and histological evaluation. Endoscopy 1984;16:15.
31. Freise J, Hofmann R, Gebel M, Huchzermeyer H. Follow-up study of chronic gastric erosions.
32. Rsch W. Erosions of the upper gastrointestinal tract. Clin Gastroenterol 1978;7:62334.
33. Karvonen AL. Occurrence of gastric mucosal erosions and their association with other upper
gastrointestinal disease: A study of patients examined by elective gastroscopy. Ann Clin Res
1981;13:15963.
34. Karvonen AL. Occurrence of gastric mucosal erosions in association with other upper
gastrointestinal diseases, especially peptic ulcer disease, as revealed by elective gastroscopy.
35.
Kawai K, Shimamoto K, Misaki F, et al. Erosion of gastric mucosaPathogenesis, incidence,
and classification of the erosive gastritis. Endoscopy 1970;3:16874.
36.
Rsch W, Ottenjann R. Gastric erosions. Endoscopy 1970;2:93.
37. Soehendra N, Rehner M, Sternberg N, Lietzke C. Die chronischen (kompletten) erosionen
des magens. Munch Med Wochenschr 1972;114:185760.
38. Braverman DZ, Morali GA, Patz JK, Jacobsohn WZ. Is duodenal ulcer a seasonal disease? A
retrospective endoscopic study of 3105 patients. Am J Gastroenterol 1992;87:15913.
39. Fich A, Goldin E, Zimmerman J, Rachmilewitz D. Seasonal variations in the frequency of
endoscopically diagnosed duodenal ulcer in Israel. J Clin Gastroenterol 1988;10:3802.
40. Cameron AJ, Higgins JA. Linear gastric erosion. A lesion associated with large diaphragmatic
hernia and chronic blood loss anemia. Gastroenterology 1986;91:33842.
41. Scolapio JS, DeVault K, Wolfe JT. Eosinophilic gastroenteritis presenting as a giant gastric
ulcer. Am J Gastroenterol 1996;91:8045.
42.
Magnus HA. The pathology of peptic ulceration. Postgrad Med J 1954;30:31136.
43.
Morson BC, Dawson IMP. Peptic ulceration. In Morson BC, Dawson IMP (eds).
Gastrointestinal Pathology. Oxford, Blackwell Scientific, 1972;11027.
44. Gelfand DW, Ott DJ. Gastric ulcer scars. Radiology 1981;140:3743.
45. Kreuning J, Bosman FT, Kuiper G, et al. Gastric and duodenal mucosa in "healthy"
individuals. An endoscopic and histopathological study of 50 volunteers. J Clin Pathol
1978;31:6977.
46. Martin TR, Vennes JA, Silvis SE, Ansel HJ. A comparison of upper gastrointestinal endoscopy
and radiology. J Clin Gastroenterol 1980;2:215.
47. Thomas J, Greig M, McIntosh J, et al. The location of chronic gastric ulcer. A study of the
relevance of ulcer size, age, sex, alcohol, analgesic intake and smoking. Digestion
1980;20:7984.
48.
Oi M, Oshida K, Sugimura S. The location of gastric ulcer. Gastroenterology 1959;36:4556.
49. Fallingborg J, Poulsen LO, Grove A, Teglbjaerg PS. Frequency of Helicobacter pylori and
gastritis in healthy subjects without gastrointestinal symptoms. Scand J Gastroenterol
1992;27:38890.
50. Barthel JS, Westblom TU, Havey AD, et al. Gastritis and Campylobacter pylori in healthy,
asymptomatic volunteers. Arch Intern Med 1988;148:114951.
51. Blomberg B, Jarnerot G, Kjellander J, et al. Prevalence of Campylobacter pylori in an

unselected Swedish population of patients referred for gastroscopy. Scand J Gastroenterol
1988;23:35862.
52. Dooley CP, Cohen H, Fitzgibbons PL, et al. Prevalence of Helicobacter pylori infection and
histologic gastritis in asymptomatic persons. N Engl J Med 1989;321:15626.
53. Sandikci MU, Doran F, Koksal F, et al. Helicobacter pylori prevalence in a routine upper
gastrointestinal endoscopy population. Br J Clin Pract 1993;47:1879.
54. Tatsuta M, Okuda S. Location, healing, and recurrence of gastric ulcers in relation to fundal
gastritis. Gastroenterology 1975;69:897902.
55.
Ida K, Kohli Y, Shimamoto K, et al. Endoscopical findings of fundic and pyloric gland area
using dye scattering method. Endoscopy 1973;5:216.
56. Axon AT. Helicobacter pylori therapy: Effect on peptic ulcer disease. J Gastroenterol Hepatol
1991;6:1317.
57. Dixon MF. Helicobacter pylori and peptic ulceration: Histopathological aspects. J
58. Graham DY, Lew GM, Klein PD, et al. Effect of treatment of Helicobacter pylori infection on
the long-term recurrence of gastric or duodenal ulcer: A randomized, controlled study. Ann
Intern Med 1992;116:7058.
59.
. Helicobacter pylori in peptic ulcer disease. NIH Consensus Statement 1994;12:123.
60. Graham DY. The relationship between nonsteroidal anti-inflammatory drug use and peptic
ulcer disease. Gastroenterol Clin North Am 1990;19:17182.
61.
Kottler RC, Tuft RJ. Benign greater curve gastric ulcer: The "sump-ulcer." Br J Radiol.
54:65141981.
62. Graham DY, Lidsky MD, Cox AM, et al. Long-term nonsteroidal anti-inflammatory drug use
and Helicobacter pylori infection. Gastroenterology 1991;100:16537.
63. Taha AS, Sturrock RD, Russell RI. Mucosal erosions in longterm non-steroidal
anti-inflammatory drug users: Predisposition to ulceration and relation to Helicobacter pylori.
Gut 1995;36:3346.
64. Borody TJ, Brandl S, Andrews P, et al. Helicobacter pylori-negative gastric ulcer. Am J
65. al-Assi MT, Genta RM, Karttunen TJ, Graham DY. Ulcer site and complications: Relation to
Helicobacter pylori infection and NSAID use. Endoscopy 1996;28:22933.
66. Saeed ZA, Evans DJ Jr, Evans DG, et al. Helicobacter pylori and Zollinger-Ellison syndrome.
Dig Dis Sci 1991;36:158.
67. Benya RV, Metz DC, Hijazi YJ, et al. Fine needle aspiration cytology of submucosal nodules in
patients with Zollinger-Ellison syndrome. Am J Gastroenterol 1993;88:25865.
68. Guerrieri C, Waxman M. Hepatic tissue in gastroscopic biopsy: Evidence of hepatic
penetration by peptic ulcer. Am J Gastroenterol 1987;82:8903.
69. Jimenez-Perez FJ, Munoz-Navas MA. Endoscopic diagnosis of gastric peptic ulcer
penetrating into the liver. Endoscopy 1991;23:989.
70. Misra PS, Muscat JL. Endoscopic recognition of duodenal ulcer penetrating to the liver.
71. Matsuoka T, Nagai Y, Muguruma K, et al. Liver penetration and gastrobronchial fistula:
Unusual complication of a peptic ulcer. Am Surg 1995;61:4924.
72. Tavenor T, Smith S, Sullivan S. Gastrocolic fistula. A review of 15 cases and an update of the
literature. J Clin Gastroenterol 1993;16:18991.
73. Levine MS, Kelly MR, Laufer I, et al. Gastrocolic fistulas: The increasing role of aspirin.
Radiology 1993;187:35961.
74. Lewis JH. Treatment of gastric ulcer. What is old and what is new?. Arch Intern Med
1983;143:26474.
75. Scheurer U, Witzel L, Halter F, et al. Gastric and duodenal ulcer healing under placebo
76. Chang FM, Saito T, Ashizawa S. Follow-up endoscopic study of gastric mucosal changes
secondary to gastric ulcer. Endoscopy 1978;10:3340.

Salter RH, Gill DK, Girdwood TG, et al. Gastric ulcer: Is endoscopy always necessary?. BMJ
1981;282:2097.
78.
Gabrielsson N. Benign and malignant gastric ulcers. Evaluation of the differential diagnostics
in roentgen examination and endoscopy. Endoscopy 1972;4:7383.
79. Bytzer P. Endoscopic follow-up study of gastric ulcer to detect malignancy: Is it worthwhile?.
80. Adachi Y, Mori M, Tsuneyoshi M, Sugimachi K. Benign gastric ulcer grossly resembling
malignancy. A clinicopathologic study of 20 resected cases. J Clin Gastroenterol
1993;16:1038.
all gastric ulcers?. Dig Dis Sci 1993;38:2848.
83. Hatfield ARW, Slavin G, Segal AW, Levi AJ. Importance of the site of endoscopic gastric
biopsy in ulcerating lesions of the stomach. Gut 1975;16:8846.
84. Martin TR, Onstad GR, Silvis SE, Vennes JA. Lift and cut biopsy technique for submucosal
sampling. Gastrointest Endosc 1976;23:2930.
85. Graham DY, Smith JL, Bouvet AA. What happens to tablets and capsules in the stomach?
Endoscopic comparison of disintegration and dispersion characteristics of microencapsulated
potassium formulations. J Pharm Sci 1990;79:20712.
86. Cardillo MR, Agnelli M. Brush cytology in the endoscopic diagnosis of benign gastric ulcers. A
useful adjunct to biopsy?. Arch Anat Cytol Pathol 1990;38:8691.
87. Keighley MRB, Thompson H, Moore J, et al. Comparison of brush cytology before or after
biopsy for diagnosis of gastric carcinoma. Br J Surg 1979;66:2467.
88. Graham DY, Spjut HJ. Salvage cytology. A new alternative fiberoptic technique. Gastrointest
Endosc 1979;25:1379.
89. Green LK, Zachariah S, Graham DY. The use of gastric salvage cytology in the diagnosis of
malignancy: A review of 731 cases. Diagn Cytopathol 1990;6:14.
90. Collen MJ, Santoro MJ, Chen YK. Giant duodenal ulcer. Evaluation of basal acid output,
nonsteroidal antiinflammatory drug use, and ulcer complications. Dig Dis Sci 1994;39:11136.
91. Hansen OH, Kronborg O, Pedersen T. The double pylorus. Scand J Gastroenterol
1972;7:6956.
92. Cappelletti F, Recchia S, Bonardi L, et al. Gastroduodenal fistula complicating a prepyloric
ulcer. Gastrointest Endosc 1983;29:1113.
93. Farack UM, Goresky CA, Jabbari M, Kinnear DG. Double pylorus: A hypothesis concerning its
pathogenesis. Gastroenterology 1974;66:596600.
94. Fayenz S. The evolution of the double pylorus. Gastrointest Endosc 1986;32:312.
95. Einhorn RI, Grace ND, Banks PA. The clinical significance and natural history of the double
pylorus. Dig Dis Sci 1984;29:2138.
96. Graham SM, Lin F, Flowers JL. Symptomatic double-channel pylorus. Successful treatment
with a biliary sphincterotome. Surg Endosc 1994;8:7923.
97. Disario JA, Fennerty MB, Tietze CC, et al. Endoscopic balloon dilation for ulcer-induced
gastric outlet obstruction. Am J Gastroenterol 1994;89:86871.
98.
Alpert LC, Graham DY, Evans DJ Jr, et al. Diagnostic possibilities for Campylobacter pylori
infection. Eur J Gastroenterol Hepatol 1989;1:1726.
99. Genta RM, Graham DY. Comparison of biopsy sites for the histopathologic diagnosis of
Helicobacter pylori: A topographic study of H. pylori density and distribution. Gastrointest
Endosc 1994;40:3425.
100.
Borsch GM, Graham DY. Helicobacter pylori. In Collen MJ, Benjamin SB (eds). Pharmacology
of Peptic Ulcer Disease. Handbook of Experimental Pharmacology. Vol 99. Berlin,
101. Graham DY, Alpert LC, Smith JL, Yoshimura HH. Iatrogenic Campylobacter pylori infection is
77.
a cause of epidemic achlorhydria. Am J Gastroenterol 1988;83:97480.

102. Cooke AR. The role of the mucosal barrier in drug-induced gastric ulceration and erosions.
Am J Dig Dis 1976;21:15564.
103.
Pfeffer CJ (ed). Drugs and Peptic Ulcer. Vol 2: Pathogenesis of Ulcer Induction Revealed by
Drug Studies in Humans and Animals. 2nd ed. Boca Raton, FL, CRC Press. 1982.
104. Wallace JL, Keenan CM, Cucala M, et al. Mechanisms underlying the protective effects of
interleukin 1 in experimental non-steroidal anti-inflammatory drug gastropathy.
105.
Douthwaite AH, Lintott GAM. Gastroscopic observation of the effect of aspirin and certain
other substances on the stomach. Lancet 1938;2:12225.
106. Graham DY, Smith JL. Aspirin and the stomach. Ann Intern Med 1986;104:3908.
107.
Graham DY. Treatment of benign and malignant strictures of the esophagus. In Silvis SE
(ed). Therapeutic Gastrointestinal Endoscopy. New York, Igaku-Shoin, 1989;141.
108.
O'Laughlin JC, Hoftiezer JW, Ivey KJ. Effect of aspirin on the human stomach in normals:
Endoscopic comparison of damage produced one hour, 24 hours, and 2 weeks after
administration. Scand J Gastroenterol 1981;67:2114.
109.
Hurley JW, Crandall LA Jr. The effect of various salicylates upon the dog's stomach: A
gastroscopic photographic evaluation. In Dixon ASTJ, Martin BK, Smith MJH, Wood RHN
(eds). Salicylates, An International Symposium. Boston, Little, Brown, 1963;2136.
110. Eastwood CL, Quimby GF. Effect of chronic aspirin ingestion on epithelial proliferation in rat
fundus, antrum, and duodenum. Gastroenterology 1982;82:8526.
111. Lanza FL, Royer GL Jr, Nelson RS. Endoscopic evaluation of the effects of aspirin, buffered
aspirin, and enteric-coated aspirin on gastric and duodenal mucosa. N Engl J Med
1980;303:1368.
112.
Schneider NA, Deluca V Jr, Gray SJ. The effect of spice ingestion upon the stomach. Am J
113.
Sanchez-Palomera E. The action of spices on the acid gastric secretion on the appetite and
on the caloric intake. Gastroenterology 1951;18:25468.
114. Desai HG, Venugopalan K, Antia FP. Effect of red chili powder on DNA content of gastric
aspirates. Gut 1973;14:9746.
115. Graham DY, Smith JL, Opekun AR. Spicy food and the stomach. Evaluation by
videoendoscopy. JAMA 1988;260:34735.
Chapter 47 Gastritis and Gastropathy

(3983)
(3984)
WILFRED M. WEINSTEIN, M.D.
Helicobacter pylori gastritis is of enormous interest because of its link with peptic ulcer, gastric
carcinoma, and gastric lymphoma. Another development, much less glamorous than H. pylori gastritis,
is the codification of the term and the concept of gastropathies. The word gastropathy denotes a type
of gastric mucosal abnormality in which some feature other than inflammatory cell infiltrates is
dominant. There are three main types of gastropathy:1(3985) reactive, congestive, and hypertrophic. In
reactive gastropathy, epithelial cell change dominates the histologic picture, often accompanied by
foveolar hyperplasia. Congestive gastropathy refers to a pattern in which vascular changes dominate,
again without significant inflammation. Hypertrophic gastropathy refers to an increase in mucosal
thickness as the dominant feature, usually with marked foveolar hyperplasia. It may also be
accompanied by inflammatory cell infiltrates. The rare disorder Mntrier's disease is the best-known
example of hypertrophic gastropathy.
Histology of the Stomach: Relation to Endoscopic Landmarks

The stomach consists of three different zones, named according to the gland type they contain: cardiac
gland zone, oxyntic (fundic) gland zone, and antral gland zone. Many types of gastric mucosal injury
have a predilection in distribution or severity for one zone over the others.
Cardiac Gland Zone and "Carditis"

The cardiac gland zone is a narrow, 1- to 2-cm-wide, band of mucosa at the squamocolumnar junction.
It contains mucous glands similar in appearance to gastric antral glands and duodenal Brunner's
glands. It usually consists of a blend of both cardiac mucous glands and oxyntic (fundic) glands, with
either an admixture of both parietal and chief cells or just parietal cells. Often, the content is not purely
mucous cells, and parietal as well as chief cells are present. Curiously, microscopic nests of pancreatic
exocrine glands may be found at the cardia, unassociated with any visible endoscopic
changes.2(3986)
Interest is growing in the gastric cardia because of the possibility that cardia inflammation and
metaplasia associated with gastroesophageal reflux may in some way predispose to cardia cancer, a
tumor of increasing prevalence in most Western countries.3(3987)
The attention of endoscopists has only recently been directed toward the gastric side of the
gastroesophageal junction in gastroesophageal reflux disease (GERD), whereas pathologists have
been aware for some time of the involvement of the cardia in this condition.4,5(3988) In GERD, the
gastric side of the gastroesophageal junction may be eroded or contain nodules; discrete inflammatory
polyps may be present within the cardiac gland mucosa just distal to the squamocolumnar junction.
Biopsy of the gastric cardia has been proposed in a preliminary study to be a more sensitive marker of
GERD than biopsy of squamous epithelium.6(3989) In a study by Clark et al.,6(3990) endoscopic
biopsies were obtained from the cardia with the endoscope retroverted within the stomach (Figure
471). With this approach, it is much easier to target the gastroesophageal junction and obtain
biopsies that contain both esophageal and gastric mucosa (Figure 471). This is important because
the changes of carditis and intestinal metaplasia are often localized right at the squamocolumnar
junction, and the mucosa is often normal just a few millimeters distal to the junction. The presence of a
hiatus hernia makes accurate targeting of the biopsy forceps much less difficult in the retroverted
position in the stomach. Because of the increasing recognition of the cardia's importance in GERD and
because of the potential connection between GERD and cardia cancer, the gastric cardia should be
evaluated at endoscopy with greater thoroughness than in the past.
(3991)Figure 471. Retroverted (turnaround, retroflexed) view of the squamocolumnar

junction from within the stomach (top) and a biopsy forcep straddling the squamocolumnar
junction (Z line) poised to obtain a biopsy (bottom). To obtain biopsies for carditis or
intestinal metaplasia of the gastroesophageal junction, it is important to have squamous and
gastric epithelium represented in tissue sections because carditis is often absent just a few
millimeters below the gastroesophageal junction.
Oxyntic (Fundic) and Antral Gland Zones

The term oxyntic gland mucosa is now preferred by many gastrointestinal (GI) pathologists to the more
commonly used fundic gland mucosa because the latter creates confusion with regard to the anatomic
fundus of the stomach. The oxyntic gland mucosa occupies both the fundus and the body of the
stomach. Its glands consist mainly of acid-secreting parietal and pepsinogen-secreting chief cells.
The antral gland mucosa occupies the gastric antrum. Its glands are of the mucous type; gastrin cells
are located at the interface of the gastric pits (foveolae) and the mucous glands.
Between the oxyntic and the antral gland zones are transition zones that combine features of the
adjacent gland types. The antral gland-oxyntic gland transition zone is often more extensive along the
lesser curvature of the stomach.

Even though the appearance of the gastric antrum is distinctly different from that of the body of the
stomach at endoscopy, its histology frequently does not conform to endoscopically defined boundaries.
The proximal half of the gastric antrum often consists exclusively or primarily of oxyntic gland mucosa.
Even in the distal half of the antrum, oxyntic elements or parietal cells alone are found among the
antral glands as far distally as the pylorus. This has practical implications for sampling of antral glands
(e.g., for H. pylori gastritis) in that biopsies should be taken no more proximally than 2 cm from the
pylorus.
Gastric Mucosal Biopsy: Pointers and Principles

Most of the following "pointers and principles" for obtaining biopsies in the stomach may be applied
throughout the GI tract. Table 471 lists the types of information that the pathologist needs when
gastric biopsies are submitted.
Information Required by the

Pathologist When Gastric Biopsies Are
Submitted
TABLE 471
SITES
Fundus
Body, antrum (indicate curvature and whether proximal,
mid, or distal)
LESION DESCRIPTION
If abnormal, use simple language: e.g., thick, thin, red,
pale.
Provide endoscopic photographs, if possible, especially
when the appearances are unusual.
HISTORY
Brief; one or two lines usually sufficient
Drugs, current or recent: e.g., antisecretory? bismuth?
nonsteroidal anti-inflammatory drugs?
immunosuppressives?
QUESTION FOR THE PATHOLOGIST
Very useful; be as specific as possible.
Reducing the Risk of Postbiopsy Bleeding

If multiple biopsies are to be taken or if procurement of snare biopsies is contemplated, it is desirable
to place the patient on potent antisecretory therapy for 1 week following the procedure. The basis of
this opinion is that multiple or single large (e.g., snare biopsy) ulcers are created. The antisecretory
therapy is intended in theory to reduce the risk of bleeding, or at the very least to promote faster
healing of these iatrogenic mucosal defects. The latter objective may be a more compelling reason for
such therapy in patients on nonsteroidal anti-inflammatory drugs (NSAIDs).
Description of Lesion or Biopsy Sites

The location of a lesion(s) or biopsy site is sometimes given in terms of centimeters from the incisor
teeth or mandibular ridge. This is well-intentioned but frequently not helpful and may even be
misleading as a method of specifying the part of the stomach sampled. For example, 50 cm from the
incisors could specify the esophagus of a basketball player or the small intestine of a jockey. Accurate
specifications for the location of biopsies are of considerable importance in a number of diseases and
clinical circumstances. In a later section on atrophic gastritis, for example, the critical importance of
such precision is discussed in detail.
In any multiuser endoscopy unit, agreed-on, standardized site descriptions for biopsy specimens are
desirable. The only way to ensure consistency is to entrust this task to the GI assistant. Our endoscopy
unit nurses know the glossary of standard locations. If an endoscopist gives a location not in the
lexicon, the nurse politely asks for a site description that conforms to the list. The location designations
we use are fundus, body (proximal, mid, or distal), and antrum (proximal, mid, or distal).7(3992)
Prepyloric may be substituted for distal antrum to denote a location within 2 cm of the pylorus. For the
body and antrum, the curvature (lesser or greater) is also noted. I prefer a curvature designation to a
wall location because agreement is often lacking concerning the locations of the anterior and posterior
walls of the stomach. Diagrams are a viable alternative to a lexicon of standard site descriptions,
provided that the diagram is combined with the final pathology report. However, conventions for the
use of a diagram must also be established for an endoscopy unit because the shape and size of the
individual patient's stomach seldom conform to a diagram that represents an idealized concept. An
indicated reference point on a diagram also suffers from the problem of subjective interpretation.
Diagrams may also be problematic with respect to electronic charting, but these limitations are easily
corrected with modern software.
Where to Obtain a Biopsy

For the diagnosis of H. pylori infection, biopsies should be confined to the mucosa within 1 to 2 cm
from the pylorus to ensure that antral gland mucosa has been sampled.
The oxyntic gland mucosa is thickest in the midbody on the greater curve of the stomach. Biopsies
should be obtained from this area when severe atrophic gastritis of the oxyntic mucosa is a
consideration. The most reliable way to approximate this location is to position the endoscope tip in the
region of the lower esophageal sphincter, inflate the stomach, and look for the area where the
cerebriform gastric body folds are thickest. The outline of the gastric wall appears to bend as if by
extrinsic compression in this region.
An appreciation of the process of mucous gland metaplasia (pseudopyloric metaplasia) is helpful in
avoiding biopsy sites that might produce a false diagnosis of severe atrophic gastritis. When severe
atrophic gastritis develops in the oxyntic mucosa, the oxyntic glands are replaced in part by two types
of metaplasia: intestinal and mucous gland. Popularly known in pathology circles as pseudopyloric
metaplasia, this mucous gland metaplasia is a process whereby the oxyntic glands disappear, partially
or completely, and are often replaced by mucous glands resembling those of the gastric antrum and of
Brunner's glands.8(3993)
In many studies of oxyntic gland mucosa, biopsies have been taken 10 cm distal to the lower
esophageal sphincter region in order to standardize the site. If a patient has a large hiatus hernia,
however, this location may be too far distal, so that biopsies may consist of mixed antral-oxyntic
mucosa. Such a biopsy with mixed gland elements and just superficial gastritis could legitimately be
interpreted as severe atrophic gastritis of the oxyntic mucosa with mucous gland metaplasia. The
lesser curve of the gastric body should also be avoided if the primary purpose is to determine whether
the patient has an oxyntic gland-predominant (type A) severe atrophic gastritis. This is because the
mucosa of the lesser curvature, as is evident at endoscopy, is thinner, lacking folds. Furthermore, the
mucosa for several centimeters proximal to the incisura may be "antralized" as part of the process of
gastritis or because the transition zone is more extensive than usual.
Biopsy Mapping
In some instances, it is desirable to obtain a detailed histologic picture of the entire stomach. My
system for this makes use of a diagram prepared beforehand (Figures 472 and 473). The scheme I
use most commonly for mapping is shown in Figure 472. The stomach is divided into five segments,
and biopsies from each segment are taken at equidistant intervals; additional biopsies are obtained
from any endoscopically visible target lesions. An alternative scheme for biopsy mapping is shown in
Figure 473.
(3994)Figure 472. Example of the diagram used at endoscopy for biopsy mapping of the
stomach. Five segments are designated, and biopsies from each segment are placed into
separate fixative bottles. The incisura and an imaginary line opposite it on the greater
curvature are taken as the junction between the antrum and the body. Most commonly, four
biopsies are taken in Segments 2 through 5, but more can be obtained in any segment of
special interest. Letters or numbers are preassigned for the specimen bottles that will contain
the biopsies from each corresponding segment. The diagram is kept within sight during
endoscopy, so the endoscopist can tell the gastrointestinal assistant which segment is being
sampled. This diagram is submitted with the specimen bottles and pathology requisition.
(Drawn by Chrissy Y. Hsieh.)
(3995)Figure 473. Alternative sketch for a protocol for biopsy mapping of the stomach.
Here a radial approach is used with four biopsies at each of three levels: midbody, incisura,
and midantrum. Any number of additional levels can be added. (Drawn by Chrissy Y. Hsieh.)
Focal Lesions
When a biopsy is being obtained from a focal lesion, a biopsy from adjacent, apparently uninvolved
mucosa may be extremely useful. This practice (1) proves that the lesion is either truly focal
histologically or part of a more diffuse process despite the focal appearance at endoscopy, (2) proves
that no associated histologic abnormality is present, or (3) provides a diagnosis from histopathologic
findings in the adjacent normal-appearing mucosa when biopsies from the focal lesion are inconclusive
or negative. When one suspects Crohn's disease or infection, for example, a biopsy of an erosion or
ulcer may reveal only exudate or granulation tissue, whereas a specimen from the adjacent mucosa
may provide more definitive findings such as the presence of granulomas or the typical cells of
cytomegalovirus infection.
Description of Lesions
The best nomenclature for the classification of endoscopic findings in gastritis consists of simple
descriptive terms that allow others to "visualize" the appearance of the mucosa. Too often
endoscopists describe not what they see but rather what they think a lesion represents. Thus,
describing a lesion as a patch of gastritis leaves open the possibilities that the endoscopist saw
erosions, bleeding, subepithelial hemorrhage, or the ubiquitous erythema! The increasing availability of
endoscopic systems that provide photographic prints and computer storage of digital images reduce
but do not eliminate some of this uncertainty.
For erosions, four pieces of information are needed: (1) size and numbers; (2) appearance of the base
(i.e., white or black); (3) appreciable depth, if any; (4) nature of the edge if applicable (e.g., rim of
erythema, raised). As discussed later, the endoscopist should not agonize excessively about whether a
lesion is an ulcer or an erosion. For other mucosal abnormalities, the shorter the descriptor, the more
likely others will be able to visualize the endoscopic findings. Examples of such simple words and less
desirable alternatives in parentheses are red (erythema), thin (atrophic), thick (hypertrophic), bumpy
(cobblestoned, nodular), blotchy (patchy erythema, patchy pallor), and tiny (diminutive).
Question(s) for the Pathologist

The most important way to ensure diagnostic accuracy and relevance in the histopathologic
assessment of endoscopic biopsy specimens is to provide the pathologist with focused questions. In
general, these reflect the thought process of the endoscopist in trying to reach a diagnosis. Even if a
definitive histopathologic diagnosis is not possible, the pathologist is frequently better able to offer an
opinion about the likelihood of one or more diagnostic possibilities. When alerted to the possibility of
Crohn's disease, for example, the pathologist might include this diagnosis in the differential diagnosis
of markedly patchy inflammation despite the absence of granulomas. When cytomegalovirus infection
is suspected, for example, the pathologist searches harder for evidence of this diagnosis if he or she is
told that the patient is immunocompromised as a result of disease or therapy.
Classification of Gastritis
Two different objectives exist with respect to the nomenclature of gastritis and gastropathy: (1) to
provide a verbal "snapshot" of the pathology (endoscopic and histopathologic) and (2) to indicate the
clinical associations and pathogenesis, if known. The Sydney system was an attempt to standardize
nomenclature of the pathology of gastritis.1,9(3996) However, the endoscopic component of this
system10(3997) has not been widely adopted, in part because attempts to classify gastritis
endoscopically are not likely to be workable,11,12(3998) as the correlation between endoscopic and
histologic appearances is often poor.
The overview of gastritis and gastropathy includes three main groupings that combine endoscopic and
histologic concepts (Table 472). These are (1) erosive and hemorrhagic, (2) nonerosive, and (3)
distinctive. The associated conditions and pathogenesis for each major group are discussed later.
TABLE 472
Overview of Gastritis and
Gastropathy
Erosive and hemorrhagic*
Nonerosive*
Distinctive
* The endoscopic and histologic picture is compatible with a
broad group of disease states.
Distinctive endoscopic and histologic features that are
diagnostic for a disorder or for a more limited group of
disorders.
Erosions and hemorrhages are endoscopically visible lesions; obtaining biopsies is not necessary
unless a distinctive type of gastritis is suspected, especially opportunistic infection in the
immunocompromised patient.
Nonerosive gastritis is a histologic diagnosis; endoscopic appearances do not predict the presence,
absence, or severity of the mucosal injury. In nonerosive gastritis, the mucosal injury is nonspecific;
this histologic diagnosis does not predict any associated clinical disorder. For example, in H. pylori
gastritis, the patient may have a peptic ulcer or carcinoma or may be entirely asymptomatic.
Distinctive types of gastritis and gastropathy are generally uncommon except in immunocompromised
patients. Their endoscopic and histologic appearances are distinctive enough to narrow the differential
diagnosis to the point that it has a bearing on therapy or prognosis or both. Examples of distinctive
types of gastritis are cytomegalovirus infection and Crohn's disease.
Erosive and Hemorrhagic Gastritis and Gastropathy

Erosions
The classic definition of an erosion is a shallow defect in the mucosa that does not extend through the
muscularis mucosae into the submucosa (Figure 474). A typical lesion is a flat or minimally
depressed tiny white spot surrounded by a red halo. These aphthous-type lesions are usually multiple
(Figure 475). When viewed closely, many erosions actually have a whitish dome rather than a
depressed base; the dome is probably due to the presence of a pseudomembrane, a feature often
seen in biopsies of erosions. If bleeding from an erosion has occurred recently, its base may be black
(Figure 476). Arbitrary criteria have been proposed to distinguish erosions from ulcers.13(3999) One
hallowed criterion is that of depth; simply stated, a lesion that has depth or is depressed must be an
ulcer. No evidence indicates that this criterion is indicative of penetration through the muscularis
mucosae into the submucosa. Of necessity, therefore, the distinction between erosion and ulcer is
often arbitrary. For example, many endoscopists are inclined to call an 8-mm-diameter, nondepressed,
white-based lesion a shallow ulcer and a depressed 2-mm lesion an erosion. This understandable
limitation is reflected in a study of interobserver variation, in which only 55% agreement was found
concerning the types of lesions present in patients taking NSAIDs.14(4000)
(4001)Figure 474. Sketch illustrating the difference between an erosion and an ulcer. An
erosion is a mucosal break that does not penetrate the muscularis mucosae, whereas an ulcer
does.
(4002)Figure 475. Multiple white-based aphthous gastric erosions surrounded by red halos.
(4003)Figure 476. Black-based gastric erosion in a patient with hematemesis.
Subepithelial Hemorrhages
The term subepithelial hemorrhage refers to the endoscopic appearance of discrete petechiae or fire
engine-red confluent streaks and patches unassociated with any visible breaks in the mucosa (Figure
477). Simply put, these lesions look like blood under a plastic wrap.15(4004) The term submucosal
hemorrhage is frequently used to describe these lesions, but this term is incorrect conceptually, as it is
impossible to see into the submucosa. Subepithelial hemorrhage is a better descriptive term because it
leaves open the possibility that the lesion is located in either the mucosa or the submucosa.
(4005)Figure 477. Diffuse subepithelial hemorrhages.

Endoscopic findings that do not qualify as either erosions or hemorrhages include the following:
erythema, mucosal swelling, friability (petechiae or blood ooze) induced by passage of the endoscope
over a segment of mucosa, and cherry-red spots resembling vascular lesions. Why is it necessary to
cite these examples when general agreement exists that an erosion represents a break in the
mucosa? The reason is that the term erosion is often used loosely in both clinical practice and
research studies. For example, in one series of studies, prepyloric erythema was considered to be one
type of erosive prepyloric change.16(4006) Another descriptor open to a variety of interpretations is
hemorrhagic erosions. Endoscopists who use this term are usually referring not to erosions but to a
variety of other possibilities, including erythema, subepithelial hemorrhage, and contact bleeding after
passage of the endoscope over an area.
Distinguishing between a subepithelial hemorrhage and erythema or a circumscribed vascular ectasia
is sometimes difficult. Furthermore, even an extensive area of apparent subepithelial hemorrhage may
on occasion represent a submucosal vascular ectasia with only minimum involvement of the overlying
mucosa (Figure 478). The apparent prominence of submucosal involvement in vascular ectasia, such
as the watermelon stomach, can be documented by endoscopic ultrasonography.17,18(4007) Prior to
extensive coagulation treatment, biopsies are sometimes obtained of lesions that appear vascular to
verify the nature of the lesion and to eliminate the possibility of an unusual or unexpected diagnosis
such as diffuse lymphoma.19,20(4008) In up to 50% of cases, however, verifying a vascular lesion on
biopsy is impossible, either because of vessel shrinkage or because of the lesion's location
predominantly in the submucosa (Figure 478).7(4009)
(4010)Figure 478. Submucosal vascular ectasia. A, Broad red streaks in the antrum radiate
to the pylorus. B, Antrectomy from A. Extensive submucosal vascular ectasia (arrow) is
present with only minimum hemorrhage in the overlying mucosa.
Biopsy Histology of Erosions and Hemorrhages

Obtaining biopsies from erosions and hemorrhages is rarely indicated in clinical practice, especially
those found in patients with clinical conditions commonly associated with this type of lesion, such as
the critically ill and those taking NSAIDs. Biopsies are usually obtained to rule out opportunistic
infection and much less commonly to obtain histologic verification that the lesion is an erosion. If
infection (e.g., with cytomegalovirus) is suspected, tissue should also be taken for cultures; biopsies
from the mucosa next to visible lesions should also be obtained for histopathologic assessment.
Biopsies from ulcerated lesions may show only exudate, whereas intact mucosa at the edge (within 0.5
cm) of the lesion may reveal evidence of the infection.
Erosions commonly have a pseudomembrane overlying a hyaline base.8(4011) The most consistent
finding in biopsies from the edges of erosions is reactive gastropathy (Figure 479), an intense
epithelial change commonly accompanied by foveolar hyperplasia. Inflammatory cell infiltrates are few
or absent. An endoscopic biopsy of an apparently typical erosion or hemorrhage often yields entirely
normal histology. In such a case, marked reactive gastropathy may be present, suggesting that the
erosion is nearby, either in unsectioned tissue within the paraffin block or in the stomach because of
mistargeting of the biopsy at endoscopy.
(4012)Figure 479. Histology of an erosion. A, Low power shows a shallow erosion with pink
hyaline material at the luminal surface. B, High power shows marked epithelial abnormalities
in the base of this erosion. Cells vary in size, and chromatin patterns in the nuclei vary in
density. Note the absence of inflammatory cells.
The hemorrhage is usually in the interpit regions, and edema may be prominent in adjacent mucosa in
biopsies of subepithelial hemorrhages from actively drinking alcoholics. Again, the inflammatory
infiltrate may be slight.8,21(4013) Biopsies from fully developed erosions or hemorrhages may reveal
partial or full-thickness mucosal necrosis with minimum inflammation (Figure 4710). If subepithelial
hemorrhages extend half the distance or more across the span of a biopsy, we term this hemorrhagic
gastropathy. If hemorrhage covers less than this distance, we do not make this histologic diagnosis
because of the possibility that the trauma of the pinch biopsy induced the localized
hemorrhage.8,21(4014)
(4015)Figure 4710. Histology of an erosion. Contrast with Figure 479. Extensive mucosal
hemorrhage is present, and virtually no residual glands are seen.
In some instances, red swollen areas of mucosa raise the question of mucosal congestion, as in portal
hypertension. The presence of mucosal congestion may be missed histologically because the trauma
induced by the pinch biopsy forceps may have converted the congestion to subepithelial hemorrhage.
Clinical Settings
The main clinical settings in which nonspecific erosions and hemorrhages may be found are outlined in
Table 473. An explanation for the presence of erosions and/or subepithelial hemorrhages is almost
always possible. When these findings are encountered unexpectedly, NSAID use is by far the main
cause. Much less common causes include alcohol ingestion and a host of other causes, including
prolapse injury. A small number of patients seem to have gastric erosions, usually in the antrum, with
no explanation. In assessing statements that erosions occur in a given condition, one should always
determine whether the investigators attempted to exclude other causes, especially NSAIDs.
Nevertheless, it is difficult to know how to place certain entities within the framework of this discussion.
For example, graft-versus-host disease may be associated with nonspecific erosions and sometimes
may also exhibit the apoptotic lesions that are seen typically in the colon.22(4016) These lesions could
be due to graft-versus-host disease, they could be related to the stress of this serious illness, or they
could be due to irradiation or chemotherapy if one of these forms of treatment preceded
transplantation.
Associations and Clinical Settings for

Erosive and Hemorrhagic Gastritis and Gastropathy
TABLE 473
STRESS LESIONS IN SERIOUSLY ILL PATIENTS

Organ system failure
Burns
Central nervous system trauma
DRUGS
Nonsteroidal anti-inflammatory drugs
Alcohol
Corrosive ingestion
Other: potassium chloride, iron, chemotherapy, cocaine, etc. (see text)
TRAUMA AND PHYSICAL AGENTS
Mechanical and interventional
Nasogastric tubes, retching
Endoscopic hemostasis: lasers, diathermy, thermal, sclerosis
Foreign body ingestion
Radiation
VASCULAR
Ischemia, distinct types: e.g., embolism (therapeutic, cholesterol),
vasculitis, and others (see text)
Congestive gastropathy
REFLUX INJURY
Duodenogastric reflux: postgastrectomy
Gastroesophageal reflux: cardia region inflammation
INFECTIONS
Epidemic(?) nephropathy
IDIOPATHIC
Chronic erosive gastritis (varioliform gastritis)
Sporadic, incidental
Stress Lesions
Erosions and diffuse subepithelial hemorrhages are very common in seriously ill patients in intensive
care units after massive burns and head injuries.2327(4017) The subepithelial hemorrhages and
erosions often cluster in the fundus and body rather than in the antrum, especially early in the course of
a patient's illness. The lesions that occur in burned patients, especially those with massive burns, have
a similar pattern, although the antrum seems to be involved more commonly. Erosive duodenitis and
duodenal ulcers may be more common accompaniments than in patients comparably ill from other
causes.28,29(4018) However, when lesions are present in the antrum or duodenum, they are always
accompanied by lesions in the gastric body.
Frank GI bleeding develops in only a small percentage of patients with stress-related erosions and
subepithelial hemorrhages:30(4019) When massive bleeding does occur, it often results from discrete
ulcers or other lesions. Mucosal ischemia is the likely cause of stress lesions.31,32(4020) Improved
cardiorespiratory support is probably the main reason that patients in intensive care units do not bleed
massively.
Drugs and Chemical Agents

Nonsteroidal Anti-Inflammatory Drugs
The erosions and subepithelial hemorrhages seen in patients taking NSAIDs are clinically
unimportant.13,33(4021) They are endoscopic findings rather than clinically significant lesions.
NSAID-induced erosions and subepithelial hemorrhages rarely if ever bleed, and the abdominal pain
noted by patients taking drugs of this class bears little relationship to the presence of these lesions.
They tend to occur with greater frequency in the fundus and body, especially during the early period
following the initiation of NSAID use; later, more lesions are seen in the antrum. If ulcers develop, they
are much more commonly located in the gastric antrum. Reports concerning the prevalence of gastric
lesions in patients taking NSAIDs usually present data in terms of numbers of ulcers and erosions.
Given the limitations of distinguishing between these two lesions at endoscopy, especially smaller
lesions, an estimate of size may be more useful than a subjective impression of which type of lesion is
present.
The presence of reactive gastropathy in patients taking NSAIDs is an unresolved issue. We have
found that reactive gastropathy (see Figure 479) does not occur as a diffuse lesion but is found only
at the edges of erosions and ulcers and only occasionally on random biopsy.34,35(4022) Others find
evidence of reactive gastropathy in apparently normal-appearing mucosa (not always specified) with
variable frequency (e.g., 2 to 45%).3638(4023)
Alcohol
Erosions and hemorrhages are not common causes of upper GI tract bleeding in alcoholics without
portal hypertension who are consuming large amounts of alcohol.21(4024) The lesions are localized,
and any surrounding histologic gastritis is due to accompanying H. pylori.39,40(4025)
Corrosives
Corrosive injury to the stomach can occur with ingestion of either strong alkalis or acids. The latter is a
more common cause of gastric injury, especially in the antrum. It is important to perform endoscopy
early in the patient's course to determine (stage) the degree of injury, along with computed tomography
(CT) of the abdomen.41(4026) The finding of a black gangrenous slough generally indicates the need
for early surgery.42(4027) Fortunately, surgery is not required in most cases.43(4028)
Other Drugs
Clinical trials of oral iron and wax-matrix potassium chloride preparations have demonstrated the
development of erosions and hemorrhages in some volunteer subjects.44,45(4029) Chronic fluoride
ingestion for otosclerosis was found to be associated with antral erosions.46(4030) Unexpected
erosions or ulcers might be explained on the basis of cocaine abuse,47,48(4031) although the large
intestine appears to be more susceptible.49(4032) The basis for the cocaine-associated lesions
appears to be ischemic. Hepatic arterial chemotherapy and an increasingly long list of other types of
chemotherapy may produce gastric erosions and large ulcers.5053(4033) In the evaluation of patients
with abdominal pain or bleeding who are undergoing chemotherapy, always consider the possibility that
observed lesions may represent opportunistic infection, especially with cytomegalovirus. Given the
seemingly ubiquitous nature and variety of toxic chemicals, herbicides for example,54(4034) it is
surprising that reports of gastric mucosal injury are not more numerous.
Localized Gastric Trauma

In patients with GI bleeding who have had nasogastric tubes in place, determining whether mucosal
hemorrhages were induced by the suction trauma of the nasogastric tube is sometimes impossible. If
the hemorrhages are arranged in a linear or geometric pattern, nasogastric tube trauma becomes a
more likely cause.
Retching prior to endoscopy or during endoscope intubation may induce dramatic-appearing showers
of petechiae or larger patches of subepithelial hemorrhage in the fundus (Figure 4711). Prolapse
gastropathy refers to the prolapse of a knuckle of gastric mucosa proximally into the distal esophagus
because of repeated retching and vomiting (Figure 4712).55,56(4035) If a hiatus hernia is pres-ent,
the subepithelial hemorrhage is located at the level of the diaphragmatic constriction. Sometimes the
whole fundus is covered with petechiae or hemorrhages, presumably via some mechanism other than
prolapse. The extent and number of subepithelial hemorrhages due to any kind of localized gastric
trauma may be greater in the presence of a coagulopathy.
(4036)Figure 4711. Marked subepithelial hemorrhage in the gastric fundus of a patient with
thrombocytopenia. Retroverted view (left); closer view of the hemorrhages (right).
(4037)Figure 4712. Prolapse of the proximal stomach into the lower esophagus. Close (left)
and distant (right) views.
Pressure trauma from the diaphragm is thought to be the mechanism for the erosions or ulcers found
in patients with large diaphragmatic hernias, especially the fixed paraesophageal type.57(4038)
Endoscopic techniques used to obliterate vascular lesions create erosions or ulcers. In a prospective
study of 40 patients undergoing extracorporeal shock wave lithotripsy for renal lithiasis, 80% of patients
developed gastric or duodenal erosions, mainly in the proximal stomach.58(4039)
Radiation
Radiation may induce gastric erosions and ulcers and result in the delayed formation of antral
strictures.59,60(4040) If erosive gastritis or ulceration is found in patients who have recently undergone
irradiation, opportunistic infection should be considered.
Ischemia
Isolated reports have been made of apparent ischemic erosions due to diverse rare causes, including
atheromatous embolization,61(4041) small vessel vasculitis,62(4042) cytomegalovirus

infection,63(4043) Henoch-Schnlein purpura, 64(4044) and amyloidosis.65,66(4045) It is difficult to
prove that persistent gastric erosions and ulcers are due to ischemia on the basis of atherosclerosis,
except for rare instances associated with revascularization operations.67(4046) The occurrence of
hemorrhagic gastropathy in cases of epidemic nephropathy may also be due to ischemia.68(4047) A
patient 70 years of age with chronic, unremitting gastric mucosal hemorrhage has been
described.69(4048) A full-thickness gastric biopsy disclosed capillary and postcapillary venule luminal
narrowing within the lamina propria, swelling and proliferation of endothelial cells, and fibrin thrombi
within some vessels. Fishbein et al.69(4049) suggested that these findings represent a new entity
(diffuse hemorrhagic gastroenteropathy), possibly due to regional small vessel disease and resultant
mucosal ischemia. Gastric erosions and subepithelial hemorrhages have been described in a small
group of individuals after vigorous running.7072(4050)
Congestive Gastropathy
The endoscopic appearance of congestive (portal hypertensive) gastropathy may be characterized by
the presence of subepithelial hemorrhages or vascular ectasia.73,74(4051) Other features include a
"snakeskin" or mosaic appearance in the mucosa of the body of the stomach and discrete red spots in
the antrum. The snakeskin appearance is often found in individuals without portal hypertension and
seems to represent an exaggeration of the area gastricae, a normal finding. In patients with severe
liver disease, severe coagulopathy is a major contributor to the development of petechiae and
hemorrhages. Congestive (portal hypertensive) gastropathy is discussed in Chapter 31: Variceal
Bleeding.
Postgastrectomy
Abnormalities of the gastric mucosa after gastric resection are considered in Chapter 52: Endoscopy in
the Postoperative Upper Gastrointestinal Tract. Erosions, especially near the stoma, may be seen in
both symptomatic and asymptomatic patients after surgery for peptic ulcer.75(4052)
Diffuse Varioliform Gastritis (Chronic Erosive Gastritis)

Diffuse varioliform gastritis is rare in North America, but some European centers report large numbers
of cases.76(4053) As originally described by Lambert et al.,76(4054) the fundus and body are involved
in the majority of cases, sometimes with extension to the antrum. The symptoms are nondescript:
abdominal pain, nausea, and vomiting. Weight loss and GI bleeding are much less common.
Symptoms may fluctuate out of synchrony with the waxing and waning gastric lesions. For erosions to
be considered idiopathic, the patient should have no history of known causes such as NSAID ingestion
or alcohol abuse. Information concerning the exclusion of known etiologic factors is not found in all
reported studies. A variety of therapies have been tried with variable success;77(4055) one report
describes the use of misoprostol as the most promising.78(4056)
The typical endoscopic features are those of small nodules (<1 cm diameter) in the gastric fundus and
on the crests of rugal folds in the body. Many of the nodules have central umbilications, some with a
whitish base (Figure 4713). The nodular lesions tend to be more numerous along the greater
curvature. Histologically, the appearance of the erosions of diffuse varioliform gastritis is the same as
that described earlier, including the presence in some cases of pseudomembranes.79(4057) Biopsies
are sometimes obtained because of a concern for the presence of an infiltrating lymphoma.
(4058)Figure 4713. Diffuse varioliform gastritis. Multiple white-based erosions are seen at
crests of prominent folds in the body of the stomach.
The "signature" histologic lesion of the gastric mucosa in diffuse varioliform gastritis is lymphocytic
gastritis.80,81(4059) This is an intense infiltrate of the surface or pit epithelium or both with
lymphocytes. The underlying lamina propria may be infiltrated by inflammatory cells, but the infiltrate is
usually not as profound as that of the overlying epithelium. H. pylori is present in half or fewer of cases
and does not seem to be important in pathogenesis.82,83(4060) Lymphocytic gastritis has
subsequently been recognized in association with other disorders, including celiac disease and
Mntrier's disease (see later).
The definition of diffuse varioliform gastritis has become problematic, as some investigators include
cases with antral lesions. Others use the term only in reference to antral lesions. The antrum-exclusive
variant is probably different pathogenetically.81,83(4061) Antral bumps, sometimes with tiny central
erosions, may be seen in patients taking NSAIDs.
The criteria I use for the diagnosis of diffuse varioliform gastritis are (1) the presence of tiny nodules,
some with central depressions or erosions; (2) involvement of the body and antrum or the body alone;
(3) the presence of lymphocytic gastritis; and (4) 30 lymphocytes per 100 epithelial cells or an
obviously intense infiltrate. For research purposes, this fourth criterion should include the analysis of
biopsies from control subjects.
Nonerosive Gastritis
Nonerosive gastritis, also known as chronic gastritis, is a diagnosis based on the histology; the
endoscopic appearances do not accurately predict whether mucosal inflammation is present. H. pylori
gastritis is the most common type of nonerosive gastritis.
Endoscopic Appearances
Endoscopists continue to find it difficult to accept the fact that the mucosal patches of red or white,
granularity, cobblestoning, and thin or thick rugal folds usually have little predictive value with regard to
histologic findings.84,85(4062) Endoscopic chromoscopy, with dyes such as Congo red and methylene
blue, has been used mainly in Japan in large surveys of patients with nonerosive gastritis and various
disease associations.86(4063) These techniques are discussed in Chapter 12: Chromoscopy.
The most important type of nonerosive gastritis is H. pylori gastritis. It is considered a nonspecific type
and not part of the group of distinctive gastritides because the demonstration of its presence does not
predict which if any associated disorder may be present. For example, a patient with H. pylori gastritis
may have an associated duodenal or gastric ulcer or may have no associated ulcer and, in the majority
of cases, may be asymptomatic.
General Histologic Patterns

Nonerosive gastritis may be confined primarily to the antral gland mucosa or the oxyntic gland mucosa,
or it may occur in both areas. With H. pylori gastritis, both zones are involved, but the antral zone is
more severely affected.
Nonerosive gastritis is traditionally divided into superficial (nonatrophic) and atrophic, depending on its
severity. In a given patient, both patterns may be present. In superficial gastritis, inflammatory cells are
confined to the regions of the pits or foveolae (Figure 4714). Atrophic gastritis is characterized by
variable degrees of gland loss and associated encroachment of inflammatory cells into the gland
zones. The severity of the superficial or atrophic gastritis is often further graded as mild, moderate, or
severe. Severe atrophic gastritis represents full-thickness mucosal involvement with disappearance of
much of the gland mass.
(4064)Figure 4714. A, Biopsy showing severe superficial antral gastritis. Numerous

inflammatory cells are seen in pit (foveolar) regions. The clear-staining antral glands are
preserved. Contrast with B. B, Normal antral gland biopsy. The arrow spans some "edema"
and hemorrhage, common findings in pinch biopsies.
Intestinal and Mucous Gland (Pseudopyloric) Metaplasia
Two types of metaplasia may occur in nonerosive gastritis. The more common is intestinal metaplasia.
This is characterized by the presence of intestinal-type goblet cells and even absorptive cells in either
antral or oxyntic mucosa. When intestinal metaplasia occurs right at the lower esophageal sphincter
region, it may represent a response to chronic gastroesophageal reflux.87(4065) In severe cases of
atrophic gastritis, intestinal metaplasia may replace the full thickness of the mucosa, with minimum
residual inflammation, making the mucosa virtually indistinguishable from that of the small bowel.
Mucous gland (pseudopyloric) metaplasia is seen in severe atrophic gastritis of the oxyntic gland
mucosa, as in pernicious anemia. Mucous glands may replace some of the oxyntic glands that have
disappeared. This mucous gland metaplasia is known more commonly by the less user-friendly term
pseudopyloric metaplasia.
Lipid Islands (Gastric Xanthelasma)
Lipid islands are lesions that represent idiosyncratic reactions to nonerosive gastritis. Endoscopically,
they are flat or slightly raised white or yellow-white lesions that range in size from pinpoint to several
millimeters in diameter (Figure 4715A). The frequency with which they are encountered is highly
variable.8890(4066) In some series, lipid islands are found in more than 50% of postgastrectomy
patients.89(4067) Histologically, these lesions consist of accumulations of foamy, fat-laden histiocytes
(Figure 4715B). Oxidized low-density lipoprotein seems to be important in the development and
maintenance of these curious lesions.91(4068) Occasionally, lipid islands are erroneously diagnosed
histologically as signet-ring carcinoma, or conversely this carcinoma may be diagnosed incorrectly as
lipid islands.92(4069)
(4070)Figure 4715. Lipid islands (gastric xanthelasma). A, Endoscopic appearance is two

tiny white bumps. B, Biopsy of one of the bumps. The lamina propria is filled with ballooned
clear-staining fat-laden histiocytes (periodic acid-Schiff [PAS] stain).
The main associations or types of nonerosive gastritis are given in Table 474. These are H. pylori
gastritis, reactive gastropathy, lymphocytic gastritis, and pernicious anemia. No association is evident
in some cases of nonerosive gastritis.
TABLE 474
Nonerosive Gastritis Types and
Associations
H. pylori gastritis
Reactive gastropathy
Lymphocytic gastritis
Pernicious anemia
Chronic active or chronic gastritis with no known associations
Therapy with drugs that eradicate H. pylori
Sampling error*
Crohn's disease
Acquired immunodeficiency syndrome
TABLE 474
Nonerosive Gastritis Types and
Associations
* Inherent or sparse organisms because of therapy with
antisecretory drugs.
H. Pylori Gastritis
The endoscopic and biopsy diagnoses of H. pylori gastritis are considered along with a brief overview
of its clinical implications.9398(4071)
H. pylori is the most common cause of nonerosive gastritis. Most patients with this infection remain
asymptomatic throughout their lifetimes. However, approximately 10% develop peptic ulcer, this being
the most important clinical implication of H. pylori infection known to date. The eradication of H. pylori
infection in patients with peptic ulcer dramatically reduces the frequency of relapse. The prevalence of
H. pylori in patients with duodenal ulcer is greater than 95%; for those with gastric ulcer, it is
approximately 80%. For patients with NSAID-associated peptic ulcer, however, the prevalence of H.
pylori is about 50%.
Carcinoma and Mucosa-Associated Lymphoid Tissue Lymphoma
Two areas of intense interest with regard to H. pylori are its relations to the pathogenesis of gastric
carcinoma and to B-cell gastric lymphoma, known as mucosa-associated lymphoid tissue lymphoma,
or more familiarly by its abbreviation, MALToma.99(4072) H. pylori gastritis is the initial risk factor for
gastric carcinoma in some patients, although other factors (environmental, genetic) must be present
before malignancy develops.93(4073) In population-based studies, intestinal metaplasia is associated
with an increased risk of gastric adenocarcinoma. The risk for the individual patient, however, is
sufficiently low that endoscopic cancer surveillance is not justified when this condition is present alone.
Approximately 75% of patients with low-grade MALToma have H. pylori; three quarters of these have
complete or nearly complete regression over the short term.100,101(4074) The expected survival of
patients with low-grade MALToma is greater than 90% at 10 years after treatment with surgery or
chemotherapy.102(4075) Therefore, any claim that H. pylori eradication cures this lymphoma must be
verified by data at 10-year follow-up.
Diagnosis
Endoscopic biopsy is still the most commonly used technique for the diagnosis of active H. pylori
infection. In actual clinical practice, some endoscopists obtain biopsies for histology and for immersion
in one of the rapid urease test wells that are available commercially. If the color indicator turns positive,
the biopsies for histology are discarded. The idea of discarding biopsies intentionally seems heretical,
but it is difficult to challenge if there is no other reason, a suspicion of neoplasia for example, to submit
biopsies for histologic examination.13C and 14C-urea breath tests are a very sensitive noninvasive
way to diagnose active infection.103(4076) In uncomplicated cases, the commercial availability of a
nonradioactive 13C breath test is certain to replace endoscopic biopsy in large numbers of patients.
Endoscopic Appearances
The only two reproducible findings of H. pylori gastritis in an endoscopic study of healthy volunteers
were antral nodularity and prominent area gastricae.85(4077) However, the sensitivity of endoscopy
was very low. In children, nodularity has been reported to be more prominent, presumably owing to the
presence of lymphoid follicles.104(4078) In adults, nodularity is sometimes prominent when atrophic
gastritis of the antrum is present, this being similar in appearance to the flat mucosa that occurs in the
duodenal bulb in patients with celiac disease. In my experience, nodularity, unless riveting in its
appearance, is often not predictive of any distinctive histologic finding(s) (Figure 4716).
(4079)Figure 4716. Bumpiness or nodularity in the gastric antrum of patient with H. pylori
gastritis after spraying with indigo carmine. Histology revealed only superficial H. pylori
gastritis without lymphoid follicles; thus no explanation for the endoscopic appearance.
A report from Brazil describes seven patients who appear to have had acute H. pylori
gastritis.105(4080) Erosions and swollen folds, especially in the antrum, were present at endoscopy.
Treatment with antibiotics led to symptomatic, endoscopic, and histologic improvement.
Helicobacter heilmanii may be associated with erosive gastritis.106(4081) Similarly, other spiral
organisms transmitted from house pets may also produce endoscopic lesions.107(4082)
The idea that H. pylori eradication shrinks thick gastric folds is gaining currency in the literature. This
inclination must be regarded with caution. Soon after endoscopy became more widely available in the
1970s, there was great interest in thick folds, and snare excision biopsies were obtained in many
patients.108(4083) Interest waned when endoscopists became aware that this endoscopic finding
disappeared in most cases with air insufflation! Many cases of "giant" gastric folds are being
reexamined for the association with H. pylori.109(4084) One case of a Mntrier's disease-like
condition improved with H. pylori eradication.110(4085) Patients with putative thick folds have
undergone eradication of H. pylori, with claims of success in shrinking the folds.111(4086) There may
be something to all of this, but long experience teaches that it is essential, at a minimum, to precisely
define "thick" folds and to study the reproducibility of this condition in patients and controls in a blinded,
prospective fashion.
Endoscopic Biopsy
For optimum accuracy in the endoscopic/histologic diagnosis of H. pylori gastritis, two biopsies should
be obtained from the prepyloric region within 1 to 2 cm of the pylorus and 2 cm from the body of the
stomach. As discussed earlier, antral biopsy sites should be restricted to the immediate prepyloric area
because much of the proximal antrum can consist of pure oxyntic gland mucosa. With regard to the
body, I take two biopsies from the midbody on the greater curvature, where the oxyntic glands are
normally thickest. When the biopsies are submitted, it is useful to advise the pathologist whether the
patient is or has undergone therapy with proton pump inhibitors, antibiotics, or bismuth.
Although diagnostic accuracy is nearly 100% with antral biopsies alone, additional biopsies from the
body ensure that the diagnosis is not overlooked if the organisms are sparse, there is intestinal
metaplasia of the antral mucosa, or the patient is taking potent antisecretory drugs. In the latter
circumstance, as discussed later, the number of organisms seems to diminish significantly in the
antrum and increase in the body.112(4087)
Histology
The H. pylori organisms are present on the surface epithelium and within the pit lumina.113(4088) The
density may be highly variable among biopsies and in different areas of a single specimen. We use
special and super stains. The special stain (30 sec Giemsa) is obtained on all gastric biopsies to
facilitate recognition of the organisms. When these are sparse or there is gastritis but no apparent
organisms, we use a super stain such as Warthin Starry, or preferably the more colorful variant (Figure
4717) described by Genta et al.114(4089)
(4090)Figure 4717. High-power photomicrograph of a Genta-stained gastric biopsy showing

numerous brown-black organisms on the surface of irregular (cupped) gastric mucous cells.
The severity of H. pylori gastritis is usually greater in the antral gland mucosa.113(4091) The
inflammatory cell infiltrate is a mix of mononuclear cells and neutrophils, hence the designation chronic
active. Lymphoid follicles are commonly found,115,116(4092) especially in children.104(4093)
Occasionally they are so prominent that the term follicular gastritis is used to describe the appearance.
This descriptive term should not be used as a diagnosis. Changes in the overlying epithelium consist of
mucin depletion, loss of nuclear polarity, and nuclear enlargement; severity ranges from mild to very
severe. The epithelium has a serrated or cupped appearance due to a loss of apical mucus (see
Figure 4717). The organisms do not reside over any areas of intestinal metaplasia, a common
histopathologic condition in atrophic gastritis.
After H. pylori eradication, the neutrophilic infiltrate and the surface epithelial abnormalities regress
fairly promptly.117,118(4094) The mononuclear cells may persist for many months, albeit in reduced
numbers. A curious phenomenon has been described in patients undergoing therapy with omeprazole.
Kuipers et al.119(4095) found that the density of organisms decreased in the antrum and increased in
the gastric body, together with parallel changes in the degree of gastritis in the two gland zones.
Chronic or Chronic Active Gastritis Without H. Pylori
Chronic or chronic active gastritis in the absence of H. pylori may be due to sampling error, either
inherent or the result of the sparseness of organisms. The latter may be due to recent ingestion of
antimicrobial agents or bismuth, either as formal therapy or for other reasons, or to ingestion of potent
antisecretory drugs. Persistent H. pylori-negative gastritis may occur when efforts at eradication have
been successful but the inflammatory response has not yet subsided. Other causes include Crohn's
disease (see later) and the acquired immunodeficiency syndrome (AIDS).120,121(4096)
Nonulcer Dyspepsia
Nonulcer dyspepsia is a dilemma.122125(4097) Patients with this condition often have persistent
symptoms despite eradication of H. pylori. Unfortunately, conflicting results are inevitable when
patients with nonulcer dyspepsia are subjected to rigorous study.124(4098) This is because nonulcer
dyspepsia is very likely a group of diseases with a host of different pathogenetic mechanisms.
Unfortunately, no method exists yet for identification of the subset of patients with symptoms that might
be due to H. pylori infection. This problem is all the more daunting because millions of people are
infected but seemingly remain asymptomatic throughout their entire lives. Given the lack of consistent
data, the role of endoscopy in nonulcer dyspepsia remains to be redefined.
Reactive Gastropathies
Reactive gastropathy is a pattern of mucosal injury in which the dominant feature is epithelial change
with minimum or no inflammatory cell infiltrates. The mucosa exhibits foveolar hyperplasia, often
marked (Figure 4718), and the epithelial cells are very reactive, with mucin depletion, large nuclei,
and prominent chromatin patterns (see Figure 479B).1(4099)
(4100)Figure 4718. Reactive gastropathy in a biopsy from a gastric stoma after antrectomy.
The center shows the typical foveolar hyperplasia that is part of the picture of reactive
gastropathy. Intestinal metaplasia (blue cells at left) and dysplasia (dark cells at right) are also
seen.
The main clinical settings for reactive gastropathy are erosions and ulcers of any cause8,15,126(4101)
and gastric surgery. Reactive gastropathy may be found in the mucosa immediately adjacent to these
lesions (see earlier under "Erosions and Hemorrhages").
Postoperative Stomach
Erythema and bile-stained fluid are common, probably inevitable, in the postoperative stomach. These
color changes, along with erosions (in 20% of cases), are most prominent at the stoma and bear little
relationship to the histologic findings or to symptoms.75(4102) For decades, endoscopists have
labeled the red bile-stained gastric remnant bile gastritis. However, even early studies demonstrated
that there was often minimum to no inflammation in the postoperative stomach, especially near the
stoma.127,128(4103) The term reactive gastropathy appropriately describes the appearance of the
gastric remnant, especially at the stoma (see Figure 4718). In addition to the reactive gastropathy,
mucosal edema and cysts may be present.128,129(4104) Sometimes the cysts become large enough
to be visible grossly and to extend into the submucosa. These cysts have been given a variety of
names, including gastritis cystica polyposa and gastritis cystica profunda.130(4105)
The ill-defined syndrome of alkaline reflux gastritis is discussed in Chapter 52: Endoscopy in the
Postoperative Upper Gastrointestinal Tract. Biopsy has no role in establishing or excluding this
diagnosis.
In earlier surveillance studies of the postoperative stomach, dysplasia was commonly diagnosed. It is
now recognized that many of the cases of so-called mild dysplasia were actually reactive
gastropathy.128(4106) Endoscopic biopsy surveillance for carcinoma is not warranted in low-risk
populations.131(4107) If surveillance is carried out in a high-risk patient (e.g., with strong family history
of carcinoma), a mapping scheme (see Figures 472 and 473) should be used, as this maximizes the
chance of finding dysplasia or intramucosal carcinoma.
Lymphocytic Gastritis
Lymphocytic gastritis is discussed in the previous section on diffuse varioliform gastritis. The lining
epithelium in lymphocytic gastritis looks similar to the lining epithelium of the duodenum in celiac
disease. Lymphocytic gastritis appears to be a novel form of injury, often H. pylori negative, that is
associated with seemingly diverse disorders, including diffuse varioliform gastritis, one third of cases of
celiac disease, and some cases of Mntrier's disease.80,81,132,133(4108) It may be present without
apparent explanation. For research purposes, the current criterion for histologic diagnosis is a
lymphocyte count of greater than 30 lymphocytes per 100 epithelial cells. Reports that claim an
association with gastric cancer include no reproducible definition of lymphocytic gastritis.134(4109)
Pernicious Anemia and Severe Atrophic Gastritis Without Pernicious Anemia

Pernicious anemia is a megaloblastic anemia with absolute achlorhydria. Vitamin B12 malabsorption
(documented by the Schilling test) is due to severe atrophic oxyntic gland gastritis, which diminishes
the ability to secrete intrinsic factor. More common than patients with frank pernicious anemia are
those patients, usually elderly, who have diffuse severe atrophic fundic gland gastritis with achlorhydria
but a residual ability to absorb vitamin B12.135,136(4110) Some of these patients may develop
pernicious anemia in the course of time. These cases are often diagnosed inadvertently because
Zollinger-Ellison syndrome is suspected on the basis of hypergastrinemia, serum gastrin
determinations having been obtained for a variety of reasons, including chronic diarrhea.
Typically, the severely atrophic gastric body is described as pale and thin with prominent vessels
(Figure 4719). In our experience, many more cases of severe atrophic oxyntic gastritis are diagnosed
in the absence of any typical endoscopic findings. Furthermore, mucosal vessels may be visible in the
absence of mucosal atrophy.137(4111) Measurement of the pH of gastric juice at endoscopy has been
proposed as a way to improve accuracy in the diagnosis of atrophic gastritis.138,139(4112)
(4113)Figure 4719. Thin-appearing mucosa. Note the prominent vascular pattern. This
endoscopic appearance in the gastric body on the greater curvature suggests severe atrophic
gastritis.
Histology
The best site for biopsies in attempting to prove the existence of severe oxyntic gland gastritis is the
midbody region on the greater curvature of the stomach. The mucosa on the lesser curvature and in
the transition zones (antral-body, cardiac-body) is often thinner than on the greater curvature, and
biopsies from these regions could be falsely interpreted as atrophic. Severe oxyntic gland atrophy is
commonly associated with only a mild inflammatory cell infiltrate and with metaplasia, as described
previously.
Nonneoplastic Polyps and Neoplasms
Severely atrophic oxyntic gastritis carries an increased risk of associated inflammatory polyps,
carcinoid tumors, and adenocarcinoma of the gastric body. Except for inflammatory polyps, these
complications are rare. Hyperplastic/inflammatory polyps, with prevalence rates of 10 to 40%, are the
most common lesions found in endoscopic surveys of patients with pernicious anemia.140142(4114)
Impressive enterochromaffin cell hyperplasia is common in association with severe atrophic oxyntic
gastritis,143147(4115) but frank gastric carcinoids are uncommon, although these are now
recognized with increasing frequency.148,149(4116) It is believed that carcinoids have the potential to
metastasize when they exceed 2 cm in diameter and should be removed when they occur as isolated
lesions.
A number of studies of endoscopic surveillance have been done in patients with atrophic
gastritis.150,151(4117) Sjoblom et al.151(4118) performed gastroscopy with biopsy in 56 patients 3
years after the diagnosis of pernicious anemia. Two small carcinoid tumors (3.6%) and two early
gastric cancers (3.6%) were detected. These investigators suggested that endoscopic surveillance
may be beneficial in patients who develop pernicious anemia at a relatively young age. In an earlier
surveillance study by Sjoblom et al.152(4119) of 71 patients with a mean follow-up of 7 years, the total
frequencies of gastric adenocarcinomas and carcinoid tumors were 3% and 4%, respectively.
An initial endoscopic examination is reasonable in patients with pernicious anemia and those with
severe atrophic gastritis and achlorhydria. Thereafter, however, regular surveillance examinations for
dysplasia are unnecessary unless neoplastic lesions (carcinoids, adenomas) are found at initial
evaluation.140(4120) Even in populations or countries where the rate of gastric cancer is high or even
endemic, the benefits of endoscopic surveillance in terms of case detection and prolongation of life are
not readily evident.141(4121) If gastric mucosal dysplasia is detected in biopsies obtained for any
reason, additional biopsies should be obtained at a second endoscopy using a mapping technique (see
Figures 472 and 473). If dysplasia is verified, the dysplastic mucosa should be removed if
possible.153(4122)
Distinctive Types of Gastritis

Distinctive types of gastritis are uncommon. Sometimes it is the endoscopic appearance that is
distinctive, as in Mntrier's disease. More often, the histologic changes are either diagnostic or highly
compatible with one of a limited group of disorders. The distinctive types of gastritis that are discussed
in this section are listed in Table 475.
Distinctive Types of
Gastritis: Erosive and Nonerosive
TABLE 475
INFECTIONS AND INFESTATIONS

Bacterial infections
Tuberculosis
Syphilis
Phlegmonous and Emphysematous Gastritis
Viral infections
Cytomegalovirus infection
Herpesvirus infections
Fungal infections
Candidiasis (Candida albicans)
Histoplasmosis
Other, e.g., zygomycosis (mucormycosis), aspergillosis
Nematosis and other parasitoses
Cryptosporidosis
Anisakidosis
Other
GASTROINTESTINAL TRACT DISEASES
Crohn's disease
Eosinophilic gastritis
SYSTEMIC DISEASE
Sarcoidosis
CONFINED TO THE STOMACH
Mntrier's disease*
Isolated unexplained granulomas
* Text includes discussion of other types of idiopathic
hypertrophic gastropathies as part of differential
diagnosis of Mntrier's disease.
Table 476 lists other disorders in which granulomas
may be found.
Infections
The most common causes of distinctive types of gastritis are infections. Gastric infections are being
recognized with increasing frequency as a result of the AIDS epidemic and of iatrogenic
immunodeficiency due to the use of chemotherapeutic drugs. When infectious gastritis is suspected,
histologic evaluation of biopsies must be complemented by culture of biopsy material (especially for
viruses) and examination of mucosal smears. The primary objective of endoscopy in
immunocompromised patients, as discussed earlier, is to obtain tissue for laboratory evaluation. In
such patients, the possibility of infection must always be eliminated when a lesion is found at
endoscopy, even when the endoscopic appearance is considered typical of a given diagnosis, such as
peptic ulcer.
Bacterial Infections
Infections with H. heilmanii and other spiral organisms were discussed in the section on the
endoscopic findings in H. pylori.
Tuberculosis
Involvement of the stomach in the presence of intraabdominal tuberculosis is
uncommon.154157(4123) A range of endoscopic findings has been described in case reports,
including generalized involvement with petechial lesions, erosions, a viscous whitish
exudate,158(4124) multiple ulcers,159(4125) and a single large ulcer.160(4126) The appearance and
location may be very similar to those of Crohn's disease, and both conditions have the potential to
produce pyloroduodenal obstruction.156,157,161(4127) Four case reports have been published of
isolated gastric tuberculosis in which there was apparently no evidence of infection elsewhere in the
body.157,158,160,161(4128) Diagnosis is by demonstration of caseating granulomas in endoscopic
biopsies and culture of Mycobacterium tuberculosis from tissue specimens obtained at endoscopy.
Because of similarities between Crohn's disease and tuberculous involvement of the GI tract, use of
the polymerase chain reaction on endoscopic biopsies has been proposed as a method of
diagnosis.162(4129)
Syphilis
Gastric syphilis is rare, but the number of cases of secondary syphilis is increasing, partly as a result of
AIDS.163166(4130) Other manifestations of syphilis may be pres-ent in 70% or more of
patients.163(4131) Criteria for diagnosis include abdominal pain, untreated syphilis, positive serology,
lack of response to conventional ulcer therapy, characteristic appearance of the lesion, and a favorable
response to specific therapy. The endoscopic appearances are those of thickened folds with or without
associated erosions and an indurated, friable mucosa;167,168(4132) in case reports, the duodenum
has not been involved.167,168(4133) The appearance may be similar to that of linitis plastica or
lymphoma.166,169171(4134)
Histology may reveal a mononuclear cell infiltrate that is often more dense than that seen in other types
of gastritis.165(4135) The Treponema pallidum spirochetes are demonstrated in gastroscopic biopsy
material by silver impregnation and fluorescent antibody techniques.172(4136) The former staining
method (Warthin-Starry) is technically difficult and not specific for T. pallidum. Although the fluorescent
antibody technique is highly specific, nonfixed fresh tissue is required, and the diagnosis must
therefore be suspected at endoscopy.
Phlegmonous and Emphysematous Gastritis
These forms of gastritis often occur in debilitated patients. Phlegmonous and emphysematous gastritis
are life-threatening entities in which extensive bacterial inflammation of the gastric submucosa is
associated with perforation as a result of necrosis.173180(4137) The term acute necrotizing gastritis
is sometimes used instead of phlegmonous gastritis.181(4138)
Phlegmonous gastritis is a purulent bacterial inflammation that primarily affects the submucosa but
may spread to all layers of the stomach, resulting in necrosis and gangrene. Because some of these
infections result in gas formation in the gastric wall, the distinction between phlegmonous gastritis and
emphysematous gastritis is arbitrary. Phlegmonous gastritis has been reported after endoscopic
tattooing with India ink and polypectomy.182,183(4139)
Emphysematous gastritis is a rarer variant of phlegmonous gastritis in which gas-forming organisms
form blebs in the submucosa.175,177179,184(4140) Prior corrosive ingestion figures prominently in
the accumulated case reports. In some cases of both phlegmonous and emphysematous gastritis, the
bacteria may not be apparent.178,185(4141) In some, prior antibiotic therapy may have eliminated the
bacteria from the wall. Gas in the gastric wall may also be due to abscess formation in the extremely
rare condition of gastric actinomycosis.186(4142)

Typically, the patient with phlegmonous gastritis presents with an acute abdomen, sometimes
accompanied by hematemesis. Ultrasonography,183(4143) CT scan, and plain radiographs of the
abdomen should all show massive thickening of the gastric wall, with or without intramural air.
Endoscopy of putative cases may reveal erosions and a rigid, thick wall.187(4144) With prompt
antibiotic therapy and surgical resection, mortality has been reduced to 20% or less.174(4145) One
patient was reported to have survived without surgery after the presumptive diagnosis of phlegmonous
gastritis was made with a snare biopsy.188(4146)
Viral Infections
Cytomegalovirus Infection
Cytomegalovirus is the most important infectious agent with regard to the stomach after H. pylori.
Cytomegalovirus infection of the stomach or duodenum may be found in a variety of clinical settings:
transplantation,189195(4147) cytomegalovirus mononucleosis,196,197(4148) or
AIDS155,198,199(4149) and with no apparent associations.200,201(4150) Mntrier's disease in
childhood, discussed later, appears to be caused by cytomegalovirus infection of the stomach.
Cytomegalovirus infection of the stomach may be pres-ent in endoscopically normal-appearing
mucosa, or it may be associated with a variety of visible lesions, including subepithelial hemorrhages,
erosions, ulcers, nonulcerated nodules, and thick folds.202204(4151) The two most common
endoscopic findings are normal mucosa and erosions. The latter are often more irregular in outline
than those associated with NSAID ingestion (Figure 4720). The most dramatic-appearing lesions are
those of hypertrophic gastropathy: localized antral thickening and diffuse or localized thickening of the
gastric body folds.205(4152) Thickening is usually due to foveolar hyperplasia, but occasionally
lymphoid hyperplasia may be prominent.206(4153) Whenever cytomegalovirus is found in biopsies of
thick folds in a patient with AIDS, underlying lymphoma and Kaposi's sarcoma must be considered in
the differential diagnosis.
(4154)Figure 4720. Irregular-appearing cytomegalovirus erosions in a patient after liver

transplantation.
The diagnosis of cytomegalovirus is established most commonly by finding the typical cytomegalic cells
with their intranuclear inclusions in tissue sections. A positive viral culture is also diagnostic. When
focal lesions are seen endoscopically and cytomegalovirus infection is suspected, biopsies should be
taken from the target lesions as well as from adjacent mucosa. As discussed earlier, biopsies from
adjacent mucosa may provide the diagnosis when directly targeted specimens do not. For example,
biopsies obtained from ulcers or erosions may contain only granulation tissue. When cytomegalovirus
infection is suspected in an immunosuppressed patient, the pathologist should be alerted to this
possibility so that more sections can be prepared for histologic examination if no typical cells are seen
on examination of the first sections.
Herpesvirus Infections
Gastric erosions associated with herpesvirus infection have been described in the stomach in the
immunocompromised patient.193,207,208(4155) When erosions or ulcers are present, the best
approach to diagnosis is to obtain a smear for cytologic examination, a smear for culture, and biopsies
for histology. Herpesvirus, unlike cytomegalovirus, is most prominent in the exudate from these
lesions, so that smears for cytologic evaluation and culture are more likely to provide a positive
diagnosis. By contrast, cytomegalovirus produces a deeper infection and is more likely to be found, for
example, in the granulation tissue at the bases of ulcers. Some common-variety ulcers may be
associated with herpes simplex virus type 1.209(4156) However, there is no indication to routinely look
for herpesvirus infection in gastric ulcers or erosions in immunocompetent individuals.
Herpes zoster rarely involves the stomach; only a few anecdotal reports of ulcerations and
gastroparesis associate infection with this virus.210212(4157)
Fungal Infections
Candidiasis (Candida Albicans)
Candida albicans and, to a lesser extent, Torulopsis glabrata are commonly found in the stomach in
the absence of any lesions.213(4158) These fungal organisms may also be found in the base of
gastric ulcers or erosions in immunocompetent as well as immunocompromised
patients.213219(4159) More than half (60%) of the ulcers in the 23 patients studied by Morishita et
al.220(4160) were present in the proximal stomach, and the majority (54%) were more than 2 cm in
diameter. In most cases, the margin of the ulcer(s) was poorly demarcated. Minoli et al.221(4161)
noted 3 endoscopic appearances in a series of 26 patients with gastric candidiasis. In 42% there was a
thrushlike white or green-white membrane that was easily removed to reveal an inflamed mucosa
beneath. Red nodules of a few millimeters in diameter, mainly in the antrum, were seen in 31% of
cases. Nondescript ulcers were present in 27%.
When Candida is found in association with a gastric ulcer or erosions in an immunocompetent patient,
it is assumed that the lesion has been colonized only by the fungus because healing is not affected by
the presence of this organism.217,222(4162) Patients with gastrostomy tubes223(4163) and liver
transplant candidates are at increased risk for colonization, but this probably does not represent a
hazard unless the patient becomes neutropenic due to disease or therapy.
Histologic examination of smears of exudate and biopsy specimens from ulcer edges for mycelia
establish the presence of C. albicans. There is no point in asking the pathologist if evidence of invasion
is present because dissemination depends on whether the patient is neutropenic.
Histoplasmosis
The colon and ileum are the most common sites of involvement of the GI tract by histoplasmosis, with
increasing numbers of cases now being diagnosed in patients with AIDS.224226(4164) The stomach
is rarely involved. Gastric involvement is usually characterized by the presence of gastric ulcers and
erosions on the crests of thickened gastric folds.227,228(4165) Patients with ulcerative lesions may
present with bleeding, or a masslike lesion may be discovered in the evaluation of abdominal
pain.229(4166) Histology may reveal the granulomas containing the fungus, but definitive diagnosis
requires culture of biopsy specimens.227(4167)
Other Fungal Infections
ZYGOMYCOSIS (MUCORMYCOSIS). The widely distributed fungi of the Phycomycetes class

produce disease primarily in immunocompromised or debilitated patients.230232(4168) The stomach
is the most frequently affected part of the GI tract in zygomycosis, the typical lesion being a deep
bleeding ulcer with dark indurated edges.
ASPERGILLOSIS. Aspergillosis is extremely rare in the GI tract.233(4169) Vascular thrombosis

with resultant ischemic necrosis accounts for some of the ulcerations that are seen.8,233(4170)
ACTINOMYCOSIS. A case of primary gastric actinomycosis has been reported.234(4171)

Nematosis and Parasitoses
Cryptosporidiosis
Cryptosporidium is seldom described in the stomach.235(4172) Until effective therapy is available for
this parasite, it is impossible to know whether its presence in the stomach in association with gross
and/or microscopic abnormalities is coincidental or the cause of the pathologic changes. The former
circumstance is usually the case. In immunosuppressed patients, even those with symptoms and
gastric lesions such as erosions, proving that the organism is actually causative is difficult.236(4173)
An antral stricture associated with Cryptosporidium was reported to resolve with paromomycin
therapy.237(4174)
Anisakidosis
This infection, formerly known as anisakiasis, represents one of the risks of eating sushi and other
types of raw fish.238240(4175) The diagnosis should be suspected in patients who have eaten raw
fish within 12 hr of the onset of abdominal pain.241(4176) A number of anisakiad genera can infect
humans, including Anisakis, Contracaecum, and Phocanema. The Ascaris-like larvae, which measure
2 to 3.5 cm, embed themselves in the gastric mucosa. Anisakidosis is usually short lived and self
limited, abdominal pain being the main clinical feature.
This infection has special appeal for endoscopists because the larvae may be "fished out" of the
mucosa, and they lend themselves to all sorts of imaging.240,242,243(4177) Endoscopic removal of
the larvae relieves abdominal pain in some patients. Because the abdominal pain may be intolerable,
endoscopy with extraction of the worms from the mucosa has been recommended by those with a
large experience with this infection.241(4178) In one report of 87 cases, the infection was localized
mainly to the gastric body in 47%, especially the greater curvature, and to the incisura in
31%.239(4179) Mucosal swelling with prominent folds is a common endoscopic feature.239,240(4180)
In some patients, erosions or ulcers may be seen, perhaps due to a greater density of the worms in a
particular area of the mucosa.239(4181) A patient has been described who had chronic dyspepsia for
almost a year that resolved after removal of a codworm-type variant (Pseudoterranova) of
anisakidosis.244(4182) Acute onset of severe abdominal pain due to gastric infestation with
Pseudoterranova decipiens has also been described.245(4183)
Anisakis is not the only infectious parasite that can be extracted from the stomach. Similar feats have
been described for hookworm246(4184) and roundworm.247(4185) Gastric ascariasis is unusual,
although gastric outlet obstruction due to the presence of large numbers of worms has been
described.248(4186)
Other Parasitic Infections
The presence of trophozoites of Giardia lamblia in the stomach was first described in
1992.249,250(4187) For this to occur, it may be necessary for intestinal metaplasia to be present in the
stomach.249(4188) If this is absent, the trophozoites may simply reflux into the stomach from the
duodenum.251(4189) In addition to the more conventional parasites in the West, other more exotic
types may be found in the stomach, especially in the immunosuppressed patient.252(4190)
Eustrongyloidiasis is another of the hazards of eating raw or poorly cooked fish.253,254(4191)
Strongyloides stercoralis is common in many parts of the world, including the southeastern United
States.255(4192) If unrecognized, strongyloidiasis may become disseminated and potentially lethal,
especially in the immunosuppressed patient.256,257(4193) Strongyloides predominates in the small
bowel but rarely is found within the gastric mucosa or elsewhere in the gastric wall, usually in
immunosuppressed patients.175,258,259(4194) The organism may be found in normal-appearing
mucosa, in ulcers, and within the wall of the stomach.175,256,258,260(4195)
Isolated reports of gastric involvement in schistosomiasis have been published;261(4196) in one
patient, Shistosoma japonicum was found in a strip biopsy resection of a gastric carcinoma.262(4197)
Amebiasis has been found in a resection specimen of a gastric lymphoma.263(4198) Comparably rare
are Acanthamoeba infection264(4199) and hydatid disease, the latter as a fistula from a hydatid cyst
into the stomach.265(4200) Gastric toxoplasmosis and leishmaniasis have been reported in patients
with AIDS as part of an apparently localized or disseminated disease process.266268(4201) In a
series of 15 cases of leishmaniasis with GI tract involvement, three quarters of the patients had gastric
involvement, and in one half of these the infection was discovered in gastric mucosa that looked
normal at endoscopy.267(4202)
Patients with malaria may have abdominal pain, nausea, and vomiting. In a report from Venezuela, a
wide variety of changes were described in the gastric mucosa of patients with malaria, half of whom
were infected with Plasmodium falciparum.269(4203) Superficial bleeding was the most objective
finding. Romero et al.269(4204) speculated that mucosal ischemia might be a factor in the
development of the gastric lesions they described. However, the observed changes might also be due
to the stress of a severe systemic, febrile illness.
Diseases of the Gastrointestinal Tract

Crohn's Disease
The antrum is the main site of gastric Crohn's disease, which often coexists with involvement of the
proximal duodenum. When the stomach or duodenum or both are involved by Crohn's disease,
evidence of the disease is almost always found in the segments of the gut that are more commonly
involved. In fact, one should be extremely wary of making a diagnosis of isolated Crohn's disease of
the stomach in the absence of involvement of the more customary sites;270(4205) if the usual sites
are uninvolved, the patient's interests are best served by a provisional diagnosis of idiopathic
granulomatous gastritis.
Endoscopy and biopsy are regarded as especially useful in children for the diagnosis of upper GI
Crohn's disease. This is attributable to the fact that this diagnosis is often more difficult in children than
in adults. The patient's history may be of limited value or not available, and there is a reluctance to
subject children to repeated radiographic studies. In adults, endoscopy and biopsy are useful if the
diagnosis is uncertain and immunosuppressive therapy is deemed necessary. Endoscopic examination
of the distal duodenum and proximal jejunum using a long endoscope is useful in this circumstance.
Biopsies should be obtained from lesions if these are present, but also in random fashion from
apparently normal mucosa when no gross abnormalities are noted. If no abnormalities are present in
the stomach, I take biopsies in random pattern from a minimum of three sites: prepyloric region,
mid-antrum, and midbody.
The stenotic transmural form of Crohn's disease is very uncommon in the stomach, but mucosal
involvement is not, especially in children.271275(4206) The endoscopic features of gastric Crohn's
disease are similar to those of disease elsewhere in the GI tract. Gastric mucosal lesions include
shallow erosions of various shapes, ulcerations (Figure 4721), and cobblestoning.273,276278(4207)
Patients may have presenting symptoms of delayed gastric emptying because of a narrowed
gastroduodenal region or dyspepsia associated with mucosal lesions. Other complications include
fistulas to the small bowel or colon.279(4208)
(4209)Figure 4721. Endoscopic views of Crohn's disease of the stomach. A, Erythema and
serpiginous ulcers in the prepyloric antrum. B, Closer view of serpiginous ulcers and
nodularity. (A and B, From the collection of Dr. M. V. Sivak, Jr.)
Although granulomas are said to be found in gastric biopsies in as many as 50% of patients with
involvement of the stomach by Crohn's disease, the yield is less than 20% in my institution. As with
Crohn's disease elsewhere in the GI tract, biopsies of normal mucosa may also reveal granulomas. A
pattern of very patchy inflammation (H. pylori negative) in biopsy material is an important clue to the
diagnosis of Crohn's disease of the stomach.271(4210) This variability in the severity of inflammatory
changes may be evident from biopsy to biopsy as well as within a single specimen.
When seriously ill patients with Crohn's disease undergo endoscopy for GI tract bleeding or severe
ulcer-type symptoms and erosions or ulcers are seen, it may be impossible to differentiate among
Crohn's lesions, stress erosions, and peptic ulcer. Opportunistic infection must also be considered if
the patient is on immunosuppressive drugs.
Eosinophilic Gastritis
Eosinophilic gastroenteritis is a rare condition that typically affects the small intestine but may involve
other areas of the GI tract, including the stomach (eosinophilic gastritis). The characteristic
histopathologic finding is masses of eosinophils, either in the mucosa or in deeper layers of the gastric
wall, that is, the muscularis propria and serosa.280284(4211) When the eosinophilic infiltrates are
predominantly in the mucosa, malabsorption becomes the primary clinical problem. If the muscle
layers are involved, obstructive symptoms are dominant. In the differential diagnosis of this condition,
the main consideration is to exclude parasitic infestation.
Symptoms due to gastric involvement are those of delayed gastric emptying or, if erosions or ulcers
are pres-ent, GI bleeding and abdominal pain. Endoscopy during the active stage when there is
mucosal involvement may reveal nodularity (Figure 4722) and sometimes persistent pyloric outlet
obstruction.285(4212) A masslike lesion has been described at endoscopy.286(4213) The diagnosis is
not always easy, and, in children, biopsies from the gastric antrum obtained at random in unselected
cases may reveal eosinophilic infiltrates.287(4214)
(4215)Figure 4722. Eosinophilic gastritis. Nodules in the antrum from a child with
eosinophilic gastroenteritis.
Systemic Disease
Sarcoidosis
GI tract involvement in sarcoidosis is rare, although the stomach is the most common
site.288290(4216) Granulomas may be found in endoscopically normal mucosa291(4217) as well as
adjacent to areas of antral narrowing, thickened folds in the gastric body, and gastric ulcers.289(4218)
If granulomas are detected in the course of evaluating a patient with known sarcoidosis, other causes,
specifically infections, should be eliminated. When marked luminal narrowing or thickened folds are
present, infiltrative neoplasms must be excluded.
Amyloidosis
Only a few descriptions of the endoscopic findings of gastric involvement in amyloidosis are available.
Infiltration of the stomach results in prominent gastric folds, ulcers, and bleeding (Figure
4723).292294(4219) Absence of antral peristaltic contractions has been noted in most cases of
advanced involvement.295(4220) There may be gastric retention with food residue present in the
stomach. Although the approach to diagnosis of amyloid involvement of the digestive tract is usually by
identification of amyloid in rectal biopsies, amyloid is also readily identifiable in gastric
biopsies.296(4221)
(4222)Figure 4723. Endoscopic views of the stomach in a patient with amyloid infiltration.
A, Antrum and pylorus. B, Prominent, deformed gastric folds in the body. (A and B, From the
Diseases Confined to the Stomach

Thick Gastric Folds (Hypertrophic Gastropathy)
Renewed interest has been expressed in thick gastric folds, an entity that first attracted attention in the
1970s, when it became possible to remove relatively large pieces of such folds with an electrocautery
snare.108,297(4223) This was followed by a period of waning interest when it proved to be impossible
in many cases to determine a cause despite the availability of large pieces of tissue for histologic
examination. The resurgence of interest arose from the use of endoscopic ultrasonography to image
the finer structure of the gastric wall itself,298(4224) vigorous attempts to identify patients with
MALTomas,299(4225) and the recognition that some cases of giant folds are associated with H. pylori
infection.300(4226) The guiding principle in the evaluation of thick folds is to make neoplasia, either an
infiltrating carcinoma or a lymphoma, the first consideration in the differential diagnosis.
No standardized criteria have been established for the definition of "thick" or "thickened" gastric folds.
The radiologist, either because of barium radiography or CT scan findings, is often the first to suggest
that thick folds are present. Endoscopy, invariably performed, then provides an additional subjective
assessment that the folds appear to be either normal or thickened. In the latter case, even a snare
biopsy often leads to disappointment because histologic examination frequently fails to provide an
explanation for the endoscopic findings.
Appelman301(4227) has pointed out that large folds are not usually produced purely by mucosal
expansion. Submucosal ridges must also contribute because they form the core of all types of gastric
folds. Thus, thick folds are the combination of an expanded mucosa covering unusually prominent
submucosal ridges. Perhaps one reason why snare biopsies fail to provide an explanation for thick
folds in many patients is the prominence of submucosal ridges with a normal-thickness overlying
mucosa.
A variety of terms have been used for thick gastric folds, including giant hypertrophic gastritis, giant
fold gastritis, giant rugal hypertrophy, chronic hypertrophic proliferative gastritis, and chronic
hypertrophic glandular gastritis.302(4228) Some of the uncertainty about thick folds stems from the
terminology. I prefer to use the generic term hypertrophic gastropathy or just thick folds.
The Endoscopic Evaluation of Thick Folds
The first step in the endoscopic evaluation of thick gastric folds is to obtain biopsies with a large
("jumbo") forceps, if available, and fine-needle aspirates,303,304(4229) assuming that the endoscopist
has adequate experience with the latter technique. A "burrowing" technique is sometimes useful when
successive biopsies are obtained from the same site. I do not find this especially helpful, as only a
blood clot is obtained after removal of the first one or two specimens.
Endoscopic ultrasonography can be helpful in determining whether seemingly thick gastric folds are
actually abnormal and in defining the extent and depth of any abnormality.298,305(4230) If initial
large-particle biopsies and needle-aspiration specimens fail to yield a diagnosis but one still suspects
neoplasia, snare biopsies of the folds are indicated. If the folds are soft and raised, grasping a portion
with a standard polypectomy snare is usually relatively easy. If the folds are more firm and difficult to
grasp, saline injection (strip biopsy) technique should be used (see Chapter 24: Early Colorectal
Cancer and Endoscopic Resection).306(4231) The saline injection should be made using a relatively
long needle so as to lift the submucosa. Otherwise, the injection merely elevates the mucosa. I use
needles designed for transbronchial biopsies that range in length from 1 to 1.3 cm. These needles are
also the most suitable for obtaining fine-needle aspirates in this circumstance.
Mntrier's Disease
Mntrier's disease is an exceptionally rare syndrome. It is not surprising, given its rarity, that the
definitions of the syndrome are incomplete. Because of its clarity, the definition I prefer, with minimum
modification, is that of Appelman.301(4232) That is, authentic cases of Mntrier's disease are those
that have (1) giant folds, especially in the fundus and body of the stomach; (2) diminished acid
secretory capacity; (3) hypoalbuminemia; and (4) histologic features of foveolar hyperplasia, atrophy of
glands, and a marked overall increase in mucosal thickness.
Patients with Mntrier's disease are generally more than 50 years of age. The disease is more
common in older men and occasionally shows a familial predisposition.307(4233) The main presenting
symptoms are weight loss, abdominal pain, and edema due to hypoalbuminemia. Many of these
patients are referred for endoscopy because massively enlarged gastric folds are detected with barium
contrast radiography during initial investigation. The fundus and body (especially the greater curvature)
are commonly affected, but concomitant involvement of the antrum may be present.
Endoscopy
The endoscopic findings are striking, with massively heaped-up folds, sometimes with associated
erosions or ulcers. The process affects the fundus and body and sometimes extends into the antrum.
The folds may be pliable, but they are not fully effaced by air insufflation (Figure 4724); crests of folds
sometimes have a confluent polypoid appearance.
(4234)Figure 4724. Mntrier's disease. A, Endoscopic view of raised polypoid folds on the
greater curvature of the stomach. B, Snare biopsy of Mntrier's disease. The polypoid lesion
contains numerous cysts with a loose stroma. (A, Courtesy of Prof. G. N. J. Tytgat; B, courtesy
of Dr. K. Lewin.)
Histologic Findings
The histology of Mntrier's disease is typified by markedly elongated gastric pits and cysts with an
edematous stroma (see Figure 4724). Normal oxyntic gland elements are replaced in the gastric
body; this accounts for the associated hypochlorhydria. In a large study (relative to the number of
documented cases of Mntrier's disease), the histologic findings in 23 patients were
reexamined.132(4235) Wolfsen et al.132(4236) found lymphocytic gastritis in 13 of these cases. Other
differences in the histology were also evident by comparison of patients with and without lymphocytic
gastritis. Wolfsen et al.132(4237) therefore assigned different names to these two groups: hypertrophic
lymphocytic gastritis (HLG) and massive foveolar hyperplasia and minimum inflammation (MFH). Until
the presence or absence of lymphocytic gastritis is confirmed and demonstrated to have a bearing on
the manifestations or course of this disease, the name Mntrier's disease should be retained, with
notation made as to whether lymphocytic gastritis and H. pylori are present.
In children with Mntrier's disease, evidence of cytomegalovirus infection should be sought because
this organism appears to be the cause of the self-limited form of the disease in this age
group.308,309(4238)
Surgery
Surgical resection was formerly carried out in most patients with Mntrier's disease for two purposes:
confirmation of the diagnosis and control of symptoms. If symptoms can be controlled and the findings
are compatible with Mntrier's disease, then surgery for confirmation of the diagnosis is not
indicated.310(4239) Biopsies obtained at endoscopy with standard pinch forceps provide a clue to the
diagnosis because marked foveolar cell hyperplasia is present. However, the relatively superficial
standard endoscopic biopsy is insufficient to establish the diagnosis with certainty. A more definitive
diagnosis can be made with an electrocautery snare biopsy, but in many circumstances even this
approach may not be necessary.
Carcinoma
A few cases of carcinoma have been reported in patients with Mntrier's disease.310312(4240)
Some of these patients have had partial gastrectomies, and a cancer in this setting must be interpreted
within the context of the cancer risk in the postoperative stomach. It has long been recommended that
patients with Mntrier's disease undergo periodic endoscopic surveillance with biopsies obtained at
random.313(4241) At first glance this seems reasonable, but it is really an impossibility given the field
of polypoid folds (see Figure 4724). If this form of surveillance is undertaken, a biopsy mapping plan
should be used. It would also be prudent, however, to advise patients about the marked limitations of
endoscopic/biopsy surveillance in this condition.
Other Types of Hypertrophic Gastropathy
The collective term I use to designate atypical or unclassified disorders in which the gastric folds are
thick is idiopathic hypertrophic gastropathy. As a group, these disorders are uncommon, although still
more common than Mntrier's disease.
The histologic picture in some patients is similar to that of Mntrier's disease, with marked foveolar
hyperplasia and some degree of atrophy of the oxyntic glands but without apparent protein loss and
hypochlorhydria.301(4242) These may be the cases that with time evolve into Mntrier's disease.
Other examples of atypical courses include spontaneous remission and progression to atrophic
gastritis.301,314(4243)
Overall, the most common form of hypertrophic gastropathy in my experience is localized thickening,
especially in the gastric antrum (Figure 4725). With snare biopsy, some of these cases prove to
resemble hyperplastic polyps; in others, overall thickening of all mucosal elements occurs. Most often,
no explanation for these findings is found except for the speculation, referred to earlier, that the
findings are due to submucosal ridges. Nonerosive gastritis is present in some of these cases, and the
term hypertrophic gastritis is sometimes used. I do not use this term because most often the gastritis
seems rather ordinary, with no firm evidence that it contributes to the gross findings of fold thickening.
(4244)Figure 4725. Localized idiopathic antral hypertrophy. A, Barium x-ray, close-up

view. Nodular-type filling defects (arrows) are seen for several centimeters proximal to the
pylorus (p). B, Red, raised serpentine folds radiate into the pylorus. Snare biopsy revealed
only nonerosive gastritis.
In cases of localized hypertrophic gastropathy, the main issue in differential diagnosis is to rule out
underlying malignancy.
Lymphoid Hyperplasia (Pseudolymphoma)
The term lymphoid hyperplasia, in its former connotation,315(4245) may now be almost obsolete,
partly because of better therapies for peptic diseasethat is, antisecretory therapy and drug regimens
for eradication of H. pylori. In effect, patients no longer have ulcers associated with lymphoid
hyperplasia for months or years. Also, some and perhaps most of the cases formerly considered to be
benign are now classified as MALTomas.99,316,317(4246) If specimens obtained at endoscopy are
interpreted as showing benign lymphoid hyperplasia or pseudolymphoma, a second opinion should be
obtained from another pathologist with experience in gastric lymphomas.
Isolated (Unexplained) Granulomatous Gastritis
Granulomas may be found in the stomach in a wide variety of conditions (Table 476). In a review of
71 patients with granulomas in the stomach, 52% had Crohn's disease, 10% were foreign body-type
granulomas, 25% were classified as isolated (unexplained), and the remainder were associated with a
variety of disorders, including Whipple's disease and vasculitis.318(4247)
Conditions Associated With

the Finding of Gastric Granulomas
TABLE 476
ISOLATED (UNEXPLAINED) GRANULOMAS

Foreign-body type
As mass or ulcer or in normal-appearing mucosa
NONINFECTIOUS
Crohn's disease
Sarcoidosis
Chronic granulomatous disease (childhood)
Allergic granulomatosis
INFECTIOUS
Tuberculosis
Histoplasmosis
Late syphilis
Parasitoses
OTHER
Whipple's disease (see text)
Modified from Lewis KJ, Riddell RH, Weinstein WM:
Stomach and proximal duodenum inflammatory
disorders. Courtesy of Klaus J. Lewin, Robert H. Riddell,
and Wilfred M. Weinstein, from Gastrointestinal
Pathology and Its Clinical Implications. New York:
Igaku-Shoin Medical Publishers, 1992.
Isolated (unexplained) granulomatous gastritis refers to the finding of granulomas in the absence of
any specific disease. It is relatively common to find isolated granulomas in gastric biopsies taken for a
variety of reasons. Some of these are of the foreign body type, containing debris, and are presumed to
represent a reaction to some form of prior mucosal injury.
The term isolated granulomatous gastritis has also been used to describe granulomatous inflammation
associated with striking structural changes in the stomach, especially the antrum. These findings may
mimic gastric carcinoma, with marked antral narrowing, mass lesions, and ulcerative lesions that may
even perforate.319321(4248)
A diagnosis of isolated granulomatous gastritis should not be made unless other associations,
especially infections, are eliminated from consideration (see Table 476).
Acknowledgments
I am indebted to Klaus Lewin, M.D., for an ongoing, stimulating collaboration to Zhang Da for superb
preparation of biopsies, to Myriam Marin for testing new histologic techniques, and to the UCLA GI
Procedures Unit staff for their consistent dedication to quality.
REFERENCES
1.
2.
3.
Price AB. The Sydney system histological division. J Gastroenterol Hepatol 1991;6:20922.
Krishnamurthy S, Dayal Y. Pancreatic metaplasia in Barrett's esophagus. An
immunohistochemical study. Am J Surg Pathol 1995;19:117280.
Chow WH, Finkle WD, McLaughlin JK, et al. The relation of gastroesophageal reflux disease
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
and its treatment to adenocarcinomas of the esophagus and gastric cardia. JAMA
1995;274:4747.
Wolf BC, Khettry U, Leonardi HK, et al. Benign lesions mimicking malignant tumors of the
esophagus. Hum Pathol 1988;19:14854.
Isaacson P. Biopsy appearances easily mistaken for malignancy in gastrointestinal
endoscopy. Histopathology 1982;6:37789.
Clark GWB, Ireland AP, Chandrasoma P, et al. Inflammation and metaplasia in the
transitional epithelium of the gastroesophageal junction: A new marker for gastroesophageal
reflux disease [Abstract]. Gastroenterology 1994;106:A63.
Lewin KJ, Riddell RH, Weinstein WM. Dialogue, handling of biopsies and resected
specimens. In Lewin KJ, Riddell RH, Weinstein WM (eds). Gastrointestinal Pathology and Its
Clinical Implications. New York, Tokyo: Igaku-Shoin, 1992;15.
Lewin KJ, Riddell RH, Weinstein WM. Stomach and proximal duodenum inflammatory and
miscellaneous disorders. In Lewin KJ, Riddell RH, Weinstein WM (eds). Gastrointestinal
Pathology and Its Clinical Implications. New York, Tokyo: Igaku-Shoin, 1992;506.
Misiewicz JJ, Tytgat GNJ, Goodwin LS, et al. The Sydney system: A new classification of
gastritis. Working Party Reports of the 9th World Congress of Gastroenterology. Melbourne:
Blackwell Scientific, 1989;110.
Tytgat GN. The Sydney system: Endoscopic division. Endoscopic appearances in
gastritis/duodenitis. J Gastroenterol Hepatol 1991;6:22334.
Stolte M. A critical look at the Sydney classification of gastritis. Endoscopy 1991;23:28990.
Kimura K, Satoh K, Taniguchi Y, et al. Some personal comments on the Sydney system for
the classification of chronic gastritis. J Gastroenterol 1994;29(suppl 7):1149.
Soll AH, Weinstein WM, Kurata J, McCarthy D. Nonsteroidal anti-inflammatory drugs and
peptic ulcer disease. Ann Intern Med 1991;114:30719.
Hudson N, Everitt S, Hawkey CJ. Interobserver variation in assessment of gastroduodenal
lesions associated with non-steroidal anti-inflammatory drugs. Gut 1994;35:10302.
Laine L, Weinstein WM. Subepithelial hemorrhages and erosions of human stomach. Dig Dis
Sci 1988;33:490503.
Stene-Larsen G, Nesland A, Berstad A. Follow-up study of erosive prepyloric changes. Scand
Parente F, Petrillo M, Vago L, et al. The watermelon stomach: Clinical, endoscopic,
endosonographic, and therapeutic aspects in three cases. Endoscopy 1995;27:2036.
Barnard GF, Colby JM, Saltzman JR, et al. Endoscopic ultrasound appearance of watermelon
stomach. Abdom Imaging 1995;20:268.
Gostout CJ, Viggiano TR, Ahlquist DA, et al. The clinical and endoscopic spectrum of the
watermelon stomach. J Clin Gastroenterol 1992;15:25663.
Labenz J, Brsch G. Bleeding watermelon stomach treated by Nd-YAG laser
photocoagulation. Endoscopy 1993;25:2402.
Laine L, Weinstein WM. Histology of alcoholic hemorrhagic "stritis" : A prospective
evaluation. Gastroenterology 1988;94:125462.
Roy J, Snover D, Weisdorf S, et al. Simultaneous upper and lower endoscopic biopsy in the
diagnosis of intestinal graft-versus-host disease. Transplantation 1991;51:6426.
Chamberlain CE. Acute hemorrhagic gastritis. Gastroenterol Clin North Am 1993;22:84373.
Reusser P, Gyr K, Scheidegger D, et al. Prospective endoscopic study of stress erosions and
ulcers in critically ill neurosurgical patients: Current incidence and effect of acid-reducing
prophylaxis. Crit Care Med 1990;18:2704.
Martin LF. Stress ulcers are common after aortic surgery. Endoscopic evaluation of
prophylactic therapy. Am Surg 1994;60:16974.
Brown TH, Davidson PF, Larson GM. Acute gastritis occurring within 24 hours of severe head
injury. Gastrointest Endosc 1989;35:3740.
Czaja AJ, McAlhany JC, Andes WA, Pruitt BA Jr. Acute gastric disease after cutaneous
thermal injury. Arch Surg 1975;110:6005.
28. Czaja AJ, McAlhany JC, Pruitt BA Jr. Acute duodenitis and duodenal ulceration after burns.
Clinical and pathological characteristics. JAMA 1975;232:6214.
29. Czaja AJ, McAlhany JC, Pruitt BA Jr. Acute gastroduodenal disease after thermal injury. An
endoscopic evaluation of incidence and natural history. N Engl J Med 1974;291:9259.
30. Cook DJ, Fuller HD, Guyatt GH, et al. Risk factors for gastrointestinal bleeding in critically ill
patients. Canadian Critical Care Trials Group. N Engl J Med 1994;330:37781.
31. Kamada T, Sato N, Kawano S, et al. Gastric mucosal hemodynamics after thermal or head
injury. A clinical application of reflectance spectrophotometry. Gastroenterology
1982;83:53540.
32. Tryba M. Stress bleeding. I. Pathogenesis, clinical picture and therapy. Anaesthesist
1994;43:82134.
33. Hirschowitz BI. Nonsteroidal anti-inflammatory drugs and the gastrointestinal tract.
Gastroenterologist 1994;2:20723.
34. Laine L, Cominelli F, Sloane R, et al. Interaction of NSAIDs and Helicobacter pylori on
gastrointestinal injury and prostaglandin production: A controlled double-blind trial. Aliment
Pharmacol Ther 1995;9:12735.
35. Laine L, Marin-Sorensen M, Weinstein WM. Nonsteroidal anti-inflammatory drug-associated
gastric ulcers do not require Helicobacter pylori for their development. Am J Gastroenterol
1992;87:1398402.
36. Taha AS, Nakshabendi I, Lee FD, et al. Chemical gastritis and Helicobacter pylori related
gastritis in patients receiving nonsteroidal anti-inflammatory drugs: Comparison and
correlation with peptic ulceration. J Clin Pathol 1992;45:1359.
37. McCarthy CJ, McDermott M, Hourihane D, O'Morain C. Chemical gastritis induced by
naproxen in the absence of Helicobacter pylori infection. J Clin Pathol 1995;48:613.
38. Quinn CM, Bjarnason I, Price AB. Gastritis in patients on non-steroidal anti-inflammatory
drugs. Histopathology 1993;23:3418.
39. Uppal R, Lateef SK, Korsten MA, et al. Chronic alcoholic gastritis. Roles of alcohol and
Helicobacter pylori. Arch Intern Med 1991;151:7604.
40.
Laine L, Marin-Sorensen M, Weinstein WM. Campylobacter pylori in alcoholic hemorrhagic
"stritis." Dig Dis Sci. 34:677801989.
41. Oakes DD. Reconsidering the diagnosis and treatment of patients following ingestion of liquid
lye. J Clin Gastroenterol 1995;21:856.
42. Vergauwen P, Moulin D, Buts JP, et al. Caustic burns of the upper digestive and respiratory
tracts. Eur J Pediatr 1991;150:7003.
43. Rappert P, Preier L, Korab W, Neubauer T. Diagnostic and therapeutic management of
esophageal and gastric caustic burns in childhood. Eur J Pediatr Surg 1993;3:2025.
44. Laine LA, Bentley E, Chandrasoma P. Effect of oral iron therapy on the upper gastrointestinal
tract. A prospective evaluation. Dig Dis Sci 1988;33:1727.
45. Moore JG, Alsop WR, Freston JW, Tolman KG. The effect of oral potassium chloride on
upper gastrointestinal mucosa in healthy subjects: Healing of lesions despite continuing
treatment. Gastrointest Endosc 1986;32:2102.
46. Das TK, Susheela AK, Gupta IP, et al. Toxic effects of chronic fluoride ingestion on the upper
gastrointestinal tract. J Clin Gastroenterol 1994;18:1949.
47. Fennell DA, Gandhi SS, Prichard BN. Gastrointestinal haemorrhage associated with
free-base (crack) cocaine. Postgrad Med J 1995;71:3778.
48. Kodali VP, Gordon SC. Gastrointestinal hemorrhage secondary to crack cocaine. Gastrointest
Endosc 1995;41:6045.
49. Gourgoutis G, Das G. Gastrointestinal manifestations of cocaine addiction. Int J Clin
Pharmacol Ther 1994;32:13641.
50. Sorbye H, Svanes K. The role of blood flow in gastric mucosal defence, damage and healing.
Dig Dis 1994;12:30517.
51. Doria MI Jr, Doria LK, Faintuch J, et al. Gastric mucosal injury after hepatic arterial infusion
chemotherapy with floxuridine. A clinical and pathologic study. Cancer 1994;73:20427.
52. Mori K, Tominaga K, Yokoyama K, et al. Efficacy of famotidine in patients with acute gastric
mucosal injury after continuous infusion of cisplatin plus vindesine. J Cancer Res Clin Oncol
1995;121:36770.
53. Wells JJ, Nostrant TT, Wilson JA, Gyves JW. Gastroduodenal ulcerations in patients
receiving selective hepatic artery infusion chemotherapy. Am J Gastroenterol 1985;80:4259.
54. Valiante F, Farinati F, Dal Santo P, et al. Upper gastrointestinal injury caused by diquaz
[Letter]. Gastrointest Endosc 1992;38:204.
55. Chen YL. Mechanical gastritis involving the cardia: The trauma of retching and vomiting. J
56. Chen YL. Mechanical gastritis as cause of upper gastrointestinal hemorrhage. Scand J
57. Boyd EJ, Penston JG, Russell RI, Wormsley KG. Hiatal hernial ulcers: Clinical features and
follow-up. Postgrad Med J 1991;67:9003.
58. Al Karawi MA, Mohamed AR, el-Etaibi KE, et al. Extracorporeal shock-wave lithotripsy
(ESWL)-induced erosions in upper gastrointestinal tract. Prospective study in 40 patients.
Urology 1987;30:2247.
59. Trott KR. Radiation risks from imaging of intestinal and abdominal inflammation. Scand J
Gastroenterol Suppl 1994;203:437.
60. Coia LR, Myerson RJ, Tepper JE. Late effects of radiation therapy on the gastrointestinal
tract. Int J Radiat Oncol Biol Phys 1995;31:121336.
61. Bourdages R, Prentice RS, Beck IT, et al. Atheromatous embolization to the stomach: An
unusual cause of gastrointestinal bleeding. Am J Dig Dis 1976;21:88994.
62. Martin de Argila C, Sanroman AL, Moreira VF, et al. Upper digestive hemorrhage as clinical
presentation of leukocytoclastic vasculitis: An unusual endoscopic diagnosis. Endoscopy
1993;25:5423.
63. Golden MP, Hammer SM, Wanke CA, Albrecht MA. Cytomegalovirus vasculitis. Case reports
and review of the literature. Medicine (Baltimore) 1994;73:24655.
64. Kato S, Shibuya H, Naganuma H, Nakagawa H. Gastrointestinal endoscopy in
Henoch-Schnlein purpura. Eur J Pediatr 1992;151:4824.
65. Menke DM, Kyle RA, Fleming CR, et al. Symptomatic gastric amyloidosis in patients with
primary systemic amyloidosis. Mayo Clin Proc 1993;68:7637.
66. Tada S, Iida M, Iwashita A, et al. Endoscopic and biopsy findings of the upper digestive tract
in patients with amyloidosis. Gastrointest Endosc 1990;36:104.
67. Casey KM, Quigley TM, Kozarek RA, Raker EJ. Lethal nature of ischemic gastropathy. Am J
Surg 1993;165:6469.
68. Nuutinen H, Vuoristo M, Farkkila M, et al. Hemorrhagic gastropathy in epidemic nephropathy.
69. Fishbein VA, Rosen AM, Lack EE, et al. Diffuse hemorrhagic gastroenteropathy: Report of a
new entity. Gastroenterology 1994;106:5005.
70. Schwartz AE, Vanagunas A, Kamel PL. Endoscopy to evaluate gastrointestinal bleeding in
marathon runners. Ann Intern Med 1990;113:6323.
71. Gaudin C, Zerath E, Guezennec CY. Gastric lesions secondary to long-distance running. Dig
Dis Sci 1990;35:123943.
72. Wright JP. Exercise and the gastrointestinal tract. S Afr Med J 1993;83:502.
73. Payen JL, Cals P, Voigt JJ, et al. Severe portal hypertensive gastropathy and antral vascular
ectasia are distinct entities in patients with cirrhosis. Gastroenterology 1995;108:13844.
74. Parikh SS, Desai SB, Prabhu SR, et al. Congestive gastropathy: Factors influencing
development, endoscopic features, Helicobacter pylori infection, and microvessel changes.
75. Hoare AM, Jones EL, Alexander Williams J, Hawkins CF. Symptomatic significance of gastric
mucosal changes after surgery for peptic ulcer. Gut 1977;18:295300.
76. Lambert R, Andre C, Moulinier B, Bugnon B. Diffuse varioliform gastritis. Digestion
1978;17:15967.
77. Rutgeerts L, Stuer A, Vandenborre K, et al. Lymphocytic gastritis. Clinical and endoscopic
presentation and long-term follow-up. Acta Gastroenterol Belg 1995;58:23842.
78. Pazzi P, Gamberini S, Scagliarini R, et al. Misoprostol for the treatment of chronic erosive
gastritis: A double-blind placebo-controlled trial. Am J Gastroenterol 1994;89:100713.
79. Franzin G, Manfrini C, Musola R, et al. Chronic erosions of the stomachA clinical,
endoscopic and histological evaluation. Endoscopy 1984;16:15.
80. Haot J, Berger F, Andre C, et al. Lymphocytic gastritis versus varioliform gastritis. A historical
series revisited. J Pathol 1989;158:1922.
81. Haot J, Jouret A, Willette M, et al. Lymphocytic gastritisProspective study of its relationship
with varioliform gastritis. Gut 1990;31:2825.
82. Haot J, Hamichi L, Wallez L, Mainguet P. Lymphocytic gastritis: A newly described entity: A
retrospective endoscopic and histological study. Gut 1988;29:125864.
83. Dixon MF, Wyatt JI, Burke DA, Rathbone BJ. Lymphocytic gastritisRelationship to
Campylobacter pylori infection. J Pathol 1988;154:12532.
84. Sauerbruch T, Schreiber MA, Schussler P, Permanetter W. Endoscopy in the diagnosis of
gastritis. Diagnostic value of endoscopic criteria in relation to histological diagnosis.
Endoscopy 1984;16:1014.
85. Laine L, Cohen H, Sloane R, et al. Interobserver agreement and predictive value of
endoscopic findings for H. pylori and gastritis in normal volunteers. Gastrointest Endosc
1995;42:4203.
86. Weinstein WM. Vital staining of esophageal and gastric mucosa not vital but may be helpful.
87.
Weinstein WM. Precursor lesions for cardiac cancer. Gastrointest Endosc Clin N Am
1997;7:1228.
88. Terruzzi V, Minoli G, Butti GC, Rossini A. Gastric lipid islands in the gastric stump and in
non-operated stomach. Endoscopy 1980;12:5862.
89. Domellof L, Eriksson S, Helander HF, Janunger KG. Lipid islands in the gastric mucosa after
resection for benign ulcer disease. Gastroenterology 1977;72:148.
90. Lechago J. Lipid islands of the stomach: An insular issue?. Gastroenterology
1996;110:6302.
91. Kaiserling E, Heinle H, Itabe H, et al. Lipid islands in human gastric mucosa: Morphological
and immunohistochemical findings. Gastroenterology 1996;110:36974.
92. Ludvikova M, Michal M, Datkova D. Gastric xanthelasma associated with diffuse signet ring
carcinoma. A potential diagnostic problem. Histopathology 1994;25:5812.
93. Correa P. Helicobacter pylori and gastric carcinogenesis. Am J Surg Pathol 1995;19(suppl
1):S3743.
94. Graham DY. Benefits from elimination of Helicobacter pylori infection include major reduction
in the incidence of peptic ulcer disease, gastric cancer, and primary gastric lymphoma. Prev
Med 1994;23:71226.
95. Goodman KJ, Correa P. The transmission of Helicobacter pylori. A critical review of the
evidence. Int J Epidemiol 1995;24:87587.
96. Blaser MJ. Helicobacter pylori phenotypes associated with peptic ulceration. Scand J
97. Tytgat GNJ, Noach LA, Rauws EAJ. Helicobacter pylori infection and duodenal ulcer disease.
Gastroenterol Clin North Am 1993;22:12739.
98. Walsh JH, Peterson WL. Treatment of Helicobacter pylori infection in the management of
peptic ulcer disease. N Engl J Med 1995;333:98491.
99. Isaacson PG. Gastrointestinal lymphoma. Hum Pathol 1994;25:10209.
100. Bayerdorffer E, Neubauer A, Rudolph B, et al. Regression of primary gastric lymphoma of
mucosa-associated lymphoid tissue type after cure of Helicobacter pylori infection. MALT
Lymphoma Study Group. Lancet 1995;345:15914.
101. Roggero E, Zucca E, Pinotti G, et al. Eradication of Helicobacter pylori infection in primary
low-grade gastric lymphoma of mucosa-associated lymphoid tissue. Ann Intern Med
1995;122:7679.
102. Montalban C, Castrillo JM, Abraira V, et al. Gastric B-cell mucosa-associated lymphoid tissue
(MALT) lymphoma. Clinicopathological study and evaluation of the prognostic factors in 143
patients. Ann Oncol 1995;6:35562.
103. Slomianski A, Schubert T, Cutler AF. [13C] urea breath test to confirm eradication of
Helicobacter pylori. Am J Gastroenterol 1995;90:2246.
104. Hassall E, Dimmick JE. Unique features of Helicobacter pylori disease in children. Dig Dis Sci
1991;36:41723.
105. Rocha GA, Queiroz DM, Mendes EN, et al. Helicobacter pylori acute gastritis: Histological,
endoscopical, clinical, and therapeutic features. Am J Gastroenterol 1991;86:15925.
106. Lavelle JP, Landas S, Mitros FA, Conklin JL. Acute gastritis associated with spiral organisms
from cats. Dig Dis Sci 1994;39:74450.
107. Yang H, Dixon MF, Li X, et al. Acute gastritis associated with infection of large spiral-shaped
bacteria. Am J Gastroenterol 1995;90:3079.
108. Komorowski RA, Caya JG, Geenen JE. The morphologic spectrum of large gastric folds:
Utility of the snare biopsy. Gastrointest Endosc 1986;32:1902.
109. Stolte M, Batz C, Eidt S. Giant fold gastritisA special form of Helicobacter pylori associated
gastritis. Z Gastroenterol 1993;31:28993.
110. Bayerdorffer E, Ritter MM, Hatz R, et al. Healing of protein losing hypertrophic gastropathy by
eradication of Helicobacter pyloriIs Helicobacter pylori a pathogenic factor in Mntrier's
disease?. Gut 1994;35:7014.
111. Avunduk C, Navab F, Hampf F, et al. Prevalence of Helicobacter pylori infection in patients
with large gastric folds: Evaluation and follow-up with endoscopic ultrasound before and after
antimicrobial therapy. Am J Gastroenterol 1995;90:196973.
112. Logan RP, Walker MM, Misiewicz JJ, et al. Changes in the intragastric distribution of
Helicobacter pylori during treatment with omeprazole. Gut 1995;36:126.
113. Robert ME, Weinstein WM. Helicobacter pylori-associated gastric pathology. Gastroenterol
114. Genta RM, Robason GO, Graham DY. Simultaneous visualization of Helicobacter pylori and
gastric morphology: A new stain. Hum Pathol 1994;25:2216.
115. Genta RM, Hamner HW, Graham DY. Gastric lymphoid follicles in Helicobacter pylori
infection: Frequency, distribution, and response to triple therapy. Hum Pathol
1993;24:57783.
116. Eidt S, Stolte M. Prevalence of lymphoid follicles and aggregates in Helicobacter pylori
gastritis in antral and body mucosa. J Clin Pathol 1993;46:8325.
117. Genta RM, Lew GM, Graham DY. Changes in the gastric mucosa following eradication of
Helicobacter pylori. Mod Pathol 1993;6:2819.
118. Witteman EM, Mravunac M, Becx MJ, et al. Improvement of gastric inflammation and
resolution of epithelial damage one year after eradication of Helicobacter pylori. J Clin Pathol
1995;48:2506.
119. Kuipers EJ, Uyterlinde AM, Pena AS, et al. Increase of Helicobacter pylori-associated corpus
gastritis during acid suppressive therapy: Implications for long-term safety. Am J
120. Marano BJ Jr, Smith F, Bonanno CA. Helicobacter pylori prevalence in acquired
immunodeficiency syndrome. Am J Gastroenterol 1993;88:68790.
121. Vaira D, Miglioli M, Menegatti M, et al. Helicobacter pylori status, endoscopic findings, and
serology in HIV-1-positive patients. Dig Dis Sci 1995;40:16226.
122. Colin-Jones DG. Dyspepsia update. Scand J Gastroenterol Suppl 1995;210:325.
123. Silverstein MD, Petterson T, Talley NJ. Initial endoscopy or empirical therapy with or without
testing for Helicobacter pylori for dyspepsia: A decision analysis. Gastroenterology
1996;110:7283.
124. Talley NJ. A critique of therapeutic trials in Helicobacter pylori-positive functional dyspepsia.
125. Patel P, Khulusi S, Mendall MA, et al. Prospective screening of dyspeptic patients by
Helicobacter pylori serology. Lancet 1995;346:13158.
126. Sobala GM, King RF, Axon AT, Dixon MF. Reflux gastritis in the intact stomach. J Clin Pathol
1990;43:3036.
127.
Sipponen P, Saukkonen M, Varis K, Siurala M. Morphometric analysis of the loss of chief and
parietal cells after partial gastrectomy for duodenal ulcer. Acta Hepatogastroenterol (Stutt)
1979;26:4938.
128. Weinstein WM, Buch KL, Elashoff J, et al. The histology of the stomach in symptomatic
patients after gastric surgery: A model to assess selective patterns of gastric mucosal injury.
Scand J Gastroenterol Suppl 1985;109:7789.
129. Dixon MF, O'Connor HJ, Axon AT, et al. Reflux gastritis: Distinct histopathological entity?. J
Clin Pathol 1986;39:52430.
130. Niv Y, Turani H. Cystic changes in gastric glands after gastric surgery and in the intact
stomach. J Clin Gastroenterol 1991;13:4659.
131. Offerhaus GJ, Tersmette AC, Giardiello FM, et al. Evaluation of endoscopy for early detection
of gastric-stump cancer. Lancet 1992;340:335.
132. Wolfsen HC, Carpenter HA, Talley NJ. Mntrier's disease: A form of hypertrophic
gastropathy or gastritis?. Gastroenterology 1993;104:13109.
133. Wolber R, Owen D, Freeman H. Colonic lymphocytosis in patients with celiac sprue. Hum
Pathol 1990;21:10926.
134. Griffiths AP, Wyatt J, Jack AS, Dixon MF. Lymphocytic gastritis, gastric adenocarcinoma, and
primary gastric lymphoma. J Clin Pathol 1994;47:11234.
135. Samloff IM, Varis K, Ihamaki T, et al. Relationships among serum pepsinogen I, serum
pepsinogen II, and gastric mucosal histology. A study in relatives of patients with pernicious
anemia. Gastroenterology 1982;83:2049.
136. Svendsen JH, Dahl C, Svendsen LB, Christiansen PM. Gastric cancer risk in achlorhydric
patients. A long-term follow-up study. Scand J Gastroenterol 1986;21:1620.
137. Meshkinpour H, Orlando RA, Arguello JF, DeMicco MP. Significance of endoscopically visible
blood vessels as an index of atrophic gastritis. Am J Gastroenterol 1979;71:3769.
138. Farinati F, Cardin F, Di Mario F, et al. Perendoscopic gastric pH determination. Simple
method for increasing accuracy in diagnosing chronic atrophic gastritis. Gastrointest Endosc
1987;33:2937.
139. Andersen J, Strom M. A technique for screening of achlorhydria and hypochlorhydria during
upper gastrointestinal endoscopy. Scand J Gastroenterol 1990;25:10848.
140. Armbrecht U, Stockbrugger RW, Rode J, et al. Development of gastric dysplasia in pernicious
anaemia: A clinical and endoscopic follow up study of 80 patients. Gut 1990;31:11059.
141. Borch K. Epidemiologic, clinicopathologic, and economic aspects of gastroscopic screening of
patients with pernicious anemia. Scand J Gastroenterol 1986;21:2130.
142. Loffeld BC, Van Spreeuwel JP. The gastrointestinal tract in pernicious anemia. Dig Dis
1991;9:707.
143. Bordi C, D'Adda T, Azzoni C, et al. Hypergastrinemia and gastric enterochromaffin-like cells.
Am J Surg Pathol 1995;19(suppl 1):S819.
144. Solcia E, Fiocca R, Villani L, et al. Hyperplastic, dysplastic, and neoplastic
enterochromaffin-like cell proliferations of the gastric mucosa. Classification and histogenesis.
Am J Surg Pathol 1995;19(suppl 1):S17.
145. Rode J, Dhillon AP, Papadaki L, et al. Pernicious anaemia and mucosal endocrine cell
proliferation of the non-antral stomach. Gut 1986;27:78998.
146. Borch K, Renvall H, Liedberg G. Gastric endocrine cell hyperplasia and carcinoid tumors in
pernicious anemia. Gastroenterology 1985;88:63848.
147. Roucayrol AM, Cattan D. Evolution of fundic argyrophil cell hyperplasia in nonantral atrophic
gastritis. Gastroenterology 1990;99:130714.
148. Modlin IM, Gilligan CJ, Lawton GP, et al. Gastric carcinoids. The Yale experience. Arch Surg
1995;130:2505, discussion 2556.
149. Kornfeld HR, Sherlock P, Hertz R. Multiple carcinoid tumors of the stomach with endoscopic
observations. Gastrointest Endosc 1971;18:747.
150. Stockbrugger RW, Menon GG, Beilby JO, et al. Gastroscopic screening in 80 patients with
pernicious anaemia. Gut 1983;24:11417.
151. Sjoblom SM, Sipponen P, Jarvinen H. Gastroscopic follow up of pernicious anaemia patients.
Gut 1993;34:2832.
152. Sjoblom SM, Sipponen P, Miettinen M, et al. Gastroscopic screening for gastric carcinoids
and carcinoma in pernicious anemia. Endoscopy 1988;20:526.
153. Weinstein WM, Goldstein NS. Gastric dysplasia and its management. Gastroenterology
1994;107:15435.
154. Fee MJ, Oo MM, Gabayan AE, et al. Abdominal tuberculosis in patients infected with the
human immunodeficiency virus. Clin Infect Dis 1995;20:93844.
155. Klatt EC, Nichols L, Noguchi TT. Evolving trends revealed by autopsies of patients with the
acquired immunodeficiency syndrome: 565 autopsies in adults with the acquired
immunodeficiency syndrome, Los Angeles, Calif, 19821993. Arch Pathol Lab Med
1994;118:88490.
156. Marshall JB. Tuberculosis of the gastrointestinal tract and peritoneum. Am J Gastroenterol
1993;88:98999.
157. Tromba JL, Inglese R, Rieders B, Todaro R. Primary gastric tuberculosis presenting as pyloric
outlet obstruction. Am J Gastroenterol 1991;86:18202.
158. Perez-Piqueras J, Coca S, Silva C, et al. Isolated gastric tuberculosis: A case report.
159. Mathis G, Dirschmid K, Sutterlutti G. Tuberculous gastric ulcer. Endoscopy 1987;19:1335.
160. Misra RC, Agarwal SK, Prakash P, et al. Gastric tuberculosis. Endoscopy 1982;14:2357.
161. Subei I, Attar B, Schmitt G, Levendoglu H. Primary gastric tuberculosis: A case report and
literature review. Am J Gastroenterol 1987;82:76972.
162. Anand BS, Schneider FE, El-Zaatari FA, et al. Diagnosis of intestinal tuberculosis by
polymerase chain reaction on endoscopic biopsy specimens. Am J Gastroenterol
1994;89:22489.
163. Greenstein DB, Wilcox CM, Schwartz DA. Gastric syphilis. Report of seven cases and review
of the literature. J Clin Gastroenterol 1994;18:49.
164. Ikebe M, Oiwa T, Mori M, et al. Gastric syphilis: Case report and review of the literature.
Radiat Med 1994;12:1715.
165. Fyfe B, Poppiti RJ Jr, Lubin J, Robinson MJ. Gastric syphilis. Primary diagnosis by gastric
biopsy report of four cases. Arch Pathol Lab Med 1993;117:8203.
166. Long BW, Johnston JH, Wetzel W, et al. Gastric syphilis: Endoscopic and histological
features mimicking lymphoma. Am J Gastroenterol 1995;90:15047.
167. Bottari M, Melina D, Napoli P, et al. Gastric lesions in secondary syphilis. Gastrointest Endosc
1988;34:4379.
168. Beckman JW, Schuman BM. Antral gastritis and ulceration in a patient with secondary
syphilis. Gastrointest Endosc 1986;32:3556.
169. Abdu RA, Carter K, Pomidor WJ. Gastric syphilis mimicking linitis plastica. Arch Surg
1993;128:1034.
170. Bigotti G, Coli A. Gastric syphilis simulating malignant lymphoma. Pathologica
1991;83:21722.
171. Anai H, Okada Y, Okubo K, et al. Gastric syphilis simulating linitis plastica type of gastric
172. Winters HA, Notar-Francesco V, Bromberg K, et al. Gastric syphilis: Five recent cases and a
review of the literature. Ann Intern Med 1992;116:3149.
173.
Starr A, Wilson JM. Phlegmonous gastritis. Ann Surg 1957;145:88.
174. Miller AI, Smith B, Rogers AI. Phlegmonous gastritis. Gastroenterology 1975;68:2318.
175. Williford ME, Foster WL Jr, Halvorsen RA, Thompson WM. Emphysematous gastritis
secondary to disseminated stronglyloidiasis. Gastrointest Radiol 1982;7:1236.
176. Mittleman RE, Suarez RV. Phlegmonous gastritis associated with the acquired
immunodeficiency syndrome/pre-acquired immunodeficiency syndrome. Arch Pathol Lab Med
1985;109:7657.
177. Stein LB, Greenberg RE, Ilardi CF, et al. Acute necrotizing gastritis in a patient with peptic
ulcer disease. Am J Gastroenterol 1989;84:15524.
178. Moosvi AR, Saravolatz LD, Wong DH, Simms SM. Emphysematous gastritis: Case report and
review. Rev Infect Dis 1990;12:84855.
179. McKelvie PA, Fink MA. A fatal case of emphysematous gastritis and esophagitis. Pathology
1994;26:4902.
180. Aviles JF, Fernandez-Seara J, Barcena R, et al. Localized phlegmonous gastritis: Endoscopic
view. Endoscopy 1988;20:389.
181. Strauss RJ, Friedman M, Platt N, et al. Gangrene of the stomach: A case of acute necrotizing
gastritis. Am J Surg 1978;135:2537.
182. Hornig D, Khn H, Stadelmann O, Btticher R. Phlegmonous gastritis after Indian ink
marking. Endoscopy 1983;15:2669.
183. Lifton LJ, Schlossberg D. Phlegmonous gastritis after endoscopic polypectomy. Ann Intern
Med 1982;97:3735.
184.
Gonzalez LL, Schowengerdt C, Skinner HH, et al. Emphysematous gastritis. Surg Gynecol
Obstet 1963;116:79.
185. Binmoeller KF, Benner KG. Emphysematous gastritis secondary to gastric infarction. Am J
186. Skoutelis A, Panagopoulos C, Kalfarentzos F, Bassaris H. Intramural gastric actinomycosis.
South Med J 1995;88:64750.
187. Wakayama T, Watanabe H, Ishizaki Y, et al. A case of phlegmonous esophagitis associated
with diffuse phlegmonous gastritis. Am J Gastroenterol 1994;89:8046.
188. Bron BA, Deyhle P, Pelloni S, et al. Phlegmonous gastritis diagnosed by endoscopic snare
biopsy. Am J Dig Dis 1977;22:72933.
189. Hackman RC, Wolford JL, Gleaves CA, et al. Recognition and rapid diagnosis of upper
gastrointestinal cytomegalovirus infection in marrow transplant recipients. A comparison of
seven virologic methods. Transplantation 1994;57:2317.
190. Chandrasekar PH, Weinmann A, Shearer C. Autopsy-identified infections among bone
marrow transplant recipients: A clinicopathologic study of 56 patients. Bone Marrow
Transplant 1995;16:67581.
191. Arabia FA, Rosado LJ, Huston CL, et al. Incidence and recurrence of gastrointestinal
cytomegalovirus infection in heart transplantation. Ann Thorac Surg 1993;55:811.
192. Komorowski RA, Cohen EB, Kauffman HM, Adams MB. Gastrointestinal complications in
renal transplant recipients. Am J Clin Pathol 1986;86:1617.
193. McDonald GB, Shulman HM, Sullivan KM, Spencer GD. Intestinal and hepatic complications
of human bone marrow transplantation. Part II. Gastroenterology 1986;90:77084.
194. Alexander JA, Cuellar RE, Fadden RJ, et al. Cytomegalovirus infection of the upper
gastrointestinal tract before and after liver transplantation. Transplantation 1988;46:37882.
195. Manez R, Kusne S, Green M, et al. Incidence and risk factors associated with the
development of cytomegalovirus disease after intestinal transplantation. Transplantation
1995;59:10104.
196. Campbell DA, Piercey JR, Shnitka TK, et al. Cytomegalovirus-associated gastric ulcer.
197. Kinoshita Y, Tojo M, Yano T, et al. Cytomegalovirus mononucleosis-associated gastric ulcers
in normal host. Gastroenterol Jpn 1993;28:8894.
198. Rotterdam H, Tsang P. Gastrointestinal disease in the immunocompromised patient. Hum
Pathol 1994;25:112340.
199. Smith MA, Brennessel DJ. Cytomegalovirus. Infect Dis Clin North Am 1994;8:42738.
200. Yokose N, Tanabe Y, An E, et al. Acute gastric mucosal lesions associated with
cytomegalovirus infection in a non-immunocompromised host. Intern Med 1995;34:8835.
201. Arnar DO, Gudmundsson G, Theodors A, et al. Primary cytomegalovirus infection and gastric
ulcers in normal host. Dig Dis Sci 1991;36:10811.
202. Laguna F, Garcia-Samaniego J, Alonso MJ, et al. Pseudotumoral appearance of
cytomegalovirus esophagitis and gastritis in AIDS patients. Am J Gastroenterol
1993;88:110811.
203. Shuster LD, Cox G, Bhatia P, Miner PB Jr. Gastric mucosal nodules due to cytomegalovirus
infection. Dig Dis Sci 1989;34:1037.
204. Kodama T, Fukuda S, Takino T, et al. Gastroduodenal cytomegalovirus infection after renal
transplantation. Fiberscopic observations. Endoscopy 1985;17:1578.
205. Rich JD, Crawford JM, Kazanjian SN, Kazanjian PH. Discrete gastrointestinal mass lesions
caused by cytomegalovirus in patients with AIDS: Report of three cases and review. Clin
Infect Dis 1992;15:60914.
206. Zucker GM, Otis C, Korowski K, Navab F. Cytomegalovirus gastritis associated with
pseudolymphoma. J Clin Gastroenterol 1994;18:2226.
207. Howiler W, Goldberg HI. Gastroesophageal involvement in herpes simplex. Gastroenterology
1976;70:7758.
208. Buss DH, Scharyj M. Herpesvirus infection of the esophagus and other visceral organs in
adults. Incidence and clinical significance. Am J Med 1979;66:45762.
209. Lohr JM, Nelson JA, Oldstone MB. Is herpes simplex virus associated with peptic ulcer
disease?. J Virol 1990;64:216874.
210. Durin S, Chapelon C, Ziza JM, et al. Gastric zoster ulcerations. Presse Med 1988;17:2112.
211. Kebede D, Barthel JS, Singh A. Transient gastroparesis associated with cutaneous herpes
zoster. Dig Dis Sci 1987;32:31822.
212. McCluggage WG, Fox JD, Baillie KE, et al. Varicella zoster gastritis in a bone marrow
transplant recipient. J Clin Pathol 1994;47:10546.
213. Kusne S, Tobin D, Pasculle AW, et al. Candida carriage in the alimentary tract of liver
transplant candidates. Transplantation 1994;57:398402.
214. Eras P, Goldstein MJ, Sherlock P. Candida infection of the gastrointestinal tract. Medicine
(Baltimore) 1972;51:36779.
215. Katzenstein AL, Maksem J. Candidal infection of gastric ulcers. Histology, incidence, and
clinical significance. Am J Clin Pathol 1979;71:13741.
216. Peters M, Weiner J, Whelan G. Fungal infection associated with gastroduodenal ulceration:
Endoscopic and pathologic appearances. Gastroenterology 1980;78:3504.
217. Gotlieb-Jensen K, Andersen J. Occurrence of Candida in gastric ulcers. Significance for the
healing process. Gastroenterology 1983;85:5357.
218. Minoli G, Terruzzi V, Rossini A. Gastroduodenal candidiasis occurring without underlying
diseases (primary gastroduodenal candidiasis). Report on two cases. Endoscopy
1979;11:1822.
219. Nelson RS, Bruni HC, Goldstein HM. Primary gastric candidiasis in uncompromised subjects.
220. Morishita T, Kamiya T, Munakata Y, Tsuchiya M. Radiologic and endoscopic studies of gastric
ulcers associated with Candida infection. Acta Gastroenterol Latinoam 1993;23:2239.
221. Minoli G, Terruzzi V, Butti G, et al. Gastric candidiasis: An endoscopic and histological study
in 26 patients. Gastrointest Endosc 1982;28:5961.
222. Di Febo G, Miglioli M, Calo G, et al. Candida albicans infection of gastric ulcer: Frequency and
correlation with medical treatment. Results of a multicenter study. Dig Dis Sci
1985;30:17881.
223.
Gottlieb K, Iber FL, Livak A, et al. Oral Candida colonizes the stomach and gastrostomy
feeding tubes. JPEN 1994;18:2647.
224. Becherer PR, Sokol-Anderson M, Joist JH, Milligan T. Gastrointestinal histoplasmosis
presenting as hematochezia in human immunodeficiency virus-infected hemophilic patients.
Am J Hematol 1994;47:22931.
225. Heneghan SJ, Li J, Petrossian E, Bizer LS. Intestinal perforation from gastrointestinal
histoplasmosis in acquired immunodeficiency syndrome. Case report and review of the

literature. Arch Surg 1993;128:4646.
226. Ankobiah WA, Vaidya K, Powell S, et al. Disseminated histoplasmosis in AIDS.
Clinicopathologic features in seven patients from a non-endemic area. N Y State J Med
1990;90:2348.
227. Orchard JL, Luparello F, Brunskill D. Malabsorption syndrome occurring in the course of
disseminated histoplasmosis: Case report and review of gastrointestinal histoplasmosis. Am J
Med 1979;66:3316.
228. Cappell MS, Mandell W, Grimes MM, Neu HC. Gastrointestinal histoplasmosis. Dig Dis Sci
1988;33:35360.
229. Boulard F, Cnudde F, Teron-Aboud B, et al. Digestive histoplasmosis in a patient with AIDS
from Guadeloupe. Ann Gastroenterol Hepatol (Paris) 1994;30:97100.
230. Lyon DT, Schubert TT, Mantia AG, Kaplan MH. Phycomycosis of the gastrointestinal tract.
231. Nagy-Agren SE, Chu P, Smith GJ, et al. Zygomycosis (mucormycosis) and HIV infection:
Report of three cases and review. J Acquir Immune Defic Syndr Hum Retrovirol
1995;10:4419.
232. Brullet E, Andreu X, Elias J, et al. Gastric mucormycosis in a patient with acquired
immunodeficiency syndrome. Gastrointest Endosc 1993;39:1067.
233. Prescott RJ, Haboubi NY, Burton IE. Gastrointestinal tract aspergilloma: Possible cause of
malabsorption. J Clin Pathol 1994;47:1701.
234. Van Olmen G, Larmuseau MF, Geboes K, et al. Primary gastric actinomycosis: A case report
and review of the literature. Am J Gastroenterol 1984;79:5126.
235. Massimillo AJ, Chang J, Freedman L, et al. Cryptosporidium gastropathy. Presentation of a
case and review of the literature. Dig Dis Sci 1995;40:18690.
236. Besasso HR, Macias J, Trombetta LA, et al. Erosive gastritis associated with Cryptosporidium
in 2 patients with AIDS. Acta Gastroenterol Latinoam 1994;24:413.
237. Forester G, Sidhom O, Nahass R, Andavolu R. AIDS-associated cryptosporidiosis with gastric
stricture and a therapeutic response to paromomycin. Am J Gastroenterol 1994;89:10968.
238. Bouree P, Paugam A, Petithory JC. Anisakidosis: Report of 25 cases and review of the
literature. Comp Immunol Microbiol Infect Dis 1995;18:7584.
239. Kakizoe S, Kakizoe H, Kakizoe K, et al. Endoscopic findings and clinical manifestation of
gastric anisakiasis. Am J Gastroenterol 1995;90:7613.
240. Songur Y, Okai T, Watanabe H, et al. Endosonographic evaluation of giant gastric folds.
241. Sugimachi K, Inokuchi K, Ooiwa T, et al. Acute gastric anisakiasis. Analysis of 178 cases.
JAMA 1985;253:10123.
242. Sakai K, Ohtani A, Muta H, et al. Endoscopic ultrasonography findings in acute gastric
anisakiasis. Am J Gastroenterol 1992;87:161823.
243. Okai T, Mouri I, Yamaguchi Y, et al. Acute gastric anisakiasis: Observations with endoscopic
244. Lopez-Velez R, Cano Ruiz A, Garcia Camacho A, et al. Pseudoterranova decipiens larva
infection mimicking a chronic peptic ulcer. Am J Gastroenterol 1992;87:9334.
245. Sohn WM, Seol SY. A human case of gastric anisakiasis by Pseudoterranova decipiens larva.
Korean J Parasitol 1994;32:536.
246. Dumont A, Seferian V, Barbier P. Endoscopic discovery and capture of Necator americanus in
the stomach. Endoscopy 1983;15:656.
247. Munoz-Navas M, Macias E, Garcia-Villarreal L, et al. Endoscopic diagnosis and extraction of
gastric parasites. Endoscopy 1993;25:491.
248. Choudhuri G, Saha SS, Tandon RK. Gastric ascariasis. Am J Gastroenterol 1986;81:78890.
249. Doglioni C, De Boni M, Cielo R, et al. Gastric giardiasis. J Clin Pathol 1992;45:9647.
250. Quincey C, James PD, Steele RJ. Chronic giardiasis of the stomach. J Clin Pathol
1992;45:103941.
251. Berney DM, Rampton D, van der Walt JD. Giardiasis of the stomach. Postgrad Med J
1994;70:2378.
252. Mannheimer SB, Soave R. Protozoal infections in patients with AIDS. Cryptosporidiosis,
isosporiasis, cyclosporiasis, and microsporidiosis. Infect Dis Clin North Am 1994;8:48398.
253. Schantz PM. The dangers of eating raw fish. N Engl J Med 1989;320:11435.
254. Wittner M, Turner JW, Jacquette G, et al. EustrongyloidiasisA parasitic infection acquired
by eating sushi. N Engl J Med 1989;320:11246.
255. Simpson WG, Gerhardstein DC, Thompson JR. Disseminated Strongyloides stercoralis
infection. South Med J 1993;86:8215.
256. Scowden EB, Schaffner W, Stone WJ. Overwhelming strongyloidiasis: An unappreciated
opportunistic infection. Medicine (Baltimore) 1978;57:52744.
257. Celedon JC, Mathur-Wagh U, Fox J, et al. Systemic strongyloidiasis in patients infected with
the human immunodeficiency virus. A report of 3 cases and review of the literature. Medicine
(Baltimore) 1994;73:25663.
258. Wurtz R, Mirot M, Fronda G, et al. Short report gastric infection by Strongyloides stercoralis.
Am J Trop Med Hyg 1994;51:33940.
259. Dees A, Batenburg PL, Umar HM, et al. Strongyloides stercoralis associated with a bleeding
gastric ulcer. Gut 1990;31:14145.
260. Barr JR. Strongyloides stercoralis. Can Med Assoc J 1978;118:9335.
261. el Shiekh Mohamed AR, al Karawi MA, Yasawy MI. Organ involvement in hepato-intestinal
schistosomiasis. Hepatogastroenterology 1994;41:3706.
262. Miyazaki A, Yajima Y, Meguro S, et al. A case of schistosomiasis japonica diagnosed by the
strip biopsy of an early gastric cancer (IIa). Nippon Shokakibyo Gakkai Zasshi
1994;91:18892.
263. Otrakji CL, Albores-Saavedra J, Martinez AJ. Gastric malignant lymphoma with superimposed
amebiasis. Am J Gastroenterol 1990;85:725.
264. Thamprasert K, Khunamornpong S, Morakote N. Acanthamoeba infection of peptic ulcer. Ann
Trop Med Parasitol 1993;87:4035.
265. Thomas S, Mishra MC, Kriplani AK, Kapur BM. Hydatidemesis: A bizarre presentation of
abdominal hydatidosis. Aust N Z J Surg 1993;63:4968.
266.
Oksenhendler E, Cadranel J, Sarfati C, et al. Toxoplasma gondii pneumonia in patients with
the acquired immunodeficiency syndrome. Am J Med 1990;88:18N-21N.
267. Laguna F, Garcia-Samaniego J, Soriano V, et al. Gastrointestinal leishmaniasis in human
immunodeficiency virus-infected patientsReport of five cases and review. Clin Infect Dis
1994;19:4853.
268. Gradoni L, Guaraldi G, Codeluppi M, et al. Gastric localization of Leishmania in a patient with
acquired immunodeficiency syndrome. A case report. APMIS 1995;103:258.
269. Romero A, Matos C, Gonzalez MM, et al. Changes in gastric mucosa in acute malaria. G E N
1993;47:1238.
270. Cary ER, Tremaine WJ, Banks PM, et al. Isolated Crohn's disease of the stomach. Mayo Clin
Proc 1989;64:7769.
271. Halme L, Karkkainen P, Rautelin H, et al. High frequency of helicobacter-negative gastritis in
patients with Crohn's disease. Gut 1996;38:37983.
272. Keller KM, Bender SW, Kirchmann H, et al. Diagnostic significance of epithelioid granulomas
in Crohn's disease in children. Multicenter Paediatric Crohn's Disease Study Group. J Pediatr
Gastroenterol Nutr 1990;10:2732.
273. Ruuska T, Vaajalahti P, Arajarvi P, Maki M. Prospective evaluation of upper gastrointestinal
mucosal lesions in children with ulcerative colitis and Crohn's disease. J Pediatr Gastroenterol
Nutr 1994;19:1816.
274. Alcantara M, Rodriguez R, Potenciano JL, et al. Endoscopic and bioptic findings in the upper
gastrointestinal tract in patients with Crohn's disease. Endoscopy 1993;25:2826.
275. Schmidt-Sommerfeld E, Kirschner BS, Stephens JK. Endoscopic and histologic findings in the
upper gastrointestinal tract of children with Crohn's disease. J Pediatr Gastroenterol Nutr
1990;11:44854.
276. Tanaka M, Kimura K, Sakai H, et al. Long-term follow-up for minute gastroduodenal lesions in
Crohn's disease. Gastrointest Endosc 1986;32:2069.
277. Mashako MN, Cezard JP, Navarro J, et al. Crohn's disease lesions in the upper
gastrointestinal tract: Correlation between clinical, radiological, endoscopic, and histological
features in adolescents and children. J Pediatr Gastroenterol Nutr 1989;8:4426.
278. Rutgeerts P, Onette E, Vantrappen G, et al. Crohn's disease of the stomach and duodenum:
A clinical study with emphasis on the value of endoscopy and endoscopic biopsies.
Endoscopy 1980;12:28894.
279. Pichney LS, Fantry GT, Graham SM. Gastrocolic and duodenocolic fistulas in Crohn's
disease. J Clin Gastroenterol 1992;15:20511.
280. Johnstone JM, Morson BC. Eosinophilic gastroenteritis. Histopathology 1978;2:33548.
281. Klein NC, Hargrove RL, Sleisenger MH, Jeffries GH. Eosinophilic gastroenteritis. Medicine
(Baltimore) 1970;49:299319.
282. Lee CM, Changchien CS, Chen PC, et al. Eosinophilic gastroenteritis: 10 years experience.
283. Talley NJ, Shorter RG, Phillips SF, Zinsmeister AR. Eosinophilic gastroenteritis: A
clinicopathological study of patients with disease of the mucosa, muscle layer, and subserosal
tissues. Gut 1990;31:548.
284. Leinbach GE, Rubin CE. Eosinophilic gastroenteritis: A simple reaction to food allergens?.
285. Navab F, Kleinman MS, Algazy K, et al. Endoscopic diagnosis of eosinophilic gastritis.
286. Ormeci N, Bayramoglu F, Tulunay O, et al. Cancer-like eosinophilic gastritis. Endoscopy
1994;26:509.
287. Katz AJ, Goldman H, Grand RJ. Gastric mucosal biopsy in eosinophilic (allergic)
gastroenteritis. Gastroenterology 1977;73:7059.
288. Sprague R, Harper P, McClain S, et al. Disseminated gastrointestinal sarcoidosis. Case report
and review of the literature. Gastroenterology 1984;87:4215.
289. Hogg SG. Case report: Gastric sarcoid simulating linitis plasticaA 5-year follow-up study.
Clin Radiol 1991;44:2778.
290. Chinitz MA, Brandt LJ, Frank MS, et al. Symptomatic sarcoidosis of the stomach. Dig Dis Sci
1985;30:6828.
291.
Palmer ED. Note on silent sarcoidosis of the gastric mucosa. J Lab Clin Med 1958;52:231.
292. Kyle RA, Greipp PR. Amyloidosis (AL). Clinical and laboratory features in 229 cases. Mayo
Clin Proc 1983;58:66583.
293. Rivera RA. Endoscopic findings in gastroduodenal amyloidosis (case report). Gastrointest
Endosc 1971;17:13744.
294. Coughlin GP, Reiner RG, Grant AK. Endoscopic diagnosis of amyloidosis. Gastrointest
Endosc 1980;26:1545.
295. Steen LE, Oberg L. Familial amyloidosis with polyneuropathy: Roentgenological and
gastroscopic appearance of gastrointestinal involvement. Am J Gastroenterol
1983;78:41720.
296. Ohno F, Numata Y, Yamano T, et al. Gastroscopic biopsy of the stomach for the diagnosis of
amyloidosis. Gastroenterol Jpn 1982;17:41521.
297. Komorowski RA, Caya JG. Hyperplastic gastropathy. Clinicopathologic correlation. Am J Surg
Pathol 1991;15:57785.
298. Tio TL. Large gastric folds evaluated by endoscopic ultrasonography. Gastrointest Endosc
299. Eidt S, Stolte M, Fischer R. Helicobacter pylori gastritis and primary gastric non-Hodgkin's
lymphomas. J Clin Pathol 1994;47:4369.
300. Stolte M, Btz C, Bayerdorffer E, Eidt S. Helicobacter pylori eradication in the treatment and
differential diagnosis of giant folds in the corpus and fundus of the stomach. Z Gastroenterol
1995;33:198201.
Appelman HD. Localized and extensive expansions of the gastric mucosa: Mucosal polyps
and giant folds. In Appelman HD (ed). Pathology of the Esophagus, Stomach and Duodenum.
Comtemporary Issues in Surgical Pathology. New York: Churchill Livingstone, 1984;79119.
302. Fieber SS, Rickert RR. Hyperplastic gastropathy. Analysis of 50 selected cases from
19551980. Am J Gastroenterol 1981;76:3219.
303. Layfield LJ, Reichman A, Weinstein WM. Endoscopically directed fine needle aspiration
biopsy of gastric and esophageal lesions. Acta Cytol 1992;36:6974.
304. Das DK, Pant CS. Fine needle aspiration cytologic diagnosis of gastrointestinal tract lesions.
A study of 78 cases. Acta Cytol 1994;38:7239.
305. Mendis RE, Gerdes H, Lightdale CJ, Botet JF. Large gastric folds: A diagnostic approach
using endoscopic ultrasonography. Gastrointest Endosc 1994;40:43741.
306. Fujimori T, Nakamura T, Hirayama D, et al. Endoscopic mucosectomy for early gastric cancer
using modified strip biopsy. Endoscopy 1992;24:1879.
307. Larsen B, Tarp U, Kristensen E. Familial giant hypertrophic gastritis (Mntrier's disease). Gut
1987;28:151721.
308. Eisenstat DD, Griffiths AM, Cutz E, et al. Acute cytomegalovirus infection in a child with
Mntrier's disease. Gastroenterology 1995;109:5925.
309. Cieslak TJ, Mullett CT, Puntel RA, Latimer JS. Mntrier's disease associated with
cytomegalovirus infection in children: Report of two cases and review of the literature. Pediatr
Infect Dis J 1993;12:3403.
310. Searcy RM, Malagelada JR. Mntrier's disease and idiopathic hypertrophic gastropathy. Ann
Intern Med 1984;100:56570.
311. Johnson MI, Spark JI, Ambrose NS, Wyatt JI. Early gastric cancer in a patient with Mntrier's
disease, lymphocytic gastritis and Helicobacter pylori. Eur J Gastroenterol Hepatol
1995;7:18790.
312. Mosnier JF, Flejou JF, Amouyal G, et al. Hypertrophic gastropathy with gastric
adenocarcinoma: Mntrier's disease and lymphocytic gastritis?. Gut 1991;32:15657.
313. Wood MG, Bates C, Brown RC, Losowsky MS. Intramucosal carcinoma of the gastric antrum
complicating Mntrier's disease. J Clin Pathol 1983;36:10715.
314. Overholt BF, Jeffries GH. Hypertrophic, hypersecretory protein-losing gastropathy.
315. Ranchod M, Lewin KJ, Dorfman RF. Lymphoid hyperplasia of the gastrointestinal tract. A
study of 26 cases and review of the literature. Am J Surg Pathol 1978;2:383400.
316. Eimoto T, Futami K, Naito H, et al. Gastric pseudolymphoma with monotypic cytoplasmic
immunoglobulin. Cancer 1985;55:78893.
317. Weston AP, Campbell DR, McGregor DH, Cherin R. Endoscopic and histologic resolution of
gastric pseudolymphoma (reactive lymphoid hyperplasia) following treatment with bismuth
and oral antibiotics. Dig Dis Sci 1994;39:256774.
318. Ectors NL, Dixon MF, Geboes KJ, et al. Granulomatous gastritis: A morphological and
diagnostic approach. Histopathology 1993;23:5561.
319. Weinstock JV. Idiopathic isolated granulomatous gastritis: Spontaneous resolution without
surgical intervention. Dig Dis Sci 1980;25:2335.
320. Compton CC, Von Lichtenberg F. Necrotizing granulomatous gastritis and gastric perforation
of unknown etiology: A first case report. J Clin Gastroenterol 1983;5:5965.
321. Hanada M, Takami M, Hirata K, Nakajima T. Hyalinoid giant cell gastritis. A unique gastric
lesion associated with eosinophilic hyalinoid degeneration of smooth muscle. Acta Pathol Jpn
1985;35:74958.
301.
Chapter 48 Benign and Malignant Tumors of the Stomach

(4249)
(4250)
RAPHAEL S. K. CHUNG, M.D.

The rational management of benign and malignant tumors of the stomach is critically dependent on
endoscopy. Although the endoscopic contribution has been mainly diagnostic, definitive treatment can
be rendered endoscopically for many lesions, particularly those that are benign.
Diagnostic endoscopy has three objectives with regard to gastric lesions: to examine the gross
pathologic characteristics of the lesion; to obtain tissue samples; and to estimate the extent of the
disease. The latter is necessary for planning treatment and to assess prognosis. Management and
outcome depend not only on a diagnostic biopsy but also on accurate delineation of the extent and
location of the tumor, which, in turn, indicate the specific operation required including choice of incision
(e.g., abdominal or thoracoabdominal). Apart from histologic grade, the prognosis differs significantly
for a carcinoma of the antrum compared with one arising in the cardia.
Although not infallible, endoscopic appearances of mucosal lesions are often diagnostic; at a minimum,
they suggest the best site for tissue sampling and the type of sampling most likely to yield a positive
diagnosis. In essence, the endoscopic recognition of benign and malignant changes is synonymous
with gross pathologic diagnosis. A malignant process often results in ulcerations with irregular,
elevated edges. Necrosis and hemorrhage lend different colors to parts of the lesion, especially in the
ulcerated area. A fungating lesion, one that is bulky and protruding into the lumen, is characteristic of
malignancy. The presence of satellite nodules, with normal or infiltrated intervening mucosa, is virtually
pathognomonic of malignancy. Endoscopic findings that suggest a benign process include rounded
contour, a smooth surface, and uniform color.
The term submucosal lesion or tumor is often used generically to refer to any lesion covered by
normal- or nearly normal-appearing mucosa and having a smoothly elevated contour that makes its
margins less sharply demarcated than lesions that arise from the mucosa. This appearance can be
associated with a wide variety of lesions that arise within and outside the wall of the gastrointestinal
tract as well as normal structures adjacent to the gut. If there is no associated mucosal abnormality, it
is difficult and usually impossible to determine the cause of the lesion at endoscopy; in such cases,
endoscopic ultrasonography (EUS) frequently provides more detailed and useful information (see
Chapter 51: Endoscopic Ultrasonography: Stomach and Duodenum). Nevertheless, some simple
observations at endoscopy can still be useful. At a minimum, it is usually possible to form a preliminary
impression as to whether the lesion is intrinsic or extrinsic to the stomach wall. In general, extrinsic
compression causes a less abrupt indentation of the wall, and therefore, the outline of the protrusion is
less discrete than that produced by a lesion intrinsic to the gut wall. If changing the patient's position or
causing the stomach to collapse by aspiration of air reduces the prominence of the lesion, or even
makes it disappear, impingement from an adjacent organ should be considered. A tortuous or
aneurysmic splenic artery may be evident as a pulsating indentation in the posterior wall body or
fundus of the stomach. Occasionally, a mass in the left lobe of the liver may be appreciated as an
indentation in the gastric body or antrum. Pancreatic pseudocyst often indents the stomach in different
areas and to different degrees; air insufflation may enhance its appreciation.
Motion of a lesion, when observed at endoscopy, is also a useful clue: expansile pulsations indicate an
arterial abnormality; transmitted aortic pulsations suggest extraluminal bulk; transmitted cardiac or
respiratory movements frequently denote attachment to the diaphragm.
Although a lesion cannot be palpated directly at endoscopy, it is often possible to appreciate whether it
is solid or soft through manipulations with a biopsy forceps or the endoscope itself. The general shape
and size of the stomach should be assessed. Its distensibility in response to air insufflation, and the
degree of deformation imparted to the lesion or the stomach should be noted. Gentle abdominal
palpation during gastroscopy, with or without transillumination, may also be rewarding (the endoscopic
equivalent of the "bimanual" examination). If a mass is palpable per abdomen, corresponding in
location to the gastric lesion being visualized, a significant extragastric component may be present.
Transillumination is occasionally helpful when used in conjunction with palpation; a palpable mass in
the region of the stomach that transilluminates suggests the presence of a cyst, most often a
pseudocyst of the pancreas.
Biopsy
It is usually possible to maneuver the endoscope in the stomach so that the trajectory of the biopsy
forceps is oriented perpendicular to the surface of a lesion. This vertical approach ensures that the
tissue will be sampled to the maximum depth possible with the forceps. The volume of tissue obtained
is dependent on the size of the forceps cups and, to some extent, the force applied.1(4251) Using a
conventional forceps of 2.4-mm diameter and 6-mm jaw span, tissue samples of 1-mm depth and 2- to
5-mm diameter (weighing 6 mg on average) can be obtained, provided the forceps are applied en face
to the tissue. An endoscopic forceps biopsy of this depth from the stomach may include a variable
amount of muscularis mucosa and sometimes a sliver of submucosa.
Histologic interpretation of the endoscopic biopsy is hampered not only by its small size but also by
artifacts due to crush and avulsion effects and to inclusion of superficial nonspecific, inflammatory
changes. This is well illustrated in the diagnosis of lymphoma where differentiation from lymphoid
hyperplasia depends on a conclusive demonstration of the absence or presence of germinal centers.
Large-particle biopsies can be obtained using a large ("jumbo") forceps. This device must be used with
an endoscope with a suitably large-diameter accessory channel. Biopsies obtained with this type of
forceps, having an average weight of 15 mg, are often of sufficient size and depth for histologic
diagnosis of many submucosal lesions. Larger specimens can be obtained by simple electrosurgical
snare excision, an approach that is suitable for most polypoid lesions. Polypectomy may be the most
expedient method of sampling such lesions, as illustrated by the differential diagnosis of polyps, for
which complete removal of the lesion by polypectomy is usually required (see later).
The lift-and-cut technique is sometimes useful for obtaining large specimens from lesions that do not
lend themselves to snaring.2(4252) A double-channel endoscope is required. A conventional biopsy
forceps, passed down one channel, is inserted through an open snare passed down the other channel.
After lifting the target tissue with the forceps, the snare is tightened around the tissue lifted, and the
tissue encircled by the snare is excised electrosurgically. The size, but more importantly the depth of
the biopsy, is controlled by the degree of lift and the diameter of the gastric tissue enclosed in the
snare loop. Experience and judgment are needed to avoid perforation, which may be due to looping too
much tissue and from excessive electrocautery. For abnormal tissue that may be friable, thickened,
and inelastic, a three-pronged grasping device may be used in place of the forceps to provide the
necessary lift.
For correct diagnosis, careful handling of specimens is at least as important as biopsy technique. It is
essential that the methods for handling and processing the specimens be discussed in detail with the
gastrointestinal pathologist who is responsible for preparation of the slides and histologic assessment
of the tissue. A variety of solutions are available for fixation of the tissue (e.g., formalin, Bouin's,
Hollande's), and generally the choice is left to the pathologist. Ideally, biopsy specimens should be
properly oriented before fixation. The specimen removed from the forceps will have curled like a ball
inside the cups, the mucosa facing outward. To obtain the correct orientation for perpendicular
suctioning, the tissue must be spread flat on a piece of filter paper with the mucosa facing upward,
before dropping it into the fixative solution. Correct technique and a degree of painstaking effort and
time as well as experience are required because haste and ineptitude damage specimens and thereby
reduce their value for diagnosis. Orientation before fixation is often not practicable in units where large
numbers of procedures are being performed by numerous individuals. If tissue is submitted for frozen
section, fixative is not used but the specimen must be kept moist with saline solution and processed
without delay.
The number of biopsies required to obtain a diagnostic sample depends on the nature of the disease
as well as the endoscopist's knowledge of gross pathology as the basis for selection of the biopsy sites
that are most likely to provide the desired information. Because invasion is a criterion for malignancy,
tissue should be taken at junctions between grossly diseased and normal tissue when this diagnosis is
suspected. In addition, tissue should be obtained from nonnecrotic, proliferative parts of the
lesion.3(4253) For example, in some types of early gastric cancer (EGC), the cancerous tissue is
confined to a rim at the edge of the lesion, and biopsies obtained from the base may overlook its
presence. In general, the margins of a malignant ulcerative lesion are most likely to be positive,
followed in order of frequency by specimens obtained from the ulcer base. Ideally, both regions of the
lesion should be sampled because this combination increases overall sensitivity of sampling.4,5(4254)
When selection of biopsy sites is optimum, multiple specimens are still required, even if the lesion is
overtly cancerous in appearance, because the conventional endoscopic biopsy forceps imposes a
considerable limitation on specimen size. In general, seven or more biopsies are needed if diagnostic
yield is to approach 100% (see also Chapter 46: Peptic Diseases of the Stomach and
Duodenum).6,7(4255) Large-particle biopsies are usually necessary for the diagnosis of lymphoma
and other predominantly submucosal lesions.
It is advisable to review the histopathology of endoscopic specimens with a pathologist, preferably one
with experience in the interpretation of endoscopic biopsies. This is mandatory whenever there is any
suspicion of malignancy based on endoscopic findings that is not confirmed by biopsies. In the study of
Jorde et al.,8(4256) for example, biopsies from 4 of 21 patients who had malignant gastric lesions (1
lymphoma, 3 carcinoma) were initially interpreted incorrectly as negative for malignancy. Occasionally,
a histologic diagnosis of malignancy may also come as a surprise to the endoscopist. Jorde et
al.8(4257) encountered 4 histologic diagnoses of cancer (0.4%) among 959 lesions that appeared
benign at endoscopy.
Brush Cytology
Conceptually, brush cytology collects a sample of cells from a wider area than that represented in
targeted biopsies. This simple device, an extendable brush within a protective sheath, is inserted
through the accessory channel of an endoscope to abrade the surface of a lesion. Obviously, necrotic
or bloody surfaces should be avoided because the diagnostic yield from such areas will be relatively
low. Once an adequate specimen has been collected, the brush is withdrawn into the sheath and the
device is removed from the endoscope. Immediate fixation of the specimen is necessary to avoid
artifacts due to dehydration. This can be accomplished in several ways according to the preference of
the cytopathologistextended again from its sheath, the brush can be cut from its wire and dropped
into a fixative solution, or the cytologic material can be brushed directly onto a glass slide that is then
placed in a fixative solution. A cytology brush should be used only once. In addition to the fact that this
type of device cannot be disinfected properly, there is also the possibility that a serious error in
diagnosis may occur should cellular material remain on the brush after collection of a specimen. Brush
cytology is not useful for submucosal lesions with overlying intact mucosa, but the yield may be good if
the mucosa has been breached as a result of ulceration or erosion or at sites where biopsies have
been obtained.
There are numerous studies comparing biopsy and cytology in the diagnosis of gastric malignancy.
The majority indicate that the combination of cytology and biopsy significantly improves the yield of
positive diagnoses compared with biopsy or cytology alone.915(4258) The increase in diagnostic yield
associated with the additional use of brush cytology ranges from about 5 to 10%. A few studies,
however, have found that collection of cytology specimens together with biopsies did not significantly
increase diagnostic yield.16,17(4259) Cook et al.16(4260) recommended that cytologic sampling be
reserved for situations in which it is not possible to obtain adequate biopsies. Nevertheless, there is a
predominance of data to indicate that the combined use of cytology and biopsy significantly improves
the yield of positive diagnoses compared with biopsy alone. The likelihood of this depends on technical
circumstances and the nature of the lesion.
It may be difficult to target biopsies where a tumor has markedly narrowed the gastric lumen, in which
case it is often easier to obtain a cytology specimen by inserting a brush into the remaining lumen.
Witzel et al.15(4261) found that brush cytology was more accurate than biopsy in the diagnosis of
carcinoma of the cardia. Although it is reasonable to conclude that brush cytology may be especially
useful for sampling of stenotic lesions, there is relatively little information concerning the yield for
brushing with respect to gross tumor morphology (see discussion of Borrmann types under Advanced
Gastric Cancer). One study did find that cytology was less sensitive in the diagnosis of infiltrative
gastric cancer compared with other types.9(4262) Also, a false-positive cytologic diagnosis of
malignancy has been more common than a false-positive biopsy result in most studies. The rate of
false-positive cytologic diagnoses in reported studies has ranged from 0 to 3.1%.10,11,13,16(4263)
Relatively few studies have assessed the order in which biopsies and cytology specimens are obtained
as a potential influence on the rate of positive diagnosis. Keighley et al.13(4264) found that the
cumulative result was significantly better when brushing was performed before obtaining biopsies, and
they recommended that brush cytology be routinely performed before biopsy.
In patients with lymphoma, the combined use of cytology and biopsy significantly improves the yield of
positive diagnoses compared with biopsy alone.18(4265) However, the yield for cytologic sampling is
substantially less in lymphoma than in gastric carcinoma.14,19(4266)
Fine-Needle Aspiration Cytology

Since its introduction in the 1930s,20(4267) the technique of fine-needle aspiration cytology has been
used successfully in the diagnosis of breast and thyroid lesions. The technique has been modified for
use at endoscopy and has been shown to be useful in the diagnosis of submucosal lesions of the
gastrointestinal tract. This technique may occasionally be useful in the diagnosis of gastric cancer,
especially necrotic tumors and those that lie deep to the normal mucosa.21,22(4268) A further
modification has been evaluated for the diagnosis of diffuse infiltrative carcinoma of the
stomach.23,24(4269)
In principle, the cytology sample is obtained by intralesional aspiration with a fine needle, allowing the
cells to stay within the needle (but not within the syringe). The cells are then quickly expelled onto the
glass slide and fixed immediately to avoid artifacts due to dehydration. Actual details of the technique
vary, but the author's preferred method is as follows: A disposable variceal injector with a 23-gauge,
5-mm-long retractable needle is used. A well-fitting 30-ml syringe is attached to the injector. The
extended needle is inserted into the lesion to be sampled under direct vision. Aspiration is performed
by creating a vacuum with the syringe while the needle is totally buried in the tissue. As long as no air
enters the system, the needle can be redirected many times while it is inside the lesion to obtain
multiple samples. Care must be taken to prevent air rushing into the injector and thereby loss of the
sample into the syringe. Before the needle is removed from the lesion, the syringe is detached. To
retrieve the cytologic material, a syringe filled with air is attached to the injector. The material inside the
needle is expelled onto two glass slides, which are then opposed to produce a thin smear. Immediate
processing is essential. Within seconds, the slide is either sprayed with a fixative or dropped into a
bottle containing the fixative solution.
Other Studies
A variety of other imaging methods may be helpful in elucidating the nature of a lesion and the extent
of involvement, including transmural involvement.
Endoscopic Ultrasonography
EUS is the logical extension of endoscopy for studying lesions deep to the mucosa and for determining
the precise depth of lesions that arise in the mucosa. Because gastric lesions are readily accessible to
the ultrasound endoscope, EUS findings are now routinely utilized in planning therapy. Before
undertaking endoscopic excision of a submucosal tumor, for example, it is extremely helpful to know
the exact location of the lesion within the wall and whether a plane exists between the lesion and the
normal structures of the gastric wall. For malignant tumors, EUS provides staging almost as accurate
as operative exploration. It is superior to computed tomography (CT) for tumor staging because it
furnishes detailed information on the depth of the lesion and involvement of the anatomic layers of the
stomach that is not attainable with CT. Furthermore, CT often understages as well as overstages the
true extent of the disease. EUS aids in the differentiation of gastric carcinoma from lymphoma and
provides data on the status of lymph nodes near the stomach.2528(4270) EUS is discussed in
Chapter 51: Endoscopic Ultrasonography: Stomach and Duodenum.
Double-Contrast Radiography
As a screening test, the barium meal still plays an important role in the diagnosis of gastric tumors.
Many patients are referred for endoscopy because of an abnormal study. With the conventional upper
gastrointestinal x-ray series technique, major deformities of the stomach (e.g., changes in outline,
space-occupying lesions) are recognized readily (Figure 481). But for detection of subtle changes in
mucosal detail, the use of a double-contrast technique, as pioneered in Japan,29(4271) is essential.
Most radiologists use a dense barium suspension of low viscosity together with an effervescent agent
that releases carbon dioxide on contact with gastric contents. Glucagon (0.1 to 0.5 mg) is sometimes
administered intravenously to reduce antral peristalsis. With the stomach distended by gas, different
x-ray projections are used to examine the different regions of the stomach, utilizing compression when
necessary. Not only will mucosal details be visualized but also extragastric lesions can sometimes be
appreciated. In experienced hands, a very high accuracy can be achieved for detection of mucosal
lesions including EGC (Figure 482).29,30(4272) As in endoscopy, however, distinguishing EGC from
benign gastric ulcer can be difficult. Endoscopic examination of patients with abnormal radiologic
findings is essential in most practice settings to obtain histopathologic confirmation.31(4273)
(4274)Figure 481. Conventional upper gastrointestinal series. A, Mass effect at lesser

curvature due to carcinoma of the stomach. B, Large ulcerated mass (arrows) in the fundus of
the stomach. (A and B, Courtesy of Dr. M. Weiner.)
(4275)Figure 482. A and B, Two examples of early gastric cancer (EGC) as demonstrated by
double-contrast barium studies illustrating the changes in the pattern of mucosal folds. (A and
B, From Marshak RH, Lindner AE, Marklansky D [eds]. Radiology of the Stomach.
Philadelphia: WB Saunders, 1983; 154, 158.)
The Dye-Scattering Technique
The different techniques of spraying dyes to enhance mucosal details are described in Chapter 12:
Chromoscopy. Dyes tend to collect in the grooves and irregular depressions and so enhance the
mucosal pattern, making subtle abnormalities more visually striking. Additionally, pH-induced color
changes tend to highlight an area with a different pH, a principle used to detect EGC. In some cases of
EGC, such as Type IIb and Type IIc+III, selective use of chromoscopy does seem to make a
difference. These techniques, however, are not widely employed outside of Japan, and experience in
interpretation is hard to come by in the West. When chromoscopy is resorted to only as a secondary
technique to help resolve questionable findings, it does not prevent missing small early cancers. The
interpretation of endoscopic appearance is always subjective and an art in itself, and no technical
embellishment can be a substitute for experience in recognition of signs of EGC (see Early Gastric
Cancer).
Computed Tomography Scan and Laparoscopy
Extension of a malignant lesion beyond the immediate environs of the stomach, such as metastasis of
gastric carcinoma, is conveniently investigated by conventional CT. Despite newer techniques,
however, the accuracy of CT is inferior to that of laparoscopy.3234(4276)
Laparoscopic staging for gastric carcinoma is useful for preoperative assessment because the patient
may be spared a major operation if the findings indicate widespread metastatic disease, at least those
patients without symptoms that would require surgical palliation. Laparoscopic staging of abdominal
neoplasms is discussed in Chapter 94: Laparoscopy of Abdominal Tumors.
Laparoscopic examination of the stomach is facilitated by the reverse Trendelenburg position,
retraction of the left lobe of the liver by means of a retractor or probe, and lifting of parts of the stomach
with grasping forceps inserted separately through accessory ports. From the gastroesophageal
junction to the duodenum and between the two curvatures, the anterior gastric wall can be inspected in
its entirety. However, the posterior wall is not directly visible. If the lesser sac is free of tumor
adhesions, the stomach may be moved with laparoscopic forceps and mobility evaluated thereby. For
anteriorly located lesions, the presence of a serosal breach, direct invasion of adjacent tissue, or
peritoneal or omental seeding can be determined with a high degree of accuracy. Subpyloric lymph
nodes, gastroepiploic nodes, and nodes along the left gastric artery (accessed via a window made in
the lesser omentum) can be sampled by direct biopsy and frozen section examination. With the patient
in the Trendelenburg position, the omentum is lifted and placed over the stomach, exposing the
infracolic compartment of the abdominal cavity. Similarly, the pelvic cavity can be explored for ovarian
and other pelvic "drop" metastasis, thereby completing an evaluation that is equivalent to that provided
by CT.
Benign Tumors of the Stomach

Most benign tumors do not cause symptoms, but larger ones may be associated with nonspecific
symptoms of epigastric pain, bloating, and early satiety. If a large tumor is strategically located in the
gastric outlet, the patient may experience symptoms due to intermittent obstruction. Bleeding and
infarction may also occur with large lesions.
Endoscopically, all benign tumors present as protrusions into the gastric lumen. Table 481 lists the
relative frequencies of all polypoid lesions as encountered at autopsy. Because most of these, unless
large, do not give rise to specific symptoms, the clinical frequency can be expected to be similar.
Benign Polypoid Tumors of the

Stomach (Autopsy Series)
TABLE 481
TYPE
Epithelial polyp
Leiomyoma
Inflammatory polyp
Heterotopic tissue
Lipoma
Neurogenic tumor
Vascular tumor
Eosinophilic granuloma
Fibroma
Miscellaneous lesions
Total
NUMBER
252
230
29
25
21
19
13
12
9
6
616
40.9
37.3
4.7
4.1
3.4
3.1
2.1
1.9
1.4
1.0
100.0
Benign Polypoid Tumors of the

Stomach (Autopsy Series)
TABLE 481
TYPE
NUMBER
From Ming SC. Epithelial polyps of the stomach. In Ming SC,

Goldman H (eds). Pathology of the Gastrointestinal Tract. 2nd
ed. Baltimore: Williams & Wilkins, 1998; modified from Ming SC.
Atlas of Tumor Pathology. 2nd series. Fascicle 7. Washington,
DC: Armed Forces Institute of Pathology, 1973; 101.
Epithelial Polyps
Gastric polyps are uncommon. The incidence as estimated by radiologic survey is about
0.4%.35(4277) Gastric epithelial polyps are classified as true adenomas, hyperplastic, hamartomatous,
and inflammatory, which is a heterogeneous group (Table 482). The clinical significance of polyps
resides in their putative relationship to gastric cancer. Hyperplastic polyps, the most commonly
encountered type, are not precancerous, but true adenomas have a malignant potential.35,36(4278)
Because true adenomas cannot be differentiated from other types of polyps by endoscopic appearance
alone, histologic studies are necessary for proper management. When histopathologic interpretation of
forceps biopsies and polypectomy specimens were compared, over 50% of the former diagnoses were
not in agreement with the diagnosis based on study of the entire specimen.37(4279) Thus, total
excision is necessary for accurate diagnosis. The inherent inaccuracy of diagnoses based on forceps
biopsies alone is understandable because hyperplastic, juvenile, and inflammatory polyps all have
varying degrees of elongated and dilated foveolae and inflammatory infiltrates, so that a clear
distinction based on small fragments of tissue can be difficult. In the case of large adenomatous
polyps, malignant degeneration may be present only in small portions of the polyp and its presence
may therefore escape detection when samples are obtained with the biopsy forceps. When multiple
polyps are present in the stomach, all should be removed and their locations mapped and coded
accurately on a diagram of the stomach.
TABLE 482
Classification of Epithelial Gastric
Polyps
NEOPLASTIC
Benign
Adenomas
Tublar (flat)
Papillary (villous)
Malignant
Carcinoma
Carcinoid
Metastatic tumors
NONNEOPLASTIC
Hyperplastic
Hamartomatous
Peutz-Jeghers
Juvenile
Fundic gland
Foveolar
TABLE 482
Classification of Epithelial Gastric
Polyps
Inflammatory
Inflammatory retention
Polyp in Cronkhite-Canada syndrome
Heterotopic
Ectopic pancreatic tissue
Ectopic Brunner's gland hyperplasia
Adenomyoma
Adenoma
Adenomatous polyps may not cause any symptoms, and they are diagnosed, most incidentally, in the
elderly, usually about the seventh decade of life. However, vague symptoms of epigastric fullness,
hunger pain or postprandial pain, or heartburn are often present. Whether these complaints are due to
associated conditions or the polyp itself is not clear in most cases.
Adenomatous polyps are usually found in the antrum (Figures 483 and 484). There are no
surrounding converging folds, a point of some importance in differential diagnosis from EGC.
Converging folds are usually seen in EGC (see later). Two histopathologic subtypes of gastric
adenomas are recognized: tubular and villous. These have different endoscopic features. Tubular
adenomas are flat lesions, appearing broad and raised, and are usually solitary but sometimes occur
as a cluster of two or three. They are small, about 1 cm in diameter; 80% have a diameter less than 2
cm. When the surface of an adenoma is irregular and nodular, the distinction from EGC may be
difficult. In fact, terms relating to the classification of EGC, such as IIa-like and I-like, are used by some
endoscopists for descriptive purposes, even after biopsies have shown the lesion to be an adenoma.
(4280)Figure 483. Adenomatous antral polyp. Solitary lesion without converging mucosal
folds. Hyperplastic polyps have a similar appearance.
(4281)Figure 484. Multiple adenomatous antral polyps.

The flat adenoma is a slow-growing subtype of adenoma that may exhibit little change in gross
appearance over many years of follow-up observations.38(4282) Nevertheless, there is strong
evidence that they are premalignant, and it has been documented that 10% progress through stages of
dysplasia and atypism to malignancy.36,38(4283) Histologically, the tubular adenomatous tissue
usually occupies the upper third of the lesion, with dilation and cystic changes in the deeper, normal
glandular layer. Pleomorphism and mitosis are typically not evident, this being consistent with the slow
growth of the lesion.
A second subtype of adenoma, the papillary adenoma, is morphologically more polypoid and can be
sessile or broad based. Typically, they are large antral lesions with an average size of 3 to 4 cm when
discovered and tend to be solitary (60%). The surface appears more deeply colored than the
surrounding mucosa, which may be atrophic. Careful inspection of the mucosal surface may show
papillary features or fissures with minor irregularities. Lobulations are common. Erosions and
sometimes small ulcerations are also common features and can be a source of significant occult blood
loss. The soap-bubble sign, described by double-contrast radiography, owes its appearance to contrast
trapped in the irregular surface. The lesion can be moved freely with the endoscope or forceps, unless
malignant change has occurred. Larger adenomatous polyps (over 4 cm diameter) may harbor
malignancy. This should be suspected if the lesion is markedly asymmetrical, if a part of it has more
irregularities than the rest, or if there is evidence that the polyp is tethered and unyielding.
Histologically, the cells are intestinal in type, often complete with striated borders; typically, they are
dysplastic. In contradistinction to the flat adenoma subtype, pleomorphism and mitoses are regular
features of the polypoid adenoma.
Gastric adenomas, like colonic adenomas, undergo malignant transformation.36,39,40(4284) The
slow-growing tubular type has a lower incidence of malignant change, but this increases with the grade
of dysplasia. For the papillary type, the incidence of malignancy increases with increasing
size.39(4285) Based on a review of the literature before 1965, Ming and Goldman36(4286) concluded
that 30 to 40% of adenomas, mostly those of the papillary type, harbor malignant change. This has
been confirmed in subsequent reviews.39,40(4287)
Apart from the potential for an adenoma to contain in situ malignancy, there is a 30% incidence of
associated carcinoma elsewhere in the stomach.39,40(4288) This risk increases with increasing age of
the patient.41(4289) Therefore, the presence of a solitary large polyp should alert the endoscopist to
search carefully for evidence of gastric cancer, especially EGC. Intestinal metaplasia, a premalignant
condition, is often found elsewhere in the adenoma-bearing stomach.42(4290)
As a rule, all gastric polyps should be removed by endoscopic polypectomy, except for large lesions
with a very broad base because the risk of a complication associated with removal of such a lesion is
high (see later). If histopathologic examination of the resected specimen discloses an adenoma without
malignant change, the patient should undergo endoscopy at regular intervals (e.g., biannually) in a
structured surveillance program.41(4291) Invasive carcinoma in an adenoma is by definition an EGC
(Type I) and should be treated as such. For the large broad-based adenoma, surgical removal is
indicated, and if malignancy is found on examination of frozen sections, the appropriate surgical
operation can be performed at the same setting.
The Hyperplastic Polyp
By far, the hyperplastic polyp is the most common type found in the stomach, being seven to eight
times more common than the true adenoma.36,37,40(4292) Although a hyperplastic polyp cannot
always be differentiated from an adenomatous polyp by endoscopic examination, certain features,
when present, are highly suggestive. Hyperplastic polyps are often small and multiple and do not have
a particular predilection for the antrum. They are especially common near a gastroenterostomy, and in
some series as much as 20% of gastric remnants 20 years after gastrectomy may bear hyperplastic
polyps.43(4293) Small hyperplastic polyps tend to be sessile. Larger lesions may be pedunculated. A
well-developed polyp is smooth and olive-shaped, but larger ones may have surface erosions. The
surface is characteristically shiny and sometimes covered with mucous streaks owing to the presence
of abundant mucous-secreting cells, a histologic feature lacking in adenomas. As in adenomas, gross
lobulation may be seen occasionally.
As much as the gross appearance is similar for hyperplastic and adenomatous polyps, the histology is
distinctly different. Evidence of secretory activity will be conspicuous in the hyperplastic polyp, which
has glands made up of surface epithelial cells. Marked branching and cystic dilation of the glandular
lumen is typical. These glands are pyloric in type even when the lesion has arisen in the fundus. The
columnar cells may be distended with mucus, having the appearance of globoid signet-ring-like cells.
However, their nuclei are uniform, small, and compressed against the base. Mitoses are only
occasionally seen. There is no evidence to suggest a malignant potential for these polyps.
The hyperplastic polyp is not considered a precursor of gastric cancer.35,36(4294) However, it has
been noted for several decades that gastric carcinoma may coexist or develop in stomachs that harbor
hyperplastic polyps. Earlier reports indicated that as high as 22% of cases are associated with the
development of gastric carcinoma elsewhere in the stomach, although the incidence appears to be
less (about 7%) in more recent reports.44(4295) If hyperplastic polyps arise as a reaction of the gastric
mucosa to chronic injury, such as chronic atrophic gastritis, it is reasonable that a field change may be
seen in such mucosa, with malignancy developing in areas of intestinal metaplasia/dysplasia.
Therefore, the clinical significance of the hyperplastic polyp is that it may draw attention to an
underlying change in the gastric mucosa. For the most part, the polyps themselves are of no clinical
significance. However, there are a few case reports in which the occurrence of dysplasia or frank
cancer within a hyperplastic polyp has been well documented.4547(4296) In some cases, malignant
transformation was discovered at endoscopic follow-up observations.4749(4297) In the majority of
cases, hyperplastic polyps harboring malignancy have been 2 cm or more in diameter.47,48(4298) It is
possible that the development of malignancy within a hyperplastic polyp may be an incidental
occurrence, and this circumstance does not establish a malignant potential for this type of polyp.
The Inflammatory Polyp
Many inflammatory entities may present as mucosal nodules that have a histology dominated by
inflammatory changes. These include granulation tissue without glandular elements (properly termed
an inflammatory pseudopolyp) or fibrous tissue with a heavy inflammatory infiltrate (inflammatory
fibroid polyp). The polypoid hyperplasia associated with many chronic inflammatory conditions (e.g.,
Crohn's disease, tuberculosis) is distinguishable from isolated polyps by virtue of the regionalized
nature of the pathologic changes. The polyp seen in the Cronkhite-Canada syndrome is one type of
inflammatory polyp with retention cysts and marked inflammation in the stroma.
Polyposis Syndromes
Familial Adenomatous Polyposis
In familial adenomatous polyposis (FAP) including the Gardner variant, polyposis may develop in the
stomach or duodenum as well as the colon (see also Chapter 49: Diseases of the
Duodenum).50(4299) In the series of 102 FAP patients of Domizio et al.,51(4300) 55% had gastric
polyps. Two patterns are recognized: fundic and antral. In most cases, involvement is of the fundic
pattern; only 10% have antral polyps, with or without fundic polyps.
In the fundic pattern, the polyps are typically small, round and sessile, 2 to 8 mm in diameter on
average, and often numerous (Figure 485). The consensus of experts holds that these are
hamartomatous lesions, best termed fundic gland polyps, consisting of an excess of epithelial
elements, and that they are not precancerous.51(4301) Histologically, the predominant finding is cystic
dilation of the superficial portion of the gastric glands, with the lining composed of a mixture of
normal-appearing chief cells, parietal cells, and mucous-secreting cells in increased numbers.
Foveolar hyperplasia without glandular dilation may be present in a small proportion of polyps, but
there is never evidence of adenoma formation. Studies of the natural history of these polyps have
documented that they may diminish in size and number or even disappear, especially after
colectomy.52,53(4302) The incidence of true adenomas in the body and fundus of the stomach in FAP
is very low, less than 1%. The incidence of gastric carcinoma in a review of over 1200 cases was also
less than 1%.54(4303)
(4304)Figure 485. A field of fundic gland polyps in familial adenomatous polyposis (FAP).
In the antral pattern, the polyps are as likely to be adenomas as hyperplastic.51,52(4305) Gastric
adenomas are relatively uncommon in patients with FAP in comparison to duodenal adenomas, which
are frequent (see also Chapter 49: Diseases of the Duodenum).55,56(4306) If adenomatous polyps
are discovered, surveillance for the onset of dysplasia is indicated. In Japan, the incidence of gastric
carcinoma in FAP is much higher than in the West. Adenomatous polyps are also found in a much
larger proportion of patients with FAP (up to 60%).57(4307)
Fundic Gland Polyposis
Multiple small, sessile polyps may arise in the gastric fundus in patients who do not have FAP. Usually,
the polyps are less numerous than with FAP, but the endoscopic appearance and histologic findings
are similar if not identical to those of FAP.58,59(4308) The great majority of patients are women of
middle age.60(4309) The cause is unknown, but this entity, usually referred to as fundic gland
polyposis, has no known relationship to FAP or its variants. The prevalence of this condition is largely
unknown, but one study from Japan identified fundic gland polyps in 26 (1.9%) of 1388 patients
examined.61(4310) Another prospective study from Puerto Rico found fundic gland polyps in 44 of
5554 (0.8%) consecutive patients undergoing endoscopy.62(4311) Several long-term follow-up studies
have demonstrated that the number of polyps typically fluctuates over the course of time. It may
increase or decrease or the polyps may completely disappear only to reappear at a later
time.58,60,63(4312) As with FAP, there is no evidence that the fundic gland polyps in this condition
have any malignant potential and longitudinal studies have not disclosed any increased incidence of
carcinoma, although the period of follow-up is relatively short.
Peutz-Jeghers Syndrome
Hamartomatous polyps can be found in the stomach in patients with the Peutz-Jeghers syndrome (25
to 50% of patients).64,65(4313) The histology of the Peutz-Jeghers polyp is similar to that of the small
intestinal hamartoma, with whirls of smooth muscle often extending to the epithelial surface. The
malignant potential of these polyps is not known, but an association with carcinoma is well
documented, albeit of low incidence.65(4314) Carcinoma tends to develop at an early age.
Juvenile Polyposis
In juvenile polyposis, gastric involvement occurs in the subtype of the syndrome in which the entire
gastrointestinal tract is affected. The patient is usually a child, although the disorder may infrequently
present during adult life. Polyps may have already been discovered elsewhere in the gastrointestinal
tract, for example, the rectum. The pathogenesis is uncertain: An inflammatory cause has been
suggested, but Morson66(4315) favors a hamartomatous concept, pointing out the similarity of the
stromal tissue and the lamina propria of the colon. In the stomach, the polyp consists of hyperplastic
foveolae and edematous stroma with a heavy infiltrate of inflammatory cells.
Cronkhite-Canada Syndrome
Gastric polyposis may be part of the generalized gastrointestinal polyposis in the Cronkhite-Canada
syndrome (Figure 486), a noninherited, noncongenital disorder that occurs primarily in the sixth to
seventh decades of life, although patient age ranges from 31 to 86 years in reported
cases.6769(4316) There is no definite sex predilection. Spontaneous remission has been
reported,67,70(4317) but the prognosis is generally poor with a mortality of 60%, most deaths
occurring within 6 to 18 months.71,72(4318) Numerous gastric polyps are usually present in the
Cronkhite-Canada syndrome. They are neither adenomas nor neoplastic and are not considered to be
premalignant. The extragastric manifestations of the Cronkhite-Canada syndrome include generalized
polyposis of the gastrointestinal tract, particularly the colon and small intestine, cutaneous
hyperpigmentation, alopecia, and onychodystrophy.69,73(4319) Protein-losing enteropathy may
occur.68(4320) Gastrointestinal bleeding has been reported,74(4321) but this appears to be unusual.
There is no concrete evidence linking the development of carcinoma to this syndrome.
(4322)Figure 486. Polyps in Cronkhite-Canada syndrome. A, Large antral polyp with

mottled appearance with smaller polyps nearby. B, Elsewhere in the stomach there are
numerous small polyps.

On gross or endoscopic examination of either the stomach or the colon, multiple sessile polyps on a
thickened mucosa are seen. The concentration of polyps within the stomach is usually heaviest in the
antrum. They range from 0.5 to 1.5 cm in diameter and may have a mottled or reticulated surface
pattern. Microscopically, there is intact surface epithelium, tortuous glands that are cystically dilated
and filled with proteinaceous fluid, and edematous chronically inflamed lamina propria. Secondary
acute inflammation may be superimposed.68(4323)
Cowden's Disease
Cowden's disease (multiple hamartoma syndrome) is an extremely rare inherited disorder that is
characterized by multiple hamartomatous polyps of the skin and mucous membranes.75,76(4324) The
disease usually manifests as multiple facial keratotic papules. Breast and thyroid lesions may be
present as well as developmental abnormalities. The pattern of inheritance is autosomal dominant, and
there is a high rate of spontaneous mutation. Approximately one third of patients have gastrointestinal
polyps. However, symptoms referable to the digestive tract are unusual, so that gastrointestinal
imaging studies have been performed in relatively few patients.75,77(4325) If present, polyps are
usually numerous, of various histologic types (lipoma, juvenile, inflammatory), and involve all levels of
the gastrointestinal tract. There is no evidence that the risk of gastrointestinal malignancy is increased
in Cowden's disease.
Mesenchymal Lesions
Leiomyoma
The leiomyoma, the most common of the benign tumors of the stomach, arises in the submucosa or
muscularis propria. Symptoms of bleeding or pain may arise when the tumor is large, but patients are
almost asymptomatic if it is small. In common with all submucosal tumors, the surface is covered by
normal gastric mucosa unless ulceration has occurred. A bridging fold, a normal-appearing gastric
mucosal fold that runs from the surrounding mucosa to the top of the lesion (Figure 487), often
graces the surface of a leiomyoma. This finding is not specific for leiomyoma but rather a sign that
indicates the submucosal location of an underlying lesion. Larger leiomyomas protrude abruptly into
the lumen, and if in the antrum, a pseudopedicle may be recognized. Intermittent prolapse into the
duodenum may occur, causing episodic abdominal cramps and vomiting. Central ulceration, due to
necrosis of the tumor, occurs in larger lesions (Figure 488) and is responsible for the clinical
presentation of bleeding. Leiomyoblastoma and leiomyosarcoma cannot be reliably distinguished from
leiomyoma on the basis of endoscopic appearance, but malignancy should be suspected when the
lesion is large (i.e., over 8 to 10 cm in diameter). Biopsies from the mucosal surface predictably yield
normal mucosa, but if adenomatous tissue is found histologically, the endoscopic diagnosis of
submucosal tumor should be reviewed and perhaps revised. Large particle biopsies or "excavating"
biopsies (acquisition of four to five successive specimens from the same area, each deeper than the
previous tissue sample) may provide information on the histopathologic nature of a submucosal lesion.
As for all submucosal lesions, EUS also provides important information that is frequently sufficient for a
tentative diagnosis. Because the true nature of a submucosal lesion cannot be determined by
endoscopic observation, it is advisable to visualize such lesions by EUS before undertaking aggressive
endoscopic biopsy or excisional procedures (see Chapter 51: Endoscopic Ultrasonography: Stomach
and Duodenum).
(4326)Figure 487. Gastric submucosal tumor showing a typical rounded lesion with the
smooth surface of normal mucosa and a "bridging fold," an endoscopic sign indicative of the
submucosal nature of the lesion. (From the collection of Dr. M. V. Sivak, Jr.)
(4327)Figure 488. A, Leiomyoma on the lesser curvature of the gastric antrum. Except for
central ulceration, the overlying mucosa is normal. Pylorus is seen distal to the lesion. B,
Leiomyoma on the greater curvature of the gastric body. Note the bridging fold. (A, Courtesy
of Dr. W. Weinstein; B, From the collection of Dr. M. V. Sivak, Jr.)
Endoscopic excision can be accomplished for smaller, predominantly luminal tumors, especially if the
tumor is pedunculated.78,79(4328) Many endoscopic techniques for resection of larger submucosal
lesions have been described (see later), but these should be utilized only when there is a reasonable
plane of cleavage between the lesion and the neighboring layers of the gastric wall as demonstrated by
EUS. For large lesions and those in which EUS suggests evidence of malignancy, surgical resection
(wedged resection or partial gastrectomy) is indicated because myosarcoma cannot be excluded in
such cases. Surgical removal is generally indicated for gastric leiomyomas, unless the patient is a poor
operative risk or the lesion is very small. Despite continuing reports of successful cases of endoscopic
resection of gastrointestinal leiomyomas, data on the efficacy and safety of these techniques are
limited and surgery remains the mainstay of treatment.
Schwannoma
Nerve sheath tumors are known under a variety of different names, such as neurofibroma,
neurilemmoma, schwannoma, or neuroma, with malignant counterparts. Histologically, they may be
indistinguishable from smooth muscle tumors. There are no endoscopic characteristics that distinguish
them from lesions that arise in smooth muscle.
Lipoma
The lipoma is typically found in the gastric body and antrum and has a characteristic yellowish hue.
Larger ones may have septations and lobulations when viewed through the translucent mucosa (Figure
489). Rarely, multiple gastric lipomas may be present.80(4329) If the lesion is probed, it will exhibit
the cushion or pillow sign (a soft, lingering indentation), indicative of its nearly fluid nature. When
covered with intact mucosa, no symptoms can be attributed to these lesions, which may also be found
elsewhere in the digestive tract of the same patient. Depending on size and location, a lipoma can be
sessile (small, located in the body) or pedunculated (large, located in the antrum). Gastrointestinal
bleeding can arise from larger lipomas, which may ulcerate.80(4330) Especially large lipomas can
cause intermittent pyloric obstruction. Small tumors are often left alone, but excision can be
accomplished. Initially, a mucosal incision is made and then the lesion can be extruded by gradually
closing a polypectomy snare around the lesion to produce a slow, squeezing action (see later).
(4331)Figure 489. Lipoma of the stomach. The yellow color is characteristic.
Angioma
Angiomas are uncommon tumors that have a characteristic gross appearance that distinguishes them
from other submucosal lesions. Although mainly submucosal, they have a mucosal component that
gives rise to the endoscopic appearance of a fiery-red, velvety, raised (1- to 2-mm) sessile lesion that
bleeds easily (Figure 4810). Often they resemble or are disguised by adherent clots. The differential
diagnosis includes a small Kaposi sarcoma, but these lesions arise in different clinical circumstances.
Successful destruction with endoscopic laser photocoagulation and electrocoagulation has been
reported.81(4332)
(4333)Figure 4810. Angioma of stomach.
Brunner's Gland Hyperplasia

Rarely, a nodule in the pyloric region may be due to hyperplasia of Brunner's glands, a diagnosis made
only on histologic examination. Brunner's gland hyperplasia occurs much more commonly in the
duodenum.
Heterotopic Tissues
Heterotopic pancreas is an uncommon incidental finding. The lesion is a submucosal nodule found
usually in the antrum (see Chapter 72: Congenital Anomalies of the Pancreas). When typical, it has a
minute central dimple that bears the vestigial duct. Sometimes an ulcer may arise in this area, in which
case the lesion may resemble a leiomyoma. EUS findings (mural, nonvascular lesion) may suggest the
diagnosis. The histology consists mainly of exocrine pancreatic tissue, although significant amounts of
islet tissue may be seen in large lesions.
Heterotopic spleen is rarer still. It tends to present as a nodule in the wall of the gastric fundus, covered
by normal mucosa (Figure 4811). EUS will demonstrate a vascular mural lesion and thereby readily
distinguishes the lesion from a leiomyoma (Figure 4812).
(4334)Figure 4811. Endoscopic view of a protruding lesion in the fundus covered with
normal mucosa that proved at surgery to be a heterotopic spleen.
(4335)Figure 4812. Endoscopic ultrasonography of the lesion in Figure 4811 shows

vascular structures in the wall of the stomach that proved at surgery to be a heterotopic spleen.
Teratoma
Teratomas of the stomach are uncommon but have dramatic clinical features because of their size.
Teratoma is a benign tumor seen mostly in infants. The large tumor presents as an abdominal mass. It
may project into the stomach or the peritoneal cavity. Associated symptoms are bleeding and
obstruction. Teeth may be demonstrated radiographically. Histologically, the presence of all three germ
layers is diagnostic.
Pseudolymphoma (Lymphoid Hyperplasia)

Pseudolymphoma is a condition that is difficult to differentiate from lymphoma (see also Chapter 47:
Gastritis and Gastropathy). Pseudolymphoma is a benign condition with a pathogenesis linked to
chronic prepyloric or gastric ulcer that has resulted in extreme reactive edema, scarring, and lymphoid
hyperplasia. Because of improvements in the treatment of peptic disease during the last several
decades, chronic gastric ulcer is now much less common in many parts of the world. Furthermore, it is
now recognized that the endoscopic and histologic appearances of pseudolymphoma are very similar
to those of the B-cell gastric lymphoma that is also known as mucosa-associated lymphoid tissue
lymphoma (MALToma). In fact, many cases previously diagnosed as benign pseudolymphoma would
be classified now as examples of gastric MALTomas (Figure 4813A-C). The relationship between
Helicobacter pylori infection and MALToma is discussed in Chapter 47.
(4336)Figure 4813. MALToma. A and B, Extensive infiltration of the gastric folds in the
body of the stomach and several well-demarcated ulcers. C, Involvement of the antrum.
Extensive involvement of the duodenum was also present in this man who had recurrent upper
gastrointestinal bleeding for more than 10 years. Helicobacter pylori was not present in biopsy
specimens, and the tumor did not respond to antimicrobial treatment for the organism. The
nature of the lesion was confirmed by gene rearrangement studies on Southern blot analysis.
D, Pseudolymphoma. Chronic gastric ulcer on a large, polypoid, and nodular fold. Biopsies
showed marked infiltration of the mucosa and submucosa by mature lymphocytes. Histologic
evaluation of the resection specimen confirmed the diagnosis of pseudolymphoma. (A-C,
Courtesy of Dr. E. Variyam; D, From the collection of Dr. M. V. Sivak, Jr.)
The ulcer in patients with pseudolymphoma may have raised edges, and the rugal folds in the region
may become large and thick, sometimes with the appearance of cobblestoning. The resulting polypoid
mass and giant folds mimic lymphomatous infiltration (see Figure 4813D). Indeed, the clinical
symptoms are those of chronic gastric ulcer. The endoscopic appearance resembles that of a benign
ulcer but of great chronicity, without healing despite long periods of pharmacologic therapy. Because of
the difficulty in differentiating benign pseudolymphoma from lymphoma, it is advisable to obtain
large-particle biopsies (jumbo forceps) at endoscopy. When biopsies of the ulcer and polypoid
component of the lesion are negative for malignancy, showing only edema, fibrosis, and excessive
inflammatory response with lymphocytic infiltration, the diagnosis of pseudolymphoma can be
entertained. With large-particle biopsy, lymphoid follicles with germinal centers are seen, findings that
clearly distinguish the lesion from lymphoma. Other histologic features that support the diagnosis of a
benign pseudolymphoma include small mature lymphocytes with a mixed infiltrate of eosinophils,
plasma cells, and neutrophils. In lymphoma, by contrast, the cells are uniform and in sheets and
masses without germinal centers. Regional nodes are not usually involved in pseudolymphoma.
Xanthoma
The gastric xanthoma (xanthelasma) is a nonneoplastic lesion. Although easily recognized at
endoscopy because of its distinctive appearance, undue significance is sometimes attached to the
lesion or it may be mistaken for a true neoplasm (see also Chapter 47: Gastritis and Gastropathy). It
appears as a yellowish, slightly raised, round lesion with an irregular, serrated margin.82,83(4337) The
lesions are seldom more than 5 mm in diameter, although xanthomas up to 10 mm in diameter have
been described.8486(4338) The presence of multiple xanthomas is relatively common.87,88(4339) In
a retrospective review of 109 cases, Moreto et al.88(4340) found a relatively even distribution
throughout the stomach. An accumulation of lipid-laden histiocytes in the lamina propria is the
characteristic histologic feature in biopsy specimens. However, the endoscopic appearance of the
lesion is so characteristic that it is seldom necessary to obtain biopsy specimens. In cases where
successive endoscopic examinations have been performed, xanthomas have been noted to decrease
in size or even disappear.87(4341) Although gastric xanthomas are uncommon lesions, they are found
more frequently in autopsy as opposed to endoscopic case series. In the latter, the incidence ranges
from 0.38 to 0.8% of cases.85,86,89(4342) No relation has been found between the presence of
gastric or upper gastrointestinal xanthomas and serum lipid levels.87(4343) Although gastric
xanthomas have no inherent clinical significance, associated mucosal abnormalities are relatively
common. For example, Moreto et al.88(4344) found evidence of chronic gastritis and intestinal
metaplasia in nearly half of the 109 cases reviewed. Gastric xanthomas are encountered more
frequently in patients who have undergone gastric resections and the incidence increases with the
number of years elapsed since operation.90(4345) This observation has led to speculation that their
occurrence may be related to bile reflux and associated mucosal injury.
The Differential Diagnosis of Giant Gastric Folds

Endoscopic and occasionally surgical biopsy specimens are required for the diagnosis of conditions
that result in giant gastric folds. Surgical excision of a small area of the stomach for final diagnosis is
seldom required, but if needed, this is a simple operation accomplished with the surgical stapler.
Recently, the operation has been performed laparoscopically, aided by gastroscopic localization (see
later). With extreme hypertrophy of gastric glandular tissue, the entire specimen may consist of only
one region of a gland, large-particle biopsy notwithstanding. Any of the techniques for endoscopic
tissue sampling discussed in this chapter may be appropriate. Obtaining biopsies repeatedly at the
same location (excavating technique) may be required for diagnosis. EUS findings may suggest a
diagnosis (see Chapter 51: Endoscopic Ultrasonography: Stomach and Duodenum).91(4346)
For the purpose of endoscopic diagnosis, it is useful to broadly classify giant gastric folds as diffuse
and localized (Table 483). The differential diagnosis of giant folds is discussed in detail in Chapter 47:
Gastritis and Gastropathy.
TABLE 483
Differential Diagnosis of
Giant Folds
DIFFUSE
Zollinger-Ellison syndrome
Mntrier's disease
Inflammatory disorders
Tuberculosis, syphilis, sarcoidosis
Isolated granulomatous gastritis
Allergic gastroenteritis
Lymphocytic gastritis
Gastritis cystica polyposa
Neoplasms
Lymphoma
Carcinoma (linitis plastica)
LOCALIZED
Polyps
Inflammatory
Hamartomatous
Hyperplastic
Neoplastic
Focal inflammatory lesions
Heterotopic pancreatic or splenic tissues
Pseudolymphoma
Modified from Goldman H. Mucosal hypertrophy and
hyperplasia of the stomach. In Ming SC, Goldman H
(eds). Pathology of the Gastrointestinal Tract. 2nd ed.
Baltimore: Williams & Wilkins, 1998.
The endoscopic appearance with the diffuse variety of giant folds is usually striking. It is ill advised to
place too much emphasis on the ability to efface enlarged folds by air insufflation within the stomach.
Although infiltrated, rigid folds will not efface, the large folds of mucosal hypertrophy due to
Zollinger-Ellison syndrome and Mntrier's disease may efface provided high intragastric pressure and
volume are attained.

Diffuse hypertrophy, involving the entire stomach or a region, is typical of the Zollinger-Ellison
syndrome and Mntrier's disease. The hypersecretory state in patients with gastrinoma may be so
pronounced that clear secretion can be seen to accumulate during endoscopy. Although erosions and
ulcers are not found in the fundus or body, abnormalities (erythema, edema, congestion, ulceration),
when present, may be noted from the antrum and for a considerable distance into the duodenum
including the third portion. In Mntrier's disease, where marked hypertrophy of the foveolar mucous
cells is the central feature, the body of the stomach is maximally involved, signs of inflammation being
absent elsewhere. Often several folds stand out from the rest, appearing polypoid and cystic; the latter
features give rise to a "beaded" appearance (Figure 4814). Even large-particle biopsies may not show
the entire gastric gland. If the samples are almost entirely occupied by hypertrophic foveolar mucous
cells with a paucity of inflammatory cells, a diagnosis of Mntrier's disease is strongly suggested.
Cumulative evidence does not favor an association between Mntrier's disease and
carcinoma.92(4347) A number of inflammatory conditions may become manifest as thickened,
enlarged gastric folds, including tuberculosis, syphilis, eosinophilic gastroenteritis, lymphocytic gastritis,
Crohn's disease, and sarcoidosis (see Chapter 47: Gastritis and Gastropathy). With all these entities,
the clinical issue is one of differentiation from malignancy including diffuse lymphoma and linitis
plastica.
(4348)Figure 4814. Surgical specimen of Mntrier's disease. Cystic and polypoid

formations are evident on the enlarged folds.
Localized enlargement of one or two gastric folds sometimes takes on the appearance of a polypoid
lesion, and thereby, the problem becomes one of differentiation from other polypoid lesions. Large
folds confined to one region may be seen in malignant infiltration such as gastric carcinoma (linitis
plastica), lymphoma, and metastatic cancers from the breast and lung, as well as other tumors with
predilection for spread to the gut such as melanomas (see later). Inflammatory lesions and other
masses occasionally have the appearance of a large fold. It is unlikely that gastric varices may be
mistaken for giant folds if the condition is borne in mind, but the varicosity may not be immediately
obvious and the lesion does not usually appear blue.
Malignant Tumors of the Stomach

A knowledge of the gross pathology of gastric cancer is essential to the endoscopic recognition of this
malignancy. EGC, defined as malignancy limited to the mucosa and submucosa, is occasionally
detected by the vigilant endoscopist but is much less commonly diagnosed than advanced gastric
cancer. The latter constitutes the majority of endoscopic experience of gastric cancer in Western
countries.
Dysplasia
Several prospective endoscopic studies have demonstrated that the presence of gastric epithelial
dysplasia indicates a significant increased risk for the development of adenocarcinoma.9395(4349)
Progression from severe dysplasia to carcinoma was noted in 8 of 26 patients (30.7%) in the study of
Coma del Corral et al.,93(4350) 6 of 10 patients (60%) studied by Di Gregorio et al.,95(4351) and 8 of
15 patients (53%) in the study of Farinati et al.94(4352) Although these studies also confirm that
dysplasia may regress regardless of its severity grade on initial biopsies, the data clearly demonstrate
that even mild dysplasia may progress to carcinoma, although the risk appears to be less than that
associated with high-grade (severe) dysplasia. In the study of Coma del Corral et al.,93(4353) 2 of 41
patients with moderate dysplasia progressed to carcinoma. Di Gregorio et al.95(4354) found that mild
dysplasia progressed to carcinoma in 4 of 73 cases (5.5%); progression to carcinoma was also noted
in 13% of patients with moderate dysplasia.

Endoscopic confirmation and surveillance are therefore indicated for patients with low-grade (mild to
moderate) dysplasia. Because of the high risk for the development of gastric carcinoma as well as the
presence of a synchronous undiagnosed cancer, gastric resection should be considered for patients
with severe dysplasia.95(4355) Careful endoscopic evaluation including confirmation and gastric
mapping of the histologic findings should be performed before surgery. Endoscopic surveillance with
meticulous examination of the mucosal surface and extensive biopsies is indicated for patients with
mild to moderate dysplasia. The most efficacious interval between surveillance examinations has not
been established. An interval of 1 year has been recommended.95(4356) To a substantial degree,
however, the interval between procedures will be determined by histologic findings on successive
examinations. Closer surveillance is advisable if there is evidence of histologic progression.

Mass screening programs for the detection of EGC have had a significant impact on the treatment of
this disease in Japan. More than 40 to 50% of gastric carcinomas treated in Japan are early-stage
lesions, the outcome of which is much more favorable than for advanced gastric cancer. Before
institution of mass screening, only about 5% of the total number of gastric cancers diagnosed in Japan
were early-stage lesions.96,97(4357) EGC represents less than 10% of all gastric cancers diagnosed
in countries other than Japan, although it appears that EGC is now being recognized more frequently
outside of Japan, a change generally attributed to the widening use of endoscopy.98102(4358)
However, an increase in diagnoses of EGC has not been demonstrated in all studies.103(4359)
Whether the marked differences among different regions of the world with respect to the frequency
with which EGC is diagnosed are attributable simply to mass surveillance programs is not clear. In any
case, mass screening is practical and cost effective only in regions where the disease is endemic.
EGC as seen in the West is usually symptomatic, although the symptoms are
nonspecific.104110(4360) In Japan, because of mass radiologic and endoscopic screening, a
substantial proportion of EGCs detected are truly asymptomatic. The ratio of EGC to advanced gastric
cancer is much higher when the diagnostic yield for mass screening is compared with that for
ambulatory care clinics because the symptoms of early and advanced cancers are similar.96,97(4361)
The most frequent symptom is epigastric pain (30 to 80%), followed by hematemesis (40%) or
vomiting (29%) and weight loss (16 to 19%). Significantly more patients with EGC have or have had
peptic ulcers than patients with advanced gastric cancer.104,105 Therefore, in the absence of mass
screening programs, early endoscopic examination of any patient with dyspepsia, especially those who
are elderly, is the only practical approach to increasing the frequency of diagnosis of EGC.108 In the
study of Lawrence and Shiu107(4362) from the United States, the interval from presentation to
diagnosis of EGC was 14 months (range 2 months to 4 years).
Of 1900 cases of EGC in the registry of the National Cancer Center in Tokyo, 95% arose against a
background of atrophic gastritis.111(4363) Chronic atrophic gastritis accompanied by intestinal
metaplasia is recognized as a precursor of EGC, and the vigilance of the endoscopist should be
heightened when atrophic gastritis is found on routine examination. In the West, chronic gastritis with
intestinal metaplasia is also commonly associated with EGC and advanced gastric
cancer.105,107(4364)
The classification of the Japanese Gastroenterological Endoscopy Society is generally used to
describe the macroscopic features of EGC.112(4365) Such lesions are classified as protruding (Type
I), superficial or flat (Type II), or excavating (Type III) (Figure 4815). Type I cancers are polypoid
lesions that protrude into the lumen; the carcinomatous tissue extends no deeper than the submucosa
(Figure 4816). In Type II, different degrees of flatness are recognized: a minimally elevated lesion,
Type IIa, appears at endoscopy as a plaque; a truly flat lesion, Type IIb, is detected only as
discoloration of the mucosal surface (pale or irregularly erythematous) and a finely nodular, irregular
surface.113,114(4366) Type IIc is minimally depressed, like a shallow ulcer, except that the edge is
irregular or discolored (Figure 4817); it is the most common type in Japan, currently accounting for
75% of EGCs. Type III occurs only rarely in Japan (1%).111(4367) This is in contrast to Europe and
the United States where Type III is the most common (30 to 65%) and Type II the least common (8 to
20%).104,107,115(4368) The gross morphology of any EGC may be completely described by a
combination of types. For example, Type IIc+III indicates a cancer with a shallow depression and a
central excavating ulcer (Figures 4818 and 4819). Although the depth of invasion of an EGC is not
determined by its diameter, these tumors are usually relatively small (less than 2.5 cm).105,107(4369)
(4370)Figure 4815. Endoscopic morphologic classification of EGC (Japan

Gastroenterological Society). Type I is protruding in appearance, a polypoid lesion elevated
from the surface. Type II is superficial or flatIIa is slightly elevated, IIb is almost flat, and
IIc is minimally depressed. Type III is an excavating type of lesion.
(4371)Figure 4816. Endoscopic view (top) and resected specimen (bottom) of Type I EGC.
(Top and bottom, From Kasugai T [ed]. Endoscopic Diagnosis in Gastroenterology. Tokyo,
New York: Igaku-Shoin, 1982; 889.)
(4372)Figure 4817. Endoscopic view (top) and resected specimen (bottom) of Type IIc
EGC. (Top and bottom, From Kasugai T [ed]. Endoscopic Diagnosis in Gastroenterology.
Tokyo, New York: Igaku-Shoin, 1982; 967.)
(4373)Figure 4818. Endoscopic view (top) and resected specimen (bottom) of Type IIc+III
EGC. The lesion is not Type III because it is mainly a superficial depression, a portion of
which is excavating (hence the designation +III). (Top and bottom, From Kasugai T [ed].
Endoscopic Diagnosis in Gastroenterology. Tokyo, New York: Igaku-Shoin, 1982; 1001.)
(4374)Figure 4819. By combining types, a complete description of all shapes of EGC is

possible. Thus, Type IIa+IIc is a minimally raised lesion with a minimally depressed center;
IIc+IIa is a minimally depressed lesion with part of it protruding minimally; IIb+IIc designates
a flat but partially excavated lesion; IIc+III is minimally depressed at the margins with a
deeper excavation in the center.
In Japan, the most common locations for EGC are the lesser curvature of the antrum (15%) and the
angulus (incisura) (28%). In about 8% of cases, two or more cancers are found, usually in the same
region of the stomach.116(4375) In a retrospective review of 73 cases of synchronous EGCs, Brandt
et al.116(4376) found that barium studies and endoscopy demonstrated 116 of the 163 malignancies
(71%). EGC may occur in association with squamous cell carcinoma of the esophagus.117(4377) In
the countries other than Japan, the distribution of EGC appears to be similar, with 45% being located in
the body of the stomach.105,107,109,110(4378)
For the endoscopist with little experience in the diagnosis of EGC, recognition of the pattern of the
folds in the mucosa surrounding the lesion is of great value. A pattern of converging folds is often the
first finding that arouses suspicion for both endoscopist and radiologist. If these folds are followed
toward the point of convergence, they either taper or become bulbous (clubbing) and terminate. In
either case, they point to the central cancerous lesion itself (Figure 4820; see also Figures 4816,
4817 and 4818), which can be raised, flat, or excavated, or a combination thereof. Ulceration of
these lesions is common (60 to 80%) and is considered a secondary peptic event. The ulcer is often
superficial, but even if it extends beyond the submucosa, the deeper part may not be cancerous.
Carcinomatous tissue may be present only in the adjacent flat portion of the lesion, which appears
discolored, finely nodular, or irregular. Healing may take place; complete healing has been observed in
20% of cases,118(4379) but ulcer recurrence is the rule. The scarring induced by the ulceration
causes the mucosal folds to converge and can thus simulate the appearance of a benign gastric ulcer.
Even though the ulcer appears deep and penetrates the submucosa, the cancer may not penetrate so
deeply. Nevertheless, Type IIc+III and Type III lesions still cause the most difficulty with differential
diagnosis. Such lesions are sometimes classified by Japanese endoscopists as being of indeterminate
type.
(4380)Figure 4820. Artist's renditions of resected specimens of EGC to illustrate converging

folds. A, Type IIc. Folds terminate with clubbing, bulbous enlargement, tapering, or an abrupt
cutoff. B, Type IIc. Many tapering terminations of converging folds. (A and B, Drawn from
photographs in Hirota T, Ming SC. Early gastric carcinoma. In Ming SC, Goldman H [eds].
Pathology of the Gastrointestinal Tract. Philadelphia: WB Saunders, 1992; 576.)
Certain macroscopic features have been shown to correlate with the histologic depth of the malignancy
(see Figures 4815 and 4819).119(4381) If the surface of a Type I lesion is smooth and devoid of
ulceration, it is likely that the malignancy remains confined to the mucosa. A Type IIb or IIc lesion
without fold convergence or bank formation (nodular, elevated margin) is most often a mucosal cancer.
If there is fold convergence in association with a Type IIc lesion but only irregular tapering of the folds,
the cancer is likely to be mucosal. However, if these folds end abruptly at the lesion, exhibit clubbing,
or have nodular margins, it is likely that the cancer has invaded the submucosa. When the base of the
depression is firm and unyielding to the biopsy forceps, the depth of involvement is at least
submucosal. If an ulcerative lesion is surrounded by an elevated, nodular margin (bank), the lesion is
likely to be an advanced cancer. These empirically derived criteria are widely used by Japanese
endoscopists to estimate the depth of penetration of the cancer. In one study of over 200 cases, these
criteria predicted the histologic depth of involvement with an accuracy of almost 70%; the depth was
overestimated in 20% and underestimated in 10%.119(4382) The diameter of the lesion alone is not a
reliable criterion for depth of invasion, except when it is smaller than 1 cm, in which case the cancer is
almost always mucosal.
Special endoscopic methods such as chromoscopy (see Chapter 12) and magnification endoscopy
(see Chapter 13) can be used to enhance the endoscopic diagnosis and staging of EGC. For example,
Iishi et al.120(4383) found that the use of chromoscopic techniques improved the ability to recognize
the presence of synchronous early lesions.
The radiographic diagnosis of EGC also depends on recognition of fold patterns using double-contrast
technique with thin barium suspension and pharmacologically induced gastric hypotony. This has
played an important role in screening programs for EGC in Japan.121(4384) The patterns of mucosal
changes due to EGC (see Figure 482) were demonstrated with double-contrast technique using thin
barium suspension in the early 1970s. In centers with a special interest in these techniques, the
accuracy of radiologic and endoscopic diagnoses of EGC is about equal,122(4385) although
endoscopy with biopsy is still required for histologic confirmation.
Some Japanese investigators have classified EGC according to growth pattern as seen in resected
specimens. The pattern is termed superficial spreading when diameter far exceeds depth; this pattern
is seen in 45% of cases. The term small mucosal type, which accounts for 37% of EGCs, describes a
heaped-up lesion that is confined to the mucosa. A flat or excavating lesion that seems to grow toward
the muscularis propria is designated penetrating, a pattern that is found in 17% of EGCs.
In contrast to advanced gastric cancer, the growth rate of EGC appears to be slow. Fujita123(4386)
estimated the doubling time for an EGC to be 18 to 108 months or 2 to 3 years on average. The tumor
may remain an early-stage malignancy for 14 to 21 years. The doubling time for advanced gastric
cancer is estimated at 2 to 10 months.123(4387) The high rate of detection of EGC in Japan may be
due to its prevalence as well as to the slow rate of growth of the lesion in addition to the efficacy of
mass-screening diagnostic techniques.
EGC tends to be tubular adenocarcinoma, the most common histologic type (52%), or signet-ring cell
carcinoma (26%). The histologic type is poorly differentiated adenocarcinoma in 13% of cases. In a
study of 156 resected specimens of EGC, Kawaura et al.124(4388) found that invasion deep to the
mucosa was more likely to be present if the cancer was undifferentiated, regardless of the diameter of
the lesion. Lymph node metastasis occurs even when the cancer is confined to mucosa, but the
incidence is low (about 10%). When there is submucosal involvement, nodal metastasis is noted in 15
to 30% of resected specimens. This is considerably higher than the reported figures from the United
States105,107,108(4389) and Europe,115(4390) where the incidence of metastasis to regional lymph
nodes is 10%.
The treatment of EGC is surgical resection with removal of at least the first echelon lymph nodes. The
5-year survival rate of patients is 80 to 90% in Japan111,125128(4391) and 60 to 75% in the
West.98,102,105,107,109,110,115 Tumors larger than 1.5 cm in diameter, those with invasion of
submucosa, and those with positive lymph nodes have significantly less favorable outcomes.107(4392)
Maehara et al.,128(4393) using multivariate analysis of data on 423 Japanese patients with EGC,
found the presence of lymph node metastasis, histopathologic differentiation of the tumor, and patient
age to be independent prognostic factors. In a clinicopathologic study of 748 solitary EGCs, Sano et
al.129(4394) found that the presence of lymph node metastasis correlated with depth of invasion and
lesion diameter of greater than 1.5 cm. All 18 patients with EGCs less than 1.5 cm in diameter in the
series of Lawrence and Shiu107(4395) survived 5 years without evidence of disease.
A variety of endoscopic methods including laser photoablation and mucosectomy have been used in
the treatment of EGC in patients who are not candidates for surgery (see later).
Advanced Gastric Cancer

The classic clinical symptoms and findings of carcinoma of the stomach are vague and nonspecific
and include epigastric pain, bloating and early satiety, anemia, melena and hematemesis, weight loss,
and sometimes dysphagia. The symptoms thus differ little from EGC, except perhaps the degree of
weight loss.105(4396) Physical signs of abdominal or pelvic masses, ascites, or nodal enlargement
indicate a late stage of the disease and unresectability. An association between gastric cancer and
certain conditions has been recognized for some time. These conditions are chronic atrophic gastritis,
adenomas, hyperplastic polyps, history of gastric ulcers, partial gastrectomy (see Chapter 52:
Endoscopy in the Postoperative Upper Gastrointestinal Tract), intestinal metaplasia, and hyperplastic
gastritis. The various types of gastritis and gastropathies are discussed in Chapter 47. Heightened
vigilance for gastric cancer is warranted whenever one of these associated conditions is recognized.
Advanced gastric cancer has several well-known gross morphologic appearances as classified by
Borrmann130(4397) in the 1920s into four types. Recognition of these morphologic types remains
useful today, especially for the endoscopist (Figure 4821).
(4398)Figure 4821. Borrmann's morphologic classification of advanced gastric cancer. A,

Type I, polypoid carcinoma, a nodular lesion with gross ulcerations. B, Type II, a fungating
carcinoma, protrudes into the lumen, but the bulk of the lesion is ulcerated. C, Type III,
ulcerated, has only minimally elevated edges, and the main portion of the cancer is an
excavating lesion with cancerous tissues at the edges as well as the base. D, Type IV, the
diffusely infiltrative carcinoma, may have some minimal elevation or bulk at the origin, and
results in the classic form of linitis plastica when most of the stomach is involved.
A Borrmann Type I cancer is polypoid and fungating, often of considerable mass (Figure 4822).
Rarely, a broad pedicle may be recognized, with the head being more massive than the base.
(4399)Figure 4822. Polypoid carcinoma with superficial erosion (Borrmann Type I). (From
Borrmann Type II gastric cancer is mainly ulcerating (Figures 4823 and 4824). The malignant ulcer
typically has an elevated base with reference to the neighboring uninvolved mucosa, although in deeply
excavating lesions, the base may be depressed, indicating deep extension, sometimes even beyond
the wall of the stomach. The edges of the ulcer are elevated, nodular, and bulky. Necrosis,
hemorrhage, and local infection give rise to a multicolored appearance, from black to dark green to
maroon. The folds surrounding a malignant ulcer tend to taper, terminate abruptly, or become bulbous
(see Early Gastric Cancer). They are rigid and cannot be effaced by distension. Such characteristics
are in contrast to the benign ulcer, which has a regular outline with sharp edges and a smooth and
whitish or yellowish base except for discrete spots representing thrombosed vessels (see Chapter 46:
Peptic Diseases of the Stomach and Duodenum). The base may be depressed, but the edges are not
elevated or nodular. The folds around a healing gastric ulcer tend to be arranged radially; they
converge right up to the edge of the ulcer and remain the same size, giving a puckered appearance to
the region of involved mucosa. Even so, it must be emphasized that although a benign ulcer never
resembles an overtly cancerous process, the malignant ulcer sometimes may appear
benign,131,132(4400) with minimally elevated edges that are more or less regular. An example is the
Type III EGC in which partial healing often occurs, which further compounds the difficulty of making a
clear distinction. Indeed, Sakita et al.118(4401) demonstrated that total healing occurred in about 20%
of Type III EGCs. Data such as these invalidate the traditional concept that an ulcer that heals is
benign. A carcinoma was found in the edges of a preexisting chronic gastric ulcer, the so-called
malignant degeneration, in only 13 of 1900 cases (0.7%) of EGC arising in precancerous lesions as
documented in the registry of the National Cancer Center Hospital in Tokyo.111(4402)
(4403)Figure 4823. Polypoid, ulcerating carcinoma in the body of the stomach (Borrmann
Type II). Superficial ulcerations with oozing of blood are seen along with a biopsy forceps in
the 7 o'clock position. Cancer developing in an adenoma may result in this morphologic type.
(Courtesy of M. Derezin, M. D.)
(4404)Figure 4824. Retroverted view of the proximal stomach shows ulcerated

adenocarcinoma (Borrmann Type II). (From the collection of Dr. M. V. Sivak, Jr.)
Because gastric carcinoma may closely resemble a benign gastric ulcer, endoscopic follow-up of
benign ulcers to document healing has been advocated despite the fact that the ulcer cancer may
occasionally heal. However, the cost-to-benefit ratio of this practice is relatively high if there is no
suspicion of malignancy based on endoscopic appearance and assessment of biopsy and cytology
specimens.133135(4405) If, however, there is any suspicion of malignancy, further endoscopic
evaluation including collection of multiple biopsies and cytology specimens is required.
Borrmann Type III cancers are infiltrating and ulcerating with a wide surrounding zone of infiltrated
gastric folds that are enlarged, rigid, and discolored (Figure 4825). The infiltration may be subtle with
normal-appearing mucosa, but the unyielding nature of the folds can usually be appreciated.
(4406)Figure 4825. Deep ulceration in an infiltrating cancer of the antrum (Borrmann Type
III). (From the collection of Dr. M. V. Sivak, Jr.)
Borrmann Type IV cancers are diffusely infiltrating, often involving the entire stomach (Figure 4826).
These lesions present the greatest degree of difficulty in terms of exactly defining the extent of
infiltration. In these tumors, a nidus of ulceration may not be obvious, and there may or may not be an
area of maximum involvement. However, when most of the stomach is infiltrated, the lumen is
nondistensible; the stomach is small and the folds large and unchanging with distension, thus meriting
the term linitis plastica. Such an appearance is so strikingly abnormal that it is rarely missed.
(4407)Figure 4826. Gastric antrum infiltrated with carcinoma (Borrmann Type IV). The
antrum was nondistensible and noncontractile with an erythematous, nodular mucosa and no
obvious ulceration or mass. (From the collection of Dr. M. V. Sivak, Jr.)
The relative frequency for the different types of advanced gastric cancers in the experience of
Ming136(4408) (probably comparable with that for the Eastern United States), in decreasing order, is
as follows: Borrmann Type III (36%), Type IV (26%), Type II (26%), and Type I (7%). EGC accounted
for about 6% of all cases. There is some evidence that gross tumor morphology as described by the
Borrmann classification may effect the rate of positive biopsy diagnosis. Awad et al.137(4409) found
that biopsies were more likely to be positive for carcinoma if the tumor was exophytic as opposed to
ulcerative or infiltrative. These investigators postulated that the infiltrative nature of some tumors may
be related to simultaneous tumor expression of epidermal growth factor and its receptor.
Advanced gastric cancers are also classified by their anatomic location, and an exact description in this
regard is of major importance in planning surgical treatmenta fact not always fully appreciated by
nonsurgeon endoscopists. For carcinoma of the gastric cardia, the extent of esophageal involvement
must be determined, not only in terms of distance from the incisor teeth but also at the distance from
the diaphragmatic hiatus because this measurement will govern the surgical approach. The sector of
the lumen involved should also be described using correct anatomic terms such as medial half,
posterior quadrant, and so on. For carcinoma of the antrum, the tumor-free length of lesser curvature
between the gastroesophageal junction and the proximal margin of the tumor, which is often
shortened, determines whether a radical subtotal gastrectomy or a total gastrectomy should be
performed. Even the choice of surgical incision depends on accurate information as to the location of
the carcinoma. For lesions extending above the diaphragmatic hiatus, a left thoracoabdominal incision
is required. For more proximal involvement up the esophagus, two separate incisions, an upper
abdominal as well as a right thoracic (Ivor Lewis), would be necessary to carry out an
esophagogastrectomy. If the lesion is at the diaphragmatic hiatus or below, an abdominal approach is
often feasible because sufficient margin may be obtainable by mobilization of the mediastinal
esophagus through the hiatus. For lesions with sufficient free margins on the lesser curvature, the
lower esophageal sphincter may be spared and total gastrectomy may not be necessary.
To establish a histologic diagnosis of gastric carcinoma, multiple biopsies should be obtained.

Specimens must be taken from both the base and the edge of the ulcerated lesion. If a conventional
forceps is used, which imposes a limit on sample size, seven or more specimens must be
obtained.3,6(4410) Large-particle biopsies, such as snare excision of a fungating portion, may at times
be more expedient.
The functional impairment caused by the tumor should also be estimated during endoscopy. Gastric
outlet obstruction, if present, may warrant palliation even if the tumor is unresectable. In the case of an
unresectable carcinoma of the gastric cardia, encroachment on the esophagus often produces
disabling dysphagia, for which endoscopic palliation may be an option (see Chapter 42: Palliation of
Malignant Obstruction: Dilation and Peroral Prosthesis and Chapter 43: Palliation of Malignant
Obstruction: Lasers and Tumor Probes). If the tumor is on the gastric fundus abutting the
gastroesophageal junction, achalasia may be diagnosed erroneously because both symptoms and
manometry may be typical of primary achalasia (see Chapter 39: Esophageal Motility and
Miscellaneous Disorders).138(4411) For lesions with extensive necrosis and ulceration, blood loss is
significant and palliative resection may improve the quality of life.139141(4412)
Treatment of advanced gastric cancer is primarily surgical. In addition to endoscopic assessment, the
tumor may be further staged by EUS, CT, and laparoscopy. When such investigations show no distant
metastasis, exploration with the aim of performing a curative resection by total removal of all gross
tumor should be undertaken. The actual operation should consist of resection of the stomach with
tumor-free resection margins together with lymphadenectomy of the first echelon of lymph nodes, even
if there is no gross evidence of nodal involvement. The lymphadenectomy should be extended to
include the second echelon if positive nodes are found.141143(4413) For carcinoma of the antrum, a
radical subtotal gastrectomy is one such operation, preferred whenever possible. When the cancer is
located high in the body or fundus of the stomach, a total gastrectomy or esophagogastrectomy may
be necessary.
Curative resection in large series has been possible in 30 to 50% of patients.143145(4414) The
5-year survival rate of patients undergoing curative resection is 25 to 35%.143145(4415) Only 3% of
patients with unresectable cancers survive 4 years and none for 5 years.144(4416) From Japan, where
radical lymphadenectomy is commonly practiced, there are some reports of curative resection rates
upward of 50% and 5-year survival rates of 35 to 45%.127,146,147(4417)
Palliative resection, when curative resection is not possible, is at times worthwhile. The quality of
palliation of dysphagia from obstruction, pain from ulceration, anemia from blood loss, and early satiety
from tumor bulk is best with surgical resection at the risk of a higher complication rate and
mortality.139141(4418) The survival rate for patients undergoing palliative resection is better than that
for patients who do not undergo resection, owing partly to patient selection, and is therefore
appropriate for patients without overt metastatic disease who are good surgical risks.141(4419)
For inoperable patients and those with metastatic disease, chemotherapy using various combinations
of agents has been shown to increase survival in responding patients.148,149(4420) Conventional
external beam radiotherapy is problematic because of the surrounding radiosensitive organs, and a
definitive treatment dose of 6000 rads cannot be safely delivered. Combined chemotherapy and
radiotherapy is also under investigation. Adjuvant chemotherapy after surgical resection has been
shown to prolong survival compared with surgery alone.150(4421) Endoscopy also plays a role in
palliation, especially for relief of dysphagia due to inoperable carcinoma of the gastric cardia. Several
endoscopic options may be considered (see later).
Prognosis is influenced by the location, gross morphology, and histopathologic type of the cancer.
Carcinoma of the gastric cardia carries a poorer prognosis than carcinoma of other areas of the
stomach, stage for stage.151,152(4422) An increase in cancers in this region has been noted since the
1970s.153(4423) However, unless the tumor is small it may not be possible to determine whether it
arises from the gastric epithelium or Barrett's epithelium within the esophagus. A true decrease in
antral gastric cancer has been noted in Japan and the United States.153(4424) In terms of gross
morphology, which reflects growth pattern, patients with an infiltrative type of tumor have a survival rate
that is half that of patients with expanding tumors.154(4425) Prognosis also depends upon histology.
The intestinal type is more favorable than the diffuse type, which is prone to develop peritoneal
involvement.155(4426) Marked infiltration of the tumor stroma with lymphoid and Langerhans cells has
a favorable effect on prognosis.156(4427) A high percentage of signet-ring cells is adverse in terms of
prognosis.
Lymphoma
Lymphoma may develop in the stomach as the primary lesion in non-Hodgkin's lymphoma without
involvement of the liver, spleen, or mediastinal or retroperitoneal lymph nodes or as a secondary
manifestation of generalized non-Hodgkin's lymphoma. Together, primary and secondary lymphomas
account for 5% of all gastric malignancies. Clinical conditions that predispose to lymphomas of the
gastrointestinal tract include celiac sprue, chronic immunosuppression after organ transplantation,
acquired immunodeficiency syndrome (AIDS), and immunoproliferative small intestinal disease.
The clinical presentation is usually that of ulcer pain that does not completely respond to medical
treatment. Bleeding and even perforation may occasionally be initial manifestations. Obstructive
symptoms occur when the tumor is large and strategically located, with nausea and vomiting, anorexia,
and weight loss.
Gastric lymphoma has several distinguishing gross morphologic characteristics that, when present,
alert the endoscopist to the possibility of the diagnosis. The presence of these features indicates that
special biopsy techniques may be required for tissue diagnosis.
Endoscopically, lymphoma presents in several forms.28,157159(4428) The massive or polypoid form
usually appears at endoscopy as a large tumor mass with ulcerations of varying depths that coalesce
to form huge cavities. The nodular form appears as multiple polypoid masses or multiple ulcerated
infiltrating lesions in close proximity to one another (Figures 4827 and 4828). The superficial
spreading form has the appearance of a diffuse, widespread mucosal abnormality consisting of a
granular and irregular surface, thick and nodular folds, with multiple superficial ulcerations (Figure
4829). It often affects one region more than other areas of the stomach. The infiltrative form appears
as giant folds, localized to one area of the stomach, often with superficial ulceration or erosions. If
generalized, the differential diagnoses would include Mntrier's disease and linitis plastica. The
infiltrative lymphoma may be especially difficult to diagnose at endoscopy.160(4429) Often a mixture of
several morphologic forms is seen (e.g., giant folds next to a large ulcerating tumor, a main polypoid
lesion in a field of smaller nodular masses).
(4430)Figure 4827. Infiltrative non-Hodgkin's lymphoma with polypoid nodules and

superficial ulcerations. (From the collection of Dr. M. V. Sivak, Jr.)
(4431)Figure 4828. Non-Hodgkin's lymphoma at the incisura seen on a retroverted view.

The ulcer is irregularly shaped with a poorly demarcated margin; the adjoining mucosa is
nodular. (From the collection of Dr. M. V. Sivak, Jr.)
(4432)Figure 4829. Lymphoma involving the greater curvature of the body. A, Giant folds
with superficial ulcerations. B, A closer view showing an enlarged, ulcerated fold with a
mottled surface. (A and B, From the collection of Dr. M. V. Sivak, Jr.)
Plasmacytoma is typically composed almost exclusively of plasma cells with relatively few inflammatory
cells. Immunohistochemical staining will usually reveal a single immunoglobulin. The lesions have
clinical features in common with multiple myeloma, but plasmacytosis is not present in the bone
marrow. In most cases of solitary plasmacytoma, the lesion is located in the stomach or small intestine.
The lesions usually appear at endoscopy as relatively discrete tumor masses, either umbilicated or
ulcerated.161,162(4433)
Because the bulk of lymphoma resides in the submucosa and deeper histologic layers of the stomach,
the conventional forceps biopsy may not provide a histologic diagnosis. It is advisable to obtain a large
number of biopsies, but even then, the diagnosis may be uncertain or evident in only one or a few
specimens.163(4434) The difficulty is compounded by the crush artifact, with severe distortion of the
architecture that occurs when specimens are obtained with the conventional forceps, and by the
presence of mixed inflammatory infiltrate due to surface necrosis or ulceration of the tumor.163(4435)
Snare excisional biopsy, large (jumbo) forceps biopsy, and other techniques of large-particle biopsy
should therefore be used.164(4436) Collection of cytologic specimens is also helpful. Examination of
frozen sections is advantageous because further tissue sampling can be performed until a positive
histologic diagnosis is obtained, thereby sparing the patient repeated endoscopic procedures, delays in
diagnosis, and anxiety. Moreover, when a diagnosis has been established by evaluation of frozen
sections, additional tissue specimens can be obtained for special purposes, such as electron
microscopy and special stains, to establish the subtype of the tumor, a factor that influences both
treatment and prognosis.
Despite the use of all available endoscopic techniques for tissue sampling, the histologic diagnosis of
lymphoma may remain elusive. In such cases, newer methods for study of endoscopic tissue
specimens may enhance diagnostic capability. For example, Fend et al.165(4437) found that
immunoglobulin gene rearrangement analysis of endoscopic gastric biopsies enhanced the ability to
diagnose malignant B-cell lymphoma. Detection of clonal rearrangements assisted in establishing the
diagnosis in three cases where histopathologic evaluation was inconclusive; in two patients, the
discovery of clonal rearrangements preceded histologic diagnosis by several months.
Once the diagnosis of gastric lymphoma has been established, the extent of disease must be
determined by evaluating the patient for extragastric and systemic involvement. EUS has contributed
significantly to the diagnosis of mural and local extramural spread of the tumor (see Chapter 51:
Endoscopic Ultrasonography: Stomach and Duodenum).2628(4438) The evaluation for systemic
spread is based on selected use of CT of the body cavities, bone marrow aspiration,
lymphangiography, laparoscopy, and occasionally laparotomy. Thus, in the Ann Arbor system of clinical
staging, non-Hodgkin's lymphoma is classified into Stages IE, IIE, III, and IV.166, 167(4439) Stage IE
is disease limited to the stomach (25 to 35% of cases). In Stage IIE, the abdominal lymph nodes are
involved as demonstrated by biopsy, lymphangiography, or EUS (50%). Stage III is defined as Stage
IIE plus nodal involvement above the diaphragm, whereas Stage IV is disseminated lymphoma (15 to
20%). The diffuse histiocytic (large cell) type is the most common histopathologically (59%), followed in
frequency by the small cleaved cell (17%), mixed (3%), small noncleaved (Burkitt's) (8%), and
miscellaneous types (9%).166(4440)
The primary treatment for localized disease is surgical resection (Stage IE), with chemotherapy and
radiotherapy usually reserved for other stages. There is some evidence that postoperative radiotherapy
or chemotherapy may be beneficial for patients with Stage IIE and even for Stage IE
disease.167(4441)
The prognosis for gastric lymphoma is better in general than that for carcinoma. An overall 5-year
survival of about 50% can be expected, but for Stages IE and IIE with smaller tumors, a 10-year
survival of 80% can be attained.168,169(4442)
Carcinoid
Gastric carcinoids are uncommon tumors. They are often asymptomatic, and the diagnosis is
frequently incidental. When symptomatic, the most common presentation is bleeding. Most are benign
lesions, but up to 25% may be malignant with metastatic potential. Interest in
hypergastrinemia-induced proliferation of enterochromaffin-like cells has heightened the awareness
that carcinoid tumors may be found in association with chronic hypergastrinemia, as in patients with
the Zollinger-Ellison syndrome, atrophic gastritis with achlorhydria, and potentially in patients taking
omeprazole for long periods of time.170(4443) A prospective endoscopic study of 123 patients with
pernicious anemia identified 5 patients with carcinoid tumors; 1 patient had multiple lesions.171(4444)
In another endoscopic study of 105 patients with pernicious anemia, carcinoid tumors were discovered
in 5 patients.172(4445) Regression of gastric carcinoid tumors located in the proximal stomach has
been described after antrectomy.173(4446)
Gastric carcinoid tumors have been described at endoscopy as both polypoid (sessile and
pedunculated) and submucosal lesions.174177(4447) There are numerous reports of multiple
carcinoid tumors in the stomach, usually in patients with hypergastrinemia.171,174,177,178(4448) As
with other submucosal tumors, the lesion may be covered with normal mucosa.176(4449) Often, a
small central dimple or ulceration may be present (Figure 4830). Although distinguishing
characteristics are lacking, the possibility should come to mind when a small submucosal lesion with a
central ulceration is seen, especially if there are multiple such lesions.
(4450)Figure 4830. Gastric carcinoid. Note the minute central ulceration.

Carcinoid syndrome, if found, may indicate the presence of other carcinoids or metastatic lesions in the
liver. A syndrome that resembles that associated with hepatic metastases from carcinoid tumors of the
small bowel has been described.179181(4451)
A diagnosis of malignant carcinoid is usually not established with biopsy because the histology is
unhelpful. Demonstration of metastatic spread is required for diagnosis. The occurrence of metastasis
appears to be related to the size of the tumor.182(4452) Because of their malignant potential, it is
advisable to remove gastric carcinoid tumors. Gastric carcinoids are highly vascular. Unless the tumors
are very small, endoscopic excision is not generally advised. For biopsy-proven carcinoids, local
surgical resection is therefore indicated.183(4453)
Metastatic Tumors
Tumors metastatic to the stomach are uncommon, being found in 0.2% of autopsies. Not counting
direct invasions from the colon, pancreas, and esophagus, hematogenous metastasis may occur with
several common carcinomas. Of these, bronchogenic carcinoma accounts for half of the cases of
metastatic gastric tumors;184,185(4454) the remainder arise from primary tumors of the pancreas,
esophagus, colon, breast,186,187(4455) and cutaneous melanoma.188190 There are case reports of
metastasis to the stomach from choriocarcinoma,191(4456) prostate carcinoma,192(4457) a
mucoepidermoid parotid tumor,193(4458) and chondrosarcoma of the right knee.194(4459) Multiple
myeloma,195(4460) chronic myelogenous leukemia,196(4461) and lymphocytic leukemia197(4462)

have been known to produce endoscopically visible lesions in the stomach. By far the majority of
metastatic tumors present as submucosal lesions, and their endoscopic detection is mostly incidental.
Larger lesions are polypoid; some develop central necrosis resulting in umbilication (Figure 4831).
Bleeding is the most common complication.187(4463) Massive bleeding tends to occur, especially in
metastatic melanoma.
(4464)Figure 4831. Metastatic breast carcinoma. Multiple similar submucosal nodules with
central ulceration were found. Biopsies were positive for carcinoma. (From the collection of
Dr. M. V. Sivak, Jr.)
Metastatic tumors in the stomach may be multiple, but more often, they are solitary. There are no
endoscopic characteristics that identify their origin except in the case of smaller melanomas, which
may be pigmented. Large melanoma metastases, however, lose their pigmentation and have no
unique characteristics.198(4465) Metastatic breast cancer sometimes causes diffuse thickening of the
gastric wall simulating linitis plastica.199(4466)
Kaposi's Sarcoma
The gastrointestinal tract of the patient with AIDS is frequently affected by Kaposi's sarcoma, and the
gastric antrum and body are two of the many favored sites. Approximately 50% of patients with skin
involvement may have gastrointestinal Kaposi's sarcomas. Endoscopically, these appear as bright red,
smooth, round, sessile nodules, 0.5 to 1 cm in diameter, located mainly in the submucosa (Figure
4832).200203(4467) Multiple lesions are common. Larger nodules may ulcerate. The histologic
appearance is diagnostic with the unique features of spindle cell proliferation, interlacing thin-walled
vascular spaces lined by endothelial cells, and extravasculated red blood cells. The endoscopic
diagnosis is not difficult if the clinical circumstances are known.
(4468)Figure 4832. Kaposi's sarcoma in the gastric antrum. (From the collection of Dr. M.
V. Sivak, Jr.)
Myosarcomas
Leiomyosarcomas and malignant leiomyoblastomas make up the subtypes of sarcoma of the stomach.
They represent about 1% of gastric malignancies. They have no distinguishing characteristics at
endoscopy. When small, they resemble leiomyomas and, indeed, most are not diagnosed as
sarcomas before operationan argument in favor of routine surgical excision for leiomyomas. Like
their benign counterpart, the myosarcomas are commonly found in the fundus and body and less often
in the antrum of the stomach. However, when they present as a large mass, they have a bosselated
surface with breaching through the mucosa in one or more areas (Figure 4833). Large blood vessels
are visible on the surface or in the submucosa. Necrosis and softening may be recognizable, and
ulcerations, present in over 80% of lesions, can be deep and hemorrhagic. The main tumor mass can
be predominantly luminal, intramural, or extragastric, depending on the direction of growth (Figure
4834). Endoscopic electrosurgical snare resection of a pedunculated and ulcerated leiomyoblastoma
in an elderly patient with upper gastrointestinal bleeding has been described.204(4469)
(4470)Figure 4833. Resected specimen containing a leiomyoblastoma. Note the bosselated

surface on one aspect of the tumor. (Courtesy of Dr. K. Lewin.)
(4471)Figure 4834. Leiomyoblastoma. A, Excisional biopsy specimen shows early serosal

encroachment. B, Mucosal surface. Note the absence of distinguishing features to separate it
from the benign counterpart.
Prognosis is related to the size of the tumor, histologic grade, the presence of invasion of adjacent
organ(s), and metastasis.205,206(4472) Tumors less than 5 cm in diameter rarely metastasize.
Treatment is surgical wedge resection with clear margins. Neither the addition of lymph-adenectomy
nor wider gastric resection, such as radical gastrectomy, confers any survival benefit.206(4473) In one
series, patients with tumors less than 5 cm in diameter had a 5-year survival rate of close to
100%.205(4474) An overall survival rate of 63% was recorded in another series after curative
resection.206(4475)
Other Malignant Tumors

There are a number of unusual malignant tumors of epithelial and mesenchymal origins that have
limited clinical relevance because of their infrequent occurrence. However, a knowledge of these
lesions is useful in the differential diagnosis of gastric masses.
Adenosquamous and pure squamous cell carcinomas are rare tumors with squamous cell
differentiation of varying degrees. Most are found in the antrum. Their biologic behavior is similar to
that of advanced gastric cancer.
Carcinosarcoma is a curious mixture of adenocarcinoma components and spindle cell sarcoma with
varying degrees of intermingling. It is believed that these tumors derive from different cell lines within
the same tumor.207,208(4476) There are no gross distinguishing characteristics.
Sarcoma of miscellaneous patterns is a convenient term for a very aggressive but fortunately rare
tumor consisting of spindle cells of pronounced pleomorphism and high mitotic rate. Histologically
difficult to clarify, they share the common characteristic of rapid dissemination in the abdomen and
widespread hematogenous metastases.208(4477)
Choriocarcinoma is rare, but it is sometimes found in association with adenocarcinoma. Although there
are no diagnostic endoscopic characteristics, the histology is typical. The tumor consists of sheets of
syncytiotrophoblasts and cytotrophoblasts, which secrete significant amounts of human chorionic
gonadotropic hormone (hCG), manifesting as gynecomastia in men. Although hCG is commonly found
in many adenocarcinomas of the stomach, only choriocarcinoma secretes enough to cause clinical
effects.
Endoscopic Treatment
Endoscopic resection of gastric lesions evolves from the technique of polypectomy. With the advent of
laparoscopic surgery and laparoscopic stapling instruments, local gastric resections can be performed
laparoscopically, often in combination with gastroscopy, which affords precise localization of the region
to be resected.
Endoscopic Excision
Mucosal lesions can be removed by electrosurgical snare polypectomy. The technique is similar to that
for colonoscopic polypectomy. Electrosurgical snare excision, with modifications of the technique, can
also be used for excision of some submucosal lesions.
Polyps
Polypectomy is preferable to forceps biopsy for the management of gastric polyps because both
diagnosis and therapy are accomplished in one step.37,209213(4478) If polyps are multiple but fewer
than about 10 in number, all of the lesions should be excised and the specimens carefully labeled and
mapped on a diagram of the stomach. Excision of all lesions is impractical in patients with diffuse
polyposis, although random sampling is appropriate. Sampling should be performed in FAP patients for
diagnosis, especially if the endoscopic appearance or location of the polyp suggests that it may be an
adenoma.
The gastric mucosa is thicker and more vascular than the colonic mucosa, and therefore certain
precautions are necessary. To resect a large polyp (2 cm or more in diameter) or a polyp with a wide
base or thick stalk, it is often advisable to utilize a double-channel endoscope. The polyp grasper can
be passed down one channel and the snare down the other. The grasper is passed through the loop of
the open snare loop, the polyp is grasped and lifted with the grasper, and the snare is closed around
the polyp. The degree of lift controls how much gastric tissue is included in the snare loop. It is possible
to include too much tissue and risk perforation. Be suspicious of this if the tissue enclosed by the snare
seems excessively bulky, if there is a "springy" feel to the closure, or if no color change is seen during
the initial period of mechanical strangulation. Incomplete excision is always preferable to perforation.
As in the colon, piecemeal excision with the snare can be used for larger sessile lesions. Although
experience with bipolar snares is limited, the risk of a full-thickness burn of the gut wall is reduced with
use of this device.
Avoidance of postpolypectomy bleeding requires careful inspection after excision and the use of
endoscopic techniques for hemostasis when indicated. If bleeding occurs after excision of a
pedunculated polyp, it is frequently useful to regrasp the remaining stalk of the polyp with the snare.
Hemostasis should be achieved by mechanical compression rather than by further application of the
coagulating current. The force of closure of the snare should be just sufficient to stem the bleeding
because excessive force will mechanically sever the stalk and result in further bleeding that may be
even more difficult to control.
The hemostatic technique of resnaring the bleeding polyp stalk is usually not suitable if the lesion was
sessile or the stalk was transected close to the gastric wall. When bleeding occurs, it often comes from
the submucosa, which is the vascular layer of the stomach, and appears endoscopically as oozing
from under the edges of the mucosal defect created by the polypectomy. Such bleeding can be treated
with injections of a solution of epinephrine (1:10,000). Unlike the heat probe and electrocoagulation
probes, epinephrine injection does not increase the area of ulceration that will result from the
polypectomy. However, there are potential complications of injection therapy (see Chapter 30: Injection
Therapy for Ulcer Bleeding). Because the ulcer resulting from polypectomy can be large and require
time to heal, antiulcer pharmacotherapy is usually prescribed in the follow-up care of patients who have
undergone gastric polypectomy.
Endoscopic treatment methods have been employed in the treatment of EGC in patients who are not
suitable candidates for surgery. These include laser photoablation, alcohol injection, polypectomy, and
mucosectomy, as well as combinations of these techniques.
Photoablation has been performed using lasers (see Chapter 43: Palliation of Malignant Obstruction:
Lasers and Tumor Probes).214(4479) Available data suggest that this method is effective provided
specific criteria are observed in the selection of lesions for treatment.215217(4480) However, there
may be residual malignancy if the tumor is incompletely treated.218(4481) Yasuda et al.216(4482)
reported their experience with photoablation of EGC in 111 patients. Surgery was performed
subsequent to endoscopic therapy in 12 patients and residual tumor was found in 3 (25%). Of the 99
patients who did not undergo surgery, endoscopic follow-up was performed in 73. There was no
residual malignancy in 81% (mean follow-up 2.7 years; range 1 to 7 years).
A major disadvantage of laser photoablation for EGC is that endoscopic assessment alone is not
perfectly accurate as a method of tumor staging. A lesion thought to be superficial may actually be
relatively invasive, in which case surgery would offer a better prospect for cure. Furthermore,
histopathologic assessment of the stage of the lesion is precluded and the effectiveness of treatment
can be determined only on the basis of follow-up endoscopy with biopsies. EUS increases the
accuracy of endoscopic staging, but the evaluation of small, early-stage and relatively noninvasive
cancers can be difficult.
Mucosectomy (strip biopsy) appears to have largely supplanted laser photoablation as the primary
method for endoscopic treatment of EGCs that are 1 cm or less in diameter.219(4483) For technical
reasons, however, mucosectomy is often used in combination with other treatment methods, especially
laser therapy. Usually, the lesion is first elevated by injecting saline solution beneath it. Special barbed
snares are often used to grasp the elevated tumor, which is then removed using electrosurgical
current. Various modifications of the basic technique have been described.220222(4484) Available
data suggest that mucosectomy is potentially curative when the EGC can be completely removed, as
determined by subsequent histopathologic assessment of the resected specimen.215,223225(4485)
Complete primary resection is usually not possible for lesions that are more than 2 cm in diameter. In
such cases, the residual cancer is often treated by laser photoablation. Based on a review of 308
cases of EGC treated by endoscopic resection, Tada et al.225(4486) found that 266 lesions could be
resected completely in a single-step procedure. Residual or recurrent cancer was subsequently found
in 44 cases (17%). Incomplete resection was least likely to occur with lesions located on the lesser
curvature as opposed to those on the anterior or posterior wall. Lesions considered as being the most
suitable for endoscopic resection were Type IIa EGC of less than 2 cm in diameter and Type IIc
differentiated adenocarcinoma of less than 1 cm.
Submucosal Tumors
The electrosurgical snare polypectomy technique can be modified for excision of submucosal lesions.
Examples of suitable lesions for excisions include symptomatic leiomyomas or lipomas, smaller than 4
to 5 cm in diameter. It is extremely helpful, perhaps necessary, to demonstrate the exact depth and
layer of origin of the lesion by EUS and that a plane exists between the tumor and the other layers of
the gastric wall.226,227(4487) A circumferential cut of controlled depth is first made with an open
snare on the gastric mucosa surrounding the lesion (Figure 4835). The plane of dissection between
the lesion and the normal submucosa can then be visualized. Dissection in this bloodless plane by
means of endoscopic scissors allows rapid freeing of the lesion, which is further facilitated by traction.
Using a double-channel endoscope, a polyp grasper, inserted through the open snare loop, provides
traction and allows closure of the snare around the elevated lesion in a manner similar to that
described for large-particle biopsy. In this fashion, closure of the snare actually squeezes the tumor
into the gastric lumen. It is helpful to inject 1:10,000 epinephrine into the submucosa before
commencing excision227(4488) because the resulting wheal facilitates snare excision in addition to
reducing hemorrhage.
(4489)Figure 4835. Endoscopic excision of submucosal tumors. A, Preliminary submucosal

infiltration with saline or 1:10,000 epinephrine solution. B, A circular incision is made with
the electrocautery snare near the lower part of the lesion. C, The plane between the lower part
of the tumor and the gastric wall is defined, aided by traction on the tumor using the snare. D
and E, Enucleation is completed by tightening the snare at the pedicle.
Laparoscopic Gastric Resections

Limited local gastric resections to remove a small tumor or obtain a full-thickness biopsy of the gastric
wall, heretofore performed as open procedures, may now be done laparoscopically.228(4490) Lesions
situated on the anterior gastric wall or along the greater curvature are suitable for this technique,
whereas lesions situated on the lesser curvature or near the esophageal entry are not. The lesion to be
resected can be accurately localized during laparoscopy by simultaneous gastroscopy with either
endoscopic transillumination, ink injection, or fine-wire puncture. Once localized, the area for resection
is devascularized with clip ligations, freed up, and grasped by forceps. A laparoscopic stapler is
inserted and placed across the base of the portion of the gastric wall held by the forceps. After firing
the stapler, which automatically excises the grasped portion of the wall and places a suture line in one
operation, the excised lesion is retrieved as the specimen.
Clinical experience with gastrointestinal stapling is extensive, and its efficacy and safety are
comparable with those of suturing. If experience with this technique proves to be similar when used for
newer laparoscopic applications, it is likely that laparoscopic gastric resection will become a widely
performed operation in the near future.
Endoscopic Palliation of Advanced Gastric Cancer

Inoperable gastric cancer is a challenging clinical problem. Technical expertise apart, the decision to
embark on endoscopic treatment is difficult and the compassionate clinician must assess the prospect
of relieving symptoms with procedures of relatively low invasiveness but unproved efficacy and
unknown complication rates. Nevertheless, the endoscopist may be called on to provide symptomatic
relief of obstruction, bleeding, and enteral nutrition.
Satiety and inanition are universal symptoms of advanced gastric malignancy. However, unless the
patient is preterminal, improved nutrition may enhance the quality of the remaining months of life.
Percutaneous endoscopic gastrostomy (PEG) (see Chapter 55: Percutaneous Endoscopic
Gastrostomy) may be feasible if the stomach is not extensively involved. If PEG is not advisable,
feeding jejunostomy can be readily accomplished laparoscopically229(4491) and is tolerated even by
patients with advanced disease.
Obstructive symptoms caused by carcinoma of the gastric cardia may be palliated in a variety of ways,
including stenting with endoprosthesis (see Chapter 42: Palliation of Malignant Obstruction: Dilation
and Peroral Prosthesis), laser photoablation (Chapter 43: Palliation of Malignant Obstruction: Lasers
and Tumor Probes), electrocautery snare excision, use of tumor probes (Chapter 43), and alcohol
injection. The last four modalities can be used to debulk a tumor, but this is rarely indicated in the
stomach unless the tumor mass is causing obstruction.
There is no effective endoscopic treatment for hemorrhage from a tumor. Often, the bleeding is slow,
chronic, or intermittent. The use of hemostatic probes (heat probe, multipolar electrocoagulation probe)
for the local treatment of areas that are oozing blood may temporarily control bleeding, but the effect is
almost always short-lived. Endoscopic therapy can be repeated for recurrent bleeding, but the interval
between procedures is necessarily short and this has an adverse effect on the quality of life. Laser
photocoagulation has been used to palliate bleeding, but there are no data to indicate that it is effective
and claims of success are largely anecdotal. Photodynamic therapy has met with limited success (see
Chapter 19). The use of hemostatic sprays and more novel approaches have yet to be evaluated.
Endoscopic palliation of gastric cancer is a rudimentary art at present. Until newer methods and
technology are introduced and tested, it is prudent to resist the temptation to attempt too much.
REFERENCES
1.
2.
3.
Danesh BJ, Burke M, Newman J, et al. Comparison of weight, depth, and diagnostic
adequacy of specimens obtained with 16 different biopsy forceps designed for upper
gastrointestinal endoscopy. Gut 1985;26:22731.
Komorowski RA, Caya JG, Geenen JE. The morphologic spectrum of large gastric folds:
Utility of the snare biopsy. Gastrointest Endosc 1986;32:1902.
Hatfield ARW, Slavin G, Segal AW, Levi AJ. Importance of the site of endoscopic gastric
biopsy in ulcerating lesions of the stomach. Gut 1975;16:8846.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Lin HJ, Lee FY, Tsai YT, et al. A prospective evaluation of biopsy site in the diagnosis of
gastric malignancy: The margin or the base?. J Gastroenterol Hepatol 1989;4:13741.
Misumi A, Mori K, Ikeda T, et al. Evaluation of fibergastroscopic biopsy in the diagnosis of
gastric cancer: A study of 339 cases. Gastroenterol Jpn 1978;13:25563.
Graham DY, Schwartz JT, Cain GD, Gyorkey F. Prospective evaluation of biopsy number in
Dekker W, Tytgat GN. Diagnostic accuracy of fiberendoscopy in the detection of upper
intestinal malignancy. A follow-up analysis. Gastroenterology 1977;73:7104.
Jorde R, Ostensen H, Bostad LH, et al. Cancer detection in biopsy specimens taken from
different types of gastric lesions. Cancer 1986;58:37682.
Shanghai Gastrointestinal Endoscopy Cooperative Group, People's Republic of China. Value
of biopsy and brush cytology in the diagnosis of gastric cancer. Gut 1982;23:7746.
Cusso X, Mones J, Ocana J, et al. Is endoscopic gastric cytology worthwhile? An evaluation
of 903 cases of carcinoma. J Clin Gastroenterol 1993;16:3369.
Moreno-Otero R, Martinez-Raposo A, Cantero J, Pajares JM. Exfoliative cytodiagnosis of
gastric adenocarcinoma. Comparison with biopsy and endoscopy. Acta Cytol 1983;27:4858.
Todorova-Dortscheva Z, Hakkiluoto A. Washing and direct vision brush cytology in the
diagnosis of gastric carcinoma. Scand J Gastroenterol 1976;11:5936.
Keighley MRB, Thompson H, Moore J, et al. Comparison of brush cytology before or after
biopsy for diagnosis of gastric carcinoma. Br J Surg 1979;66:2467.
Waldron R, Kerin M, Ali A, et al. Evaluation of the role of endoscopic biopsies and cytology in
the detection of gastric malignancy. Br J Surg 1990;77:623.
Witzel L, Halter F, Gretillat PA, et al. Evaluation of specific value of endoscopic biopsies and
Cook IJ, de Carle DJ, Haneman B, et al. The role of brushing cytology in the diagnosis of
gastric malignancy. Acta Cytol 1988;32:4614.
Kiil J, Andersen D, Jensen OM. Biopsy and brush cytology in the diagnosis of gastric cancer.
Kobayashi S, Yoshii Y, Kasugai T. Biopsy and cytology in the diagnosis of early gastric
cancer: 10-year experience with direct vision techniques at a Japanese institution. Endoscopy
1977;8:538.
Sherman ME, Anderson C, Herman LM, et al. Utility of gastric brushing in the diagnosis of
malignant lymphoma. Acta Cytol 1994;38:16974.
Martin LF. Stress ulcers are common after aortic surgery. Endoscopic evaluation of
prophylactic therapy. Am Surg 1994;60:16974.
Ingoldby CJ, Mason MK, Hall RI. Endoscopic needle aspiration cytology: A new method for
Lange P, Kock K, Laustsen J, et al. Endoscopic fine-needle aspiration cytology of the
stomach. A new diagnostic procedure. Endoscopy 1987;19:723.
Iishi H, Yamamoto R, Tatsuta M, Okuda S. Evaluation of fine-needle aspiration biopsy under
direct vision gastrofiberscopy in diagnosis of diffusely infiltrative carcinoma of the stomach.
Cancer 1986;57:13659.
Singh S, Macleod G, Walker T, et al. Endoscopic fine-needle aspiration cytology in the
diagnosis of linitis plastica. Br J Surg 1994;81:1010.
Bolondi L, Casanova P, Caletti GC, et al. Primary gastric lymphoma versus gastric carcinoma:
Endoscopic US evaluation. Radiology 1987;165:8216.
Caletti GC, Lorena Z, Bolondi L, et al. Impact of endoscopic ultrasonography on diagnosis and
treatment of primary gastric lymphoma. Surgery 1988;103:31520.
Tio TL, den Hartog Jager FC, Tytgat GN. Endoscopic ultrasonography of non-Hodgkin
Suekane H, Iida M, Yao T, et al. Endoscopic ultrasonography in primary gastric lymphoma:
Correlation with endoscopic and histologic findings. Gastrointest Endosc 1993;39:13945.
Shirakabe H. Double Contrast Studies of the Stomach. Tokyo: Bunkoda. 1971.
30. Tsuchigame T, Ogata Y, Sumi M, et al. Differential diagnosis of gastric adenoma and type IIa
early gastric cancer. Acta Radiol 1991;32:3741.
31. Barentaz JO, Rosenbusch GR, Strijk SP, et al. Radiological examination in gastric cancers: A
retrospective study of 188 patients. Acta Radiol Diagn (Stockh) 1986;27:54752.
32. Watt I, Stewart I, Anderson D, et al. Laparoscopy, ultrasound and computed tomography in
cancer of the esophagus and gastric cardia: A prospective comparison for detecting
intra-abdominal metastases. Br J Surg 1989;76:10369.
33. Possik RA, Franco EL, Pires DR, et al. Sensitivity, specificity, and predictive value of
laparoscopy for the staging of gastric cancer and for the detection of liver metastasis. Cancer
1986;58:18.
34. Kriplani AR, Rapur BM. Laparoscopy for preoperative staging and assessment of operability
in gastric carcinoma. Gastrointest Endosc 1991;37:4416.
35.
Ming SC. The classification and significance of gastric polyps. In Yardley JH, Morson BM
(eds). The Gastrointestinal Tract. Baltimore: Williams & Wilkins, 1977;14975.
36.
Ming SC, Goldman H. Gastric polyps: A histogenetic classification and its relation to
carcinoma. Cancer 1965;18:7216.
37. Seifert E, Elster K. Gastric polypectomy. Am J Gastroenterol 1975;63:4516.
38. Kamiya T, Morishita T, Asakura H, et al. Long term follow-up on gastric adenoma and its
relation to gastric protruded carcinoma. Cancer 1982;50:2493503.
39.
Hirota T, Okada T, Itabashi M, et al. Histogenesis of human gastric cancerWith special
reference to the significance of adenoma as a precancerous lesion. In Ming SC (ed).
Precursors of Gastric Cancer. New York: Praeger, 1984;23352.
40. ReMine SG, Hughes RW Jr, Weiland LH. Endoscopic gastric polypectomies. Mayo Clin Proc
1981;56:3715.
41. Cristallini EG, Ascani S, Bolis GB. Association between histologic type of polyp and carcinoma
in the stomach. Gastrointest Endosc 1992;38:4814.
42. Nakano H, Persson B, Slezak P. Study of the gastric mucosal background in patients with
gastric polyps. Gastrointest Endosc 1990;36:3942.
43. Stemmermann GN, Hayashi T. Hyperplastic polyps of gastric mucosa adjacent to
gastroenterostomy stomas. Am J Clin Pathol 1979;71:3415.
44. Snover DC. Benign epithelial polyps of the stomach. Pathol Annu 1985;20:30329.
45. Daibo M, Itabashi M, Hirota T. Malignant transformation of gastric hyperplastic polyps. Am J
46. Dirschmid K, Walser J, Hugel H. Pseudomalignant erosion in hyperplastic gastric polyps.
Cancer 1984;54:22903.
47. Rosen S, Hoak D. Intramucosal carcinoma developing in a hyperplastic gastric polyp.
48. Toshiaki K, Morishita T, Asakura H, et al. Histoclinical long-standing follow-up study of
hyperplastic polyps of the stomach. Am J Gastroenterol 1981;75:27581.
49. Mizuno H, Kobayashi S, Kasugai T. Endoscopic follow-up of gastric polyps. Gastrointest
Endosc 1975;21:1125.
50. Eichenberger P, Hammer B, Gloor F, et al. Gardner's syndrome with glandular cysts of the
fundic mucosa. Endoscopy 1980;12:637.
51. Domizio P, Talbot IC, Spigelman AD, et al. Upper gastrointestinal pathology in familial
adenomatous polyposis: Results from a prospective study of 102 patients. J Clin Pathol
1990;43:73843.
52. Jarvinen HJ, Nyberg M, Peltokallio P. Upper gastrointestinal tract polyps in familial
adenomatosis coli. Gut 1983;24:3339.
53. Iida M, Yao T, Itoh H, et al. Natural history of fundic gland polyposis in patients with familial
adenomatosis coli/Gardner's syndrome. Gastroenterology 1985;89:10215.
54. Jagelman DG, Decosse JJ, Bussey HJR. Upper gastrointestinal cancer in familial
adenomatous polyposis. Lancet 1988;1:114951.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.

Jarvinen HJ, Sipponen P. Gastroduodenal polyps in familial adenomatous and juvenile
polyposis. Endoscopy 1986;18:2304.
Iida M, Yao T, Itoh H, et al. Natural history of gastric adenomas in patients with familial
adenomatosis coli/Gardner's syndrome. Cancer 1988;61:60511.
Iida M, Yao T, Watanabe H, et al. Spontaneous disappearance of fundic gland polyposis:
Report of three cases. Gastroenterology 1980;79:7258.
Tatsuta M, Okuda S, Tamura H, Taniguchi H. Gastric hamartomatous polyps in the absence
of familial polyposis coli. Cancer 1980;45:81823.
Iishi H, Tatsuta M, Okuda S. Clinicopathological features and natural history of gastric
hamartomatous polyps. Dig Dis Sci 1989;34:8904.
Kinoshita Y, Tojo M, Yano T, et al. Incidence of fundic gland polyps in patients without familial
adenomatous polyposis. Gastrointest Endosc 1993;39:1613.
Marcial MA, Villafana M, Hernandez-Denton J, Colon-Pagan JR. Fundic gland polyps:
Prevalence and clinicopathologic features. Am J Gastroenterol 1993;88:17113.
Hizawa K, Iida M, Matsumoto T, et al. Natural history of fundic gland polyposis without familial
adenomatosis coli: Follow-up observations in 31 patients. Radiology 1993;189:42932.
Bartholomew LG, Moore CE, Dahlin D, et al. Intestinal polyposis associated with
mucocutaneous pigmentation. Surg Gynecol Obstet 1962;115:111.
Utsunomiya J, Gocho H, Miyanaga T, et al. Peutz-Jeghers syndrome: Its natural course and
management. Johns Hopkins Med J 1975;136:7182.
Morson BC. Some peculiarities in the histology of intestinal polyps. Dis Colon Rectum
1962;5:33744.
Peart AG Jr, Sivak MV Jr, Rankin GB, et al. Spontaneous improvement of Cronkhite-Canada
syndrome in a postpartum female. Dig Dis Sci 1984;29:4704.
Daniel ES, Ludwig SL, Lewin KJ, et al. The Cronkhite-Canada syndrome: An analysis of
clinical and pathologic features and therapy in 55 patients. Medicine 1982;61:293309.
Johnson GK, Soergel KH, Hensley GT, et al. Cronkhite-Canada syndrome: Gastrointestinal
physiology and morphology. Gastroenterology 1972;63:14052.
Russell DM, Bhathal PS, St. John DJ. Complete remission in Cronkhite-Canada syndrome.
Cotterill JA, Day JL, Hughes JP, et al. The Cronkhite-Canada syndrome. Postgrad Med J
1973;49:26873.
Jarnum S, Jensen H. Diffuse gastrointestinal polyposis with ectodermal changes. A case with
severe malabsorption and enteric loss of plasma proteins and electrolytes. Gastroenterology
1966;50:10718.
Cronkhite LW Jr, Canada WJ. Generalized gastrointestinal polyposis: Unusual syndrome of
polyposis, pigmentation, alopecia, and onychotrophia. N Engl J Med 1955;252:10115.
Rrengol-Miro JR, Pedreira JD, Guardia J. Unusual cause of gastric bleeding:
Cronkhite-Canada polyposis. Endoscopy 1977;9:1779.
Salem OS, Steck WD. Cowden's disease (multiple hamartoma and neoplasia syndrome): A
case report and review of the English literature. J Am Acad Dermatol 1983;8:68696.
Starink TM. Cowden's disease: Analysis of fourteen new cases. J Am Acad Dermatol
1984;11:112741.
Hizawa K, Iida M, Matsumoto T, et al. Gastrointestinal manifestations of Cowden's disease.
Report of four cases. J Clin Gastroenterol 1994;18:138.
Binmoeller KF, Grimm H, Soehendra N. Endoscopic closure of a perforation using metallic
clips after snare excision of a gastric leiomyoma. Gastrointest Endosc 1993;39:1724.
Fujisaki J, Mine T, Akimoto K, et al. Enucleation of a gastric leiomyoma by a combined laser
and snare electrocutting technique. Gastrointest Endosc 1988;34:12830.
Skinner MS, Broadaway RK, Grossman P, Seckinger D. Multiple gastric lipomas. Dig Dis Sci
1983;28:11479.
81.
Jensen DM, Bown S. Gastrointestinal angioma: Diagnosis and treatment with laser therapy
and other endoscopic modalities. In Fleischer D, Jensen D, Bright-Asare P (eds). Therapeutic
Laser Endoscopy and Gastrointestinal Disease. Boston: Martinus Nij- hoff, 1983;15160.
82. Javdan P, Pitman ER, Schwartz IS. Gastric xanthelasma: Endoscopic recognition.
83. Kunze KC, Baum RA, Nasrallah SM. Gastric xanthoma. Gastrointest Endosc 1987;33:1145.
84. Kimura D, Hiramoto T, Buncher CR. Gastric xanthelasma. Arch Pathol 1969;87:1107.
85.
Kameya S, Nakamura S, Mizutani K, et al. Xanthoma in the stomach. Gastrointest Endosc
1963;5:3741.
86.
Aste H, Cheli R, Nicolo G, Santi L. Gastric xanthelasma. Acta Endosc 1978;8:14.
87. Mast A, Eelwaut A, Mortier G, et al. Gastric xanthoma. Am J Gastroenterol 1976;65:3117.
88. Moreto M, Ojembarrena E, Zaballa M, et al. Retrospective endoscopic analysis of gastric
xanthelasma in the non-operated stomach. Endoscopy 1985;17:2101.
89. Terruzzi V, Minoli G, Butti GC, Rossini A. Gastric lipid islands in the gastric stump and in
non-operated stomach. Endoscopy 1980;12:5862.
90. Domellof L, Eriksson S, Helander HF, Janunger KG. Lipid islands in the gastric mucosa after
resection for benign ulcer disease. Gastroenterology 1977;72:148.
92.
Appelman HD. Localized and extensive expansions of the gastric mucosa: Mucosal polyps
and giant folds. In Appelman HD (ed). Pathology of the Esophagus, Stomach and Duodenum.
Contemporary Issues in Surgical Pathology. New York: Churchill Livingstone, 1984;79119.
93. Coma del Corral MJ, Pardo-Mindan FJ, Razquin S, Ojeda C. Risk of cancer in patients with
gastric dysplasia. Follow-up study of 67 patients. Cancer 1990;65:207885.
94. Farinati F, Rugge M, Di Mario F, et al. Early and advanced gastric cancer in the follow-up of
moderate and severe gastric dysplasia patients. A prospective study.
I.G.G.E.D.Interdisciplinary Group on Gastric Epithelial Dysplasia. Endoscopy
1993;25:2614.
95. Di Gregorio C, Morandi P, Fante R, De Gaetani C. Gastric dysplasia. A follow-up study. Am J
96. Raneko E, Nakamura T, Umeda N, et al. Outcome of gastric carcinoma detected by gastric
mass survey in Japan. Gut 1977;18:62630.
97.
Iwatsuki S. Type of gastric carcinoma detected by mass survey. Gastric Cancer 1972;23:55.
98. Cohen MM, Zoeter MA, Loar C. Survival following surgical treatment of gastric cancer. A
challenge for the community endoscopist. Surg Endosc 1994;8:8626.
99. Evans DM, Craven JL, Murphy F, Cleary BK. Comparison of "|rly gastric cancer" in Britain and
Japan. Gut 1978;19:19.
100. Longo WE, Zucker KA, Zdon MJ, Modlin IM. Detection of early gastric cancer in an
aggressive endoscopy unit. Am Surg 1989;55:1004.
101. Paulino F, Roselli A. Early gastric cancer: Report of twenty-five cases. Surgery
1979;85:1716.
102. Belcastro G, Ferri M, Chiavellati L, et al. Early gastric cancer. Ital J Surg Sci 1988;18:22732.
103. Newbold KM, Thompson H, Dykes PW. The effect of routine endoscopy on the detection rate
of T1 gastric cancer (early gastric cancer) in Birmingham. Endoscopy 1989;21:569.
104.
Fielding JWL, Ellis DJ, Jones BG, et al. Natural history of "|rly" gastric cancer: Results of a
10-year regional survey. BMJ 1980;281:9657.
105. Green PH, O'Toole KM, Weinberg LM, Goldfarb JP. Early gastric cancer. Gastroenterology
1981;81:24756.
106. Moreaux J, Bougaran J. Early gastric cancer: A 25-year experience. Ann Surg
1993;217:34755.
107. Lawrence M, Shiu MH. Early gastric cancer: 28-year experience. Ann Surg 1991;213:32734.
108. Green PH, O'Toole KM, Slonim D, et al. Increasing incidence and excellent survival of
patients with early gastric cancer: Experience in a United States medical center. Am J Med
1988;85:65861.
109. Chia MM, Langman JM, Hecker R, et al. Early gastric cancer: 52 cases of combined
experience of two south Australian teaching hospitals. Pathology 1988;20:21626.
110. Oleagoitia JM, Echevarria A, Santidrian JI, et al. Early gastric cancer. Br J Surg
1986;73:8046.
111.
Hirota T, Ming SC. Early gastric cancer. In Ming SC, Goldman H (eds). Pathology of the
Gastrointestinal Tract. Philadelphia: WB Saunders, 1992;570582.
112. Japanese Research Society for Gastric Cancer. The general rules for the gastric cancer study
in surgery and pathology. Jpn J Surg 1981;11:12745.
113. Misumi A, Misumi K, Murakami A, et al. Endoscopic diagnosis of minute, small, and flat early
gastric cancers. Endoscopy 1989;21:15964.
114. Mori M, Adachi Y, Kakeji Y, et al. Superficial flat-type early carcinoma of the stomach. Cancer
1992;69:30613.
115. Gentsch HH, Groitl H, Giedl J. Results of surgical treatment of early gastric cancer in 113
patients. World J Surg 1981;5:1037.
116. Brandt D, Muramatsu Y, Ushio K, et al. Synchronous early gastric cancer. Radiology
1989;173:64952.
117. Matsuura H, Morita M, Tsutsui S, et al. Esophageal carcinoma is frequently accompanied with
early gastric carcinoma. Int Surg 1992;77:6971.
118. Sakita T, Oguro Y, Takasu S, et al. Observations on the healing of ulcerations in early gastric
cancer. The life cycle of the malignant ulcer. Gastroenterology 1971;60:8359 passim.
119. Sano T, Okuyama Y, Kobori O, et al. Early gastric cancer. Endoscopic diagnosis of depth of
invasion. Dig Dis Sci 1990;35:13404.
120. Iishi H, Tatsuta M, Okuda S. Diagnosis of simultaneous multiple gastric cancers by the
endoscopic Congo red-methylene blue test. Endoscopy 1988;20:7882.
121.
Shirakabe H, Ichikawa H, Rumakura R, et al. Atlas of X-Ray Diagnosis of Early Gastric
Cancer. Tokyo: Igaku Shoin. 1966.
122. Craanen ME, Dekker W, Ferwerda J, et al. Early gastric cancer: A clinicopathologic study. J
123.
Fujita S. Natural history of human gastric carcinoma in terms of their genesis and
progression. Asian Med J 1983;26:787805.
124. Kawaura Y, Kanehira E, Ohta Y, Nakano I. Predicting histologic findings from endoscopic
appearance of early gastric cancer. J Clin Gastroenterol 1991;13:51720.
125. Tsukuma H, Mishima T, Oshima A. Prospective study of "|rly" gastric cancer. Int J Cancer
1983;31:4216.
126. Fukutomi H, Sakita T. Analysis of early gastric cancer cases collected from major hospitals
and institutes in Japan. Jpn J Clin Oncol 1984;14:16979.
127. Nakamura K, Ueyama T, Yao T, et al. Pathology and prognosis of gastric carcinoma: Findings
in 10,000 patients who underwent primary gastrectomy. Cancer 1992;70:10307.
128. Maehara Y, Okuyama T, Oshiro T, et al. Early carcinoma of the stomach. Surg Gynecol
Obstet 1993;177:5937.
129. Sano T, Kobori O, Muto T. Lymph node metastasis from early gastric cancer: Endoscopic
resection of tumour. Br J Surg 1992;79:2414.
130.
Borrmann R. Geshwelste des Magens und Duodenums. In Henke F, Lubarsch O (eds).
Handbuch der Spezieller Patholgischen Anatomie und Histologie. Vol. 4.<\f>Berlin:
Springer-Verlag,. 8651926.
131. Podolsky I, Storms PR, Richardson CT, et al. Gastric adenocarcinoma masquerading
endoscopically as benign gastric ulcer. A five-year experience. Dig Dis Sci 1988;33:105763.
132. Chan TO, Kuo YC, Chien RN, et al. Gastric adenocarcinoma simulating benign gastric ulcer.
Chang Keng I Hsueh 1992;15:5963.
133. Erickson RA. Impact of endoscopy on mortality from occult cancer in radiographically benign
gastric ulcers. A probability analysis model. Gastroenterology 1987;93:83545.
all gastric ulcers?. Dig Dis Sci 1993;38:2848.

135. Bytzer P. Endoscopic follow-up study of gastric ulcer to detect malignancy: Is it worthwhile?.
136.
Ming SC. Adenocarcinoma and other malignant epithelial tumors of the stomach. In Ming SC,
Goldman H (eds). Pathology of the Gastrointestinal Tract. Philadelphia: WB Saunders,
1992;596.
137. Awad JA, Lee EY, Wang TC, et al. Effect of mucosal thickening near gastric carcinoma on the
endoscopic diagnosis of malignancy. Dig Dis Sci 1990;35:31720.
139. Boddie AW, McMurtry MJ, Glacco GG, et al. Palliative total gastrectomy and
esophagogastrectomy: A reevaluation. Cancer 1983;51:11959.
140.
Lawrence W Jr. Gastric cancer. CA: Cancer J Clin 1986;36:21624.
141. Douglass HO, Nava HR. Gastric adenocarcinomaManagement of the primary disease.
Semin Oncol 1985;12:3240.
142. Boddie AW Jr, McBride CM, Balch CM. Gastric cancers. Am J Surg 1989;157:595609.
143. Heberer G, Teichmann RK, Kramling HJ, et al. Results of gastric resection for carcinoma of
the stomach: The European experience. World J Surg 1988;12:37481.
144. Yap P, Pantangco E, Yap A, et al. Surgical management of gastric carcinoma: Follow-up
results in 465 consecutive cases. Am J Surg 1982;143:2847.
145. Meyers WC, Damiano RJ Jr, Rotolo FS, et al. Adenocarcinoma of the stomach: Changing
patterns over the last four decades. Ann Surg 1987;205:18.
146. Soga J, Kobayashi K, Saito J, et al. The role of lymphadenectomy in curative surgery for
gastric cancer. World J Surg 1979;3:7018.
147. Soga J, Ohyama S, Myashita K, et al. A statistical evaluation of advancement in gastric
cancer surgery with special reference to the significance of lymphadenectomy for cure. World
J Surg 1988;12:398405.
148.
MacDonald JS, Gohmann JJ. Chemotherapy of advanced gastric cancer: Present status,
future prospects. Semin Oncol 1988;253:20618.
149. Moertel C, Rubin J, O'Connel MJ, et al. A phase II study of combined 5-FU, doxorubicin and
cisplatinum in the treatment of advanced upper gastrointestinal adenocarcinomas. J Clin
Oncol 1986;4:10539.
150. The Italian Gastrointestinal Tumor Study Group. Adjuvant treatment following curative
resection for gastric cancer. Br J Surg 1988;75:11004.
151.
Sjostedt S, Pieper R. Gastric cancer: Factors influencing long term survival and postoperative
mortality. Acta Chir Scand 1986;530(suppl):259.
152. Moreaux J, Miska S. Carcinoma of the gastric cardia: Surgical management and long-term
survival. World J Surg 1988;12:22935.
153. Antonioli DA, Goldman H. Changes in the location and the type of gastric adenocarcinoma.
Cancer 1982;50:77581.
154. Shennib H, Lough J, Klein HW, et al. Gastric carcinoma: Intestinal metaplasia and tumor
growth patterns as indicators of prognosis. Surgery 1986;100:77480.
155. Kaibara N, Iitsuka Y, Kimura A, et al. Relationship between area of serosal invasion and
prognosis in patients with gastric carcinoma. Cancer 1987;60:1369.
156. Watanabe H, Enjoji M, Imai T. Gastric carcinoma with lymphoid stroma: Its morphological
characteristics and prognostic correlation. Cancer 1976;38:23243.
157. Taal BG, Burgers JM. Primary non-Hodgkin's lymphoma of the stomach: Endoscopic
diagnosis and the role of surgery. Scand J Gastroenterol Suppl 1991;188:337.
158. al Mofleh IA. Endoscopic features of primary upper gastrointestinal lymphoma. J Clin
Gastroenterol 1994;19:6973; discussion 734.
159. Swaroop VS, Mohandas KM, Swaroop VD, et al. Comparative endoscopic study of primary
gastric lymphoma vs. gastric carcinoma. J Surg Oncol 1994;56:947.
160. Seifert E, Schulte F, Weismuller J, et al. Endoscopic and bioptic diagnosis of malignant
non-Hodgkin's lymphoma of the stomach. Endoscopy 1993;25:497501.

161. Kaufman JC, Fierst SM, Novetsky AD, Kim DS. Plasmacytoma of the stomach: Endoscopic
diagnosis. Gastrointest Endosc 1984;30:1867.
162. Sloyer A, Katz S, Javors FA, Kahn E. Gastric involvement with excavated plasmacytoma:
Case report and review of endoscopic criteria. Endoscopy 1988;20:2679.
163. Arista-Nasr J, Jimenez A, Keirns C, et al. The role of the endoscopic biopsy in the diagnosis
of gastric lymphoma: A morphologic and immunohistochemical reappraisal. Hum Pathol
1991;22:33948.
164. Suekane H, Iida M, Kuwano Y, et al. Diagnosis of primary early gastric lymphoma. Usefulness
of endoscopic mucosal resection for histologic evaluation. Cancer 1993;71:120713.
165. Fend F, Schwaiger A, Weyrer K, et al. Early diagnosis of gastric lymphoma: Gene
rearrangement analysis of endoscopic biopsy samples. Leukemia 1994;8:359.
166. The Non-Hodgkin's Lymphoma Classification Project. National Cancer Institute-sponsored
study of classifications of non-Hodgkin's lymphomas: Summary and description of working
formulation for clinical usage. Cancer 1982;49:211235.
167. Haber DA, Mayer RJ. Primary gastrointestinal lymphoma. Semin Oncol 1988;15:15461.
168. Shimm DS, Dosoretz DE, Anderson T, et al. Primary gastric lymphoma: An analysis with
emphasis on prognostic factors and radiation therapy. Cancer 1983;52:20448.
169. Lewin KJ, Ranchod M, Dorfman RF. Lymphomas of the gastrointestinal tract. A study of 117
cases presenting with gastrointestinal disease. Cancer 1978;42:693707.
170.
Creutzfeldt W. The achlorhydria-carcinoid sequence: Role of gastrin. Digestion
1988;26:124854.
171. Borch K, Renvall H, Liedberg G. Gastric endocrine cell hyperplasia and carcinoid tumors in
pernicious anemia. Gastroenterology 1985;88:63848.
172. Sjoblom SM, Sipponen P, Miettinen M, et al. Gastroscopic screening for gastric carcinoids
and carcinoma in pernicious anemia. Endoscopy 1988;20:526.
173. Hirschowitz BI, Griffith J, Pellegrin D, Cummings OW. Rapid regression of
enterochromaffin-like cell gastric carcinoids in pernicious anemia after antrectomy.
174. Kornfeld HR, Sherlock P, Hertz R. Multiple carcinoid tumors of the stomach with endoscopic
observations. Gastrointest Endosc 1971;18:747.
175. Gueller R, Haddad JK. Gastric carcinoids simulating benign polyps. Two cases diagnosed by
endoscopic biopsy. Gastrointest Endosc 1975;21:1535.
176. DeLuca RF, Ferrer JP, Gambescia RA, Raskin JB. Gastric carcinoid endoscopically
simulating leiomyoma. Am J Gastroenterol 1978;70:1636.
177. Nakamura S, Iida M, Yao T, Fujishima M. Endoscopic features of gastric carcinoids.
178. Abrams JS. Multiple malignant carcinoids of the stomach. Arch Surg 1980;115:121921.
179.
Christodoulopoulos JB, Klotz AP. Carcinoid syndrome with primary carcinoid tumor of the
stomach. Gastroenterology 1961;40:42940.
180.
Oates JA, Sjoerdsma A. A unique syndrome associated with secretion of 5-hydroxytryptophan
by metastatic gastric carcinoids. Am J Med 1962;32:33342.
181. Giordano JA. Gastric carcinoid syndrome diagnosed by endoscopy. Am J Gastroenterol
1981;76:413.
182. Morise K, Kusugami K, Hayakawa M, et al. Minute carcinoid tumor of the stomach: Report of
two cases and review of the Japanese literature. Gastroenterol Jpn 1985;20:596603.
183. Moertel CG. Treatment of the carcinoid tumor and the malignant carcinoid syndrome. J Clin
Oncol 1983;1:72781.
184. Borsch G, Schmidt G, Sperling J, et al. Gastroduodenal metastasesAn unusual initial
manifestation of bronchogenic carcinoma. Endoscopy 1984;16:11821.
185. Green LK. Hematogenous metastases to the stomach. A review of 67 cases. Cancer
1990;65:1596600.
186. Catino A, Lorusso V, Gargano G, et al. Metastatic involvement of the stomach secondary to
breast carcinoma. A case report. Eur J Gynaecol Oncol 1992;13(suppl):858.

187. Dupas JL, Delamarre J, Marti R, et al. Gastric metastasis from breast carcinoma: Emergency
snare-ectomy for control of hemorrhage. Endoscopy 1976;8:357.
188. Morini S, Bassi O, Colavolpe V. Malignant melanoma metastatic to the stomach: Endoscopic
diagnosis and findings. Endoscopy 1980;12:869.
189. Richter R, Panish J, Berci G. Endoscopic findings in melanoma metastatic to the stomach.
190. Shah SM, Smart DF, Texter EC Jr, Morris WD. Metastatic melanoma of the stomach: The
endoscopic and roentgenographic findings and review of the literature. South Med J
1977;70:37981.
191. Eskreis D, Zinberg J, Manzione NC, Jones J. Metastatic choriocarcinoma to the stomach
presenting as hematemesis. Dig Dis Sci 1988;33:24750.
192. Holderman WH, Jacques JM, Blackstone MO, Brasitus TA. Prostate cancer metastatic to the
stomach. Clinical aspects and endoscopic diagnosis. J Clin Gastroenterol 1992;14:2514.
193. Kibsgaard K. Gastric metastases from muco-epidermoid parotid tumor: Endoscopic
diagnosis. Gastrointest Endosc 1979;25:1067.
194. Konishi H, Isetani K, Satoh T, et al. A case of metastatic chondrosarcoma of the stomach. J
195. Feingold ML, Goldstein MJ, Lieberman PH. Multiple myeloma involving the stomach. Report
of a case with gastroscopic observations. Gastrointest Endosc 1969;16:10710.
196. Snyder N III, Villar H, Patterson M, Hughes WS. Chronic myelogenous leukemia presenting
as gastric polyps: Diagnosis by endoscopic polypectomy. Gastrointest Endosc
1974;20:1667.
197. Bynum TE. Gastroscopic appearance of multiple gastric ulcers associated with lymphatic
leukemia. Gastrointest Endosc 1970;17:3840.
198. Nelson RS, Lanza F. Malignant melanoma metastatic to the upper gastrointestinal tract.
199.
Choi SH, Sheehan FR, Pickren JW. Metastatic involvement of the stomach by breast cancer.
Cancer 1964;17:7917.
200. Friedman SL, Wright TL, Altman DF. Kaposi's sarcoma in patients with acquired
immunodeficiency syndrome: Endoscopic and autopsy findings. Gastroenterology
1985;89:1028.
201. Ahmed N, Nelson RS, Goldstein HM, et al. Kaposi's sarcoma of the stomach and duodenum:
Endoscopic and roentgenologic correlations. Gastrointest Endosc 1975;21:14952.
202.
Barrison IG, Foster S, Harris JW, et al. Upper gastrointestinal Kaposi's sarcoma in patients
positive for HIV antibody without cutaneous disease. BMJ (Clin Res Ed) 1988;296:923.
203. Boyd GD, Shabot JM, Pollard RB, Gourley WK. Endoscopic evolution of gastric Kaposi's
sarcoma. Gastrointest Endosc 1987;33:4503.
204. Papazian A, Gineston JL, Capron JP, Quenum C. Leiomyoblastoma of the stomach:
Endoscopic treatment. Endoscopy 1984;16:1579.
205. Shiu MH, Farr GH, Papachristou DN, Hajdu SI. Myosarcomas of the stomach: Natural history,
prognostic factors and management. Cancer 1982;49:17787.
206. Grant CS, Rim CH, Farrugia G, et al. Gastric leiomyosarcoma: Prognostic factors and surgical
management. Arch Surg 1991;126:98590.
207. Tanimura H, Furuta M. Carcinosarcoma of the stomach. Am J Surg 1967;113:7029.
208. Appleman HD, Hjelwig EB. Sarcomas of the stomach. Am J Clin Pathol 1977;67:210.
209. Asaki S. Endoscopical polypectomy using high frequency current: Double-snare method of
polypectomy for prevention of incidental bleeding and perforation. Tohoku J Exp Med
1982;136:2158.
210. Brandt L, Frankel A, Waye JD. Endoscopic nonoperative gastric polypectomy. Am J Dig Dis
1973;18:108790.
211. Chua CL. Gastric polyps: The case for polypectomy and endoscopic surveillance. J R Coll
Surg Edinb 1990;35:1635.
212. Hughes RW Jr. Gastric polyps and polypectomy: Rationale, technique, and complications.
213. Seifert E, Gail K, Weismuller J. Gastric polypectomy. Long-term results (survey of 23 centres
in Germany). Endoscopy 1983;15:811.
214. Takemoto T. Laser therapy of early gastric carcinoma. Endoscopy 1986;18(suppl 1):326.
215. Hiki Y, Sakakibara Y, Mieno H, et al. Endoscopic treatment of gastric cancer. Surg Endosc
1991;5:113.
217. Watanabe M, Sugino Y, Imai Y, et al. Reevaluation of endoscopic laser therapy for treatment
of early cancer in the stomach. Keio J Med 1993;42:2068.
218. Hiki Y, Shimao J, Yamao Y, et al. The concepts, procedures, and problems related in
endoscopic laser therapy of early gastric cancer. A retrospective study on early gastric
cancer. Surg Endosc 1989;3:16.
219. Hirao M, Masuda K, Asanuma T, et al. Endoscopic resection of early gastric cancer and other
tumors with local injection of hypertonic saline-epinephrine. Gastrointest Endosc
1988;34:2649.
220. Fujimori T, Nakamura T, Hirayama D, et al. Endoscopic mucosectomy for early gastric cancer
using modified strip biopsy. Endoscopy 1992;24:1879.
221. Takechi K, Mihara M, Saito Y, et al. A modified technique for endoscopic mucosal resection of
small early gastric carcinomas. Endoscopy 1992;24:2157.
222. Karita M, Tada M, Okita K. The successive strip biopsy partial resection technique for large
early gastric and colon cancers. Gastrointest Endosc 1992;38:1748.
223. Haruma K, Sumii K, Inoue K, et al. Endoscopic therapy in patients with inoperable early
gastric cancer. Am J Gastroenterol 1990;85:5226.
224. Imaoka W, Ida K, Katoh T, et al. Is curative endoscopic treatment of early gastric cancer
possible?. Endoscopy 1987;19(suppl 1):711.
225. Tada M, Murakami A, Karita M, et al. Endoscopic resection of early gastric cancer.
Endoscopy 1993;25:44550.
226. Spinelli P, Cerrai FG, Cambareri AR, et al. Two-step endoscopic resection of gastric
leiomyomas. Surg Endosc 1993;7:902.
227. Nakamura S, Iida M, Suekane H, et al. Endoscopic removal of gastric lipoma: Diagnostic
value of endoscopic ultrasonography. Am J Gastroenterol 1991;86:61921.
228. Goh PMY, Alponat A, Kum CK. Early international results of laparoscopic gastrectomies. Surg
Endosc 1997;11:6502.
229. Duh QY, Way LW. Laparoscopic jejunostomy using T-fasteners as retractors and anchors.
Arch Surg 1993;128:1058.
Chapter 49 Diseases of the Duodenum

(4492)
BERNARD M. SCHUMAN, M.D.
Diseases of the duodenum, other than bulbar peptic ulcer, are uncommon and may be overlooked at
endoscopy. Moreover, the major duodenal papillathe most interesting duodenal structureis
ordinarily out of view or can be seen only tangentially with forward-viewing endoscopes and, thus,
cannot be adequately studied (see also Chapter 38: Technique of Upper Gastrointestinal Endoscopy).
It is also unlikely that more than half of the linear extent of the duodenum is ever visualized by routine
endoscopic examination, and what is confidently declared a normal duodenoscopy is in reality a normal
semiduodenoscopy. Not only is the duodenum a repository for peptic ulcer but it may also harbor a
wide variety of diseases, both primary and secondary. Inflammatory, neoplastic, or infectious disorders
are readily identifiable endoscopically, and some may be definitively diagnosed by endoscopic
biopsy.1(4493)
Inflammatory Diseases
Crohn's Disease
Crohn's disease of the duodenum is uncommon, occurring in only 2% of patients with this form of
inflammatory bowel disease.2,3(4494) More than 200 cases have been reported,2(4495) most with
involvement elsewhere in the digestive tract. Rutgeerts et al.4(4496) estimated that about 3% of their
patients with Crohn's disease had gastroduodenal involvement. Almost all patients with duodenal
Crohn's disease have had epigastric pain as the major symptom. The diagnosis is usually made
radiologically in the course of a small bowel x-ray series. Endoscopy is performed to confirm the
diagnosis and to obtain biopsies.
The endoscopic features of duodenal Crohn's disease were described in a study of 14 patients, 3 of
whom had involvement of the duodenum alone.5(4497) The characteristic findings were similar to
those of Crohn's disease elsewhere in the intestine. The mucosal surface is granular and often
nodular. Erosions and aphthoid ulcers are frequent, but the long, deep linear ulcers with a white base
and narrow red margin are seen less often (Figure 491). The wall of the bowel lacks distensibility, and
contractions are not evident because of the stiffened folds. Indeed, mural inflammation commonly
results in stenosis of the proximal duodenum that is sufficient to prevent passage of the endoscope
(Figure 492).3(4498) By the time characteristic findings of Crohn's duodenitis are present, stenosis
may already be a major feature. When there is associated gastric disease, which usually involves the
antrum, duodenoscopy may be impossible because of antral and pyloric stenoses. Hydrostatic balloon
dilation of a gastroduodenal stricture has been reported.6(4499) Although this appeared to provide
temporary relief of the obstruction, antral bypass surgery became necessary 1 year later.
(4500)Figure 491. Crohn's ulcer of duodenum. The narrowed duodenal lumen is in the lower
right of the field.
(4501)Figure 492. Severe stricture of the descending duodenum in a patient with duodenal
Crohn's disease. (From the collection of Dr. Michael V. Sivak, Jr.)
A rare complication of Crohn's duodenitis is the formation of an enterocutaneous fistula. Of two such
reported cases, the fistula was identified by duodenoscopy in one, although mucosal biopsies
disclosed only nonspecific inflammation.5(4502) Fistula formation to the biliary system has also been
described.4(4503) Endoscopic evaluation of duodenum involved with enterocolonic fistulas is essential
to determine whether the duodenal process is secondary to contiguous colonic granulomatous
inflammation or a manifestation of primary Crohn's involvement.
Endoscopic duodenal biopsies in Crohn's disease have been disappointing for the most part. Nugent et
al.3(4504) did not encounter any granulomas in endoscopic biopsies from the original 17 patients in
their series, although granulomas were found in 3 patients who had capsule biopsies. This suggests
that larger and perhaps deeper endoscopic biopsies will have a higher yield of granulomas. This was
verified in their further experience (recounted in an addendum to their paper) with 8 additional patients
with duodenal Crohn's disease, in 4 of whom granulomas were demonstrated in biopsies obtained,
presumably, with endoscopes containing larger accessory channels. In a later report by Nugent and
Roy,7(4505) endoscopic biopsies revealed granulomas or "granulomatous inflammation" in only 8
patients (15%). In this series of 89 patients, which was compiled over more than 30 years, Nugent and
Roy observed that gastroduodenal Crohn's disease was responsible for comparatively less patient
morbidity than the presence of the disease in more distal sites. Surgery for gastroduodenal disease
was performed in 33 patients, usually because of obstructive symptoms. Higher yields for granulomas
in endoscopic forceps biopsies have been noted in earlier studies. Gad8(4506) reported a yield of
33.3% in patients with gastroduodenal Crohn's disease, and Rutgeerts et al.4(4507) reported a yield of
68% (15 of 22 patients). In all likelihood, the lower figures reported by Nugent and Roy are more
representative of the general experience.
Eosinophilic Gastroenteritis
Marked eosinophilic infiltration of the wall of the stomach and small intestine is the hallmark of
eosinophilic gastroenteritis. Patients with this uncommon disorder will often have a history of allergic
disorders, and marked peripheral eosinophilia is invariably present.9(4508) Although the stomach is
primarily involved, in one disease pattern there is multifocal involvement of the duodenum and jejunum
in addition to the extensive submucosal infiltration of the antrum. Endoscopically, the pattern of
mucosal folds is accentuated and the mucosa hyperemic, but these findings are usually marked in the
antrum and less obvious in the duodenum, where there may be only focal erythema and erosion.
Another pattern in adultsnot distinguishable endoscopically from the first pattern, although
characterized by both mucosal and submucosal eosinophilic infiltrationlends itself more readily to
definitive diagnosis by endoscopic biopsy, especially in the stomach but also in the duodenum and
jejunum.10(4509) Dense collections of eosinophilsin effect, eosinophilic microabscessescan be
found by microscopic study of biopsy material, although the absence of this finding does not exclude
the disease. In fact, parasitic infection, radiation enteritis, lymphoma, and polyarteritis nodosa must be
eliminated as diagnostic possibilities when biopsies show nonspecific inflammation.9(4510)
Pancreatitis
The question of whether alcohol has a noxious effect on duodenal mucosa remains unanswered. Lev
et al.11(4511) could not correlate histologic change with alcohol consumption. However, in a group of
alcoholic patients with chronic pancreatitis, endoscopic biopsy of the duodenal bulb showed an
increased frequency of isolated duodenitis compared with that in patients with chronic alcoholism or
liver cirrhosis alone.12(4512) Duodenal mucosal abnormalities appear to be associated with the
inflammatory process of both acute and chronic pancreatitis, although mucosal alterations may not be
evident endoscopically.
The inflammatory process may be progressive, so that actual stenosis of the second or third portion of
the duodenum ensues. This finding on upper gastrointestinal series should not preclude endoscopic
evaluation, even when stricturing is apparently severe, because duodenoscopy is nevertheless
successful in most such cases.13(4513) However, this usually yields meager findings other than
confirmation of the narrowed caliber because in most cases the mucosa is intact and normal in
appearance,14(4514) although occasionally erythema and friability will be present. Failure to identify
the major duodenal papilla is not unusual, since it may be retracted and not visible even with the
side-viewing duodenoscope.15(4515) Although the stenotic segment may be long, the stricture usually
resolves as the contiguous inflammatory process subsides. A small percentage of patients will develop
a fibrosing pancreaticoduodenitis and will ultimately require surgical bypass.14(4516) Considerable
rigidity of the stenotic area may be appreciated by palpation with the endoscope, but it cannot be
assumed that mural fibrosis is the cause because a similar abnormality may occur with a pseudocyst
or pancreatic cancer. Before surgical intervention, duodenos-copy should be complemented by a
thorough evaluation that may include ultrasonography, computed tomography (CT), and endoscopic
retrograde cholangiopancreatography (ERCP).
Infectious Diseases
Mycobacteria
Tuberculosis
Duodenal tuberculosis, when it is found in conjunction with ileocecal involvement, is frequently
mistaken for Crohn's disease in the United States, where gastrointestinal tuberculosis is rare.16(4517)
In Africa, Asia, and India, where intestinal tuberculosis is common,17(4518) duodenal tuberculosis is
still infrequently encountered and constitutes only 3% of gastrointestinal tuberculosis. Because only
50% of patients with duodenal tuberculosis have involvement elsewhere in the digestive tract or other
viscera, diagnosis is difficult, and laparotomy and surgical biopsy are necessary for histologic
confirmation.
Duodenoscopy is undertaken to explain findings on upper gastrointestinal series of thickened folds,
ulcerations, and stricture consistent with Crohn's disease, peptic ulcer, or neoplasm. In one report
specifically related to duodenoscopic evaluation, the finding was a stricture 2 to 3 cm long, surrounded
by hyperemic and granular mucosa.17(4519) Six endoscopic biopsies disclosed nonspecific chronic
inflammation and fibrosis, but caseating granulomas were seen in subsequent surgical material. In
another patient, marked bulbar duodenitis and a postbulbar stricture were found, but again biopsy
specimens were not diagnostic and surgery was required to obtain histologic verification.16(4520)
Pyloric obstruction may interfere with endoscopic evaluation of the duodenum.18(4521) Endoscopic
biopsy was also disappointingly negative in the evaluation of an irregular ulcer 2 cm in widest diameter
with a necrotic base in the distal transverse duodenum; again, surgical material from the duodenum
contained granulomas in the submucosa and serosa.19(4522)
Hydrostatic balloon dilation of tuberculous strictures of the duodenum has been reported in two
patients.20(4523) Both patients, who also underwent treatment with multiple antituberculous agents,
were asymptomatic at 8 and 10 months follow-up.
Infection with Mycobacterium Avium-Intracellulare
There is an increasing number of reports of patients with acquired immunodeficiency syndrome (AIDS)
and infection of the duodenum by Mycobacterium avium-intracellulare.2124(4524) The endoscopic
appearance resembles the mucosal change seen in Whipple's disease (see later in this chapter) rather
than the florid inflammatory reaction usually seen with Mycobacterium tuberculosis. The main finding is
the presence of multiple white or yellow nodules from 2 to 4 mm in diameter. Intervening mucosa
between the nodules, which may become confluent, is erythematous, and there may be hemorrhagic
erosions. Histopathologically, the nodules are composed of macrophages within the lamina propria and
numerous intracellular and extracellular acid-fast organisms.
Parasites
Although in many countries the prevalence of parasitic infection is low, the endoscopist should
consider the presence of parasitic infection, which is often amenable to therapy, when unexplained
mucosal changes are encountered in the duodenum.
A worm may rarely be found in the duodenum.21(4525) The worm should be removed with the biopsy
forceps for specific identification. It is more likely that only the nonspecific duodenal mucosal
inflammation that is a reaction to the parasite or its eggs will be seen, but the endoscopist should be
alert to the possibility of a parasitosis, in order that aspiration of luminal fluid and mucosal biopsies may
be appropriately performed.
Giardia Lamblia
Giardia lamblia is a protozoan of cosmopolitan distribution that is disseminated by contaminated water

and fecal-oral spread. The diagnosis is ordinarily confirmed by finding the characteristic flagellate in
diarrheal stool specimens. However, the prominence of upper gastrointestinal symptoms such as
anorexia and nausea as well as loss of weight may initiate radiologic and endoscopic study of the
stomach and duodenum. At duodenoscopy, the valvulae conniventes may be thickened, although the
mucosa itself may appear relatively normal.25(4526) In a patient with hypogammaglobulinemia,
however, the mucosa may show numerous punctate nodules, a manifestation of lymphoid nodular
hyperplasia. If giardiasis is suspected, the duodenal luminal contents should be aspirated to examine
the fluid for the trophozoite. Endoscopic biopsy should have a high positive return, based on the
excellent results achieved with capsule biopsy.25(4527) Giardiasis is not a condition that requires
endoscopy for diagnosis, but if the opportunity arises incidentally during evaluation of a patient, it
should not be lost. When there is no suspicion of giardiasis, however, the routine aspiration of luminal
contents is unwarranted because the diagnostic yield is low.26,27(4528)
Worms
Strongyloides stercoralis infests the duodenum and may cause a myriad of symptoms, the most
common being epigastric pain. There is a report of a patient with severe recurrent upper
gastrointestinal hemorrhage due to massive infestation by S. stercoralis in which the diagnosis was
made only by endoscopic biopsy and aspiration of duodenal contents.28(4529) Upper gastrointestinal
roentgenography may show minimal mucosal changes or, with severe involvement occurring in
endemic areas or in immunosuppressed patients, thickened duodenal folds and mural rigidity.29(4530)
These latter findings usually require endoscopy for definitive diagnosis, but the endoscopic appearance
varies from "marked edema" in mild infections30(4531) to gross inflammatory hypertrophy of the
mucosa, which in advanced disease nearly obliterates the lumen.29(4532) Strongyloidiasis must be
first suspected if it is to be diagnosed endoscopically. The diagnosis depends on aspiration of luminal
mucus and juice, which will contain rhabditiform larvae of S. stercoralis, and on duodenal biopsy, which
results in a high yield for the parasite buried in the mucosa.28(4533)
A small luminal helminth such as hookworm may be aspirated via the suction channel of an endoscope
into a trap, or it may be gently picked off the mucosa with a biopsy forceps.21(4534) Identification of
any rhabditiform or filariform larvae obtained at biopsy, however, requires a parasitologist to
differentiate hookworm from S. stercoralis and Trichostrongylus spp.30(4535) The large helminth
Ascaris lumbricoides presents no difficulty in recognition because of its characteristic gross external
features. A heavy infection of A. lumbricoides may result in migration of the worm into the ampulla,
from which it may be extracted if it is visible.31(4536) Pancreatitis has been documented as a
complication of ductal invasion by the roundworm.32(4537)
A bizarre encounter with a tapeworm in the duodenum led to extraction of the 250-cm worm with its
head intact.33(4538) There is a similar case of removal of a beef tapeworm (Taenia saginata) from the
duodenal bulb of a 72-year-old woman with epigastric pain.34(4539) Because of increased immigration
of people from Asia, endoscopists in the United States should be prepared to recognize such
trematodes as Fasciola hepatica and Clonorchis sinensis.35(4540)
Tumors
Tumors, benign and malignant, that arise primarily in the main duodenal papilla are discussed in
Chapter 69: Tumors of the Main Duodenal Papilla. Tumors that arise in all other areas of the denum
are considered in the following sections.
Benign Neoplasms
The duodenum constitutes only 8% of the small bowel by length but harbors 10 to 20% of small bowel
tumors.36(4541) Because only 3 to 6% of all gastrointestinal benign and malignant neoplasms
originate in the small intestine, the actual number is relatively small. Moreover, only 50% of the benign
tumors ever become symptomatic; the remainder are found incidentally at endoscopy, surgery, or
autopsy.37(4542)
The diagnosis of a duodenal tumor is usually first made by upper gastrointestinal x-rays. Although the
finding may be unexpected if not incidental, a duodenal defect does warrant further investigation.
The most direct approach for evaluation is duodenoscopy and biopsy. If the lesion appears
pedunculate by radiographic study, the patient and duodenoscopist should be prepared for
polypectomy in order to avoid an additional procedure. Selection of an endoscope for the evaluation
depends on the location and size of the tumor. A forward-viewing, small-diameter endoscope may
permit the best view of bulbar tumors, but a lesion on the medial wall of the descending duodenum
may be best observed with a side-viewing duodenoscope. When adequate biopsy specimens are
essential, as in the case of a large tumor, or when polypectomy is anticipated, a forward-viewing
instrument with a large-caliber channel will allow adequate access to a lesion that is large enough to
project into the lumen. On occasion, both forward-viewing and side-viewing endoscopes are required
for satisfactory study of the lesion.
In most instances, routine upper gastrointestinal endoscopy probably does not include bowel beyond
the second portion of the duodenum. To reach and examine a more distally located lesion identified on
barium contrast study, endoscopy under fluoroscopy is preferable to be certain that the tip of the
endoscope crosses the spine as it traverses the third portion of the duodenum. Enteroscopy is
frequently required for the evaluation of lesions at or distal to the ligament of Treitz (see Chapter 50:
Endoscopy of the Small Intestine).
A variety of benign polypoid lesions may be found in the duodenum. In a series of 8802 patients,
Matsuura et al.38(4543) encountered the following lesions in the duodenal bulb: hyperplastic polyps in
15 patients, adenomas in 2, Brunneroma in 1, and heterotopic gastric mucosa in 3. In a study of 9
patients whose largest polyp was in the duodenal bulb, 13 of 15 of these patients' other polyps were
also in the duodenal bulb, suggesting that profound spatial clustering occurs on the basis of local
stimuli.39(4544) The most common histologic diagnosis was gastric tissue in a study by Matsui et
al.40(4545) in which endoscopic biopsies were obtained from polyps in the duodenal bulb in 263
patients. Duodenal polyps were found in 27 of 584 patients (prevalence 4.6%) referred for
esophagogastroduodenoscopy (EGD) in the prospective study of Jepsen et al.41(4546) In 16 cases,
endoscopic biopsies disclosed either ectopic gastric mucosa (7 polyps) or inflammation (9 polyps).
Invariably, these small, sessile lesions were multiple and located in the duodenal bulb. Three of 7
polyps covered by normal-appearing mucosa were diagnosed endoscopically as lipomas. Two
adenomas (0.4% of patients; 7% of polyps) were found in the descending duodenum.
Patients referred for endoscopic evaluation of duodenal polyps identified radiologically may prove to
have pseudotumors such as excluded gastric mucosa in the duodenal stump after Billroth II gastric
resection,42(4547) schistosomiasis,43(4548) or an impacted common bile duct stone.44(4549) Thus,
polypoid lesions take many forms and can have bizarre etiologies. The importance of endoscopic
biopsy, even for apparently innocuous polyps, cannot be overestimated.
Adenomas
Adenomas, the most common duodenal neoplasm, occur as single or multiple lesions and are usually
pedunculate. Occult bleeding is the most frequent sign of an adenoma but, as in one of our
cases,36(4550) massive hemorrhage may occur. Endoscopically, the lesion projects into the lumen
and may be a few to 50 mm in diameter.45(4551) The surface of the tumor is the same color as the
surrounding mucosa, but it is usually lobulated, with particularly prominent lobules apparent in larger
adenomas.
Adenomas of the duodenum can be classified histologically in the same manner as colonic adenomas.
The tumor presumably progresses from one composed entirely of tubular glands to one that shows a
florid or papillary glandular pattern. Thus, at one end of the spectrum is the tubular adenoma and, at
the other end, the villous adenoma, with a mixed tubulovillous type occupying an intermediate position.
Large villous adenomas seem to have the same high degree of malignant potential as their colonic
counterparts. For this reason, pedunculate adenomas should be removed endoscopically. Those that
are too large or that cannot be removed endoscopically for other technical reasons should be removed
by surgery.46(4552)
Endoscopic biopsies usually provide an accurate diagnosis for duodenal neoplasms (85%
accuracy),47(4553) but areas of in situ or infiltrating carcinoma can be missed in 60% of villous
adenomas.48(4554)
Of our 10 cases of villous adenoma diagnosed by endoscopy,49(4555) we were impressed with the
variable clinical presentation (bleeding in 5 cases, obstruction in 4, jaundice in 2), as well as the
variation in endoscopic appearance. A villous adenoma in the duodenum may take the form of
prominent folds or a verrucous encircling lesion (Figure 493); it is usually sessile, but 2 of our cases
were pedunculate polyps. The pedunculate polyp should be removed in toto by duodenoscopic
polypectomy, and the specimen should be carefully studied for evidence of invasive cancer. The
sessile lesion, which is usually shaggy and multilobulated, should be thoroughly studied by biopsy.
Pancreaticoduodenectomy is indicated only if an invasive cancer is identified; otherwise, a wide local
excision is appropriate.46(4556) In a review of 44 cases of villous adenomas treated surgically, only 5
required pancreaticoduodenectomy, and 3 were amenable to endoscopic polypectomy alone.50(4557)
In one series that spanned 21 years, 34 villous adenomas were identified, although 22% were
associated with familial adenomatous polyposis (FAP).51(4558) Cancer was found in 16 (47%) of the
villous lesions, 5 of which were removed endoscopically. Only patients with FAP developed recurrent
villous adenomas.51(4559) Invasive carcinoma was present in one third of the 36 duodenal villous
adenomas in the series of Bjork et al.52(4560) Five-year survival after radical
pancreaticoduodenectomy was 45%. A review of 192 villous tumors of the duodenum reported up to
1992 found that 42% contained malignancy at the time of diagnosis.53(4561)
(4562)Figure 493. A, Radiographic views of a villous tumor of the descending duodenum. B,

Endoscopic view of a villous tumor. (A and B, From the collection of Dr. Michael V. Sivak,
Jr.)
Villous adenomas not involving the main duodenal papilla and less than 2 to 3 cm in diameter have
been treated endoscopically by laser photoablation in a few cases.54,55(4563) The tumor recurred in
two of the six patients in the series of Ghilain and Dive54(4564) and was treated again by laser
photoablation. Definitive tissue diagnosis is precluded by laser photoablation, and there is always the
possibility that malignant tissue may remain viable or that malignancy may have already spread beyond
the mucosa. It would therefore be prudent to reserve laser photoablation for patients who are not good
candidates for surgery. Continuing endoscopic surveillance including endoscopic biopsy is also
mandatory in these patients.
Leiomyoma
The leiomyoma is the most frequently occurring benign tumor in the small intestine and is found mainly
in the jejunum.56(4565) Although least commonly found in the duodenum, the leiomyoma is,
nevertheless, the second most prevalent neoplasm in the duodenum.57(4566)
Most leiomyomas arise from the muscularis propria and vary in size from a few millimeters to several
centimeters.37(4567) Exoenteric tumors may become large enough to bleed or less commonly to
produce obstruction and thus come to the attention of the endoscopist.58(4568)
Leiomyomas are submucosal tumors and offer no distinct mucosal feature that permits endoscopic
diagnosis. The lesion bulges into the lumen as a round or ovoid sessile polyp with a smoothly tapering
border (Figure 494); rarely, it is pedunculate.36(4569) As the polyp enlarges, it may lead to pressure
necrosis of the overlying mucosa and a resultant central ulceration. Bleeding from this ulcer usually
leads to endoscopy, and this central hemorrhagic ulcer overlying a submucosal tumor should arouse a
suspicion of leiomyoma.59(4570)
(4571)Figure 494. Proximal (A) and distal (B) views of a large submucosal duodenal tumor
that proved to be a leiomyoma at surgery. (A and B, From the collection of Dr. Michael V.
Sivak, Jr.)
In our endoscopic series of 45 duodenal polyps, 19 biopsy specimens showed normal mucosa or
mucosal inflammation.36(4572) It is likely that many small submucosal bulges generally ignored by the
endoscopist are leiomyomas that probably originate in the muscularis mucosae. Endoscopic
ultrasonography (EUS) is frequently of value in the differential diagnosis of submucosal tumors,
leiomyomas in particular (see Chapter 51: Endoscopic Ultrasonography: Stomach and Duodenum).
Brunner's Gland Adenoma
Polyps that arise from Brunner's gland of the duodenum are rare lesions found proximal to the major
duodenal papilla, most often in the bulb. Although the adenoma is the most common form of Brunner's
gland hyperplasia, there are diffuse nodular and circumscribed nodular forms (Figure 495) as
well.60,61(4573) Because of the bulbar location and the propensity of these tumors to bleed or
obstruct the lumen, Brunner's gland adenoma, or Brunneroma, may be mistaken for peptic ulcer
disease if not well visualized endoscopically.
(4574)Figure 495. Circumscribed nodular form of Brunner's gland hyperplasia. (From the
collection of Dr. Michael V. Sivak, Jr.)
The endoscopic findings may vary from multiple 2- to 3-mm nodules carpeting the mucosa of the
proximal duodenum to a solitary polyp of a few millimeters to several centimeters in diameter (Figure
496). The diffuse nodular pattern accounts for the characteristic "Swiss-cheese" radiographic
appearance. Although Brunner's glands are submucosal, primarily below the muscularis mucosae,
endoscopic biopsies are frequently diagnostic because hyperplasia results in extension of the glands
into the mucosa.36,60,62(4575) Brunner's gland hyperplasia is not a premalignant lesion, and surgical
excision is unnecessary unless tumor size becomes sufficient to cause obstruction by intermittent
intussusception.63(4576)
(4577)Figure 496. Pedunculated Brunner's gland adenoma in the proximal descending

duodenum. After injecting the stalk with epinephrine, the lesion was removed by
electrosurgical snare polypectomy. (From the collection of Dr. Michael V. Sivak, Jr.)
Brunner's gland adenoma may bleed64,65(4578) and require excision. If pedunculate, these polyps
are best removed by endoscopic polypectomy; seven cases of successful endoscopic removal have
been described.36,64,6670(4579)
Some caveats should be observed with snare excision of a duodenal polyp. The major duodenal
papilla must be identified, so that if it is prominent or protuberant, it is not inadvertently snared.
Excessive coagulation in the thin-walled duodenum increases the risk of perforation. Conversely,
inadequate coagulation increases the danger of excessive bleeding from the richly vascular
duodenum, which is bathed in acid-peptic juices.
Lipoma
Lipomas are found three times more commonly in the small intestine than in the stomach but less than
half as often as in the colon.36(4580) These submucosal tumors may become several centimeters in
diameter and form a pseudopedicle as they enlarge. With a sessile lesion, a positive tissue diagnosis
can be made by the so-called drill technique. This biopsy method involves taking several biopsies from
the same site, so that underlying submucosal fat composing the lesion is retrieved.
The endoscopic appearance is similar to that of other submucosal duodenal tumors (Figure 497).
However, the polypoid form lends itself to palpation with the blunt tip of the biopsy forceps, by means of
which the soft nature of the tumor is appreciated and the characteristic "pillow" signa smooth
indentation made by the forcepsis elicited.
(4581)Figure 497. A and B, Two views of a soft, whitish, polypoid submucosal duodenal
tumor presumed to be a lipoma. (A and B, From the collection of Dr. Michael V. Sivak, Jr.)
The finding of negative absorption coefficients for a duodenal tumor noted on CT may actually provide
a diagnosis of lipoma more reliably than biopsy.71(4582) Attenuation values of 61 Hounsfield units
were observed in a patient who had an obstructing duodenal lipoma.72(4583)
Intraluminal polypoid lipomas may intussuscept, producing intermittent obstruction; also, bleeding may
occur as a result of superficial necrosis.71(4584) When the lipoma is symptomatic, endoscopic
polypectomy is feasible if a sufficiently developed pseudopedicle has formed.73(4585)
Neurofibroma
Solitary neurofibromas are usually found in the ileum. They occur in 15% of patients with von
Recklinghausen's disease, with the most frequently involved intestinal segments being (in descending
order) the jejunum, the ileum, and the duodenum.56(4586) In a review of 28 cases of neurogenic
tumors of the small intestine reported since 1955, Sivak et al.74(4587) recorded 10 duodenal
neoplasms, all of which had ulcerated and bled; 4 were neurofibromas, 2 neurofibromosarcomas, 2
paragangliomas, and 2 neuromas. In a patient with von Recklinghausen's disease, these workers
performed endoscopic polypectomy of an ulcerated 2.5-cm round mass 4 cm distal to the main
duodenal papilla. The submucosal polyp was a neurilemoma. In another patient with von
Recklinghausen's disease, multiple submucosal elongated polypoid tumors were present in the pylorus
and the first and second portions of the duodenum.75(4588) Several of these tumors had a central
umbilication similar to that seen with leiomyoma. The duodenal mucosa had a diffuse reticular brown
pigmentation. Endoscopic biopsies of the polyps were consistent with diagnosis of a neurofibroma.
The duodenal neurogenic tumor is exceedingly rare. It is dangerous because of its propensity to bleed
and to undergo malignant transformation. When solitary, it may be amenable to endoscopic
polypectomy, but the upper gastrointestinal neurogenic tumor is usually one of several, and surgical
resection is necessary.
Lymphangioma
Lymphatic cysts, or lymphangiomas, are rarely seen in the duodenum.7678(4589) These lesions are
located in the submucosa but give the overlying mucosa a yellow color. The polypoid tumor is soft and
can be removed by endoscopic polypectomy,76(4590) or the cyst can be drained by means of multiple
biopsies.79(4591) Ultrastructural study of two duodenal lymphangiomas diagnosed by endoscopy
suggests a hamartomatous origin, and the lymphangioma may be nothing more than a
lymphangiectatic cyst.80(4592)
Pancreatic Rest
Heterotopic pancreas may be found in the duodenum, more commonly in the bulb. These formations
are usually noted incidentally at endoscopy as submucosal nodules with a central umbilication. Biopsy
specimens infrequently contain pancreatic acini and ducts; therefore, diagnosis is essentially by
endoscopic observation. Pancreatic rests have been responsible for duodenal obstruction, massive
hemorrhage, and common bile duct obstruction,81(4593) but none of these major complications has
as yet been reported as an endoscopic diagnosis. The pancreatic rest is also discussed in Chapter 72:
Congenital Anomalies of the Pancreas.
Telangiectasias
Telangiectasias may be seen in the duodenum either as solitary lesions or as part of the multiorgan
involvement of Osler-Weber-Rendu disease. In the latter condition, it would be unusual to see
telangiectasias in the duodenum without also finding them in the stomach. Bleeding is a frequent
complication, and endoscopic evaluation provides definitive diagnosis. These lesions may be pale and
not readily visible if the patient is anemic. Patients should be reexamined after blood replacement if
initial studies fail to explain the gastrointestinal hemorrhage. Telangiectasias have the characteristic
raised central arteriolar bleb. Thin tendrils radiate outward from this central point (Figure 498). It
would appear to be unusual to witness spontaneous bleeding from a gastrointestinal telangiectasia.
(4594)Figure 498. Distant (A) and close-up (B) views of duodenal telangiectasia. (A and B,
From the collection of Dr. Michael V. Sivak, Jr.)
Telangiectasias of the duodenum can be eradicated by endoscopic electrocoagulation or
photocoagulation, but this may not solve the problem of blood loss if there are more vascular lesions
distally in the intestine.
Polyposis Syndromes
Adenomas of both stomach and duodenum occur with surprising frequency in patients with FAP and
the Gardner syndrome variant (GS).8284(4595) In patients with FAP or GS, endoscopy will reveal
duodenal adenomas in about a third;82,85(4596) the adenomas are usually multiple, small, white, and
located in the second portion of the duodenum.82(4597) Yao et al.83(4598) noted similar "minute-size
adenomas" as well as larger lesions that ranged from 5 to 22 mm (Figure 499). Duodenal
adenocarcinoma may also be encountered at EGD in patients with FAP (Figure 4910).86(4599) Using
a side-viewing duodenoscope, Iida et al.87(4600) studied the major duodenal papilla in 24 patients with
major FAP, and in 10 patients there was a granular-to-nodular surface that, on biopsy, proved to be
adenomatous change. In a similar study of FAP patients, almost half had an abnormal-appearing major
duodenal papilla.88(4601)
(4602)Figure 499. Small, irregular-shaped sessile and semisessile polyps in patients with
familial adenomatous polyposis (FAP) (A) and Gardner's syndrome variant (B and C). D,
Confluent whitish sessile duodenal polyps in a patient with FAP. (A-D, From the collection of
Dr. Michael V. Sivak, Jr.)
(4603)Figure 4910. A-E, Adenocarcinoma of the duodenum in five patients. The lesion seen
in A was found in the patient with FAP whose more proximal polyp is seen in Figure 499A.
The cancer shown in B also arose in a patient with FAP. (A-E, From the collection of Dr.
Michael V. Sivak, Jr.)
Gastric and duodenal polyps are found in patients with GS, and in a review of the literature to 1976, 36
cases of gastric or small bowel (mostly duodenal) adenomas were collected.88(4604) Prospective
endoscopic studies indicate that duodenal adenomas develop frequently in patients with GS, especially
in the region of the major duodenal papilla.89(4605) In one case with only duodenal adenomas,
superficial carcinoma was identified in a large adenoma.90(4606) Endoscopic removal of duodenal
polyps, which proved to be adenomatous, was accomplished in another patient with GS, and severe
dysplasia was evident on microscopic study of the adenomas.91(4607)
Subsequent reports confirm that patients with FAP and GS are at increased risk for the development of
carcinoma from duodenal or periampullary adenomas.85,92,93(4608) There is a 3 to 12% incidence of
duodenal carcinoma in patients with GS.93(4609) In a study that compared the incidence rate of upper
gastrointestinal cancer in patients with FAP with that for the general population through person-year
analysis, Offerhaus et al.94(4610) found an increased relative risk of duodenal carcinoma and
carcinoma of the major duodenal papilla but no significant increase in risk for cancer of the stomach,
jejunum, or ileum. Recurrence of villous adenomas after surgical or endoscopic removal may occur in
patients with FAP. This is uncommon in patients with solitary or sporadic duodenal
adenomas.51(4611)
The exclusion of gastric and duodenal adenomas and surveillance for carcinoma are, therefore,
sufficient reasons to perform EGD routinely on patients with FAP and GS.9496(4612) Offerhaus et
al.94(4613) suggested that prophylactic duodenectomy may be indicated when endoscopy
demonstrates the presence of large adenomas with high-grade dysplasia. Data concerning the
long-term benefit of surgery for such patients have not been presented. Because malignant change
has been demonstrated with duodenal adenomas, endoscopic polypectomy, when feasible, is
appropriate for large polyps96(4614) and electrocoagulation or excision for small polyps. Adenomas of
the papilla are a special problem; endoscopic follow-up with biopsy at intervals of 1 to 2 years to
exclude dysplasia may be a reasonable approach.
Cronkhite-Canada syndrome is a rare disorder characterized by alopecia, skin pigmentation, nail
atrophy, and diffuse gastrointestinal polyposis.97(4615) The polyps are an inflammatory type and
contain mucous retention cysts; they are located mostly in the stomach and colon, but the small bowel,
particularly the duodenum, may be affected. Small sessile polyps seen on endoscopy almost
completely replace the mucosa, and biopsy reveals the characteristic retention cysts. These polyps
have no malignant potential, and thus, nothing more than correct identification is required.
Spontaneous remission of the syndrome, with complete disappearance of the gastrointestinal polyps,
has been reported.97(4616)
Hamartomatous polyps develop throughout the entire digestive tract in Peutz-Jeghers syndrome, an
autosomal dominant disorder. The diagnosis is confirmed by the radiographic finding of multiple small
bowel polyps, usually jejunal, in a patient with melanin spots on the lips, buccal mucosa, and tips of the
fingers and toes. Endoscopy is indicated in the unusual circumstance of bleeding localized to the
duodenum or when duodenal intussusception from a large polyp is suspected. Endoscopic
polypectomy may avoid surgical intervention. Microscopically, the retrieved polyp will have the
distinctive branching stalk of smooth muscle fibers that is thicker at the center of the polyp than at its
end.37(4617)
The small intestine is the most common site for the occurrence of polyps in patients with the
Peutz-Jeghers syndrome. Push enteroscopy or intraoperative enteroscopy with or without polypectomy
is frequently helpful in the management of these patients (see also Chapter 50: Endoscopy of the
Small Intestine).98,99(4618)
Malignant Tumors
Adenocarcinoma
About 50% of all primary malignancies of the small intestine are adenocarcinomas.100(4619)
Carcinomas of the duodenum account for only 0.35% of all gastrointestinal carcinomas, but relative to
its length, the duodenum harbors the largest number of cancers in the small bowel.101(4620)
Duodenal carcinoma is usually classified according to its anatomic relationship to the major duodenal
papilla. Because carcinoma of the bulb is very rare, supraampullary tumors are essentially found in the
second portion of the duodenum, and in a study of 71 cases of duodenal carcinoma, these constituted
7% of cases;102(4621) infraampullary and periampullary tumors made up 56% and 32%, respectively.
Discrete ampullary cancers are not always clearly separated from the periampullary group in many
reports. For some authors,100,103(4622) the term periampullary has traditionally included ampullary
adenocarcinoma. More often than not, it is impossible to decide, either endoscopically or at surgery,
whether the cancer arose from the duodenal mucosa and encroached on the papilla or vice versa (see
also Chapter 69: Tumors of the Main Duodenal Papilla). The other problem with classification of
periampullary adenocarcinoma is the determination of the tissue origin because the cancer may arise
not only from contiguous duodenal mucosa but also from the distal bile and pancreatic ducts, ampulla,
and pancreas. To further complicate the matter, duodenal adenocarcinoma may also cause biliary
obstruction and jaundice.104(4623) The most accurate diagnosis of ampullary cancer is by gross
pathologic and microscopic study of resected or autopsy specimens, although uncertainty may still
persist in some cases.
Duodenoscopy has had an important impact on preoperative diagnosis of duodenal adenocarcinoma.
In one series of 17 patients, endoscopy provided a histologic diagnosis of cancer in 5 of the 7 cases in
which it was performed.100(4624) In another series of 12 adenocarcinomas, 2 cases not diagnosed by
radiographic study were confirmed by duodenoscopic biopsy.101(4625) The combination of brush
cytology and biopsy provided a correct diagnosis in all 4 patients reported by King et al.105(4626)
There is one report of duodenoscopic excision of an "early" duodenal carcinoma by means of the "strip
biopsy" technique.106(4627) The small lesion was discovered incidentally in a 67-year-old man;
follow-up was relatively short at 10 months.
The duodenoscopic appearance of adenocarcinoma is not specific and cannot be differentiated with
confidence at endoscopy from leiomyosarcoma or lymphoma. The lesion is usually nodular and
polypoid, but it may also be ulcerated or friable (see Figure 4910). It is most often confined to the
second or third portion of the duodenum, whereas lymphoma may be diffuse.101(4628) Because
biopsies are generally easy to obtain, a tissue diagnosis can be made in most cases.105(4629) Walsh
et al.107(4630) failed to obtain malignant tissue by endoscopic biopsy in three of six cases of
adenocarcinoma of the major duodenal papilla, but the papilla was visually abnormal in all six patients.
Leiomyosarcoma
In the small bowel, myogenic cancers are most common in the jejunum and least common in the
duodenum (10%).108(4631) There are parallel percentages of leiomyoma (50%) and leiomyosarcoma
(53%) in the descending portion of the duodenum.109(4632) One case was reported in which a
biopsy-proven 4-cm diameter leiomyoma that was followed became a 7-cm leiomyosarcoma that was
surgically excised.110(4633) Thus, circumstantial evidence suggests that the leiomyoma is the
precursor of the leiomyosarcoma, indicating that large leiomyomas should be excised even if found
only incidentally. Until 1981, 123 cases of duodenal leiomyosarcoma were reported in a review of the
literature.111(4634)
The endoscopic appearance will vary depending on the size of the tumor and whether it is intraluminal,
intramural, or exoenteric. In one instance, the endoscopist saw an umbilicated polypoid mass that was
friable and bled profusely on biopsy.112(4635) However, the tumor may be submucosal and develop a
central ulceration as it enlarges (Figure 4911).109(4636) The ulcerated form would presumably be the
most common, because bleeding is the chief complaint in over half the cases.37(4637)
(4638)Figure 4911. A, Radiograph showing an ulcerated leiomyosarcoma of the duodenum

with a fistulous track. B, Endoscopic view with the duodenum to the left of the field, the tumor
in the center, and the fistula to the right. C, Closer view of the lesion, with the duodenal lumen
to the left. D, Close-up view of the lesion, with the fistula to the right. (A-D, From the
Endoscopic biopsy of duodenal leiomyosarcoma is not reliable or, more accurately, not interpretable as
benign or malignant.93,109,112(4639) The differential diagnosis by EUS is also difficult (see Chapter
51: Endoscopic Ultrasonography: Stomach and Duodenum). Although large tumors are most often
malignant, standard endoscopic biopsies are inadequate for a count of mitotic figures, on which the
diagnosis of malignancy rests. Investigational endoscopic techniques for obtaining biopsies from
submucosal lesions may prove to be useful (see Chapter 23: Needle Biopsies of Submucosal
Lesions). At times, the diagnosis of leiomyosarcoma must await the demonstration of metastases
during the clinical course, because even the surgically excised tumor may not meet the pathologic
criteria for cancer.111(4640)
Lymphoma
Although lymphomas are rare in the gastrointestinal tract of adults, they are the most common
digestive tract tumor in children. Lymphomas occur least frequently in the duodenum, except in
patients with immunoproliferative small intestinal disease (IPSID).37(4641)
The duodenum was involved in all six cases of IPSID studied endoscopically.113(4642) The mucosa
was thickened and covered by many sessile polyps of varying sizes. In one patient, scattered ulcers
were seen, and in another, a large ulcerative mass was found in the third portion of the duodenum
(Figure 4912). Unfortunately, this report does not indicate whether endoscopic biopsies were
diagnostic.
(4643)Figure 4912. A, Secondary involvement of the duodenum by lymphoma with several

discrete nodules. B, More extensive and confluent involvement of the same area 2 months
later. C, Lymphoma in the second portion of the duodenum in a patient with acquired
immunodeficiency syndrome (AIDS). (A-C, From the collection of Dr. Michael V. Sivak, Jr.)
In another retrospective study,114(4644) primary malignant lymphoma was diagnosed preoperatively
in only one of six patients. There were no characteristic endoscopic findings. Although duodenoscopic
biopsy may provide a diagnosis of lymphoma, the differentiation into Hodgkin's and non-Hodgkin's
types of lymphoma requires surgical biopsy.114(4645)
Lymphosarcoma of the duodenum has been described, but less than 75 cases have been collected
since 1877.115(4646) Endoscopy in 1 patient showed a "fungating ulcerating lesion of the second
portion of the duodenum."115(4647) The biopsy was originally interpreted as adenocarcinoma, but on
comparison with the surgical biopsy, it was considered consistent with lymphosarcoma.
Metastases
Malignant invasion of the medial wall of the duodenum most likely originates in the pancreas, stomach,
or biliary tree, but contiguous spread from the right kidney and the hepatic flexure of the colon either
directly or via lymphatic channels may also occur.37(4648) Excluding the "obvious" carcinomas of the
pancreas, stomach, and biliary tract, Veen et al.116(4649) collected 14 cases of duodenal metastases
over a 10-year span: 5 from the colon; 2 each from the kidney and pancreas; 2 from cutaneous
melanoma; and 1 each from the uterus, esophagus, and gallbladder.
Pancreatic cancer compromising the duodenal lumen is a common development in patients who
survive 10 months from presentation.117(4650) The usual endoscopic appearance is that of a soft
polypoid tumor involving the second or third portion of the duodenum (Figure 4913).118,119(4651)
Biopsies are almost always confirmatory of pancreatic cancer.
(4652)Figure 4913. A, Duodenal invasion by carcinoma of the pancreas. B, Ulcerated,

bleeding carcinoma of the pancreas is visible in the medial wall of the duodenum. C,
Cystadenocarcinoma of the pancreas invading the duodenum. (A-C, From the collection of Dr.
Cancer of the stomach only rarely spreads to the duodenum, perhaps because of the poor lymphatic
connections between the two organs.120(4653) Lymphatic invasion by gastric cancer cells turned the
color of the duodenal mucosa to white in one case,120(4654) but polypoid tumors surrounded by
mucosa of a more normal color were also identified, biopsies from the latter yielding a histologic
diagnosis (Figure 4914).
(4655)Figure 4914. Involvement of the duodenum by metastatic gastric adenocarcinoma.

Note the very small whitish nodules that cover much of the mucosal surface. Biopsy was
positive for carcinoma. (From the collection of Dr. Michael V. Sivak, Jr.)
Four cases of hypernephroma metastatic to the duodenum were discovered endoscopically because of
the onset of upper gastrointestinal bleeding.121(4656) A nodular mass with ulceration was noted at
duodenoscopy in two cases, a multilobulated smooth mass in one case, and a circumferential ulcer in
another case (Figure 4915). Biopsies in all cases were negative for malignancy, although cytologic
findings were positive for malignant cells in one. The hypervascularity of the hypernephroma impedes
retrieval of adequate biopsy material and contributes to the failure to obtain an appropriate pathologic
diagnosis.121(4657)
(4658)Figure 4915. Hypernephroma metastatic to the duodenum. (From the collection of Dr.
Malignant melanoma metastasizes frequently to the stomach and duodenum. Bleeding or obstruction
related to large ulcerated tumors warrants endoscopic examination. If the duodenal tumor is
pigmented, it is recognized readily at endoscopy (Figure 4916). The nodules vary in size and are
usually ulcerated. Biopsy specimens are usually positive for melanoma.122124(4659) The
gastrointestinal metastasis may be the only one apparent, and endoscopic examination becomes an
important evaluation because surgical resection may result in prolonged palliation.122(4660)
(4661)Figure 4916. Focal metastatic nodule of malignant melanoma in the duodenum. Note
the brownish color of the lesion. (From the collection of Dr. Michael V. Sivak, Jr.)
Metastases from carcinoma of the breast,123(4662) testis,124(4663) ovary,124(4664) and uterine
cervix125(4665) have been identified at endoscopy. Recently, gastrointestinal tract involvement with
hepatocellular carcinoma was observed in 8 of 396 patients, and the metastatic site was the
duodenum in 4 patients.126(4666) In the patient with duodenal metastasis from an hepatocellular
carcinoma reported by Humbert et al.,127(4667) the endoscopic appearance was similar to that of the
surface of the liver. A patient with metastatic bronchogenic carcinoma to the stomach and duodenum
has been described; the findings included multiple polypoid lesions in the stomach and duodenum as
well as a large gastric ulceration.128(4668)
Miscellaneous Malignant Tumors
An extramedullary plasmacytoma, a lesion usually associated with multiple myeloma, was identified in
the first portion of the duodenum as a primary isolated tumor.129(4669) Endoscopic biopsy of this
fungating mass was followed by severe hemorrhage that necessitated laparotomy.
Visceral involvement may occur in 10% of patients with Kaposi's sarcoma; the gastrointestinal tract is
commonly affected, the majority of lesions being found in the stomach.130(4670) There may be
duodenal extension of the maculopapular-to-polypoid tumor that takes on a dark red color as it
enlarges. Lesions of 1 cm or greater may have a central umbilication of variable depth. Endoscopic
biopsy is needed for tissue diagnosis, but because of the submucosal location of the tumor, the tissue
samples are often negative.131(4671)
Carcinoid tumors of the gastrointestinal tract are found mainly in the appendix (50%) and next most
commonly in the small intestine (25%). The most frequent site in the small bowel is the distal 50 cm of
ileum.37(4672) The duodenum is the second least common site. Since foregut carcinoids exhibit an
argyrophil reaction but not argentaffin staining, they rarely cause carcinoid syndrome. Endoscopically,
they are typically small (less than 2 cm) submucosal lesions covered by normal mucosa that may have
a yellow-orange tint (Figure 4917).36,132(4673) In three of six patients, the lesion showed an
irregularly shaped erythematous depression.133(4674) Bleeding from the tumor may occur.134(4675)
Biopsies are usually diagnostic.133(4676) Carcinoid tumors, although indolent, are malignant, and
surgical excision is considered mandatory, but in the elderly patient with relative contraindications to
surgery and an asymptomatic lesion of less than 1 cm, a tumor size that rarely metastasizes, a
nonoperative approach would be reasonable.135(4677)
(4678)Figure 4917. Carcinoid tumor in the duodenal bulb. (From the collection of Dr.
Gastrinomas, of which two thirds are malignant, are most often localized to the pancreas, but the
duodenal wall is a frequent extrapancreatic site. Because these tumors are small and submucosal,
they may be overlooked at endoscopy. The opportunity for an early biopsy diagnosis of a duodenal
gastrinoma afforded by endoscopy should not be missed, since management will be governed by the
endoscopic finding.136138(4679) Indeed, Zollinger-Ellison syndrome has been "cured" by endoscopic
polypectomy of a submucosal 3- to 4-mm tumor.136,139,140(4680) However, if the serum gastrin

does not return to normal, then the polypectomy was incomplete or there is a pancreatic gastrinoma
present as well.136,141(4681) In patients who undergo laparotomy for isolation of gastrinoma,
intraoperative endoscopic transillumination is essential to effect a cure, since in a prospective study of
26 patients with Zollinger-Ellison syndrome, 39% had duodenal gastrinomas.142(4682) Because these
lesions are often small and difficult to palpate, intraoperative endoscopic transillumination can be of
considerable assistance to the operating surgeon.
Miscellaneous Disease
Foreign Bodies
More than 99% of foreign bodies that enter the duodenum pass through the small intestine. The
junction of the second and third portions of the duodenum is the first anatomic flexure at which the
foreign body may become lodged.143(4683) Toothpicks account for 10% of intestinal
perforations143(4684) and are particularly difficult to localize because the wooden (now plastic as well)
toothpick is not radiopaque. Toothpicks, however, can be easily removed endoscopically from the
duodenum with forceps, even if they are already perforating the wall (Figure 4918).144,145(4685)
(4686)Figure 4918. Toothpick in the duodenum. It was readily grasped and retrieved with
the biopsy forceps.
Pins and needles have a high risk of perforation if they are retained in the duodenum, and immediate
endoscopic extraction is indicated. A needle can be grasped with forceps and pulled safely into the
accessory channel of the endoscope. An open safety pin is best retrieved by putting the closed biopsy
forceps through the loop and then opening the jaws of the forceps; the safety pin is captured and can
be pulled out with the endoscope.146,147(4687) Use of an overtube with the endoscope also adds a
margin of safety and may make the procedure technically easier. A snare is another useful retrieval
device. A thermometer was extracted from the duodenum by placing a polypectomy snare around the
neck of the thermometer.148(4688) Endoscopic removal of an hepatic artery infusion catheter that has
penetrated the duodenal wall is feasible if it can be ascertained the catheter is free of all vascular
ligatures and the origin of the artery catheterized has thrombosed.149(4689) To remove impacted
foreign bodies from the duodenum endoscopically requires innovative technique, patience, and the
ability to analyze the situation and the operation that best handles it.150(4690)
A duodenal bezoar was encountered between the second and the third segments of the duodenum in
a patient who had a gastroduodenostomy. It was conjectured that narrowing of the flexure was a result
of the surgery, although two diverticula were present in the area.151(4691) Duodenal bezoars are
almost always associated with obstruction and have been reported in patients with duodenal
web152(4692) and diaphragm.153(4693) Foreign bodies and their extraction by endoscopic methods
are considered in detail in Chapter 54: Foreign Body Extraction.
Diverticula
Duodenal diverticula are found during 10 to 20% of endoscopic examinations and, when juxtapapillary,
are thought to be associated with an increased incidence of gallstones and pancreatitis (see Chapter
62: Calculus Disease of the Bile Ducts).154,155(4694) The diverticulum often encompasses the papilla
at its superior border, but rarely, the papilla may be located entirely within the diverticulum. In one fifth
of cases, the diverticula are multiple and may, in a few cases, border either side of the papilla, giving
the so-called pantaloon effect.156(4695) Although diverticulitis with subsequent perforation156(4696)
and diverticular bleeding157,158(4697) have been recorded, diverticula are not known to interfere with
or complicate duodenoscopy.
An unusual variant of duodenal diverticula is the intraluminal duodenal diverticulum (IDD), a
developmental abnormality resulting from the effect of peristaltic activity on a duodenal
diaphragm.159(4698) The most common symptoms are abdominal pain and nausea, often associated
with meals. The IDD has a diagnostic radiographic appearance, and duodenoscopy is invoked to
evaluate the pouch for bleeding or inflammation and to localize the major duodenal papilla.160(4699) A
double aperture, one for the pouch and the other for the duodenal lumen, is better appreciated with the
side-viewing duodenoscope. However, only the pouch will be seen if it is circumferentially attached.
Duodenoscopy may be useful for removal of impacted debris in the pouch or for dilation of the outflow
ostium of the IDD.139,161(4700) Hajiro et al.,162(4701) using a polypectomy snare, endoscopically
excised the tip of an IDD that had been inverted and thus provided an adequate opening in the center
of the diverticulum. A similar procedure was reported by Hiraoka et al.163(4702) Ravi et al.164(4703)
incised an IDD using a needle knife and then a standard papillotome. Although only a small number of
patients will be found, endoscopic techniques are probably applicable in most cases of obstructed IDD
and should replace surgical duodenotomy and excision.
Diaphragms, Congenital Stenosis, and Duplication Cysts

Duodenal diaphragm (web) is regarded as a congenital anomaly. In more than half of patients, the
diagnosis has been made during infancy or childhood. However, the age range is wide, and many case
reports pertain to older patients. For example, Cooperman et al.165(4704) reported the case of a
woman with symptoms of postprandial nausea and vomiting who presented at the age of 73 years with
massive dilation, presumably longstanding, of the stomach and duodenal bulb (Figure 4919). It is
supposed that some patients maintain adequate nutrition and remain asymptomatic into adulthood
because the degree of obstruction produced by the diaphragm can vary from case to case. However,
the aperture has been found at surgery to be exceedingly small, so that the exact explanation for the
wide age range remains uncertain. Endoscopic diagnosis in an adult patient was reported by Ahmed et
al.166(4705) in 1972. Suarez and Bolden152(4706) reported endoscopic evaluation of a patient in
which they were unable to pass the endoscope through the eccentrically positioned 10-mm aperture. A
forme fruste diaphragm was found by the author at duodenoscopy (Figure 4920). Therapeutic
intervention has been surgical in the past, but there are now several case reports of endoscopic
management by various methods including laser ablation.167170(4707)
(4708)Figure 4919. Duodenal diaphragm in an elderly woman. Note the eccentrically placed
opening, about 6 or 7 mm in diameter, through the diaphragm. (From the collection of Dr.
(4709)Figure 4920. Duodenal diaphragm. This incomplete diaphragm was nonobstructing

and so was not divided by endoscopic electrocoagulation.
Congenital stenosis has been identified endoscopically at the apex of the duodenal bulb, which had a
circumferential cuffing of the mucosa "giving the appearance of a cervix."171(4710) The stenosis did
not relax in response to intravenous glucagon. I have encountered a postsurgical, postbulbar stenosis
that balloon dilation improved temporarily (Figure 4921).
(4711)Figure 4921. Left and right, Postsurgical, postbulbar stenosis. Balloon dilation was
effective only temporarily.
Duplication cysts of the gastrointestinal tract are uncommon congenital anomalies characterized by the
presence of a smooth muscle layer and a lining of gastrointestinal mucosa. Although these are usually
recognized during childhood, they occasionally remain asymptomatic until later in life. The most
common symptoms are those of duodenal obstruction, but duplication cysts can also cause abdominal
pain, jaundice, pancreatitis, and hemorrhage.172(4712) Typically, these cysts contain clear fluid, but
those that arise in the duodenum may communicate with the main duodenal papilla and can therefore
contain bile, pancreatic juice, and even gallstones.173(4713)
At endoscopy, a duplication cyst appears as a round or elongated structure covered by
normal-appearing mucosa. When the cyst is found within the second portion of the duodenum,
choledochocele must be included in the differential diagnosis.174(4714) Because the management of
these two lesions differs, accurate diagnosis is essential. Several imaging studies may be of
assistance in demonstrating the cystic nature of the lesion, but precise diagnosis depends on the
demonstration that the cyst consists of smooth muscle with a mucosal lining. EUS could potentially be
of assistance, but the endosonographic appearance of a duodenal duplication cyst has not been
reported.
Surgical management is traditional, but a few duodenal duplication cysts have been treated
successfully by endoscopic means including incision with drainage175,176(4715) and complete
excision.177(4716) Biopsies obtained from within the cyst after endoscopic incision revealed intestinal
mucosa in the case reported by Johanson et al.176(4717)
Varices
Duodenal varices may be found in patients with cirrhosis or extrahepatic portal hypertension (Figure
4922). In the latter, esophageal and gastric varices are often absent. In the series of 13 cases
reported from the Middle East by Al-Mofarreh et al.,178(4718) the etiology in 9 cases was liver fibrosis
due to schistosomiasis. Isolated duodenal varices have also been described. In the case reported by
Gushurst and Lesesne,179(4719) a duodenal varix, diagnosed at endoscopy, arose directly from the
inferior vena cava and did not communicate with the portal venous system.
(4720)Figure 4922. Duodenal varices. (From the collection of Dr. Michael V. Sivak, Jr.)
The endoscopic diagnosis of duodenal varices was first reported by Kunisaki et al.180(4721) in 1973
and by Kunert and Ottenjann181(4722) in 1977. Bleeding from duodenal varices can be severe and is
frequently misinterpreted endoscopically as arising from a duodenal ulcer.182,183(4723) The varices
are postbulbar and may not be suspected if the endoscopist has not already encountered esophageal
varices. When the source of active duodenal bleeding remains undiagnosed, the patient should
undergo arteriography184(4724) or hepatic portography,182(4725) which will demonstrate clearly, as a
rule, the duodenal varices. On one occasion, however, arteriography was negative and a large, dilated,
blue varix of the second portion of the duodenum was easily seen at duodenoscopy.185(4726) It might
be anticipated that duodenal varices would become a more common diagnostic problem in view of the
number of patients with portal hypertension undergoing endoscopic sclerotherapy for complete
obliteration of esophageal varices. In point of fact, if recollateralization of portal flow through the
duodenal venous system occurs at all, it is unusual to see bleeding duodenal varices following
esophageal variceal obliteration.186(4727) Of the 30 cases reported in the literature, about 40% occur
in the setting of extrahepatic obstruction.186(4728)
The usual treatment of bleeding duodenal varices is surgical;182,185,187(4729) portacaval shunt,
suture ligation, or excision of the varix is usually carried out. However, Sauerbruch et al.188(4730)
successfully performed sclerotherapy in a patient who had a portacaval shunt and had bled from a
duodenal varix. Subsequently, there have been several reports of good outcomes with sclerotherapy in
individual patients.189192(4731)
Hemangiomatosis and Hematoma

Capillary hemangiomas of the duodenum are not rare, but cavernous hemangiomas are distinctly
unusual. They are congenital and may develop in other organs, particularly the skin. Their endoscopic
identification as a vascular lesion should not be difficult. A submucosal tumor with bridging folds and a
lustrous, slightly blue color is characteristic.193(4732) However, unless the entire duodenum is
examined, even large hemangiomas will be missed, only to be demonstrated by
angiography.194(4733)
Intramural duodenal hematoma results primarily from blunt abdominal trauma, but anticoagulant
therapy195(4734) and chronic pancreatitis196(4735) have also been incriminated. It has also been
reported as a complication of ERCP,197(4736) the hematoma forming in the peripapillary area as a
result of accidental intramural injection of contrast material. With spontaneous intramural duodenal
hematoma associated with anticoagulant therapy, the duodenal lumen at duodenoscopy has been
found to be concentrically narrowed by purple ecchymotic folds of varying thickness; "the duodenal
bulb is never involved."195(4737) In one report of intramural duodenal hematoma,198(4738) the fixed,
thickened folds were misinterpreted as a secondary feature of pancreatic cancer at
gastroduodenoscopy. In the other cases,196,197(4739) only a purplish bleb was evident
endoscopically. Following endoscopic duodenal mucosal biopsies, complete duodenal obstruction due
to a postbiopsy hematoma developed in an 8-year-old child199(4740) and a 23-year-old
man.200(4741) It is conjectured that undue traction on the relatively fixed duodenum by the biopsy
forceps may tear the rich submucosal vascular plexus. Aizawa et al.201(4742) described the case of a
52-year-old man with an intramural duodenal hematoma that was successfully managed by a
combination of ultrasonographically guided aspiration, endoscopic balloon dilation of the duodenal
narrowing, and parenteral nutrition.
Fistulas
Biliary-Duodenal Fistula
Cholecystoduodenal fistula is the most common type of biliary enteric fistula, but choledochoduodenal
or cholecystoduodenocolic fistulas also occur. Most of these benign fistulas are caused by erosion of a
large gallstone, but peptic ulcer may be responsible in a small number of cases. A stone larger than 2
cm may become impacted in the small bowel, usually the terminal ileum, and only rarely in the
duodenum.202,203(4743)
Acute pyloroduodenal obstruction, a condition usually referred to as Bouveret's syndrome, may occur
when a large gallstone gains access to the duodenum via a cholecystoduodenal fistula and becomes
impacted in the duodenal bulb. In addition to symptoms of gastric outlet obstruction, some patients
have presented with upper gastrointestinal hemorrhage.203205(4744) If the stone is impacted in the
duodenum, it may be viewed endoscopically.202,203,206,207(4745) However, a stone impacted in the
third portion of the duodenum was not found at fiberoptic duodenoscopy because the examination was
terminated at the bulb.208(4746) I have seen one case in which attempts to trap the stone in a Dormia
basket for withdrawal were unsuccessful (Figure 4923).203(4747) Disimpaction usually requires
surgical duodenotomy.209(4748) Attempts to remove an impacted stone by endoscopic means have
met with variable success.210,211(4749)
(4750)Figure 4923. Gallstone impacted in the descending duodenum. This stone could not
be trapped in a Dormia basket. (Courtesy of Dr. M. Ibrahim.)
Duodenocolic Fistula
Duodenocolic fistulas most often represent erosion from carcinoma of the hepatic flexure of the
colon212,213(4751) but may also be caused by carcinoma of the gallbladder or duodenum.214(4752)
Less frequently, benign conditions such as peptic ulcer disease, diverticulitis, and granulomatous or
ulcerative colitis may cause such fistulas. The diagnosis is most often confirmed by barium enema
studies; upper gastrointestinal endoscopy ordinarily plays no role. Ergin et al.215(4753) observed
endoscopically a necrotic mass with central ulceration in the duodenum in a patient with a fistulous
communication between the second portion and the hepatic flexure. However, the communication had
been documented by barium enema and upper gastrointestinal series prior to endoscopy.
Nevertheless, endoscopic biopsy revealed adenocarcinoma.
Aortoenteric Fistula
About 80% of aortoenteric fistulas communicate with the duodenum. Because the third portion is
relatively fixed retroperitonally between the abdominal aorta posteriorly and the mesenteric artery and
vein anteriorly, it is subject to pressure necrosis from an infrarenal aneurysm (primary type of
aortoenteric fistula) or from an aortic graft (secondary type).216(4754) An extensive review published
in 1980 identified 186 patients with primary aortoenteric fistulas.217(4755) The diagnosis of a
secondary-type aortoduodenal fistula can be made endoscopically if a bile-stained vascular prosthesis
is recognized, a pulsatile mass is appreciated, or arterial bleeding is encountered in the second or third
part of the duodenum.218(4756) Walsh et al.107(4757) and Perdue et al.219(4758) stressed the
importance of recognition of the early signs of back pain, fever, and intermittent bleeding in alerting the
physician to the possibility of impending aortoenteric hemorrhage. Walsh et al. noted that the gallium
scan and CT were each positive in all five patients examined,107(4759) whereas endoscopy was
diagnostic in only three of eight cases reported by Perdue et al.219(4760) It is mandatory that
endoscopy include examination of the third and fourth portions of the duodenum; otherwise, the
diagnosis of aortoduodenal fistula will not be made.216(4761) Some surgeons believe that upper
gastrointestinal endoscopy is needed only to exclude another cause of hemorrhage in a patient with a
known aortic prosthesis and that visualization of the fistula should not even be attempted.220(4762) It
is difficult, however, to discourage definitive diagnosis if the endoscopist sees blood in the distal
duodenum.
Champion et al.221(4763) insisted that endoscopy be repeated as often as necessary to establish the
diagnosis of aortoduodenal fistula. However, significant upper gastrointestinal hemorrhage and the
absence at endoscopy of other potential causes in a patient with an aortic graft can be reason enough
for surgical exploration because exsanguination is a distinct possibility in patients with this condition.
Many reports of endoscopic diagnosis of aortoduodenal fistulas have been published since
1973.222232(4764) In some cases, endoscopic diagnosis was made by recognition of the
bile-stained graft (Figure 4924),221,222,225,227,229(4765) a late stage of pressure necrosis, but in
most instances, the lesion appeared as a smooth mass that was either ulcerated or pulsatile, or
both.223226,228(4766) The aortoenteric fistula is considered further in Chapter 26: Acute Upper
Gastrointestinal Bleeding.
(4767)Figure 4924. Aortoduodenal fistula. The bile-stained, ribbed prosthesis is readily

identifiable.
Two cases have been reported in which a graft constructed for a mesocaval shunt for control of
esophageal variceal bleeding was seen protruding into the second portion of the duodenum at
endoscopy.233,234(4768)
Celiac Disease
Stevens and McCarthy235(4769) concluded in 1977 that endoscopic biopsy of the duodenal bulb
excludes celiac disease if villi are present. Villi were recognizable endoscopically in 25 controls, but not
in 11 untreated patients with celiac disease; spraying indigo carmine on the duodenal mucosa
exaggerated the abnormal mosaic pattern for directed biopsy.235(4770) Gillberg and Ahren236(4771)
found that biopsies of the first and second parts of the duodenum in 14 patients with celiac disease
correlated well with proximal jejunal suction biopsies. In another study, 27 patients with suspected
celiac disease had duodenoscopic biopsies; 11 biopsies considered compatible with celiac disease
were confirmed by suction jejunal biopsy, and 13 normal biopsies were also verified by subsequent
suction biopsy.237(4772) Other comparative studies confirmed the accuracy and reliability of
duodenoscopic biopsy for the diagnosis of villous atrophy.238,239(4773) Although specimens may be
somewhat smaller, duodenoscopic biopsy offers several advantages compared with the capsule
method: multiple specimens can be obtained and sampling can be directed at mucosal areas that
appear abnormal.
Biopsies taken with the smaller forceps used with the smaller-diameter endoscopes are not as easy to
interpret with respect to celiac disease. The mucosal abnormality in celiac disease, being patchy and
variable, requires, moreover, several biopsies to ensure a reliable diagnosis, although it appears that
the more severe changes are found twice as often in the duodenum as in the jejunum (Figure
4925).240(4774) Nevertheless, endoscopic biopsy is the most dependable, safe, and expeditious way
to obtain duodenal tissue, particularly in children.241,242(4775) Ladas et al.243(4776) compared
paired duodenal biopsy specimens obtained with a "standard" forceps with those obtained with a large,
so-called jumbo forceps of the type that requires use of a large-channel instrument. The latter
instrument provided specimens that were significantly larger and longer. When the large specimens
were oriented by means of stereomicroscopy, 72% contained at least four villi in a row compared with
39% obtained with the standard forceps and oriented without a microscope.
(4777)Figure 4925. A-C, Severe celiac disease with involvement of the distal duodenum.
(A-C, From the collection of Dr. Michael V. Sivak, Jr.)
Duodenal biopsies have been recommended to exclude subclinical celiac sprue in any patient whose
duodenum shows the endoscopic sign of scalloped valvulae conniventes244,245(4778) or,
contrariwise, absent duodenal folds.246(4779) In a study that compared the upper gastrointestinal
endoscopic findings in 46 children with celiac disease in various stages with those in 27 children with
other conditions, Corazza et al.245(4780) found that scalloped duodenal folds had a sensitivity of 88%
and a specificity of 87% with respect to the diagnosis of subtotal villous atrophy. Maurino et
al.247(4781) prospectively evaluated 100 patients referred for endoscopic duodenal biopsy. Of the 36
patients found to have celiac disease on biopsy, 34 had endoscopic findings including loss or reduction
in number of the valvulae conniventes (27 patients), a mucosal mosaic pattern (14 patients), scalloped
folds (12 patients), and visible submucosal blood vessels (5 patients). For the diagnosis of celiac
disease, these endoscopic findings had a sensitivity of 94%, a specificity of 92%, and a positive
predictive value of 84%. Loss or reduction in number of the valvulae conniventes was the most
sensitive (76%) and specific (98%) single endoscopic finding. Maurino et al.247(4782) also made a
duodenoscopic diagnosis of celiac disease in 3 of 4 asymptomatic patients in a group of 24 first-degree
relatives of patients with celiac disease who underwent EGD. In a similar study by McIntyre et
al.,248(4783) 11 of 15 patients with celiac disease had reduced or absent valvular folds compared with
2 of 60 patients who did not have celiac disease (sensitivity 73%, specificity 97%, positive predictive
value 85%).
Iron deficiency anemia is often the indication for EGD. On occasion, this may be the initial and only
clinical manifestation of celiac disease.249,250(4784) Patients with celiac disease are at increased risk
for the development of small intestinal lymphoma; adenocarcinoma of the small intestine has also been
reported.251(4785)
Whipple's Disease
Duodenoscopic study of patients with untreated Whipple's disease reveals in most cases a typical
pattern of thickened mucosa with discrete or confluent yellow-white granules between the folds (Figure
4926).252,253(4786) These pinhead-size projections actually represent distended villi. In the study of
Geboes et al.,254(4787) these characteristic postbulbar duodenal lesions were found in 9 of 12
patients. The mucosa is also friable255(4788) or has areas of petechial hemorrhage.253(4789)
Biopsies correlate well with duodenoscopic findings because the granular mucosa shows distended villi
packed with periodic acid-Schiff (PAS)-positive macrophages, whereas the more normal-appearing
mucosa has slender villi and submucosal PAS-positive macrophages.252,253(4790) After treatment,
both duodenoscopic and histologic evaluations are essentially normal.256(4791) A case can be made
for duodenoscopic biopsy as the diagnostic method of choice because duodenal and jejunal
involvement by Whipple's disease may be patchy and the single biopsy obtained with the blind suction
biopsy capsule may fail to retrieve tissue containing PAS-positive macrophages. At duodenoscopy,
several biopsies of abnormal areas should result in a high percentage of positive findings.
(4792)Figure 4926. Whipple's disease involving the distal duodenum. (From the collection
of Dr. Michael V. Sivak, Jr.)
Nodular Lymphoid Hyperplasia

Patients with acquired hypogammaglobulinemia may develop a nodular mucosal pattern of the small
bowel. These nodules are the result of lymphoid hyperplasia, which represents a response to
deficiency or absence of IgA (Figure 4927). Although the nodules are seen predominantly in the
jejunum or ileum, on one occasion they were found endoscopically in the descending duodenum,
where they appeared as 1- to 3-mm, soft, polypoid excrescences covering the mucosa.257(4793) The
biopsy specimens showed lymphoid hyperplasia, and aspirates from the duodenum revealed Giardia
lamblia, a commonly associated infection.
(4794)Figure 4927. A, Nodular lymphoid hyperplasia of the duodenum in a patient with

immunoglobulin A deficiency. B, Close-up view of the mucosal nodules. (A and B, From the
Pseudomelanosis
A patchy or spotty black discoloration of the duodenal mucosa, usually termed pseudomelanosis, has
been reported in at least 11 patients.258265(4795) The pigment is found mainly in the lysosomes of
macrophages in the lamina propria.263(4796) At first, these pigment granules were thought to be
melanin pigment,257(4797) but electron microscopy and electron probe x-ray analysis established that
they are not melanin but iron sulfide261,262(4798) or a variety of iron and sulfide
compounds.266,267(4799)
Amyloidosis
The duodenum is the most common site for deposition of amyloid in the gastrointestinal tract. The
duodenal mucosal appearance, predominantly in the second portion, is that of a fine granularity,
although erosions, ulcers, and polyps may also be seen.268(4800) Tada et al.269(4801) found a
correlation between duodenoscopic findings and the accumulation pattern of specific amyloid proteins
in a study of 30 patients with amyloidosis of the small intestine. A fine granular appearance was found
significantly more frequently where there was deposition of amyloid A protein, whereas accumulation of
light chain protein correlated with the presence of multiple polypoid protrusions and thickening of the
valvulae conniventes.
REFERENCES
1.
2.
Schuman BM. Beyond the bulb. Gastrointest Endosc 1985;31:4023.

Frandsen PJ, Jarnum S, Malstrom J. Crohn's disease of the duodenum. Scand J
3. Nugent FW, Richmond M, Park SK. Crohn's disease of the duodenum. Gut 1977;18:11520.
4. Rutgeerts P, Onette E, Vantrappen G, et al. Crohn's disease of the stomach and duodenum:
A clinical study with emphasis on the value of endoscopy and endoscopic biopsies.
Endoscopy 1980;12:28894.
5. Danzi JT, Farmer RG, Sullivan BH Jr, Rankin GB. Endoscopic features of gastroduodenal
Crohn's disease. Gastroenterology 1976;70:913.
6. Murthy UK. Repeated hydrostatic balloon dilation in obstructive gastroduodenal Crohn's
disease. Gastrointest Endosc 1991;37:4845.
7. Nugent FW, Roy MA. Duodenal Crohn's disease: An analysis of 89 cases. Am J
8. Gad A. The diagnosis of gastroduodenal Crohn's disease by endoscopic biopsy. Scand J
9.
Greenberger N. Allergic disorders of the intestine and eosinophilic gastroenteritis. In
Sleisenger MH, Fordtran JS (eds). Gastrointestinal Disease: Pathophysiology, Diagnosis and
Management. Philadelphia: WB Saunders, 1983;106976.
10. Robert F, Omura E, Durant JR. Mucosal eosinophilic gastroenteritis with systemic
involvement. Am J Med 1977;62:13943.
11. Lev R, Thomas E, Parl FF, et al. Pathological and histomorphometric study of the effects of
alcohol on the human duodenum. Digestion 1980;20:20713.
12. Piubello W, Vantini I, Scuro LA, et al. Gastric secretion, gastroduodenal histological changes,
and serum gastrin in chronic alcoholic pancreatitis. Am J Gastroenterol 1982;77:10510.
13. Makrauer FL, Antoniali DA, Banks PA. Duodenal stenosis in chronic pancreatitis.
Clinicopathological correlation. Dig Dis Sci 1982;27:52532.
14. Bradley EL III, Clements JL Jr. Idiopathic duodenal obstruction: An unappreciated
complication of pancreatitis. Ann Surg 1981;193:63848.
15.
Grodsinsky G, Schuman BM, Black MA. Absence of pancreatic duct dilation in chronic
pancreatitis. Arch Surg 1976;112:4449.
16.
Black GA, Carsky EW. Duodenal tuberculosis. Am J Roentgenol 1978;131:32930.
17. Tandon RK, Pastakia B. Duodenal tuberculosis as seen by duodenoscopy. Am J
18. Tishler JM. Duodenal tuberculosis. Radiology 1979;130:5935.

19. Gleason T, Prinz RA, Kirsch EP, et al. Tuberculosis of the duodenum. Am J Gastroenterol
1979;72:3640.
20. Vij JC, Ramesh GN, Choudhary V, Malhotra V. Endoscopic balloon dilation of tuberculous
duodenal strictures. Gastrointest Endosc 1992;38:5101.
21. Monroe LS. The endoscopic encounter with parasites. Gastrointest Endosc 1984;30:1134.
22. Bosch O, Porres JC, Martinez Quesada G, et al. Endoscopic appearance of a duodenal
infection by Mycobacterium avium-intracellulare in AIDS. Endoscopy 1994;26:506.
23. Monsour HP Jr, Quigley EM, Markin RS, et al. Endoscopy in the diagnosis of gastrointestinal
Mycobacterium avium-intracellulare infection. J Clin Gastroenterol 1991;13:204.
24. Vazquez-Iglesias JL, Yanez J, Durana J, Arnal F. Infection by Mycobacterium
avium-intracellulare in AIDS: Endoscopic duodenal appearance mimicking Whipple's disease.
Endoscopy 1988;20:27980.
25. Brandborg LL, Owens R, Fogel R, et al. Giardiasis and traveler's diarrhea. Gastroenterology
1980;78:160214.
26. McHenry R, Bartlett MS, Lehman GA, O'Connor KW. The yield of routine duodenal aspiration
for Giardia lamblia during esophagogastroduodenoscopy. Gastrointest Endosc
1987;33:4256.
27. Ramirez FC, Clarridge JE, Heiser MC, et al. A study of the frequency of recovery of
unexpected Giardia lamblia and Cryptosporidium from duodenal aspirates taken during
routine upper endoscopy. Am J Gastroenterol 1993;88:5524.
28. Bhatt BD, Cappell MS, Smilow PC, Das KM. Recurrent massive upper gastrointestinal
hemorrhage due to Strongyloides stercoralis infection. Am J Gastroenterol 1990;85:10346.
29. Bone MF, Chesner IM, Oliver B, et al. Endoscopic appearance of duodenitis due to
strongyloidiasis. Gastrointest Endosc 1982;28:1901.
30. Peralta NR, Rodrigues MA. Strongyloides stercoralis larva in gastric and duodenal aspirates.
Acta Cytol 1978;22:613.
31. Chen CH, Wong JM, Hwang YT, et al. Biliary ascariasis: Report of a case. J Formos Med
Assoc 1994;93:43790.
32. Winters C Jr, Chobanian SJ, Benjamin SB, et al. Endoscopic documentation of
Ascaris-induced acute pancreatitis. Gastrointest Endosc 1984;30:834.
33.
Jager G, Voigtsberger P, Bohm S. Endoscopic removal of a tapeworm. Z Arztl Fortbild
1978;72:337.
34. Descombes P, Dupas JL, Capron JP. Endoscopic discovery and capture of Taenia saginata.
35. Hauser SC, Bynum TE. Abnormalities on ERCP in a case of human fascioliasis. Gastrointest
Endosc 1984;30:802.
36. Reddy RR, Schuman BM, Priest RJ. Duodenal polyps: Diagnosis and management. J Clin
37. Herbsman H, Wetstein L, Rosen Y, et al. Tumors of the small intestine. Curr Probl Surg
1980;17:12182.
38. Matsuura H, Kuwano H, Kanematsu T, et al. Clinicopathological features of elevated lesions
of the duodenal bulb. J Surg Oncol 1990;45:7984.
39. Cappell MS. Profound spatial clustering of polyps in individuals with multiple nonmalignant
polyps of the stomach and duodenal bulb: A combined endoscopic and radiographic study.
40. Matsui K, Kitagawa M. Biopsy study of polyps in the duodenal bulb. Am J Gastroenterol
1993;88:2537.
41. Jepsen JM, Persson M, Jakobsen NO, et al. Prospective study of prevalence and endoscopic
and histopathologic characteristics of duodenal polyps in patients submitted to upper
endoscopy. Scand J Gastroenterol 1994;29:4837.
42. Borsch G, Ernst R, Saedi F. Excluded gastric mucosa presenting as a polypoid tumor in the
duodenal stump after B II resection. Z Gastroenterol 1983;21:5934.
43. Thatcher BS, Fleischer D, Rankin GB, et al. Duodenal schistosomiasis diagnosed by
endoscopic biopsy of an isolated polyp. Am J Gastroenterol 1984;79:9279.
44. Finch PJ, Haddock J, Grundy A. Duodenal pseudotumor: A silent impacted bile duct calculus.
45. Bar-Meir S, Hallak A, Baratz M. Endoscopic removal of a giant duodenal polyp. Endoscopy
1983;15:2930.
46. Cooperman M, Clausen KP, Hecht C, et al. Villous adenomas of the duodenum.
47. Blackman E, Nash SV. Diagnosis of duodenal and ampullary epithelial neoplasms by
endoscopic biopsy: A clinicopathologic and immunohistochemical study. Hum Pathol
1985;16:90110.
48. Chappuis CW, Divincenti FC, Cohn I Jr. Villous tumors of the duodenum. Ann Surg
1989;209:5938; discussion 5989.
49. Batra SK, Schuman BM, Reddy RR. The endoscopic variety of duodenal villous
adenomaAn experience with ten cases. Endoscopy 1983;15:8992.
50. Delpy JC, Bruneton JN, Drouillard J, Lecomte P. Nonvaterian duodenal adenomas: Report of
24 cases and review of the literature. Gastrointest Radiol 1983;8:13541.
51. Galandiuk S, Hermann RE, Jagelman DG, et al. Villous tumors of the duodenum. Ann Surg
1988;207:2349.
52. Bjork KJ, Davis CJ, Nagorney DM, Mucha P Jr. Duodenal villous tumors. Arch Surg
1990;125:9615.
53. Witteman BJ, Janssens AR, Griffioen G, Lamers CB. Villous tumours of the duodenum. An
analysis of the literature with emphasis on malignant transformation. Neth J Med
1993;42:511.
54. Ghilain JM, Dive C. Endoscopic laser therapy for small villous adenomas of the duodenum.
Endoscopy 1994;26:30810.
55. Paraf F, Naveau S, Zourabichvili O, et al. Endoscopic laser therapy for duodenal villous
adenoma. Dig Dis Sci 1989;34:14667.
56. Wilson JM, Melvin DB, Gray G, et al. Benign small bowel tumors. Ann Surg 1975;181:24750.
57. Anderson JR, Ford MJ. Leiomyoma of the duodenumAn unusual presentation of a rare
tumor. Postgrad Med J 1979;55:21820.
58. Morris SJ, Shifrin H, Feinberg A, et al. Duodenal leiomyoma as a cause of gastrointestinal
hemorrhage. South Med J 1978;71:4703.
59. Serraf A, Klein E, Schneebaum S, et al. Leiomyomas of the duodenum. J Surg Oncol
1988;39:1836.
60. Farkas I, Patko A, Kovacs L, et al. The brunneroma, the adenomatous hyperplasia of the
Brunner's glands. Acta Gastroenterol Belg 1980;43:17986.
61. Franzin G, Musola R, Ghidini O, et al. Nodular hyperplasia of Brunner's glands. Gastrointest
Endosc 1985;31:3748.
62.
Maratka Z, Kocianova J, Kudrmann J, et al. Hyperplasia of Brunner's glands. Radiology,
endoscopy and biopsy findings in 11 cases of diffuse, nodular and adenomatous form. Acta
Hepatogastroenterol 1979;26:649.
63. Maglinte DD, Mayes SL, Ng AC, et al. Brunner's gland adenoma: Diagnostic considerations. J
64. Appel MF, Bentlif PS. Endoscopic removal of bleeding Brunner gland adenoma. Arch Surg
1976;111:3012.
65. Barnart GR, Maull KI. Brunner's gland adenomas: Clinical presentation and surgical
management. South Med J 1979;72:15379.
66. Alper IA, Haubrich WS. Duodenoscopic removal of a Brunner's gland adenoma. Gastrointest
Endosc 1973;20:73.
67. Kawaja HT, Deakin M, Colin-Jones DG. Endoscopic removal of a large ulcerated Brunner's
gland adenoma. Endoscopy 1986;18:199201.
68. Kehl O, Buhler H, Stamm B, et al. Endoscopic removal of a large obstructing and bleeding
duodenal Brunner's gland adenoma. Endoscopy 1985;17:2312.

69. Matsui T, Iida M, Fujischima M, et al. Brunner's gland hamartoma associated with
microcarcinoids. Endoscopy 1989;21:378.
70. Matsumoto T, Iida M, Matsui T, et al. A large Brunner's gland adenoma removed by
endoscopic polypectomy. Endoscopy 1990;22:1923.
71. Michel LA, Ballet T, Collard JM, et al. Severe bleeding from submucosal lipoma of the
duodenum. J Clin Gastroenterol 1988;10:5415.
72. Whetstone MR, Zuckerman MJ, Saltzstein EC, Boman D. CT diagnosis of duodenal lipoma.
73. Imamura K, Fuchigami T, Iida M, et al. Duodenal lipomaA report of three cases.
74. Sivak MV Jr, Sullivan BH Jr, Farmer RG. Neurogenic tumors of the small intestine: Review of
the literature and report of a case with endoscopic removal. Gastroenterology
1975;68:37489.
75. Rutgeerts P, Hendricks H, Geboes K, et al. Involvement of the upper digestive tract by
systemic neurofibromatosis. Gastrointest Endosc 1981;27:225.
76.
Sauerbruch T, Keiditsch E, Wotzka R, et al. Lymphangioma of the duodenum. (Diagnosis by
endoscopic resection.) Endoscopy. 9:179821977.
77. Salata HH, Mercader J, Navarro A, et al. Lymphangioma of the duodenum. Endoscopy
1984;16:302.
78. Plavsic B, Jereb-Provic B. Radiologic and endoscopic diagnosis of duodenal angioma. Acta
Radiol 1987;28:7358.
79. Gangl A, Polterauer P, Krepler R, Kumpan W. A further case of submucosal lymphangioma of
the duodenum diagnosed during endoscopy. Endoscopy 1980;12:18890.
80. Aneiros J, Pleguezuelos J, Garcia del Moral R, et al. Lymphangioma of the duodenum: An
ultrastructural study. Endoscopy 1986;18:2458.
81. Laughlin EH, Keown ME, Jackson JE. Heterotopic pancreas obstructing the ampulla of Vater.
Arch Surg 1983;118:97980.
82. Sivak MV Jr, Jagelman DG. Upper gastrointestinal endoscopy in polyposis syndromes:
Familial polyposis coli and Gardner's syndrome. Gastrointest Endosc 1984;30:1024.
83. Yao T, Iida M, Ohsato K, et al. Duodenal lesions in familial polyposis of the colon.
84. Gahtan V, Nochomovitz LE, Robinson AM, et al. Gastroduodenal polyps in familial polyposis
coli. Am Surg 1989;55:27880.
86. Jarvinen HJ, Sipponen P. Gastroduodenal polyps in familial adenomatous and juvenile
polyposis. Endoscopy 1986;18:2304.
87. Iida M, Yao T, Itoh H, et al. Endoscopic features of adenoma of the duodenal papilla in familial
polyposis of the colon. Gastrointest Endosc 1981;27:68.
88. van Stolk R, Sivak MV Jr, Petrini JL, et al. Endoscopic management of upper gastrointestinal
polyps and periampullary lesions in familial adenomatous polyposis and Gardner's syndrome.
89. Shemesh E, Bat L. A prospective evaluation of the upper gastrointestinal tract and
periampullary region in patients with Gardner syndrome. Am J Gastroenterol 1985;80:8257.
90. Schulman A. Gastric and small bowel polyps in Gardner's syndrome and familial polyposis
coli. J Can Assoc Radiol 1976;27:2069.
91. Sweeney BF, Anderson DS. Endoscopic removal of duodenal polyp in a patient with
Gardner's syndrome. Dig Dis Sci 1982;27:55760.
92. Kurtz RC, Sternberg SS, Miller HH, Decosse JJ. Upper gastrointestinal neoplasia in familial
polyposis. Dig Dis Sci 1987;32:45965.
93. Kotzampassi K, Eleftheriadis E, Zissiadis A, Aletras H. Duodenal polyposis in Gardner's
syndrome: The significance of endoscopic surveillance. Am J Gastroenterol 1990;85:15412.

94. Offerhaus GJA, Giardiello FM, Krush AJ, et al. The risk of upper gastrointestinal cancer in
familial adenomatous polyposis. Gastroenterology 1992;102:19802.
95. Spigelman AD, Williams CB, Talbot IC, et al. Upper gastrointestinal cancer in patients with
familial adenomatous polyposis. Lancet 1989;2:7835.
96. Iida M, Yao T, Itoh H, et al. Natural history of duodenal lesions in Japanese patients with
familial adenomatosis coli (Gardner's syndrome). Gastroenterology 1989;96:13016.
97. Russell DM, Bhathal PS, St. John DJ. Complete remission in Cronkhite-Canada syndrome.
98. Fujita S, Kusunoki M, Sakanoue Y, et al. Preoperative endoscopy in Peutz-Jeghers
syndrome. Lancet 1990;335:4156.
99. Petroianu A, De Carvalho DG. Perioperative endoscopy in Peutz-Jeghers syndrome. Lancet
1989;2:403.
100. Lillemoe K, Imbembo AL. Malignant neoplasms of the duodenum. Surg Gynecol Obstet
1980;150:8226.
101. Kerremans RP, Lerut J, Pennincks FM. Primary malignant duodenal tumors. Ann Surg
1979;190:17982.
102. Spira IA, Ghazi A, Wolff WI. Primary adenocarcinoma of the duodenum. Cancer
1977;39:17216.
103. Langer B, Lipson R, McHattie JD, et al. Periampullary tumors: Advances in diagnosis and
surgical treatment. Can J Surg 1979;22:347.
104. van Ooijen B, Kalsbeek HL. Carcinoma of the duodenum. Surg Gynecol Obstet
1988;166:3437.
105. King AJ, McCabe J, Waldron M, et al. Combined endoscopic cytology and biopsies in the
diagnosis of duodenal malignancy. Br J Clin Pract 1990;44:2645.
106. Obata S, Suenaga M, Araki K, et al. Use of strip biopsy in a case of early duodenal cancer.
Endoscopy 1992;24:2324.
107. Walsh DB, Eckhauser FE, Cronewett JL, et al. Adenocarcinoma of the ampulla of Vater:
Diagnosis and treatment. Ann Surg 1982;195:1527.
108. Hauswald KR, Griffen WO. Smooth muscle tumors of the duodenum. Rev Surg
1977;34:647.
109. Sato A, Hikosaka O, Koike Y, et al. Diagnostic and therapeutic considerations of myogenic
tumors of the duodenum. Gastroenterol Jpn 1978;13:6571.
110. Wong J, Stephen MS, Ong GB, et al. Leiomyosarcoma of the duodenum: A report of three
cases. Aust N Z J Surg 1977;47:7938.
111. Barkan A, Wolloch Y, Dintsman M, et al. Leiomyosarcoma of the duodenum. Two case
reports and a literature review. Am J Proctol Gastroenterol Colon Rectal Surg 1981;32:1821,
28.
112. Yassinger S, Imperato TJ, Midgley R Jr, et al. Leiomyosarcoma of the duodenum.
113. Helmy I. Endoscopic diagnosis of immunoproliferative small intestinal disease (IPSID).
Endoscopy 1980;12:1146.
114.
Kobler E, Buhler H, Fehr H, et al. Primary malignant lymphoma of the duodenum. Praxis
1982;71:9247.
115. Payson BA, Weingarten LA, Pollack J. Lymphosarcoma of the duodenum associated with
carcinoma of the lung. Am J Gastroenterol 1979;71:295300.
116. Veen HF, Oscarson JE, Malt RA. Alien cancers of the duodenum. Surg Gynecol Obstet
1976;143:3942.
117. Bungay K, Dennistone S, Hunt PS. Duodenal obstruction and carcinoma of the head of the
pancreas. Med J Aust 1980;2:1501.
118. Sharon P, Stalnikovicz R, Rachmilewitz D. Endoscopic diagnosis of duodenal neoplasms
causing upper gastrointestinal bleeding. J Clin Gastroenterol 1982;4:358.
119. Ashkenazi S, Sharom P, Levij IS, et al. Giant cell carcinoma of the pancreas: Report of a case
with upper gastrointestinal bleeding diagnosed by endoscopic brush cytology. Am J

120. Cronstedt JL, Kalczynski J, Jonsson NGE. Involvement of the duodenum by gastric
carcinoma. Gastrointest Endosc 1982;28:445.
121. Theodors A, Sivak MV Jr, Carey WD. Hypernephroma with metastasis to the duodenum:
Endoscopic features. Gastrointest Endosc 1980;26:4850.
122. Sivak MV Jr, Sullivan BH Jr. Endoscopic diagnosis of malignant melanoma metastatic to the
duodenum. Gastrointest Endosc 1975;22:368.
123. Coughlin GP, Bourne AJ, Grant AK. Endoscopic diagnosis of metastatic disease of the
stomach and duodenum. Aust N Z J Med 1977;7:525.
124.
Lammli J, Buhler H, Bosseckert H, et al. Metastasen im duodenum. Praxis 1982;71:10547.
125. Gurian L, Ireland K, Petty W, Katon R. Carcinoma of the cervix involving the duodenum: Case
report and review of the literature. J Clin Gastroenterol 1981;3:2914.
126. Chen LT, Chen CY, Jan CM, et al. Gastrointestinal tract involvement in hepatocellular
carcinoma: Clinical, radiological and endoscopic studies. Endoscopy 1990;22:11823.
127. Humbert P, Sarmiento J, Boix J, et al. Hepatocellular carcinoma presenting with bleeding due
to duodenal perforation by the tumor. Endoscopy 1987;19:378.
128. Borsch G, Schmidt G, Sperling J, et al. Gastroduodenal metastasesAn unusual initial
manifestation of bronchogenic carcinoma. Endoscopy 1984;16:11821.
129. Mannell A. Primary isolated extramedullary plasmacytoma of the duodenum. Aust N Z J Surg
1979;49:57780.
130. Ahmed N, Nelson RS, Goldstein HM, et al. Kaposi's sarcoma of the stomach and duodenum:
Endoscopic and roentgenologic correlations. Gastrointest Endosc 1975;21:14952.
131. Safai B, Johnson KG, Myskowski PL, et al. The natural history of Kaposi's sarcoma in the
acquired immunodeficiency syndrome. Ann Intern Med 1985;103:74450.
132. Merrell DE, Mansbach C, Garbutt JT Jr. Carcinoid tumors of the duodenum: Endoscopic
diagnosis of two cases. Gastrointest Endosc 1985;31:26971.
133. Hirakawa K, Iida M, Matsui T, et al. Endoscopic findings in carcinoid tumor of the duodenum.
134. Gencsi E, Lux E, Kaduk B, et al. Upper gastrointestinal bleeding as an unusual presentation
of a duodenal carcinoid. Endoscopy 1986;18:1057.
135. Wengrower D, Fich A. Primary duodenal carcinoid. Am J Gastroenterol 1987;82:106970.
136. Donovan DC, Dureza R, Jain U. Gastrinoma of duodenum. Diagnosis by endoscopy. N Y
State J Med 1979;79:17668.
137. Wu WC, Kengis J, Whalen GE, et al. Endoscopic localization of a wall tumor in
Zollinger-Ellison syndrome. Gastroenterology 1974;66:12379.
138. Wadas DD, Foutch PG, Manne RK, Sanowski RA. Endoscopic diagnosis of a duodenal
gastrinoma. Gastrointest Endosc 1988;34:4301.
139. Otten MH, Berkenhager JC, van Blankenstein M. Zollinger-Ellison syndrome treated by
endoscopic removal of a duodenal gastrinoma. Neth J Med 1978;21:24851.
140. Gilhool WJ. Endoscopic diagnosis and removal of a duodenal wall gastrinoma. Am J
141. Rolny P, Ekman R, Sjolund K, et al. Endoscopic diagnosis and treatment of duodenal
gastrinoma. Gastrointest Endosc 1990;36:5113.
142. Frucht H, Norton JA, London JF, et al. Detection of duodenal gastrinomas by operative
endoscopic transillumination. A prospective study. Gastroenterology 1990;99:16227.
143. Wiest JW, Follette DM, Traverso LW. Toothpick perforation of the duodenum. West J Med
1980;132:1579.
144. Honaas TO, Shaffer EA. Endoscopic removal of a foreign body perforating the duodenum.
Can Med Assoc J 1977;116:1649.
145. Schwartz JT, Graham DY. Toothpick perforation of the intestines. Ann Surg 1977;185:646.
146. Thompson MH. Endoscopic removal of a duodenal foreign body. Br Med J 1975;4:5023.
147. Anderson FH. Removal of a safety pin from the duodenum with the fiberoptic endoscope. Am
148. Kerlin P, Paull A. Removal of duodenal foreign body with endoscopic snare. Med J Aust
1978;2:2767.
149. Weese JL, Wissler DW, Magary JA, Ramirez G. Endoscopic transduodenal removal of an
hepatic artery infusion catheter. Gastrointest Endosc 1987;33:2467.
150. Joseph AE, Crampton AR, Agha FP, et al. Impacted foreign bodies in the duodenum. Am J
151. Madura MJ, Naughton BJ, Craig RM. Duodenal bezoar: A case report and review of the
literature. Gastrointest Endosc 1982;28:268.
152. Suarez LA, Bolden EI. Congenital duodenal web in an adult. Curr Surg 1978;35:3669.
153. Turnbull A, Kussin S, Bains M. Radiographic and endoscopic features of a congenital
duodenal diaphragm in an adult. A case report and review of the literature. Gastrointest
Endosc 1980;26:468.
154. Leinkram C, Roberts-Thomson IC, Kume GA. Juxtapapillary duodenal diverticula. Association
with gallstones and pancreatitis. Med J Aust 1980;1:20910.
155. van der Spuy S. The relationship between juxtapapillary diverticula and biliary calculi. An
endoscopic study. Endoscopy 1979;11:197202.
156. Cox CL. Perforated duodenal diverticulitis. South Med J 1980;73:830.
157. Franceschi D, Castillo W, Yuh JN, Chen PF. Endoscopic diagnosis of upper gastrointestinal
tract bleeding from a duodenal diverticulum. Gastrointest Endosc 1987;33:3834.
158. Lapin R, Kamath ML, Engler J, Friedman H. Massive gastrointestinal hemorrhage from
duodenal diverticula. Am J Gastroenterol 1974;61:1859.
159. Widmann M, Schuman B, Poznanski A. Intraluminal duodenal diverticulum. Gastroenterology
1966;50:57881.
160. Wobser EC, Koischwitz D, Katz K, et al. Radiographical and endoscopical studies on two
cases of intraluminal diverticulum of the duodenum. Endoscopy 1977;8:1015.
161. Karoll MP, Ghahremani GG, Port RB, Rosenberg JL. Diagnosis and management of
intraluminal duodenal diverticulum. Dig Dis Sci 1983;28:4116.
162. Hajiro K, Yamamoto H, Matsui HJ, et al. Endoscopic diagnosis and excision of intraluminal
duodenal diverticulum. Gastrointest Endosc 1979;25:1514.
163. Hiraoka T, Nakamura M, Ohno K, et al. Endoscopic excision of intraluminal duodenal
diverticulum. Dig Dis Sci 1985;30:27481.
164. Ravi J, Joson PM, Ashok PS. Endoscopic incision of intraluminal duodenal diverticulum. Case
report of a new technique. Dig Dis Sci 1993;38:7626.
165. Cooperman AM, Adachi M, Rankin GB, Sivak MV Jr. Congenital duodenal diaphragms in
adults: A delayed cause of intestinal obstruction. Ann Surg 1975;182:73942.
166. Ahmed MY, Kepkay DL, Beck IT, McCorriston JR. Congenital duodenal diaphragm in an
adult: Diagnosis by duodenoscopy. Gastrointest Endosc 1972;18:1201.
167. al-Kawas FH. Management of a duodenal web by endoscopic laser therapy. Gastrointest
Endosc 1989;35:1135.
168. Gertsch P, Mosimann R. Endoscopic laser treatment of a congenital duodenal diaphragm in
an adult. Gastrointest Endosc 1984;30:2534.
169. Jex RK, Hughes RW Jr. Endoscopic management of duodenal diaphragm in the adult.
170. Kay GA, Lobe TE, Custer MD, Hollabaugh RS. Endoscopic laser ablation of obstructing
congenital duodenal webs in the newborn: A case report of limited success with criteria for
patient selection. J Pediatr Surg 1992;27:27981.
171. Halko M, Lo Presti PA, Ratel HD, et al. Congenital duodenal stenosis. Am J Gastroenterol
1978;69:3237.
172. Duque JJ, Strum WB, Kuster GG, Kaufmann D. Upper gastrointestinal bleeding from a
duodenal duplication: An unusual complication of a rare disease. Gastrointest Endosc
1989;35:4712.
173. Holcomb GW III, Gheissari A, O'Neill JA Jr, et al. Surgical management of alimentary tract
duplications. Ann Surg 1989;209:16774.

174. Bulajic M, Savic-Perisic M, Korneti V, et al. Use of endoscopy to diagnose symptomatic
duodenal duplication cyst in an adult. Endoscopy 1991;23:2346.
175. al Traif I, Khan MH. Endoscopic drainage of a duodenal duplication cyst. Gastrointest Endosc
1992;38:645.
176. Johanson JF, Geenen JE, Hogan WJ, Huibregtse K. Endoscopic therapy of a duodenal
duplication cyst. Gastrointest Endosc 1992;38:604.
177. Dave P, Romeu J, Clary S, et al. Endoscopic removal of an obstructing duodenal duplication
cyst. Endoscopy 1984;16:756.
178. Al-Mofarreh M, Al-Moagel-Alfarag M, Ashoor T, Shadoochy F. Duodenal varices. Report of 13
cases. Z Gastroenterol 1986;24:67380.
179. Gushurst TP, Lesesne HR. Isolated duodenal varix: An unusual cause of gastrointestinal
hemorrhage. South Med J 1984;77:9158.
180.
fiberduodenoscope. Endoscopy 1973;5:1014.
181. Kunert H, Ottenjann R. Endoscopy in bleeding duodenal varices: A rare cause of massive
upper GI hemorrhage. Endoscopy 1977;8:99101.
182. Aagaard J, Burcharth F. Bleeding duodenal varices demonstrated by transhepatic
portography: Report of a case misinterpreted as bleeding duodenal ulcer. Acta Chir Scand
1980;146:7778.
183. Rappazzo JA, Kozarek RA, Altman M. Duodenal varices: Endoscopic diagnosis of an unusual
source of upper gastrointestinal hemorrhage. Gastrointest Endosc 1981;27:2278.
184. Itzchak Y, Glickman MG. Duodenal varices in extrahepatic portal obstruction. Radiology
1977;124:61924.
185. Richardson JD, McInnis WD, Pestana C. Duodenal varices. Am Surg 1976;42:2013.
186. Eleftheriadis E. Duodenal varices after sclerotherapy for esophageal varices. Am J
187. Khouqeer F, Morrow C, Jordan P. Duodenal varices as a cause of massive upper
gastrointestinal bleeding. Surgery 1987;102:54852.
188. Sauerbruch T, Weinzierl M, Dietrich HP, et al. Sclerotherapy of a bleeding duodenal varix.
Endoscopy 1982;14:1879.
189. Kirkpatrick JR, Shoenut JP, Micflikier AB. Successful injection sclerotherapy for bleeding
duodenal varix in intrahepatic portal obstruction. Gastrointest Endosc 1985;31:25960.
190. Gertsch P, Blumgart LH. Cure of a bleeding duodenal varix by sclerotherapy. Br J Surg
1988;75:717.
191. Barbish AW, Ehrinpreis MN. Successful endoscopic injection sclerotherapy of a bleeding
duodenal varix. Am J Gastroenterol 1993;88:902.
192. Tsuji H, Okano H, Fujino H, et al. A case of endoscopic injection sclerotherapy for a bleeding
duodenal varix. Gastroenterol Jpn 1989;24:604.
193. Ikeda K, Murayama H, Takano H, et al. Massive intestinal bleeding in hemangiomatosis of the
duodenum. Endoscopy 1980;12:30610.
194. Sutton D, Murfitt J, Howarth F. Gastrointestinal bleeding from large angiomas. Clin Radiol
1981;32:62932.
195.
Loison F, Patri B. Intramural hematoma of the duodenum with anticoagulantsRole of
endoscopy. Acta Endosc 1983;13:2257.
196. van Spreeuwel JP, van Garp LH, Bast TJ, et al. Intramural hematoma of the duodenum in a
patient with chronic pancreatitis. Endoscopy 1981;13:2468.
197. Patel R, Shaps J. Intramural duodenal hematomaA complication of ERCP. Gastrointest
Endosc 1982;28:2189.
198. Freed JS, Roe G, Szuchmacher PH. Intramural duodenal hematoma mimicking carcinoma of
the head of the pancreas. Mt Sinai J Med 1979;46:5647.
199. Ghishan FK, Werner M, Vieira P, et al. Intramural duodenal hematoma: An unusual
complication of endoscopic small bowel biopsy. Am J Gastroenterol 1987;82:36870.
200. Zinelis SA, Lee M, Hershenson MD, et al. Intramural duodenal hematoma following upper
gastrointestinal endoscopic biopsy. Dig Dis Sci 1989;34:28991.
201. Aizawa K, Tokuyama H, Yonezawa T, et al. A case of traumatic intramural hematoma of the
duodenum effectively treated with ultrasonically guided aspiration drainage and endoscopic
balloon catheter dilation. Gastroenterol Jpn 1991;26:21823.
202. Ramanujam P, Shabeeb N, Silver JM. Unusual manifestations of gallstone migration into the
gastrointestinal tract. South Med J 1983;76:302.
203. Salah-Eldin AA, Ibrahim MAH, Alapati R, et al. The Bouveret syndrome: An unusual cause of
hematemesis. Henry Ford Hosp Med J 1990;38:524.
204. Chait MM, Lerner AG. Bouveret's syndrome presenting as upper gastrointestinal hemorrhage.
205. Heyd RL, Solinger MR, Howard AL, Rosser JC Jr. Acute upper gastrointestinal hemorrhage
caused by gallstone impaction in the duodenal bulb. Dig Dis Sci 1992;37:4525.
206. Bottari M, Pallio S, Scribano E, Certo A. Pyloroduodenal obstruction by a gallstone: Bouveret's
syndrome. Gastrointest Endosc 1988;34:4402.
207. Grove O. Acute pyloric obstruction by gallstone: Report of a case diagnosed by gastroscopy.
208. Thorpe JA. Gallstone ileus. J R Coll Surg Edinb 1979;24:299300.
209. Argyropoulos GD, Velmachos G, Axenidis G. Gallstone perforation and obstruction of the
duodenal bulb. Arch Surg 1979;114:3335.
210. Bedogni G, Contini S, Meinero M, et al. Pyloroduodenal obstruction due to a biliary stone
(Bouveret's syndrome) managed by endoscopic extraction. Gastrointest Endosc
1985;31:368.
211. Van Dam J, Steiger E, Sivak MV Jr. Giant duodenal gallstone presenting as gastric outlet
obstruction: Bouveret's syndrome. J Clin Gastroenterol 1992;15:1503.
212. Barton W, Walsh TN, Keane T, et al. Malignant duodenocolic fistula: Report of a case and
review of the literature. Dis Colon Rectum 1987;30:6367.
213. Roberts PL, Coller JA, Corriveau S, et al. Malignant duodenocolic fistula diagnosed by
endoscopy. Surg Endosc 1989;3:1124.
214. Welch JP, Warshaw AL. Malignant duodenocolic fistulas. Am J Surg 1977;133:65861.
215. Ergin MA, Alfonso A, Auda SP, et al. Primary carcinoma of the duodenum producing a
malignant duodenocolic fistula. Dis Colon Rectum 1978;21:40812.
216. Gregson R, Craig O. Aorto-enteric fistulae: The role of radiology. Clin Radiol 1983;34:6572.
217. Steffes BC, O'Leary JP. Primary aortoduodenal fistulae: A case report and review of the
literature. Am Surg 1980;46:1219.
218. Baker MS, Fisher JH, van der Reis L, et al. The endoscopic diagnosis of an aortoduodenal
fistula. Arch Surg 1976;111:304.
219. Perdue GD Jr, Smith RB III, Ansley JD, et al. Impending aortoenteric hemorrhage: The effect
of early recognition on improved outcome. Ann Surg 1980;192:23743.
220. Hill SL, Knott LH, Alexander RH. Recurrent aortoduodenal fistula: A lesson in management.
Am Surg 1982;48:13740.
221. Champion MC, Sullivan SN, Watson WC. Aortoduodenal fistula: Endoscopic diagnosis. Dig
Dis Sci 1980;25:8112.
222. Brady PG. Aortoduodenal fistula: Role of endoscopy in diagnosis. Am J Gastroenterol
1978;69:7057.
223. Ott W, Kerr RM, Gelfand DW. Aortoduodenal fistula: An unusual endoscopic and radiographic
appearance simulating leiomyoma. Gastrointest Endosc 1978;24:2968.
224. Brand EJ, Sivak MV Jr, Sullivan BH Jr. Aortoduodenal fistula: Endoscopic diagnosis. Dig Dis
Sci 1979;24:9404.
225. Puppala AR, Munaswamy M, Doshi AM, Steinheber FU. Endoscopic diagnosis of
aortoduodenal fistula. Complication of abdominal aortic by-pass grafts. Am J Gastroenterol
1980;73:4147.
226. Martin J, Cano N, Di-Costanzo J, et al. Aortoduodenal fistula: Endoscopic diagnosis. Dig Dis
Sci 1981;26:9567.
227. Banai J, Salfay G, Szeleczky M, et al. An uncommon endoscopic finding: Bleeding secondary
to aortoenteric fistula. Endoscopy 1982;14:1856.
228. Baker BH, Baker MS, van der Reis L, Fisher JH. Endoscopy in the diagnosis of aortoduodenal
fistula. Gastrointest Endosc 1977;24:357.
229. Hardin RD, Foster RD, Meissner FW. Endoscopic diagnosis of a paraprosthetic-enteric
fistula. Gastrointest Endosc 1987;33:3189.
230. Mir-Madjlessi SH, Sullivan BH Jr, Farmer RG, Beven EG. Endoscopic diagnosis of
aortoduodenal fistula. Gastrointest Endosc 1973;19:1878.
231. Bedogni G, Meinero M, Barbieri I, et al. Gastrointestinal hemorrhage from an aortoduodenal
fistula: Endoscopic considerations. Ital J Surg Sci 1985;15:2136.
232. Konigsrainer A, Pointner R, Reissigl H, Weimann S. Endoscopic aspects of aortointestinal
fistula. Surg Endosc 1987;1:3740.
233. Wexler RM, Falchuk KR, Horst DA, et al. Duodenal erosion of a mesocaval graft: An unusual
complication of mesocaval shunt interposition surgery. Gastroenterology 1980;79:72930.
234. Goldschmid S, Brady PG, Slone FL, Farber MS. Duodenal erosion caused by a mesocaval
graft. Gastrointest Endosc 1986;32:4124.
235.
Stevens FM, McCarthy CF. The endoscopic demonstration of coeliac disease. Endoscopy
1977;8:17780.
236. Gillberg R, Ahren C. Coeliac disease diagnosed by means of duodenoscopy and endoscopic
duodenal biopsy. Scand J Gastroenterol 1977;12:9116.
237. Holdstock G, Eade OE, Isaacson P, Smith CL. Endoscopic duodenal biopsies in coeliac
disease and duodenitis. Scand J Gastroenterol 1979;14:71720.
238. Gillberg R, Kastrup W, Mobacken H, et al. Endoscopic duodenal biopsy compared with biopsy
with the Watson capsule from the upper jejunum in patients with dermatitis herpetiformis.
239.
Mee AS, Burke M, Vallon AG, et al. Small bowel biopsy for malabsorption: Comparison of the
diagnostic adequacy of endoscopic forceps and capsule biopsy specimens. Br Med J
1985;291:76972.
240. Scott BB, Losowsky MS. Patchiness and duodenal-jejunal variation of the mucosal
abnormality in coeliac disease and dermatitis herpetiformis. Gut 1976;17:98492.
241. Kirberg A, Laterre JJ, Hartord ME. Endoscopic small intestinal biopsy in infants and children:
Its usefulness in the diagnosis of celiac disease and other enteropathies. J Pediatr
242. Scott BB, Jenkins D. Endoscopic small intestinal biopsy. Gastrointest Endosc 1981;27:1627.
243. Ladas SD, Tsamouri M, Kouvidou C, Raptis SA. Effect of forceps size and mode of
orientation on endoscopic small bowel biopsy evaluation. Gastrointest Endosc 1994;40:515.
244. Jabbari M, Wild G, Goresky CA, et al. Scalloped valvulae conniventes: An endoscopic marker
of celiac sprue. Gastroenterology 1988;95:151822.
245. Corazza GR, Caletti GC, Lazzari R, et al. Scalloped duodenal folds in childhood celiac
246. Brocchi E, Corazza GR, Caletti G, et al. Endoscopic demonstration of loss of duodenal folds
in the diagnosis of celiac disease. N Engl J Med 1988;319:7414.
247. Maurino E, Capizzano H, Niveloni S, et al. Value of endoscopic markers in celiac disease. Dig
Dis Sci 1993;38:202833.
248. McIntyre AS, Ng DP, Smith JA, et al. The endoscopic appearance of duodenal folds is
predictive of untreated adult celiac disease. Gastrointest Endosc 1992;38:14851.
249.
Schmitz U, Ko Y, Seewald S, et al. Iron-deficiency anemia as the sole manifestation of celiac
disease. Clin Invest 1994;72:51921.
250. Gostout CJ. Enteroscopy for unexplained iron-deficiency anemia: Identifying the patient with
sprue. Gastrointest Endosc 1993;39:769.
251. Hall MJ, Cooper BT, Rooney N, et al. Coeliac disease and malignancy of the duodenum:
Diagnosis by endoscopy, successful treatment of the malignancy, and response to a gluten
free diet. Gut 1991;32:902.

252. Volpicelli NA, Salyer WR, Milligan FD, et al. The endoscopic appearance of the duodenum in
Whipple's disease. Johns Hopkins Med J 1976;138:1923.
253. Riemann JF, Rosch W. Synopsis of endoscopic and related morphological findings in
Whipple's disease. Endoscopy 1978;10:98103.
254. Geboes K, Ectors N, Heidbuchel H, et al. Whipple's disease: The value of upper
gastrointestinal endoscopy for the diagnosis and follow-up. Acta Gastroenterol Belg
1992;55:20919.
255. Crane S, Schlippert W. Duodenoscopic findings in Whipple's disease. Gastrointest Endosc
1978;24:2489.
256. Geboes K, Ectors N, Heidbuchel H, et al. Whipple's disease: Endoscopic aspects before and
after therapy. Gastrointest Endosc 1990;36:24752.
257. Feller ER, Weiser MM, Schapiro RH. Endoscopic visualization of nodular lymphoid
hyperplasia. Gastrointest Endosc 1977;24:378.
258. Breslaw L. Melanosis of the duodenal mucosa. Gastrointest Endosc 1980;26:456.
259. Cowen ML, Humphries TJ. Pseudomelanosis of the duodenum. Endoscopic and histologic
observations on a unique case of pigmentation in the duodenum. Gastrointest Endosc
1980;26:1078.
260. Ganju S, Adomavicius J, Salgia K, Steigmann F. The endoscopic picture of melanosis in the
duodenum. Gastrointest Endosc 1980;26:445.
261. Yamase H, Norris M, Gillies C, et al. Pseudomelanosis duodeni: A clinicopathologic entity.
262. Lee HH, O'Donnell DB, Keren DF. Characteristics of melanosis duodeni: Incorporation of
endoscopy, pathology, and etiology. Endoscopy 1987;19:1079.
263. Castellano G, Canga F, Lopez I, et al. Pseudomelanosis of the duodenum. Endoscopic,
histologic, and ultrastructural study of a case. J Clin Gastroenterol 1988;10:1504.
264. Kuo YC, Wu CS. Duodenal melanosis. J Clin Gastroenterol 1988;10:1604.
265. Sharp JR, Insalaco SJ, Johnson LF. "Melanosis" of the duodenum associated with a gastric
ulcer and folic acid deficiency. Gastroenterology 1980;78:3669.
266. Rex DK, Jersild RA Jr. Further characterization of the pigment in pseudomelanosis duodeni in
three patients. Gastroenterology 1988;95:17782.
267. Lin H-S, Tsay S-H, Chiang H, et al. Pseudomelanosis duodeni: Case report and review of
literature. J Clin Gastroenterol 1988;10:1559.
268. Tada S, Iida M, Iwashita A, et al. Endoscopic and biopsy findings of the upper digestive tract
in patients with amyloidosis. Gastrointest Endosc 1990;36:104.
269. Tada S, Iida M, Yao T, et al. Endoscopic features in amyloidosis of the small intestine: Clinical
and morphologic differences between chemical types of amyloid protein. Gastrointest Endosc
1994;40:4550.
Chapter 50 Endoscopy of the Small Intestine

(4802)
(4803)
BLAIR S. LEWIS, M.D.
The standard endoscopic examinations, esophagogastroduodenoscopy (EGD) and colonoscopy,

visualize only limited areas of the small intestine. Upper endoscopy reaches the junction of the second
and third portions of the duodenum, and the terminal ileum can be intubated for up to 30 cm at
colonoscopy. Endoscopic evaluation of the distal duodenum and large areas of the jejunum and ileum
has been termed enteroscopy.1,2(4804) At present, several methods are available: push enteroscopy,
sonde enteroscopy, intraoperative enteroscopy, as well as the rope-way technique, now largely
abandoned, and the investigational technique of endostomy.
Push Enteroscopy
Technique
Advancing a long endoscope into the small bowel during upper intestinal endoscopy is the most
common form of enteroscopy. During "push" enteroscopy, an endoscope is "pushed" beyond the
ligament of Treitz into the proximal jejunum. Push enteroscopy is also termed deep upper endoscopy,
extended EGD, or "enteroscopy." Although designated instruments have been made for this
examination,35(4805) most experience has been based on peroral passage of an "adult" or
"pediatric" colonoscope.611(4806) When using a colonoscope passed orally, it is possible to intubate
the jejunum for approximately 40 to 60 cm beyond the ligament of Treitz. Video push enteroscopes are
now available.12,13(4807) A comparison of the various instruments used for push enteroscopy is
shown in Table 501.
TABLE 501
INSTRUMENT
Pediatric colonoscope
Adult colonoscope
SIF 100 (Olympus)
VSB-P2900 (Pentax)
Specifications of Instruments Used for Push Enteroscopy

DIAMETER
(mm)
11
14
10.5
11
LENGTH
(cm)
135
165
217
240
TYPE
Video/fiberoptic
Video/fiberoptic
Video
Video
OVERTUBE
(cm)
100
60
DEPTH OF
INSERTION
(cm)
60
90
120
120
Recently, longer instruments have been developed. Together with the development of an overtube that
limits looping of the endoscope insertion tube within the stomach, deeper intubation of the jejunum with
these instruments is now possible.14,15(4808) The new push enterscopes measure 200 to 225 cm in
length. Push enterscopy with an overtube requires the use of fluoroscopy. Currently available
overtubes are 60 or 100 cm in length. Some have a Gore-Tex tip 10 cm in length. The Gore-Tex is soft
and pliable, limiting the mucosal trauma when advancing the overtube over the endoscope.
The overtube (also termed a stiffening tube) is initially backloaded onto the insertion tube of the orally
passed enteroscope. Once the endoscope tip is within the jejunum, the overtube is advanced down the
esophagus until its distal tip rests within the second portion of the duodenum for the 60-cm overtube,
and to or beyond the ligament of Treitz for the 100-cm version. Fluoroscopy is used to ensure intestinal
placement of the overtube because prepyloric placement of the overtube tip does little to aid deep
intubation. This is especially true when using an overtube with a soft Gore-Tex tip that may buckle and
not provide any rigidity if within the antrum. During advancement of the overtube, the insertion tube of
the enteroscope must be in a shortened and straightened configuration. This is achieved by torquing
the insertion tube clockwise while pulling it backward. This maneuver is very similar to that of positing a
side-viewing duodenoscope within the duodenum during endoscopic retrograde
cholangiopancreatography and is performed simultaneously with advancement of the overtube under
fluoroscopic guidance.
The stiffening tube (overtube) acts to limit intragastric and intraduodenal bulb looping of the
enteroscope, previously a limitation to deep small bowel intubation. The push enteroscope can then be
advanced 80 to 120 cm beyond the ligament of Treitz, effectively bringing endoscopic evaluation and
therapeutic intervention to the entire jejunum. In many cases, deep intubation can also be achieved
with this instrument without use of an overtube (Figure 501). Barkin et al.14(4809) described extreme
patient discomfort in 13 of the 38 patients examined with the use of the overtube and noted that deep
sedation was necessary in these patients.
(4810)Figure 501. Plain x-ray of the abdomen showing deep intubation of the small intestine
with the push enteroscope.
Push enteroscopy using a pediatric colonoscope is a relatively simple procedure. Sterilization of the
instrument is not necessary, and a properly cleaned and disinfected colonoscope may be used safely.
Colonoscopes provide four-way directional tip deflection of 160 to 180 degrees, unlike gastroscopes
that offer this range of deflection in only one direction. Greater tip deflection, along with increased
length, makes the pediatric colonoscope the best readily available instrument for push enteroscopy. An
"adult" colonoscope can prove to be too wide and with too great an angle of curvature at the tip to
permit peroral passage, especially in the elderly patient or the patient with a short oropharynx. The
examination is performed in similar fashion to standard EGD. The patient is sedated with the usual
medications and placed in the left lateral position. A complete examination of the esophagus and
stomach should be performed. Limiting air insufflation and thus distention of the stomach facilitates
deeper insertion of the instrument. The pylorus usually offers no impediment to passage of the
larger-diameter instrument. The patient may feel discomfort as the instrument is advanced, distending
the greater curvature of the stomach. Paradoxical or arrested tip motion can be noted if an instrument
loop develops within the fundus. With gentle pressure, it is generally possible to "push through" the
loop. The ligament of Treitz is usually encountered 80 to 100 cm from the incisors; at this point,
extreme tip deflection is typically required to find the lumen. Once around the ligament, the first jejunal
loop is readily identified owing to its straight configuration as it drops to the pelvis.
Indications and Results

There are several indications for push enteroscopy of the proximal small intestine. This is now the
accepted method for obtaining small bowel biopsies because the examination time is short, the tissue
specimens are of good quality, and the procedure is relatively easy for both patient and
physician.6,7(4811)
Push enteroscopy is often used to search for a cause of obscure gastrointestinal bleeding. Reported
diagnostic yields have varied from 13 to 38%.811,16,17(4812) Messer et al.11(4813) found the site of
bleeding in 20 of 52 patients (38%); findings included angiodysplasias in 9 and small bowel tu-mors in
11. Foutch et al.10(4814) reported a yield of 38% in 39 patients; angiodysplasia accounted for 80% of
diagnoses. Chong et al.16(4815) performed push enteroscopy in 68 patients, 55 of whom underwent
the procedure because of obscure gastrointestinal bleeding. Bleeding lesions distal to the ligament of
Treitz were found in 14 patients (25%), angiodysplasia again being the most common diagnosis. An
abnormality was found in the small intestine in 19 of 31 patients (61%) who underwent push
enteroscopy for obscure gastrointestinal bleeding or unexplained iron deficiency anemia in the series
of Harris et al.18(4816) Small bowel angiodysplasia, the most common diagnosis, was found in 12
patients.
The value of enteroscopy in patients with iron deficiency anemia and no overt evidence of bleeding is
less certain than for patients with outright manifestations of bleeding. EGD or colonoscopy will
demonstrate a possible cause of iron deficiency anemia in significant numbers of patients. In the study
of Rockey and Cello,19(4817) standard endoscopic investigations revealed a potential source of blood
loss in 62 of 100 iron-deficient patients. Small bowel x-rays obtained using enteroclysis technique failed
to disclose a cause in 26 patients in whom EGD and colonoscopy were negative. In a follow-up study
conducted by sending questionnaires to primary care physicians, Sahay and Scott20(4818) found that
iron deficiency anemia disappeared in 89% of patients when a prior EGD (with small bowel biopsy) and
barium enema did not identify a cause. Nevertheless, enteroscopy may be of value when anemia is
profound and persistent. Patients with sprue may occasionally present with only iron deficiency anemia,
and a few cases have been reported in which the diagnosis was made at enteroscopy (see Chapter
49: Diseases of the Duodenum).21(4819) Ulcerative jejunitis has also been described at enteroscopy
in a patient with refractory celiac disease.22(4820)
Push enteroscopy can be therapeutic as well as diagnostic; coagulation of bleeding sites and
polypectomy are readily accomplished during the procedure. Using bipolar cautery, Foutch et
al.10(4821) coagulated angiodysplasias in 11 of 12 patients; hemostasis was achieved in 8 of the 11
treated patients. Askin and Lewis23(4822) retrospectively studied the clinical course of 83 of 112
patients with gastrointestinal bleeding of unknown etiology and a diagnosis of small intestinal
angiodysplasia at either push or sonde enteroscopy. Patients not lost to follow-up were grouped into
those who did and those who did not undergo endoscopic coagulation. The group of patients who
underwent electrocoagulation (n = 55) was similar to the group (n = 28) in which no intervention was
performed with regard to average age (73 vs. 71 years), mean number of units transfused (21.4 vs.
15.8), and duration of bleeding (22 months for each group). The mean duration of follow-up for the
patients who did not undergo endoscopic treatment was 26 months; there was no significant difference
in the rates of bleeding before and after enteroscopy in this group. The mean duration of follow-up for
patients who underwent endoscopic coagulation was 30 months; in this group, there was a significant
difference in the number of units transfused per month before (2.40 2.97 units) and after (0.32 0.91
units) treatment.
Push enteroscopy has also been used in various ways to place jejunal feeding tubes. It has been
advocated for placement of a transgastric jejunal tube through a previous gastrostomy.24(4823) With
the use of push enteroscopy, the Ponsky technique for the placement of percutaneous gastrostomies
has been extended to permit direct percutaneous jejunostomies (see Chapter 55: Percutaneous
Endoscopic Gastrostomy).25,26(4824) Nasojejunal feeding tubes have been placed using a Seldinger
technique.27(4825) The push enteroscope is positioned in the jejunum initially and a guidewire is then
advanced through the accessory channel of the instrument. Finally, the endoscope is removed, leaving
the guidewire in place, over which a feeding tube is advanced. This technique has also been used to
place tubes to facilitate barium enteroclysis radiography of the small intestine.28,29(4826)
Patients with polyposis syndromes have been evaluated by push enteroscopy for the presence of
jejunal polyps. Bertoni et al.30(4827) found jejunal adenomas in 8 of 16 affected patients, usually within
the first 20 cm beyond the ligament of Treitz. One lesion was a 4-cm diameter severely dysplastic
tubulovillous adenoma. Polypectomy has been performed at push enteroscopy in patients with familial
adenomatous polyposis (FAP).31(4828) Push enteroscopy has been used to obtain a biopsy diagnosis
in cases where small bowel x-rays raise a suspicion of a malignant tumor in the jejunum.32(4829)
Push enteroscopy has been used to obtain cholangiograms in patients after Roux-en-Y
hepaticojejunostomy.14,33(4830)
Complications
It can be assumed that the complication rate and types of complications associated with push
enteroscopy will be the same as those for EGD. There is no reported evidence that untoward events
are more common by comparison. In fact, complications are rare with push enteroscopy and have not
been reported with peroral passage of a colonoscope for this purpose.2(4831)
Complications have been associated with the use of an overtube to facilitate push enteroscopy. Barkin
et al.14(4832) reported two major complications with use of the longer push enteroscope with the
overtube in a series of 37 patient examinations. A Mallory-Weiss tear occurred in 1 patient, and 1 case
of pancreatitis most likely secondary to trauma of the main duodenal papilla was described. Three
patients complained of abdominal pain after the procedure, although this resolved spontaneously within
2 hours in all cases. One of 68 patients undergoing push enteroscopy with use of an overtube in the
series of Chong et al.16(4833) sustained a pharyngeal tear. Yang and Laine34(4834) described 3
cases in which a strip of gastric mucosa was avulsed along the greater curvature, presumably during
insertion of the overtube. Hematemesis occurred in 1 patient, but the bleeding stopped spontaneously
and transfusion was not required. These complications were related to the use of an overtube before
the addition of a Gore-Tex tip to the device.
Sonde Enteroscopy
Technique
Sonde enteroscopy, also termed small bowel enteroscopy, can provide endoscopic visualization of the
entire small intestine. The technologic basis for the procedure is the placement of a balloon at the
distal tip of a long, narrow-diameter instrument. Once the balloon is distended with air or water,
peristalsis carries the long flexible endoscope to the distal small bowel. Endoscopic evaluation is
performed during withdrawal.
Early prototype sonde enteroscopes were too stiff to permit distal passage, and forward angles of view
were narrow (60 degrees).3537(4835) Peroral passage, which was required with these relatively
large-diameter instruments, was associated with patient salivation, gagging, and considerable
discomfort.38(4836)
A thin, flexible, enteroscope suitable for transnasal passage (tip diameter 5 mm; working length 2560
mm) was developed in 1986 (Figure 502).39(4837) The initial prototype had a forward angle of view
of 90 degrees, but this has been increased to 120 degrees. This instrument, in contrast to a push
enteroscope, does not have an accessory channel. Therefore, it is not possible to obtain biopsies or
perform therapeutic maneuvers. Furthermore, the sonde endoscope does not have mechanisms for tip
deflection, and thus visualization is invariably incomplete, even when total small bowel intubation is
achieved. Abdominal palpation along with inflation and deflation of the balloon is the only means for
manipulating the intraluminal view. Moreover, it is difficult and frequently impossible to control the
motion of the instrument during withdrawal, that is, the instrument tip may slip backward over long
segments of intestine in response to even the slightest degree of traction. This phenomenon markedly
amplifies the difficulty of endoscopic observation.
(4838)Figure 502. The sonde fiberoptic enteroscope. Note the balloon and suture attached at
the distal tip of the instrument.
Prototype sonde enteroscopes that have an accessory channel have been developed. Using this type
of instrument, Gostout40(4839) described a technique to reduce the problem of rapid retrograde
slippage during instrument withdrawal in which a biopsy forceps is used to push successive loops of
intestine away from the tip of the instrument. Despite such efforts, however, endoscopic observation
with currently available sonde enteroscopes remains less than ideal. Lewis and Waye41(4840)
estimated that approximately 50 to 70% of small bowel mucosa is observed during a standard
examination. Usually, the tip of the instrument does not reach the ileocecal valve; the ileum is reached
in about 75% of examinations, the ileocecal valve in 10% (Figure 503). Failure to achieve total small
bowel intubation can be attributed to various factors including intraabdominal adhesions, endoscope
malfunction, or the presence of an obstructing lesion.
(4841)Figure 503. Plain x-ray of the abdomen showing total small bowel intubation with the
sonde enteroscope.
A technique for rapid placement of the sonde enteroscope into the jejunum was developed to shorten
the examination time.41(4842) In this rapid technique, a push enteroscope grasps a suture affixed to
the tip of the sonde enteroscope while both instruments lay within the stomach. Then, both instruments
are advanced into the small bowel. This technique permits total to near-total small bowel intubation
within 8 hr and allows the procedure to be performed on an ambulatory basis. Procedure times range
from 6 to 8 hr; withdrawal times range from 45 to 60 min. Dabezies et al.42(4843) reported on the
development of a video sonde enteroscope. This instrument was used in seven patients, but owing to
its larger-diameter size, peroral placement was required and depth of intubation was limited.

At present, the only indication for sonde enteroscopy is the evaluation of patients with gastrointestinal
bleeding of obscure etiology. Initially, Lewis and Waye43(4844) reported results of this procedure in 60
patients with obscure gastrointestinal bleeding; a small bowel site of bleeding was detected in 33%.
Similar experience was reported by Barthel et al.,44(4845) with a yield of 27.8% in 18 patients, and by
Gostout et al.,45(4846) with a yield of 26% in 35 patients. In the latter report, an orally passed
enteroscope was used along with a team approach to enhance patient care and shorten physician
time. The sonde enteroscope was used by Morris et al.46(4847) to evaluate 15 patients with obscure
bleeding who were taking nonsteroidal anti-inflammatory drugs (NSAIDs). Seven patients were found
to have either jejunal or ileal ulceration. An abnormality was found by sonde enteroscopy in 14 of 16
patients with obscure gastrointestinal bleeding in the retrospective series of Lopez et al.47(4848) Push
enteroscopy was negative in all cases. Positive findings included ileal ulcers (6 patients),
angiodysplasia (3 patients), and tumors (2 patients).
Despite advances in sonde enteroscopy, some authorities have questioned its clinical
usefulness.48(4849) It has been stated that the development of sonde enteroscopy was unnecessary
and, most likely, too expensive.49(4850) It is clear, however, that the deeper one looks into the small
bowel, the more one sees. Patients with obscure gastrointestinal bleeding constitute only a small
fraction of the total number of patients with bleeding. By comparison, however, the diagnosis and
treatment of patients with obscure bleeding consumes a considerably larger proportion of health care
resources over the course of time.50(4851) Berner et al.51(4852) retrospectively reviewed the results
of 553 enteroscopic examinations performed for gastrointestinal bleeding of obscure origin in which
both sonde and push techniques were employed. The combined studies identified a potential source of
bleeding in 58%. Abnormalities were found distal to the limits of EGD in 40%. In 26%, a lesion was
detectable only by sonde enteroscopy. Mucosal vascular lesions, the most common finding, were
discovered in 31%; tumors were diagnosed in 6%. Angiography and small bowel radiography
performed using the technique of enteroclysis may be falsely negative in patients with small bowel
tumors.52(4853)
Complications
Sonde enteroscopy appears quite safe, and no serious complications have been described. Among 60
patients undergoing the procedure, Lewis and Waye41(4854) encountered 2 cases of epistaxis related
to transnasal passage of the enteroscope. Lewis53(4855) described enteroscopically induced trauma
within the small bowel; this "notching" was similar in appearance to that of NSAID-induced ulcers.
Intraoperative Enteroscopy
Technique
Because sonde enteroscopy is not widely available, intraoperative enteroscopy remains the most
common form of total small bowel endoscopic examination. Colonoscopes are routinely employed in
this examination, although a push enteroscope may also be used. Use of a sonde enteroscope has
also been reported.54(4856) Access to the small intestine may be via the mouth, the colon, or an
enterotomy. The use of video endoscopes allows both the surgeon and the endoscopist to see the
intraluminal view.55(4857)
With the peroral approach, the endoscope need not be sterilized. The proximal jejunum is intubated
prior to the performance of the laparotomy because it may be difficult, once the abdomen is opened, to
advance the instrument around the ligament of Treitz owing to excessive, and unopposed, bowing of
the insertion tube along the greater curvature of the stomach. This can be controlled to some extent if
the surgeon exerts counterpressure along the greater curvature. For peroral passage of an "adult"
colonoscope, it may be necessary to temporarily deflate the cuff of the endotracheal tube to permit
passage of this relatively wide-diameter instrument. Once the colonoscope is placed within the
proximal jejunum, laparotomy is performed. A noncrushing clamp is placed across the ileocecal valve
to prevent distention of the colon with insufflated air. Excessive distention of the colon and small
intestine can lead to subsequent difficulties with closure of the abdomen.
The endoscopic examination is performed by having the surgeon grasp the endoscope tip and hold a
short segment of bowel (about 10 cm) straightforward from the instrument lens to allow endoscopic
inspection. The view is enhanced by dimming the overhead lights. This also allows the surgeon to look
at the transilluminated bowel, a technique that makes even the smallest vascular lesions readily visible
(Figure 504). Small amounts of blood in the lumen can be an important clue to the presence of a
lesion. Once a segment of bowel has been examined both internally and externally, it is pleated onto
the insertion tube and the next section is examined (Figure 505). Generally, the examination should
be performed only during intubation because the mucosal trauma that occurs with pleating causes
artifacts that can have the appearance of vascular ectasias.56(4858) As an aid in differentiating trauma
from angiodysplasias, Bowden57(4859) described a technique ("reverse transillumination") in which
the endoscopic light source is turned off and the operating room lights are focused on the intestine.
Lesions identified during the course of enteroscopy should be marked by placing a suture in the
serosa. At the end of the examination, the endoscope is withdrawn and as much air as possible is
aspirated from the gut. The remaining serosal sutures serve to identify sites for resection.
(4860)Figure 504. Normal transilluminated bowel. The operating suite has been darkened,
and the bowel is transilluminated by the endoscope. Vascular lesions are readily demonstrated
with this technique.
(4861)Figure 505. The small bowel is totally intubated by an orally passed endoscope at
intraoperative enteroscopy.
The results of intraoperative enteroscopy by peroral placement of a colonoscope in 23 patients were
reported by Lewis et al.58(4862) The ileocecal valve was reached in 60% of cases; the distal ileum
was reached within 2 feet of the valve in 15%.
A long endoscope (e.g., push enteroscope, "pediatric" colonoscope) can also be used to examine the
small intestine via a laparotomy. With the peroral technique, it is sometimes difficult or impossible to
reach the ileocecal valve, especially if there are adhesions. Formation of an instrument loop along the
greater curvature of the stomach is usually a constant problem during the procedure. Instrument
insertion through an enterotomy, usually placed around the midpoint of the small bowel, frequently
circumvents many of these problems. However, a sterilized endoscope or an endoscope covered by a
sterile plastic sheath is required for this procedure. Otherwise, the technique is virtually the same as
that for peroral intraoperative endoscopy.
Retrograde intubation of the small bowel can also be performed as an alternative by initially performing
a colonoscopy. This technique is the most involved and time-consuming and generally only a small
segment of the ileum can be examined.

There are several indications for intraoperative enteroscopy: to diagnose the site and cause of obscure
bleeding, to investigate lesions discovered incidentally during operation, to provide endoscopic access
to lesions otherwise inaccessible owing to adhesions, and to aid surgeons confronted with dense
adhesions in finding abnormalities.57(4863) Intraoperative enteroscopy has been used to evaluate
and, in some cases treat, patients with polyposis syndromes.
Intraoperative enteroscopy is at present the most widely used method for identification of small
intestinal sites of bleeding.5557,5964(4864) Reported diagnostic yields for this indication range from
83 to 100%.55,56,6366(4865) Angiodysplasia is the most common operatively nonpalpable cause of
bleeding, but radiation enteritis, ulcerations, and even strictures may require endoscopic identification.
Although coagulation of a bleeding site by thermal noncontact (laser) and contract (heat probe,
multipolar probe) methods can be performed endoscopically during intraoperative enteroscopy,
surgical resection is considered definitive. Treatment of diffuse angiodysplasias of the small bowel by
laser photocoagulation during intraoperative enteroscopy has not resulted in long-term
hemostasis.58(4866) During long-term follow-up of patients after intraoperative enteroscopy, the rate
of recurrent bleeding has ranged from 0 to 60%.56,58,59,64,66(4867)
Lewis et al.58(4868) compared the results of sonde and intraoperative enteroscopy in 23 patients who
underwent both procedures. Intraoperative and sonde enteroscopic findings of angiodysplasias
matched exactly in number and location in 77% of cases. Preoperative sonde enteroscopy missed
angiodysplasias in 3 patients, in 1 instance within an area of traversed bowel, and within areas of the
small bowel not reached by the sonde enteroscope in the other 2 cases. Intraoperative enteroscopy
missed an angiodysplasia in 1 patient; acute bleeding reoccurred postoperatively and necessitated
repeat intraoperative enteroscopy.
Mathus-Vliegan and Tytgat61(4869) performed enteroscopic polypectomy in five patients with
Peutz-Jeghers syndrome because of persistent bleeding or intussusception. Several of these patients
had undergone multiple prior operations including surgical resections of the small intestine. These
authors noted that in many cases the procedure was long and tedious. The long-term effectiveness of
the procedure was not evaluated. A similar experience in five patients with Peutz-Jeghers polyps was
reported by Webb,67(4870) who found that intraoperative enteroscopy identified polyps that were
overlooked by palpation of the bowel. In nine patients with either familial adenomatous polyposis (FAP)
(n = 7) or familial juvenile polyposis (n = 2), Rodriguez-Bigas et al.68(4871) found polyps in the jejunum
or ileum, or both, in five patients by intraoperative enteroscopy. These were adenomatous polyps in
three patients, each of whom had adenomas in the upper gastrointestinal tract at EGD. A juvenile
polyp from one 14-year-old patient contained intramucosal carcinoma. These authors recommended
that a baseline intraoperative enteroscopy be obtained in any polyposis patient undergoing surgery, as
for example, prophylactic colectomy in patients with FAP.
Endoscopic access to an hepaticojejunostomy is usually precluded by the presence of a
Roux-en-Y-type anastomosis. However, the anastomosis is accessible by intraoperative enteroscopy.
Asbun et al.69(4872) inserted an endoscope via an enterotomy made in the jejunal limb connected to
the bilioenteric anastomosis in three patients, all of whom had recurrent cholangitis and a history of
multiple biliary procedures. Endoscopic revision of the anastomosis was successful in two patients; the
anastomosis was found to be patent in the third patient.
Intraoperative enteroscopy has been used to evaluate the small intestine in patients undergoing
surgery for Crohn's disease.7072(4873) In the studies of Smedh et al.70(4874) and Lescut et
al.,71(4875) mucosal abnormalities due to Crohn's disease were more extensive than recognized
preoperatively and by operative inspection of the serosal surface of the small intestine. However,
Smedh et al.70(4876) found that mural thickening was generally more extensive than mucosal
inflammation. These investigators found that the endoscopic findings influenced surgical decisions in
20 of 33 operations performed in 31 patients.
Localization of a stricture by intraoperative enteroscopy has been reported in a 4-month-old infant with
prolonged obstruction.73(4877) Resection of the stricture, which eluded diagnosis by customary
methods, resulted in immediate clinical improvement. Interoperative enteroscopy diagnosed
heterotopic gastric mucosa in a 4-year-old boy with recurrent small bowel intussusception.74(4878)
Three prior surgeries without enteroscopy had not demonstrated the lesion, which was not associated
with a Meckel diverticulum.
Complications
Lewis et al.,58(4879) in a series of 23 patients, noted that pleating the small bowel produced mucosal
lacerations in 50% of patients; a perforation occurred in 5%. Prolonged ileus was not encountered, but
this has been described after intraoperative enteroscopy with cessation of bowel function lasting as
long as 11 days postoperatively.75(4880) Intraoperative enteroscopy can be ineffective in the actively
bleeding patient because intraluminal blood can obscure the view.76(4881) Intraabdominal adhesions
from a previous laparotomy can also make intraoperative enteroscopy difficult.
Investigational Methods of Enteroscopy

Investigational techniques of enteroscopy include the so-called rope-way method, endostomy, and
laparoscopy-assisted enteroscopy.
The rope-way method is the oldest technique for intubation of the total small intestine.37,77,78(4882)
The patient swallows a guide string, which is allowed to pass through the rectum. The string is then
exchanged for a somewhat stiffer Teflon tube over which an endoscope is passed. A complete
endoscopic examination of the small intestine can be obtained with this method. Unfortunately, the
procedure is painful owing to tightening of the guide tube and often requires general anesthesia. Owing
to patient discomfort, length of time necessary for string passage, and development of better tolerated
procedures, the rope-way method has largely been abandoned.
As described by Frimberger et al.,79(4883) endostomy involves the creation of enterocutaneous
fistulas through which a narrow-diameter endoscope is used to intubate the small bowel. These
authors described the use of their technique in one patient. The fistulas were created in the cecum and
jejunum by using standard Ponsky gastrostomy technique. After allowing 8 to 10 days for the tracts to
mature, 4-mm-diameter prototype endoscopes were inserted through the jejunostomy and cecostomy
to evaluate the small intestine. Although this technique must still be considered experimental, it offers
therapeutic capabilities throughout the small intestine.
Laparoscopy-assisted enteroscopy has been performed successfully in animals.80(4884) Following
laparoscopic inspection of the surface of the small bowel, a segment is exteriorized, an enterotomy is
performed, and then an endoscope is passed into the small intestine. Phillips et al.80(4885) found it
possible to examine the small bowel from the ligament of Treitz to the ileocecal valve by means of this
experimental technique. Furthermore, an arbitrarily selected segment of the bowel could also be
resected. No complications of the procedure were noted in any of the animals.
REFERENCES
1.
2.
Lewis B, Waye J. Small bowel enteroscopy in 1988: Pros and cons. Am J Gastroenterol
1988;83:799802.
Gilbert D, Buelow R, Chung R, et al. Status evaluation: Enteroscopy. Gastrointest Endosc
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
1991;37:6737.
Ogoshi K, Hara Y, Ashizawa S. New technique for small intestinal fiberoscopy. Gastrointest
Endosc 1973;20:645.
Morishita T, Hibi T, Asakura H. Endoscopy of the jejunal mucosa in human cholera.
Hibi T, Asakura H, Kobayashi K, et al. Alpha heavy chain disease lacking secretory alpha
chain with cobblestone appearance of the small intestine and duodenal ulcer demonstrated by
endoscopy. Gut 1982;23:4227.
Parker H, Agayoff J. Enteroscopy and small bowel biopsy utilizing a peroral colonoscope.
Barkin J, Schonfeld W, Thomsen S, et al. Enteroscopy and small bowel biopsy: An improved
technique for the diagnosis of small bowel disease. Gastrointest Endosc 1985;31:2157.
Foutch P, Sanowski R, Kelly S. Enteroscopy, a method for detection of small bowel tumors.
Lewis B, Waye J. Small bowel enteroscopy: A comparison of findings with push and sonde
enteroscopy in 81 patients with GI bleeding of obscure origin [Abstract]. Gastrointest Endosc
1988;34:207.
Foutch PG, Sawyer R, Sanowski R. Push-enteroscopy for diagnosis of patients with
gastrointestinal bleeding of obscure origin. Gastrointest Endosc 1990;36:33741.
Messer J, Romeu J, Waye J, et al. The value of proximal jejunoscopy in unexplained
gastrointestinal bleeding [Abstract]. Gastrointest Endosc 1984;30:15194.
Dykman DD, Killian SE. Initial experience with the Pentax VSB-P2900 enteroscope. Am J
Barkin JS, Chong J, Reiner DK. First-generation video enteroscope: Fourth-generation
push-type small bowel enteroscopy utilizing an overtube. Gastrointest Endosc 1994;40:7437.
Barkin J, Lewis B, Reiner D, et al. Diagnostic and therapeutic jejunoscopy with a new, longer
enteroscope. Gastrointest Endosc 1992;38:558.
Shimizu S, Tada M, Kawai K. Development of a new insertion technique in push-type
enteroscopy. Am J Gastroenterol 1987;82:8447.
Chong J, Tagle M, Barkin JS, Reiner DK. Small bowel push-type fiberoptic enteroscopy for
patients with occult gastrointestinal bleeding or suspected small bowel pathology. Am J
Davies GR, Benson MJ, Gertner DJ, et al. Diagnostic and therapeutic push type enteroscopy
in clinical use. Gut 1995;37:34652.
Harris A, Dabezies MA, Catalano MF, Krevsky B. Early experience with a video push
enteroscope. Gastrointest Endosc 1994;40:624.
Rockey DC, Cello JP. Evaluation of the gastrointestinal tract in patients with iron-deficiency
anemia. N Engl J Med 1993;329:16915.
Sahay R, Scott BB. Iron deficiency anaemiaHow far to investigate?. Gut 1993;34:14278.
Gostout CJ. Enteroscopy for unexplained iron-deficiency anemia: Identifying the patient with
sprue. Gastrointest Endosc 1993;39:769.
Green JA, Barkin JS, Gregg PA, Kohen K. Ulcerative jejunitis in refractory celiac disease:
Enteroscopic visualization. Gastrointest Endosc 1993;39:5845.
Askin MP, Lewis BS. Push enteroscopic cauterization: Long-term follow-up of 83 patients with
bleeding small intestinal angiodysplasia. Gastrointest Endosc 1996;43:5803.
Lewis B. Perform PEJ not PED. Gastrointest Endosc 1990;36:3113.
Shike M, Wallach C, Likier H. Direct percutaneous endoscopic jejunostomies. Gastrointest
Endosc 1991;37:625.
Lewis B. Direct percutaneous endoscopic jejunostomy. Gastrointest Endosc 1991;37:493.
Lewis B, Mauer K, Bush A. The rapid placement of jejunal feeding tubes: The Seldinger
technique applied to the gut. Gastrointest Endosc 1990;36:73940.
Cohen M, Barkin J. Enteroscopy and enteroclysis: The combined procedure. Am J
29. McGovern R, Barkin J. Enteroscopy and enteroclysis: An improved method for combined
procedure. Gastrointest Radiol 1990;15:3278.
30. Bertoni G, Sassatelli R, Tansini P, et al. Jejunal polyps in familial adenomatous polyposis
assessed by push-type endoscopy. J Clin Gastroenterol 1993;17:3437, discussion 3478.
31. Pennazio M, Arrigoni A, Risio M, et al. Clinical evaluation of push-type enteroscopy.
Endoscopy 1995;27:16470.
32. Nakamura S, Iida M, Nakao Y, et al. Diagnostic value of push-type jejunal endoscopy in
primary jejunal carcinoma. Surg Endosc 1993;7:18890.
33. Gostout C, Bender C. Cholangiopancreatography, sphincterotomy and common duct stone
removal via Roux-en-Y limb enteroscopy. Gastroenterology 1988;95:15663.
34. Yang R, Laine L. Mucosal stripping: A complication of push enteroscopy. Gastrointest Endosc
1995;41:1568.
35. Tada M, Akasaka Y, Misaki F, et al. Clinical evaluation of a sonde-type small intestinal
fiberscope. Endoscopy 1977;9:338.
36. Tada M, Misaki F, Kawai K. Pediatric enteroscopy with a sonde-type small intestinal
fiberscope (SSIF VI). Gastrointest Endosc 1983;29:447.
37. Tada M, Kawai K. Small bowel endoscopy. Scand J Gastroenterol Suppl 1984;102:3952.
38.
Lewis B, Waye J. A comparison of 2 sonde-type small bowel enteroscopes in 100 patients:
The SSIF VII and the SSIF VI KAI. Gastroenterology 1988;94:A261.
39. Tada M, Shimizu S, Kawai K. A new transnasal sonde-type fiberscope (SSIF VII) as a
pan-enteroscope. Endoscopy 1986;18:1214.
40. Gostout CJ. Improving the withdrawal phase of sonde enteroscopy with the push-away
method. Gastrointest Endosc 1993;39:6972.
41. Lewis B, Waye J. Total small bowel enteroscopy. Gastrointest Endosc 1987;33:4358.
42. Dabezies M, Fisher R, Krevsky B. Video small bowel enteroscopy: Early experience with a
prototype instrument. Gastrointest Endosc 1991;37:602.
43. Lewis BS, Waye JD. Chronic gastrointestinal bleeding of obscure origin: Role of small bowel
enteroscopy. Gastroenterology 1988;94:111720.
44.
Barthel J, Vargo J, Sivak M. Assisted passive enteroscopy and gastrointestinal tract bleeding
of obscure origin [Abstract]. Gastrointest Endosc 1990;36:222.
45. Gostout C, Schroeder K, Burton D. Small bowel enteroscopy: An early experience in
gastrointestinal bleeding of unknown origin. Gastrointest Endosc 1991;37:58.
46. Morris AJ, Madhok R, Sturrock RD, et al. Enteroscopic diagnosis of small bowel ulceration in
patients receiving nonsteroidal anti-inflammatory drugs. Lancet 1991;337:520.
47. Lopez MJ, Cooley JS, Petros JG, et al. Complete intraoperative small-bowel endoscopy in the
evaluation of occult gastrointestinal bleeding using the sonde enteroscope. Arch Surg
1996;131:2727.
48. Morrissey J. Small intestinal fiberscope. Gastrointest Endosc 1973;20:76.
49. Ament M. A large investment for small intestinal endoscopy. Gastrointest Endosc
1983;29:5960.
50. Spiller RC, Parkins RA. Recurrent gastrointestinal bleeding of obscure origin: Report of 17
cases and a guide to logical management. Br J Surg 1983;70:48993.
51. Berner JS, Mauer K, Lewis BS. Push and sonde enteroscopy for the diagnosis of obscure
gastrointestinal bleeding. Am J Gastroenterol 1994;89:213942.
52. Lewis B, Kornbluth A, Waye J. Small bowel tumors: The yield of enteroscopy. Gut
1991;32:7635.
53. Lewis B. Small bowel enteroscopy. Lancet 1991;337:10934.
54. Hoffman JS, Cave DR, Birkett D. Intra-operative enteroscopy with a sonde intestinal
fiberscope. Gastrointest Endosc 1994;40:22930.
55. Flickinger EG, Stanforth AC, Sinar DR, et al. Intraoperative video panendoscopy for
diagnosing sites of chronic intestinal bleeding. Am J Surg 1989;157:13744.
56. Frank M, Brandt L, Boley S. Iatric submucosal hemorrhage: A pitfall of intraoperative
57. Bowden T. Endoscopy of the small intestine. Surg Clin North Am 1989;69:123747.
58. Lewis B, Wenger J, Waye J. Intraoperative enteroscopy versus small bowel enteroscopy in
patients with obscure GI bleeding. Am J Gastroenterol 1991;86:1714.
59. Bowden TA Jr, Hooks VH III, Teeslink CR, et al. Occult gastrointestinal bleeding: Locating the
cause. Am Surg 1980;46:807.
60. Strodel WE, Eckhauser FE, Knol JA, et al. Intraoperative fiberoptic endoscopy. Am Surg
1984;50:3404.
61. Mathus-Vliegen E, Tytgat G. Intraoperative endoscopy: Technique, indications and results.
62. Apelgren KN, Vargish T, Al-Kawas F. Principles for use of intraoperative enteroscopy for
hemorrhage from the small bowel. Am Surg 1988;54:858.
63. Besa L, Ohri S, Hutton K, et al. Role of intraoperative enteroscopy in obscure gastrointestinal
bleeding of small bowel origin. Br J Surg 1991;78:1925.
64. Lau W, Wong S, Yuen W, et al. Intraoperative enteroscopy for bleeding angiodysplasias of
the small intestine. Surg Gynecol Obstet 1989;168:3414.
65. Lau WY. Intraoperative enteroscopyIndications and limitations. Gastrointest Endosc
1990;36:26871.
66. Ress AM, Benacci JC, Sarr MG. Efficacy of intraoperative enteroscopy in diagnosis and
prevention of recurrent, occult gastrointestinal bleeding. Am J Surg 1992;163:949.
67. Webb W. Intraoperative endoscopy of the total gastrointestinal tract in familial polyposis
68. Rodriguez-Bigas MA, Penetrante RB, Herrera L, Petrelli NJ. Intraoperative small bowel
enteroscopy in familial adenomatous and familial juvenile polyposis. Gastrointest Endosc
1995;42:5604.
69. Asbun HJ, Villa-Gomez G, Foianini J, et al. Operative enteroscopy. A useful tool in the
evaluation and intervention of bilioenteric anastomoses. Surg Endosc 1995;9:10935.
70. Smedh K, Olaison G, Nystrom PO, Sjodahl R. Intraoperative enteroscopy in Crohn's disease.
Br J Surg 1993;80:897900.
71. Lescut D, Vanco D, Bonniere P, et al. Perioperative endoscopy of the whole small bowel in
Crohn's disease. Gut 1993;34:6479.
72. Klein O, Colombel JF, Lescut D, et al. Remaining small bowel endoscopic lesions at surgery
have no influence on early anastomotic recurrences in Crohn's disease. Am J Gastroenterol
1995;90:194952.
73. Duggan C, Shamberger RC, Antonioli D, Leichtner AM. Intraoperative enteroscopy in the
diagnosis of partial intestinal obstruction in infancy. Dig Dis Sci 1995;40:22368.
74. Turck D, Bonnevalle M, Gottrand F, Farriaux JP. Intraoperative endoscopic diagnosis of
heterotopic gastric mucosa in the ileum causing recurrent acute intussusception. J Pediatr
75. Whelan RL, Buls JG, Goldberg SM, Rothenberger DA. Intraoperative endoscopy. University of
Minnesota experience. Am Surg 1989;55:2816.
76. Bowden TA Jr, Hooks VH III, Mansberger AR Jr. Intraoperative gastrointestinal endoscopy.
Ann Surg 1980;191:6807.
77.
Deyhle P, Jenny S, Fumagalli J. Endoscopy of the whole small intestine. Endoscopy
1972;4:1557.
78.
Classen M, Fruhmorgen P, Koch H. Peroral enteroscopy of the small and large intestine.
Endoscopy 1972;4:15762.
79. Frimberger E, Hagenmuller F, Classen M. Endostomy: A new approach to small-bowel
80. Phillips E, Hakim MH, Saxe A. Laparoendoscopy (laparoscopy assisted enteroscopy) and
partial resection of small bowel. Surg Endosc 1994;8:6868.
Chapter 51 Endoscopic Ultrasonography: Stomach and Duodenum

(4886)
(4887)
GIANCARLO CALETTI, M.D.
PAOLO BOCUS, M.D.
ENRICO RODA, M.D.
Early studies on endoscopic ultrasonography (EUS) were dedicated primarily to obtaining a better view
of the internal organs, especially those more difficult to explore by conventional
ultrasonography.1(4888) This initial work did not stress the possibility of studying the gastric wall, and
the display of the wall structure was in fact regarded as incidental.25(4889) As it became apparent
that EUS could define the gastric wall with greater resolution than any other imaging technique, the
focus of attention began to shift toward the intestinal wall itself.68(4890)
Stomach
The Gastric Wall
Ultrasonographically, the stomach wall comprises five layers of differing echogenicities (Figure 511).
By conventional usage, these are usually numbered from innermost (first layer) to outermost (fifth
layer). Because the intestinal wall also histologically comprises five layers (mucosa, muscularis
mucosae, submucosa, muscularis propria, and serosa), the simplest interpretation of the
ultrasonographic pattern was that the sonographic layers correspond directly to the histologic layers.
The hypothesis did not account for the fact that an ultrasound image is derived from echoes produced
by reflections of an ultrasound wavefront at interfaces between the anatomic tissue layers as well as
echoes created by the internal structure of the tissue layers.
(4891)Figure 511. Endoscopic ultrasonography (EUS) image (frequency 7.5 MHz, range 6
cm) of normal gastric wall. The probe (P) is in the body of a water-filled stomach (S). The
five-sonographic-layer structure is well visualized (arrows).
The model that has been proposed to explain visualization of the gastrointestinal (GI) tract wall as a
succession of distinct concentric layers is based on two assumptions: (1) The junction surface between
two histologic layers is an acoustic interface that generates echoes of high intensity and (2) a histologic
layer itself is presumed to be relatively homogeneous, with regard to ultrasound, and echo-poor
compared with the two limiting frontiers (junction surfaces) that define it. Consequently, the GI tract
wall should be simply conceived as a succession of hyperechogenic histologic discontinuities
separated by the histologic layers themselves, which are hypoechoic or nonechogenic.
Another problem is that all layers distinguishable by ultrasound have approximately the same
thickness. This is not true by actual measurement of the corresponding anatomic structures. The
muscularis mucosae and the serosa are only a few microns thick, and it is unlikely that they can
produce ultrasonographically evident layers at the frequencies employed in EUS. The third
ultrasonographic layer corresponds spatially to the submucosa, but this structure appears thicker by
ultrasound than by actual measurement. The fourth ultrasound layer corresponds spatially to the
muscularis propria, but it is thinner than the actual proper muscle layer. Furthermore, it is clear that a
one-to-one comparison of ultrasound and histologic layer thickness is not possible because the total
thickness of the ultrasound image is not the same as the total thickness of the histologic section.
"Depth" or "thickness" on the ultrasound image is actually a measurement of time, not distance.
Ultrasonographic thickness represents the time for an ultrasound pulse emitted by a transducer to
travel and return from a reflective tissue interface rather than the real distance in the tissue. Time (t) is
related to distance (d) by the equation d = vt, where v is the velocity of sound in the tissue being
imaged (see Chapter 11: Principles of Endoscopic Ultrasonography). Consequently, an infinitely thin
interface appears on the image as an echogenic layer with an apparent thickness "d." The latter value
(d) is about 300 to 600 m in endosonography, less with some new equipment, and constitutes one
possible definition of the "axial resolution" of the system. The shorter the pulse, the better the axial
resolution of the machine. Moreover, acoustic velocity is not equal in all tissues, and it is probable that
the different histologic layers of the gut have different acoustic impedances and thus different
thicknesses by ultrasound.
Based on careful correlation of ultrasound images with corresponding histologic sections, an
interpretation of ultrasound images of the GI tract has been proposed that takes into account the tissue
structure, the ultrasound properties of the GI tissue, and the physics of ultrasound. Although this
explanation is based on studies made on resected tissue, an interpretation of in vivo ultrasound images
can be extrapolated.912(4892)
If a more superficial layer is echo-rich and a deeper layer echo-poor, then the interface echo adds
thickness to the superficial layer and subtracts thickness from the deeper layer. If the more superficial
layer is echo-poor and the deeper layer echo-rich, then the interface echo is not distinguishable from
the deeper layer and therefore does not change the relative layer thickness.
Five major layers are seen on ultrasound images of normal stomach. The layer nearest to the mucosal
surface, usually seen as a thin echogenic line, is much thinner than the full thickness of the mucosa. It
corresponds to the superficial mucosa. The second layer, thicker than the first, is hypoechoic. This
layer corresponds to the deep mucosa, including the muscularis mucosae. The third central layer is the
most hyperechoic. It corresponds to the submucosa plus the acoustic interface between the
submucosa and the muscularis propria. This explains why it appears to be slightly thicker on
ultrasound than by histology. Its hyperechoic pattern is due to "backscattered" echoes. Small structures
(microvessels, lymphatics, muscular fibers) rich in collagen and other substances, such as fibrin,
elastin, and some lipids, add to the echogenicity of the layer. The fourth hypoechoic layer corresponds
to the muscularis propria minus the acoustic interface between the submucosa and the muscularis
propria. The fifth layer is hyperechoic and corresponds to the serosa and subserosal fat.
The thickness of echographic layers changes in vivo according to the degree of distention or the
functional state of the gastric wall. The wall of the stomach in the fundus and body may be thinner than
in the antrum, and the muscularis propria in the antropyloric region may be more evident compared
with other portions of the stomach.
Instrumentation
Ultrasound endoscopes may be classified according to the direction and method of scanning as either
mechanical sector scanning or linear array instruments. By means of a rotating acoustic mirror, the
former provides a 360-degree scan arranged perpendicular to the insertion tube of the instrument; the
latter may employ a rotating acoustic mirror or a transducer array to provide a scan in a plane oriented
parallel to the insertion tube. There is much greater experience with the sector scan echoendoscope.
The Olympus GF UM20/EU M20 (Olympus Optical Ltd, Tokyo, Japan) and the more recently
introduced GF UM3O/EU M30, with a accessory channel for endosonography-guided puncture and
switchable frequencies (7.5 and 12 MHz) are the most reliable instruments and the most suitable for
clinical practice.
Echoprobes (catheter probes) that can be introduced through the accessory channel of conventional
endoscopes have also been developed.1315(4893) These provide satisfactory images of relatively
small or flat lesions that are visualized endoscopically. They do not adequately visualize gastric
neoplasms of large diameter, the spread of such lesions into neighboring organs, or perimural lymph
nodes at some distance from the wall. Compared with these probes, conventional echoendoscopes
have a greater depth of penetration and usually permit more complete delineation of the gastric wall
circumference.
Technique of Exploration
The forward oblique optics of the Olympus GF UM20 ultrasound endoscope make it difficult to perform
an adequate visual assessment of the upper GI tract. For a complete study, it is recommended that a
standard endoscopic examination be performed before EUS. After anesthetizing the throat with
tetracaine (tablets) and administering intravenous diazepam (5 to 10 mg) and pentazocine (15 to 30
mg), the patient is placed in the left lateral decubitus position. The instrument is advanced into the
esophagus or stomach and, if necessary, into the descending duodenum. Glucagon (0.5 to 1 mg) can
be administered intravenously to achieve gastric atony and satisfactory distention of the wall.
EUS may be performed three ways: (1) by direct apposition of the transducer on the mucosa, (2) with a
small water-filled balloon covering the ultrasonic probe, or (3) with instillation of deaerated water (500
to 600 ml) into the stomach. The first is most useful for the investigation of extraluminal structures. The
water-filled balloon separates the transducer probe itself from the intestinal wall, which in turn focuses
the ultrasound scan on the wall and adjacent organs. With this technique, the esophageal wall is
adequately explored, but extraluminal structures and the gastric wall are only partially observed. In fact,
organs outside the gastric wall are often too far from the focus for satisfactory exploration with this
method. Furthermore, the gastric wall is not distended, so its characteristic layers may not be clearly
defined.
Direct instillation of deaerated water into the stomach results in satisfactory distention of the stomach,
and it also enhances the visualization of the entire structure of the gastric wall through the water
interface. This technique is also very useful for localization and study of small structures outside the
wall. EUS examination can sometimes be hampered by the presence of mucus and air bubbles within
the stomach. This phenomenon is commonly associated with deep ulcerative lesions or diffuse
submucosal processes. To reduce these artifacts, it is necessary to achieve nearly complete aspiration
of the gastric contents at the end of the routine endoscopic examination, if performed just before the
EUS. Moreover, a mucolytic-antifoam solution ingested a few minutes before beginning the EUS helps
to circumvent this problem.16(4894) The endosonographer should be aware that failure to aspirate all
of the air from the stomach prior to installation of water creates an air/fluid interface. This type of
interface strongly reflects sound and can therefore produce unusual artifacts (Figure 512).17(4895)
(4896)Figure 512. EUS image (frequency 7.5 MHz, range 9 cm) of an artifact produced by
reflection of ultrasound at the air/fluid interface within the stomach. (From the collection of
Dr. M. V. Sivak, Jr.)
For a complete EUS examination of the stomach, we use five standardized positions within the gastric
lumen.18(4897) In the first position, the transducer is placed at the pylorus, to provide a 360-degree
scan of the distal antrum. The second position is achieved by withdrawing the instrument tip slightly
with deflection toward the lesser curvature. This maneuver usually causes the scanning plane to pass
through the lesser curvature at the angulus to demonstrate the antrum as well as the distal body of the
stomach. Withdrawal to the level of the angulus with straightening results in position three; again,
scanning is through the angulus, so that both the body and the antrum are visualized but their relative
positions are reversed compared with the image display for position two. Position four, a sector scan of
the body alone, is achieved by further withdrawal. The fifth position is at the cardia of the stomach. Just
prior to reaching this position, the instrument tip can be deflected toward the fundus for an en face view
of the wall in this region.
The position of the patient can be changed to facilitate scanning for a particular lesion. The right
decubitus position provides good visualization of the antral region and the greater curvature. The
supine position allows study of the anterior wall, and the prone position is useful for imaging the
posterior wall and pancreatic region. It is sometimes difficult to find and scan a small gastric lesion,
even though it is seen endoscopically, because the optical and ultrasound components of EUS
instruments have different imaging planes.
When the structure to be evaluated is some distance from the echoendoscope, the 7.5-MHz frequency
must be used, but the 12-MHz probe is better for detecting fine structural detail close to the instrument.
Generally, we use the 7.5-MHz frequency first, in order to locate the lesion and evaluate its size and
contour, and then use the 12-MHz frequency to precisely define changes within each sonographic
layer.
Technical Problems
To achieve satisfactory EUS exploration of the gastric wall, two technical problems must be
considered: (1) Two regions are not adequately visualized with the water-filled stomach method, even
with ideal techniquethe antral prepyloric region and the proximal part of the lesser curvature just
distal to the cardia.18(4898) (2) The minimal focal distances of the available instruments are 10 mm at
7.5 MHz and 15 mm for 12 MHz. We demonstrated in an in vitro study that visualization of the
sonographic layering is best when the probe is at least at 30 mm from the mucosa.9(4899) This ideal
distance is sometimes difficult to achieve in peculiar situations or in the blind areas mentioned
previously.
Gastric Tumors
EUS diagnosis of GI diseases is based on changes in the layer structure of the GI wall as
demonstrated by this imaging technique. Neoplasms appear as a disruption in the continuity of an
ultrasound layer or as diffuse thickening of one or more layers. Infiltrating gastric neoplasms produce
diffuse thickening of all five ultrasound layers without causing layer disruption. A change in the
echogenicity of the ultrasound layers is another sign of tissue pathology. Neoplasms are usually of
intermediate echogenicity: less echogenic than the central layer but more echogenic than the second
and fourth layers.
A uniform staging system for cancer, including GI cancers, was formulated by the American Joint
Committee on Cancer (AJCC) in cooperation with the International Union Against Cancer (UICC; Union
Internationale Contre le Cancer) in 1987.19,20(4900) Known as the TNM classification, this refers to
depth of infiltration of the primary tumor into or through the gut wall (T), the presence or absence of
malignant involvement of regional lymph nodes (N), and the presence or absence of distant
metastases (M). Tumors classified according to the TNM system are staged from stage 0 (best
prognosis) to stage IV (worst prognosis). The recently (1987) modified TNM classification is widely
accepted as the standard for clinical staging of malignant tumors.
According to clinical endosonographic staging, tumor infiltration (T) of the first three layers is called T1
(tumor involving mucosa and submucosa); invasion of the fourth layer, T2 (tumor involving muscularis
propria); invasion through the fifth layer, T3 (tumor penetrating the serosa); and direct invasion of
adjacent organs or structures, T4 (Table 511). It is usually impossible to differentiate the subserosa
from the serosa with currently available echoendoscopes. It can therefore be especially difficult to
differentiate T2 from T3 gastric cancers by EUS. The presence of ascites is the one exception, as the
outer margin of the serosa is well defined by the tissue/fluid interface.
TABLE 511
TNM Classification of Gastric Carcinoma
T: PRIMARY TUMOR
Tumor invades lamina propria or submucosa
T1
Tumor invades muscularis propria and subserosa
T2
Tumor invades serosa
T3
Tumor invades adjacent structures
T4
TABLE 511
TNM Classification of Gastric Carcinoma

Tx
N: REGIONAL LYMPH NODES
No regional lymph node metastasis
N0
Metastasis in perigastric lymph nodes within 3 cm of edge of
N1
tumor
Metastasis in perigastric lymph nodes more than 3 cm from the
N2
edge of the primary tumor, or in lymph nodes along left gastric,
common hepatic, mesenteric, splenic, or celiac arteries
M: DISTANT METASTASIS
No distant metastasis
M0
Distant metastasis: hepatic metastases, peritoneal dissemination
M1
Mx
STAGE GROUPING
Stage IA
T1 N0 M0
Stage IB
T1 N1 M0, T2 N0 M0
Stage II
T1 N2 M0, T2 N1 M0, T3 N0 M0
Stage IIIA
T2 N2 M0, T3 N1 M0
Stage IIIB
T3 N2 M0, T4 N1 M0
Stage IV
T4 N2 M0, any T, any N, M1
Carcinoma
It is important to assess preoperatively the depth of tumor infiltration and the extent of lymph node
involvement.2124(4901) Conventional endoscopy permits an examination of only the inner surface of
stomach. EUS shows not only tumor depth and local spread but also the transition from pathologic to
normal wall. Moreover, EUS displays the relationship of the tumor to neighboring organs and the
presence of lymph node metastasis.2532(4902)
Cancer of the stomach usually appears at EUS as a hypoechoic, irregular-shaped mass. There are
characteristic changes in the sonographic layers (Figure 513).33(4903) Early changes include fusion
and thickening of the second and third layers. When there is a diffuse thickening, this is strictly located
to the visible area and a coarse, irregular hypoechoic mass is present. In cases where cancer invasion
is limited to the mucosa, the abnormal findings are limited sonographically to the second hypoechoic
layer, which corresponds to the mucosa, whereas the central hyperechoic (third) layer remains intact.
When a tumor has spread into the submucosa but has not yet reached the muscularis propria, the
central hyperechoic (third) layer is altered but is not yet completely disrupted. In cases where the
cancer has invaded the muscularis propria, the central hyperechoic layer is interrupted. If the
outermost hyperechoic (fifth) layer is disrupted, the cancer has reached the serosa. In infiltrative
carcinoma, involvement of the gastric wall is always transmural and more echogenic than in gastric
lymphoma; involvement of extramural structures and perigastric lymph nodes will also be detected by
EUS much more commonly.33(4904)
(4905)Figure 513. EUS image (frequency 7.5 MHz, range 9 cm) of T3-stage cancer. The
lesion is a transmural hypoechoic tumor with penetration into the serosa (black arrows). The
outermost hyperechoic layer (fifth) is disrupted (white arrows show the normal wall).
Pprobe; Swater-filled stomach.
A correct preoperative assessment of the extent of cancer invasion is a major factor in determining the
feasibility not only of surgery but also of the endoscopic management of early gastric cancer. In this
regard, particular attention must be given to the third or submucosal layer. In 24 of 42 cases of
advanced gastric carcinoma, Okai et al.34(4906) found that EUS demonstrated a growth pattern with a
liplike thickening of the gastric wall but no gross interruption of submucosal structures. This means that
whenever the tumor growth is visible on both sides of the submucosa leading to a liplike growth pattern
on EUS, the submucosa is invaded by tumor growth histologically.
On EUS, lymph nodes are seen as rounded, discrete structures. They are usually differentiated without
difficulty from blood vessels during real-time scanning. Malignant infiltration usually makes lymph
nodes more hypoechoic, round, and sharply defined. Inflammatory nodes tend to be more echogenic
and more ellipsoidal and to have a less-distinct margin. The larger the node, the more likely it is
malignant.35(4907) Ultrasonographic features of lymph nodes predictive of metastases in a study of
100 patients with esophageal cancer by Catalano et al.36(4908) were echo-poor (hypoechoic)
structure, sharply demarcated borders, rounded contour, and size greater than 10 mm (arranged in
increasing order of importance). The more of these features that were present, the greater the
likelihood that metastasis was present.
EUS has been shown in numerous studies to be highly accurate for assessment of depth of tumor
invasion and lymph node status in patients with gastric carcinoma (Table 512). Furthermore,
comparison studies have consistently demonstrated that EUS is more accurate than computed
tomography (CT) in the evaluation of depth of tumor infiltration and metastatic involvement of lymph
nodes (Table 513). There is some variation in accuracy with respect to assessment of depth of
invasion; accuracy is highest for T3 and lowest for T2 stage lesions (Table 514). This discrepancy is
due at least in part to the criteria for these two stages as set forth in the TNM classification (see
previously).
Reported Accuracy of Endoscopic Ultrasonography in

Locoregional Staging of Gastric Cancer
TABLE 512
AUTHOR
YEAR
STAGE (%)
N STAGE (%)
Kimura and Yamanaka118

Murata et al.119
1987
100
85
1988
146
79
Ohashi et al.120
Tio et al.29
1989
174
67
1989
72
84
68
Aibe et al.121
Saito et al.43
1989
67
73
69
1991
110
81
Akahoshi et al.37
Botet et al.122
1991
74
81
50
1991
50
92
78
Heintz et al.123
Kida et al.124
1991
19
79
72
1991
321
82
Rsch et al.125
Yasuda et al.126
1992
41
71
75
1992
500
79
Dittler et al.127
Grimm et al.128
1993
254
83
66
1993
147
78
87
Ziegler et al.129
1993
108
86
74
Nattermann et al.130
1993
50
82
78
Reported Accuracy of Endoscopic Ultrasonography in

Locoregional Staging of Gastric Cancer
TABLE 512
AUTHOR
YEAR
STAGE (%)
N STAGE (%)
1993
42
91
69
Caletti et al.39
2275
80
All cases T Stage
924
71
N Stage
Adapted from Rsch T. Endosonographic staging of gastric cancer: A review of literature
Endoscopic Ultrasonography Versus Computed

Tomography in the Staging of Gastric Carcinoma
TABLE 513
T STAGE (%)
AUTHOR
Botet et al.122
Nattermann et
al.130
Ziegler et al.129
YEAR
EUS
CT
N STAGE (%)
EUS
CT
1991
50
92
42
78
48
1993
50
82
25
77
33
1993
108
86
43
74
51
Adapted from Rsch T. Endosonographic staging of gastric cancer: A review of literature

EUSendoscopic ultrasonography; CTcomputed tomography.
Accuracy of Endoscopic Ultrasonography

Staging of Gastric Cancer According to T Stage
TABLE 514
STAGE
ACCURACY (%)
483
86
T1
301
64
T2
500
91
T3
143
80
T4
282
85
N0
311
71
N1
232
65
N2
Based on analysis of data from 12 reported series. Adapted from Rsch T.
Endosonographic staging of gastric cancer: A review of literature results.
Tio et al.29(4909) prospectively studied 72 patients with gastric cancer before surgery. The staging
accuracy of EUS was 76.9% for T1 tumors, 93.3% for T2, 81.3% for T3, and 87.5% for T4, with an
overall accuracy rate of 83.8%. Understaging occurred in 5.6% of cases and overstaging in 9.7%. With
regard to detection of metastatic lymph nodes, a false-positive diagnosis was made in 50% of patients,
whereas a false-negative diagnosis was made in only 14.3%. The accuracy of EUS in staging lymph
nodes was 50% for N0, 61.5% for N1, and 89.7% for N2. Somewhat later, Tio et al.31(4910) reported
an overall accuracy of 83% for staging tumor (T) infiltration (79% for T1, 89% for T2, 84% for T3, and
78% for T4). Overstaging and understaging occurred in 9% and 8% of cases, respectively. For
detection of regional lymph node metastasis, these investigators found overall diagnostic accuracy to
be 73%, with a sensitivity of 87%, specificity of 48%, and positive and negative predictive values of
74% and 68%, respectively. In detecting distant metastasis, EUS was found to have an overall
accuracy of 95%, with a sensitivity of 43%, a specificity of 100%, and positive and negative predictive
values of 100% and 95%, respectively. Similar results were reported by Akahoshi et al.37(4911) in a
series of 74 patients with gastric cancer; overall accuracy rate was 81.1% (92.5% for T1, 57.1% for T2,
100% for T3, and 60% for T4). EUS had a success rate of 50% in detecting perigastric lymph node
metastasis. Smith et al.38(4912) found a correlation between T stage of gastric carcinoma as
demonstrated by EUS and cancer recurrence after surgery with curative intent. Only 2 of 13 patients
with T1 or T2 lesions were found to have recurrent cancer (median follow-up 25 months), whereas 23
of 30 patients with either T3 or T4 lesions had recurrent cancer at initial EUS; 22 of these patients with
advanced tumors died during follow-up.
In our prospective study of 42 patients with gastric cancer, EUS correctly assessed depth of invasion in
91%: 83% for T1, 100% for T2 (including three linitis plastica), 86% for T3, and 100% for T4.39(4913)
One early gastric cancer (type III) was overstaged; understaging occurred in 2 patients. Sensitivities for
detection of metastatic and reactive lymph nodes were 56% and 29%, respectively. For the metastatic
group, positive predictive value was 93%, specificity was 93%, and negative predictive value was 54%.
Diagnostic accuracy was 69%.
EUS is assuming an increasingly important role in some centers in the management of early gastric
cancer.40,41(4914) Specifically, the decision for local endoscopic therapy versus standard surgical
treatment is being based on EUS findings. Several endoscopic methods of treatment are available
including mucosectomy ("strip biopsy technique") and laser photoablation. Mucosectomy provides a
resection specimen for histopathologic assessment of depth of tumor invasion, whereas primary laser
treatment does not (see also Chapter 48: Benign and Malignant Tumors of the Stomach). Standard
sector and linear scanning echoendoscopes, which operate at the more familiar frequencies of 7.5 and
12 MHz, may not be ideally suited for the evaluation of small, superficial cancers. Newer dedicated
instruments and catheter probes that utilize higher frequencies (e.g., 20 MHz) provide better resolution,
albeit with lesser depth of penetration.15,41,42(4915) The catheter probe appears to have a particular
advantage in that it is relatively easy to place this type of device on the lesion.
The previously discussed studies demonstrated that EUS is very accurate in the assessment of depth
of tumor infiltration. Depending on whether the muscularis propria is intact or disrupted, early-stage
cancers can be distinguished from advanced carcinomas. However, inflammation around an
early-stage lesion may result in overstaging or understaging.34,39,43(4916) Because of the limited
depth of penetration at the ultrasound frequencies employed, EUS is not accurate in the diagnosis of
distant metastases, particularly liver metastases and peritoneal dissemination. Thus, transcutaneous
abdominal ultrasonography or CT is generally recommended for the complete staging of gastric
cancer.
Evaluation of regional lymph node status in esophageal or gastric tumors continues to be problematic.
The sensitivity and negative predictive value of EUS for the detection of metastatic lymph nodes is low
in our experience and that of other investigators,27,44,45(4917) although Tio et al.26,46(4918)
reported better results. This disparity is probably explained by the use of different criteria for detection
and assessment of lymph node involvement. In fact, Tio et al.26,46(4919) found EUS sensitivity to be
high and specificity low, whereas the opposite was true in our series and that of other
investigators.47,48(4920) Various methods have been proposed to enhance the detection of regional
lymph nodes by EUS. Aibe et al.49(4921) presented a novel approach in which oral administration of a
20% sesame seed oil emulsion appeared to enhance detection.
Gastric Lymphoma
Lymphomatous involvement of the gastric wall proceeds mainly by longitudinal (i.e., horizontal) growth
within the wall, whereas the growth of carcinoma is predominantly vertical, with early invasion of the
deep layers of the wall. Mucosal involvement in lymphoma is often less extensive than the infiltration of
the underlying layers. Extended infiltration of the second and third layers with localized mucosal
ulcerations is virtually pathognomonic at EUS of lymphoma. In the early stages of gastric lymphoma,
EUS discloses abnormal thickening of the second or the second plus the third layers of the wall, with
preservation of these layers as distinct structures (Figure 514). Thickening is diffuse with a typical
hypoechoic pattern in advanced cases, and the individual sonographic layers are no longer
distinguishable. These pathologic alterations are found not only with and around the lesions observed
at endoscopy but also in areas where the mucosa appears at endoscopy to be normal (Figure
515).33,50,51(4922)
(4923)Figure 514. EUS image (frequency 12 MHz, range 6 cm) of T1-stage lymphoma
(arrows) involving the greater curvature and the two faces of the gastric body. There is
thickening of the second layer; the third, fourth, and fifth layers are normal. Pprobe;
Swater-filled stomach.
(4924)Figure 515. EUS image (frequency 7.5 MHz, range 12 cm) of T4-stage lymphoma.
The lesion is a transmural hypoechoic tumor with penetration into adjacent structures. There is
diffuse thickening of the entire gastric wall (white arrows) with a typical hypoechoic pattern.
The individual sonographic layers are no longer distinguishable (block arrowheads show the
normal wall). Pprobe; Swater-filled stomach.
Several studies with EUS for the evaluation of gastric lymphoma have been
conducted.33,46,50,52(4925) In our series, EUS correctly determined the depth of invasion in 92% of
patients (96% for T1, 67% for T2, 100% for T3, 80% for T4).39(4926) Overstaging occurred in 2 cases;
understaging occurred in 1. Sensitivities for detection of metastatic and reactive lymph nodes were
44% and 64%, respectively. Positive predictive value was 100%, there being no false-positive result.
Specificity was 100%, negative predictive value was 72%, and diagnostic accuracy was 77%.
EUS can be useful in differentiating cancer from lymphoma. Stage for stage, lymphoma has a much
better prognosis than carcinoma, particularly the more advanced stages.53(4927) Preoperatively,
however, gastric carcinoma is sometimes difficult to differentiate from lymphoma.5456(4928) In one
study, contrast roentgenography detected a potentially malignant abnormality in 75% of cases, but it
was not specific for lymphoma in any instance.56(4929) The sensitivity of endoscopy was only slightly
higher in this study (85% vs. 75%). Moreover, only 60% of biopsy specimen gave a positive result for
lymphoma.56(4930)
In our prospective study of 86 patients with endoscopic findings that raised a suspicion of cancer or
lymphoma, EUS made a correct diagnosis of cancer in 35 of 42 patients (sensitivity 83%, positive
predictive value 87%, specificity 97%, negative predictive value 96%, diagnostic accuracy
95%).39(4931) The diagnosis of lymphoma was correct in 39 of 44 patients (sensitivity 89%, positive
predictive value 87%, specificity 97%, negative predictive value 97%, diagnostic accuracy 95%). Thus,
EUS together with endoscopic forceps biopsy resulted in an overall correct preoperative diagnosis in
83 of 86 patients (97%). Despite these results, the diagnosis of gastric malignancies is not the most
important role for EUS. It must never supplant biopsy, but in difficult cases, it can be of significant
assistance in making the correct diagnosis.
Differentiation at EUS between lymphoma, linitis plastica, and Mntrier's disease can be difficult. All
three lesions have more or less the same appearance; that is, extended thickening of the second and
third layers with preservation of layers as distinctive structures. In linitis, the thickening is hypoechoic
as in lymphoma, but it has a much greater longitudinal extent and sometimes circumferentially involves
the entire gastric wall (Figure 516). The thickening of Mntrier's disease is more localized and
hyperechoic rather than hypoechoic (Figure 517).
(4932)Figure 516. EUS findings (frequency 7.5 MHz, range 6 cm) in linitis plastica. There
is extended thickening of the second and third layers with preservation of the fourth and fifth
layers as distinctive structures (arrows). The thickened layers appear hypoechoic as in
lymphoma, and there is nearly complete circumferential involvement of the gastric wall.
Pprobe; Aascites.
(4933)Figure 517. EUS findings (frequency 7.5 MHz, range 9 cm) in Mntrier's disease
(arrows). There is extended thickening of only the second layer, with preservation of layers as
distinctive structures. Pprobe; Swater-filled stomach.
The response to chemotherapy can also be demonstrated by EUS.57(4934) Some investigators have
suggested that there may be a correlation between the pattern of gastric involvement as demonstrated
by EUS and the histopathologic cell type of the lymphoma.58(4935)
Palazzo et al.51(4936) compared endoscopic and endosonographic findings with histopathologic
analysis of resected specimens in 24 patients with primary gastric lymphoma. The depth of invasion
was correctly determined by EUS in 91.5% of patients. In this study, the overall accuracy for EUS in the
identification of lymph node involvement was 83% (sensitivity 100%, specificity 80%). EUS also
demonstrated that the extent of gastric wall involvement was greater than appreciated by endoscopy in
58% of cases. By comparison with resection specimens, however, EUS still underestimated the extent
of involvement, mainly with lesions of low-grade malignancy, in 37.5% of cases.
Schder et al.59(4937) found that preoperative assessment of patients with primary gastric lymphoma
resulted in a decrease in total gastrectomies and an increase in the number of resection specimens
with tumor-free margins. One of 10 patients studied preoperatively by EUS required total gastrectomy,
whereas 9 underwent partial gastric resection. Intraoperative EUS was additionally performed in 3
patients. All resection specimen had tumor-free margins. These results were compared with those for
a group of 23 patients who underwent operation prior to the use of EUS for staging. In this historical
comparison group, total gastrectomy was performed in 65%; margins of resection specimens were
tumor-free in 72%.
Other Gastric Tumors

EUS findings have been described in a variety of other upper GI tumors.6062(4938) Yoshikane et
al.61(4939) performed EUS in 29 patients with carcinoid tumors (5 gastric, 7 duodenal, 17 rectal). The
tumors were described as being typically hypoechoic, well-demarcated, homogeneous, round to oval
lesions located within the third layer or between the second and the third layers. Overall accuracy with
regard to assessment of the depth of invasion was 75%. EUS also detected the presence of
metastasis in nearby lymph nodes in 3 of 4 patients. EUS was performed in 22 patients with acquired
immunodeficiency syndrome (AIDS) and Kaposi's sarcoma by Zoller et al.62(4940) EUS findings were
that of a hypoechoic, heterogeneous lesion, with involvement mainly of the mucosal and submucosal
layers.
Large Gastric Folds

Gastric rugal folds may appear at endoscopy or upper GI x-ray series to be enlarged. Because there
are no endoscopic criteria for assessing the thickness of rugal folds, and because their apparent size is
influenced by several factors, such as distention of the stomach, in addition to involvement by
pathologic disorders, an impression that a fold is enlarged or "thickened" is purely subjective. In the
absence of any additional endoscopic abnormalities, the likelihood that this is due to disease is
relatively low. Nevertheless, any of several well-known disorders may result in enlargement of rugal
folds. EUS may be of value in cases where the significance of this endoscopic observation remains
uncertain. EUS may be especially useful when endoscopic biopsies have been negative or
inconclusive despite the suspicion of malignancy.63,64(4941) In such cases, it has sometimes been
necessary to resort to surgery to obtain a full-thickness biopsy of the stomach to establish a diagnosis.
Mendis et al.65(4942) performed EUS in 28 patients who were thought to have enlarged gastric folds
noted at endoscopy or by radiography. Large forceps biopsy obtained at the time of EUS was positive
for malignancy in 4 patients (16%), but negative in 3 further patients in whom endosonography had
disclosed thickening of the third and fourth hypoechoic layers. On the basis of the EUS findings, the
latter 3 patients underwent surgery, at which adenocarcinoma was found. EUS also demonstrated
gastric varices in 4 patients; biopsies were not obtained in these patients. During follow-up (mean 35
months), no patient with a negative biopsy and EUS findings developed evidence of malignancy.
Avunduk et al.66(4943) found that 18 of 32 patients who underwent EUS because of enlarged gastric
folds had Helicobacter pylori infection and thickening of the first three sonographic layers of the gastric
wall. After treatment of the infection, EUS demonstrated resolution of the thickening for all three
involved layers.
Gastritis cystica profunda is an unusual condition characterized by small cysts in the submucosal layer
of the stomach.6769(4944) It is known in the literature by several terms, including diffuse submucosal
cysts, submucosal heterotopic gastric glands, gastritis cystica polyposa, and diffuse heterotopic cystic
malformation. Pathogenesis is uncertain. Although the lesion is thought not to be precancerous, in
some cases gastritis cystica profunda has been found in association with adenocarcinoma.68,69(4945)
In one reported case, carcinoma was multifocal within the stomach.68(4946) Enlarged gastric folds
have been noted at endoscopy and on x-ray contrast studies, either as a solitary finding or associated
with other abnormalities. The diagnosis can only be suspected at endoscopy, but EUS demonstrates
the characteristic findings of small cysts within a thickened submucosal layer.6769(4947)
Gastric Ulcer
A role of EUS in imaging gastric ulcers has not been well defined. EUS may detect the presence of
malignancy in an ulcer by demonstrating disruption of layers and infiltration into surrounding tissue.
However, some of these ultrasonic findings are nonspecific and could be due to edema beneath the
base of a benign ulcer. Furthermore, it may be difficult to differentiate the endosonographic
appearance of an ulcer scar from that of a submucosal tumor.70(4948)
EUS can be used to demonstrate the successive stages of ulcer healing.71(4949) Some investigators
also believe that information provided by EUS may prove useful in differentiating between intractable
ulcers and those that can be expected to heal.72(4950) Niwa et al.,73(4951) in a complex and unusual
study, attempted to assess this point. EUS features of intractable gastric ulcers included depth
extending through the muscularis propria, persistent edema in the ulcer base following therapy,
disruption of the submucosa and muscularis propria away from the ulcer, and a slow
ultrasonographically measured contraction rate during healing. Kimura et al.74(4952) found that EUS
findings could be correlated with the quality of ulcer healing. The relapse rate for ulcers judged by EUS
to be well healed was only 4.5%, whereas there was a relapse rate of 75% of the ulcers judged by EUS
to be poorly healed. Niwa et al.75(4953) used EUS to quantitate differences in gastric ulcer healing
between groups of patients receiving different therapy regimens. Further prospective EUS studies of
ulcer healing, including response to different antiulcer medications, are indicated.
Submucosal Tumors
The diagnosis of a gastric submucosal tumor has always been problematic. The typical appearance at
endoscopy of the gastric wall at the internal face of the lesion can be caused not only by benign or
malignant pathologic growths arising within the wall itself but also by compression of the wall by
adjacent organs that are either normal or diseased. Except to note the presence of a submucosal
lesion, endoscopy is of limited value in defining the true nature and origin of such a lesion and, in the
case of a nonepithelial neoplasm, its histopathology. Biopsy specimens, even when numerous, often
do not provide a histopathologic diagnosis because they are relatively small and because an
insufficient depth of tissue is sampled with standard biopsy techniques. Various techniques have been
developed for obtaining larger and deeper samples (see Chapter 23: Needle Biopsies of Submucosal
Lesions; see also Chapter 21: Endoscopic Ultrasonography with Linear Array Instruments: Anatomy
and EUS-Guided Fine-Needle Aspiration), but these are of no value in cases of extrinsic compression
and may not provide sufficient tissue for certain diagnoses.
EUS readily differentiates submucosal tumor from extrinsic compression of the upper digestive tract. It
provides precise information about the site, size, and nature (solid or cystic) of the lesion. The margins
of the lesion can be defined more clearly than with other imaging modalities. Furthermore, recognition
of acoustic differences between lesions and identification of the layer of origin within the wall frequently
suggest a histopathologic diagnosis.7681(4954)
Yasuda et al.76(4955) performed EUS in 80 patients with gastric submucosal lesions, including 55 with
subsequent histopathologic confirmation (24 leiomyomas, 3 leio-myosarcomas, 12 cysts, 7 aberrant
pancreas, 4 lipomas, 2 carcinoids, 1 fibroma, 1 granular cell myoblastoma, 1 eosinophilic granuloma),
and 59 patients with extraluminal compression. Based on the gastric wall structure demonstrated by
EUS, Yasuda et al.76(4956) were able to determine the origin of the lesion within the wall and to
confirm its submucosal nature. The most characteristic EUS appearance of a leiomyoma is that of a
tumor that is continuous with the muscularis propria (fourth hypoechoic layer) beneath the submucosa
(third hyperechoic layer) (Figure 518). Echograms of tumors under 3 cm in diameter demonstrated a
hypoechoic, homogeneous mass. For lesions more than 3 cm in diameter, the echogram usually
showed spotty patterns with hypo-echoic areas that correspond to necrotic regions of the tumor. All
three leiomyosarcomas in this series were over 3.5 cm in diameter, but there were no other
endosonograhic characteristics that differentiated this malignant tumor from a leiomyoma.
(4957)Figure 518. EUS findings (frequency 7.5 MHz, range 6 cm) in gastric leiomyoma (L).
A solid hypoechoic mass originates from the fourth layer; the first three layers are normal
(arrows). Pprobe; Swater-filled stomach.
Yasuda et al.76(4958) described aberrant pancreas as a solid mass with an echogram intermediate
between that of the hyperechoic submucosal layer and that of the hypoechoic muscularis propria. The
two carcinoid tumorsthe granular cell myoblastoma and the fibromawere all observed in the
submucosa, but their echopattern was not analyzed. Yasuda et al.76(4959) stated that the EUS
diagnosis of lipoma and leiomyoma is easy, but that differentiation between leiomyoma and
leiomyosarcoma is impossible. These investigators also found that EUS could not determine the
histopathologic nature of tumor masses observed in the submucosa. However, all of the 59 cases of
extraluminal compression were correctly defined by EUS. The cause of extrinsic gastric compression
was correctly defined by EUS in all 19 patients in the series of Motoo et al.82(4960) Adjacent normal
organs or structures accounted for the endoscopic appearance of a submucosal lesion in 16 cases (7
splenic artery, 5 spleen, 2 normal pancreas, 1 gallbladder, 1 colon); extragastric malignant tumors
were demonstrated in the remaining 3 patients (2 hepatic hemangiomas, 1 neurogenic tumor of the
omentum).
We reported our experience with 25 patients with submucosal lesions, including 10 with extrinsic
compressions and 15 in whom a tumor arose within the gastric wall.77(4961) EUS successfully
visualized the lesion and contributed to the correct diagnosis in 24 patients. In 10 of these patients,
EUS demonstrated clearly that the findings at endoscopy were attributable to extrinsic compression by
an adjacent organ that was either normal (2 left lobe of liver, 3 spleen, 1 abdominal aorta) or abnormal
(2 pancreatic cysts, 1 endocrine pancreatic tumor, 1 liver metastasis). In all of these patients, the
endosonographic appearance of the gastric wall overlying the lesion was normal. In another 10 of the
patients, EUS revealed a solid echogenic mass embedded within the fourth hypoechoic layer, an
appearance compatible with leiomyoma. In 1 patient, a round hyperechoic mass was seen in front of
the fourth hypoechoic layer. The latter was of normal thickness and appearance, and the lesion was
diagnosed as a lipoma (Figure 519). In 2 patients, EUS demonstrated that polypoid lesions,
seemingly solid at endoscopy, were actually liquid-filled structures, that is, gastric varices. An
observation such as this is of great significance when endoscopic polypectomy is under consideration.
(4962)Figure 519. EUS findings (frequency 7.5 MHz, range 9 cm) in gastric lipoma (L). A
solid hyperechoic mass originates from the third layer; the first two layers (arrows) are
normal. The fourth hypoechoic layer is clearly depicted and is normal. Pprobe;
Swater-filled stomach.
Tumors arising within the smooth muscle of the gastric wall can have different growth patterns. An
intrinsic growth pattern is manifested as compression of the gastric lumen. However, the lesion may
increase in size extrinsically with growth away from the serosal surface of the stomach. Large,
extrinsically growing tumors often cause considerable diagnostic confusion, particularly when they
exhibit prominent central cystic degeneration. Because the large size of these tumors tends to obscure
their site of origin from the gastric wall, they may occasionally be confused with a variety of unrelated
disorders including pseudocysts or tumors of the pancreas.83(4963)
Attempts to differentiate leiomyosarcoma from leiomyoma on the basis of size, shape, and
sonographic appearance have been generally disappointing. Unfortunately, the EUS appearance of
these two lesions is virtually the same, and the absence of distinguishing features makes differential
diagnosis difficult and relatively inaccurate.76,81,84(4964)
Aibe and Takemoto85(4965) found that malignant gastric submucosal tumors destroy the
endosonograhic layering pattern of the wall, whereas the layers remain intact if the tumor is benign.
Nakazawa et al.86(4966) affirm that irregularly shaped sonolucent areas within the tumor correspond
to liquefaction necrosis and that this appearance may be indicative of leiomyosarcoma. Tio et
al.87(4967) assert that leiomyosarcoma or leiomyoblastoma are visualized at EUS as a
heterogeneous, hypoechoic large mass with sharply demarcated boundaries or irregular margins,
occasionally associated with a central ulcer or fistulous tract. Five expert endosonographers reviewed
videotapes of 25 benign and 10 malignant gastric smooth muscle tumors in the study of Chak et
al.88(4968) Features predictive of malignancy included cystic spaces greater than 4 mm in diameter
within the tumor, echogenic foci more than 3 mm in diameter, and an irregular outer border. Other
features such as echogenicity and homogeneity of the tumor as well as an irregular luminal border
were not found to be as useful. Overall accuracy in predicting malignancy ranged from 56 to 77%,
indicating a certain degree of interobserver variation. However, when two or more predictive features
were present, accuracy in the diagnosis of malignancy ranged from 80 to 100% among experts.
Although large tumor size and the occurrence of gross necrosis or fistulization are also suggestive of
malignancy, only the biologic behavior of the tumor (frank local infiltration or remote metastasis)
defines true malignancy. In the absence of a histopathologic diagnosis, every investigator has
recommended either careful follow-up, usually to include endosonography, or, occasionally, surgical
resection.
It can be extremely difficult at EUS to differentiate pancreatic rest and leiomyoma. The usual
endosonographic features of the pancreatic rest are (1) mass in the third (submucosal) layer of the
stomach or duodenum; (2) a solid, homogeneous echopattern with some cystic or echoless areas that
correspond to drainage ducts; (3) possible thickening of the fourth layer (muscularis propria)
surrounding the mass.89(4969)
New devices are under development, such as the guillotine needle, that make it possible to obtain
larger, deeper tissue samples from solid submucosal tumors (see Chapter 23: Needle Biopsies of
Submucosal Lesions).9093(4970) Safe and effective use of these instruments depends on prior
demonstration by EUS of the solid nature of the lesion as well as its size and position in relation to the
gastric wall. Initial results are encouraging; early-stage malignancy is sometimes found when EUS
detects no evidence of malignancy. Rarely, a lesion arising within the gastric wall may prove at EUS to
be cystic (e.g., gastric duplication cyst), in which case endoscopic needle aspiration may be
appropriate.94(4971) However, submucosal lesions may also be shown by EUS to be vascular in
nature, as for example, gastric varices or even arteriovenous malformations;95(4972) attempts to
aspirate or obtain tissue specimens from such lesions are hazardous and contraindicated. Guillotine
needle biopsies are useful in deciding the management of asymptomatic patients with solid gastric
submucosal tumors. If the biopsy is negative for malignancy, such patients can be followed by EUS
and surgery avoided in many cases. Because many different types of lesions may produce the typical
appearance of a submucosal lesion at endoscopy, EUS should be performed before undertaking any
type of puncture of such a lesion.
Polyps With Arteriovenous Malformations

Tio and Tytgat96(4973) demonstrated that EUS can visualize vascular abnormalities in a gastric polyp
prior to endoscopic polypectomy. Aberrant vessels can be seen as echo-free structures in the
submucosa and are readily differentiated from the other nonvascular components of a polyp. Tio and
Tytgat96(4974) pointed out that EUS can demonstrate the size of the vascular structures within the
pedicle of a large polypoid lesion prior to endoscopic transection. This may call attention to the need for
additional techniques at polypectomyfor example, injection of alcohol or epinephrine into the
pedicleso that unexpected, profuse bleeding can often be avoided.
Portal Hypertension
EUS visualizes a large part of the portal venous system and has been proposed for the assessment of
portal hypertension.97,98(4975)
Gastric Varices
With the ultrasound probe positioned just below the gastroesophageal junction and the stomach filled
with water, EUS displays gastric varices as rounded, echo-free structures beneath the mucosa and
submucosa of the fundus (Figure 5110). This echo pattern may be found in patients with portal
hypertension in whom gastric varices are not visible at endoscopy.98(4976) In fact, EUS detects fundal
varices with greater frequency than endoscopy. Furthermore, it is frequently difficult at endoscopy to
distinguish varices from normal gastric rugal folds.99(4977) Boustiere et al.100(4978) compared
endoscopic and EUS findings in 80 patients with cirrhosis with those of 50 control subjects. Although
endoscopy provided better visualization of esophageal varices, gastric varices and evidence of portal
hypertension in its early stages were best detected by endosonography.
(4979)Figure 5110. EUS image (frequency 10 MHz, range 6 cm) of gastric fundus with
gastric varices (V; arrows). Pprobe; Swater-filled stomach.
Portal Hypertensive Gastropathy
EUS will visualize multiple small, rounded echo-free structures within the submucosa of the stomach in
all patients with portal hypertension and endoscopic findings of portal hypertensive gastropathy (see
Chapter 31: Variceal Bleeding) (see Figure 516).98102(4980)
Perigastric Veins
Dilated vessels can be demonstrated as numerous echo-free structures just external to the gastric wall
in nearly half of patients with portal hypertension.99(4981)
In the future, EUS could assume a significant clinical role in the evaluation of patients with portal
hypertension, especially in treatment planning. No single treatment plan can be suitable for all patients.
On the contrary, failure of a particular treatment may be attributable to incorrect selection of patients.
Liver function tests, abdominal ultrasonography, and endoscopy may not provide sufficient information
for making correct therapeutic decisions. In such cases, EUS provides the most detailed assessment
of the gastroesophageal venous circulation.
Infiltrative Disorders
Andriulli et al.103(4982) described the EUS findings in a patient with eosinophilic gastroenteritis that
primarily involved the muscularis propria of the antrum and distal ileum. EUS demonstrated thickening
of the fourth hypoechoic layer of the gastric antrum. Despite clinical improvement, the EUS findings
remained unchanged 6 months later and eosinophilia persisted.
EUS findings in a patient with systemic amyloidosis and GI involvement were described by Gandolfi et
al.104(4983) as extensive thickening of the gastric wall with loss of the normal layer structure of the
first three layers.
Upper GI Bleeding
There are a few reports of the use of transendoscopic Doppler ultrasound in the management of
patients with upper GI bleeding.
In a prospective study of 114 patients with acute ulcer bleeding, Kohler and Riemann105(4984) found
that transendoscopic Doppler ultrasound demonstrated blood vessels within the ulcer base in 62% of
cases where complete assessment was technically possible. Endoscopic and ultrasound findings were
in agreement in only 52% of cases. Injection therapy was performed in 65 of 66 Doppler-positive
cases. A second, follow-up endoscopy was performed 24 to 48 hr later, at which time a positive
Doppler signal was found in 22% of patients, and in these patients, further injection therapy was
performed. Five patients (7.7%) had recurrent bleeding. Unfortunately, this study did not include a
control group. Transendoscopic Doppler ultrasound was used by Jaspersen et al.106(4985) to identify
and guide endoscopic injection treatment of nonbleeding Dieulafoy's lesions in five patients. Doppler
ultrasound was also employed to assess the results of therapy, which was successful in all cases.
EUS findings in three patients with watermelon stomach, before and after laser photocoagulation, were
described by Parente et al.107(4986) Prior to treatment, a "spongy" appearance was noted involving
the second and third sonographic layers in areas where vascular structures were seen endoscopically.
This appearance was no longer evident after treatment.
Miscellaneous
Sakai et al.108(4987) found thickening of the third (submucosal) sonographic layer when EUS was
performed in 10 patients with acute gastric anisakiasis.
Duodenum
The Duodenal Wall
It is rarely possible to distinguish the five layers of the duodenal wall at EUS, probably due to focusing
problems. When the probe is within the duodenal cavity, it is often too close to the mucosa, so that the
optimum focusing distance cannot always be obtained (Figure 5111).18(4988)
(4989)Figure 5111. EUS image (frequency 7.5 MHz, range 6 cm) of T3-stage duodenal
lymphoma (arrows). This lesion is transmural and hypoechoic with penetration into the serosa.
Thickening is diffuse with a typical distinguishable. Pprobe; Dwater-filled duodenum.
Instrumentation
Duodenal endosonographic exploration can be performed with the standard sector scanning
echoendoscope. An endosonographic duodenoscope (JF-UM20; Olympus Optical, Tokyo, Japan) is
also available. Compared with other sector scanning instruments offered by Olympus Optical, this
instrument is longer and has a smaller distal tip that facilitates passage through the cricopharyngeus
and pylorus. It has an 80-degree field of view with a 15-degree lateral-retrograde orientation that
facilitates visualization of the main duodenal papilla. However, the scanning frequency of this
instrument is fixed, that is, it cannot be changed between 7.5 and 12 MHz, although it is available in
either of these frequencies.
Technique of Exploration
The three techniques described previously for endosonographic exploration of the stomach can also
be used for duodenal imaging. With direct apposition of the transducer on the mucosa, only the
structures outside the wall are visualized. Extraluminal organs and the duodenal wall are adequately
observed with the water-filled balloon method. In either case, air must be completely aspirated from the
lumen before beginning endosonography. It may be necessary to change the patient's position to
obtain a complete exploration of the second portion of the duodenum. In the water-filled lumen
technique, the instrument is first advanced to the descending duodenum with the patient in the
standard left lateral decubitus position; the patient is then turned to the supine or right lateral position
and 400 to 600 ml of deaerated water is infused into the lumen. With any of these techniques,
glucagon must never be administered because close contact of the ultrasonic probe with the wall is
difficult to achieve when the duodenum is hypotonic.
Indications
Duodenal Tumors (Ampullary Carcinoma)
Ampullary carcinoma has a more favorable prognosis compared with the other periampullary
cancers.109(4990) Preoperative differentiation between distal biliary, pancreatic, and ampullary
carcinomas can be difficult because they all arise within a small anatomic region. However, accurate
preoperative diagnosis and staging are essential to the appropriate selection of patients for surgery.
EUS diagnosis of ampullary carcinoma is based on demonstration of a hypoechoic tumor in the region
of the main duodenal papilla, with destruction or alteration of the normal ampullary anatomy, with or
without penetration into adjacent structures. Penetration is defined as a continuation of the main
hypoechoic tumor into surrounding structures (Figure 5112).110112(4991) The TNM classification of
ampullary tumors is shown in Table 515.
TABLE 515
TNM Classification of Ampullary
Cancer
T: PRIMARY TUMOR
No tumor assessable
T0
Tumor limited to ampulla of Vater
T1
Tumor invading duodenal wall, in particular the
T2
muscularis of the duodenum
Tumor invading 2 cm or less into pancreas
T3
Tumor invading more than 2 cm into pancreas or other
T4
adjacent organs, in particular the major blood vessels
Tx
N: REGIONAL LYMPH NODES
N0
Regional lymph node metastasis
N1
Nx
M: DISTANT METASTASIS
M0
Distant metastasis: hepatic metastasis, peritoneal
M1
dissemination, or lymph node metastasis along splenic
vein or at splenic hilum
Mx
(4992)Figure 5112. EUS image (frequency 7.5 MHz, range 6 cm) showing clearly
demarcated hypoechoic ampullary carcinoma (T) with penetration into the adjacent dilated
common bile duct (CBD) but with no evidence of infiltration into the adjacent pancreas (P).
Bwater-filled balloon.
Tio et al.113(4993) used EUS for the preoperative TNM staging of tumors in 43 patients with
pancreatic cancer and 24 patients with ampullary carcinomas. EUS results were correlated with the
histopathology of resected specimens. Early-stage tumors could be distinguished from advanced
stages of cancer with EUS. Demonstration of detailed images of ductular and parenchymal
abnormalities made it possible to distinguish pancreatic and ampullary carcinomas. The overall
accuracy of EUS for tumor classification was 92% and 88% for pancreatic and ampullary carcinoma,
respectively. The overall accuracy of EUS in assessing regional lymph node status in patients with
pancreatic carcinoma was 74% (sensitivity 91%, specificity 42%, positive predictive value 75%,
negative predictive value 71%). The overall accuracy rate of EUS in assessing regional lymph nodes in
ampullary carcinoma was 54% (sensitivity 80%, specificity 36%, positive predictive value 47%,
negative predictive value 71%). Distant metastasis are not demonstrable because of the limited
penetration depth of EUS.
Submucosal Tumors
EUS is useful in the localization and diagnosis of submucosal mass lesions by virtue of its ability to
separate the wall of the gut into sonographic layers and to define such lesions as arising within or
outside the wall (Figure 5113).80,114(4994) EUS findings associated with various types of lesions are
the same as those associated with the same tumor when it arises in the stomach and are described
earlier in this chapter.
(4995)Figure 5113. EUS image (frequency 7.5 MHz, range 12 cm) of submucosal mass in
the second part of the duodenum. The lumen is markedly narrowed, and a cystic lesion (C) is
behind the normal duodenal wall (arrows). At surgery, a cystoadenocarcinoma of the
pancreatic head was discovered. Bwater-filled balloon.
Other Tumors
EUS can accurately stage duodenal lymphomas and villous adenomas.115(4996) The decision for
local surgical resection versus pancreaticoduodenectomy can be guided by EUS findings.
Localization of neuroendocrine tumors involving the upper GI tract can be especially difficult, as they
are usually small and sometimes multiple. In several studies, EUS has been shown to be useful in
locating these lesions (see also Chapter 78: Endoscopic Ultrasonography of the Retroperitoneal
Organs).
Ruszniewski et al.116(4997) performed EUS and other imaging studies in 22 patients with the
Zollinger-Ellison syndrome who subsequently underwent exploratory surgery, which included
intraoperative ultrasonography and duodenal transillumination. EUS sensitivity for the detection of
duodenal wall tumors was 50%, compared with a sensitivity of 40% for endoscopy. The comparative
imaging study of Zimmer et al.117(4998) included 18 patients with a total of 25 primary neuroendocrine
tumors (1 stomach, 6 duodenum, 17 pancreas, 1 liver). Six of the 8 extrapancreatic tumors were
identified by EUS, 5 by somatostatin receptor scintigraphy, and only 1 by CT, transabdominal
ultrasonography, and magnetic resonance imaging.
REFERENCES
1.
Strohm WD. Limits of conventional abdominal sonography and feature of endoscopic

sonography. Scand J Gastroenterol Suppl 1982;94:712.
2. DiMagno EP, Buxton GL, Regan PT, et al. Ultrasonic endoscopy. Lancet 1980;1:62931.
3. DiMagno EP, Regan PT, Clain JE, et al. Human endoscopic ultrasonography.
4.
Lux G, Heyder N, Demling L. Endoscopic ultrasonographyTechnique, orientation and
diagnostic possibilities. Endoscopy 1982;4:2205.
5. Strohm WD, Phillip J, Hagenmller F, Classen M. Ultrasonic tomography by means of an
ultrasonic fiberendoscope. Endoscopy 1980;12:2414.
6. Caletti G, Bolondi L, Labo G. Anatomical aspects in ultrasonic endoscopy for the stomach.
7. Tanaka Y, Yasuda K, Aibe T, et al. Anatomical and pathological aspects in ultrasonic
endoscopy for GI tract. Scand J Gastroenterol Suppl 1984;94:4350.
8. Caletti G, Bolondi L, Labo G. Ultrasonic endoscopyThe gastrointestinal wall. Scand J
9. Bolondi L, Casanova P, Santi V, et al. The sonographic appearance of the normal gastric wall:
An in vitro study. Ultrasound Med Biol 1986;12:9918.
10. Aibe T, Fuji T, Okita T, Takemoto T. A fundamental study of normal layer structure of the
gastrointestinal wall visualized by endoscopic ultrasonography. Scand J Gastroenterol Suppl
1986;123:615.
11. Bolondi L, Caletti G, Casanova P, et al. Problems and variations in the interpretation of the
ultrasound feature of the normal upper and lower GI tract wall. Scand J Gastroenterol Suppl
1986;123:1626.
12. Kimmey MB, Martin RW, Hagitt RC, et al. Histologic correlates of gastrointestinal ultrasound
images. Gastroenterology 1989;96:43341.

13. Rsch T, Classen M. A new ultrasonic probe for endosonographic imaging of the upper GI
14. Kimmey MB, Martin RW, Silverstein FE. Endoscopic ultrasound probes. Gastrointest Endosc
1990;36:S406.
15. Maruta S, Tsukamoto Y, Niwa Y, et al. Evaluation of upper gastrointestinal tumors with a new
endoscopic ultrasound probe. Gastrointest Endosc 1994;40:6038.
16. Yiengpruksawan A, Lightdale CJ, Gerdes H, et al. Mucolytic-antifoam solution for reduction of
artifacts during endoscopic ultrasonography: A randomized controlled trial. Gastrointest
Endosc 1991;37:5436.
17. Grech P. Mirror-image artifact with endoscopic ultrasonography and reappraisal of the fluid-air
interface. Gastrointest Endosc 1993;39:7003.
18.
Caletti GC, Bolondi L, Barbara L. Instrumentation and scanning techniques. In Kawai K (ed).
Endoscopic Ultrasonography in Gastroenterology. Tokyo, New York: Igaku-Shoin, 1988;117.
19.
Sobin LH, Hermanek P, Hutter RVP. TNM classification of malignant tumors: A comparison
between the new (1987) and the old editions. Cancer 1988;6:23104.
20.
Hermanek P, Sabin LH. TNM Classification of Malignant Tumours. 4th ed. Heidelberg, Berlin,
New York: Springer-Verlag. 1987.
21.
Maruyama K. The most important prognostic factors for gastric cancer patients. A study using
univariate and multivariate analyses. Scand J Gastroenterol Suppl 1987;133:638.
22.
Sato T, Sakai Y, Ishiguro S, et al. Radiologic manifestations of early gastric lymphoma. Am J
Roentgenol 1986;146:5137.
23.
Hornig D, Hermanek P, Gall FP. The significance of the extent of proximal margins of
clearance in gastric cancer surgery. Scand J Gastroenterol Suppl 1987;133:6971.
24.
Kirchner R, Henke W, Wittekind C, et al. The surgeon's estimation of radicality and results of
pathologic examination at margins of resection in gastric cancer surgery. Scand J
25.
Caletti GC, Bolondi L, Brocchi E, et al. Staging of gastric cancer by means of endoscopic
26. Tio TL, den Hartog FCA, Tytgat GN. The role of endoscopic ultrasonography in assessing
local resectability of oesophagogastric malignancies. Accuracy, pitfalls, and predictability.
27. Aibe T, Ito T, Ioscida T, et al. Endoscopic ultrasonography of lymph nodes surrounding the
upper GI tract. Scand J Gastroenterol Suppl 1986;123:1649.
28. Heyder N. Endoscopic ultrasonography of tumours of the oesophagus and the stomach. Surg
Endosc 1987;1:1723.
29. Tio TL, Schouwink MH, Cikot RJ, Tytgat GN. Preoperative TNM classification of gastric
carcinoma by endosonography in comparison with the pathological TNM system: A
prospective study of 72 cases. Hepatogastroenterology 1989;36:516.
30. Tio TL, Coene PP, Schouwink MH, Tytgat GNJ. Esophagogastric carcinoma: Preoperative
TNM staging with endosonography. Radiology 1989;173:4117.
31. Tio TL, Coene PP, Luiken GJ, Tytgat GN. Endosonography in the clinical staging of
esophagogastric carcinoma. Gastrointest Endosc 1990;36(suppl 2):S210.
32. Ziegler K, Sanft C, Zeitz M, et al. Preoperative staging of gastrointestinal tumors by
endosonography. Surg Endosc 1990;4:15460.
33. Bolondi L, Casanova P, Caletti GC, et al. Primary gastric lymphoma versus gastric carcinoma:
Endoscopic US evaluation. Radiology 1987;165:8216.
34. Okai T, Yamakawa O, Matsuda N, et al. Analysis of gastric carcinoma growth by endoscopic
ultrasonography. Endoscopy 1991;23:1215.
35. Tio TL, Tytgat GNJ. Endoscopic ultrasonography in analyzing peri-intestinal lymph node
abnormality. Scand J Gastroenterol Suppl 1986;123:15863.
36. Catalano MF, Sivak MV Jr, Rice T, et al. Endosonographic features predictive of lymph node
metastasis. Gastrointest Endosc 1994;40:4426.
37. Akahoshi K, Misawa T, Fujishima H, et al. Preoperative evaluation of gastric cancer by

endoscopic ultrasound. Gut 1991;32:47982.
38. Smith JW, Brennan MF, Botet JF, et al. Preoperative endoscopic ultrasound can predict the
risk of recurrence after operation for gastric carcinoma. J Clin Oncol 1993;11:23805.
39. Caletti G, Ferrari A, Brocchi E, Barbara L. Accuracy of endoscopic ultrasonography in the
diagnosis and staging of gastric cancer and lymphoma. Surgery 1993;113:1427.
40. Motoo Y, Okai T, Songur Y, et al. Endoscopic therapy for early gastric cancer. Utility of
endosonography and evaluation of prognosis. J Clin Gastroenterol 1995;21:1723.
41. Takemoto T, Yanai H, Tada M, et al. Application of ultrasonic probes prior to endoscopic
resection of early gastric cancer. Endoscopy 1992;24(suppl 1):32933.
42. Yanai H, Fujimura H, Suzumi M, et al. Delineation of the gastric muscularis mucosae and
assessment of depth of invasion of early gastric cancer using a 20-megahertz endoscopic
ultrasound probe. Gastrointest Endosc 1993;39:50512.
43. Saito N, Takeshita K, Habu H, Endo M. The use of endoscopic ultrasound in determining the
depth of cancer invasion in patients with gastric cancer. Surg Endosc 1991;5:149.
44. Sugimachi K, Ohno S, Fujishima H, et al. Endoscopic ultrasonographic detection of
carcinomatous invasion and of lymph nodes in the thoracic esophagus. Surgery
1990;107:36671.
45. Lightdale CJ, Botet JF. Esophageal carcinoma: Pre-operative staging and evaluation of
anastomotic recurrence. Gastrointest Endosc 1990;36(suppl 2):S116.
46. Tio TL, den Hartog Jager FC, Tytgat GN. Endoscopic ultrasonography of non-Hodgkin
47.
Caletti GC, Brocchi E, Gibilaro M, et al. Sensitivity specificity and predictive value of
endoscopic ultrasonography (EUS) in the diagnosis and assessment of gastric cancer.
48. Dancyger H, Classen M. Endoscopic ultrasonography in esophageal diseases. Gastrointest
Endosc 1989;35:2205.
49. Aibe T, Fujimura H, Yanai H, et al. Endosonographic diagnosis of metastatic lymph nodes in
gastric carcinoma. Endoscopy 1992;24(suppl 1):3159.
50. Caletti GC, Lorena Z, Bolondi L, et al. Impact of endoscopic ultrasonography on diagnosis and
treatment of primary gastric lymphoma. Surgery 1988;103:31520.
51. Palazzo L, Roseau G, Ruskone-Fourmestraux A, et al. Endoscopic ultrasonography in the
local staging of primary gastric lymphoma. Endoscopy 1993;25:5028.
52. Tio TL, den Hartog Jager FC, Tytgat GN. Endoscopic ultrasonography in detection and
staging of gastric non-Hodgkin lymphoma. Comparison with gastroscopy, barium meal, and
computerized tomography scan. Scand J Gastroenterol Suppl 1986;123:528.
53. Hockey MS, Powell J, Crocker J, et al. Primary gastric lymphoma. Br J Surg 1987;74:4837.
54. Jones RE, Willis S, Innes DJ, et al. Primary gastric lymphoma. Problems in staging and
management. Am J Surg 1988;155:11823.
55. Fleming ID, Mitchell S, Dilawari RA. The role of surgery in the management of gastric
lymphoma. Cancer 1982;49:113541.
56. Shutze WP, Halpern NB. Gastric lymphoma. Surg Gynecol Obstet 1991;172:338.
57. Fujishima H, Misawa T, Maruoka A, et al. Staging and follow-up of primary gastric lymphoma
by endoscopic ultrasonography. Am J Gastroenterol 1991;86:71924.
58. Suekane H, Iida M, Yao T, et al. Endoscopic ultrasonography in primary gastric lymphoma:
Correlation with endoscopic and histologic findings. Gastrointest Endosc 1993;39:13945.
59. Schder G, Hildebrandt U, Kreissler-Haag D, et al. Role of endosonography in the surgical
management of non-Hodgkin's lymphoma of the stomach. Endoscopy 1993;25:50912.
60. Imamura A, Tochihara M, Natsui K, et al. Glomus tumor of the stomach: Endoscopic
ultrasonographic findings. Am J Gastroenterol 1994;89:2712.
61. Yoshikane H, Tsukamoto Y, Niwa Y, et al. Carcinoid tumors of the gastrointestinal tract:
Evaluation with endoscopic ultrasonography. Gastrointest Endosc 1993;39:37583.
62. Zoller WG, Bogner JR, Powitz F, et al. Endoscopic ultrasound in the diagnosis and staging of
gastrointestinal Kaposi's sarcoma. Endoscopy 1995;27:1916.

63. Andriulli A, Recchia S, De Angelis C, et al. Endoscopic ultrasonographic evaluation of patients
with biopsy negative gastric linitis plastica. Gastrointest Endosc 1990;36:6115.
64. Fujishima H, Misawa T, Chijiwa Y, et al. Scirrhous carcinoma of the stomach versus
hypertrophic gastritis: Findings at endoscopic US. Radiology 1991;181:197200.
66. Avunduk C, Navab F, Hampf F, Coughlin B. Prevalence of Helicobacter pylori infection in
patients with large gastric folds: Evaluation and follow-up with endoscopic ultrasound before
and after antimicrobial therapy. Am J Gastroenterol 1995;90:196973.
67. Okada M, Iizuka Y, Oh K, et al. Gastritis cystica profunda presenting as giant gastric mucosal
folds: The role of endoscopic ultrasonography and mucosectomy in the diagnostic work-up.
68. Nakamura T, Kobayashi S, Nakamura S. Diffuse submucosal cysts of the stomach with
multifocal gastric carcinoma: Diagnostic role of endoscopic ultrasonography. Gastrointest
Endosc 1993;39:879.
69. Niizawa M, Ishida H, Morikawa P, et al. Diffuse heterotopic submucosal cystic malformation of
the stomach: Ultrasonographic diagnosis. Gastrointest Radiol 1992;17:912.
70. Okai T, Songur Y, Watanabe H, et al. Endosonography of a submucosal tumor-like gastric
lesion caused by fibrotic ulcer healing. Endoscopy 1995;27:12831.
71. Yoshikane H, Tsukamoto Y, Niwa Y, et al. Sequential observation of gastric ulcer healing by
endoscopic ultrasonography. Scand J Gastroenterol 1994;29:66570.
72.
Nakazawa S, Segawa K, Yamanaka T. Evaluation of endoscopic ultrasonography on the
histological diagnosis of gastric ulcer. Curr Ther Res Clin Exp 1986;40:60412.
73. Niwa Y, Nakazawa S, Yoshino J, et al. Quantification of gastric ulcer healing by endoscopic
74. Kimura K, Yoshida Y, Kihira K, et al. Endoscopic ultrasonographic (EUS) evaluation of the
quality of gastric ulcer healing. Gastroenterol Jpn 1993;28(suppl 5):17885.
75. Niwa Y, Nakazawa S, Tsukamoto Y, et al. A new method for evaluating gastric ulcer healing
by endoscopic ultrasonography. Scand J Gastroenterol 1991;26:45764.
77. Caletti G, Zani L, Bolondi L, et al. Endoscopic ultrasonography in the diagnosis of gastric
submucosal tumor. Gastrointest Endosc 1989;35:4138.
78. Nakamura S, Iida M, Suekane H, et al. Endoscopic removal of gastric lipoma: Diagnostic
value of endoscopic ultrasonography. Am J Gastroenterol 1991;86:61921.
79. Rsch T, Lorenz R, Wichert AV, et al. Gastric fundus impression caused by splenic vessels:
Detection by endoscopic ultrasound. Endoscopy 1991;23:857.
80. Boyce GA, Sivak MV Jr, Rsch T, et al. Evaluation of submucosal upper gastrointestinal tract
lesions by endoscopic ultrasound. Gastrointest Endosc 1991;37:44954.
81. Hashimoto H, Mitsunaga A, Suzuki S, et al. Evaluation of endoscopic ultrasonography for
gastric tumors and presentation of three-dimensional display of endoscopic ultrasonography.
82. Motoo Y, Okai T, Ohta H, et al. Endoscopic ultrasonography in the diagnosis of extraluminal
compressions mimicking gastric submucosal tumors. Endoscopy 1994;26:23942.
83. Malhotra R, Evans R, Bhawan J, et al. A malignant gastric leiomyoblastoma presenting as an
infected pseudocyst of the pancreas. Am J Gastroenterol 1988;83:4526.
84. Yasuda K, Cho E, Nakajima M, Kawai K. Diagnosis of submucosal lesions of the upper
gastrointestinal tract by endoscopic ultrasonography. Gastrointest Endosc 1990;36(suppl
2):S1720.
85.
Aibe T, Takemoto T. Benign lesions of gastrointestinal tract. In Kawai K (ed). Endoscopic
Ultrasonography in Gastroenterology. Tokyo, New York: Igaku-Shoin, 1988;4455.
86. Nakazawa S, Yoshino J, Nakamura T, et al. Endoscopic ultrasonography of gastric myogenic
tumor. A comparative study between histology and ultrasonography. J Ultrasound Med

1989;8:3539.
87. Tio TL, den Hartog Jager FCA, Tytgat GNJ. Endoscopic ultrasonography for the evaluation of
smooth muscle tumors in the upper gastrointestinal tract. Gastrointest Endosc
1990;36:34250.
88.
Chak A, Canto M, Rsch T, et al. Endosonographic (EUS) differentiation of benign and
malignant smooth muscle tumors [Abstract]. Gastrointest Endosc 1996;43:417.
89. Minamoto T, Ueda H, Ooi A, et al. A limitation of endoscopic ultrasound: An unusual case of
early gastric cancer overlying a pancreatic rest. Am J Gastroenterol 1991;86:6226.
90.
Caletti GC, Gullini S, Brocchi E, et al. Endoscopic ultrasonography (EUS) with the guillotine
needle biopsy in the diagnosis of gastric submucosal tumors (GSMT). Gastrointest Endosc
1990;36:222.
91. Caletti GC, Brocchi E, Ferrari A, et al. Guillotine needle biopsy as a supplement of
92. Vilmann P, Hancke S, Henriksen FW, Jacobsen GK. Endosonographically-guided fine needle
aspiration biopsy of malignant lesions in the upper gastrointestinal tract. Endoscopy
1993;25:5237.
93. Wiersema MJ, Wiersema LM, Khusro Q, et al. Combined endosonography and fine-needle
aspiration cytology in the evaluation of gastrointestinal lesions. Gastrointest Endosc
1994;40:199206.
94. Ferrari AP Jr, Van Dam J, Carr-Locke DL. Endoscopic needle aspiration of a gastric
duplication cyst. Endoscopy 1995;27:2702.
95. Kono K, Sekikawa T, Iino H, et al. A case of arteriovenous malformation in the submucosal
layer of the stomach. J Gastroenterol 1994;29:3403.
96. Tio TL, Tytgat GNJ. Endoscopic ultrasonography of an arteriovenous malformation in a
gastric polyp. Endoscopy 1986;18:1568.
97. Caletti GC, Bolondi L, Zani L, et al. Detection of portal hypertension and esophageal varices
by means of endoscopic ultrasonography. Scand J Gastroenterol Suppl 1986;123:747.
98. Caletti GC, Brocchi E, Zani L, et al. The important role of EUS in the assessment of patients
with portal hypertension. Gastrointest Endosc 1988;34:1545.
99. Caletti G, Brocchi E, Baraldini M, et al. Assessment of portal hypertension by endoscopic
ultrasonography. Gastrointest Endosc 1990;36(suppl 2):S217.
100. Boustiere C, Dumas O, Jouffre C, et al. Endoscopic ultrasonography classification of gastric
varices in patients with cirrhosis. Comparison with endoscopic findings. J Hepatol
1993;19:26872.
101. McCormack TT, Sims J, Eyre-Brook I, et al. Gastric lesions in portal hypertension:
Inflammatory gastritis or congestive gastropathy?. Gut 1985;26:122632.
102. Papazian A, Braillon A, Dupas JL, et al. Portal hypertensive gastric mucosa: An endoscopic
study. Gut 1986;27:1199203.
103. Andriulli A, Recchia S, Valente G, et al. Endoscopic ultrasonography in eosinophilic infiltration
of gastric wall. Ital J Gastroenterol 1990;22:12932.
104. Gandolfi L, Colecchia A, Leo P, et al. Endoscopic ultrasonography in the diagnosis of
gastrointestinal myloid deposits: Clinical case report. Endoscopy 1995;27:1324.
105. Kohler B, Riemann JF. Endoscopic injection therapy of Forrest II and III gastroduodenal
ulcers guided by endoscopic Doppler ultrasound. Endoscopy 1993;25:21923.
106. Jaspersen D, Gaster CB, Koerner T, Hammar CH. Doppler-controlled injection treatment of
Dieulafoy's disease. J Gastroenterol Hepatol 1993;8:2679.
107. Parente F, Petrillo M, Vago L, Porro GB. The watermelon stomach: Clinical, endoscopic,
endosonographic, and therapeutic aspects in three cases. Endoscopy 1995;27:2036.
108. Sakai K, Ohtani A, Muta H, et al. Endoscopic ultrasonography findings in acute gastric
anisakiasis. Am J Gastroenterol 1992;87:161823.
109. Hayes DH, Bolton JS, Willis GW, et al. Carcinoma of the ampulla of Vater. Ann Surg
1987;206:5727.
110. Sivak MV Jr, Kaufman A. Endoscopic ultrasonography in the differential diagnosis of

pancreatic disease: A preliminary report. Scand J Gastroenterol Suppl 1986;123:1304.
111. Tio TL, Tytgat GNJ. Endoscopic ultrasonography in staging local resectability of pancreatic
and periampullary malignancy. Scand J Gastroenterol Suppl 1986;123:13542.
112. Yasuda K, Mukai H, Cho E, et al. The use of endoscopic ultrasonography in the detecting and
staging of carcinoma of the papilla of Vater. Endoscopy 1988;20:21822.
113. Tio TL, Tytgat TNJ, Cikot RJLM, et al. Ampullopancreatic carcinoma: Preoperative TNM
classification with endosonography. Radiology 1990;175:45561.
114. Stelling T, von Rooij WJ, Tio TL, et al. Pancreatitis associated with congenital duodenal
duplication cyst in an adult. Endoscopy 1987;19:1713.
115. Tio TL, Sie LH, Verbeek PC, et al. Endosonography in diagnosing and staging duodenal
villous adenoma. Gut 1992;33:5678.
116. Ruszniewski P, Amouyal P, Amouyal G, et al. Localization of gastrinomas by endoscopic
ultrasonography in patients with Zollinger-Ellison syndrome. Surgery 1995;117:62935.
117. Zimmer T, Ziegler K, Bader M, et al. Localisation of neuroendocrine tumours of the upper
gastrointestinal tract. Gut 1994;35:4715.
118.
Kimura K, Yamanaka T. Endoscopic ultrasonography in the assessment of depth of invasion
of gastric cancer. In Takemoto T, Kawai K (eds). Recent Topics of Digestive Endoscopy.
Amsterdam: Excerpta Medica, 1987;706.
119. Murata Y, Suzuki S, Hashimoto H. Endoscopic ultrasonography of the upper gastrointestinal
tract. Surg Endosc 1988;2:1803.
120. Ohashi S, Nakazawa S, Yoshino J. Endoscopic ultrasonography in the assessment of
invasive gastric cancer. Scand J Gastroenterol 1989;24:103948.
121.
Aibe T, Fujimura H, Noguchi T, et al. Endosonographic detection and staging of early gastric
cancer. In Dancygier H, Classen M (eds). Fifth International Symposium on Endoscopic
Ultrasonography (suppl). Munich: Verlag Demeter, 1989;718.
122. Botet JF, Lightdale CJ, Zauber AG, et al. Preoperative staging of gastric cancer: Comparison
of endoscopic US and dynamic CT. Radiology 1991;181:42632.
123. Heintz A, Junginger T. Endosonographisches staging von karinomen in speiserhre und
magen. Bildgebung 1991;58:48.
124.
Kida M, Yamada Y, Sakaguchi T, et al. The diagnosis of gastric cancer by endoscopic
ultrasonograph. Stomach Intestine 1991;26:6170.
125. Rsch T, Lorenz R, Zenker K, et al. Local staging and assessment of resectability in
carcinoma of esophagus, stomach and duodenum by endoscopic ultrasonography.
126.
Yasuda K, Mukai H, Cho E, et al. Evaluation of the degree of gastric cancer invasion by
endoscopic ultrasonography (EUS) for endoscopic treatment of early gastric cancer. Stomach
Intestine 1992;27:116774.
127. Dittler HJ, Siewert JR. Role of endoscopic ultrasonography in gastric carcinoma. Endoscopy
1993;25:1626.
128. Grimm H, Binmoeller KF, Hamper K, et al. Endosonography for preoperative locoregional
staging of esophageal and gastric cancer. Endoscopy 1993;25:22430.
129. Ziegler K, Sanft C, Zimmer T, et al. Comparison of computed tomography, endosonography,
and intraoperative assessment in TN staging of gastric carcinoma. Gut 1993;34:60410.
130. Nattermann C, Galbenu-Grunwald R, Nier H, et al. Endoskopischer ultraschall im TN-staging
des magenkarzinoms. Ein vergleich mit der computertomographie und der konventionellen
sonographie. Z Gesamte Inn Med 1993;48:604.
131. Rsch T. Endosonographic staging of gastric cancer: A review of literature results.
Chapter 52 Endoscopy in the Postoperative Upper Gastrointestinal

Tract
(4999)
(5000)
JOHN I. ALLEN, M.D.
MELODY O'CONNOR ALLEN, M.D.
There have been fundamental changes in the medical and surgical treatment of gastrointestinal
disorders since the first edition of this text was published. Two examples are the degree to which
gastric acid can be suppressed and the development of minimally invasive surgical techniques such as
laparoscopic cholecystectomy (LC).
Prolonged and virtually complete suppression of acid production by the stomach can be achieved using
proton pump inhibitors or high doses of H2 antagonists. Thus, fewer patients are being referred for
surgical treatment of acid-related disorders, and those who are tend to have complex or complicated
acid-peptic disease. In this subset of patients, accurate preoperative and postoperative endoscopic
evaluation is often invaluable. Unfortunately, with fewer patients undergoing surgery (such as Nissen
fundoplication), endoscopists may lack the experience needed to appreciate complex postoperative
anatomic alterations.
We surveyed the endoscopic experience at the Minneapolis Veterans Administration Medical Center
for the years 1977 to 1988; of 13,000 esophagogastroduodenoscopy (EGD) procedures performed,
10% were on patients who had previously undergone gastrointestinal surgery.1(5001) These data are
in agreement with data for the same period from other centers.2(5002) An informal poll of an
18-member Minneapolis-based private practice gastroenterology group indicated that fewer than 5% of
the esophagogastroduodenoscopies they performed in 1991 (n = 5219) were in patients who had
undergone previous intestinal surgery. Figure 521 illustrates the 10-year utilization rates for
Diagnosis-Related Group 155 (stomach, esophageal, and duodenal procedures) in the United
States.3(5003) Thus, several data sources suggest that fewer patients are undergoing upper intestinal
tract surgery, and thus fewer such patients will be seen by endoscopists.
(5004)Figure 521. Ten-year utilization rate of diagnosis-related group (DRG) 155

(esophageal, stomach, and duodenal procedures in patients greater than 17 years of age
without comorbidity).
The field of abdominal surgery is undergoing fundamental changes with the continuing development of
laparoscopic surgical techniques. Since 1987, when Mouret4(5005) performed the first LC, utilization of
this and other laparoscopic techniques has expanded rapidly.57(5006) It is estimated that half of the
37,500 general surgeons in the United States now are trained in LC. Of 535 cholecystectomies
performed in 1991 at the largest private hospital in Minneapolis (Abbott Northwestern), 56% were
laparoscopic procedures. There are intense efforts to develop laparoscopic common bile duct stone
removal, inguinal herniorrhaphy, appendectomy, fundoplication, pyloroplasty, highly selective
vagotomy, and bowel resection.7,8(5007) So also, the role of endoscopy in caring for patients
undergoing laparoscopic surgery is still evolving.
This chapter discusses the indications for and the technique of endoscopy in patients who have
undergone operations on the esophagus, stomach, duodenum, biliary tree, pancreas, and abdominal
aorta. Symptoms for which endoscopy are performed after surgery are summarized in Table 521.
Basic surgical principles as well as the anatomic and physiologic consequences of commonly
performed abdominal operations are described. Operative procedures are reviewed briefly, with
exclusion of finer details. There are many variations in surgical approach, technique, and materials
used; thus, knowledge of the operative anatomy of each patient is essential so that accurate
information can be obtained at endoscopy.
undergone operations on the esophagus, stomach, duodenum, biliary tree, pancreas, and abdominal
aorta. Symptoms for which endoscopy are performed after surgery are summarized in Table 521.
Basic surgical principles as well as the anatomic and physiologic consequences of commonly
performed abdominal operations are described. Operative procedures are reviewed briefly, with
exclusion of finer details. There are many variations in surgical approach, technique, and materials
used; thus, knowledge of the operative anatomy of each patient is essential so that accurate
information can be obtained at endoscopy.
Major Symptoms of
Postoperative Patients Referred for Endoscopy*
TABLE 521
SYMPTOM
Pain
Vomiting
Bleeding
Weight loss
Dumping
Anemia
Diarrhea
Early satiety or anorexia
Esophageal reflux
FREQUENCY
(%)
5166
4148
1439
33
32
18
17
12
48
* Data collected from published reports2 and unpublished data

from the authors' experience.
Postsurgical Endoscopy of the Esophagus

Antireflux Procedures
Surgical Techniques and Normal Endoscopic Appearance
Gastroesophageal reflux occurs at least once a month in 15 to 40% of adults.9,10(5008) With use of
high-dose H2-antagonist and proton pump inhibitors, medical therapy achieves adequate symptom
control in well over 90% of patients. However, 5 to 10% of patients have chronic, complex, or
complicated disease, the subgroup who may choose antireflux surgery over lifelong medical therapy. In
the hands of a skilled surgeon, the outcome is good in 80 to 90% of patients for up to 10 years.9(5009)
A randomized trial comparing medical with surgical therapy in complicated reflux disease confirmed
the efficacy of both approaches.11(5010)
Currently used antireflux operations include the Nissen fundoplication,1216(5011) Belsey Mark IV
procedure,17,18(5012) Hill repair,19,20(5013) Collis repair,21,22(5014) and placement of implantable
collars (Angelchik) that fit around the abdominal esophagus.23,24(5015) Landreneau et al.25(5016)
described a combined transthoracic parietal cell vagotomy with fundoplication. End-stage, complex
esophageal strictures, especially those causing a shortened esophagus, are treated by esophageal
resection.26(5017)
The Nissen fundoplication, the antireflux procedure most frequently chosen, is performed via an
abdominal approach. A fundoplication is created by wrapping the gastric fundus 360 degrees around
the distal esophagus (Figure 522). This creates an intraabdominal esophageal segment and a
"flap-valve" mechanism at the gastroesophageal junction. The wrap should be created with a
60-French (20-mm-diameter) dilator in the esophagus to ensure an adequate lumenal diameter, and
the external width of the wrap should be limited to 1 to 2 cm.12(5018) Both of these maneuvers
decrease the incidence of postoperative dysphagia. The gastric fundus is sutured to the anterior
serosal surface of the esophagus to prevent cephalad slippage of the stomach beneath the wrap. The
Belsey Mark IV repair differs from the Nissen procedure in that the fundic wrap extends only 270
degrees around the distal esophagus and the procedure is performed via a thoracotomy incision. Data
from manometric studies suggest that both a mechanical and a functional lower esophageal sphincter
result with the Nissen and the Belsey repairs.2731(5019) Not only are resting lower esophageal
sphincter pressures increased above preoperative levels, but there is also appropriate elevation of
sphincter pressure in response to cholinergic or pentagastrin stimulation.
(5020)Figure 522. Nissen fundoplication, with a 360-degree wrap of the gastric fundus
around the intraabdominal esophagus. Sutures fix the wrap to the anterior surface of the
esophagus to prevent intussusception of the stomach underneath the wrap ("slipped Nissen").
The Hill and Collis repairs are rarely used and have a higher failure rate than the techniques described
in the previous paragraph.20,22(5021) Both anchor the cardia of the stomach posteriorly to the
preaortic fascia and involve closure of the esophageal hiatus. The Angelchik prosthesis is a
doughnut-shaped silicone collar designed to create an artificial sphincter mechanism at the distal
esophagus. Experience with this device is limited, but reports of dislodgment or disruption of the
prosthesis suggest a high failure rate.3235(5022)
Intact Nissen and Belsey repairs have a similar endoscopic appearance. The esophageal mucosa
should be normal, and the distal esophagus must be carefully examined for residual Barrett's
epithelium. The gastroesophageal junction is narrowed, and the normal angulation between the distal
esophagus and the stomach lumen is lost. Normally, there is no protrusion of gastric mucosa into the
distal esophagus. The endoscope may be gently advanced through the lower sphincter, and the
position of the sphincter at rest and its approximate diameter and length should be estimated.
A retroflexed endoscopic view of a Nissen repair (Figure 523) shows a thickened fold of stomach wall
adjacent to the lumen of the fundus that spirals clockwise and posteriorly around the distal esophagus.
The wrap should not be patulous around the endoscope. Redundant folds of mucosa should be
flattened by air insufflation and carefully examined for paraesophageal herniation. Other than
diminished gastric volume, the stomach and duodenum are not altered in appearance.
(5023)Figure 523. Retroflexed close-up view of the 360 degree wrap of Nissen
fundoplication. Note that the wrap spirals in a clockwise and posterior manner around
endoscope.
Endoscopic Findings in Symptomatic Patients
Previously, clinicians caring for patients eligible for antireflux surgery were confronted by a bewildering
array of published studies dealing with therapeutic results and complications of antireflux
operations.3653(5024) Most of these earlier studies were anecdotal in nature and relied on the return
of reflux symptoms to define recurrence. Now, endoscopy, manometry, pH monitoring, barium contrast
studies, and radioisotope procedures more accurately define both symptom recurrence and problems
with the surgical repair.12,20,2730,39,42,43,54,55(5025) There is a prospective randomized trial
comparing medical and surgical therapy11(5026) and also attempts to standardize postoperative
evaluation by combining endoscopic findings with 24-hr pH monitoring.44(5027)
For postoperative patients in whom symptoms have developed, precise causative diagnosis is
mandatory because failure of antireflux operations may be due to errors in diagnosis, suboptimal
timing of surgery (too early or too late in the course of the disease), inadequate preoperative treatment
of esophagitis or strictures, technical mistakes, perioperative complications, or disease progression
after surgery.26,41(5028) After failed antireflux surgery, the major decision faced by physician and
patient is whether reoperation is appropriate or medical therapy combined with endoscopic treatments
will be adequate. Reoperation is often difficult and success rates are lower than with initial
operations.26,3640,4453,56(5029) Several causes of a failed fundoplication are illustrated in Figure
524.
(5030)Figure 524. Complications of a Nissen fundoplication. A, "Slipped Nissen." B, Tight

wrap. C, Paraesophageal hernia through wrap. D, Disrupted Nissen fundoplication.
Postoperative symptoms and conditions that mark the failure of antireflux surgery and usually
necessitate endoscopic evaluation include recurrence or persistence of reflux symptoms, dysphagia,
odynophagia, "gas-bloat" syndrome, epigastric pain, and Barrett's esophagus. The time at which these
symptoms or conditions develop after surgery suggests a probable cause.
Recurrence of reflux symptoms may result from a technically inadequate ("loose") wrap or from
postoperative disruption of the wrap (see Figure 524D).11,12,22,2634,36,40,43,56(5031)
Endoscopically, a wide-open lumen at the gastroesophageal junction and mucosal changes consistent
with reflux esophagitis suggest a disrupted wrap. A loose wrap may initially appear relatively intact but
may be patulous when viewed by retroflexion. Symptom relief in the early postoperative period followed
by recurrence of similar complaints suggests an initially successful repair that has become
disrupted.12,26(5032) Mean time to reflux recurrence in patients with a disrupted wrap was 7 months
in one study, and endoscopy accurately diagnosed this condition.49(5033) When there are recurrent
symptoms of reflux, endoscopy is the appropriate procedure for diagnosis and it may be followed by a
trial of medical therapy, although reoperation may be necessary if symptoms are severe.
When dysphagia or odynophagia occurs after an antireflux procedure, these causes should be
considered: (1) an intact antireflux procedure with an associated unrecognized motility disorder, (2) a
gastric wrap that is too tight, (3) a peptic stricture, (4) "slipped" Nissen, and (5)
cancer.1,11,12,22,2634,36,40,43,56(5034)
Mild dysphagia within 4 weeks after operation is usually transient, and patients can be reassured and
observed without treatment. If dysphagia persists beyond 1 to 2 months, or if symptoms are severe,
endoscopy should be performed first and then manometry, barium swallow with fluoroscopy, or cine
esophagography should be obtained as needed. Intractable symptoms occur if the original diagnosis
was incorrect (e.g., the patient actually has achalasia), if an existing peptic stricture was not adequately
dilated preoperatively, or because of problems related to the operation (e.g., slipped Nissen, disrupted
wrap, stricture due to a perioperative infection). Treatment is determined by cause and might include
medical therapy, dilation, myotomy, or another fundoplication.
A motor disorder of the esophagus may coexist with significant reflux esophagitis yet remain clinically
silent until a fundoplication is performed. It is mandatory that primary motility disorders (e.g.,
scleroderma) be differentiated from motility problems secondary to the acid reflux itself. Therefore, all
patients with reflux who are being considered for surgery must have manometric evaluation, and
preferably 24-hr pH monitoring as well, before fundoplication.12,26(5035)
If esophageal motility is abnormal and there is no endoscopic evidence of esophagitis, we assume that
a primary motility disorder is present and recommend that a loose surgical wrap or a Belsey repair be
performed. In patients with severe esophagitis and abnormal motility, we institute aggressive medical
therapy and document resolution of the esophagitis (by repeated endoscopy) prior to surgery. Once the
esophagitis has resolved, esophageal manometry is repeated; if normal, a standard Nissen
fundoplication is performed. If the motility study remains abnormal (but peristaltic contractions are
present), a coexisting primary motility disorder is suspected and the surgical plan is modified (see
earlier). In the rare patient in whom peristalsis is found to be absent at any point during evaluation, a
variant of achalasia is diagnosed and fundoplication is not performed.53(5036) By using this intensive
preoperative evaluation program, we are rarely confronted with a postoperative patient in whom
dysphagia develops because of an unsuspected motility disorder.
Other causes of dysphagia in the early postoperative period are a tight Nissen (see Figure 524B) or a
stricture. At endoscopy, both appear as a narrowing at the gastroesophageal junction with or without
evidence of esophagitis and scarring. The occurrence of a tight Nissen is minimized by appropriate
surgical technique (see earlier).12(5037) Peptic strictures can continue to cause dysphagia after
fundoplication if inadequately treated preoperatively. In addition to healing esophagitis preoperatively
by aggressive medical therapy, we routinely dilate peptic strictures until dysphagia resolves before
referring patients to surgery. Dilation of a peptic stricture after fundoplication is difficult and is
associated with an increased risk for disruption of the wrap. A rarer cause of postoperative stricture is
a perioperative abscess or a leak that heals with resultant scarring around the distal esophagus.
Treatment for a tight Nissen or stricture begins with careful dilation under endoscopic control.
"Through-the-scope" (TTS) hydrostatic balloon dilators or polyvinyl-type dilators for passage of a
guidewire are suitable. The use of dilators larger than 45 French (15 mm) increases the chance that
the wrap will be disrupted. Persistent dysphagia not responsive to dilation may require resection and
reconstruction of the esophagus by colonic interposition. The changes in surgical technique described
previously have reduced the frequency of persistent dysphagia after Nissen fundoplication to 5 to
10%.1,9,12,26(5038)
A slipped Nissen (see Figure 524A) results from intussusception of the body of the stomach through
the fundic wrap.53,56(5039) This produces an "hourglass" deformity on barium swallow (Figure 525).
Slippage occurs when the wrap is not adequately fixed to the esophagus during operation. Endoscopic
examination reveals that the gastroesophageal junction is proximal to the diaphragm and the fundic
wrap. Operative repair is necessary and dilation is not a therapeutic option.
(5040)Figure 525. Barium contrast radiograph of a slipped Nissen. Note the hourglass
appearance of the distal esophagus. (From Allen JI, Allen MO. Endoscopy as an alternative
and aid to re-operation. In McQuarrie DG, Humphrey EW [eds]. Reoperative General
Surgery. St. Louis, Mosby-Year Book, 1992; 10343.)
The "gas-bloat" syndrome (postfundoplication syndrome), consisting of early satiety, abdominal
bloating, and weight loss, occurs in 10 to 40% of patients after a Nissen fundoplication.12,26,57(5041)
Usually it is transient, of uncertain cause, and resolves spontaneously. It may be caused by a tight
fundic wrap or by gastric hypomotility resulting from intraoperative damage to the vagal nerve. The
presence of a tight wrap or persistent stricture is recognizable endoscopically, and patients may
respond dramatically to gentle dilation. Only 2% of patients with this syndrome have intractable
symptoms requiring reoperation.12,26,57(5042)
Epigastric pain after fundoplication may result from gastric ulcers58(5043) or a paraesophageal hernia
(see Figure 524C).59(5044) Ulceration may occur anywhere in the stomach. Initial therapy should be
pharmacologic, but if ulceration is related to gastric stasis as a result of inadvertent vagotomy during
fundoplication, a pyloroplasty may be required. Paraesophageal herniation of gastric fundus through an
intact wrap occurs rarely after fundoplication and is diagnosed by barium radiography (Figure 526)
and endoscopy. The potential for strangulation necrosis of the incarcerated segment usually
necessitates revision of the fundoplication.
(5045)Figure 526. Paraesophageal hernia through a Nissen fundoplication. Barium contrast

study demonstrates an intact Nissen fundoplication through which the anterior gastric wall has
herniated (underneath the Nissen wrap). Note the diminished luminal diameter of the distal
esophagus. Metallic clips mark the area of the wrap.
Barrett's esophagus is an acquired change in the mucosal lining of the distal esophagus as a result of
long-term gastroesophageal reflux (see also Chapter 40: Benign and Malignant Tumors of the
Esophagus).6063(5046) The risk of esophageal adenocarcinoma is reported to be 30- to 40-fold
greater in patients with Barrett's esophagus compared with that in patients with reflux without Barrett's
esophagus.6072(5047) Of patients with Barrett's esophagus, 10% also have distal esophageal
adenocarcinoma at the time of diagnosis.6072(5048) The frequent coincidence of Barrett's
esophagus with cancer has led to recommendations for endoscopic surveillance.63,6976(5049)
However, results from a prospective surveillance program have suggested that surveillance endoscopy
is ineffective or too costly.77(5050) There are no studies that demonstrate a reduction in the mortality
from esophageal cancer as a result of surveillance endoscopy in patients with Barrett's esophagus,
although a cooperative study is currently under way.
The role of antireflux surgery for patients with uncomplicated Barrett's esophagus is not well defined.
Patients with intractable reflux symptoms (without dysplasia or cancer) or a benign ulcer within
Barrett's epithelium may benefit from standard antireflux surgery.78(5051) Despite a report
demonstrating regression of Barrett's epithelium after antireflux surgery in 4 of 10 patients,79(5052)
regression after antireflux surgery does not occur with sufficient frequency to justify operative therapy
for all patients.60,63,65(5053) Adenocarcinoma has occurred within Barrett's mucosa after antireflux
procedures.63,65(5054) Management of patients with Barrett's esophagus is discussed in Chapter 40:
Benign and Malignant Tumors of the Esophagus.
Esophagomyotomy
Esophagomyotomy involves surgical sectioning of the lower esophageal sphincter. The endoscopic
appearance of the esophagus after an esophagomyotomy depends on preexisting conditions (see also
Chapter 39: Esophageal Motility and Miscellaneous Disorders). With longstanding achalasia, the
esophagus is cavernous and dilated; the folds and bends may simulate the sigmoid colon, and
passage of the endoscope into the stomach may be difficult. Most patients with achalasia have a
normal-appearing esophagus; after a successful myotomy, passage through the gastroesophageal
junction is easily performed. Difficulty passing through the gastroesophageal junction suggests the
presence of an underlying malignancy.
Surgical treatment of achalasia is successful in 80 to 94% of patients and achieves longer-lasting
results than pneumatic dilation.80,81(5055) Esophagomyotomy, however, has been reported to be
complicated by postoperative esophageal reflux in up to 25% of patients, usually when the myotomy is
extended into the proximal stomach.8084(5056) Rarely, esophagomyotomy is combined with an
antireflux procedure, but in patients with longstanding achalasia, this can result in dysphagia because a
physiologic obstruction is being replaced with a surgically constructed sphincter in patients with poor or
absent esophageal motility.8084(5057)
Mercer and Hill81(5058) emphasized the need for an adequate myotomy during the primary operation
for achalasia because failure rates for reoperation are high (25% of patients have only fair or poor
results after reoperation). Complications of esophagomyotomy for achalasia that should prompt
endoscopy include dysphagia, regurgitation, aspiration, hemorrhage, and development of squamous
cell carcinoma of the esophagus. Suspected esophageal perforation is a contraindication to endoscopy
and is diagnosed radiographically.
Most patients with a failed myotomy present with recurrent dysphagia or reflux symptoms. Endoscopy
and esophageal manometry are indicated to distinguish the various causes for recurrent symptoms
(failed myotomy, peptic stricture, stricture due to a perioperative leak, cancer). The demonstration at
manometry of a high-pressure lower esophageal sphincter shortly after operation suggests an
inadequate myotomy. Dysphagia that develops long after surgery may be due to healing of the
myotomy or the development of a stricture, especially if the patient has a history of chronic reflux
esophagitis.
Endoscopic treatment after a failed myotomy can be beneficial in patients who have strictures due to
reflux esophagitis or perioperative leaks. Such strictures can be managed using standard methods of
dilation. Pneumatic dilation is not appropriate for patients with a failed myotomy because of the high
risk of perforation.
It is uncertain whether achalasia should be regarded as a premalignant condition of the esophagus. It
has been reported that squamous cell carcinoma develops in 1 to 10% of patients, with an average
reported frequency of 3%.85(5059) Cancer is usually discovered at a late stage because the enlarged
esophagus accommodates the bulky tumor without causing symptoms. The question of cancer risk in
patients with achalasia is discussed in Chapter 39: Esophageal Motility and Miscellaneous Disorders.
Esophageal Resection
Esophageal resection is performed for carcinoma of the esophagus, gastroesophageal junction, or
proximal stomach and for intractable strictures of the esophagus resulting from caustic substance
ingestion or (rarely) severe reflux esophagitis. The type of disease (malignant vs. benign) and its
location determine the extent of esophageal resection. Extent of resection, in turn, dictates the method
for restoration of gastrointestinal continuity. Three methods of reconstruction are used: (1)
esophagogastrostomy, (2) esophagojejunostomy, or (3) esophageal replacement.
The tendency for esophageal cancer to spread rapidly to lymphatics or adjacent vital structures (e.g.,
aorta, trachea) limits the number of patients in whom curative resection is possible.8692(5060) If
curative resection is undertaken, a proximal margin of 10 cm is required to ensure removal of the
cancer because of its tendency for submucosal spread. Palliative resection or a bypass operation for
extensive carcinoma of the esophagus may be indicated. However, a variety of nonsurgical methods of
palliation are available (see also Chapter 42: Palliation of Malignant Obstruction: Dilation and Peroral
Endoprosthesis and Chapter 43: Palliation of Malignant Obstruction: Lasers and Tumor Probes).
Esophageal resection, which frequently requires both abdominal and thoracic incisions, has significant
morbidity and mortality. Early anastomotic leakage with sepsis and respiratory complications account
for an operative mortality (30 days) of 10 to 20%.87,8992(5061) Because of the high early morbidity
rate and a 5-year survival rate of only 10 to 15%, surgery for esophageal cancer should be attempted
only in certain patients selected on the basis of careful preoperative evaluation (see also Chapter 41:
Endoscopic Ultrasonography of the Esophagus).81,8992(5062)
Esophagogastrostomy is the most commonly used method of reconstruction after resection of the
middle or distal third of the esophagus for cancer (Figure 527).8992(5063) The anastomosis may be
hand-sewn or made with a stapling device.93(5064) Resection with esophagogastrostomy is not
performed for benign strictures because of severe postoperative reflux.94,95(5065) Because resection
of the distal esophagus includes resection of the vagi, concomitant pyloroplasty or pylorotomy is often
performed.
(5066)Figure 527. Proximal gastrectomy and esophagogastrostomy. Resection of the

proximal stomach and the gastroesophageal junction included resection of the vagi. Therefore,
pyloroplasty was necessary to prevent gastric stasis. Position of the gastroesophageal
anastomosis is on the anterior wall of the stomach, with a small gastric pouch extending
superiorly toward the staple line.
Esophagojejunostomy is performed after total gastrectomy for cancer96,97(5067) or for the
Zollinger-Ellison syndrome (Figure 528).98(5068) Free reflux of duodenal contents through the
esophagojejunostomy into the distal esophagus may cause severe esophagitis and late development
of a stricture.99(5069) If a loop esophagojejunostomy (Figure 528A) is performed, a concomitant
jejunojejunostomy is placed proximal to the esophagojejunostomy to allow duodenal juices to "bypass"
the anastomosis. If a Roux-en-Y reconstruction (Figure 528B) is performed, alkaline juices are
diverted into the jejunum 40 cm distal to the anastomosis to minimize reflux and the potential for
esophagitis.
(5070)Figure 528. Total gastrectomy with esophagojejunostomy. A, Gastrectomy with loop

esophagojejunostomy plus distal jejunojejunostomy. B, Gastrectomy with Roux-en-Y
esophagojejunostomy. This variation emphasizes the need to know surgical anatomy prior to
endoscopy.
When a Roux-en-Y reconstruction is performed, the jejunal limb is anastomosed to the distal
esophagus in an end-to-side manner. This creates an efferent limb that courses directly in a caudal
direction from the anastomosis plus a short afferent segment that terminates blindly in an oversewn
end (Figure 529). Knowledge of the surgical anatomy is critical to avoid mistaking the Roux-en-Y limb
for a loop esophagojejunostomy, which in turn can lead to forceful attempts to intubate the blind limb.
(5071)Figure 529. Esophagojejunostomy. Endoscope is positioned in the distal esophagus

looking into the Roux-en-Y limb. Note the dominant efferent limb that courses in a straight
caudad direction (left limb) and the short blind limb of the jejunum to the right. This view is
typical of an esophagojejunal anastomosis (end-to-side, respectively) of a Roux-en-Y
reconstruction (for both esophagojejunostomy and gastrojejunostomy) and should not be
mistaken for a loop jejunostomy.
Resection of all or most of the esophagus for carcinoma or a long corrosive stricture necessitates
esophageal replacement. Skin tubes and jejunal grafts have been used for this purpose, but colonic
interposition is preferred (Figure 5210).100102(5072) The interposed segment may be intrathoracic,
that is, in the position of the normal esophagus, substernal, or subcutaneous.103(5073)
(5074)Figure 5210. Esophageal resection. Esophagectomy, proximal gastrectomy, and

reconstruction with colonic interposition.
The normal endoscopic appearance of the interposed segment is similar to that of the normal colon as
seen colonoscopically. The proximal anastomosis is easily traversed (Figure 5211), whereupon a
long, sometimes convoluted segment of colon is seen. The triangular configuration of the haustral folds
of the transverse colon is usually apparent (Figure 5211). The distal anastomosis may be widely
patent and free reflux may be noted. Despite the alterations in anatomy and function, it is relatively
uncommon to discover inflammatory changes in the interposed colon.104(5075)
(5076)Figure 5211. Esophageal resection with colonic interposition. Endoscopic view from
the proximal anastomosis into the lumen of the interposed colonic segment.
There are few indications for endoscopy in the early postoperative period. When anastomotic leakage
is suspected (pain, fever, leukocytosis, prolonged ileus, and pleural effusion), endoscopy is
contraindicated until this diagnosis is eliminated by radiographic contrast studies.105(5077)
Late complications manifesting as pain or dysphagia after esophageal resection include alkaline reflux
esophagitis, anastomotic stricture, and recurrence of carcinoma.94,95,106109(5078) Alkaline or bile
reflux esophagitis is problematic after esophagogastrostomy. It results from the loss, inherent in the
surgery, of the normal esophageal and pyloric sphincters. Diagnosis is best confirmed by endoscopic
examination, with biopsies, when appropriate. Alkaline reflux esophagitis has the endoscopic
appearance of bile-stained, edematous, friable mucosa, often with ulceration and localized
hemorrhage in both the esophagus and the gastric remnant. Conser-vative measures and
pharmacologic therapy may be inadequate, and patients with severe symptoms may require complete
gastrectomy and conversion to a Roux-en-Y esophagojejunostomy for diversion of duodenal secretions
from the esophagus.94,95(5079) Alkaline reflux esophagitis is discussed in Chapter 44: Hiatus Hernia
and Peptic Diseases of the Esophagus.
Benign anastomotic strictures occur in 5 to 10% of patients who have undergone esophageal resection
and primary reconstruction (esophagogastrostomy or jejunostomy).8892(5080) Most strictures
respond to the standard methods for dilation of the esophagus (bouginage with mercury-filled dilators,
guidewire-directed passage of polyvinyl dilators, dilation with TTS hydrostatic balloon dilators; see
Chapter 44: Hiatus Hernia and Peptic Diseases of the Esophagus) (Figure 5212).95,105117(5081)
(5082)Figure 5212. X-ray film showing dilation of an esophagojejunostomy. The dilating

balloon is shown in the proximity of the end-to-end anastomosis staple line. (From Allen JI,
Allen MO. Endoscopy as an alternative and aid to re-operation. In McQuarrie DG, Humphrey
EW [eds]. Reoperative General Surgery. St. Louis, Mosby-Year Book, 1992; 10343.)
The most frequent late-occurring complication of colonic interposition is the development of a stricture
at the proximal anastomosis. It is generally attributed to anastomotic leakage early in the postoperative
course. Periodic dilation is effective and reoperation is seldom required. Strictures sometimes can be
tortuous in their course and the esophageal lumen difficult to identify. Diagnosis and TTS balloon
dilation of a complex proximal stricture in a patient with colonic interposition is illustrated in Figure
5213. After two operative attempts to correct this stricture, luminal patency was maintained long-term
by repeated balloon dilation.
(5083)Figure 5213. Contrast barium radiography of a proximal colonic anastomosis of a

colonic interposition and subsequent balloon dilation. A and B, Two views of the tortuous
angulated proximal anastomosis (imaged with barium contrast). C, Guidewire passed through
the anastomosis. D, Partially inflated dilating balloon. E, Fully inflated balloon with a small
residual waist at the stricture. (A-E, From Allen JI, Allen MO. Endoscopy as an alternative
and aid to re-operation. In McQuarrie DG, Humphrey EW [eds]. Reoperative General
Surgery. St. Louis, Mosby-Year Book, 1992; 10343.)
Late development of a stricture of the distal portion of the colonic segment owing to reflux of gastric
contents is rare, and endoscopic evaluation is indicated to exclude the possibility of recurrent tumor.
Endoscopic incision of densely fibrotic strictures that fail to respond to attempts at dilation by
established methods has been reported but experience with this technique is limited.118,119(5084)
When cancer recurs at the suture line, widespread recurrence or metastatic disease is probable.
Endoscopic dilation may be palliative in some patients.110(5085) Recurrent anastomotic carcinoma
frequently responds to laser (see Chapter 43: Palliation of Malignant Obstruction: Lasers and Tumor
Probes).
An unusual complication of colonic interposition occurs as a result of redundancy of the colonic
segment in its proximal portion, which may result in mechanical obstruction to swallowing at the point
where the colon passes underneath the clavicle (Figure 5214). Manual manipulation of the
subcutaneous portion of the colon may facilitate passage of food under the bony prominence. This
complication may not respond to dilation and resection of a small portion of interposed colon, and
reanastomosis to distal esophagus may be necessary.
(5086)Figure 5214. Colonic interposition. Protrusion of the proximal segment of the colon
over the clavicle produces mechanical obstruction and dysphagia.
Polyps may arise in the interposed colonic segment; these may be removed endoscopically.120(5087)
Postsurgical Endoscopy of the Stomach

Peptic Ulcer Surgery
Accurate endoscopic examination of patients who have undergone ulcer surgery requires a knowledge
of the procedures performed and postoperative anatomy. Surgical treatment for peptic ulcer disease is
based on elimination of cholinergic and hormonal stimuli of acid secretion and reduction of parietal cell
mass. Based on these principles, ulcer surgery involves truncal vagotomy with a drainage procedure or
vagotomy with gastric resection.121,122(5088) Highly selective vagotomy, an alternative to truncal
vagotomy, eliminates vagal stimulation in the parietal cell mass but maintains innervation of the antral
pump mechanism. For this reason, drainage procedures are not necessary with highly selective
therapy.
Ulcer surgery causes major alterations in gastric function, although the resultant symptoms are usually
mild and transient. Therefore, it is to be expected that patients will have complaints such as mild
diarrhea during the early postoperative period. However, the patient's symptomspersistent chronic
pain, for examplemay also be due to a problem unrelated to the surgery. Or, postoperative
symptoms may indicate a pathologic process such as postgastrectomy obstruction or ulceration.
Clinical evaluation includes precise documentation of symptoms, careful physical examination,
appropriate laboratory tests, and possibly endoscopy. Although radiographic examination is
occasionally helpful in certain situations, the accuracy of barium contrast studies is compromised by
the surgically altered anatomy and rapid emptying of the stomach.1,123,124(5089) False-positive and
false-negative radio-graphic results are most common with respect to mucosal abnormalities. For
these reasons, endoscopy has become the primary diagnostic procedure for evaluation of most
complications after ulcer surgery.1,123132(5090)
Close attention with specific notation should be given to the following points when endoscopy is
performed in any patient who has undergone ulcer surgery: type of anastomosis (Billroth I or II); size of
the gastric pouch; position and diameter of the gastrointestinal stoma; estimated quantity of bile
refluxed into the gastric pouch and esophagus; degree and anatomic distribution of any inflammation;
presence of retained food or secretions; any abnormal mucosal appearance suggesting intestinal
metaplasia, cancer, or ulceration; intraluminal sutures or staples; and the presence or absence of
antral motility.
Peptic Ulcer Surgery without Gastrectomy
The drainage procedures used in patients who undergo truncal vagotomy without resection include
pyloroplasty and gastroenterostomy. The Heineke-Mikulicz pyloroplasty (Figure 5215)133,134(5091)
involves a longitudinal incision through the muscular pylorus with approximation of the two ends of the
surgical incision to form a suture line that is perpendicular to the original incision. This procedure
disrupts the continuity of the pyloric sphincter mechanism and widens the gastric outlet.
Gastroenterostomy drainage (Figure 5216) is accomplished by anastomosing proximal jejunum to the
gastric antrum or body in a dependent area along the greater curvature.
(5092)Figure 5215. A and B, Truncal vagotomy and pyloroplasty.
(5093)Figure 5216. Truncal vagotomy and gastroenterostomy. A, Loop gastrojejunostomy

with the afferent limb (proximal jejunum) on the patient's right and the efferent limb (distal
jejunum) on the left. B, Loop gastrojejunostomy with the afferent and efferent limbs reversed.
After pyloroplasty, the appearance of the stomach is normal, with an identifiable antrum and a widely
patent, somewhat distorted pylorus (Figure 5217). Patients with active ulcer disease at the time of
operation may have gastric outlet obstruction during the early postoperative period. On endoscopy, the
pyloric channel may be edematous and difficult to intubate. By following the muscular folds of the
antrum, gentle pressure can be exerted to traverse the pyloric channel and enter the duodenum.
Prolonged nasogastric suction will usually resolve early postoperative outlet obstruction.
(5094)Figure 5217. Normal-appearing pyloroplasty with a triangular, widely patent pylorus.

Suture line of the Heineke-Mikulicz incision is on the anterior wall of the pyloric channel (left
upper edge of the photograph).
Endoscopy in a patient who has undergone a gastroenterostomy without resection requires specific
knowledge of the surgical anatomy before intubation (Figure 5218). The antrum and remaining
pylorus must be identified, and a thorough examination must be made of the duodenum for residual or
recurrent ulcers. In the early postoperative period, edema and friability may be present in the antrum
and around the gastrointestinal stoma. The pylorus may be inflamed and difficult to intubate (Figure
5219). The pylorus lies in a posterosuperior position relative to the stoma, the latter usually being
located in a dependent position on the greater curvature of the stomach. The efferent limb is often
identifiable by the ease with which it is entered. The afferent limb will usually contain bile.
(5095)Figure 5218. Gastroenterostomy. Pylorus appears on the right and gastroenterostomy

stoma is to the left (on the anterior gastric wall). Both are patent, with no inflammation. Antral
mucosa is normal.
(5096)Figure 5219. Gastroenterostomy, seventh postoperative day. Pylorus (superior margin

of photograph) is scarred and inflamed; surrounding antral mucosa is friable.
Gastroenterostomy stoma is in a dependent position on the greater curvature (lower margin of
photograph).
Gastrectomy with Billroth I, Billroth II, or Roux-en-Y Anastomosis
The operative techniques in partial gastrectomy are similar for a Billroth I (gastroduodenostomy) and
Billroth II (gastrojejunostomy) anastomosis.134(5097) The proximal stomach is transected at the
antrum, and the duodenum is divided distal to the pyloric ring. The usual method of constructing a
Billroth I anastomosis is the Hofmeister technique in which the upper half of the gastric resection line
beginning at the lesser curvature is closed and the lower half is used for anastomosis to the remaining
duodenum (Figure 5220A).134(5098)
(5099)Figure 5220. Antrectomy and gastroenterostomy (vagotomy not illustrated). A,

Antrectomy with Billroth I gastroduodenostomy. B, Antrectomy with Billroth II loop
gastrojejunostomy (afferent limb left, efferent limb right). C, Antrectomy with Billroth II loop
gastrojejunostomy (afferent and efferent limbs reversed).
t endoscopy, the size of the gastric pouch, the healed gastrotomy scar running longitudinally along the
lesser curvature (Figure 5221), and the stoma should be noted. Normally the stoma is widely patent,
is in dependent position, and leads directly into duodenum (Figure 5222). The presence of suture
material, reflux of bile into the gastric pouch or esophagus, ulceration, or inflammation should be
recorded and biopsies obtained if mucosal abnormalities are seen.
(5100)Figure 5221. Well-healed gastrotomy scar courses longitudinally along the lesser
curvature (see Figure 5220 for further illustration of the location of the gastrotomy scar).
(5101)Figure 5222. Gastrectomy with Billroth I anastomosis. Normal-appearing Billroth I

stoma leads directly to the duodenum.
After gastric resection, a Billroth II anastomosis (see Figure 5220B and C) is constructed using either
the Hofmeister or the Polya technique to connect the distal stomach to a proximal loop of
jejunum.134(5102) In the Polya technique, the anastomosis is formed by using the entire width of the
original transverse gastrostomy. With either the Hofmeister or the Polya technique, the two jejunal
limbs may be separated by a wide distance or narrow raphe depending on the width of the actual
stoma (Figure 5223).
(5103)Figure 5223. Antrectomy with a Billroth II gastrojejunostomy. A, Wide anastomosis

results in a wide separation of the jejunal limbs. B, Narrow anastomosis results in closely
approximated limbs separated by a narrow raphe.
At endoscopy, the area around the gastrointestinal stoma should be carefully inspected for sutures,
staples, and ulceration. An endoscopic description of the anatomic findings resulting from the surgery
is essential. Complete examination of the anastomosis, particularly the jejunal side, may require gentle
retroflexion of the endoscope within the stoma, a maneuver facilitated by the use of small-diameter
endoscopes with a 135-degree angle of view and a tight bending radius (Figure 5224). The efferent
limb is dependent, is the most easily entered, and appears to continue in a straight caudad direction
(Figure 5225). The opening of the afferent limb is more difficult to enter, may be tucked behind a fold
of the stoma, and may be situated to the left or right of the efferent limb, depending on the method of
surgical anastomosis. The afferent limb is identified by the presence of bile and the ampulla. With
intubation to the proximal duodenum, the closure site may be examined and a biopsy obtained to
diagnose retained antrum.135(5104)
(5105)Figure 5224. Jejunal side of a Billroth II anastomosis. Endoscope is retroflexed in the

stoma of the Billroth II anastomosis. Jejunal side of the anastomosis may be inspected for
inflammation or recurrent ulcers with this technique.
(5106)Figure 5225. Gastrectomy with a Billroth II anastomosis (3 years postresection) with

the jejunal limbs widely separated. Dominant limb (most easily entered) is usually efferent,
whereas the afferent limb may be situated on either side of the efferent limb and is often
tucked under a fold.
The Roux-en-Y gastrojejunostomy is primarily performed as a second operation when antrectomy and
Billroth II (sometimes Billroth I) anastomosis results in symptomatic alkaline
reflux.95,106,107,109,136138(5107) The operation involves division of the jejunum about 20 cm
distal to the ligament of Treitz, anastomosis of the proximal end of the efferent jejunal limb to the
gastric pouch, and a jejunojejunostomy (end-to-side) between the proximal intestinal limb and the distal
jejunum 40 cm beyond the gastrojejunostomy. The distance between the gastric pouch and the
jejunojejunostomy is a critical factor if bile reflux (too short) and recurrent ulceration (too long) are to be
avoided.
Endoscopic examination of the gastric remnant after resection with Roux-en-Y anastomosis is the
same as described for the Billroth II anastomosis. The gastrojejunostomy is a side-to-side anastomosis
with a short blind limb that must not be mistaken for a stenotic afferent jejunal limb (see Figure 529).
The anastomosis should be examined for the evidence of obstruction, inflammation, recurrent
ulceration, dysplasia, and carcinoma.
Endoscopy in the Early Postoperative Period
Bleeding and obstruction are the two major complications in the early postgastrectomy period for which
endoscopy may be requested.139,140(5108) Endoscopy is usually indicated for significant bleeding,
but early gastric outlet obstruction is often diagnosed radiographically, endoscopy being reserved for
selected patients.
Although endoscopy in early postoperative patients is not technically difficult, care must be exercised to
avoid disruption of the fresh surgical anastomosis (Figure 5226). A well-constructed anastomosis can
withstand gentle endoscopic examination even immediately after surgery if minimal insufflation is
used.139,140(5109) However, if an anastomotic leak is suspected, endoscopy is contraindicated; this
diagnosis must be established radiographically.
(5110)Figure 5226. Fresh gastrotomy scar at 24 hours postsurgery is not bleeding and is
well secured. This area can be a source of postoperative bleeding and should be carefully
examined if hemorrhage occurs in the immediate postoperative period.

Endoscopic localization of the site of bleeding that occurs in the early postgastrectomy period is
essential for diagnosis and appropriate choice of therapy. Minimum blood loss can occur as oozing at
the anastomotic site; this is of little consequence. Significant postgastrectomy bleeding (more than 4 U
of blood loss within 48 hr) may originate from operative suture lines, visible intraluminal sutures (Figure
5227), active ulcer disease, erosions caused by a nasogastric tube, gastritis, or varices. Bleeding may
be exacerbated by coagulopathy. Carefully performed endoscopy, with gentle washing and inspection
of the anastomoses and remaining gastric pouch, will usually identify bleeding sites.
(5111)Figure 5227. Gastrectomy with a Billroth I anastomosis. This is a fresh anastomosis

with blood oozing from stoma where suture material is visible.
Outlet obstruction in the early postoperative period may be mechanical (tight anastomosis and edema)
or functional (prolonged gastric atony or a nondependent stoma). It is suspected when attempts to
discontinue postoperative nasogastric drainage fail or when drainage is excessive.141(5112) A barium
contrast study is the initial diagnostic step for differentiation between mechanical and functional
obstruction. Patients with prolonged gastric outlet obstruction or marked malnutrition preoperatively
may require from weeks to months to regain normal gastric motility, especially after vagotomy. If
mechanical obstruction is suspected, endoscopy is indicated to evaluate stomal size and position, to
search for suture material that could compromise the lumen, or to diagnose severe inflammation and
edema at the stoma. Rarely, intractable postoperative anastomotic obstruction may require
reoperation.
Endoscopy in the Late Postoperative Period
The most common symptoms in the late postoperative period that prompt endoscopy are summarized
in Table 522. About 80% of patients with postoperative complications require endoscopy within the
first year after surgery.1,127(5113) Because specific symptom patterns often suggest particular
pathologic findings, the endoscopist should be familiar with the patient's symptoms, such as the pattern
of pain and its relation to food or medications, the nature of vomitus (undigested food, bilious
secretions), and the duration and rate of bleeding.
Indications for Endoscopy in the

Postoperative Patient
TABLE 522
Evaluation of appropriate symptoms
Evaluation of weight loss, anemia, and malabsorption
Definition of postoperative anatomy
Diagnosis of mucosal or submucosal lesions
Control of bleeding mucosal lesions
Dilation of strictures
Removal of foreign objects (suture, staples)
Disruption of bezoars
Gastrostomy tube placement
Laser therapy of obstructing lesions
Detection of primary or recurrent cancers
Surveillance of premalignant lesions
Aid to laparascopic surgical techniques
ENDOSCOPIC RETROGRADE CHOLANGIOPANCREATOGRAPHY
Biliary Duct
Indications for Endoscopy in the

Postoperative Patient
TABLE 522
Definition of postoperative anatomy

Diagnosis of intraductal abnormalities
Diagnosis of biliary fistulas
Sphincterotomy for retained stones
Stricture dilation
Stent placement
Gallstone dissolution or disintegration
Pancreatic Duct
Preoperative definition of duct dilation or pseudocysts
Definition of postoperative anatomy and anastomoses
Stent placement
Stricture dilation
Specific considerations having a bearing on endoscopy in postgastrectomy patients include the

presence of duodenogastric bile/alkaline reflux, gastric mucosal changes, recurrent ulceration,
mechanical obstruction, bleeding, and weight loss or malnutrition.121,122,142(5114) Gastric
carcinoma is occasionally found in the postgastrectomy patient.
The following common endoscopic findings should not be considered pathologic. The usual angulation
of the esophagogastric junction to the left may be lost. Although gastric mucosa may be normal with
normal rugal folds, frequently the rugae are absent, there is mucosal pallor, and finely arborizing
mucosal vessels are readily distinguished. Distensibility is normal, although maximum insufflation is
difficult to maintain with an open stoma. On the gastric side of the stomal ring, small outpouchings,
recesses, or nodules of granulation tissue are often noted. Portions of the stomach or jejunal limbs
may be bound down or "tented" by postoperative adhesions; these alter the normal course of the
intestine, for which reason the endoscope should be advanced with caution and gentleness.
Duodenogastric Reflux
Terminology with respect to postoperative gastritis due to duodenogastric reflux is imprecise and
confusing. Although postoperative gastritis is often referred to as bile gastritis, it is by no means certain
that bile is the cause, or at least the exclusive causative factor. The term alkaline reflux gastritis is also
used. Such terminology must not be mistakenly accepted as indicative of the pathophysiology of this
poorly understood condition. For convenience, the term bile reflux is used here.
Duodenogastric reflux, manifested as bilious vomiting often with concurrent epigastric pain, may occur
after all types of ulcer surgery.143151(5115) The pain of bile reflux differs from that of recurrent ulcer
in that it is relieved by bilious vomiting. Bile reflux is normal in patients with incompetent or absent
pyloric sphincters but is not clinically significant unless it produces symptoms.152,153(5116)
Endoscopy is useful for determining the cause of bilious vomiting and also the mucosal effects of
duodenogastric reflux. In nearly all patients, bile reflux causes the gastric mucosa to appear
erythematous and friable (Figure 5228),152155(5117) but in some patientsperhaps those with a
greater degree of refluxperistomal ulcers of esophageal inflammation occur. These patients require
aggressive medical therapy; if this fails, a second operation (Roux-en-Y gastrojejunostomy) may be
necessary to divert the normal duodenal juices from the stomach. Endoscopic evaluation of the
effectiveness of medical therapy is mandatory in assessing healing and the need for reoperation.
(5118)Figure 5228. Gastric erythema. Note the intense erythema in the gastric pouch after
gastrectomy with a Billroth I anastomosis. Erythema is most intense around the stoma.
Minimum gastritis was present histologically.
Gastric Mucosal Changes
Important alterations of the gastric mucosa in the postgastrectomy patient include erythema, gastritis,
intestinal metaplasia, dysplasia, and cancer.1,127,132,156160(5119) Dysplasia and cancer are
discussed in a subsequent section.
Mucosal erythema develops in nearly all patients who undergo ulcer surgery; this cannot be quantified
or correlated with clinical factors.159163(5120) Although erythema is usually associated with bile
staining (bile gastritis), gastritis cannot always be proved histologically. Erythema is usually most
intense around the stoma; the mucosa can be extremely friable (see Figure 5228). Erythema is
abolished by biliary diversion surgery.146,148(5121)
Histologic evidence of gastritis is found in 60 to 100% of patients who have undergone gastric
resection for peptic ulcer disease (Figure 5229).1,128,130,150,154(5122) There is no evidence that
gastritis can be accurately quantified or implicated in specific clinical symptoms.1,128,130(5123) In our
experience, 75 to 80% of patients with Billroth II reconstruction after gastric resection will have
histologically verifiable gastritis for prolonged periods that may extend to many years after surgery,
regardless of clinical symptoms. In animal studies, gastritis has been induced by duodenogastric
reflux,164(5124) but a similar relationship has not been demonstrated in humans. Long-term sequelae
of gastritis include mucosal atrophy and occasionally deficiency of intrinsic factor and hence a lack of
vitamin B12.165,166(5125) Such gastritis is rarely the source of significant bleeding.
(5126)Figure 5229. Histologically proven severe gastritis is seen diffusely throughout the
gastric pouch. Mucosa is friable and bile-stained.
Intestinal metaplasia refers to the development of intestinal epithelium in gastric
mucosa.167,168(5127) The degree of intestinalization may be mild, with few intestinal elements, or
severe, with formation of villi. Any cell type found in the intestine can also be found in metaplastic areas
within the stomach. It is not known whether this lesion is heterotopic, that is, the result of misplacement
of intestinal epithelium, or truly metaplastic, that is, a transformation of gastric into intestinal
epithelium.168(5128)
Some investigators believe that intestinal metaplasia is a premalignant lesion because of the frequent
endoscopic finding of intestinal epithelium in association with carcinoma.167169(5129) No long-term
study in humans has prospectively followed metaplastic lesions to determine whether they degenerate
into cancer. Animal experiments studying chemically induced intestinal metaplasia suggest that
malignant degeneration does occur, but the extent to which these findings may be correlated with
postgastrectomy metaplasia in humans is uncertain.163(5130)
The endoscopic appearance of intestinal metaplasia is distinctive (Figure 5230). There is usually a
superficial, flat, whitish, feathery edge of tissue beginning at or near the stoma that extends proximally
into the gastric pouch. Occasionally, lesions occur as white spots or islands in the peristomal area. Our
policy for patients with intestinal metaplasia is annual or biannual endoscopic examinations with
repeated biopsies.
(5131)Figure 5230. Intestinal metaplasia. Feathery white border of the metaplastic intestinal
epithelium can be seen on the gastric side of a Billroth II anastomosis. Intestinal metaplasia
may also appear as white patches or larger white areas and should not be mistaken for cancer.
Note the eventration of the efferent limb through the stoma.
Recurrent Ulcer
Recurrence rates after peptic ulcer surgery are as follows: vagotomy and pyloroplasty, 5 to 8%;
vagotomy and antrectomy, 2 to 4%; highly selective vagotomy, 10 to 20%; and subtotal gastrectomy
(without vagotomy) 20%.129,130,170176(5132) The presenting symptoms in 75 to 90% of patients
with marginal ulcers include pain, and in over 50%, concurrent bleeding. Of all patients who present to
the endoscopist with typical dyspeptic pain, 58% will have recurrent ulcers.1(5133) Unlike primary ulcer
disease, which is characterized by intermittent and cyclical symptoms, recurrent ulcers after surgery
tend to cause continuous symptoms that do not remit until effective therapy is instituted. Symptoms
often persist despite medical therapy, and a second operation becomes
necessary170176(5134)thus the need for timely reevaluation of patients and accurate diagnosis of
recurrent ulcer. Because radiographic examination of the postgastrectomy stomach is highly
inaccurate,123(5135) endoscopy is the mandatory procedure of choice when ulcer recurrence is a
consideration.
Peptic ulcers almost always occur on the stomal ring or the jejunal limb adjacent to the stoma; they
rarely occur in gastric mucosa,175(5136) although this point is disputed by one study.176(5137)
Several biopsies of gastric ulcers, especially in the absence of mechanical obstruction or bezoar, must
be obtained to eliminate the possibility of carcinoma. Ulcers on the anastomotic ring are flat and linear
with an overlying pseudomembrane (Figure 5231). Stomal ulcers in jejunal mucosa are
characteristically round and shallow (Figure 5232). Stomal ulcers may be circumferential and may
bleed easily on contact by the endoscope (Figure 5233). Marked edema and consequent obstruction
are unusual but do occur (Figure 5234).
(5138)Figure 5231. Recurrent ulcer. Ulcer at the stomal ring is flat and superficial and has
an adherent white pseudomembrane. Minimum gastritis was noted.
(5139)Figure 5232. Recurrent ulcer. Extreme close-up view of an ulcer found in the jejunal
mucosa just beyond the stoma in a patient who had a gastrectomy and a Billroth II
anastomosis. Typically, the ulcer is small and shallow.
(5140)Figure 5233. Recurrent ulcer surrounding the stoma of a Billroth II anastomosis.

Ulcer was rather shallow but bled easily. Edema was not severe enough to cause outlet
obstruction in this patient. Note the visible suture at the inferior edge (6 o'clock position) of
the stoma.
(5141)Figure 5234. Ulceration and outlet obstruction. Deep ulceration and intense
inflammation are present in a patient with a Billroth II anastomosis who presented with outlet
obstruction.
A recurrent ulcer requires careful investigation to determine its cause (Figure 5235), the most
common being incomplete vagotomy.170(5142) The need for thorough endoscopy in a stomach that is
clear of retained food or bezoar cannot be overstated; this includes careful inspection of remaining
gastric mucosa for the cobblestoned ap-pearance that suggests continued function or hypertrophy of
acid-producing epithelium; estimation of stomal diameter; and investigation of the closed afferent limb
for retained antrum.
(5143)Figure 5235. Conditions causing recurrent ulcers. Ulceration may also be due to
development of a stump cancer. Esophageal ulceration may be secondary to alkaline reflux
(not illustrated).
Mechanical Obstruction
Gastric outlet obstruction that occurs late in the postoperative course is usually due to anastomotic
stricture or recurrent ulceration with edema. Endoscopy is the examination of choice, with particular
attention to the stomal area. The presence of a bezoar (Figure 5236) or retained food may require
lavage or a period of clear liquid intake before endoscopy will be successful. Occasionally, mechanical
obstruction may be associated with other changes in the gastric pouch (Figure 5237).
(5144)Figure 5236. Moderate-size gastric pouch with erythema and a phytobezoar. Patient
had a narrowed anastomosis that required surgical revision.
(5145)Figure 5237. Bizarre-appearing gastric mucosa in an elderly patient with abdominal

pain and anorexia who had a strictured Billroth I anastomosis. Biopsies revealed gastritis and
Candida infection. The endoscopic appearance reverted to normal with balloon dilation and
antifungal therapy.
Bleeding
The approach to management and therapy for hemorrhage in the postgastrectomy patient is similar to
that for bleeding in the nonsurgical patient.177(5146) This is considered in detail in Chapter 26: Acute
Upper Gastrointestinal Bleeding. If hemorrhage occurs, emergency endoscopy is indicated, during
which a meticulous search is carried out for the following: esophageal varices, esophagitis,
Mallory-Weiss tear, gastritis, intraluminal sutures, and ulcers. Particular attention must be directed
toward the jejunal side of the anastomosis, where stomal ulcers are most likely to occur. Gentle
retroflexion of the endoscope within the stoma may reveal blind areas not readily seen when
approaching the anastomosis in a forward-viewing direction (see Figure 5224).
Endoscopy may be performed in postgastrectomy patients with subclinical bleeding or anemia to
search for recurrent ulcer, severe gastritis, or the rare gastric carcinoma. Anemia may develop in
postgastrectomy patients from loss of blood, malabsorption of iron, or intrinsic factor deficiency.
Recurrent ulcer is suspect in the anemic patient with dyspepsia-type pain. Even in the absence of this
type of pain, recurrent ulceration, gastritis, or other causes of blood loss, including colonic lesions,
must be eliminated before assuming anemia is due to iron or intrinsic factor deficiency.
Malnutrition and Weight Loss
Weight loss and malnutrition are sometimes indications for endoscopy in the postgastrectomy patient,
although the examination is usually not helpful except to rule out bleeding lesions, as discussed
previously. The incidence of gastric remnant cancer is small, but it is necessary to exclude this lesion
from consideration when there has been significant loss of weight.

Cancer
Balfour178(5147) reported in 1922 the case of a patient who developed adenocarcinoma in the gastric
remnant after surgery for a nonneoplastic ulcer. Subsequently, many investigators reported an
increased incidence of so-called gastric stump cancer, compared with nonoperated
controls.179202(5148) The increase in cancer incidence becomes evident approximately 2 decades
after operation. There is some evidence that postgastrectomy patients have a higher incidence of
nongastric cancers (especially lung cancer) 1 to 2 decades after operation.181,182(5149) Some have
advocated routine endoscopic screening with biopsy for all patients beginning 10 to 15 years after
operation.179201,203(5150)
The range of histologic changes in the gastric mucosa of postgastrectomy patients includes chronic
atrophic gastritis, intestinal metaplasia, gastric gland dilation (cystic dilation or cystification), papillary
and glandular hyperplasia, and adenomatous transformation.159163(5151) The severity of these
histologic changes may be related to the amount of duodenogastric reflux; animal studies suggest that
the frequency of gastric remnant cancer correlated quantitatively with bile reflux.164(5152) In an
endoscopic study of 87 patients who had undergone surgery for peptic ulcer disease more than 20
years earlier, the Billroth II operation correlated much more strongly with the occurrence of dysplasia
than other operations.204(5153) Dysplasia was most likely to be present around the stoma.
Lundegardh et al.,205(5154) using data from a national cancer registry in Sweden, found no increase
overall in the risk for gastric cancer among 6459 patients followed for 25 to 33 years after partial
gastrectomy for benign peptic ulcer. However, differences were evident for certain subgroups of
patients. Beginning at 19 years' follow-up, the risk of gastric cancer increased progressively with each
interval of follow-up. Prior to this point, the risk actually decreased in patients who had undergone
partial gastrectomy. Furthermore, there was an inverse association between age at operation and risk
of stomach cancer, independent of duration of follow-up; an increased risk was found for patients less
than 50 years of age at operation. The risk among women who underwent operation was higher than
that for men. At the time of the study, the incidence of gastric cancer among men in the general
population was approximately twice that for women. The risk for patients who had undergone a Billroth
II was four times higher than for those who had a Billroth I. The risk was also higher for patients who
had surgery for gastric as opposed to duodenal ulcer. Lundegardh et al.205(5155) concede certain
potential sources of bias in this study. During certain time periods, there was greater interest in the
possibility of an association between gastric surgery and increased cancer risk. Although the incidence
of gastric cancer varies with socioeconomic status, it was not possible to adjust the data for this factor.
Although the results of retrospective studies suggest an increased risk for gastric cancer among
postgastrectomy patients, the results of other retrospective studies191194(5156) and several
population-based studies195197(5157) do not support this conclusion.
Stael von Holstein et al.206(5158) performed endoscopic surveillance on patients who had undergone
surgery for peptic ulcer at a single institution in Sweden between 1930 and 1960. In this prospective
study, 682 of 838 living patients were offered screening examinations including endoscopic biopsies
beginning in January 1973. For geographic reasons, surveillance was not offered to some subjects.
There were 354 patients who entered the study. Those who did not accept screening constituted a
control group; data on these patients were obtained through a national cancer registry. Depending on
the endoscopic and histologic findings at the initial procedure, endoscopy was performed at 1- to
3-year intervals. During the study, 17 cases of early-stage gastric cancer were discovered in patients
undergoing endoscopic surveillance compared to 2 in the control group. However, at 17 years'
follow-up, 14 patients in the control group versus 12 patients in the screening group had died of gastric
carcinoma. Because the mortality from gastric cancer was not reduced, Stael von Holstein et
al.206(5159) concluded that regular endoscopic surveillance is not indicated in patients who have
previously undergone gastric surgery for benign disease.

The results from a subsequent study by Stael von Holstein et al.,207(5160) however, suggest that
endoscopic surveillance with biopsies should be performed at short intervals if dysplasia is discovered
at endoscopy, especially if graded as severe. Stump carcinoma was diagnosed in 3 of 22 patients who
were discovered to have moderate dysplasia at an initial screening endoscopy; cancer was diagnosed
in these patients between 2 and 6 years after entering a study of endoscopic surveillance. The
remaining patients in this group were followed for up to 18 years, and mild to moderate dysplasia
persisted except for 2 patients who had severe dysplasia on one occasion. Subsequently, both had
only mild or moderate dysplasia. Severe dysplasia, not present at the initial screening examination,
was discovered at later procedures in another group of 17 patients. Carcinoma developed in the gastric
remnant in 7 (41%) of these patients at a median of 2 years (range 1 to 11 years) after discovery of the
presence of severe dysplasia. Two other patients underwent surgery for suspected carcinoma but only
dysplasia was discovered. Three patients died of unrelated causes; no evidence of carcinoma was
found in the remaining 5 patients who were followed from 6 and 18 years.
Based on a consensus of experts, The American Society for Gastrointestinal Endoscopy does not
recommend routine endoscopic screening of asymptomatic postgastrectomy patients.208(5161)
Decision analysis methods suggest that in the United States, where the incidence of gastric cancer is
low, minimum if any benefit would result from routine screening.198(5162) Based on currently available
data, endoscopy with multiple biopsies of abnormal-appearing mucosa199201(5163) and recurrent
gastric ulcers is indicated in postgastrectomy patients only if new or recurrent symptoms develop. If
gastric ulcers are found, biopsy for evidence of carcinoma is indicated (Figure 5238).
(5164)Figure 5238. Malignant ulcer. Deep, irregular gastric ulcer in a patient who had
undergone gastrectomy with a Billroth I anastomosis. Biopsies revealed adenocarcinoma.
Other Problems
Other problems that may prompt endoscopy include retained sutures, postgastrectomy dumping
syndrome, afferent loop syndrome, symptoms suggesting esophageal reflux, and dysphagia.
Nonabsorbable suture material or staples may appear on the luminal surface of the gastric mucosa.
These can cause pain, ulceration, and bleeding.209(5165) At endoscopy, sutures are usually found
within ulcers in the region of the anastomosis or at the gastrotomy closure (Figure 5239). Symptoms
may be relieved and bleeding may cease after removal of suture material. When sutures are noted
during endoscopy, a gentle attempt at removal should be made even in the absence of ulceration.
(5166)Figure 5239. Suture ulcer. Suture material is visible within the ulcer bed. Endoscopic
removal of the suture resulted in healing of the ulcer.
The dumping syndrome is usually a transient problem in the postgastrectomy patient and is alleviated
by appropriate dietary manipulation. Endoscopic examination rarely is helpful in defining the problem or
directing therapy.
The afferent loop syndrome, now considered extremely rare, results from a narrowed afferent loop
outlet.129(5167) Postprandial accumulation of pancreatic and biliary secretions within the limb causes
pain and nausea. These symptoms are relieved when the secretions suddenly enter the stomach and
induce bilious vomiting (Figure 5240). Experimental studies using balloon occlusion of the afferent
loop or radionucleotide imaging of biliary secretions have failed to demonstrate an association between
symptoms and a dysfunctional afferent loop.210(5168) Endoscopy assists in evaluating the patency of
afferent and efferent limbs and in defining other more probable causes of postgastrectomy pain such
as recurrent ulcer.
(5169)Figure 5240. Computed tomography of a dilated afferent loop in a patient who

underwent antrectomy and Billroth II reconstruction 6 years prior to presentation. The
massively dilated loop reached into the pelvis at laparotomy. (From Allen JI, Allen MO.
Endoscopy as an alternative and aid to re-operation. In McQuarrie DG, Humphrey EW [eds].
Reoperative General Surgery. St. Louis, Mosby-Year Book, 1992; 10343.)
Symptoms referable to the esophagus are common after ulcer surgery. Usually, symptoms of reflux
and dysphagia are transient and related to physiologic adaptation to vagotomy. If dysphagia continues
beyond the immediate postoperative period, endoscopy should be performed to rule out severe reflux
esophagitis or stricture, both of which are managed by conventional means.
Endoscopic Retrograde Cholangiopancreatography in the Postgastrectomy Patient
Patients who have undergone ulcer surgery often present unique technical problems to the
endoscopist attempting endoscopic retrograde cholangiopancreatography (ERCP) with or without
sphincterotomy. Although esophageal surgery, pyloroplasty, or gastrectomy with Billroth I anastomosis
does not ordinarily add to the technical difficulty in cannulation of the ampulla, ERCP after gastrectomy
and Billroth II anastomosis can be extremely difficult. In particular, it is often difficult to locate the
afferent limb with a side-viewing endoscope. For this reason, an end-viewing endoscope may be used
to establish the position of the afferent limb and the ampulla of Vater and to attempt cannulation. The
length of the afferent loop is rarely the reason for failure of ERCP. Because the courses from the
stomach of the afferent and efferent limbs of a loop jejunostomy vary (see Figure 5220), the position
of the endoscope should be confirmed fluoroscopically before concluding that the length of the afferent
limb makes ERCP impossible. Technique of ERCP and endoscopic papillotomy in patients who have
undergone gastric surgery is discussed in Chapter 61: Special Papillotomy Techniques.
Cancer Operations
The extent of resection for gastric carcinoma depends on tumor size and location. Proximal gastric
carcinomas are usually treated by esophagogastrectomy with anastomosis of the remaining
esophagus to stomach or with colonic interposition if esophageal involvement is extensive (see Figures
527 and 5210).211,212(5170) Carcinoma of the midstomach is treated by total or near-total
gastrectomy (see Figure 527), whereas distal cancer is treated by distal gastrectomy with
gastrojejunostomy (see Figure 5220).211214(5171) Billroth II (gastrojejunostomy) anastomosis is
most often used rather than Billroth I (gastroduodenostomy) to avoid obstruction of the anastomosis
after resection if the tumor should recur in the duodenum.
Postoperative problems usually relate to recurrence of carcinoma at the site of anastomosis. The
presenting features may be upper gastrointestinal bleeding, abdominal pain, or obstruction.
Endoscopic examination of the remaining gastric pouch and stoma should be performed to look for
evidence of recurrent carcinoma. In symptomatic patients who present for endoscopy, we routinely
obtain biopsies in the area of the stoma in patients with gastric cancer even when there is no visible
evidence of recurrent tumor. Obstructing carcinoma at the anastomosis occasionally can be treated by
hydrostatic balloon dilation, although the risk of bleeding or perforation is significant.
Weight Reduction (Bariatric) Surgery

The complications and decreased life expectancy of patients with morbid obesity may warrant
operative intervention. Although the jejunoileal bypass is no longer done because of severe
postoperative metabolic problems,215,216(5172) gastric bypass or gastroplasty is currently used for
weight reduction.217219(5173) These procedures reduce the size of the gastric pouch and gastric
outlet, thereby limiting caloric intake. From an endoscopic standpoint, it is important to understand the
difference between a gastroplasty and a gastric bypass (Figure 5241). A gastroplasty operation
potentially would allow endoscopic examination of the distal stomach, whereas examination of the
distal stomach is not possible after a gastric bypass operation.
(5174)Figure 5241. Gastric bypass operations. A, An 85% gastric bypass with a horizontal
double row of staples and a Roux-en-Y gastrojejunostomy. B, Gastric bypass with loop
gastrojejunostomy. C, An 85% bypass with a diagonal row of staples and a Roux-en-Y
gastrojejunostomy (in a direct line with the distal esophagus). D, Vertical banded gastroplasty.
Four commonly performed bariatric procedures are illustrated in Figure 5241. Three of the
procedures illustrated involve creation of a small proximal gastric pouch and bypass of the distal
stomach. The proximal pouch is then drained by a variety of techniques (see Figure 5241A-C).
Endoscopic evaluation after gastric bypass should include estimations of pouch size and stomal
diameter in addition to examination for mucosal abnormalities such as gastritis or ulcer. Examination of
the bypassed segment is difficult but may be accomplished using a pediatric colonoscope in patients in
whom drainage is carried out via a loop gastrojejunostomy. The vertical-banded gastroplasty, or its
sister operation, the "ring" gastroplasty, does not exclude the distal stomach. Instead, a permanently
narrowed stoma is created, allowing slow passage of food from the proximal pouch to the distal
stomach (see Figure 5241D).220(5175)
Postoperative problems that require both radiographic and endoscopic investigations include failure of
weight reduction, stomal obstruction, reflux esophagitis, and upper gastrointestinal
bleeding.220223(5176) Failure to lose weight may be due to staple line disruption, gastric pouch
dilation, stomal dilation, and dietary indiscretion.219228(5177) A large gastric pouch or a
large-diameter anastomosis requires reoperation with replacement of the staple line or reduction in
pouch size. Stomal obstruction results in frequent emesis and regurgitation (Figure 5242) and is
sometimes amenable to hydrostatic balloon dilation.223,225,226,229(5178) The incidence of stomal
ulceration, which produces pain and bleeding, is less than 2%.227(5179) Stomal ulceration in
association with a small gastric pouch usually responds to pharmacologic therapy; ulceration
associated with a large pouch may require pouch revision and vagotomy. Symptomatic bleeding from
the bypassed stomach and duodenum rarely oc-curs (0.3%).228(5180) Marlex mesh, used to reinforce
the stoma of a horizontal-banded gastroplasty, may erode into the stomach.230,231(5181) Fardy et
al.231(5182) described endoscopic findings in 12 patients; the mesh had eroded into the stomach at a
mean of 23 months after operation (range 5 weeks to 55 months) and was associated with dehiscence
of the staple line in 3 and stomal stenosis in 9 patients.
(5183)Figure 5242. Large gastric pouch and stenotic stoma in a patient who had undergone
gastric bypass. Three pills are seen adjacent to the pinhole-sized stoma. Surgical revision with
reduction of the pouch size was necessary.
Gastrostomy Tubes
Gastrostomy tubes are used for enteric feeding or, less frequently, for gastric decompression. Most
gastrostomy tubes are now placed using endoscopic techniques, although a few are placed during
surgical procedures (see also Chapter 55: Percutaneous Endoscopic Gastrostomy). Gastrostomy
tubes typically are placed in the dependent portion of the stomach near the fundus-antrum junction on
the greater curvature (Figure 5243). The catheter is pulled toward the skin and anchored with external
devices to approximate the stomach wall and the abdominal peritoneal surface.
(5184)Figure 5243. Percutaneous endoscopic gastrotomy tube within the stomach lumen.
Typically, these enter the stomach at the antrum-fundus junction.
Patients with gastrostomy tubes undergo endoscopy primarily for bleeding or obstruction. The
technique of endoscopy is not difficult, although the anatomy of the stomach is somewhat distorted by
the gastrostomy tube. The usual method of following rugal folds to the antrum may not be helpful
because of the tension produced by the gastrostomy tube on the stomach wall. Endoscopy can be
performed immediately after placement of the tube. Maturation of the tube tract requires about 6
weeks, after which time most tube manipulation or exchange can be accomplished without endoscopy.
Endoscopy After Choledochoenterostomy

Choledochoenterostomy is used for permanent decompression of a biliary tree obstructed by stones or
cancer.232239(5185) This involves anastomosis of the common bile duct to the posterosuperior
aspect of the duodenal bulb (Figure 5244A). Choledochojejunostomy, used when duodenal
abnormalities or cancer precludes duodenal anastomosis, may be either a loop
choledochojejunostomy or a Roux-en-Y (Figure 5244B and C).
(5186)Figure 5244. Types of choledochoenterostomies. A, Choledochoduodenostomy. B,

Loop choledochojejunostomy. C, Roux-en-Y choledochojejunostomy.
Complications of choledochoenterostomy include cholangitis, biliary obstruction, and "sump"
syndrome.232239(5187) Each is related to stomal stenosis or ob-struction, which requires endoscopy
with cholangiogra-phy for diagnosis.232239(5188) A choledochoduodenostomy is located in the
duodenum immediately beyond the pylorus on the superior aspect of the bulb. At endoscopy, which is
facilitated by use of a side-viewing instrument, this normally appears as a patent, round anastomosis
approximately 1 cm in diameter (Figure 5245). A choledochojejunostomy can be examined with a
pediatric colonoscope, although even this endoscope may not be long enough to reach the
anastomosis. Cholangiography of the proximal biliary tract is performed using a balloon-tipped catheter
to occlude the anastomosis and permit adequate filling of the biliary system with a contrast agent
(Figure 5246A). Injection of contrast agent at the ampulla will demonstrate the distal blind segment of
the common duct (Figure 5246B). If stones or debris is shown in the distal duct, as occurs in the
"sump" syndrome, endoscopic sphincterot-omy may be beneficial (Figure 5247).239(5189) Because
recurrent cholangitis, biliary obstruction, and the sump syndrome are probably related to stricture of the
choledochoduodenostomy, endoscopic dilation of a strictured anastomosis with removal of debris may
be curative.238(5190)
(5191)Figure 5245. Choledochoduodenostomy seen with a side-viewing endoscope. Stoma

is located in the duodenal bulb just beyond the pylorus. Diameter of the stoma is
approximately 1 cm. (Courtesy of Dr. J. Segal.)
(5192)Figure 5246. Endoscopic cholangiography after choledochoduodenostomy. A,

Injection of contrast material into the common bile duct via a papilla (bottom arrow)
demonstrates free flow through the choledochoduodenostomy and into the proximal biliary
radicals. A stone is visible within the common bile duct at the level of the anastomosis (top
arrow). B, Contrast material is injected into the proximal biliary radicals after inflation of an
extraction balloon catheter to occlude the choledochoduodenostomy (top arrow). Air outlines
the distal common duct and a portion of the pancreatic duct (bottom arrow).
(5193)Figure 5247. "Sump" syndrome. Position of the stoma is superoposterior on the wall
of the duodenal bulb just distal to the pyloric sphincter. In the "sump" syndrome," stones or
debris accumulate in the distal blind end of the common bile duct.
Endoscopy After Pancreatic Operation

Chronic Pancreatitis
Surgery for chronic pancreatitis involves resection of portions of the pancreas and pancreatic duct
drainage.240243(5194) The most commonly performed procedures are the
Puestow-Gillesby240(5195) and the DuVal (Figure 5248).240242(5196) Symptomatic relief after
surgical therapy can be expected in 75% of patients if rigid indications for operation are imposed.
(5197)Figure 5248. Pancreatic drainage procedures. A, Longitudinal pancreaticojejunostomy

with Roux-en-Y reconstruction (Puestow-Gillesby procedure). B, End-to-end
pancreaticojejunostomy (DuVal procedure).
The Puestow-Gillesby procedure is used to treat patients with long, dilated segments of the main
pancreatic duct and involves incision of the entire length of the duct with drainage by means of a
side-to-side anastomosis of a jejunal limb to the gland (see Figure 5248A). Chronic pancreatitis
involving the body or tail of the gland can be treated by resection of the tail and end-to-end
anastomosis of the pancreatic duct to a Roux-en-Y limb (DuVal procedure) (see Figure 5248B).
Preoperative ERCP is mandatory when considering surgical therapy for chronic pancreatitis. This
outlines the specific anatomic features and therefore aids in planning the surgical
approach.240242(5198) Endoscopy should be considered postoperatively when symptoms recur after
a pain-free period. Postoperative retrograde pancreatography will define ductal anatomy and establish
the patency of an anastomosis, or the lack thereof, although anastomotic patency does not always
correlate with the presence or absence of symptoms (Figure 5249).
(5199)Figure 5249. Endoscopic retrograde cholangiopancreatography in a patient who has

undergone a DuVal procedure. Pancreatography after a DuVal end-to-end
pancreaticojejunostomy shows a patent proximal pancreatic duct with free flow of contrast
material into the jejunal limb (arrow).
Pancreatic Pseudocysts
Surgical procedures for internal drainage of pseudocysts include cystgastrostomy (Figure 5250) and
cystjejunostomy.134,240245(5200) For cystgastrostomy, an incision is made via an anterior
gastrotomy through the posterior wall of the stomach into the pseudocyst cavity located directly
posterior and adherent to the stomach wall.246(5201) The cystgastrostomy is oversewn
circumferentially to ensure hemostasis of the highly vascular gastric wall. After this type of surgery,
endoscopy will demonstrate the anterior gastrotomy suture line (Figure 5251) and the entrance to the
cyst cavity, which is usually on the posterior gastric wall. Pseudocysts rapidly collapse after drainage.
Thus, endoscopy in the later postoperative period will fail to demonstrate the cyst anastomosis (Figure
5252). Endoscopy is indicated for postoperative upper gastrointestinal bleeding. Early-onset bleeding
after operation may originate from the margin of the cystgastrostomy, the gastrotomy suture line, the
interior of the cyst, or from sources unrelated to the procedure. Sometimes angiography is required to
localize bleeding sites.
(5202)Figure 5250. Pancreatic pseudocystgastrostomy. Illustration of the operative

appearance of a cystgastrostomy performed via an anterior gastrotomy.
(5203)Figure 5251. Pseudocystgastrostomy. Gastrotomy scar can be seen on the anterior

side of the stomach with the cystgastrostomy just below and to the right.
(5204)Figure 5252. Pseudocystgastrostomy. Stoma is still visible in this 7-day-old

cystgastrostomy. Usually, the cyst rapidly collapses after decompression and the stoma
disappears.
Whipple Operations
Less than 10% of patients with pancreatic cancer have tumors that are resected for cure. The surgical
procedures are distal pancreatectomy for lesions in the tail of the pancreas and
pancreatoduodenectomy (Whipple's procedure) for lesions of the pancreatic head and periampullary
region (Figure 5253).134,247(5205) The Whipple operation involves resection of the pancreatic head,
distal common bile duct, and duodenum along with adjacent node-bearing tissue.247(5206) A
gastrojejunostomy, choledochojejunostomy, and pancreatojejunostomy are performed to reconstitute
intestinal continuity and biliary and pancreatic drainage. The endoscopic appearance of the stomach is
identical to that of a gastric resection with Billroth II anastomosis performed for ulcer disease.
Endoscopy is occasionally performed for recurrent biliary obstruction or pancreatitis to determine the
patency of the biliary and pancreatic anastomoses. A pediatric colonoscope is recommended.
Considerable skill and patience are required for this type of complex endoscopic examination.
(5207)Figure 5253. Pancreaticoduodenectomy (Whipple procedure). A, Structures resected

include the antrum, entire duodenum, head of the pancreas, gallbladder, and distal common
bile duct. B, Reconstruction of sequential choledochojejunostomy, pancreaticojejunostomy
and gastrojejunostomy.
Endoscopy After Vascular Surgery

Graft-enteric fistulas occur in 1 to 4% of patients treated for abdominal aortic
aneurysms.248,249(5208) Mortality rates of 100% in unoperated patients and 50 to 75% in surgically
treated patients warrant an aggressive search for this lesion in patients with intraabdominal grafts who
present with gastrointestinal bleeding.248254(5209) Successful management depends on early,
accurate diagnosis (see also Chapter 26: Acute Upper Gastrointestinal Bleeding).248(5210)
Graft-enteric fistulas result in part from mechanical forces between the proximal suture line of the graft
and the overlying third portion of the duodenum. A second important pathogenetic component is
probably subclinical infection of the graft.249(5211) At the initial episode, upper gastrointestinal
bleeding is small to moderate in volume, and it often stops spontaneously (the "herald bleed"). Ulcers
are often present concomitantly and should not be cause for lessening the intensity of diagnostic
evaluation. Barium contrast studies, computed tomography, and angiography may fail to diagnose the
fistula, especially in the absence of active bleeding.252254(5212)
Every patient with an abdominal aortic graft and upper gastrointestinal bleeding requires emergency
endoscopy to the ligament of Treitz (Figure 5254).252254(5213) We recommend a 135-cm-length
pediatric colonoscope for this purpose. Because bleeding from a graft fistula always recurs and is
potentially lethal, rapid diagnosis is essential. At endoscopy, one or more of the following may be
found: absence of any bleeding source, extrinsic compression of the third portion of the duodenum,
bleeding in the distal duodenum, superficial erosions of the distal duodenum, and graft material
protruding into the lumen of the duodenum (Figure 5255). Extreme care must be taken not to
dislodge clot or impacted graft material; otherwise massive hemorrhage may occur.
(5214)Figure 5254. Aortograft duodenal fistula. Fistula from the proximal aortic graft suture
line into the fourth portion of the duodenum near the ligament of Treitz. Pediatric colonoscope
is needed to provide sufficient length for endoscopic examination.
(5215)Figure 5255. A and B, Aortoenteric graft fistula. Dacron graft is visible in the lumen
of the fourth part of the duodenum viewed with a pediatric colonoscope. (A and B, From the
Endoscopic findings suggesting a graft-enteric fistula should prompt emergency operative exploration
for confirmation and treatment. Some authors advocate endoscopy in the operating suite just prior to
anesthesia induction to minimize delay in definitive management and avoid massive
bleeding.250254(5216) Our policy is to consider each patient individually, but always to err on the side
of aggressive surgical intervention should a question concerning therapy arise. Every patient with
upper gastrointestinal bleeding should be questioned specifically about prior intraabdominal vascular
surgery.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
Clark JL, Allen JI, Allen MO. Endoscopy as an alternative and aid to reoperation. In McQuarrie
DG, Humphrey EW (eds). Reoperative General Surgery. St. Louis, Mosby-Year Book,
1992;10343.
Hirschowitz BI, Luketic GC. Endoscopy in the post-gastrectomy patient. An analysis of 580
patients. Gastrointest Endosc 1971;18:2730.
Everhart JE. Overview. In Everhart JE (ed). Digestive Diseases in the United States:
Epidemiology and Impact. National Institutes of Health Publication No. 941447. Bethesda,
MD, National Institutes of Health, 1994;154.
White JV. Laparascopic cholecystectomy: The evolution of general surgery. Ann Intern Med
1991;115:6513.
Graber JN, Schultz LS, Pietrafitta JJ, et al. Complications of laparascopic cholecystectomy: A
prospective review of an initial 100 consecutive cases. Lasers Surg Med 1992;12:927.
Olsen DO. Laparoscopic cholecystectomy. Am J Surg 1991;161:33944.
Perissat J, Vitale GC. Laparoscopic cholecystectomy: Gateway to the future. Am J Surg
1991;161:408.
Pietrafitta JJ, Schultz LS, Graber JN, et al. Laser laparoscopic vagotomy and pyloromyotomy.
Richter JE. Surgery for reflux diseaseReflections of a gastroenterologist. N Engl J Med
1992;326:8258.
10. Richter JE, Castell DO. Gastroesophageal reflux. Pathogenesis, diagnosis, and therapy. Ann
Intern Med 1982;97:93103.
11. Spechler SJ. Comparison of medical and surgical therapy for complicated gastroesophageal
reflux disease in veterans. The Department of Veterans Affairs Gastroesophageal Reflux
Disease Study Group. N Engl J Med 1992;326:78692.
12.
DeMeester TR, Stein HJ. Gastroesophageal reflux disease. In Moody FG, Carey LC, Jones
RS, et al. (eds). Surgical Treatment of Digestive Disease. Chicago, Year Book Medical,
1986;65108.
13.
Nissen R. Gastropexy and "fundoplication" in surgical treatment of hiatal hernia. Am J Dig Dis
1961;6:95461.
14. Henderson RD, Marryatt GV. Total fundoplication gastroplasty (Nissen gastroplasty):
Five-year review. Ann Thorac Surg 1985;39:749.
15. Negre JB, Markkula HT, Keyrilainen O, et al. Nissen fundoplication: Results at 10 year
follow-up. Am J Surg 1983;146:6358.
16.
DeMeester TR, Bonavina L, Albertucci M. Nissen fundoplication for gastroesophageal reflux
disease: Evaluation of primary repair in 100 consecutive patients. Ann Surg 1986;240:920.
17.
Baue AE, Belsey RH. The treatment of sliding hiatus hernia and reflux esophagitis by the
Mark IV technique. Surgery 1967;62:396406.
18. DeMeester TR, Johnson LF, Kent AH. Evaluation of current operations for the prevention of
gastroesophageal reflux. Ann Surg 1974;180:51125.
19. Hill LD. An effective operation for hiatal hernia: An eight year appraisal. Ann Surg
1967;166:68192.
20. Woodward ER. Surgical treatment of reflux peptic esophagitis. Am Surg 1982;48:64751.
21.
Collis JL, Kelly TD, Wiley AM. Anatomy of the crura of the diaphragm and the surgery of
hiatus hernia. Thorax 1954;9:17589.
22. Henderson RD, Marryatt G. Total fundoplication gastroplasty: Long-term follow-up in 500
patients. J Thorac Cardiovasc Surg 1983;85:817.
23. Angelchik J-P, Cohen R. A new surgical procedure for the treatment of gastroesophageal
reflux and hiatal hernia. Surg Gynecol Obstet 1979;148:2468.
24. Starling JR, Reichelderfer MO, Pellett JR, et al. Treatment of symptomatic gastroesophageal
reflux using the Angelchik prosthesis. Ann Surg 1982;195:68691.
25. Landreneau RJ, Marshall JB, Johnson JA, et al. A new balanced operation for complex
gastroesophageal reflux disease. Ann Thorac Surg 1991;52:3257.
26. Gott JP, Polk HC. Repeat operation for failure of anti-reflux procedures. Surg Clin North Am
1991;71:1331.
27. Lipshutz WH, Eckert RJ, Gaskins RD, et al. Normal lower-esophageal reflux. N Engl J Med
1974;291:110710.
28. Higgs RH, Castell DO, Farrell RL. Evaluation of the effect of fundoplication on the
incompetent lower esophageal sphincter. Surg Gynecol Obstet 1975;141:5715.
29. Bowes KL, Sarna SK. Effect of fundoplication of the lower esophageal sphincter. Can J Surg
1975;18:32833.
30. Hallgrimsson JG, Linnet H, Jonasson H, et al. Postoperative evaluation after correction of
oesophageal hiatus hernia. Scand J Thorac Cardiovasc Surg 1976;10:25761.
31. Noble HGS, Christie DL, Cahill JL. Follow-up studies on patients undergoing Nissen
fundoplication utilizing intraoperative manometry. J Pediatr Surg 1982;17:4903.
32. Lackey C, Potts J. Penetration into the stomach: A complication of the anti-reflux prosthesis.
JAMA 1982;248:350.
33. Condon RE. More misadventures with the esophageal collar. Surgery 1983;93:4778.
34. Lilly MP, Slafsky F, Thompson WR. Intraluminal erosion and migration of the Angelchik
anti-reflux prosthesis. Arch Surg 1984;119:84953.
35. Stelling T, Ferwerda J, Mulder CJ, et al. Complications of the Angelchik antireflux prosthesis:
Endoscopic management. Gastrointest Endosc 1987;33:4536.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
Thompson WM. Diagnostic imaging and interventional techniques in re-operations in general

surgery. In McQuarrie DG, Humphrey EW (eds). Reoperative General Surgery. St. Louis,
Mosby-Year Book, 1992;21111.
Shirazi SS, Schulze K, Soper RT. Long-term follow-up for treatment of complicated chronic
reflux esophagitis. Arch Surg 1987;122:54852.
Negre JB. Post-fundoplication symptoms: Do they resist the success of Nissen
fundoplication?. Ann Surg 1983;198:698700.
Behar J, Sheahan DG, Biancani P, et al. Medical and surgical management of reflux
esophagitis. A 38-month report of a prospective clinical trial. N Engl J Med 1975;293:2638.
Leonardi HK, Ellis FH Jr. Complications of the Nissen fundoplication. Surg Clin North Am
1983;63:115565.
Boesby S, Sorenson HR, Madsen T, et al. Failures after surgical treatment of patients with
hiatus hernia and reflux symptoms: A pathophysiological study. Scand J Gastroenterol
1982;17:21924.
Fuchs KH, DeMeester TR, Albertucci M. Specificity and sensitivity of objective diagnosis of
gastroesophageal reflux disease. Surgery 1987;102:57580.
Thoeni RF, Moss AA. The radiographic appearance of complications following Nissen
fundoplication. Radiology 1979;13:1721.
Feussner H, Petri A, Walker S, et al. The modified AFP score: An attempt to make the results
of anti-reflux surgery comparable. Br J Surg 1991;78:9426.
Hill LD, Ilves R, Stevenson JK, et al. Reoperation for disruption and recurrence after Nissen
fundoplication. Arch Surg 1979;114:5428.
Leonardi HK, Crozier RE, Ellis FH Jr. Reoperation for complications of the Nissen
fundoplication. J Thorac Cardiovasc Surg 1981;81:506.
Stirling MC, Orringer MB. Surgical treatment after the failed anti-reflux operation. J Thorac
Cardiovasc Surg 1986;92:66772.
Henderson RD. Surgical management of the failed gastroplasty. J Thorac Cardiovasc Surg
1986;91:4652.
O'Hanrahan T, Marples M, Bancewicz J. Recurrent reflux and wrap disruption after Nissen
fundoplication: Detection, incidence and timing. Br J Surg 1990;77:5457.
Keenan DJM, Hamilton JRL, Gibbons J, et al. Surgery for benign esophageal stricture. J
Thorac Cardiovasc Surg 1984;88:18299.
Rees JR. The surgical treatment of complicated peptic esophagitis. Am Surg
1987;53:497500.
Hatton PD, Selinkoff PM, Hartford FJ Jr. Surgical management of the failed Nissen
fundoplication. Am J Surg 1984;148:7603.
Mattox HE, Albertson DA, Castell DO, et al. Dysphagia following fundoplication: "Slipped"
fundoplication versus achalasia complicated by fundoplication. Am J Gastroenterol
1990;85:146872.
Luostarinen M, Isolauri J, Laitinen J, et al. Fate of Nissen fundoplication after 20 years. A
clinical, endoscopical, and functional analysis. Gut 1993;34:101520.
Luostarinen M. Nissen fundoplication for reflux esophagitis. Long-term clinical and endoscopic
results in 109 of 127 consecutive patients. Ann Surg 1993;217:32937.
Kuhnke M, Folse R. Herniation of the gastric wall through a Nissen fundoplication. Arch Surg
1987;122:10857.
Hocking MP, Maher JW, Woodward ER. Definitive surgical therapy for incapacitating
"gas-bloat" syndrome. Am Surg 1982;48:1313.
Herrington JL Jr, Meacham PW, Hunter RM. Gastric ulceration after fundic wrapping: Vagal
nerve entrapment, a possible causative factor. Ann Surg 1982;195:57481.
Festen C. Paraesophageal hernia: A major complication of Nissen's fundoplication. J Pediatr
Surg 1981;16:4969.
Spechler SJ, Goyal RK. Barrett's esophagus. N Engl J Med 1986;315:36271.
Berardi RS, Devaiah KA. Barrett's esophagus. Surg Gynecol Obstet 1983;156:52138.
62. Bozymski EM, Herlihy KJ, Orlando RC. Barrett's esophagus. Ann Intern Med 1982;97:1037.
63. DeMeester TR, Attwood SA, Smyrk TC, et al. Surgical therapy in Barrett's esophagus. Ann
Surg 1990;212:52842.
64. Haggitt RC, Tryzelaar J, Ellis FH, Colcher H. Adenocarcinoma complicating columnar
epithelium-lined (Barrett's) esophagus. Am J Clin Pathol 1978;70:15.
65. Hamilton SR, Hutcheon DF, Ravich WJ, et al. Adenocarcinoma in Barrett's esophagus after
elimination of gastroesophageal reflux. Gastroenterology 1984;86:35660.
66. Spechler SJ, Robbins AH, Rubins HB, et al. Adenocarcinoma and Barrett's esophagus. An
overrated risk?. Gastroenterology 1984;87:92733.
67. Cameron AJ, Ott BJ, Payne WS. The incidence of adenocarcinoma in columnar-lined
(Barrett's) esophagus. N Engl J Med 1985;313:8579.
68. Van Der Veen AH, Dees J, Blankensteijn JD, Van Blankenstein M. Adenocarcinoma in
Barrett's oesophagus: An overrated risk. Gut 1989;30:148.
69. Atkinson M. Barrett's oesophagusTo screen or not to screen?. Gut 1989;30:25.
70. Lightdale CJ. Endoscopy in premalignant conditions of the esophagus. Gastrointest Endosc
1984;30:30810.
71. Schmidt HG, Riddell RH, Walther B, et al. Dysplasia in Barrett's esophagus. J Cancer Res
Clin Oncol 1985;110:14552.
72. Rogers EL, Goldkind SF, Iseri OA, et al. Adenocarcinoma of the lower esophagus: A disease
primarily of white man with Barrett's esophagus. J Clin Gastroenterol 1986;8:6138.
or early adenocarcinoma in Barrett's esophagus without grossly recognizable neoplastic
74. Haggitt RC, Reid BJ, Rabinovitch PS, et al. Barrett's esophagus: Correlation between mucin
histochemistry, flow cytometry, and histologic diagnosis for predicting increased cancer risk.
Am J Pathol 1988;131:5361.
75. Lightdale CJ, Winawer SJ. Screening diagnosis and staging of esophageal cancer. Semin
Oncol 1984;11:10112.
76. Spechler SJ. Endoscopic surveillance for patients with Barrett esophagus: Does the cancer
risk justify the practice?. Ann Intern Med 1987;106:9024.
77. Achkar E, Carey W. The cost of surveillance for adenocarcinoma complicating Barrett's
esophagus. Am J Gastroenterol 1988;83:2914.
78. Pearson FG, Cooper JD, Patterson GA, Prakash D. Peptic ulcer in acquired columnar-lined
esophagus: Results of surgical treatment. Ann Thorac Surg 1987;43:2414.
79. Brand DL, Ylvisaker JT, Gelfand M, Pope CE. Regression of columnar esophageal (Barrett's)
epithelium after anti-reflux surgery. N Engl J Med 1980;302:8448.
80. Donahue PE, Schlesinger PK, Bombeck CT, et al. Achalasia of the esophagus: Treatment
controversies and the method of choice. Ann Surg 1986;203:50511.
81.
Mercer CD, Hill LD. Reoperation after failed esophagomyotomy for achalasia. Can J Surg
1986;3:17780.
82. Murray GF. Operation for motor dysfunction of the esophagus. Ann Thorac Surg
1980;29:18491.
83. Henderson RD, Ryder DE. Reflux control following myotomy in diffuse esophageal spasm.
84.
Gallone L, Peri G, Galliera M. Proximal gastric vagotomy and anterior fundoplication as
complementary procedures to Heller's operation for achalasia. Surg Gynecol Obstet
1982;34:2306.
85.
Tytgat GNJ. Benign and malignant tumors of the esophagus. In Sivak MV Jr (ed).
Gastroenterologic Endoscopy. Philadelphia, WB Saunders, 1987;373400.
86.
Postlethwait RW, Sealy WC. Surgery of the Esophagus. New York, Appleton-Century-Crofts.
1979.
87. Cukingnan RA, Carey JS. Carcinoma of the esophagus. Ann Thorac Surg 1978;26:27486.
88. Le-Tian X, Zhen-Fu S, Ze-Jian L, et al. Surgical treatment of carcinoma of the esophagus and
cardiac portion of the stomach in 850 patients. Ann Thorac Surg 1983;35:5427.
89. Postlethwait RW. Complications and deaths after operations for esophageal carcinoma. J
Thorac Cardiovasc Surg 1983;85:82731.
90. Ellis FH Jr, Gibb SP, Watkins E Jr. Esophagogastrectomy: A safe, widely applicable, and
expeditious form of palliation for patients with carcinoma of the esophagus and cardia. Ann
Surg 1983;198:53140.
91. Molina JE, Lawton BR, Myers WO, Humphrey EW. Esophagogastrectomy for
adenocarcinoma of the cardia: Ten years' experience and current approach. Ann Surg
1982;195:14651.
92. Gunnlaugsson GH, Wychulis AR, Roland C, et al. Analysis of the records of 1,657 patients
with carcinoma of the esophagus and cardia of the stomach. Surg Gynecol Obstet
1970;130:9971005.
93. Hopkins RA, Alexander JC, Postlethwait RW. Stapled esophagogastric anastomosis. Am J
Surg 1984;147:2837.
94. Payne WS. Surgical treatment of reflux esophagitis and stricture associated with permanent
incompetence of the cardia. Mayo Clin Proc 1970;45:55362.
95. Herrington JL Jr, Mody B. Total duodenal diversion for treatment of reflux esophagitis
uncontrolled by repeated anti-reflux procedures. Ann Surg 1976;183:63644.
96.
Inberg MV, Heinonen R, Rantakokko V, et al. Surgical treatment of gastric carcinoma: A
regional study of 2,590 patients over a 27-year period. Arch Surg 1975;110:7307.
97. Shiu MH, Papacristou DN, Kosloff C, et al. Selection of operative procedure of
adenocarcinoma of the midstomach. Ann Surg 1980;192:7307.
98. Jensen RT, Gardner JD, Raufman J-P, et al. Zollinger-Ellison syndrome: Current concepts
and management. Ann Intern Med 1983;98:5975.
99. Morrow D, Passaro ER Jr. Alkaline reflux esophagitis after total gastrectomy. Am J Surg
1976;132:28791.
100. May IA, Samson PC. Esophageal reconstruction and replacements. Ann Thorac Surg
1969;7:24977.
101.
Mansour KA, Hansen HA II, Hersh T, et al. Colon interposition for advanced nonmalignant
esophageal stricture: Experience with 40 patients. Ann Thorac Surg 1991;32:58491.
102. Hankins JR, Cole FN, McLaughlin JS. Colon interposition for benign esophageal disease:
Experience with 23 patients. Ann Thorac Surg 1984;37:1926.
103. Postlethwait RW, Sealy WC, Dillon ML, et al. Colon interposition for esophageal substitution.
104. Isolauri J, Helin H, Markkula H. Colon interposition for esoph-ageal disease: Histologic finding
of colonic mucosa after a follow-up of 5 months to 15 years. Am J Gastroenterol
1991;86:27780.
105. Ancona E, Bardini R, Nosadini A, et al. Esophagogastric anastomotic leakage. Int Surg
1982;67:1435.
106. Anderson HN. Postoperative alkaline reflux gastritis and esophagitis. Am Surg
1977;43:6707.
107. Herrington JL Jr, Sawyers JL, Whitehead WA. Surgical management of reflux gastritis. Ann
Surg 1974;180:52637.
108. Nath BJ, Warshaw AL. Alkaline reflux gastritis and esophagitis. Annu Rev Med
1984;35:38396.
109. Payne WS. Prevention and treatment of biliary-pancreatic reflux esophagitis: The role of
long-limb Roux-Y. Surg Clin North Am 1983;63:8518.
110. Keshishian JM, Smyth NP, Maxwell DD, Chua L. Dilation of difficult strictures of the
esophagus. Surg Gynecol Obstet 1984;158:815.
111.
Graham DY. Treatment of benign and malignant strictures of the esophagus. In Silvis SE
(ed). Therapeutic Gastrointestinal Endoscopy. New York, Igaku-Shoin, 1989;141.
112. Patterson DJ, Graham DY, Smith JL, et al. Natural history of benign esophageal stricture
treated by dilatation. Gastroenterology 1983;85:34650.
114. London RL, Trotman BW, DiMarino AJ, et al. Dilatation of severe esophageal strictures by an
inflatable balloon catheter. Gastroenterology 1981;80:1735.
115. Cox JGC, Winter RK, Maslin SC, et al. Balloon or bougie for dilatation of benign oesophageal
stricture? An interim report of a randomised controlled trial. Gut 1988;29:17417.
116.
Cox JGC, Buckton GK, Bennett JR. Radial pressures during stricture dilatation with bougie
and balloons. Gut 1987;28:A13701.
117. Kozarek RA. Esophageal dilation. Mayo Clin Proc 1992;67:299300.
118. Thorsen G, Rosseland AR. Endoscopic incision of postoperative stenoses in the upper
gastrointestinal tract. Gastrointest Endosc 1983;29:269.
119. Weiss W, Hold H, Neumayr A, Schuller J. Endoscopic treatment of stenosized anastomosis
after gastrectomy. Endoscopy 1977;9:2424.
120. Szanto I, Kiss J, Vamosi-Nagy I, Voros A. Endoscopic polypectomy in the segment of colon
used for oesophageal replacement. Endoscopy 1981;13:134.
121.
Jordan PH. Surgery for peptic ulcer disease. Surg Clin North Am 1991;28:265330.
122. Sachdeva AK, Zaren HA, Sigel B. Surgical treatment of peptic ulcer disease. Med Clin North
Am 1991;75:9991012.
123. Martin TR, Vennes JA, Silvis SE, Ansel HJ. A comparison of upper gastrointestinal endoscopy
and radiology. J Clin Gastroenterol 1980;2:215.
124. Knutson CO, Max MH, Ahmad W, et al. Should flexible fiberoptic endoscopy replace barium
contrast study of the upper gastrointestinal tract?. Surgery 1978;84:60915.
125. Max MH, West B, Knutson CO. Evaluation of postoperative gastroduodenal symptoms:
Endoscopy or upper gastrointestinal roentgenography?. Surgery 1979;86:57882.
126. Gohel VK, Laufer I. Double-contrast examination of the postoperative stomach. Radiology
1978;129:6017.
127. Cotton PB, Rosenberg MT, Axon ATR, et al. Diagnostic yield of fibreoptic endoscopy in the
operated stomach. Br J Surg 1973;60:62932.
128. Bowden TA Jr, Hooks VH III, Mansberger AR Jr. The stomach after surgery: An endoscopic
perspective. Ann Surg 1983;197:63744.
129. Koelz HR, Gewertz BL. The stomach. Part I: Vagotomy. Clin Gastroenterol 1979;8:30521.
130. Alexander-Williams J, Hoare AM. The stomach. Part II: Partial gastric resection. Clin
131. Spainhour JB, Webster PD. Endoscopy in the evaluation of patients following gastrectomy.
132. Donahue PE, Nyhus LM. Surgeon-endoscopists and the assessment of postoperative
patients. South Med J 1982;75:15705.
133.
Hayden WF, Read RC. A comparative study of Heineke-Mikulcz and Finney pyloroplasty. Am
J Surg 1986;116:7559.
134.
Maingot R. Abdominal Operations. 7th ed. New York, Appleton-Century-Crofts. 1980.
135. Sakai P, Filho UL, Cabrera PAI, et al. Fiberoptic endoscopy in the diagnosis of retained
gastric antrum. Endoscopy 1983;15:2468.
136. de Langen ZJ, Slooff MJM, Jansen W. The surgical treatment of postgastrectomy reflux
gastritis. Surg Gynecol Obstet 1984;158:3226.
137. Powell DC, Bivins BA, Bell RM, et al. Technical complications of Roux-en-Y
gastrojejunostomy. Arch Surg 1983;118:9225.
138. Ferguson GH, Rose M, MacLennan I, et al. Vomiting after Roux-en-Y biliary diversion:
Relationship to surgical technique. Br J Surg 1990;77:54850.
139. Waldmann D, Ruckauer K, Salm R. Early postoperative endoscopy of the operated intestine.
Endoscopy 1981;13:10812.
140. Manegold BC. Early postoperative endoscopy in the operated stomach. Endoscopy
1981;13:1047.
141. Waye JD. Surgical problems with the postoperative stomach: A technique for meaningful
endoscopic evaluation. Mt Sinai J Med 1976;43:2913.

142. Delcore R, Cheung LY. Surgical options in postgastrectomy syndromes. Surg Clin North Am
1991;71:5776.
143. Burden WR, Hodges RP, Hsu M, et al. Alkaline reflux gastritis. Surg Clin North Am
1991;71:3344.
144. Clark CG. Medical complications of gastric surgery for peptic ulcer. Compr Ther
1981;7:2632.
145.
Reber HA, Way LW. Surgical treatment of late postgastrectomy syndrome. Am J Surg
1985;129:717.
146. Davidson ED, Hersh T. The surgical treatment of bile reflux gastritis: A study of 59 patients.
Ann Surg 1980;192:1758.
147. Ritchie WP Jr. Alkaline reflux gastritis: An objective assessment of its diagnosis and
treatment. Ann Surg 1980;192:28898.
148. Fiore C, Malangoni MA, Broadie TA, et al. Surgical management of alkaline reflux gastritis.
Arch Surg 1982;117:68994.
149. Boren CH, Way LW. Alkaline reflux gastritis: A re-evaluation. Am J Surg 1980;140:406.
150. Van Heerden JA, Phillips SF, Adson MA, McIlrath CC. Postoperative reflux gastritis. Am J
Surg 1975;129:828.
151. Berardi RS, Siroospour D, Ruiz R, et al. Alkaline reflux gastritis. A study in forty postoperative
duodenal ulcer patients. Am J Surg 1976;132:5527.
152. Hoare AM, Keighley MRB, Starkey B, et al. Measurement of bile acids in fasting gastric
aspirates: An objective test for bile reflux after gastric surgery. Gut 1978;19:1669.
153. Keighley MRB, Asquith P, Edwards JAC, et al. The importance of an innervated and intact
antrum and pylorus in preventing postoperative duodenogastric reflux and gastritis. Br J Surg
1975;62:8459.
154. Keighley MRB, Asquith P, Alexander-Williams J. Duodenogastric reflux: A cause of gastric
mucosal hyperaemia and symptoms after operations for peptic ulceration. Gut
1975;16:2832.
155.
Moismann F, Burri B, Diserens H, et al. Enterogastric reflux: Experimental and clinical study.
Scand J Gastroenterol 1981;16(Suppl 67):14952.
156. Geboes K, Rutgeerts P, Broeckaert L, et al. Histologic appearances of endoscopic gastric
mucosal biopsies 1020 years after partial gastrectomy. Ann Surg 1980;192:17982.
157. Domellof L, Eriksson S, Janunger K-G. Carcinoma and possible precancerous changes of the
gastric stump after Billroth II resection. Gastroenterology 1977;73:4628.
158. Johnston DH. A biopsy study of the gastric mucosa in postoperative patients with and without
marginal ulcer. Am J Gastroenterol 1966;46:10318.
159. Burbige EJ, French SW, Tarder G, Belber JP. Correlation between gross appearance and
histologic findings in the postoperative stomach. Gastrointest Endosc 1979;25:35.
160. Stempien SJ, Dagradi AE, Tan DT. Endoscopic aspects of the gastric mucosa ten years or
more after vagotomy-pyloroplasty. Gastrointest Endosc 1971;18:212.
161.
Lees F, Grandjean LC. The gastric and jejunal mucosae in healthy patients with partial
gastrectomy. Arch Intern Med 1958;101:94351.
162. Hoare AM, Jones EL, Alexander-Williams J, Hawkins CF. Symptomatic significance of gastric
mucosal changes after surgery for peptic ulcer. Gut 1977;18:295300.
163. Savage A, Jones S. Histological appearances of the gastric mucosa 1527 years after partial
gastrectomy. J Clin Pathol 1979;32:17986.
164. Langhans P, Heger RA, Hohenstein J, et al. Gastric carcinoma after gastric surgeryAn
experimental study. Hepatogastroenterology 1981;28:347.
165. Siurala M, Lehtola J, Ihamaki T. Atrophic gastritis and its sequelae: Results of 1923 years
follow-up examinations. Scand J Gastroenterol 1974;9:4416.
166. Kliems G, Paquet KJ, Lindstaedt H, et al. Atrophic gastritis after Billroth I gastrectomy.
Endoscopy 1979;11:12730.
167. Reynolds KW, Johnson AG, Fox B. Is intestinal metaplasia of the gastric mucosa a
premalignant lesion?. Clin Oncol 1975;1:1019.

168. Korn ER. Intestinal metaplasia of the gastric mucosa: Endoscopic recognition and review of
the literature. Am J Gastroenterol 1974;61:2704.
169.
Morson BC. Carcinoma arising from areas of intestinal metaplasia in gastric mucosa. Br J
Cancer 1955;9:37785.
170. McFadden DW, Zinner MJ. Reoperation for recurrent peptic ulcer disease. Surg Clin North
Am 1991;71:7791.
171. Siim C, Lublin HK, Jensen HE. Selective gastric vagotomy and drainage for duodenal ulcer: A
1013-year follow-up study. Ann Surg 1981;194:68791.
172. Welch CE, Malt RA. Medical progress. Abdominal surgery. N Engl J Med 1983;308:62432,
68595, 75360.
173. Grossman MI (Moderator). Peptic ulcer: New therapies, new diseases. Ann Intern Med
1981;95:60927.
174. Thompson BW, Read RC. Secondary operations for duodenal ulcer. Am J Surg
1977;134:75862.
175. Stabile BE, Passaro E Jr. Recurrent peptic ulcer. Gastroenterology 1976;70:12435.
176. Sharaiha ZK, Smith JL, Cain GD, et al. Recurrent ulcers after gastric surgery: Endoscopic
localization to gastric mucosa. Am J Gastroenterol 1983;78:26971.
177. Proceedings of the Consensus Conference on therapeutic endoscopy in bleeding ulcers.
Gastrointest Endosc 1990;36:S1-S63.
178.
Balfour DC. Factors influencing the life expectancy of patients operated on for gastric ulcer.
Ann Surg 1922;76:4058.
179. Northfield TC, Hall CN. Carcinoma of the gastric stump: Risks and pathogenesis. Gut
1990;31:12179.
180.
Elder JB, Knight T. Surgical suppression of gastric acid secretion. Scand J Gastroenterol
1991;26(Suppl 188):2632.
181. Stemmermann GN, Nomura AMY, Chyou PH. Cancer incidence following subtotal
gastrectomy. Gastroenterology 1991;101:7115.
182. Termesmette AC, Offerhaus GJA, Giardiello FM, et al. Long-term prognosis after partial
gastrectomy for benign conditions. Gastroenterology 1991;101:14853.
183. Eberlein TJ, Lorenzo FV, Webster MW. Gastric carcinoma following operation for peptic ulcer
disease. Ann Surg 1978;187:2516.
184. Kobayashi S, Prolla JC, Kirsner JB. Late gastric carcinoma developing after surgery for
benign conditions. Endoscopic and histologic studies of the anastomosis and diagnostic
problems. Am J Dig Dis 1970;15:90512.
185. Domellof L, Janunger K-G. The risk for gastric carcinoma after partial gastrectomy. Am J Surg
1977;134:5814.
186. Klarfeld J, Resnick G. Gastric remnant carcinoma. Cancer 1979;44:112933.
187. Osnes M, Lotveit T, Myren J, Serck-Hanssen A. Early gastric carcinoma in patients with a
Billroth II partial gastrectomy. Endoscopy 1977;9:459.
188. Totten J, Burns HJG, Kay AW. Time of onset of carcinoma of the stomach following surgical
treatment of duodenal ulcer. Surg Gynecol Obstet 1983;157:4313.
189. Orlando R III, Welch JP. Carcinoma of the stomach after gastric operation. Am J Surg
1981;141:48791.
190. Mortensen NJMcC, Thomas WEG, Jones SM. Endoscopic screening for premalignant
changes 25 years after gastrectomy: Results of a five-year prospective study. Br J Surg
1984;71:36371.
191. Kivilaakso E, Hakkiluoto A, Kalima TV, et al. Relative risk of stump cancer following partial
gastrectomy. Br J Surg 1977;64:3368.
192.
DeJode LR. Gastric carcinoma following gastroenterostomy and partial gastrectomy. Br J
Surg 1961;48:5124.
193.
Liavaag K. Cancer development in gastric stump after partial gastrectomy for peptic ulcer.
Ann Surg 1962;155:1036.
194. Dougherty SH, Foster CA, Eisenberg MM. Stomach cancer following gastric surgery for
benign disease. Arch Surg 1982;117:2948.
195. Logan RFA, Langman MJS. Screening for gastric cancer after gastric surgery. Lancet
1983;2:66770.
196. Schafer LW, Larson DE, Melton LJ III, et al. The risk of gastric carcinoma after surgical
treatment for benign ulcer disease. N Engl J Med 1983;309:12103.
197. Ross AHM, Smith MA, Anderson JR, et al. Late morality after surgery for peptic ulcer. N Engl
J Med 1982;307:51922.
198. Sonnenberg A. Endoscopic screening for gastric stump cancerWould it be beneficial? A
hypothetical cohort study. Gastroenterology 1984;87:48995.
200.
Winawer SJ. Tissue diagnosis in upper gastrointestinal malignancy. Gastroenterology
1982;82:37082.
201. Green PH, O'Toole KM, Weinberg LM, Goldfarb JP. Early gastric cancer. Gastroenterology
1981;81:24756.
202. Pickford IR, Craven JL, Hall R, et al. Endoscopic examination of the gastric remnant 3139
years after subtotal gastrectomy for peptic ulcer. Gut 1984;25:3937.
203. Schrumpf E, Serck-Hanssen A, Stadaas J, et al. Mucosal changes in the gastric stump 2025
years after partial gastrectomy. Lancet 1977;2:4679.
204. Savalgi RS, Corbishley CM, Caygill C, et al. Relation between severity and extent of
precancerous lesions in the postoperative stomach. Lancet 1990;336:4136.
205. Lundegardh G, Adami H-O, Helmick C, et al. Stomach cancer after partial gastrectomy for
benign ulcer disease. N Engl J Med 1988;319:195200.
206. Stael von Holstein C, Eriksson S, Huldt B, Hammar E. Endoscopic screening during 17 years
for gastric stump carcinoma. A prospective clinical trial. Scand J Gastroenterol
1991;26:10206.
207. Stael von Holstein C, Hammar E, Eriksson S, Huldt B. Clinical significance of dysplasia in
gastric remnant biopsy specimens. Cancer 1993;72:15325.
208.
American Society for Gastrointestinal Endoscopy. Appropriate Use of Gastrointestinal
Endoscopy. Manchester, MA: ASGE, 1992;68.
209. Bono JA. Upper gastrointestinal fiberoptic endoscopy reveals the silk suture-line ulcer. Am
Surg 1978;44:2825.
210. Sivelli R, Farinon AM, Sianesi M, et al. Technetium-99m HIDA hepatobiliary scanning in
evaluation of afferent loop syndrome. Am J Surg 1984;148:2625.
211. Papchristou DN, Fortner JG. Adenocarcinoma of the gastric cardia: The choice of
gastrectomy. Ann Surg 1980;192:5864.
212. Zacho A, Cederqvist C, Fischerman K. Surgical treatment of gastric malignancies. A
twenty-year series comprising mainly far advanced and high-seated tumors. Ann Surg
1974;179:94101.
213. Cady B, Ramsden DA, Stein A, et al. Gastric cancer. Contemporary aspects. Am J Surg
1977;133:4239.
214. Dupont JB Jr, Lee Jr, Burton GR, et al. Adenocarcinoma of the stomach: Review of 1,497
cases. Cancer 1978;41:9417.
215. Nachlas MM, Crawford DT, Pearl JM. Current status of jejunoileal bypass in the treatment of
morbid obesity. Surg Gynecol Obstet 1980;150:25670.
216. Hocking MP, Duerson MC, O'Leary JP, et al. Jejunoileal bypass for morbid obesity: Late
follow-up in 100 cases. N Engl J Med 1983;308:9959.
217. Halverson JD, Zuckerman GR, Koehler RE, et al. Gastric bypass for morbid obesity: A
medical-surgical assessment. Ann Surg 1981;194:15260.
218. Thompson WR, Amaral JF, Caldwell MD, et al. Complications and weight loss in 150
consecutive gastric exclusion patients: A clinical review. Am J Surg 1983;146:60212.
219. NIH Conference. Gastrointestinal surgery for severe obesity. Consensus Development
Conference Panel. Ann Intern Med 1991;115:95661.

220. Kim CH, Sarr MG. Severe reflux esophagitis after vertical banded gastroplasty for treatment
of morbid obesity. Mayo Clin Proc 1992;67:335.
221. Halverson JD, Koehler RE. Assessment of patients with failed gastric operations for morbid
obesity. Am J Surg 1983;145:35763.
222. Mason EE, Scott DH. Reoperation for failed gastric bypass procedures for obesity. Surg Clin
North Am 1991;71:4555.
223. Kretzschmar CS, Hamilton JW, Wissler DW, et al. Balloon dilation for the treatment of stomal
stenosis complicating gastric surgery for morbid obesity. Surgery 1987;102:4436.
224. Ellison EC, Martin EW Jr, Laschinger J, et al. Prevention of early failure of stapled gastric
partitions in treatment of morbid obesity. Arch Surg 1980;115:52833.
225. Wolper JC, Messmer JM, Turner MA, Sugerman HJ. Endoscopic dilation of late stomal
stenosis. Its use following gastric surgery for morbid obesity. Arch Surg 1984;119:8367.
226. Paulk SC. Formal dilation after gastric partitioning. Surg Gynecol Obstet 1983;156:5024.
227. Printen KJ, Scott D, Mason EE. Stomal ulcers after gastric bypass. Arch Surg
1980;115:5257.
228. Printen KJ, LeFavre J, Alden J. Bleeding from the bypassed stomach following gastric bypass.
Surg Gynecol Obstet 1983;156:656.
229. Sataloff DM, Lieber CP, Seinige UL. Strictures following gastric stapling for morbid obesity.
Results of endoscopic dilatation. Am Surg 1990;56:16774.
230. Grace DM. Recognition and management of Marlex erosion after horizontal gastroplasty for
morbid obesity. Can J Surg 1987;30:2825.
231. Fardy J, Holliday R, Sullivan S. Marlex mesh erosion: A complication of horizontal banded
gastroplasty. Gastrointest Endosc 1988;34:4612.
232. Madden JL, Chun JY, Kandalaft S, et al. Choledochoduodenostomy: An unjustly maligned
surgical procedure?. Am J Surg 1970;119:4554.
233. Hallstone A, Triadafilopoulos G. "Spontaneous sump syndrome": Successful treatment by
duodenoscopic sphincterotomy. Am J Gastroenterol 1990;85:151820.
234. McSherry CK, Fisher MG. Common bile duct stones and biliary-intestinal anastomoses. Surg
235. Akiyama H, Ikezawa H, Kameya S, et al. Unexpected problems of external
choledochoduodenostomy: Fiberscopic examination in 15 patients. Am J Surg
1980;140:6605.
236.
Siegel JH. Duodenoscopic sphincterotomy in the treatment of the "sump syndrome." Dig Dis
Sci. 26:92281981.
237. Barkin JS, Silvis S, Greenwald R. Endoscopic therapy of the "sump" syndrome. Dig Dis Sci
1980;25:597601.
238. Teplick SK, Wolferth CC Jr, Hayes MF Jr, et al. Balloon dilatation of benign postsurgical
biliary-enteric anastomotic strictures. Gastrointest Radiol 1982;7:30710.
239.
Marbet UA, Stalder GA, Faust H, et al. Endoscopic sphincterotomy and surgical approaches
in the treatment of the "sump syndrome." Gut. 28:14251987.
240. Traverso LW, Tompkins RK, Urrea PT, et al. Surgical treatment of chronic pancreatitis:
Twenty-two years' experience. Ann Surg 1979;190:3129.
241. Potts JR III, Moody FG. Surgical therapy for chronic pancreatitis: Selecting the appropriate
approach. Am J Surg 1981;142:6549.
242. Bolman RM III. Surgical management of chronic relapsing pancreatitis. Ann Surg
1981;193:12531.
243. Cooperman AM. Chronic pancreatitis. Surg Clin North Am 1981;61:7183.
244.
Puestow CB, Gillesby WJ. Retrograde surgical drainage of pancreas for chronic relapsing
pancreatitis. AMA Arch Surg 1958;76:698707.
245.
DuVall MJ Jr. Caudal pancreatico-jejunostomy for chronic relapsing pancreatitis. Ann Surg
1954;140:77585.
246. Kugelberg CH, Whelin L, Arnesjo B, et al. Endoscopic pancreatography in evaluating results
of pancreaticojejunostomy. Gut 1976;17:26772.

Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Ann
Surg 1935;102:7639.
248. Connolly JE, Kwaan JH, McCart PM, et al. Aortoenteric fistula. Ann Surg 1981;194:40212.
249.
Champion MC, Sullivan SN, Coles JC, et al. Aortoenteric fistula: Incidence, presentation,
recognition, and management. Ann Surg 1982;195:347.
250. Perdue GD Jr, Smith RB III, Ansley JD, et al. Impending aortoenteric hemorrhage: The effect
of early recognition on improved outcome. Ann Surg 1980;192:23743.
251. Mir-Madjlessi SH, Sullivan BH Jr, Farmer RG, Beven EG. Endoscopic diagnosis of
aortoduodenal fistula. Gastrointest Endosc 1973;19:1878.
252. Baker MS, Fisher JH, van der Reis L, et al. The endoscopic diagnosis of an aortoduodenal
fistula. Arch Surg 1976;111:304.
253. Kleinman LH, Towne JB, Bernhard VM. A diagnostic and therapeutic approach to aortoenteric
fistulas: Clinical experience with twenty patients. Surgery 1979;86:86880.
254. Martin J, Cano N, Di-Costanzo J, et al. Aortoduodenal fistula: Endoscopic diagnosis. Dig Dis
Sci 1981;26:9567.
247.
Chapter 53 Esophagogastro-Duodenoscopy in the Pediatric Patient

(5217)
(5218)
JOHN TIMOTHY BOYLE, M.D.
Esophagogastroduodenoscopy (EGD) is an integral part of the practice of pediatric

gastroenterology.15(5219) Endoscopic technology has continued to improve over the years. The
indications for upper endoscopy are broad, and the diagnostic accuracy and therapeutic potential are
high. The value of diagnostic accuracy, however, is only relevant if it is consistently associated with
improvements in management and outcome.
This chapter focuses on the clinical indications for and practical technical aspects of upper
gastrointestinal endoscopy in children. Indications for pediatric EGD are listed in Table 531. The
major differences between EGD in pediatric and in adult patients are in sedation and analgesia,
limitations imposed by accessory channel diameter, the absence of cancer in most differential
diagnoses, and emphasis on histologic diagnosis.
Indications for Upper Gastrointestinal Endoscopy in

Pediatric Patients
TABLE 531
Upper gastrointestinal bleeding

Diagnostic evaluation
Control of bleeding
Sclerosis of varices
Diagnostic evaluation of reflux esophagitis in infancy
Dysphagia, odynophagia, persistent refusal to eat, chest pain
Unexplained, persistent vomiting
Dyspepsia or suspected acid-peptic disease
Acute upper abdominal pain associated with anorexia, weight loss despite a course of
therapy for suspected acid-peptic disease
Chronic dyspepsia associated with significant morbidity (e.g., anorexia, weight loss,
school absenteeism, limitation of activity)
Abnormal or equivocal imaging studies of the gastrointestinal tract (e.g., barium meal x-ray
film, gastric scintigraphy)
Iron deficiency anemia with negative results on colon evaluation
Diagnosis that affects patient management
Indications for Upper Gastrointestinal Endoscopy in

Pediatric Patients
TABLE 531
When sampling duodenal and jejunal iissue or fluid is indicated

Early diagnosis of esophageal varices
Caustic ingestion
Therapeutic endoscopy
Removal of foreign body
Dilation of stricture, web, rings
Percutaneous gastrostomy
Placement of a feeding tube in the duodenum or jejunum
Facilities, Sedation, Instruments, and Procedure

Endoscopic examinations on infants and children should only be undertaken in a setting in which
physician and support staff are trained to deal with the routine and emergency care of the pediatric
patient.2(5220) The endoscopy suite should be designed with the safety and the physical and
emotional comfort of the patient in mind.1(5221) All equipment must be appropriate in size for the
patient, including suction devices, mouth guards, restraints, and intravenous setups. Resuscitative
equipment, including Ambu bag and masks, endotracheal tubes, and laryngoscopes, must be of
appropriate size for the pediatric patient, readily available, and checked regularly to ensure that all
necessary devices are in good working order.
Three individuals are required when pediatric endoscopy is performed in the endoscopy suite: the
physician, at least one registered nurse with training in pediatrics, and a medical technician. The
physician and nurse must be familiar with appropriate dosages of sedative and resuscitative
medications. Pulse oximetry should be standard for all pediatric endoscopic procedures, because it
has been shown to be more sensitive than clinical observation in detecting hypoxia.6,7(5222) Oxygen
and wall suction should be readily available. Automated monitoring of blood pressure is also desirable.
As in adults, attention must be given before endoscopy to respiratory status, potential cardiopulmonary
problems, and skeletal abnormalities.
Although it is controversial, my colleagues and I routinely administer antibiotic prophylaxis before
diagnostic EGD to patients with valvular congenital heart disease. As recommended by the American
Heart Association, the standard antibiotic regimen before the procedure is intravenous ampicillin and
gentamicin.8(5223) For patients with a history of a penicillin allergy, vancomycin is substituted for
ampicillin. A second dose of oral amoxicillin is given 6 hr after the procedure.
The duration of fasting before EGD depends on the patient's age: 4 hr before the procedure for infants
younger than 6 months of age, 6 hr for patients between the ages of 6 and 36 months, and 8 hr for
those older than 36 months of age. Data suggest that it may be safe to give clear liquids up to 2 hr
before general anesthesia without affecting the volume or pH of gastric contents at the time of
anesthesia.9(5224) This is not our routine practice, unless there is a concern regarding delay of the
procedure.
Topical oropharyngeal anesthetics are useful in older children undergoing EGD.10(5225) Younger
patients may be frightened or agitated by the loss of oral sensation.
The goals of sedation and analgesia are to relax the patient, minimize discomfort, facilitate safe and
efficient performance of the procedure, and induce amnesia.
Although conscious sedation is usually adequate for performing procedures in older children and
adolescents, deep sedation or general anesthesia is required for EGD in young children.1,11(5226) A
significant advantage of general anesthesia is that a larger-diameter endoscope can be safely used
because the airway is protected. As discussed later, this becomes an important issue in therapeutic
endoscopy.
Although a few centers still routinely use general anesthesia for diagnostic EGD, most infants and
children can be adequately sedated with a combination of intravenous narcotic and benzodiazepine.
The exact choice of drugs varies widely among centers.
Drug metabolism is highly variable in children. Children frequently require much higher doses of
sedatives on a milligram-per-kilogram basis than adults.
Diazepam (0.2 to 0.4 mg/kg, maximum of 10 mg) and midazolam (0.1 to 0.2 mg/kg, maximum of 5
mg) are the two most commonly used benzodiazepines. Both are sedatives and amnestics, but neither
provides analgesia. Both drugs must be infused slowly and carefully titrated to avoid respiratory
depression. Both agents are reversible with the benzodiazepine antagonist flumazenil (0.2 mg given
over 60 sec). The major difference between these two agents is duration of sedation. Midazolam is
water soluble and therefore shorter acting. The choice of a benzodiazepine may reflect anticipated
procedure duration, potentially an important factor in centers with training programs.
Narcotics provide analgesia and sedation when used in combination with benzodiazepines. Meperidine
(1 to 2 mg/kg, maximum of 100 mg) and fentanyl (2 to 3 g/kg) are the two most commonly used
narcotics. More centers are beginning to use ketamine in toddlers and young children, who are often
difficult to sedate because of fear and concern.12,13(5227) Intravenous ketamine (0.25 mg/kg, titrating
up to a maximum dose of 2 mg/kg) is a dissociative anesthetic agent that produces analgesia and
amnesia. An adequately sedated patient has a blank stare and nystagmus. Unlike other anesthetics,
ketamine is a cardiopulmonary stimulant. Oral and tracheobronchial secretions are increased but can
be controlled by premedication with an anticholinergic agent, glycopyrrolate (6 to 10 g/kg).
Pretreatment with intravenous diazepam (0.1 to 0.2 mg/kg) helps to prevent distressing emergence
phenomena such as hallucinations, nightmares, and delirium. The latter are rare in children younger
than 10 years, which is the upper age limit for administration of ketamine.
A fully equipped pediatric endoscopy suite has flexible instruments in three ranges of tip diameters for
upper gastrointestinal endoscopy: 5.3 mm (2-mm accessory channel diameter), 7.5 to 7.9 mm (2-mm
accessory channel diameter), and 9.8 mm (2.8-mm accessory channel diameter). Larger-diameter
endoscopes provide more light and therefore better visualization, improved ability to aspirate fluid and
blood, better-quality biopsies, and a greater range of options for therapeutic interventions. The
disadvantage is that larger endoscopes, by causing partial obstruction of the oral airway and
compressing the trachea in younger patients, may produce dyspnea, especially if there is associated
obstruction of the nasal airway, which may reduce the margin of safety during conscious sedation. As a
general rule, my colleagues and I use endoscopes with 9.8-mm diameters in children who weigh more
than 10 kg, 7.9-mm diameters for patients between 2500 g and 10 kg, and 5.3-mm diameters for
premature infants less than 2500 g. The key is to adapt the choice of instrument to the patient. Each
patient should be treated as an individual, and decisions regarding instrument size and the type of
sedation and analgesia should be based on the goals of the procedure, patient comfort, and skill of the
endoscopist.
Video endoscopic technology is invaluable for procedures such as foreign body extraction,
sclerotherapy, and percutaneous gastrostomy, which require the endoscopy assistant to work in
concert with the endoscopist. Video technology is available only for endoscopes with insertion tubes
with diameters of 9.8 mm or greater.
Upper gastrointestinal endoscopy in infants and children is performed with a technique similar to that
used in adults. With the patient positioned on his or her left side, the endoscope is gently advanced
through the hypopharynx, preferably under direct vision (see Chapter 38: Technique of Upper
Gastrointestinal Endoscopy). The endoscopist must be familiar with the appearance of the
hypopharynx. The piriform fossae lie on opposite sides of the esophageal opening. In older children,
the cricopharyngeal sphincter can be opened by asking the patient to swallow. The piriform fossae are
obliterated by this maneuver, allowing the instrument to be safely advanced into the esophagus. In
infants and heavily sedated patients, the tip should be directed posterior to visualize the slitlike aperture
of the esophagus. At this point, gentle pressure with the tip of the endoscope, a puff of air, or squirt of
water usually triggers the gag reflex, thereby opening the upper esophageal sphincter and allowing the
tip to be advanced into the esophagus.
The tips of the small-caliber endoscopes used for EGD in infants are very flexible. I refer to them as
lumen-seeking instruments. Use of the deflection knobs to drive the instrument tip into the esophagus
should be avoided. The instrument should be passed in the same manner as an orogastric or
nasogastric tube, with the operator feeling his or her way with the right hand in guiding the instrument
into the esophagus. Rarely, in heavily sedated or neurologically impaired children, the tip of the
endoscope inadvertently slips into the upper trachea. This usually elicits coughing, but not always. The
horizontal tracheal rings then become an important landmark and indicate that the endoscope is in the
wrong place.
The esophagus is best examined during insertion. As in adults, the mucosal surface of the partially
distended, normal esophagus may appear lumpy or irregular at first, suggesting abnormal mucosa.
Further inflation defines the smooth, featureless tube lined by whitish-pink epithelium. Peristaltic
movements are commonly seen in the middle and lower third of the esophagus. Transmitted cardiac
pulsations are usually prominent in the distal esophagus. The observation of free gastroesophageal
reflux (GER) during upper gastrointestinal endoscopy is a nonspecific finding.
As in adults, small intraepithelial venules are commonly seen in the esophagus of the pediatric patient.
A fine, longitudinally running venous palisade is often present just proximal to the squamocolumnar
junction. This has no clinical significance. The squamocolumnar junction line may or may not be
demarcated by an acute transition from the pale esophageal mucosa to the rubric gastric columnar
epithelium. The squamocolumnar junction may have a ringlike or undulating configuration. It is
important to distinguish swallowed lidocaine (Xylocaine) anesthetic, which may look like whitish
patches.
The endoscope should enter the stomach without resistance. On entry, the tip should be deflected
downward (posteriorly) and slightly to the left (toward the spleen) before inflating the stomach. It is
often difficult to suction saliva or regurgitated bile-stained froth through the small-diameter accessory
or suction channel of small-caliber endoscopes. The significance of the gastric bile-stained froth in a
deeply sedated patient is debatable, but a note should be made of its presence. The instrument can
usually be advanced with ease along the greater curvature of the stomach while gradually angulating
the tip upward and rotating as necessary to the right to bring the pylorus and angulus (incisura) into
view.
As in adults, the normal pylorus in children is a ring or diaphragm rather than a channel. The presence
of a channel usually indicates ulceration, edema, or scarring. Cicatricial deformity of the pylorus
suggests current or previous ulceration. A sentinel prepyloric fold implies an active ulcer or prior
ulceration distal to the pyloric ring.
It is easier to examine the stomach during withdrawal of the endoscope. The incisura, lesser curve,
cardia, gastroesophageal junction, and fundus are best viewed on withdrawal of the instrument with the
tip in the upward deflected (i.e., retroverted, retroflexed) or J-position. Mucosal vessels are not
normally visible in the antrum. The relative prominence of gastric rugal folds varies between
individuals. It is common to find punctate erythema or red lines along the crests of gastric rugae. These
are of no particular significance, although they may be a sign of portal gastropathy in patients with
portal hypertension. The fundus shows a more marked vascular pattern than the body.
A variety of mucosal ecchymoses may be produced when the endoscope is maneuvered roughly over
the mucosa or forcibly pushed against it. Suction artifacts are easy to induce during aggressive
suctioning and are often interpreted by inexperienced endoscopists as abnormalities when viewed
during retroflexion. These appear as cherry red, round, raised lesions or flattened lesions with a red
ring. It is important to differentiate these from traumatic erosions caused by a previous indwelling
nasogastric tube, which are often multiple and superficially eroded.
The view of the duodenal mucosa on advancement of the tip of the endoscope through the pylorus is
usually obscured until the endoscope is withdrawn slightly. At this point, the superior duodenal fold
usually becomes obvious, marking the junction between the first and second portions of the
duodenum. The instrument tip is advanced to the fold separating first and second parts of the
duodenum and then deflected rightward and upward while rotating (torquing) the instrument 90
degrees to the right. When the endoscope first passes the superior duodenal fold, the tubular structure
of the duodenum becomes apparent with the characteristic valvulae conniventes. These are usually
thin, ringlike structures, although some assume irregular patterns. With the levels of magnification
afforded by video endoscopes, the digitate villi and leaf and ridge forms are appreciated. The papilla of
Vater is usually easily identified between folds.
Some incidental findings that may be discovered during EGD include the following:
1. Inlet patches of pink gastric mucosa in the proximal esophagus, not to be confused with Barrett's
esophagus
2. Glycogenic acanthosis in the distal esophagus, characterized by round, pale, elevated plaques
3. Heterotopic pancreatic tissue, characterized by an umbilicated dimple in the gastric antrum
4. Cavernous hemangioma or angioma, characterized by a submucosal flat or raised, bluish lesion
5. Congenital gastric diverticula, usually situated in the proximal stomach and best seen on
retroflexion (a wide mouth is more common than a narrow mouth)
6. Congenital double pylorus
7. Congenital duodenal diverticula, commonly found on the medial wall of the duodenum, adjacent to
the papilla of Vater
8. Multiple small sessile polyp-like structures in the duodenum, indicating Brunner's gland
hypertrophy, nodular lymphoid hyperplasia, or gastric metaplasia
9. A linear scar in an otherwise normal stomach, indicating the site of a previous gastrostomy
10. Evidence of prior pyloroplasty, characterized by a gastric antrum that funnels directly into the
duodenum
The incidence of significant complications from EGD in pediatric patients ranges from 0.5 to
l.7%.1,14(5228) Most EGD complications are related to administration of sedative medications rather
than the procedure itself. Complications related to sedation include respiratory depression, paradoxical
excitation, bronchospasm, aspiration, allergic reactions to drugs, venous irritation and phlebitis related
to intravenous infusion of drugs, and flushing or nasal pruritus due to meperidine. Topical
oropharyngeal analgesia may trigger laryngospasm. Reactions after sedation include nausea, vomiting,
and disorientation. Mechanical complications from the procedure include sore throat, broken or loose
teeth, trauma to the piriform sinus, retropharyngeal hematoma, duodenal hematoma, and perforation.
Complications of biopsy sampling include intraluminal bleeding, intramural hematoma, and perforation.
Indications for Diagnostic Upper Gastrointestinal Endoscopy

The clinical presentation together with nasogastric lavage can delineate the level of bleeding within the
gastrointestinal tract in most cases. The presence of bright red or coffee-grounds blood in the stomach
is diagnostic of an upper gastrointestinal bleed, and gastric lavage allows the clinician to estimate the
volume of blood lost and whether the bleeding has stopped.
The age-related causes of upper gastrointestinal bleeding in pediatric patients are listed in Table
532.15(5229) Two diagnostic categories account for 95% of cases: mucosal lesions and esophageal
varices. The object of history, physical examination, and laboratory tests is primarily to distinguish
between these two categories. Although lesions in both categories may present with severe bleeding
and hemodynamic instability, bleeding from mucosal lesions is usually self-limited and responds to
conservative measures. Variceal bleeding also usually stops spontaneously, but it carries a worse
prognosis for recurrent bleeding and requires early therapy. As a general rule, all pediatric patients with
a history suggesting upper gastrointestinal bleeding should be hospitalized for observation.
Causes of Upper Gastrointestinal

Bleeding Based on Age
TABLE 532
INFANT (< 2 YR)

Swallowed maternal blood
Bleeding diathesis
Stress ulcer
Hemorrhagic gastritis
Esophagitis
Mallory-Weiss tear
Foreign body irritation
(nasogastric tube)
Pyloric stenosis
Vascular malformation
Duplication
CHILD OR ADOLESCENT (216 YR)

Gastric ulcer
Duodenal ulcer
Helicobacter pylori gastritis
Mallory-Weiss tear
Esophagitis
Esophageal varices
Gastric varices
Portal hypertensive gastropathy
Foreign body
Crohn's disease
Vascular malformation
Hemobilia
Toxic ingestion
Esophageal pill-induced ulcer
Henoch-Schnlein purpura
Leiomyoma (sarcoma)
Gastric lipoma
The goals of endoscopy in upper gastrointestinal bleeding are to localize the site of bleeding to the
esophagus, stomach, or duodenum; to diagnose the specific cause of bleeding; and to initiate
therapeutic intervention when indicated. Emergency endoscopy is only necessary when the patient
continues to bleed at a rate considered to be life-threatening (e.g., copious witnessed blood,
continuous transfusion requirement, hemodynamic instability) despite aggressive medical therapy. The
diagnostic yield for endoscopy performed 6 to 12 hr after presentation is as good as when performed
immediately on admission. If the patient responds quickly to resuscitation and the nasogastric lavage
clears, it is unnecessary to perform endoscopy on an emergency basis in the middle of the night.
Gastrointestinal bleeding may be a complication of all acute and chronic upper gastrointestinal
mucosal lesions, including esophagitis, Mallory-Weiss tears, gastritis, duodenitis, stress ulcers, peptic
ulceration, lipomas, and vascular malformations. The diagnosis is usually suspected by an antecedent
history of vomiting or abdominal pain or by the absence of a history and physical findings suggestive of
chronic liver disease or portal hypertension.
Endoscopy, when performed by a well-trained physician, is the most sensitive and specific diagnostic
procedure for determining the cause and site of upper gastrointestinal mucosal bleeding. A specific
diagnosis should be pursued in patients with:
1. Active bleeding documented by nasogastric lavage
2. Evidence of severe hemorrhage, including hemodynamic instability requiring resuscitation with
intravenous fluids or an equilibrated hemoglobin level less than 10 g/dl
3. Conditions that affect clotting, such as a catabolic state or serious acute or chronic disease
4. A history of previous unexplained gross or occult bleeding or unexplained iron deficiency anemia
5. A history of chronic dyspepsia (e.g., vomiting, abdominal pain, nausea, oral regurgitation,
heartburn, dysphagia)
Acute, self-limited blood loss in patients with a history of forceful emesis, symptoms of infectious
gastroenteritis, or recent ingestion of ulcerogenic drugs justifies medical therapy based on a
presumptive diagnosis alone (e.g., Mallory-Weiss tear, acute gastritis, drug-induced ulcer) without
endoscopy.
Diagnostic and therapeutic endoscopy should only be performed after the patient's vital signs have
been stabilized. If active bleeding has stopped, EGD usually can be accomplished with intravenous
conscious sedation. Intubation for protection of the airway is indicated in actively bleeding infants or
small children or neurologically impaired children in whom therapeutic endoscopy is contemplated. It is
relatively common for blood clots to be vomited during endoscopy. General anesthesia may be
particularly useful in patients with profuse bleeding, because bleeding often stops during anesthesia.
Lavage with saline at room temperature using a large-bore sump or Ewald tube is essential to
evacuate as much blood and clot as possible from the stomach before introduction of the endoscope.
The selection of an endoscope must be tailored to the individual patient. However, an endoscope with
the largest possible diameter should be used to facilitate washing of the mucosal surface and suction
of blood, which otherwise may obscure the bleeding site. The first goal is to localize the site of
bleeding. Fresh red blood quickly changes color to brown in an acid environment. It is common to find
a large amount of old, reddish-appearing clot lying in a pool of brownish or purplish blood in the fundus.
Such clots cannot be aspirated through the endoscope. It is best to distend the stomach by air
insufflation and maneuver the endoscope past the clot to inspect the greater part of the stomach and
duodenum. If bleeding or oozing from an area obscured by clot is suspected, the area can often be
effectively washed clean by a jet of water from a heat probe passed through the accessory channel of
the endoscope.
The possibility of bleeding esophageal varices is suggested by a medical history of jaundice, hepatitis,
blood transfusion, chronic right heart failure, pulmonary hypertension, exchange transfusion,
omphalitis, or umbilical vein catheterization and by physical findings of splenomegaly,
hepatosplenomegaly, jaundice, or ascites. Endoscopy should only be performed after the patient has
been resuscitated and vital signs are stable. Children with known varices who are receiving propranolol
therapy may be unable to mount a compensatory increase in cardiac output in response to a major
hemorrhage.
Esophageal varices run upward from the gastro-esophageal junction. Their surface tends to have a
blue tint, and their outlines are usually beaded. The greater their diameters, the prouder they stand.
Gastric varices, which seldom occur in the absence of esophageal varices, may not be easy to
recognize because they may resemble normal gastric rugae. Actively bleeding esophageal or gastric
varices or an overlying clot on a varix confirms variceal bleeding. Other possible sites of bleeding in
children with portal hypertension include portal (hypertensive, congestive) gastropathy, gastric or
duodenal ulcer, esophagitis, and gastric varices.16(5230) Mild portal gastropathy is characterized by a
mosaic or snakeskin pattern of superficial reddening, pink speckling, or petechial hemorrhages. The
severe form mimics hemorrhagic gastritis. Portal gastropathy may involve any area of the
stomach.17(5231)
Reflux Esophagitis in Infancy

In infancy, GER is perceived as a functional developmental disorder.18(5232) The cardinal symptom is
effortless, painless regurgitation, which usually begins between birth and 3 months of age. Most infants
"outgrow" reflux by 18 months of age, although symptoms rarely persist into early childhood. GER in
infancy does not predict GER disease in later childhood or adulthood.
Functional GER may become pathogenic at any time during its course. The development of
complications is not necessarily related to the frequency or volume of emesis or to the duration of
symptoms. Signs of pathogenic GER include failure to thrive, recurrent pneumonia, and Sandifer's
syndrome. The latter refers to a movement disorder with unusual head cocking movements that
characteristically involve extension of the neck and rotation of the head, usually during or immediately
after meals. Symptoms of pathogenic GER include excessive irritability, feeding difficulty, sleep
disturbance, hematemesis, bronchospasm, and chronic cough, especially nocturnal cough. Evidence
of esophagitis has been reported in 60 to 80% of infants with signs and symptoms of pathogenic
GER.1923(5233)
The classic gross criteria for the diagnosis of esophagitis (e.g., mucosal erosions, exudate, ulcers,
stricture) are rare in infancy. In contrast, esophagitis in infancy is based on histologic criteria for
biopsies obtained during endoscopy. The sine qua non of histologic esophagitis in chi1dren is
intraepithelial eosinophils.23,24(5234) The small forceps biopsy samples obtained through pediatric
endoscopes are most often inadequate to evaluate hyperplastic epithelial changes, including basal
hyperplasia and papillary elongation, that are described in adult patients with esophagitis.25(5235)
Although there are no accepted criteria for diagnosis, four to six intraepithelial eosinophils per
high-power field is generally accepted as more than sufficient for the histologic diagnosis of reflux
esophagitis. If the number of intraepithelial eosinophils exceeds 20 per high-power field, an allergic
cause for the esophagitis should be suspected and a change to an elemental formula considered,
especially if the response to antisecretory therapy is poor.26(5236) Intraepithelial neutrophils are
regarded as less specific than eosinophils, but in practice they are usually given equal weight to
eosinophils in the diagnosis of esophagitis.27(5237) Some pathologists further expand the criteria for
the histologic diagnosis of esophagitis by attaching importance to the presence of intraepithelial
lymphocytes or squiggle cells, telangiectasia, and basal nuclear changes (i.e., spongiosis).24(5238)
Upper gastrointestinal endoscopy is recommended for the diagnosis of esophagitis in infancy. Although
blind suction biopsy of the distal esophagus has been reported to be a safe and cost-effective
alternative, the necessary equipment (i.e., Rubin tube) is no longer manufactured or generally
available.27(5239) Endoscopy with biopsy is indicated for all infants with clinical emesis and a history
of hematemesis, guaiac-positive stools, anemia, deceleration of weight gain velocity, or unremitting
chronic lung disease. More controversial is the diagnostic approach to the otherwise well infant with
clinical GER and excessive irritability and sleep disturbance. Given the rarity of gross findings and the
absence of reports of late stricture formation in this group of children, it seems reasonable to treat for 4
to 6 weeks with positional, dietary, and antacid or antisecretory therapy. Endoscopy is reserved for
patients who fail to respond and those in whom symptoms promptly recur after pharmacologic therapy.
Repeat endoscopy is rarely indicated in a patient with histologic esophagitis unless symptoms persist
to the point that the relevance of the diagnosis is questioned or an escalation in therapy, including
antireflux surgery, is being considered.
Moderate to severe neurologic impairment is a special circumstance for children with GER.18(5240)
Current speculation holds that onset in these patients occurs later in childhood and that GER is
acquired rather than congenital or developmental. Hiatus hernia, otherwise rare in pediatrics, occurs in
this patient population and is perhaps related to the abnormal intraabdominal pressure dynamics
associated with spasticity, abnormal posturing, scoliosis, and seizures. Most of these children have
pathogenic GER. They are more prone to develop the classic gross features of esophagitis. Because
of the high incidence of oropharyngeal incoordination and the child's inability to describe the symptoms
of odynophagia and heartburn, endoscopy should be performed when such patients present with
new-onset anorexia, excessive irritability, or sleep disturbance.
Dysphagia, Odynophagia, Persistent Refusal to Eat, or Chest Pain

Esophageal dysphagia may result from structural or propulsive abnormalities of the esophagus. The
identification of esophageal dysphagia depends on recognition of certain features in the patient's
history: slow feeding, sensation of "food getting stuck," heartburn, odynophagia, oral regurgitation,
chest pain, vomiting of undigested food, and chronic unexplained reactive airway disease or recurrent
pneumonia. In an infant or young child, heartburn may present as excessive irritability, anorexia, slow
feeding, and sleep disturbance. Odynophagia may present with irritability or arching during meals or an
unexplained refusal to eat. The differential diagnoses for esophageal dysphagia are listed in Table
533.
Differential Diagnosis of
Esophageal Dysphagia
TABLE 533
INTRINSIC MECHANICAL DEFECTS

Foreign body
Esophageal stricture
Peptic esophagitis
Caustic ingestions
Postsurgical anastomosis
Tracheobronchial remnants
Esophageal web
Lower esophageal rings
Muscular ring
Mucosal (Schatzki) ring
EXTRINSIC MECHANICAL DEFECTS
Vascular anomalies (e.g., aberrant right subclavian artery,
double aortic arch, right aortic arch with ligamentum)
Mediastinal abnormalities
ESOPHAGEAL MOTOR DISORDERS
Achalasia
Diffuse esophageal spasm
Scleroderma
Esophagitis
Reflux esophagitis
Infectious esophagitis
Pill-induced esophagitis
Dysautonomia
S/P tracheoesophageal fistula or esophageal atresia
All patients with suspected esophageal dysphagia should first undergo barium contrast radiography for
detection of an esophageal web, stricture, vascular ring, extrinsic compression, and achalasia. In
patients with significant dysphagia, endoscopy is indicated even when the result of barium contrast
radiography is normal. If residual contents are discovered in the esophagus, assuming the patient has
fasted as instructed, a motility disorder should be suspected. Endoscopy facilitates the diagnosis of
gross peptic esophagitis, infectious esophagitis, Barrett's esophagus, and pill-induced esophagitis.
Esophagoscopy also facilitates sizing and dilation of webs, rings, and strictures. In patients with
achalasia, endoscopy before pneumatic dilation eliminates or confirms the presence of significant
stasis esophagitis and provides an assessment of the ease of passage through the esophagogastric
junction, thereby confirming characteristic x-ray and manometric findings.
Gross peptic esophagitis in its mildest form is characterized by mucosal congestion and edema, which
obscure the normal fine vascular pattern of the distal esophagus, or red streaks running along the
crests of longitudinal folds in the distal esophagus. In places, these folds may be superficially
ulcerated. Esophagitis of moderate severity manifests with irregular tonguelike or flamelike areas of
shallow ulceration overlain by yellow slough and surrounded by an erythematous zone. Contact with
these areas causes bleeding. Severe esophagitis is characterized by deep irregular ulcers and edema
of the surrounding mucosa; linear or patchy superficial ulceration is more common than single
esophageal ulcers.
Barrett's esophagus in patients younger than 4 to 6 years of age is rare.28(5241) In older children,
Barrett's esophagus is associated with hiatus hernia and peptic stricture and more frequently occurs in
neurologically impaired children with chronic erosive esophagitis. This diagnosis should be suspected if
pink mucosa is present more than 2 cm proximal to the lower esophageal junction, but the diagnosis is
definitive only if multiple biopsies from this mucosa demonstrate the presence of specialized
metaplastic epithelium (i.e., with goblet cells stained by alcian blue, pH 2.5).28,29(5242) High-grade
dysplasia and adenocarcinoma have been described in pediatric patients.30(5243)
For the patient with immunodeficiency who develops vomiting, food refusal, dysphagia, or
odynophagia, including patients undergoing cancer chemotherapy, endoscopy should be considered
early in the course. Even in the absence of oral lesions, severe esophagitis secondary to infection with
Candida, cytomegalovirus (CMV), herpes simplex virus (HSV), or other opportunistic organisms may
be found.31(5244) Factors that are well known as predisposing in the development of esophageal
candidiasis include the acquired immunodeficiency syndrome (AIDS), malignancy (especially
lymphoma and leukemia), organ transplantation, prolonged antibiotic and corticosteroid therapy,
immunosuppressive therapy, and malnutrition.
Although Candida-induced esophagitis is commonly suggested by the concomitant presence of oral or
pharyngeal thrush, extensive esophageal candidiasis may exist in the absence of oral-pharyngeal
lesions.32(5245) Endoscopy with brushings and biopsy is the standard for diagnosis of Candida
esophagitis. Raised, white candidal plaques in the esophagus cannot be washed away; when they are
brushed, there is usually bleeding at the site of attachment. An inflammatory stricture with no hint of a
fungal cause may prove to have invasive fungal elements on biopsy.31(5246) Histopathologically,
invasive disease implies hyphal penetration into intact mucosa as opposed to colonization, which
implies the presence of yeast on an intact mucosal surface or necrotic tissue. Endoscopic brushings
and biopsies should be obtained, even when findings are typical for Candida-induced esophagitis,
because they may also yield presence of Aspergillus.
CMV infection of the esophagus is typically most prominent in the middle to distal esophagus, with a
characteristic early appearance of superficial erosions with geographic, serpiginous, nonraised
borders.31,33(5247) Because CMV-infected fibroblasts and endothelial cells are found in the base of
esophageal ulcers and never in squamous epithelium, multiple biopsies should be taken from the
center of the ulcer crater. Columnar epithelial cells of the stomach and duodenum do support CMV
infection, and mucosal biopsies should be obtained when enteric CMV disease is suspected.
Characteristic histologic features of CMV infection in esophageal biopsy specimens include large cells
in the subepithelial layer with amphophilic intranuclear inclusions, halo surrounding the nucleus, and
multiple, small cytoplasmic inclusions. A biopsy specimen from an ulcer base should also be obtained
for viral culture.
The endoscopic appearance of HSV esophagitis depends on the duration of the infection.31(5248)
Rounded, 1- to 3-mm vesicles are the earliest lesions, occurring predominantly in the middle to distal
esophagus. Characteristic sharply demarcated ulcers with raised margins and yellow-gray bases are
formed as vesicles slough. These punched-out ulcers may have surrounding erythema, but uninvolved
mucosa appears normal. Confluent ulcers may mimic Candida infection. Denudement of the
esophagus may occur in severe HSV esophagitis. To confirm the diagnosis of HSV esophagitis,
biopsies must be obtained from the margins of ulcers, and the ulcers should be brushed. Biopsies from
the ulcer base are usually inadequate for the diagnosis of HSV. Immunohistochemical stains of the
debris in a denuded esophagus may be useful in identifying sloughed HSV infected cells. Viral culture
is more sensitive than microscopic examination for the diagnosis of HSV infection. HSV esophagitis
should also be suspected in an immunocompetent patient with a history of cold sores who presents
with nasolabial herpetic lesions concurrent with the onset of dysphagia, nausea, vomiting, or
hematemesis.34(5249)
Several medications, including tetracycline, doxycycline, clindamycin, aspirin, slow-release potassium
preparations, and extended-release theophylline preparations, can cause erosion and ulceration when
lodged in the esophagus.35(5250) Endoscopy is only indicated when clinical symptoms of dysphagia or
odynophagia do not improve within several days after stopping use of the offending agent.
Unexplained, Persistent Vomiting

It is always important in pediatrics to distinguish between vomiting and regurgitation. Vomiting implies
the forceful expulsion of stomach contents by powerful contraction of the abdominal muscles.
Regurgitation, in contrast, is an effortless ejection of stomach contents. Vomiting is a nonspecific
symptom that rarely occurs alone. Intraabdominal causes of vomiting are usually associated with other
gastrointestinal symptoms, including dysphagia, abdominal pain associated with symptoms of
dyspepsia, gastrointestinal bleeding, altered bowel pattern, or jaundice. In the absence of these
symptoms, endoscopy has no role in the evaluation of acute, episodic, or cyclical vomiting until
gastrointestinal anatomic, infectious, metabolic, neurologic, toxic, cardiac, and psychogenic causes
and pregnancy are eliminated.36(5251) Endoscopy should never be performed before barium contrast
radiography to rule out upper gastrointestinal anatomic disorders.
Rarely, vomiting may be the predominant symptom of peptic ulcer, food allergy, eosinophilic
gastroenteritis, infectious esophagitis, GER, or an eating disorder. In difficult cases, the endoscopist
may be called on to evaluate a patient with persistent vomiting but without associated gastrointestinal
symptoms for gross and histologic abnormalities of the upper gastrointestinal tract. The major concern
in such cases is that finding a grossly normal mucosa but nonspecific histologic changes on biopsy
may direct management toward the gastrointestinal tract, thereby delaying establishment of a correct
diagnosis. A gastrointestinal diagnosis that does not adequately explain the patient's chief complaint
should always be regarded with skepticism.
Dyspepsia or Suspected Acid Peptic Disease

Acute Upper Abdominal Pain
Symptoms of fever, vomiting, diarrhea, or upper respiratory illness preceding or occurring
simultaneously with the onset of diffuse or epigastric abdominal pain is characteristic of viral
gastroenteritis. In cases associated with severe pain (e.g., excessive irritability, anorexia, sleep
disturbance in an infant), persistent emesis, or weight loss, the initial evaluation should be directed
toward acute medical disorders (e.g., pneumonia, pyelonephritis, biliary tract disease, pancreatitis,
hepatitis, Henoch-Schnlein purpura, hemolytic-uremic syndrome, diabetic ketoacidosis, primary
peritonitis) and surgical disorders (e.g., appendicitis, intussusception, incarcerated hernia, torsion of
testes).37(5252) Acute gastritis, duodenitis, or peptic ulcer is suggested by the constellation of
epigastric or periumbilical abdominal pain and vomiting, usually associated with eating. In the context
of symptoms of an acute viral syndrome, acute stress, or ingestion of ulcerogenic drugs, a diagnostic
EGD is rarely indicated. For most patients, endoscopy can be deferred for at least 2 weeks, pending a
trial of antacid or antisecretory drug therapy. The endoscopic differential diagnosis is the same as that
for chronic dyspepsia, discussed in the next section. The only exception to the conservative approach
is the immunocompromised patient because of the frequency with which opportunistic infection is
documented by endoscopy.
Chronic Dyspepsia
There is no consensus of expert opinion regarding the indications for endoscopy in the evaluation of
recurrent abdominal pain in children because of a lack of controlled prospective data. Diagnostic
accuracy has value and relevance only if it is consistently associated with improvement in patient
outcome.
Recurrent abdominal pain (RAP) is a common descriptor that has evolved from the seminal definition
by Apley of intermittent abdominal pain in children between the ages of 4 and 16 years that persists for
longer than 3 months and affects normal activity.38(5253) RAP has been reported to occur in 10 to
15% of children between the ages of 4 and 14 years. Children with RAP tend to exhibit one of three
clinical presentations: isolated paroxysmal abdominal pain, abdominal pain associated with symptoms
of dyspepsia, or abdominal pain associated with altered bowel pattern.39(5254) Most patients in all
three categories of presentation have functional pain, which is defined as an absence of a specific
anatomic, infectious, inflammatory, or biochemical cause. Criteria for the diagnosis of functional
disease include a characteristic history; normal findings for the physical examination, including rectal
examination (except for abdominal pressure tenderness); and negative routine diagnostic study
results, including those for the complete blood cell count, erythrocyte sedimentation rate, and stool
tests for ova and parasites. Upper gastrointestinal endoscopy has no role in the diagnostic evaluation
of the pediatric patient with isolated paroxysmal abdominal pain or chronic abdominal pain associated
with altered bowel pattern who fulfills the diagnostic criteria for functional pain.
An upper gastrointestinal x-ray series with small bowel follow-through is indicated in any patient with
altered bowel pattern undergoing evaluation for possible Crohn's disease. Although studies have
suggested that 10 to 40% of pediatric patients with Crohn's disease may have involvement of the
esophagus, stomach, or duodenum, routine EGD has no place in the evaluation of a patient in whom a
diagnosis has been established by colonoscopy.40,41(5255) EGD may be indicated in the patient with
Crohn's disease in whom chronic anorexia, poor weight gain, or symptoms of dyspepsia persist despite
a normal air contrast upper gastrointestinal x-ray series and apparent disease remission by objective
physical and biochemical parameters.
Symptoms of dyspepsia in association with chronic abdominal pain suggest upper gastrointestinal tract
dysfunction. Abdominal pain is often localized to the epigastrium, right or left upper quadrants. In
younger children, however, pain is more likely to be periumbilical. A history of occasional vomiting is
relatively common, but vomiting as a rule is not a major component of the clinical presentation.
Nausea, anorexia, heartburn, oral regurgitation, early satiety, postprandial abdominal bloating or
distention, increased hiccups, and increased belching are common associated problems. The
differential diagnoses for RAP associated with symptoms of dyspepsia are listed in Table 534. In
terms of sensitivity and specificity, there are no signs or symptoms which reliably differentiate
functional dyspepsia from upper gastrointestinal inflammatory, structural, or motility disorders.
Differential Diagnosis of Recurrent

Abdominal Pain Associated With Symptoms of
Dyspepsia
TABLE 534
ASSOCIATED WITH UPPER GASTROINTESTINAL INFLAMMATION

Gastroesophageal reflux disease
Peptic ulcer
Helicobacter pylori gastritis
Nonsteroidal anti-inflammatory druginduced ulcer
Crohn's disease
Eosinophilic gastroenteritis
Mntrier's disease
Cytomegalovirus gastritis
Parasitic infection (e.g., Giardia, Blastocystis hominis)
Varioliform gastritis
Lymphocytic gastritis or celiac disease
Henoch-Schnlein purpura
MOTILITY DISORDERS
Differential Diagnosis of Recurrent

Abdominal Pain Associated With Symptoms of
Dyspepsia
TABLE 534
Idiopathic gastroparesis
Biliary dyskinesia
Intestinal pseudoobstruction
OTHER DISORDERS
Chronic pancreatitis
Chronic hepatitis
Chronic cholecystitis
Ureteropelvic junction obstruction
Abdominal migraine
Partial small bowel obstruction
Psychiatric disorders
Criteria for the diagnosis of functional dyspepsia in children include a characteristic history; normal
physical examination, except for abdominal pressure tenderness; a normal laboratory evaluation that
includes a complete blood cell count, erythrocyte sedimentation rate, chemistry profile, and stool tests
for ova and parasites; a normal barium contrast upper gastrointestinal x-ray series and small bowel
series if vomiting is a significant part of the history; and EGD that reveals no specific gross or histologic
evidence of an upper gastrointestinal inflammatory process.39(5256) Upper gastrointestinal endoscopy
is the best and most specific means for identification of inflammation in the upper gastrointestinal tract,
including an assessment of severity and the recognition of associated complications.
Objective findings by gross inspection include superficial erosions, ulcer, stricture, antral nodularity
associated with Helicobacter pylori gastritis,42(5257) gastric rugal hypertrophy associated with CMV
gastritis and Mntrier's disease,43,44(5258) superficial ulceration or intramural hemorrhage
associated with Henoch-Schnlein purpura,45,46(5259) scalloped duodenal folds associated with
celiac disease,47(5260) and the small heaped-up, volcano-like mounds with central craters associated
with chronic varioliform gastritis.48(5261) Less specific gross findings, including erythema, edema,
increased vascularity, and friability, become meaningful only in the context of histopathologic findings.
Crohn's disease of the upper gastrointestinal tract has no pathognomonic gross features. With
esophageal involvement, the mucosa in the distal esophagus may appear lumpy or irregular.
Superficial ulceration, linear ulcers, or polypoid mucosa may be present with gastric Crohn's disease.
Endoscopic findings of Crohn's disease in the duodenum include aphthoid ulcers, linear ulcers, and
stenosis.
Recognizable objective histologic findings include the following:
1. Intraepithelial eosinophils or polymorphonuclear cells, basal hyperplasia, or papillary elongation on
esophageal biopsy, indicating reflux esophagitis23(5262)
2. Intraepithelial eosinophils in antral or duodenal biopsies, indicating possible eosinophilic
gastroenteritis49(5263)
3. Intraepithelial plasma cells, polymorphonuclear leukocytes, microerosions, crypt abscesses, and
pyloric gland metaplasia with or without noncaseating granuloma, suggesting Crohn's
disease50(5264)
4. Characteristic CMV intranuclear inclusions
5. Lymphocytic infiltrate within surface and pit gastric epithelium associated with spiral bacterium
indicative of H. pylori gastritis51(5265)
6. Dense lymphocytic infiltrate within the surface and pit gastric epithelium associated with a
uniformly flat duodenal mucosa, indicating celiac disease52(5266)
In the absence of gross ulceration, H. pylori infection, or celiac disease, the diagnostic value of
identifying superficial gastritis, as characterized by inflammatory infiltration of the upper lamina propria,
is less clear, especially in the context of a gastric mucosa that appears grossly normal at endoscopy.
Chronic superficial antral gastritis is extremely common in asymptomatic adults and in patients with
adult irritable bowel syndrome. Controversy also exists about whether chronic epithelial or interstitial
inflammation of the duodenum (i.e., duodenitis) without discrete ulceration or evidence of H. pylori
infection causes symptoms of dyspepsia. For adults or pediatric patients, no evidence indicates that
nonspecific antral gastritis or duodenitis confers an increased risk for peptic ulcer disease. Until more
outcome data are available, the documentation of mild, nonspecific superficial antral or duodenal
inflammation does not preclude a diagnosis of functional dyspepsia.
In adults, abdominal pain associated with dyspepsia is often treated empirically with antacids or
antisecretory drugs, with the assumption that a favorable clinical response reflects healing of minor
upper gastrointestinal inflammation.53(5267) Endoscopy is performed only if response to treatment is
poor or if symptoms recur. This approach is also reasonable in children older than 10 years of age in
whom symptoms have persisted for less than 3 months, are triggered by an apparent viral infection,
and do not affect function, particularly school attendance. However, if symptoms recur, exist for longer
than 3 months, or result in altered function (e.g., weight loss, school absence, altered activity), I feel
that the diagnostic information gained from upper gastrointestinal endoscopy improves the specificity of
diagnosis and compliance with treatment. Although more costly at the outset, a positive diagnosis,
backed by endoscopy, allows the primary physician to resume management of the patient, reduces
medication costs, and reduces repetition of or escalation of testing.
Unexplained Iron Deficiency Anemia

Blood loss must be considered a possible cause of iron deficiency anemia in every case, particularly
the older child. If occult blood is detected in the stools, the major differential includes peptic ulcer,
Meckel's diverticulum, polyp, hemangioma, parasitic infection in specific geographic locations, and
inflammatory bowel disease. Gastrointestinal occult bleeding and hypochromic anemia have also been
reported as the presenting signs of gastroduodenitis associated with cow's milk protein
intolerance.54(5268) In the absence of symptoms of dyspepsia, the initial investigation should be
directed toward lower gastrointestinal causes.
Indications that Affect Patient Management

Sampling Duodenal or Jejunal Tissue or Fluid
Tissue or fluid samples from the duodenum are often needed for the diagnostic evaluation of patients
with chronic diarrhea or malabsorption. EGD allows collection of pancreaticobiliary secretions after
provocative stimulation with secretin or cholecystokinin octapeptide for the diagnosis of pancreatic
insufficiency. Duodenal aspirates and brushings may also aid in the diagnosis of giardiasis,55(5269)
Cryptosporidium infection,56(5270) and bacterial overgrowth. Endoscopic biopsies may confirm or
exclude a variety of enteropathies, including celiac disease,57(5271) lymphangiectasia,58,59(5272)
Crohn's disease,60(5273) eosinophilic gastroenteritis,49(5274) and sucrase-isomaltase
deficiency.61(5275)
Despite the ease of obtaining duodenal tissue samples by EGD, pathologists often express concern
regarding sample size, orientation, and crush artifact. In children of large enough size to accommodate
passage of an endoscope with a 2.8-mm diameter accessory channel, a Crosby-Kugler or Watson
capsule placed by means of endoscopy may improve the biopsy size and ease of orientation.62(5276)
Granot et al.63(5277) compared biopsies obtained at endoscopy with biopsies obtained with a suction
capsule for the adequacy for histopathologic assessment and found the specimens to be comparable.
Early Diagnosis of Esophageal Varices
Endoscopy may be performed to assess risk factors for bleeding in the asymptomatic child with portal
hypertension if medical treatment of the portal hypertension is considered.16(5278) Endoscopic signs
used to predict risk of variceal hemorrhage include variceal size, the presence of cherry red spots, the
blue color of varices, and the red markings over the varices. It is common to find punctuate erythema
along the crests of gastric rugae. This is of no particular significance, although it may be a sign of
portal gastropathy in patients with portal hypertension.17(5279)
Suspected Ingestion of Caustic Substances

There are approximately 5000 reports each year in the United States of accidental caustic ingestion by
children, although it is estimated that only 10% of all ingestions are actually reported.64(5280) Most
accidental ingestions occur in children younger than the age of 5 years. The extent and severity of
injury to the gastrointestinal tract depends on the type of agent, concentration, pH, physical state, and
quantity ingested. Solid agents in the form of granules, pellets, pastes, and powders cause immediate
pain and usually are quickly expelled. If swallowed, they adhere to the oropharynx and proximal
esophagus, causing focal injury. In contrast, alkaline and acid solutions are often odorless and
tasteless and do not cause immediate discomfort, even after several swallows.
The concentration of caustic agents in household products is limited to 10% by law in the United States
(i.e., Poison Prevention Packaging Act, 1970). This has reduced the incidence of acute full-thickness
burns.65(5281) However, a 10% alkaline solution is still capable of causing liquefaction necrosis
extending to the outer muscle layer of the gastrointestinal tract. Data indicate that the pH of an alkaline
solution is also an important factor in the degree of esophageal ulceration. A pH of 12.5 or greater,
regardless of caustic concentration, may result in esophageal ulceration.64(5282)
It is important to understand the natural history of an alkaline burn. Initially, erythema is seen at the site
of injury. Ulceration may not develop until 24 hr after ingestion. This has important implications
regarding early endoscopic assessment of the severity of the injury. Over the next several days,
bacterial invasion may extend the depth of the ulceration. Necrotic sloughing usually occurs 5 to 7 days
after ingestion. Stricture formation, resulting from subsequent fibroblast infiltration, first becomes
evident by day 21 after the ingestion; complete stricture formation requires 4 to 6 weeks.66(5283)
Acids cause injury by coagulation necrosis with eschar formation. The esophagus is spared in more
than 80% of acid ingestion cases because of the alkaline pH of the esophagus and resistance of the
squamous epithelium to acid.67(5284) The site of acute injury is usually the prepyloric area of the
stomach. Skip areas are commonly detected by endoscopy. Prepyloric stenosis may be a late
complication during healing.
Diagnostic evaluation of a child with suspected caustic injury begins with identifying and determining
the composition of the ingested substance. It is extremely helpful to have the container. Poison Control
Centers have been established in many countries, and it is important to telephone this agency to obtain
the latest information regarding concentration of the caustic substance in the ingested solution,
presence or absence of balance builders (i.e., sodium salts of carbonates, phosphates, and silicates
that increase the pH of the solution), and pH of the solution.
Predicting injury on the basis of clinical signs and symptoms is difficult. Thirty to 50% of patients with
oral burns have esophageal burns.68(5285) The absence of oral burns, however, does not exclude the
presence of an esophageal or gastric burn.69(5286) Symptoms including drooling, vomiting, abdominal
pain, or stridor should raise concern regarding possible esophageal injury. Flexible endoscopy using a
small-diameter pediatric endoscope is the only reliable way to evaluate the upper gastrointestinal tract
after caustic ingestion. Endoscopy is recommended if it is certain by history that a significant volume of
true alkali (pH > 12.5) or acid (pH < 2.0) has been ingested; for oral burns; for symptoms of drooling,
vomiting, abdominal pain, or stridor; if intentional ingestion is suspected; and if there is a concern about
the reliability of follow-up. Because many of these children are seen in an emergency room setting,
identification of the primary caregiver is not always possible. If the dependability of follow-up is an
issue, it can be argued that the major benefit of early endoscopy is to eliminate the possibility of
esophageal and gastric injury.
Endoscopy should be performed 12 to 24 hr after the ingestion.64(5287) Nothing by mouth is given
before endoscopy. If there are significant oral burns or symptoms of drooling, vomiting, abdominal
pain, or stridor, the procedure should be performed under general anesthesia. Otherwise, conscious
sedation is adequate because the damage is likely to be mild. General anesthesia allows careful
evaluation of the oral cavity and hypopharynx and makes it easier to pass the endoscope under direct
vision.
Esophageal burns are classified as first degree based on the presence of erythema, edema, or
superficial sloughing. If there is exudate overlying small erosions or shallow ulcers, the burn is
classified as second degree. Exudate overlying large ulcers, black coagulum, or circumferential
ulceration warrant classification as a third-degree burn. The presence of a third-degree burn can
usually be suspected based on knowledge of the caustic substance and the presence of symptoms. In
such cases, a pediatric surgeon must be available, because a third-degree burn necessitates
placement of a surgical gastrostomy and an esophageal stent or passage of a string for future dilation.
After a third-degree burn is identified, further advancement of the endoscope is contraindicated.
However, a guidewire may be passed under fluoroscopy into the stomach pending gastrostomy. A
third-degree burn is a contraindication for placement of a percutaneous gastrostomy.
If no esophageal or gastric lesions are seen, the child is discharged as soon as adequate oral intake is
established. Children with first-degree burns are usually observed for 24 hr and discharged after oral
intake is established. The management of second- and third-degree burns is controversial. A 5- to
7-day period of bowel rest, together with intravenous antibiotics (ampicillin, 200 mg/kg/day
administered intravenously every 6 hr) may prevent secondary infection and hasten tissue repair but
does not prevent stricture formation.66(5288)
Further endoscopic procedures should be delayed for a minimum of 3 weeks after the initial diagnosis.
A follow-up barium contrast x-ray study should be performed at that time in all patients with secondand third-degree burns to determine whether a stricture has formed and to assess esophageal motility.
In the absence of a gastrostomy, dilation of an esophageal stricture should be initiated 4 weeks after
the ingestion and be performed under endoscopic guidance using a guidewire with Savary dilators, the
Puestow over-the-wire technique with metal olives, or inflatable dilators.
Therapeutic Upper Gastrointestinal Endoscopy

When an actively bleeding lesion is encountered during diagnostic endoscopy, the endoscopist should
be prepared to initiate some form of therapy. An actively bleeding ulcer in children may benefit from
early endoscopic therapy with epinephrine injection or endoscopic heat probe coagulation.70(5289) A
proud-standing visible vessel or fresh adherent clot should alert the endoscopist to the fact that the
patient is likely to rebleed.71(5290)
The published experience with endoscopic techniques and devices for hemostasis is limited, because
ulcer bleeding that does not stop spontaneously is relatively rare in pediatrics. The aggressive use of
antisecretory drugs prophylactically in pediatric intensive care units has dramatically reduced the
incidence of severe bleeding associated with acute or stress ulcers. The heat probe is a hemostatic
device for programmable energy delivery to tissue. It is ideal for treating spurting or oozing lesions
because it allows washing with water, vessel tamponade, and thermal coagulation. Injection of a
solution of 1:10,000 epinephrine into and around an oozing bleeding ulcer or ulcer with a nonbleeding
visible vessel is also effective and technically easier to perform. Using a flexible, 25-gauge needle
injector inserted through the accessory channel, aliquots of 0.3 to 0.5 ml are injected by multiple
punctures until the bleeding stops. The total dose, scaled down from those recommended for adults, is
in the range of 0.2 ml/kg.70(5291)
The endoscopist may choose to perform sclerotherapy during active variceal bleeding. Alternatively,
sclerotherapy may be delayed until after hemorrhaging has been controlled by pharmacologic agents
such as vasopressin, somatostatin, or metoclopramide or by balloon tamponade, especially if there is
difficulty obtaining a clear endoscopic field of vision.15,16(5292) Gastroesophageal balloon tamponade
should only be employed when bleeding gastric or esophageal varices have been documented by
EGD. Acute bleeding can often be arrested by using the gastric balloon alone pulled taut at the
gastroesophageal junction (i.e., 150 ml of air in a pediatric balloon or 250 ml in an adult balloon after
ensuring correct position below the diaphragm). Endoscopic sclerotherapy should be performed on an
elective basis within 6 to 24 hr after active bleeding has been controlled by pharmacologic therapy or
balloon tamponade.
The techniques of sclerotherapy used in children are similar to those employed in adults. Prophylactic
administration of antibiotic is recommended for patients with significant hypersplenism, congenital
heart disease, and central venous catheters. Before any injection of an active bleeding varix, it is
important to establish a clear endoscopic field by washing blood into the stomach. It may be helpful to
elevate the head of the endoscopy table 30 degrees to promote esophageal clearance.
No particular sclerosant has been accepted as generally preferable for use in children. Those used
generally reflect the choice of agents in corresponding adult endoscopy units in individual hospitals.
Most North American centers use sodium tetradecyl sulfate, usually mixed with ethanol, saline, or both
agents.
Intravariceal injections are preferred, although inadvertent paravariceal injections are relatively
common. Intravariceal injection is usually performed using a 25-gauge injection needle. Circumferential
intravariceal injections in the distal 3 to 5 cm of the esophagus are made with a dose of 0.5 to 3.0 ml of
sclerosant per varix (until resistance is encountered or the varix blanches) and a cumulative dose of up
to 0.8 ml/kg per session, with a total volume not to exceed 20 ml.72,73(5293)
As in adults, a venous jet of blood or oozing of blood may be seen in children after injection and on
withdrawing the sclerotherapy needle. This indicates the varix is still patent, but such bleeding usually
stops spontaneously or responds to tamponade with the endoscope. If not, the varix is reinjected at
adjacent sites until hemostasis is achieved. Intravenous metoclopramide (0.1 mg/kg) has been
recommended in such cases to increase lower esophageal sphincter pressure and promote gastric
emptying of blood.16(5294) After the procedure, my colleagues and I routinely treat patients for 5 days
with sucralfate. Proujansky et al.74(5295) found that bleeding after sclerotherapy is more common in
patients with a platelet count of less than 100,000 cells/mm3. Outpatient sclerotherapy sessions are
continued every 2 to 4 weeks until varices are obliterated. Thereafter, surveillance endoscopy and, if
necessary, repeated sclerotherapy are carried out at intervals of 6 to 9 months.
Many series have documented that injection sclerotherapy is as successful in children as adults in
stopping bleeding from esophageal varices.16,73,7580(5296) Long-term follow-up of patients with
chronic liver disease after sclerotherapy is limited because variceal bleeding in children with portal
hypertension secondary to cirrhosis is an accepted indication for early listing for liver transplantation. In
children with extrahepatic portal hypertension or congenital hepatic fibrosis, prevention of recurrent
bleeding is the definitive treatment. The rate of recurrent bleeding after complete obliteration of
esophageal varices in children with extrahepatic portal hypertension (i.e., cavernous transformation of
the portal vein secondary to portal vein obstruction) has been reported to be low, possibly because
sclerotherapy enhances the natural tendency for development of spontaneous portosystemic shunts
over time.72(5297) In the consecutive series of Stringer and Howard81(5298) of 36 children with
variceal bleeding due to extrahepatic portal hypertension, injection sclerotherapy alone provided
effective hemostasis. Bleeding was recurrent in 31%, but further sclerotherapy was effective for one
half of cases. Four patients underwent portosystemic shunt surgery because of gastric variceal
bleeding.
The most common significant complications of injection sclerotherapy in children are esophageal
ulceration and stricture.7274(5299) Esophageal perforation is rare. Patients with more than mild chest
pain and fever should undergo an esophagogram on an emergency basis with meglumine diatrizoate
(Gastrografin) as the contrast agent to detect a leak. Mild dysphagia during the early period after
sclerosis is common, but esophageal motility is generally well preserved, and long-lasting symptoms
are rare. Stricture formation is related to the use of excessive volumes of sclerosant per session,
paravariceal injection, and the need for more than six sclerotherapy sessions to eradicate varices.
Esophageal stricture is usually easily managed by esophageal dilation.
Transient hemoglobinuria in children has been ascribed to intravascular hemolysis caused by
sclerosant within the systemic circulation, especially in sessions using volumes of sclerosant that
exceed recommendations. False submucosal tracts and respiratory complications are more common
in children who weigh less than 12 kg or when the dosage of sclerosant exceeds recommendations.
Endoscopic band ligation of esophageal varices has been reported in children, but experience is
limited.82,83(5300)
Foreign Bodies
The peak age at which foreign body ingestion occurs is between 6 months and 3 years. Mentally
retarded and psychiatrically disturbed children are at risk at older ages. The approach to each patient
must be individualized and is based on the patient's symptoms, medical history, the type of foreign
body, and its location in the upper gastrointestinal tract. Eighty to 90% of foreign bodies that reach the
stomach pass through the gastrointestinal tract spontaneously.84(5301) A guideline for objects that
require endoscopic removal is presented in Table 535.
Guidelines for Foreign Bodies

That Require Endoscopic Removal
TABLE 535
Coins that remain in the esophagus for more than 24 hr

Coins or objects with a diameter greater than 25 mm (e.g., U.S.
quarter)
Coins that fail to pass from the stomach after 4 weeks
Disk batteries in esophagus
Disk batteries remaining in the stomach for more than 48 to 72
hr
Sharp or elongated objects wider than 2 cm or longer than 5 cm
(3 cm long and 2 cm wide for infants)
Meat impaction if signs of esophageal obstruction or dysphagia
persist for more than 24 hr
Coins
Coins account for most esophageal foreign bodies. Up to half of children who swallow coins are
asymptomatic.64,8588(5302) Symptomatic patients usually present with chest pain, excessive
salivation, or frequent gagging. Chest and abdominal x-ray films should be obtained for all patients
suspected of coin ingestion. The presence of symptoms or a coin of 25 mm or greater diameter
warrants endoscopic removal as soon as possible. Asymptomatic patients who have ingested
smaller-diameter coins may be given liquids orally and observed. Persistence of the coin in the
esophagus for 24 hr necessitates removal.
As a rule, smaller coins such as the U.S. penny (18 mm), dime (17 mm), and nickel (21 mm) pass
spontaneously after they reach the stomach.64(5303) Larger coins are less likely to pass, and a case
can be made for their early removal. Symptoms usually are related to gastric outlet obstruction and
include vomiting and crampy abdominal pain. The presence of symptoms warrants endoscopic
removal of the coin as soon as possible. Such patients are rare, however, and it is generally agreed
that it is safe to leave a coin in the stomach up to 4 weeks while waiting for it to pass, provided the
patient remains asymptomatic.85(5304) An abdominal x-ray film should be obtained every 7 to 10 days
to monitor passage of the coin.
Coins should be removed with the patient under general anesthesia. Coins are best retrieved by an
alligator-type grasping forceps designed with recurved tips or jaws that easily grip the edge of the
coin.85(5305) Because this forceps requires a 2.8-mm accessory channel, it is essential to use an
endoscope with an outer diameter of 9.8 mm. A McGill or Kelly forceps must be available in case the
coin is lost in the hypopharynx during extraction.
Disk or Button Batteries
Disk or button batteries have become an important cause of caustic esophageal injury in infants and
children.64,85,89,90(5306) There are four types of disk batteries: mercuric oxide, silver oxide,
manganese oxide, and lithium. All four types contain a 20 to 45% solution of potassium hydroxide or
sodium hydroxide. Tissue injury occurs as a result of pressure necrosis, low-voltage burns, or
corrosive injury from leakage of the alkaline solution. A battery lodged in the esophagus warrants
endoscopic removal as soon as possible, because severe tissue injury can occur within 4 hr.64(5307)
In asymptomatic patients with suspected ingestion, it is essential to differentiate disk batteries from
coins. Disk batteries usually show a double-density shadow on an anteroposterior chest x-ray film
because of their bilaminar structure. Rounded edges and a step-off contour at the anode-cathode
junction are clues on the lateral film.
Endoscopic retrieval using a foreign body grasping (i.e., rat-tooth) forceps, polyp snare, or Dormia
basket is the procedure of choice. Disk batteries stuck in the esophagus should be removed with the
patient under general anesthesia. The esophageal mucosa should be inspected for tissue damage
before removal of the endoscope. In cases of suspected severe mucosal injury, a follow-up
radiographic contrast study should be performed 3 to 6 weeks after extraction to exclude stricture
formation.
More than 90% of the disk batteries that reach the stomach pass through the gastrointestinal tract
within 72 hr.89(5308) The patient should be instructed to eat a normal diet. Cathartics such as
magnesium sulfate or magnesium citrate or a prokinetic agent (e.g., metoclopramide, cisapride) may
be given to reduce transit time. Continued presence of the battery in the stomach beyond 48 to 72 hr
warrants endoscopic removal.64(5309) Passage through the small bowel or colon rarely results in a
complication. If the battery remains in the same segment of colon for more than 5 days, colonoscopic
retrieval is indicated. Surgical removal becomes necessary at any time the child develops signs of
peritoneal irritation, regardless of location.
Sharp or Elongated Foreign Bodies
Surprisingly, most small sharp objects such as nails, screws, straight pins, jewelry, and thumbtacks
pass spontaneously. If ingestion of a radiolucent object is suspected, an x-ray study using a small
amount of dilute barium may be helpful in determining the presence, size, location, and orientation of
the object. In the older child or adolescent, objects longer than 5 cm or wider than 2 cm usually warrant
endoscopic removal.85(5310) For infants and young children, objects longer than 3 cm or wider than 2
cm should be removed.64(5311) Objects such as chicken or fish bones, long straight pins, toothpicks,
or open safety pins should be removed as soon as possible. If it is decided to observe the patient, a
high-fiber diet should be started and the parents instructed to screen the stools. In cases of radiopaque
objects, plain x-ray films should be obtained at 5-day intervals until passage is confirmed.
The approach to endoscopic removal depends on the object's size, shape, and location. If possible, the
object should be duplicated and the best retrieval device determined by dry runs before the actual
procedure. Safe removal requires axial orientation of the object, with the sharp or pointed edge located
distally. When possible, an overtube should be used. In some cases, rigid esophagoscopy may be the
procedure of choice. The approach to an open safety pin depends on orientation. If stuck in the
esophagus with the open end oriented proximally and the spring distal and if there is no obstruction
distal to the pin, it is carefully pushed into the stomach, turned, then grasped by the spring and pulled
into an overtube before extraction. The overtube should be positioned with its distal end just within the
stomach. When the spring end is proximal in the esophagus, the spring end may be grasped within the
esophagus and pulled into the overtube before extraction.85(5312)
Meat Impaction
Meat impaction usually occurs in patients with esophageal motility disorders, peptic esophageal
stricture, esophageal web, or lower esophageal rings (e.g., Schatzki's ring) or in those who have a
history of prior esophageal surgery, usually tracheoesophageal fistula repair. Meat impaction usually
occurs in older children and adolescents. Urgent endoscopic removal is warranted if the patient
experiences difficulty handling secretions. If the patient is experiencing mild chest pain but is able to
handle secretions, endoscopic removal may be delayed 8 to 12 hr to see if the impacted food will pass
spontaneously. Light sedation or glucagon (0.51.0 mg intravenously) may facilitate passage.90(5313)
Disruption of a meat impaction with papain is not recommended in children.
Meat stuck in the esophagus should be removed with the patient under general anesthesia. If the
lumen distal to the impaction cannot be easily viewed, the bolus should be extracted rather than
pushed into the stomach. The rat tooth jaw forceps and polyp snare are the most suitable devices for
removal of food from the esophagus. After the impacted material has been removed, the esophagus
should be carefully inspected for the presence of associated abnormalities.
Bezoars
Treatment of phytobezoars, composed of plant and vegetable material, can be facilitated by
endoscopic disruption. Most occur in healthy girls between the ages of 10 and 19 years, patients with
gastroparesis, and rarely in association with chronic use of H2-receptor blocking agents. The onset of
symptoms is usually insidious. The clinical presentation is that of gastric outlet obstruction and includes
vomiting, anorexia, abdominal pain, and weight loss. Severe halitosis may result from the putrefying
material in the stomach. Diagnosis is usually made by upper gastrointestinal contrast radiography.
Endoscopy distinguishes between trichobezoars (i.e., casts of hair) and phytobezoars if the nature of
the bezoar is uncertain. Surgical removal is usually required for trichobezoars. Phytobezoars may be
disrupted with various endoscopic devices. Further dissolution has been described using papain,
acetylcysteine, or cellulase. Gastric emptying is then facilitated by lavage with clear liquids plus
administration of a prokinetic agent.
Lactobezoars, which may occur in premature infants, do not require disruption. They are treated by
withholding feeding for 48 hr and maintaining hydration with intravenous fluids.
Dilation of Esophageal Strictures, Webs, and Rings

Esophageal strictures in children may follow reflux esophagitis, caustic ingestion, surgical
anastomosis, chemotherapy, radiation therapy, epidermolysis bullosa, and rarely, infectious
esophagitis. Peptic esophagitis, which produces a marked inflammatory reaction and fibrosis, may
produce a stricture that usually extends over several centimeters in the esophageal wall, forming a
tubular stricture. No data indicate that histologic esophagitis of infancy predisposes to stricture
formation. In addition to dilation, fundoplication is considered an integral part of therapy for peptic
strictures complicating GER in pediatric patients.91(5314)
It is important to estimate the lumen diameter and length of an esophageal stricture. An estimate of the
diameter can be obtained by opening a biopsy forceps near the proximal end of the stricture. Dilation
may be accomplished in pediatric patients by blind or fluoroscopically guided passage of soft,
mercury-weighted, tapered, rubber bougies; by wire-guided passage of tapered, polyvinyl bougies; and
by hydrostatic balloons passed over a guidewire or through the accessory channel of an
endoscope.92,93(5315) Mercury-weighted bougies can be used for straight, smooth, concentric
strictures, and wire-guided methods are best for long, tight, or tortuous strictures. Fluoroscopic
monitoring is indicated during dilation of stricture with one or more of the following characteristics: long
strictures; those that involve the distal end of the esophagus; angulated or tortuous strictures; those
with an associated ulcer or diverticula proximal, within, or distal to the stricture; and strictures
associated with hiatus hernia. Wire-guided dilation is best performed with the patient under general
anesthesia. Dilation with mercury bougies may be accomplished in awake or sedated
children.94(5316)
The first dilation is usually performed with a bougie or hydrostatic balloon with a diameter equal to the
estimated diameter of the stricture. The size of the dilator is increased in increments of 2 French. After
mild resistance is encountered, no more than two dilators of progressively increasing diameter (2
French) should be passed in a single session. Dilation is considered adequate in infants when the
esophageal lumen can be dilated to a diameter of 18 to 25 French (6 to 8 mm).95(5317) For older
children, a diameter of 36 to 46 French (12 to 15 mm) is considered adequate.95(5318) The use of
intralesional injections of corticosteroids in the treatment of refractory esophageal strictures has been
reported as successful in pediatric patients but remains controversial.96(5319)
The principal complications of dilation include perforation, bleeding, and pulmonary aspiration. Mild
bleeding is invariable but seldom of consequence. Persistent chest pain or fever after dilation warrants
emergency contrast radiography with water-soluble contrast to look for evidence of a perforation. In the
absence of x-ray evidence of perforation, blood cultures should be obtained, and the patient should be
observed while given nothing by mouth for the next 12 hr. Spontaneous resolution of fever during this
time is an indication to cautiously resume feeding.
Congenital esophageal stenosis and lower esophageal rings are rare in pediatric patients. Three types
of congenital esophageal stenosis have been described: membranous webs, fibromuscular stenosis,
and tracheobronchial remnants. Membranous webs are treated by bougienage.97(5320)
Radiographically, muscular stenoses have smooth contours and are located in the distal esophagus
proximal to the gastroesophageal junction. They usually appear as contractile rings and may disappear
completely on maximum distention of the esophagus during barium contrast studies. In contrast,
tracheobronchial remnants are usually irregular and located in the mid-esophagus.98(5321) Dilation of
a stenosis associated with tracheobronchial remnants carries a high risk of perforation. Surgery is
usually required for tracheobronchial remnants. Fibromuscular stenosis, characterized histologically by
submucosal proliferation of smooth muscle, does respond to serial dilation.97,99(5322)
A lower esophageal mucosal ring, frequently referred to as a Schatzki ring, is a thin annular membrane
of mucosa that occurs at the gastroesophageal junction and projects into the lumen perpendicular to
the esophageal wall.97(5323) The pathogenesis is unknown. The rings have been associated with
hiatus hernia and peptic esophagitis in older children and adolescents.100(5324) Complications are
purely mechanical, related to the size of the ring. Symptoms occur when a food bolus lodges at the
ring. The ring may be easily stretched by transesophageal dilation and then ruptured by passage of a
46-French bougie.
Esophageal dilation for achalasia is unique in that it involves forceful disruption of the lower
esophageal sphincter.101(5325) Before dilation, it is critical to confirm the diagnosis by standard
radiographic, manometric, and endoscopic methods of evaluation. The treatment of achalasia is strictly
palliative. Sustained relief of symptoms in children has been well-described after pneumatic dilation.
The choice of dilators for achalasia has become limited because of manufacturing changes. The most
widely available hydrostatic dilator has a special inelastic balloon that, when filled with a radiopaque
fluid (1 part Hypaque, 4 parts water), exerts a radial force.102(5326) Pneumatic dilation is best
performed using deep conscious sedation and topical anesthesia in the hypopharynx to suppress the
gag reflex. Endoscopy is always performed before dilation to completely exclude obstruction and gross
Candida infection or peptic esophagitis in the distal esophagus. Mild stasis esophagitis is invariably
present. Biopsies should never be obtained before dilation.
The dilation balloon is placed in the sphincter zone using fluoroscopic guidance with half the bag above
and half below the diaphragm. This position is monitored fluoroscopically as the bag is gradually
inflated up to a pressure of 7 to 8 pounds per square inch (1 psi = 51.7 mm Hg) and maintained in
position for 30 to 45 sec. Expertise is required to hold the bag in position because of its tendency to be
pulled into the stomach (i.e., path of least resistance) as the pressure is increased. The fluoroscopic
appearance of the bag and the presence of pain should also guide the extent of dilation and not the
pressure readings alone. At maximal pressure, the bag should have an hourglass shape, as observed
fluoroscopically, with the indentation caused by the lower esophageal sphincter. No further pressure
should be exerted if one side of the hourglass straightens. At maximal pressure, the patient usually
exhibits evidence of substernal discomfort. This sign, together with small amounts of blood streaking
on the balloon after withdrawal, indicate an adequate dilation.
After hydrostatic dilation, contrast radiography using barium or water-soluble contrast is routinely
obtained to exclude esophageal perforation. Food and fluids are withheld for 8 to 12 hr, and vital signs
are monitored frequently. When feedings are resumed, the patient can tell immediately whether the
procedure was successful. The usual cautions related to esophageal dilation apply to forceful dilation
of achalasia. As in adults, a second attempt at dilation is warranted in children before considering
surgical myotomy if a single session does not produce satisfactory relief of dysphagia. No published
reports of intrasphincteric injection of botulinum toxin in pediatric patients are available.
Percutaneous Endoscopic Gastrostomy

Percutaneous endoscopic gastrostomy (PEG) is based on the simple principle of sutureless
approximation of the stomach to the peritoneum by a catheter.103,104(5327) Continuous apposition
leads to the development of firm adhesions between the stomach and abdominal wall around the tract.
Among the 220 children reported by Gauderer,103(5328) ages ranged from 1 month to 21 years.
Forty-two patients (19%) weighed less than 5 kg, including three infants weighing only 2.5 kg. The
major indication for PEG in pediatric patients is swallowing or feeding difficulty resulting in grossly
inadequate oral intake or recurrent oral aspiration. Categories of patients include those with severe
birth injury, cerebral palsy, myopathies, complex oropharyngeal and laryngotracheal malformations,
chromosomal and metabolic abnormalities, congenital heart disease, bronchopulmonary dysplasia,
esophageal dysmotility, malignancy, and AIDS. Although some patients exhibit evidence of significant
swallowing dysfunction from birth or have had a limited oral feeding experience because of critical
neonatal illness or chronic gastrointestinal disease, most who are referred for evaluation have failed
attempts at oral feeding over an extended period at home or in the hospital.
Patients should be considered for PEG if the time to substitute complete oral feedings is expected to
exceed 3 to 4 months; supplemental nasogastric feedings are having a negative influence on oral
feedings; oral feedings are consistently difficult, such that meal duration always exceeds 40 to 60
minutes or the caregiver's whole day is consumed in trying to feed the child; or symptoms of aspiration
during feeding or episodes of recurrent pneumonia are frequent.105(5329) PEG may also be indicated
for patients with complex intestinal disorders (e.g., short-gut syndrome, malabsorption, Crohn's
disease, intestinal pseudoobstruction) to deliver enteral nutrients by continuous intragastric infusion.
Rarely, patients may require PEG to deliver unpalatable medications or diets.
Many pediatric surgeons and gastroenterologists are hesitant to perform PEG alone in patients with
severe neurologic impairment who need supplemental nutrition as part of a long-term plan of nutritional
support because of the reported common occurrence of late-onset pathogenic GER. In the past, many
surgeons recommended that a protective antireflux procedure be performed in any patient with severe
neurologic impairment requiring gastrostomy. It is only recently, however, that the significant morbidity
(20 to 50%, with up to 20% requiring reoperation for recurrence of GER) of antireflux surgery in this
patient population has been recognized.106(5330) Even patients with successful surgical outcomes
have a high mortality rate within 1 year of surgery resulting from complications of their underlying
disorders, independent of reflux.107(5331) The experience of my group supports the observation of
Langer et al.108(5332) that standard diagnostic tests for GER, including barium contrast radiography,
gastroesophageal scintography, and prolonged intraesophageal pH monitoring do not predict which
patients will tolerate gastrostomy feeding. The preliminary observations suggest that the ability to gain
weight without clinical exacerbation of GER while undergoing supplemental nasogastric feedings for a
defined period (at least 4 weeks) before PEG predicts a successful outcome for subsequent long-term
gastrostomy feedings. Preoperative pH monitoring and endoscopic esophageal biopsies obtained at
the time of PEG are used to determine whether medical management of GER is required after the
PEG procedure. My colleagues and I have found that most patients require long-term use of prokinetic
or antisecretory medications to successfully control reflux.
We routinely obtain upper gastrointestinal contrast radiography preoperatively to assess stomach size
and position and to check for malrotation. The procedure is performed in the operating room with the
patient under general anesthesia. A single dose of a broad-spectrum antibiotic is given intravenously
shortly before the procedure is begun. PEG is a therapeutic procedure and should never be performed
after a diagnostic endoscopy, because the air-filled small bowel will push the colon and stomach in a
cephalad direction and thereby increase the risk of colonic puncture or tangential puncture of the
stomach. An endoscope with a diameter of 7.8 mm or 9.8 mm is used. The gastrostomy tube itself
should have a tapered tip with a loop of fine, flexible stainless steel wire, a nondetachable gastric
retainer, and markings on the shaft to allow the endoscopist to determine the length of the remaining
tube segment during the pulling process.104(5333) Two operators are required.
The prospective best-possible target site for the gastrostomy is marked at the junction of the lower two
thirds and upper third on a line between the umbilicus and the midportion of the left inferior costal
margin. This is primarily to remind the endoscopist that the midline and rib cage are not good sites for
PEG location. The abdominal wall in the midline is thinner, and this may predispose to leaks.
Gastrostomies too close to the rib also have a tendency to leak because of increased catheter motion
during breathing. Proximity to the rib may also lead to chondritis.
The endoscope should be directed in the stomach toward the anterior wall. Gastric insufflation
mobilizes the liver, spleen, and colon away from the gastrostomy site and brings the anterior gastric
wall into contact with the abdominal wall. Gastric underinflation does not push the colon caudally,
making it more vulnerable to puncture. Gastric overinflation may rotate the colon anteriorly, again
making it more vulnerable to puncture. Unequivocal transillumination of the abdominal wall and digital
pressure that indents the central portion of the stomach are the two most important safeguards for the
prevention of mishaps during this procedure.104(5334) Failure to demonstrate either should lead the
endoscopist to consider aborting the procedure and recommending an open operation. Epigastric
vessels, if visualized during transillumination, should be avoided.
After the appropriate site has been determined, a small transverse skin incision (about 1 cm) is made.
This must be long enough to allow egress of any subcutaneous fluid accumulation around the tube.
The use of a curved Kelly hemostat to maintain the intragastric mound facilitates placement of the
endoscopic snare around the mound. The abdominal wall is pierced with a 16-gauge, tapered,
intravenous cannula using a quick, deliberate motion directed at the tip of the hemostat to ensure that
puncture of the stomach occurs at the apex of the intragastric mound. The endoscopist then snares
the heavy silk suture or plastic-coated wire that has been passed through the cannula, the endoscope
is withdrawn, the PEG catheter is interlooped with the guidewire or silk suture, and the well-lubricated
catheter is pulled through the patient's mouth, esophagus, upper stomach, and across the gastric and
abdominal walls. The markings approximating the abdominal wall thickness allow the endoscopist to
estimate the location of the inner retainer. An external crossbar reinforced with tape is used to secure
the catheter. The inner retainer and outer crossbar should not be tight, because this may result in
pressure necrosis, pain, and leakage. The stomach is decompressed for 12 to 24 hr after the
procedure, and regular feedings are commenced the next day.
There is minimal or no ileus after the PEG procedure. Pneumoperitoneum is common. There is no
reason to obtain a post-PEG abdominal radiograph. The most serious complication of PEG, gastrocolic
fistula, is avoided by strict adherence to the rule of unequivocal transillumination and observed digital
pressure.104(5335) In the retrospective series of Marin et al.109(5336) of 70 patients between the
ages of 3 months and 24 years, the major complication rate was 13%, and multiple organ system
failure was associated with a high risk of complications.
Placement of Transpyloric Feeding Tubes

Intragastric feeding may be poorly tolerated or pose an undue risk of aspiration in certain patients. If
patients do not tolerate intragastric feeding after gastrostomy, a small-diameter polyurethane tube can
be passed through the gastrostomy tube and placed endoscopically in the small intestine to permit
jejunal feedings.110(5337) In an intensive care setting, the endoscope can also be used to facilitate
passage of a polyurethane nasojejunal tube to allow enteral feedings to be initiated in a patient with
postsurgical gastroparesis or gastroparesis associated with critical illness.
REFERENCES
1.
Benaroch LM, Rudolph CD. Introduction to pediatric esophagogastroduodenoscopy and

enteroscopy. Gastrointest Endosc Clin North Am 1994;4:12142.
2. Caulfield M, Wyllie R, Sivak MV Jr, et al. Upper gastrointestinal tract endoscopy in the
pediatric patient. J Pediatr 1989;115:33945.
3. Van Stiegmann G, Stellin GP. Emergent and therapeutic upper gastrointestinal endoscopy in
children. World J Surg 1985;9:2949.
4.
Moyer MS, Gryboski JD. Pediatric endoscopy: Gastroscopy. In Barkin J, O'Phelan CA (eds).
Advanced Therapeutic Endoscopy. New York: Raven Press, 1990;30919.
5. Dedinsky GK, Vane DW, Black CT, et al. Complications and reoperation after Nissen
fundoplication in childhood. Am J Surg 1987;153:17783.
6. Bendig DW. Pulse oximetry and upper intestinal endoscopy in infants and children. J Pediatr
7. Gilger MA, Jeiven SD, Barrish JO, McCarroll LR. Oxygen desaturation and cardiac
arrhythmias in children during esophagogastroduodenoscopy using conscious sedation.
8. Dajani AS, Bisno AL, Chung KJ, et al. Prevention of bacterial endocarditis. Recommendations
by the American Heart Association. JAMA 1990;264:291922.
9.
Schreiner MS. The preop fast: Not quite so fast. Contemp Pediatr 1992;June:4552.
10. Jameson JS, Kapadia SA, Polson RJ, et al. Is oropharyngeal anaesthesia with topical
lignocaine useful in upper gastrointestinal endoscopy?. Aliment Pharmacol Ther
1992;6:73944.
11. Kauffman RE, Banner W, Berlin CM, et al. Guidelines for monitoring and management of
pediatric patients during and after sedation for diagnostic and therapeutic procedures.
Pediatrics 1992;89:11105.
12. Green SM, Johnson NE. Ketamine sedation for pediatric procedures: Part 1, a prospective
series. Ann Emerg Med 1990;19:102432.
13. Green SM, Johnson NE. Ketamine sedation for pediatric procedures: Part 2, review and
implications. Ann Emerg Med 1990;19:103346.
14.
Ament ME. Prospective study of the risks of complication in 6424 procedures in pediatric
gastroenterology [Abstract]. Pediatr Res 1981;15:524A.
15.
Boyle JT. Gastrointestinal emergencies. In Fleisher GR, Ludwig S (eds). Textbook of Pediatric
Emergency Medicine. 3rd ed. Philadelphia: Williams & Wilkins, 1993;896913.
16. Hassall E. Nonsurgical treatments for portal hypertension in children. Gastrointest Endosc
17. Hyams JS, Treem WR. Portal hypertensive gastropathy in children. J Pediatr Gastroenterol
Nutr 1993;17:138.
18. Boyle JT. Gastroesophageal reflux in the pediatric patient. Gastroenterol Clin North Am
1989;18:31537.
19. Baer M, Maki M, Nurminen J, et al. Esophagitis and findings of long-term esophageal pH
recording in children with repeated lower respiratory tract symptoms. J Pediatr Gastroenterol
Nutr 1986;5:18790.
20. Biller JA, Winter HS, Grand RJ, Allred EN. Are endoscopic changes predictive of histologic
esophagitis in children?. J Pediatr 1983;103:2158.
21. Hyams JS, Ricci AJ, Leichtner AM. Clinical and laboratory correlates of esophagitis in young
children. J Pediatr Gastroenterol Nutr 1988;7:526.
22. Shub MD, Ulshen MH, Hargrove CB, et al. Esophagitis: A frequent consequence of
gastroesophageal reflux in infancy. J Pediatr 1985;107:8814.
23. Winter HS, Madara JL, Stafford RF, et al. Intraepithelial eosinophils: A new diagnostic
criterion for reflux esophagitis. Gastroenterology 1982;83:81823.
24. Dahms BB, Rothstein FC. Mucosal biopsy of the esophagus in children. Perspect Pediatr
Pathol 1987;11:97123.
25. Ismail-Beigi F, Horton PF, Pope EC 2d. Histological consequences of gastroesophageal reflux
in men. Gastroenterology 1970;58:16374.
26. Kelly KJ, Lazenby AJ, Rowe PC, et al. Eosinophilic esophagitis attributed to gastroesophageal
reflux: Improvement with an amino acid-based formula. Gastroenterology 1995;109:150312.
27. Black DD, Haggitt RC, Orenstein SR, Whitington PF. Esophagitis in infants: Morphometric
histological diagnosis and correlation with measures of gastroesophageal reflux.
28. Hassall E. Barrett's esophagus: New definitions and approaches in children. J Pediatr
29. Hassall E, Weinstein WM, Ament ME. Barrett's esophagus in childhood. Gastroenterology
1985;89:13317.
30. Hassall E, Dimmick JE, Magee JF. Adenocarcinoma in childhood Barrett's esophagus: Case
documentation and the need for surveillance in children. Am J Gastroenterol 1993;88:2828.
31. Baehr PH, McDonald GB. Esophageal infections: Risk factors, presentation, diagnosis, and
32. Braaegger CP, Albisetti M, Nadal D. Extensive esophageal candidiasis in the absence of oral
lesions in pediatric AIDS?. J Pediatr Gastroenterol Nutr 1995;21:1046.
33. Theise ND, Rotterdam H, Dieterich D. Cytomegalovirus esophagitis in AIDS: Diagnosis by
endoscopic biopsy. Am J Gastroenterol 1991;86:11236.
34. Ashenburg C, Rothstein FC, Dahms BB. Herpes esophagitis in the immunocompetent child. J
Pediatr 1986;108:5847.
35. McCord GS, Clouse RE. Pill-induced esophageal strictures: Clinical features and risk factors
for development. Am J Med 1990;88:5128.
36.
Boyle JT. Vomiting. In Reece RM (ed). Manual of Emergency Pediatrics. Philadelphia: WB
37.
Boyle JT. Abdominal pain. In Walker WA, Durie PR, Hamilton JR, Walker-Smith JA, Watkins
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
JB (eds). Pediatric Gastrointestinal Disease. PathophysiologyDiagnosisManagement. 2nd

ed. St. Louis: Mosby-Year Book, 1996;20527.
Apley JS. The Child with Abdominal Pains. London: Blackwell Scientific. 1975.
Boyle JT. A new look at recurrent abdominal pain in children by subdivision of patients into
symptomatic subgroups: Simplifying the role of endoscopy in the diagnostic evaluation.
Baillieres Clin Pediatr 1994;2:787806.
Treem WR, Ragsdale BD. Crohn's disease of the esophagus: A case report and review of the
literature. J Pediatr Gastroenterol Nutr 1988;7:4515.
Griffiths AM, Alemayehu E, Sherman P. Clinical features of gastroduodenal Crohn's disease
in adolescents. J Pediatr Gastroenterol Nutr 1989;8:16671.
Bunjanover Y, Konikoff F, Baratz M. Nodular gastritis and Helicobacter pylori. J Pediatr
Chouraqui JP, Roy CC, Brochu P, et al. Mntrier's disease in children: A report of a patient
and review of sixteen other cases. Gastroenterology 1981;80:10427.
Occena RO, Taylor SF, Robinson CC, Sokol RJ. Association of cytomegalovirus with
Mntrier's disease in childhood: Report of two cases with review of literature. J Pediatr
Banerjee B, Rashid S, Singh E, Moore J. Endoscopic findings in Henoch-Schnlein purpura.
Tomomasa T, Hsu JY, Itoh K, Kuroume T. Endoscopic findings in pediatric patients with
Henoch-Schnlein purpura and gastrointestinal symptoms. J Pediatr Gastroenterol Nutr
1987;6:7259.
Corazza GR, Caletti GC, Lazzari R, et al. Scalloped duodenal folds in childhood celiac
Couper R, Laski B, Drumm B, et al. Chronic varioliform gastritis in childhood. J Pediatr
1989;115:4414.
Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Pediatr
Schuffler MD, Chaffee RG. Small intestine biopsy in a patient with Crohn's disease of the
duodenum. Gastroenterology 1979;76:100914.
Peterson WL. Helicobacter pylori and peptic ulcer disease. N Engl J Med 1991;324:10438.
DeGiacomo C, Gianetti A, Negrini R, et al. Lymphocytic gastritis: A positive relationship with
celiac disease. J Pediatr 1994;124:5762.
Talley NJ, Phillips SF. Non-ulcer dyspepsia: Potential causes and pathophysiology. Ann Intern
Med 1988;108:86579.
Coello-Ramirez P, Larrosa-Haro A. Gastrointestinal occult hemorrhage and gastroduodenitis
in cow's milk protein intolerance. J Pediatr Gastroenterol Nutr 1984;3:2158.
Bendig DW. Diagnosis of giardiasis in infants and children by endoscopic brush cytology. J
Pediatr Gastroenterol Nutr 1989;8:2046.
Gupta SK, Fitzgerald JF, Chong SKF. Endoscopically assisted diagnosis of cryptosporidiosis
in two immunocompetent children. J Pediatr Gastroenterol Nutr 1995;21:4747.
Oderda G, Forni M, Morra I, et al. Endoscopic and histologic findings in the upper
gastrointestinal tract of children with coeliac disease. J Pediatr Gastroenterol Nutr
1993;16:1727.
Spodaryk M, Mrukowicz J, Stopyrowa J, et al. Severe intestinal nodular lymphoid hyperplasia
in an infant. J Pediatr Gastroenterol Nutr 1995;21:46873.
Abramowsky C, Hupertz V, Kilbridge P, Czinn S. Intestinal lymphangiectasia in children: A
study of upper gastrointestinal endoscopic biopsies. Pediatr Pathol 1989;9:28997.
Schmidt-Sommerfeld E, Kirschner BS, Stephens JK. Endoscopic and histologic findings in the
upper gastrointestinal tract of children with Crohn's disease. J Pediatr Gastroenterol Nutr
1990;11:44854.
Treem WR. Congenital sucrase-isomaltase deficiency. J Pediatr Gastroenterol Nutr
1995;21:114.
62. Maksimak M, Cera PJ Jr. Mucosal biopsy. Gastrointest Endosc Clin North Am
1994;4:195221.
63. Granot E, Goodman-Weill M, Pizov G, Sherman Y. Histological comparison of suction
capsule and endoscopic small intestinal mucosal biopsies in children. J Pediatr Gastroenterol
Nutr 1993;16:397401.
64. Byrne WJ. Foreign bodies, bezoars, and caustic ingestion. Gastrointest Endosc Clin North Am
1994;4:99119.
65. Walton WW. An evaluation of the Poison Prevention Packaging Act. Pediatrics
1982;69:36370.
66.
Wason S. The emergency management of caustic ingestions. J Emerg Med 1987;79:4137.
67. Maull KI, Scher LA, Greenfield LJ. Surgical implications of acid ingestion. Surg Gynecol
Obstet 1979;148:8958.
68. Anderson KD, Rouse TM, Randolph JG. A controlled trial of corticosteroids in children with
corrosive injury of the esophagus. N Engl J Med 1990;323:63740.
69. Gaudreault P, Parent M, McGuigan MA, et al. Predictability of esophageal injury from signs
and symptoms: A study of caustic ingestion in 378 children. Pediatrics 1983;71:76770.
70.
71. Storey DW, Bown SG, Swain CP, et al. Endoscopic prediction of recurrent bleeding in peptic
72. Hassall E, Berquist WE, Ament ME, et al. Sclerotherapy for extrahepatic portal hypertension
in childhood. J Pediatr 1989;115:6974.
73. Howard ER, Stringer MD, Mowat AP. Assessment of injection sclerotherapy in the
management of 152 children with oesophageal varices. Br J Surg 1988;75:4048.
74. Proujansky R, Orenstein SR, Kocoshis SA. Patient and procedure variables associated with
complications following variceal sclerotherapy in children. J Pediatr Gastroenterol Nutr
1991;12:338.
75. Atkinson JB, Woolley MM. Treatment of esophageal varices by sclerotherapy in children. Am
J Surg 1983;146:1036.
76. Dall'Oglio L, Bagolan P, Ferro F, et al. Endoscopic injection sclerosis of oesophageal varices
in childrenIndications and techniques. Endoscopy 1984;16:98100.
77. Paquet KJ, Lazar A. Current therapeutic strategy in bleeding esophageal varices in babies
and children and long-term results of endoscopic paravariceal sclerotherapy over twenty
years. Eur J Pediatr Surg 1994;4:16572.
78. Supe AN, Mathur SK, Borwankar SS. Esophageal endoscopic sclerotherapy in children using
3% aqueous phenol. Indian J Gastroenterol 1994;13:14.
79. Thapa BR, Mehta S. Endoscopic sclerotherapy of esophageal varices in infants and children.
J Pediatr Gastroenterol Nutr 1990;10:4304.
80. Stellen GP, Lilly JR. Esophageal endosclerosis in children. Surgery 1985;98:9705.
81. Stringer MD, Howard ER. Long-term outcome after injection sclerotherapy for oesophageal
varices in children with extrahepatic portal hypertension. Gut 1994;35:2579.
82. Fox VL, Carr-Locke DL, Connors PJ, Leichtner AM. Endoscopic ligation of esophageal
varices in children. J Pediatr Gastroenterol Nutr 1995;20:2028.
83. Hassall E, Treem WR. To stab or strangle: How best to kill a varix?. J Pediatr Gastroenterol
Nutr 1995;20:1214.
84.
Pellerin D, Fortier-Beaulieu M, Gueguen J. The fate of swallowed foreign bodies: Experience
of 1250 instances of sub-diaphragmatic foreign bodies in children. Prog Pediatr Radiol
1969;2:286302.
85. Webb WA. Management of foreign bodies of the upper gastrointestinal tract.
86. Caravati EM, Bennett DL, McElwee NE. Pediatric coin ingestion. A prospective study on the
utility of routine roentgenograms. Am J Dis Child 1989;143:54951.
87. Schunk JE, Corneli H, Bolte R. Pediatric coin ingestion: A prospective study of coin location
and symptoms. Am J Dis Child 1989;143:5468.

88. Hodge D, Tecklenberg F, Fleisher G. Coin ingestion: Does every child need a radiograph?.
Ann Emerg Med 1985;14:4436.
89. Litovitz TL. Battery ingestions: Product accessibility and clinical course. Pediatrics
1985;75:46976.
90. Litovitz T, Schmitz BF. Ingestion of cylindrical and button batteries: An analysis of 2382 cases.
Pediatrics 1992;89:74757.
91. O'Neill JA, Betts J, Ziegler MM, et al. Surgical management of reflux strictures of the
esophagus in childhood. Ann Surg 1982;196:4539.
92. Shah MD, Berman WF. Endoscopic balloon dilation of esophageal strictures in children.
93. Tam PK, Sprigg A, Cudmore RE, et al. Endoscopy-guided balloon dilatation of esophageal
strictures and anastomotic strictures after esophageal replacement in children. J Pediatr Surg
1991;26:11013.
94. Orenstein SR, Whitington PF. Esophageal stricture dilatation in awake children. J Pediatr
95. Goldthorn JF, Ball WS, Wilkinson LG, et al. Esophageal strictures in children: Treatment by
serial balloon catheter dilatation. Radiology 1984;153:6558.
96. Berenson GA, Wyllie R, Caulfield M, Steffen R. Intralesional steroids in the treatment of
refractory esophageal strictures. J Pediatr Gastroenterol Nutr 1994;18:2502.
97. Goyal RK, Glancy JJ, Spiro HM. Lower esophageal ring. N Engl J Med 1970;282:135562.
98. Shoshany G, Bar-Maor JA. Congenital stenosis of the esophagus due to tracheobronchial
remnants: A missed diagnosis. J Pediatr Gastroenterol Nutr 1986;5:9779.
99. Garau P, Orenstein SR. Congenital esophageal stenosis treated by balloon dilatation. J
Pediatr Gastroenterol Nutr 1993;16:98101.
100. Marshall JB, Kretschmar JM, Diaz-Arias AA. Gastroesophageal reflux as a pathogenic factor
in the development of symptomatic lower esophageal rings. Arch Intern Med
1990;150:166972.
101. Boyle JT, Cohen S, Watkins JB. Successful treatment of achalasia in childhood by pneumatic
dilatation. J Pediatr 1981;99:3540.
102. Neumann G, Samani U, Nas S. A new endoscopic technique for introducing pneumatic
dilators in patients with achalasia. Gastrointest Endosc 1992;38:598601.
103. Gauderer MWL. Percutaneous endoscopic gastrostomy: A 10-year experience with 220
children. J Pediatr Surg 1991;26:28894.
104.
Gauderer MWL, Stellato TA. Percutaneous endoscopic gastrostomy in children: The
technique in detail. Pediatr Surg Int 1991;6:827.
105.
Boyle JT. Nutritional management of the developmentally disabled child. Pediatr Surg Int
1991;6:7681.
106. Byrne WJ, Euler AR, Ashcraft E, Nash DG. Gastroesophageal reflux in the severely retarded
who vomit: Criteria for and results of surgical intervention in twenty-two patients. Surgery
1982;91:958.
107. Turnage WP, Oldham KT, Coran AG, Blane CE. Late results of fundoplication for
gastroesophageal reflux in infants and children. Surgery 1989;105:45764.
108. Langer JC, Wesson DE, Ein SH, et al. Feeding gastrostomy in neurologically impaired
children: Is an antireflux procedure necessary?. J Pediatr Gastroenterol Nutr 1988;7:83741.
109. Marin OE, Glassman MS, Schoen BT, Caplan DB. Safety and efficacy of percutaneous
endoscopic gastrostomy in children. Am J Gastroenterol 1994;89:35761.
110. Strife JL, Dunbar JS, Rice S. Jejunal intubation via gastrostomy catheters in pediatric patients.
Radiology 1985;154:249.
Chapter 54 Foreign Body Extraction

(5338)
ROBERT A. SANOWSKI, M.D.

M. EDWYN HARRISON, M.D.
MARK F. YOUNG, M.D.
PAUL J. BERGGREEN, M.D.
Ingestion of a foreign body is a common occurrence. Fortunately, most foreign objects pass without
causing symptoms. Only 10 to 20% fail to traverse the entire gastrointestinal tract,1(5339) but
impaction in the alimentary tract is relatively common because of the high frequency of ingestion. In a
Swedish study, the annual incidence of hospitalization for foreign body ingestion was 122 per million
persons.2(5340) In many cases, retention of a foreign body can pose a serious threat to health. An
estimated 1500 deaths occur each year in the United States as a result of foreign body
ingestions.3(5341)
Because a remarkable variety of articles may be ingested, the endoscopist must be able to use various
instruments to remove objects that lodge in the esophagus, stomach, and duodenum. The imagination
of the physician is often tested by marked variations in size, shape, and potential for injury. The
decision to remove a foreign body is influenced by the age and clinical status of the patient; the size,
shape, and nature of the material ingested; the anatomic area in which the object is lodged; and the
technical ability of the endoscopist. Fiberoptic instruments with accessories facilitate removal of foreign
bodies with minimal mortality and morbidity.
Those who swallow foreign objects are most commonly children, psychotic patients, manipulative
individuals seeking some secondary gain, and careless or reckless individuals who place objects in
their mouths. Smooth, round objects usually pass through the alimentary tract, but endoscopic removal
of the object is required when a foreign body has lodged proximal to the diaphragmatic hiatus.4(5342)
Psychotic and mentally retarded patients may ingest great quantities of bizarre materials, necessitating
endoscopic or surgical removal.5(5343) Prisoners may swallow foreign bodies as a manipulative
measure. In this group, diagnosis is not difficult because the patient is aware of the occurrence and
communicates the fact for secondary gain (e.g., hospitalization and endoscopic therapy are preferable
to prison). Acute and chronic forms of alcoholism are frequently associated with foreign body ingestion
and food bolus impaction.
Elderly individuals with dentures may have poor gingival sensation and be unable to detect swallowed
bones. Loose or broken dentures may become dislodged and lodge in the upper gastrointestinal tract.
Poor vision and rapid eating associated with poor mastication can compound the problem.
The materials that are ingested consist of a wide variety of substances limited only by human
imagination and the size of the oropharyngeal cavity. Generally, four types of material can lodge in the
gastrointestinal tract: bezoars, most commonly composed of hair or vegetable matter; food; true
foreign bodies such as bones, coins, dentures, or pins; and a variety of medical prostheses. The types
of true foreign bodies ingested vary according to the patient's age, occupation, and dietary habits. Fish
bone and chicken bone impactions are common in Asian populations.6(5344) Dentures, toothpicks,
wire, pins, and tacks may be swallowed accidentally by the older patient or by those who use any of
these objects in their occupations. Most patients ingesting true foreign bodies are younger than 40
years. Food impaction usually occurs in those older than 60 years.7(5345)
Anatomy of Impaction
The sites of impaction in the gastrointestinal tract are the normal anatomic and pathologic points of
narrowing. Beyond the oropharynx, pointed objects may lodge in the piriform sinuses and valleculae.
Approximately 50 to 80% of esophageal foreign bodies become arrested in the cervical esophagus,
and the remainder lodge in the upper thoracic and distal esophagus.6(5346) The normal anatomic
points of narrowing at the level of the cricopharyngeus muscle, aortic arch, left main stem bronchus,
and cardioesophageal junction can entrap sharp foreign bodies. Objects impacted at the aortic arch
may perforate into the aorta, leading to an aorto-esophageal fistula and exsanguination.6(5347)
After passing through the esophagus, most foreign bodies move through the remainder of the
alimentary tract. However, if an object is long (e.g., spoons, wires, springs) or composed of material
that cannot pass readily from the stomach (e.g., hair), impaction and obstruction may occur at the
pylorus. Other points of anatomic narrowing are the duodenum, ileocecal valve, Meckel's diverticulum,
and anus.
Impaction and obstruction may occur at pathologic points of narrowing in the stomach, duodenum, and
ileum. Esophageal stricture, hiatus hernia, and tortuousity of the cardioesophageal junction are
frequently associated with food impaction. Carcinoma of the esophagus uncommonly results in the
arrest of a foreign body,6(5348) probably because the stenosis caused by the growing tumor develops
gradually and patients typically exercise caution while eating to avoid obstruction. Obstruction of an
esophageal prosthesis placed for palliation of encroaching carcinoma occasionally occurs (see
Chapter 42: Palliation of Malignant Obstruction: Dilation and Peroral Prosthesis). Stenosis of the
pylorus, alterations in the stomach after operation, and impaired gastric emptying are associated with
retention of foreign bodies. Ileal strictures can entrap swallowed materials.
Sharp objects, including needles, pass through the alimentary tract in 90% of patients without injury.
This phenomenon can be explained by a reflex relaxation of the intestinal musculature that occurs in
response to the presence of the object. Axial flow in the intestinal lumen, combined with reflex mural
relaxation and slowing of peristalsis, tends to turn these objects so that the sharper end trails as the
object moves through the gut.1(5349) In the colon, foreign objects become covered with fecal material,
and the bowel wall is thereby protected from injury. Foreign bodies in the colon and rectum are
discussed in Chapter 87: Miscellaneous Disorders of the Colon.
Clinical Features
The presenting symptoms that result from a gastrointestinal foreign body vary widely. The presentation
may be acute, and the patient may recall in detail the time and type of foreign body ingested. However,
in children, the mentally deranged or retarded, or those ingesting objects for secondary gain,
complications related to the presence of a foreign body may bring the patient to the physician.
Symptoms depend on the area of lodgment. In adults, foreign bodies in the esophagus cause pain and
discomfort on swallowing. There may be a persistent sensation or awareness on swallowing of a
foreign body (e.g., fish bone). The saliva may become blood tinged, and the patient may have a history
of gagging or choking during meals. Children may refuse to take feedings and may have increased
salivation, pain and discomfort on swallowing, and vomiting. Fever, pain, and tenderness in the neck
may indicate perforation, and a high index of suspicion of foreign body ingestion must be entertained
for patients with severe psychotic or emotional problems or mental retardation.
Sudden onset of dysphagia while eating or drinking indicates lodgment of a foreign body or impacted
food. The latter is commonly meat with a stringy texture (e.g., roast beef, steak). Repeated foreign
body ingestion or food bolus impaction is more likely to occur in patients with symptoms of esophageal
reflux. Gastric foreign bodies are frequently associated with partial or complete outlet obstruction, with
accompanying characteristic symptoms.
Physical findings are not remarkable unless the foreign body can be palpated in the pharynx or a mass
is present in the abdomen, as occurs with a trichobezoar. Nonspecific, mild abdominal tenderness and
a decrease in bowel sounds may be found.
Diagnostic Evaluation
A plain roentgenogram should be obtained first if the object is thought to be radiopaque, because it
may be localized easily with anteroposterior and lateral views. Care must be taken to ascertain whether
the object is in the airway or extraluminal air is present. Objects such as fish and chicken bones, glass,
wood, and plastic are not radiopaque. Special radiographic techniques for imaging of the soft tissues of
the neck may be necessary to outline thin metallic objects.
After a screening roentgenogram, endoscopy is the preferred method of evaluation. Endoscopy often
can provide accurate diagnosis and effective treatment. In the past, a radiographic contrast
examination with barium or Gastrografin (meglumine diatrizoate) has been suggested as the second
step in the diagnosis and localization of foreign bodies. Contrast studies provide little useful
information, except in specific circumstances in the pediatric population, and they may compromise the
endoscopic view of the foreign body.
Treatment of Ingested Foreign Bodies

After the diagnosis of foreign body ingestion has been made, the necessity for removal is considered.
The indication and timing of endoscopy depend on the type of foreign body ingested. Eighty to 90% of
materials pass spontaneously through the gastrointestinal tract, and endoscopy is unnecessary when
this is the case.1(5350) Approximately 20% of patients need endoscopy, and only 1% require surgical
intervention.8,9(5351) The method of removal depends on the type of foreign body ingested (i.e., size,
shape, and whether sharp or pointed) and whether it is located in the esophagus, stomach, or
duodenum. The success rates for endoscopic extraction of foreign bodies impacted in the esophagus
is essentially the same whether performed with rigid and flexible instruments, although the
complication rate is significantly higher with rigid endoscopes (10% vs. 5.1%).10(5352)
Meat Bolus Removal

The most common esophageal foreign body in the adult population is impacted meat.11(5353)
Predisposing factors include narrowing of the lower esophagus secondary to peptic esophageal
stricture and Schatzki's ring. Patients with meat impaction should not undergo barium contrast
radiographic studies of the esophagus. Little in the way of information is obtained, and the barium
obscures the esophageal lumen during endoscopy. The use of Gastrografin is also discouraged
because a severe pneumonitis may result if aspiration occurs.12(5354) The timing of endoscopy
depends on the clinical status of the patient. Those with copious amounts of saliva and secretions are
at increased risk for aspiration and should undergo endoscopy on an emergent basis. Patients with
chronic strictures and recurrent meat impaction who are not having difficulty with secretions do not
require immediate intervention.
The use of papain meat tenderizer has no role in the management of meat impaction in the
esophagus. This agent was first used by Richardson in 1945 in the treatment of 17 patients who
presented with this condition.13(5355) Several types of complications have resulted from its use.
Esophageal perforation and frank necrosis of most of the thoracic esophagus was reported by
Holsinger et al.14(5356) Hyperosmolar coma related to the high content of sodium chloride in this
agent has also been reported.15(5357) Goldner and Danley16(5358) found that papain had no
inherent capacity in vitro to digest or reduce the size of meat cubes. These investigators found that the
tenderizer significantly worsened the degree of esophagitis in previously inflamed rabbit mucosa.
Glucagon reduces lower esophageal sphincter pressure and may enhance passage of an impacted
meat bolus. Success rates as high as 50% have been reported in radiologic studies.17(5359) With the
use of flexible endoscopes, however, this agent is usually not needed.
Endoscopy is the method of choice for removal of an impacted meat bolus. Rigid endoscopy is rarely
needed, and fiberoptic endoscopy is safer and can be performed without general anesthesia. Because
patients with an impacted meat bolus are usually uncomfortable and anxious, a somewhat deeper level
of conscious sedation than that employed for routine diagnostic esophagogastroduodenoscopy may be
required. The endoscope is passed under direct vision, and the location of the bolus within the
esophagus is determined. The endoscope is then withdrawn, and an overtube is placed. Use of this
device substantially reduces the number of times the esophagus must be intubated.12(5360) An
overtube can be constructed from polyvinyl tubing. A 40-French Maloney dilator is lubricated and
inserted in the overtube.18(5361) The dilator and overtube are then passed as a unit beyond the
cricopharyngeus muscle, the dilator is removed, and the endoscope is passed through the overtube
and into the esophagus.
If endoscopy is performed in the first few hours after ingestion, the bolus can usually be removed as a
single piece with a polypectomy snare. After several hours, the meat is often fragmented due to partial
enzymatic digestion. The fragmented bolus may be removed piecemeal with a foreign body forceps or
polypectomy snare. Another option is to gently push the meat bolus into the stomach. Before this is
attempted, the esophageal lumen distal to the impaction must be fully visualized to determine that the
luminal diameter is adequate. Otherwise, there is an increased risk of perforation.
Any patient who presents with an initial episode of impaction should undergo a complete
esophagogastroduodenoscopic examination to determine the cause of esophageal narrowing. Dilation
is indicated for patients with a peptic stricture or Schatzki's ring.
A life-threatening situation may result if the meat bolus lodges at or just below the level of the
cricopharyngeus and compresses the adjacent trachea. The immediate result is the same as that for
lodgment of a food bolus within the trachea: acute obstruction of the airway and impending suffocation.
This situation is sometimes referred to as a cafe coronary because bystanders often assume that the
victim is suffering an acute myocardial infarction.19(5362) Immediate action to relieve the obstruction
is required in such situations.
In 1975, Heimlich20(5363) described a maneuver for resuscitation of individuals with acute airway
obstruction. This is performed by standing behind the patient with arms wrapped around the victim.
The rescuer's fists are placed in vertical apposition high in the patient's epigastrium, and forcible
pressure is applied. This action usually dislodges the impaction. Although the Heimlich maneuver is
frequently lifesaving, complications related to its use, whether performed correctly or incorrectly, have
been described. Based on numbers of case reports, the most common of these are rupture of a hollow
organ such as the stomach,2126(5364) esophagus,27(5365) or jejunum;28(5366) injury to a major
vascular structure;2932(5367) and pneumomediastinum.3335(5368)
Sharp Foreign Bodies

Sharp and pointed foreign bodies pose a greater risk to the patient, with perforation occurring in 35% of
cases.36(5369) The most common sharp objects ingested include glass, needles, pins, and dental
bridges. Because of the high rate of laceration in the stomach and intestine, sharp foreign bodies
should be removed after identification. For extraction of such objects, the use of an overtube is
mandatory. This device affords excellent protection to the esophageal mucosa and greatly decreases
the risk of aspiration. The first use of an overtube was for extraction of razor blades from the
stomach.37(5370) The endoscope is then passed and the object identified. Various types of
endoscopic devices for grasping foreign bodies are available (Figure 541).
(5371)Figure 541. Six commmonly used endoscopic forceps and grasping devices that are
available for foreign body removal.

Small, sharp foreign bodies can be withdrawn safely through the overtube.18(5372) When an overtube
is used to extract larger foreign objects, it may be possible and necessary to remove the overtube,
endoscope, and the object together as a unit if the object is too large to fit through the overtube.
However, this depends on the nature and shape of the object. In some cases, the potentially
dangerous portion of the foreign body can be held within the overtube, or it may be possible to keep
the pointed end of an object oriented opposite to the direction of movement as the overtube is being
removed.
A latex hood that is fitted on to the distal end of the endoscope has been used to remove sharp or
pointed foreign bodies.38(5373) After the sharp object is retrieved, the latex hood everts to cover the
object when the endoscope is withdrawn into the esophagus, protecting the mucosa.
Removal of an open safety pin requires particular skill. If lodged in the stomach, the safety pin should
be grasped at the hinged end. It can then be withdrawn into the overtube; once inside, the overtube
itself is withdrawn. A similar method can be employed to remove a safety pin lodged in the esophagus
when the hinged end is positioned proximally. If the hinged end is located distally, the safety pin must
first be pushed into the stomach. It can then be removed hinged end first through the overtube (Figure
542).39(5374)
(5375)Figure 542. Illustration of the overtube-endoscope technique for removing sharp

objects from the esophagus and stomach. A, Large straight pin is grasped by the forceps and
pulled into the overtube. B, Flip-top ring from a soft drink container is pulled into the
overtube. C, Large open safety pin is pulled back into overtube. Once a sharp object has been
completely withdrawn into the overtube, the overtube and endoscope may be withdrawn safely
at the same time.
If a sharp or pointed object passes beyond the stomach, daily plain radiographs are obtained to
document progression of the object. Surgical intervention is advocated if the foreign body fails to
advance after 72 hr or the patient becomes symptomatic.39(5376)
Proximal Esophageal Foreign Bodies

Patients with foreign bodies lodged at or above the cricopharyngeus require a different approach. The
patient is placed in a supine position and intravenous sedation is given. A laryngoscope is placed
(preferably a Miller blade) and the foreign body located. A long surgical grasping clamp (Kelly or
McGill) may be used to withdraw the object. Most are located anterior to the epiglottis at the base of
the tongue or in the tonsillar pillar.40(5377) If laryngoscopic removal is unsuccessful, the patient is
placed in the left lateral decubitus position, and the flexible endoscope is introduced under direct vision.
A thorough examination of the hypopharynx is then performed. The identified object is removed with
foreign body forceps. If the object is not identified, the patient should return when symptoms recur.
Body Packers
A novel problem with foreign bodies has been created by drug smuggling and the use of body orifices
to conceal illegal drugs, an activity described by the terms body packer and body bagger.17(5378)
Drugs are placed in condoms, balloons, or plastic bags that are then ingested or inserted into the
vagina or rectum. The illicit drugs most commonly smuggled in this fashion include cocaine (i.e.,
boletas), marijuana, and heroin.41(5379) Possible complications include fatal intoxication from leakage
of the drug from the bag and bowel obstruction necessitating surgery.42,43(5380)
The evaluation of such patients should begin with a history and physical examination, including rectal
and vaginal inspection for hidden materials. Plain abdominal x-ray films or computed tomography (CT)
scans may reveal concealed packets in the gastrointestinal tract or pelvis. These appear as multiple,
homogeneous, oval or cylindrical structures surrounded by a thick radiolucent halo.42(5381) The rate
of positive abdominal x-ray films for individuals who engage in this activity ranges from 72 to
92%.43(5382) Blood and urine drug screens are then obtained.44(5383)
Patients who are asymptomatic at the time of evaluation may be observed if less than 24 hr has
elapsed since ingestion. The use of endoscopy to remove packets is discouraged. Biopsy forceps can
tear the containers, resulting in acute intoxication.45(5384) The use of enemas, cathartics, and
emetics is also not recommended because these agents may also rupture the packets. If the packets
fail to pass spontaneously after 48 hr or signs of bowel obstruction develop, surgical intervention is
recommended. Abdominal x-ray films should be obtained before discharge to document the removal of
all packets.
Button Batteries
The increasingly common ingestion of button batteries has paralleled the increased production of
digital watches, calculators, hearing aids, and cameras. These tiny power units contain metallic salts
and alkali. Metallic salts include mercuric oxide, silver oxide, manganese dioxide, zinc oxide, and
lithium hydroxide, and typical alkali are sodium or potassium hydroxide.4650(5385)
Ingestion of button batteries poses several potential problems to patients. Local damage to the
gastrointestinal mucosa may result from pressure necrosis, direct corrosive action, and low-voltage
currents.51(5386) Disintegration of the battery casing may result in toxic serum concentrations of
mercury. Major complications have occurred when the battery became lodged in the esophagus,
including several deaths related to the development of tracheoesophageal fistula.52(5387)
The approach to the management of patients who ingest button batteries is determined by the location
of the foreign object.53(5388) Most of those located in the stomach and intestine pass spontaneously
without therapeutic intervention. Barium contrast radiography of the esophagus should be obtained at
48 hr and 14 days after removal to document that there is no fistula formation.54(5389)
A button battery lodged in the esophagus should be removed as rapidly as possible because of the risk
of fistula development. Two approaches are possible. In the first, an endotracheal tube is placed,
general anesthesia is administered, and the battery is captured and removed with a Dormia-type
basket or polypectomy snare. Attempts to capture a button battery with commonly available
endoscopic grasping devices are usually ineffective because the edges of the object are very smooth.
In the second approach, a flexible endoscope is passed, and a through-the-scope balloon is placed
beyond the battery and inflated. The battery, balloon, and endoscope may then be removed as a unit. If
the disk cannot be removed safely from the esophagus, an alternative is to push it under direct
visualization into the stomach. Attempts at removal may then proceed with a polypectomy snare or
basket (Figure 543).54(5390)
(5391)Figure 543. Retrieval baskets and a snare suitable for foreign body removal (left to
right): A coin is contained in a basket retriever; a button battery is in the basket of a polyp
retriever; a spiral six-wire retrieval basket is suitable for coin removal; a polypectomy snare is
used for toothbrush extraction.
H2-receptor blocking drugs are often administered to decrease gastric acidity in the hope of lessening
the likelihood of corrosion of the battery.53(5392) Metoclopramide has been used to enhance gastric
emptying and promote passage of the battery.53(5393) These practices usually cause no harm, but
they do not have proven benefit. Most batteries pass within 72 hr.55(5394) Stools should be strained
and follow-up x-ray films obtained to document passage. If the patient develops abdominal pain and
the battery fails to progress through the intestine, surgery must be considered.
Toothbrush Extraction
A toothbrush may be accidentally ingested during a seizure, while attempting to induce vomiting, or as
it is used to brush teeth.5658(5395) Diagnosis is possible by plain roentgenogram because the metal
sheaths holding the bristles are radiopaque. Although the toothbrush occasionally passes through the
alimentary tract spontaneously, it usually becomes trapped in the stomach and must be
removed.57(5396)
An impacted toothbrush is the prototype of a technically challenging foreign body extraction: a long,
irregularly shaped, pointed object. Removal should be undertaken cautiously. If possible, a comparable
toothbrush should be obtained and the extraction technique practiced and refined using the endoscope
and various accessories before approaching the patient.57(5397) After this rehearsal, the endoscope
is passed through an overtube into the stomach. The accessory (e.g., polypectomy snare) that proved
to be most suitable for grasping the toothbrush in the practice attempts is then inserted through the
endoscope. After the toothbrush has been grasped, the endoscope, snare, and toothbrush are
withdrawn into the overtube, and the entire assembly is carefully removed.
Atypical Foreign Bodies

An astonishing variety of foreign bodies become lodged in the alimentary tract (Figures 544 and
545). Those discussed previously are only the relatively common objects ingested. Psychiatric
patients may ingest almost any object that can pass the oropharynx. A number of unusual and bizarre
foreign bodies have also been found in patients without psychiatric histories. Their only common
characteristic is that most are iatrogenic.
(5398)Figure 544. A, Plain radiograph of the abdomen of a 40-year-old prisoner with

multiple metallic foreign bodies in the stomach. B, Twenty-one pieces of bedsprings, wires,
toothpaste tubes, and shower grating were removed endoscopically without complications
from the patient described in A.
(5399)Figure 545. A, Endoscopic view of a swallowed penny in a 53-year-old patient with

partial gastric outlet obstruction (the coin was removed with foreign body forceps). B,
Endoscopic view of a hairy, foul-smelling mass in the stomach of a 15-year-old girl who had
been ingesting her hair. C, Gross surgical specimen, a large trichobezoar, was removed from
the stomach of the patient in B. D, Endoscopic view of a bone impacted proximal to the
cardioesophageal junction in a 60-year-old woman with dysphagia and chest pain. She had
ingested fried chicken the previous night. E, Large chicken bone was removed from the
esophagus of the patient in D. F, Endoscopic view of a piece of glass ingested by a
72-year-old woman. A polypectomy snare encircles the broken glass. G, The piece of glass
shown in F is grasped in its long axis with a snare seen at the edge of a black overtube. H,
Colonoscopic view of a straight pin lodged in the ileocecal valve. It was removed with a
forceps. I, Endoscopic view of a nail in the antrum of the stomach as it is grasped with a snare.
J, Colonoscopic view of an impacted dental bridge in the ileum prior to removal.
Endoscopic removal of unusual gastric foreign bodies has been relatively successful. These have
included a gastric balloon, a Marlex ring eroding into the gastric wall, an Atkinson tube, and other
objects.5963(5400) An Angelchik ring found penetrating through the gastric wall had to be removed
surgically.64(5401)
Surgical removal has been required for a variety of objects impacted distally in the intestine and colon.
These have included biliary stents, button batteries trapped in the ileum or colon, a surgical pack that
migrated into the ileum, a hepatic portal vein catheter, and a tooth.6570(5402) Percutaneous
gastrostomy and jejunostomy tubes, like their surgically placed counterparts, have been successfully
removed endoscopically.7173(5403)
Removal of Gastroduodenal Foreign Bodies

Impaired progression of a foreign body through the stomach necessitates its removal, especially if the
object is large, long, sharp, or potentially toxic. Most gastroduodenal foreign bodies can be removed
endoscopically in a manner similar to that used for toothbrush extraction. Before endoscopy, it may be
helpful to practice grasping an object similar to the one ingested. This allows the endoscopist to
determine which endoscopic accessory is best for grasping the object, thereby increasing the safety
and shortening the duration of the procedure.
The overtube is especially useful when several objects must be recovered. Small objects can be
grasped and pulled into the overtube. For larger ones, such as pens, pencils, spoons, and wires, which
are more difficult to remove, a snare or special grasping or foreign body forceps should be passed
ahead of the endoscope; if grasped properly, these objects can frequently be pulled into the overtube.
It may be necessary to hyperextend the patient's neck to bring a large or long foreign body through the
posterior hypopharynx.
Objects that lodge in the duodenum should be pulled back into the stomach with a polypectomy snare
and then be removed as for a gastric foreign body. Glucagon can be administered intravenously to
induce relaxation of the antrum and pylorus. A long overtube may be used to keep the pylorus open
while extracting an object from the duodenum with a polypectomy snare.
Bezoars
A bezoar is a gastric foreign body formed by a conglomeration of plant fiber, hair, food, or any of a
startling variety of less common substances. These masses tend to accrete over time to form an
ever-expanding bulk and worsening clinical problem. In much of history, medicinal powers have been
attributed to the concretion taken from the stomachs of goats and gazelles, and these masses were
given the name bezoar (from the Arabic badzehr and Persian padzahr) to mean antipoison.74(5404)
Modern experience shows instead that bezoars can cause gastric ulcers, small bowel obstruction,
gastrointestinal hemorrhage, and significant discomfort.
Types of Bezoars
In 1938, DeBakey and Ochsner75(5405) proposed three categories of bezoars: phytobezoars (i.e.,
plant bezoars), trichobezoars (i.e., hair bezoars), and phytotrichobezoars. Several newly recognized
types may be added: lactobezoars derived from milk products; concretions composed of medicines,
sand, cement, or shellac; and bezoars composed of intact foreign bodies.
Phytobezoars are the most common of all bezoars.76(5406) Surgery for peptic ulcer with resulting
alterations in gastric anatomy has led to a dramatic increase in prevalence.77(5407) Phytobezoars
occur especially in individuals with impaired gastric emptying, diminished grinding of the antral mill,
loss of pyloric function, or achlorhydria. All of these conditions may occur after vagotomy and
pyloroplasty.78(5408) Patients who have undergone a Billroth type I gastrectomy seem particularly
prone to bezoar formation. Although partial gastrectomy and vagotomy is the most common cause of
gastric bezoars,79(5409) any cause of hypomotility of the stomach can lead to bezoar
formation.80(5410)
Phytobezoars have been reported in patients with insulin-dependent diabetes mellitus, hypothyroidism,
mixed connective tissue disease, idiopathic gastric stasis, myotonic dystrophy,81(5411) gastric cancer,
damage to the vagus by esophageal cancer or esophageal sclerotherapy, or duodenal stenosis and in
those undergoing treatment with anticholinergic drugs.77,8189(5412) Poor mastication of food owing
to defective teeth or dentures, intake of a high-fiber diet, and achlorhydria also may predispose to
bezoars.90(5413)
The mechanism of bezoar formation is best understood for fruit bezoars, particularly those derived
from persimmons. When emptying of the stomach is impaired, gastric acid can promote polymerization
of plant fiber, creating a tannin-cellulose-hemicellulose-protein complex.91(5414) The resulting
phytobezoars usually are compact and somewhat abrasive.
Trichobezoars, or hairballs, were the most common form of bezoar before the advent of modern
surgical techniques. Ingested hair does not readily pass through the pylorus but collects in the
stomach.92(5415) With prolonged trichophagia, or hair ingestion, the retained hair gradually accretes
into a large, dark mass with intermeshed food materials and debris.93(5416) It is characterized by a
fetid odor, probably the product of bacterial colonization.94(5417) Patients may ingest their own hair
and hair from other persons, hair from animals, carpet fibers, and other synthetic or natural
fibers.95(5418) Trichobezoars are also associated with other abnormal intragastric objects, such as
polyps in gastric polyposis, percutaneous endoscopic gastrostomy tubes, and gastric
staples.9698(5419)
Lactobezoars are found primarily in low-birth-weight infants. As advances in neonatal medical science
have improved the survival of these infants, the incidence of lactobezoars has risen. Premature infants
are fed a concentrated formula that is composed partly of milk products. In the neonate, gastric
emptying may be impaired, and dehydration is common. These forces allow the already concentrated
milk products to congeal into a milk bezoar.99(5420)
Bezoars may form as the product of simple concretions in the stomach. Numerous medications have
been implicated in the formation of gastric bezoars. Antacid tablets, aluminum hydroxide gel, Gaviscon,
Kayexalate, cholestyramine, sucralfate, enteric-coated aspirin, slow-release theophylline, lupini beans
(i.e., a homeopathic treatment for arthritis), and long-acting nifedipine (i.e., nifedipine gastrointestinal
therapeutic system) have each been reported as the nidus for bezoars.99105(5421) Shellac bezoars
have been found in alcoholic furniture finishers who drank shellac for its alcohol content, only to have it
congeal in their stomachs.99(5422) Adolescents may ingest glue, rather than sniffing it, and so create
an insoluble mass. Poorly chewed and undigested food, such as pits of fruit, seeds, and nuts, can form
a bezoar. Even concrete and sand have been found as the primary component of gastric
bezoars.106(5423)
Diagnosis
Symptoms and signs of bezoars are often nonspecific and vary with the location of the bezoar. Patients
with gastric bezoars most commonly present with vague epigastric distress (80%); frank epigastric pain
occurs nearly as often (70%).99107(5424) They may complain of dyspepsia, food intolerance,
pyrexia, nausea, or vomiting.108(5425) Gastric ulcer with hemorrhage occurs in as many as 24% of
patients, particularly if the bezoar is the more abrasive phytobezoar.90(5426) Gastric or intestinal
perforation with peritonitis is less frequent. Perforation is often the result of phytobezoars.7578,8088
(5427)and 90109(5428) Gastric bezoars may cause gastric outlet obstruction; the small intestine may
become obstructed by a bezoar. Thirty percent of patients with gastric bezoars experience weight
loss.99(5429) This observation led to the development of the once popular Garren intragastric balloon
for weight loss.99(5430) Despite these many possible symptoms, some patients are entirely
asymptomatic, and their bezoars are discovered only incidentally.90(5431)
Physical signs of gastric bezoar are few but can be helpful if present. In the appropriate setting, the
presence of a gastric mass can be nearly diagnostic. a mass is palpable in 57% of patients with
phytobezoars and in 88% with trichobezoars.107(5432) Some patients have hypochromic, microcytic
anemia, mild leukocytosis, or hypoproteinemia, but laboratory test abnormalities are not found for most
patients.110,111(5433)
Gastroscopy is the most effective method for the diagnosis of bezoars. A phytobezoar appears as a
yellow, green, or brown mass, often with an associated gastric ulcer. Trichobezoars tend to be larger,
black or dark brown, mucoid covered masses that may extend through the full length of the stomach
and into the intestine. The appearance of other bezoars varies with their composition.
Traditionally, the diagnosis of bezoars was made by barium or Gastrografin contrast radiography.
Bezoars may still be discovered incidentally in this way.90(5434) Fluoroscopy may disclose a freely
movable mass in the gastric barium pool; plain x-ray films may show a bubbly mass disrupting the
gastric air shadow.99,112(5435) However, the upper gastrointestinal contrast roentgenography has
proven to be insufficiently sensitive in detecting these varied lesions, missing as many as 55% of all
bezoars.90(5436) Barium contrast x-ray films are not reliable in detecting the gastritis and gastric
ulcers caused by bezoars. Failure to diagnose gastritis or ulcer can have significant consequences
because enzyme therapy for bezoars is contraindicated when these conditions are present.
Alternatively, ultrasonography has been used to establish the diagnosis of bezoars. The bezoar is seen
as an echogenic, solid mass with high amplitude internal echoes.113,114(5437) However,
ultrasonography does not distinguish a phytobezoar from a trichobezoar and, as with barium contrast
roentgenography, will not reliably detect gastritis or an ulcer. When ultrasonography is used as an initial
diagnostic test, it must be followed by endoscopy for a definitive diagnosis.
Treatment
The approach to removal of bezoars varies with their composition and size. Phytobezoars are
effectively treated without surgery in most cases. Simple lavage with saline can be effective, although
the volumes of water required may be prohibitive.115(5438) The addition of organic solvents has been
more successful, using papain, N-acetylcysteine, or cellulase to dissolve proteins, proteinaceous
mucus, or cellulose, respectively. The most effective enzyme is cellulase, which has a reported
dissolution rate of 83 to 100%.99(5439) Metoclopramide has also been administered with some
success.
Endoscopic treatment of phytobezoars is effective. The bezoar can be fragmented in any of several
ways and then removed by lavage of the stomach through a large overtube. Fragmentation of the
bezoar was first performed using uncomplicated techniques and readily available accessories or by
simple suction through the endoscope.
For masses resistant to simple methods, fragmentation may be achieved with jets of water through an
endoscopic catheter. Water pressure in excess of that which can be generated by hand with a syringe
is required. This pressure is provided by attaching the catheter to various devices (e.g., Water
Pik).76,83,116(5440) Water jet fragmentation is highly effective, although it occasionally
fails.117(5441) Alternative methods of endoscopic fragmentation include treatment with the
neodymium:yttrium-aluminum-garnet (Nd:YAG) laser, an endoscopic electric drill, electrohydraulic
lithotripsy, a laser-ignited miniexplosive device, or an endoscopic lithotriptor.78,118122(5442)
Trichobezoars are more resistant to treatment than phytobezoars. Neither lavage nor enzyme therapy
is successful. Endoscopic removal through the repeated and tedious application of Nd:YAG laser
treatment has been reported but has not been universally successful.123(5443) Surgical removal of
the bezoar is the standard approach, although care must be taken to avoid contaminating the
peritoneal cavity with the bacteria-rich material enmeshed in the hairball.75(5444)
Lactobezoars are readily treated if gastric perforation has not occurred. Discontinuation of the
nasogastric feedings together with hydration and, in some cases, gastric lavage can effectively
dissolve the bezoar.
Concretions of medicines can be removed by lavage or by endoscopy with the use of a forceps or
snare. Lavage can precipitate drug toxicity if high doses of the medication are potentially hazardous;
toxicity has been reported with treatment of salicylate and theophylline concretions.100,101(5445)
Concretions that are solid must be removed surgically.
Prevention
After removal of a bezoar, prevention of recurrence is paramount. The physiologic abnormalities that
promoted formation of a phytobezoar persist and predispose the patient to recurrent bezoar formation.
Prophylactic enzyme therapy with each meal has been recommended for disruption of nascent
bezoars.124(5446) Metoclopramide can be used to treat underlying gastric stasis and thereby indirectly
prevent recurrent bezoar formation.125,126(5447) Trichobezoars recur with the resumption of
trichophagia. Patients who persist in this type of ingestion should undergo counseling and be observed
closely in an effort to prevent recurrence.
Gastrointestinal Foreign Bodies in Children

Children compose a large portion of patients who ingest foreign objects, and most of these objects are
true foreign bodies. As in adults, approximately 70% of impactions occur in the proximal third, 20% in
the middle third, and 10% in the distal third of the esophagus.127(5448) After a foreign object has
passed the esophagus, it usually traverses the remainder of the alimentary tract. However, it may also
become lodged at anatomic points of narrowing, including the pylorus, ileocecal valve, Meckel's
diverticulum, and anus. Most ingestions occur in children younger than 5 years of age. Coins are the
most common foreign objects ingested by children.128,129(5449) Other common objects include
bones, plastic or metal toys, buttons, tacks, pins, and ballpoint pen caps. The management of foreign
body ingestion in pediatric patients is discussed in Chapter 53: Esophagogastroduodenoscopy in the
Pediatric Patient.
Button battery ingestion poses a special problem in the pediatric patient (see Chapter 53:
Esophagogastroduodenoscopy in the Pediatric Patient). The alkaline solution in these batteries causes
rapid liquefaction necrosis of gastrointestinal mucosa. Larger batteries (>21 mm in diameter) most
frequently become impacted. Esoph-ageal impaction is a true emergency; endoscopic removal is
optimally accomplished under general anesthesia. After the battery has passed into the stomach, it is
likely to clear the gastrointestinal tract, but daily radiographs are recommended to assess its
progress.54(5450) If the battery remains in the stomach longer than 48 hr or if epigastric pain or
discomfort develops, endoscopic removal is mandatory. After it passes the duodenum, the battery is
likely to traverse the remainder of the gastrointestinal tract. Radiographs every 3 to 4 days can
document its passage.
Intravenous administration of glucagon, a hormone that relaxes gastrointestinal smooth muscle, is
successful in some children with foreign bodies lodged in the middle or distal esophagus because
these portions are composed of smooth muscle. This drug has little or no effect in the presence of
esophageal webs, rings, or strictures (e.g., those resulting from repair of a tracheoesophageal fistula).
The side effects of glucagon are limited to nausea and vomiting. It is contraindicated in patients with
pheochromocytoma, insulinoma, or glucagon allergy. Although success rates range from only 29 to
50%, glucagon is a safe drug and may be used in children before more invasive methods.54(5451)
Controversy exists regarding the use of the Foley catheter in the extraction of esophageal foreign
bodies in children. Campbell and Condon130(5452) reported a series of more than 2500 cases in
which the overall complication rate was 0.4%. Success rates with this technique vary widely, ranging
from 37.5% in one small series to 98% in a larger study.131(5453) Contraindications include airway
obstruction, impaction for longer than 24 hr (2 hr for button batteries), or impaction of an object with
unknown characteristics or one that is not rounded and smooth. Some authorities discourage the use
of this method.54(5454) Potential disadvantages include a lack of airway control, patient discomfort,
exposure to radiation, and an inability to use this technique in patients with preexisting esophageal
structural abnormalities. Its advantages include rapidity, a shortened stay in the hospital, relatively low
cost, and elimination of the need for general anesthesia or intubation. However, the procedure should
only be performed by experienced pediatric radiologists or surgeons, and emergency resuscitation
equipment should always be on hand.
Flexible fiberoptic esophagoscopy is the procedure of choice for removal of most esophageal foreign
bodies in children. For very young children, endotracheal intubation and general anesthesia are
recommended for endoscopic removal of foreign bodies that have been impacted for more than 24 hr,
sharp or pointed objects, button batteries, or objects prone to fragmentation. The perforation rate with
this technique is 0.25%.132(5455) Less serious complications include esophageal laceration and
problems related to general anesthesia or underlying respiratory illness.
Complications
A retained foreign body in the esophagus can cause mucosal edema, ulceration, perforation,
mediastinitis, or fistula formation. Stricture formation is a late complication. The incidence of symptoms
and complications increases with the duration of impaction. Chaikhouni et al.133(5456) found that the
only factor consistently correlated with the occurrence of major complications was duration of
impaction over 24 hr.
After removal of a foreign body, the patient should be carefully observed for fever, chest or abdominal
pain, dyspnea, tachycardia, or crepitation over the neck. Esophageal perforation requires early
diagnosis for management to be successful.
REFERENCES
1.
Davidoff E, Towne JB. Ingested foreign bodies. N Y State J Med 1975;75:1037.

Haglund S, Haverling M, Kuylenstierna R, Lind MG. Radiographic diagnosis of foreign bodies
in the oesophagus. J Laryngol Otol 1978;92:111725.
3. Devanesan J, Pisani A, Sharma P, et al. Metallic foreign bodies in the stomach. Arch Surg
1977;112:6645.
4. Erbes J, Babbitt DP. Foreign bodies in the alimentary tract of infants and children. Appl Ther
1965;7:11039.
5. Bitar DE, Holmes TW Jr. Polybezoar and gastrointestinal foreign bodies in the mentally
retarded. Am Surg 1975;41:497504.
6.
Nandi P, Ong GB. Foreign bodies in the oesophagus: Review of 2,394 cases. Br J Surg
1978;65:69.
7. Vizcarrondo FJ, Brady PG, Nord HJ. Foreign bodies of the upper gastrointestinal tract.
8. Schwartz GE, Polsky HS. Ingested foreign bodies of the gastrointestinal tract. Am Surg
1976;42:2368.
9.
Perelman H. Toothpick perforation of the gastrointestinal tract. J Abdom Surg 1962;4:513.
10. Berggreen PJ, Harrison E, Sanowski RA, et al. Techniques and complications of esophageal
foreign body extraction in children and adults. Gastrointest Endosc 1993;39:62630.
11. Webb WA, McDaniel L, Jones L. Foreign bodies of the upper gastrointestinal tract: Current
2.
management. South Med J 1984;77:10836.

12. Gelfand DW. Complications of gastrointestinal radiologic procedures: I. Complications of
routine fluoroscopic procedures. Gastrointest Radiol 1980;5:293315.
13.
Richardson JR. A new treatment for esophageal obstruction due to meat impaction. Ann Otol
Rhinol Laryngol 1945;54:32848.
14.
Holsinger JW Jr, Fuson RL, Sealy WC. Esophageal perforation following meat impaction and
papain ingestion. JAMA 1968;203:7345.
15.
Zarling EJ, Moeller DD. Bezoar therapy: Complications using Adolph's Meat Tenderizer and
alternatives from literature review. Arch Intern Med 1980;141:166970.
16. Goldner F, Danley D. Enzymatic digestion of esophageal meat impaction. A study of Adolph's
meat tenderizer. Dig Dis Sci 1985;30:4569.
17.
Ferrucci JT, Long JA Jr. Radiologic treatment of esophageal food impaction using intravenous
glucagon. Radiology 1977;125:238.
18. Rogers BHG, Kot C, Meiri S, et al. An overtube for the flexible fiberoptic
esophagogastroduodenoscope. Gastrointest Endosc 1982;28:2567.
19.
Haugen RK. The cafe coronary: Sudden death in restaurants. JAMA 1963;186:142.
20. Heimlich HJ. A life-saving maneuver to prevent food-choking. JAMA 1975;234:398401.
21. Bintz M, Cogbill TH. Gastric rupture after the Heimlich maneuver. J Trauma 1996;40:15960.
22. Dupre MW, Silva E, Brotman S. Traumatic rupture of the stomach secondary to Heimlich
maneuver. Am J Emerg Med 1993;11:6112.
23. van der Ham AC, Lange JF. Traumatic rupture of the stomach after Heimlich maneuver. J
Emerg Med 1990;8:7135.
24. Cowan M, Bardole J, Dlesk A. Perforated stomach following the Heimlich maneuver. Am J
Emerg Med 1987;5:1212.
25. Croom DW. Rupture of stomach after attempted Heimlich maneuver. JAMA
1983;250:26023.
26. Visintine RE, Baick CH. Ruptured stomach after Heimlich maneuver. JAMA 1975;234:415.
27. Haynes DE, Haynes BE, Yong YV. Esophageal rupture complicating Heimlich maneuver. Am
J Emerg Med 1984;2:5079.
28. Razaboni RM, Brathwaite CE, Dwyer WA Jr. Ruptured jejunum following Heimlich maneuver.
J Emerg Med 1986;4:958.
29. Passik CS, Ackermann DM, Piehler JM, Edwards WD. Traumatic rupture of Ionescu-Shiley
aortic valve after the Heimlich maneuver. Arch Pathol Lab Med 1987;111:46970.
30. Kirshner RL, Green RM. Acute thrombosis of abdominal aortic aneurysm subsequent to
Heimlich maneuver: A case report. J Vasc Surg 1985;2:5946.
31. Chapman JH, Menapace FJ, Howell RR. Ruptured aortic valve cusp: A complication of the
Heimlich maneuver. Ann Emerg Med 1983;12:4468.
32. Roehm EF, Twiest MW, Williams RC Jr. Abdominal aortic thrombosis in association with an
attempted Heimlich maneuver. JAMA 1983;249:11867.
33. Rich GH. Pneumomediastinum following the Heimlich maneuver. Ann Emerg Med
1980;9:27980.
34.
Agia GA, Hurst DJ. Pneumomediastinum following the Heimlich maneuver. J Am Coll Emerg
Physicians 1979;8:4735.
35. Fink JA, Klein RL. Complications of the Heimlich maneuver. J Pediatr Surg 1989;24:4867.
36.
Rsch W, Classen M. Fiber endoscopic foreign body removal from the upper gastrointestinal
37. Witzel L, Scheurer U, Muhlemann A, Halter F. Removal of razor blades from stomach with
fibreoptic endoscope. Br Med J 1974;2:539.
38. Bertoni G, Pacchione D, Sassatelli R, et al. A new protector device for safe endoscopic
removal of sharp gastroesophageal foreign bodies in infants. J Pediatr Gastroenterol Nutr
1993;16:3936.
39. Altman AR, Gottfried EB. Intragastric closure of an ingested open safety pin. Gastrointest
Endosc 1978;24:2945.
40. McCarron MM, Wood JD. The cocaine "body packer" syndrome. Diagnosis and treatment.
JAMA 1983;250:141720.
41. Hartoko TJ, Demey HE, DeSchepper AM. The body packer syndromeCocaine smuggling in
the gastrointestinal tract. Klin Wochenschr 1988;66:111620.
42. Jonsson S, O'Mearu M, Young JB. Acute cocaine poisoning. Importance of treating seizures
and acidosis. Am J Med 1983;75:10614.
43. Carvana DS, Weinbach B, Goerg D, et al. Cocaine-packet ingestion: Diagnosis, management
and natural history. Ann Intern Med 1984;100:734.
44.
Gherardi RK, Baud FJ, Leporc P, et al. Detection of drugs in the urine of body-packers.
Lancet 1988;1:10268.
45. Suarez CA, Aranjo A, Lester JL III. Cocaine-condom ingestion: Surgical treatment. JAMA
1977;238:13912.
46. Litovitz TL. Button battery ingestions. A review of 56 cases. JAMA 1983;249:2495500.
47. Kulig K, Rumack CM, Rumack BH, et al. Disk battery ingestion, elevated urine mercury levels
and enema removal of battery fragments. JAMA 1983;249:25024.
48. David TJ, Ferguson AP. Management of children who have swallowed button batteries. Arch
Dis Child 1986;61:3212.
49. Kost KM, Shapiro RS. Button battery ingestion: A case report and review of the literature. J
Otolaryngol 1987;16:2527.
50. Votteler TP, Nash JC, Rutledge JC. The hazard of ingested alkaline disk batteries in children.
JAMA 1983;249:25046.
51. Mant TG, Lewis JL, Mattoo TK, et al. Mercury poisoning after disc-battery ingestion. Hum
Toxicol 1987;6:17981.
52. Blatnik DS, Toohill RJ, Lehman RH. Fatal complication from an alkaline battery foreign body
in the esophagus. Ann Otol Rhinol Laryngol 1977;86:6115.
53. Studley JGW, Linehan IP, Ogilvie AL, et al. Swallowed button batteries: Is there a consensus
on management?. Gut 1990;31:86770.
54. Webb WA. Management of foreign bodies of the upper gastrointestinal tract.
55. Litovitz TL. Battery ingestions: Product accessibility and clinical course. Pediatrics
1985;75:46976.
56. Kirk AD, Bowers BA, Moylan JA, et al. Toothbrush swallowing. Arch Surg 1988;123:3824.
57. Riddlesberger MM Jr, Cohen HL, Glick PL. The swallowed toothbrush: A radiographic clue of
bulimia. Pediatr Radiol 1991;21:2624.
58. Mughal M. Accidental ingestion of a toothbrush. Arch Emerg Med 1986;3:11923.
59. Mirich DR, Gray RR, Haber GB. Percutaneous deflation of a gastric balloon: Technical note.
Cardiovasc Intervent Radiol 1989;12:1645.
60. Subramanyam K, Robbins HT. Erosion of Marlex band and plastic ring into the stomach after
gastroplasty: Endoscopic recognition and management. Am J Gastroenterol
1989;84:131921.
61. Larson JV. Wing nut obstruction. Gastrointest Endosc 1989;35:68.
62.
Bendig D, Mackie GG. Management of smooth-blunt gastric foreign bodies in asymptomatic
patients. Clin Pediatr 1990;29:6435.
63. Jantasch H, Lechner G, Wittich GR, et al. Dislocated Atkinson tubes: Removal and
repositioning with a balloon catheter. Radiology 1989;170:8856.
64. Massaioli N, Bertero D, Buzio M, et al. Endogastric migration of an Angelchik prosthesis. A
case report and review of the literature. Minerva Chir 1990;45:18994.
65. Elewaut A, De Vos M, Huble F, et al. Unusual migration of a straight Amsterdam-type
endoprosthesis for bile duct stones. Am J Gastroenterol 1989;84:6746.
66. Lowe GM, Bernfield JB, Smith CS, et al. Gastric pneumatosis: Sign of biliary stent-related
perforation. Radiology 1990;174;10378.
67.
Dresing K, Lottner C, Stock W. Port-catheter perforation into the duodenum and other early
complications after port implantation before intra-arterial infusion therapy of the liver with
chemotherapeutic drugs. Med Klin 1991;15:24550.

68. Mizrahi S, Eyal I, Shtamler B. Foreign body removal through an appendicostomy. Dis Colon
Rectum 1990;33:902.
69. Dharamsi RD, Jesudason SR, Rolston DD. Chronic watery diarrhea due to a surgical pack in
the ileal lumen. J Clin Gastroenterol 1990;12:23941.
70. Zabala S, Insignares E, Miguelena JM, et al. Intestinal obstruction caused by foreign body in a
patient with colon neoplasm. Rev Esp Enferm Dig 1991;79:35960.
71. McGovern R, Barkin JS, Goldberg RI, Phillips RS. Duodenal obstruction: A complication of
percutaneous endoscopic gastrostomy tube migration. Am J Gastroenterol 1990;85:10378.
72. O'Dell KB, Gordon RS, Becker LB. Gastrostomy tube transmigration: A rare cause of small
bowel obstruction. Ann Emerg Med 1991;20:8179.
73. Pereira MC, Mersih K. Foley catheter gastrostomy tube migration: Small bowel obstruction
relieved by percutaneous balloon aspiration. Gastrointest Endosc 1991;37:3724.
74. Goldstein SS, Lewis JH, Rothstein R. Intestinal obstruction due to bezoars. Am J
75.
DeBakey M, Ochsner A. Bezoars and concretions. Surgery 1938;4:93463.
76. Lange V. Gastric phytobezoar: An endoscopic technique for removal. Endoscopy
1986;18:1956.
77. Tsianos EB, Drosos AA. Gastroscopic removal of a phytobezoar formed in a patient with
mixed connective tissue disease. Endoscopy 1988;20:1223.
78. Lubke HJ, Winkelmann RS, Berges W, et al. Gastric phytobezoar: Endoscopic removal using
the gallstone lithotriptor. Z Gastroenterol 1988;26:3936.
79. Robles R, Parrilla P, Escamilla C, et al. Gastrointestinal bezoars. Br J Surg 1994;81:10001.
80. Brady PG. Gastric phytobezoars consequent to delayed gastric emptying. Gastrointest
Endosc 1978;24:15961.
81. Brady PG, Richardson R. Gastric bezoar formation secondary to gastroparesis diabeticorum.
82. Allred-Crouch AL, Young EA. BezoarsWhen the "knot in the stomach" is real. Postgrad Med
1985;78:2615.
83. Madsen R, Skibba RM, Galvan A, et al. Gastric bezoars. A technique of endoscopic removal.
Am J Dig Dis 1978;23:7179.
84. Hayes PG, Rotstein OD. Gastrointestinal phytobezoars: Presentation and management. Can
J Surg 1986;29:41920.
85. Kuiper D. Gastric bezoar in a patient with myotonic dystrophy. Am J Dig Dis 1971;16:52934.
86. Van Thiel DH, De Belle RC, Painter TB, et al. Phytobezoar occurring as a complication of
gastric carcinoma. Gastroenterology 1975;68:12926.
87. Davion T, Delamarre J, Reix N, et al. Gastric bezoar: Another side effect of endoscopic
variceal sclerotherapy. Scand J Gastroenterol 1989;24:81820.
88.
Edell S, Wagner DK. Duodenal stenosis contributing to bezoar formation. Clin Pediatr
1971;10:5435.
89. Ahn YH, Maturu P, Steinheber FU, Goldman JM. Association of diabetes mellitus with gastric
bezoar formation. Arch Intern Med 1987;147:5278.
90. Stein DT, Ballin R, Stone BG. Endoscopic removal of gastric phytobezoars. J Clin
91. Holloway WD, Lee SP, Nicholson GI. The composition and dissolution of phytobezoars. Arch
Pathol Lab Med 1980;104:15961.
92.
Davies I. Hairballs or hair casts of stomach and gastrointestinal tract. Lancet 1921;2:791.
93. Sanowski RA, DiBianco J Jr. Pseudotumors of the stomach. Am J Dig Dis 1966;11:60714.
94. Bacer D, D'Angio GJ, Evans AE. Proceedings of the Tumor Board of the Children's Hospital
of Philadelphia. Trichobezoar. Med Pediatr Oncol 1988;16:3413.
95. Deslypere JP, Praet M, Verdonk G. An unusual case of the trichobezoar: The Rapunzel
syndrome. Am J Gastroenterol 1982;77:46770.
96. Bates CK, Keefe EB, Sahin A. Trichobezoar associated with gastric polyposis. West J Med
1988;149:2203.
97. Reeves-Darby V, Halpert R. Gastric bezoar complicating gastric stapling. Am J Gastroenterol
1990;85:3267.
98. Evans JR, Mitchell RD. Tissue paper bezoar associated with percutaneously placed
gastrostomy tube. Gastrointest Endosc 1988;34:1424.
99. Andrus CH, Ponsky JL. Bezoars: Classification, pathophysiology, and treatment. Am J
100.
Cereda J, Scott J, Quigley EMM. Endoscopic removal of pharmacobezoar of slow release
theophylline. Br Med J 1986;29:369.
101. Bogacz K, Caldron P. Enteric-coated aspirin bezoar: Elevation of serum salicylate level by
barium study. Am J Med 1987;83:7836.
102. Prisant LM, Carr AA, Bottini PB, et al. Nifedipine GITS (gastrointestinal therapeutic system)
bezoar. Arch Intern Med 1991;151:18689.
103. Reddy AN. Sucralfate gastric bezoar. Am J Gastroenterol 1986;81:14950.
104. Quigley MA, Flicker M, Caldwell EG. Sucralfate bezoarTheory or fact?. Am J Gastroenterol
1986;81:7245.
105. Meeroff JC, Squires PA, Grasman K. Gastric lupinoma: A new variety of phytobezoar. Am J
106.
Visvanathan R. Cement bezoars in the stomach. Br J Surg 1986;33:1234.
107.
Raffin SB. Bezoars. In Schlesinger MH, Fordtran JS (eds). Gastrointestinal Disease. 4th ed.
Vol I. Philadelphia: WB Saunders, 1989;7415.
108. Amjad H, Kumar GK, McCaughey R. Postgastrectomy bezoars. Am J Gastroenterol
1975;64:32731.
109. Mangold D, Woolam GL, Garcia-Rinaldi R. Intestinal obstruction due to phytobezoars. Arch
Surg 1978;113:10013.
110. Bernstein LH, Gutstein S, Efron G, et al. TrichobezoarAn unusual cause of megaloblastic
anemia and hypoproteinemia in childhood. Am J Dig Dis 1973;18:6771.
111. Hossenbocus A, Colin-Jones D. Trichobezoar, gastric polyposis, protein-losing
gastroenteropathy and steatorrhea. Gut 1973;14:7302.
112. Rigler RG, Grininger DR. Phytobezoars following partial gastrectomy. Surg Clin North Am
1970;50:3816.
113. Bidula MM, Rifkin MD, McCoy RI. Ultrasonography of gastric phytobezoar. JCU J Clin
Ultrasound 1986;14:49.
114. McCracken S, Jongeward R, Silver TM, et al. Gastric trichobezoar: Sonographic findings.
Radiology 1986;161:1234.
115. Miller G, Clemencon G. Vagotomie und Magenbezoare. Dtsch Med Wochenschr
1973;98:19656.
116. Rider JA, Foresti-Lorente RF, Garrido J, et al. Gastric bezoars: Treatment and prevention. Am
117. Van Gossum A, Delhaye M, Cremer M. Failure of non-surgical procedures to treat gastric
trichobezoar. Endoscopy 1989;21:113.
118. Chen GH. Removal of bezoars from stomach using an endoscopic drill. Gastrointest Endosc
1985;31:355.
119. Ying-cai H, Zhong-he G, Ying G, et al. Endoscopic lithotripsy of gastric bezoars using a
laser-ignited mini-explosive device. Chin Med J 1990;103:1525.
120. Naveau S, Poynard T, Zourabichvili O, et al. Gastric phytobezoar destruction by Nd:YAG laser
therapy. Gastrointest Endosc 1986;32:4301.
121. Huang YC, Guo ZH, Gu Y, et al. Endoscopic lithotripsy of gastric bezoars using a laser-ignited
mini-explosive device. Chin Med J 1990;103:1525.
122. Kuo JY, Mo LR, Tsai CC, et al. Endoscopic fragmentation of gastric phytobezoar by
electrohydraulic lithotripsy. Gastrointest Endosc 1993;39:7068.
123. Soehendra N. Endoscopic removal of a trichobezoar. Endoscopy 1989;21:201.
124. Klamer TW, Max MH. Recurrent gastric bezoars. A new approach to treatment and
prevention. Am J Surg 1983;145:4179.

125. Winkler WP, Saleh J. Metoclopramide in the treatment of gastric bezoars. Am J Gastroenterol
1983;78:4035.
126. Delpre G, Kadish U, Glanz I. Metoclopramide in the treatment of gastric bezoars. Am J
127. Ratcliff KM. Esophageal foreign bodies. Am Fam Physician 1991;44:82431.
128. Crysdale WS, Sendi KS, Yoo J. Esophageal foreign bodies in children: 15 year review of 484
cases. Ann Otol Rhinol Laryngol 1991;100:3204.
129. Baraka A, Bikhazi G. Oesophageal foreign bodies. Br Med J 1975;1:5613.
130. Campbell JB, Condon VR. Catheter removal of blunt esophageal foreign bodies in children.
Survey of the Society for Pediatric Radiology. Pediatr Radiol 1989;19:3615.
131. Mariani PJ, Wagner DK. Foley catheter extraction of blunt esophageal foreign bodies. J
Emerg Med 1986;4:3016.
132. Jackson RM, Hawkins DB. Coins in the esophagus. What is the best management?. Int J
Pediatr Otorhinolaryngol 1986;12:12735.
133. Chaikhouni A, Kratz JM, Crawford FA. Foreign bodies of the esophagus. Am Surg
1985;51:1739.
Chapter 55 Percutaneous Endoscopic Gastrostomy

(5457)
RICHARD C. K. WONG, M.B., B.S., F.A.C.P.
JEFFREY L. PONSKY, M.D., F.A.C.S.
Nearly 2 decades have elapsed since the original description by Gauderer et al.1(5458) of the
endoscopically guided technique, now known as percutaneous endoscopic gastrostomy (PEG), for
creating a tube gastrostomy for enteral feeding. Since then, many centers throughout the world have
published favorable results with this technical concept, which has been used successfully in children
and infants,15(5459) adults,617(5460) and the elderly.18(5461) During the years since its
introduction, numerous reports have described modifications of the procedure as well as complications
that may be expected and their management.1922(5462) Advances in technology and design have
also led to suggestions for improvements in the materials used in the performance of the
procedure.23(5463)
The essential role of enteral nutrition in maintaining gastrointestinal integrity and mucosal defense
against external pathogens has become more widely recognized since the mid-1980s. For instance, it
has been shown in mice that progressive enteral protein malnutrition results in atrophy of the small
intestinal and cecal mucosa with reduction in both the density and the height of the intestinal villi. This
becomes most pronounced at 21 days.24(5464) Furthermore, bacterial translocation across the gut
wall in response to an inflammatory stimulus was limited to the mesenteric lymph nodes in enterally fed
mice, whereas lethal systemic sepsis occurred in protein-malnourished animals.25(5465) The scientific
basis for this observation is probably multifactorial, but one component may be a reduction in levels of
secretory immunoglobulin A (IgA) seen in parenterally fed (but enterally malnourished) rats.26(5466)
This is important because secretory IgA plays a major role in the maintenance of the gut mucosal
defense system. For trauma patients, a metaanalysis has demonstrated that early enteral nutrition
(compared with only parenteral) significantly reduces postoperative septic complications.27(5467)
Thus, in most patients, it is recommended that enteral nutrition be started after 1 to 2 weeks without
nutrient intake.28(5468)
Indications
The original objective of the PEG procedure was to provide an effective means for long-term enteral
alimentation in patients who were unable to swallow.1(5469) This remains the major indication for
placement of a tube gastrostomy for patients with a functional gastrointestinal tract who are unable or
unwilling to maintain an oral intake sufficient for their caloric requirements. In general, PEG should be
considered only in patients who are thought to require long-term enteral tube feedings. Nasogastric or
nasoenteric tubes are preferred for short-term management. Long-term has been defined by the
American Gastroenterological Association as more than 30 days.28(5470) Such patients may have
severe neurologic or developmental impairment,2933(5471) mechanical obstruction of the upper
aerodigestive tract,34(5472) severe facial trauma,2(5473) or may have undergone maxillofacial tumor
surgery.35(5474) The placement of such feeding tubes has also been reported in burn patients,
through either burn wound areas or donor sites.36(5475) Many investigators have recommended early
PEG tube placement in patients with head and neck cancers.3742(5476) However, implantation
metastases at the gastrostomy insertion site have been reported, albeit rarely.43(5477) Long-term
enteral hyperalimentation via PEG tubes has also proved useful in children with growth failure due to
Crohn's disease or cystic fibrosis.44,45(5478) The technique of combined tracheostomy and PEG has
also been reported for selected patients who have long-term requirements for both airway protection
and enteral nutrition.46,47(5479) PEG tubes have also been used successfully in the treatment by
enteral alimentation of the wasting syndrome seen in patients with acquired immunodeficiency
syndrome (AIDS).48,49(5480) However, it is not known whether improvement of AIDS-associated
wasting lengthens patient survival. Long-term nutritional support of a comatose pregnant woman until
delivery of a normal infant at 37 weeks' gestation has been reported.50(5481)
Non-alimentation-related applications of PEG have been described more recently. These include
chronic enteral administration of unpalatable medications or diets,22(5482) bile replacement in cases
of external biliary fistula,51,52(5483) and gastric decompression in patients with carcinomatosis,
gastric atony, or recalcitrant intestinal obstruction.5357(5484) Other novel uses have included
nonsurgical treatment of the gas-bloat syndrome after Nissen fundoplication,58(5485) facilitated
dilation and stenting of obstructing esophageal neoplasms,59(5486) therapeutic endoscopic retrograde
cholangiopancreatography with biliary stent placement,60(5487) and fixation of the stomach in patients
with recurrent gastric volvulus.61(5488)
Contraindications
Standard PEG is contraindicated whenever the anterior abdominal wall cannot be brought into close
proximity with the anterior gastric wall. This can occur when there is significant ascites, morbid obesity,
prominent enlargement of the left lobe of the liver, or splenomegaly. Indeed, one study found that
cancer patients with ascites are at increased risk of complications (80%) and death (20%) directly
related to the standard PEG procedure.62(5489) In certain instances, some of the aforementioned
contraindications may be considered relative if the procedure is performed with a small surgical
incision63(5490) or an additional imaging modality such as laparoscopy,64(5491) transcutaneous
ultrasound,65(5492) endoscopic ultrasound,66(5493) or computed tomography.67(5494) PEG is
contraindicated in patients with gastric varices, uncorrectable coagulopathy, a nonfunctional
gastrointestinal tract, and those undergoing chronic peritoneal dialysis. Portal hypertensive
(congestive) gastropathy represents a relative contraindication to PEG.68(5495)
Mechanical gastrointestinal obstruction is a contraindication to PEG unless the gastrostomy is for
decompressive purposes, although obstructed loops of bowel may increase the technical difficulty of
the procedure. It has been suggested in one anecdotal report that subcutaneous administration of
octreotide may reduce the degree of obstruction by decreasing the volume of gastrointestinal
secretions before gastrostomy tube placement.69(5496)
Previous abdominal surgery is not an absolute contraindication to PEG if there is meticulous
adherence to details of proper technique.70(5497) Indeed, PEG has been successfully performed in
the early postoperative period (i.e., within 14 days of surgery) in patients with unexpected surgical
complications that necessitate long-term enteral nutrition.71(5498) Identification of a "safe tract" by
aspiration of air from the stomach with a syringe and needle may be considered in patients with prior
abdominal surgery.72(5499) Certainly, a previous subtotal or total gastrectomy suggests that the
procedure may be impossible to perform.
The presence of a ventriculoperitoneal shunt does not preclude placement of a PEG. However, the site
of the shunt should be carefully avoided.73(5500) Data from one retrospective study suggest that a
PEG can be safely placed within 1 month of acute myocardial infarction in otherwise stable patients
when enteral feeding is considered absolutely essential.74(5501) The most important absolute
contraindications to the use of a feeding gastrostomy tube are short life expectancy as a result of
severe disease and any condition that makes it unlikely that a patient will benefit from enteral feeding.
Technique
The patient is prepared by withholding feedings for 8 hours before the procedure. A single prophylactic
dose of an antibiotic, usually a cephalosporin, is administered intravenously just before the
procedure.75(5502) However, there is controversy regarding the choice of antibiotic and number of
doses needed to effect adequate prophylaxis against infection. Data from one study suggest that
administration of wider-spectrum antibiotics for a longer period of time may be required.76(5503) The
original "pull" method of PEG tube placement is the most widely used.77(5504) Two modified
techniques, the "push"78(5505) and the "introducer"79(5506) methods, have also become popular. A
single-step button gastrostomy has also been developed.80(5507) Regardless of the technique for
PEG tube placement, a thorough diagnostic esophagogastroduodenoscopy (EGD) should be
performed initially because this may reveal significant upper gastrointestinal abnormalities.81(5508) In
one study, the findings at the initial EGD led to major changes in subsequent management in 36% of
patients.82(5509)
The "Pull" Method

The patient is placed in the supine position on the endoscopy table. The posterior pharynx is
anesthetized with a topical agent and then swabbed with an antiseptic solution. Also, sedative drugs
may be administered intravenously. The abdomen is prepared and draped in a sterile fashion.
Introduction of the endoscope may occasionally be difficult with the patient supine and may be
facilitated, if necessary, by rolling the patient to his or her side for passage of the instrument. After
introduction of the endoscope, the upper tract is carefully surveyed, and the stomach is fully inflated.
The latter measure displaces the liver in a cephalad direction and the colon toward the pelvis. Also, as
a result of maximum distention, the anterior gastric wall comes to lie in contact with the peritoneal
surface of the anterior abdominal wall. The endoscope should be positioned proximal to the gastric
angulus, if possible, as puncture of the antrum may damage the gastric pacemaker, resulting in atony.
As the endoscopist manipulates the instrument in the body of the stomach, the room lights are
dimmed, and the patient's abdominal wall is observed for transillumination by the endoscope light.
Transillumination ensures that the stomach is in contact with the abdominal wall and that there are no
intervening viscera. The site of brightest transillumination should be chosen for placement of the
gastrostomy. It is best to avoid close proximity to the rib cage or xiphoid process, if possible. When
finger pressure is applied to the abdominal wall at the selected site, the endoscopist should observe a
definite indentation in the anterior gastric wall (Figure 551). A polypectomy snare is opened over this
area in preparation for capture of the puncturing needle. The skin at the selected point is anesthetized
with a local anesthetic, and a skin incision of approximately 1 cm is made. The puncturing cannula is
then thrust through the incision into the gastric lumen. The snare loop is tightened around the cannula,
and the inner stylet of the cannula is removed. A good length of suture or looped wire is passed
through the cannula into the stomach, and the snare is selectively tightened around this material
(Figure 552). The suture is then pulled, with the gastroscope, from the patient's mouth (Figure 553).
(5510)Figure 551. Finger pressure at the site selected for puncture should produce a clear
indentation of the gastric wall. This area is surrounded with a snare. (From Ponsky JL. Atlas
of Surgical Endoscopy. St Louis: Mosby-Year Book, 1992.)
(5511)Figure 552. After the stomach is punctured, the snare is tightened around the needle
and the inner stylet is removed. A suture or wire is passed through the needle into the gastric
lumen. (From Ponsky JL. Atlas of Surgical Endoscopy. St Louis: Mosby-Year Book, 1992.)
(5512)Figure 553. The snare is repositioned and tightened selectively on the suture itself.
The suture is then pulled, with the endoscope, from the patient's mouth. (From Ponsky JL.
Atlas of Surgical Endoscopy. St Louis: Mosby-Year Book, 1992.)
After fixation of the suture to the dilating end of the gastrostomy tube, the complex is well lubricated
and pulled (hence, "pull" technique) by traction at the abdominal end of the suture down through the
esophagus and into the stomach. After several inches of the tube have exited the abdomen, the
gastroscope is reinserted to observe the seating of the head of the gastrostomy tube against the
gastric wall (Figure 554). Contact between the head of the tube and the gastric mucosa should be
without tension. The gastroscope is removed, and an outer crossbar is applied to prevent migration of
the tube. This crossbar must not be placed too tightly against the skin, as this will result in ischemia
and necrosis of the underlying tissue (Figure 555). A distance of several millimeters between the skin
and the crossbar is optimum and will not impair gastrostomy tract formation.83(5513) It has been
suggested that the second insertion of the endoscope be omitted, provided abdominal wall thickness
does not exceed 6 cm as determined by marks on the gastrostomy tube.84(5514) However, the only
certain method for confirmation of optimum head (or internal bumper) positioning is endoscopic. The
use of stabilizing sutures and nylon T-fasteners has also been suggested, although these are not
usually required.85,86(5515)
(5516)Figure 554. The tube is affixed to the suture and pulled into the esophagus. The
gastroscope is reintroduced to follow the tube, and traction is continued until the head of the
catheter just meets the gastric wall. (From Ponsky JL. Atlas of Surgical Endoscopy. St Louis:
Mosby-Year Book, 1992.)
(5517)Figure 555. An outer crossbar is applied to prevent migration of the tube. The
crossbar should rest a millimeter or two from the skin to avoid producing ischemia and
necrosis of the underlying tissue.
The "Push" Method

This modification of PEG was first described by Sacks et al.78(5518) The technique is exactly the
same as the pull method, except that a sturdier wire is passed into the stomach instead of a flexible
suture. Once retrieved from the patient's mouth, a gastrostomy tube with a rigid dilator end is threaded
over the wire. Both ends of the wire are held taut as the tube is pushed (hence, "push" technique) over
the wire into the stomach (Figure 556). As the dilator end of the tube emerges from the abdominal
wall, it is grasped and pulled the rest of the way to its final position. Results of this method are similar
to those of the pull technique.87,88(5519)
(5520)Figure 556. The guidewire is held taut while the tube is pushed over it and emerges
from the abdominal wall. (From Ponsky JL. Percutaneous endoscopic stomas. Surg Clin North
Am 1989; 69[6].)
The "Introducer" Method

This technique, described by Russell et al.79,89(5521) establishes an endoscopically guided
gastrostomy using principles common to the insertion of central venous lines and cardiac pacemakers.
A modification of this technique has also been used successfully in infants and children.90(5522) Once
the site for gastrostomy placement has been chosen, a needle is thrust into the stomach under
endoscopic visualization. A short guidewire is passed through the needle, and the needle is removed.
An introducer with an outer sheath is passed over the wire with a twisting motion until the sheath is
clearly visible within the stomach (Figure 557). The introducer is withdrawn, leaving only the sheath in
the stomach, through which is passed a balloon-tipped gastrostomy tube. The balloon is inflated under
endoscopic visualization, and the sheath is peeled away (Figure 558). The tube is sutured to the skin,
completing the procedure.
(5523)Figure 557. Under endoscopic visualization, the introducer and sheath are passed
over the guidewire into the stomach. (From Ponsky JL. Percutaneous endoscopic stomas. Surg
Clin North Am 1989; 69[6].)
(5524)Figure 558. A balloon catheter is inserted through the sheath and inflated. The sheath
is then stripped away. (From Ponsky JL. Percutaneous endoscopic stomas. Surg Clin North
Am 1989; 69[6].)
Early results with the introducer technique suggested a somewhat higher incidence of life-threatening
complications compared with the other methods.91(5525) This was due in part to the instability of the
balloon catheter in the stomach. Deflation often occurred, leading to premature separation of the
gastric and abdominal walls. This problem may be obviated by newer tube designs. In addition, a larger
(20-French) introducer is now available, which allows placement of an 18-French gastrostomy
tube.92(5526)
The "introducer" technique requires only one pass of the endoscope and hence is valuable in patients
with tight stenotic lesions of the esophagus that make repeated passage of the endoscope and
gastrostomy tube ill advised. Furthermore, there is the theoretical advantage that the gastrostomy tube
does not come into contact with the bacterial flora of the upper aerodigestive tract. However, this
method may be technically more difficult than the push method.93(5527)
The Single-Step Button Method

The gastrostomy button is a skin-level, continent gastrostomy device of variable shaft length. It was
introduced in the 1980s as a PEG tube replacement device for use after a mature gastrocutaneous
tract has formed (hence, a two-stage procedure).94,95(5528) Since then, it has been further
developed and marketed for use as a single- or one-step procedure.80(5529) The technique of
placement is similar to that of the push method (see earlier). In essence, this involves pushing the
button device (enclosed in a silicone tube) over a stiff guidewire, through the patient's mouth, into the
stomach, and through the anterior abdominal wall. However, there is one additional step of measuring
the tissue thickness at the tube exit site just before inserting the apparatus, so that a button of
appropriate shaft length can be selected. At the end of the procedure, the button is deployed by
releasing it from the dilating portion of the device. Accurate intragastric placement can be confirmed
endoscopically.
Despite the advantages of a skin-level device, the largest prospective multicenter evaluation to date
(with 86 patients) of a single-step button reported serious problems with placement and a high 90-day
complication rate.96(5530) Nevertheless, prior studies have reported satisfactory outcomes in
adults97(5531) as well as infants and children.98(5532) Furthermore, laparoscopically guided
techniques have been described for placing a single-step gastrostomy button.99,100(5533) Further
evaluation of these new techniques in prospective, randomized, comparative trials is needed before
widespread usage can be recommended.
Feeding and Local Care

Tube feeding and oral intake of fluid or food should not be permitted until the patient is examined. In
particular, the abdomen should be examined for tenderness, distention, and bowel sounds. The
gastrostomy tube insertion site should be closely inspected for signs of infection, and the outer
crossbar should be loosened if it is too tight. The latter is especially important because local
complications are more common if excessive traction is applied to the tube.101(5534) Local skin care
can be maintained by cleaning the tube insertion site daily with soap and water. Also, exposure to the
air is often helpful. An occlusive dressing is not recommended, although a simple dressing can be
applied to the tube site daily, if needed.102(5535) Hydrogen peroxide (on a cotton swab) can be used
to periodically remove the buildup of debris around the insertion site.103(5536)
Traditionally, enteral gastrostomy tube feedings were begun 24 hr after PEG tube placement. However,
more recent studies comparing early tube feedings (3 to 6 hr after placement) with delayed feedings
(24 hr after placement) have shown no differences in complication rates.104106(5537) Tube feedings
can be administered by rapid bolus. More commonly, they are given intermittently or continuously by
gravity feedings or by means of a mechanical pump. Feeding must not be initiated until bowel sounds
are heard, and initial feedings (after tube placement) should be by continuous intragastric infusion.
Once this is tolerated at the optimum infusion rate, the regimen can be changed to intermittent or
bolus. The particular feeding regimen will depend on the individual patient's overall state of health and
activity as well as the functionality of the gastrointestinal tract. For instance, the rapid-bolus method
may be preferable for the alert and active patient, whereas continuous intragastric feedings may be
more suitable for patients (the majority) who are at risk for aspiration.103,107(5538)
To reduce the risk of aspiration during tube feeding, the head of the patient's bed should be raised
between 30 and 45 degrees during feeding and for 1 hr afterward. One suggested regimen of
continuous tube feeding involves the initial infusion of water at 50 ml/hr for 4 hr. If this is tolerated, an
isotonic full-strength polymeric formula is begun at 50 ml/hr, with incremental advancements of 25 ml
every 12 hr as tolerated.108(5539) It is often beneficial that a dietitian and a nurse see both the patient
and the caregiver after placement of the gastrostomy tube to provide advice on appropriate feeding
and the general care of the tube. Instructions should also be provided as to when to call the physician
should problems arise. The use of a multidisciplinary nutrition support team to manage enterally fed
patients may reduce complications and be more effective in attaining nutritional goals.109111(5540)
Aspiration is an important, common, and controversial complication encountered during the
subsequent care of the tube-fed patient. Its incidence may actually vary over time owing to alterations
in the underlying disease processes, the feeding regimen, or the body posture of the patient.108(5541)
There have been studies purporting to show that PEG actually reduces the incidence of
gastro-esophageal reflux.112,113(5542) However, data to the contrary also exist.114116(5543) A
distinction must be made between oropharyngeal and gastric aspiration. Some authors have
suggested that all patients be evaluated for gastro-esophageal reflux before a decision is made
concerning placement of a gastrostomy tube.3(5544) Others have proposed a 4- to 6-week trial period
of nasogastric feedings, at least in the pediatric population.2(5545) Still others have advocated
simultaneous laparoscopic Nissen fundoplication and PEG in children with documented preprocedure
gastroesophageal reflux.117(5546) In addition, the potential role for prokinetic agents such as
cisapride is yet to be determined.118(5547)
Percutaneous Endoscopic Gastrostomy Tube Replacement

PEG tubes should be removed when they are no longer needed, are causing complications, or require
replacement.119(5548) To avoid the need for a repeat gastrostomy, the physician must be satisfied
that the patient can maintain adequate oral nutrition before removing the tube. In general, this can be
accomplished by occluding the tube for a short trial period and documenting caloric intake. However,
patients with chronic neurologic impairment may have a waxing and waning disease course. In such
individuals, it is probably wise to extend the trial period for 6 months or more before removing the
tube.120(5549)
Various methods of tube removal have been described.119(5550) In general, PEG tubes of more
recent design have a soft, deformable inner bumper that can be removed (nonendoscopically) by
external traction. The tube is first rotated in its tract to release all encrustations, and then steady, gentle
traction is applied until the tube exits the abdominal wall. However, feeding tubes that have a rigid inner
bumper must be removed endoscopically. This is done by cutting the tube off at the skin level and
carefully retrieving the inner bumper with a snare. Use of an overtube or foreign body sheath may also
be advantageous. In most instances, we do not recommend cutting the gastrostomy tube externally
and allowing the bumper and the remaining tube to pass in the stool. This practice has led to
numerous reports of complications related to internally retained tube components,121124(5551)
especially with small children.125(5552) Exploratory laparotomy was required in some
cases.126,127(5553)
Once the gastrostomy tube has been removed, intentionally or inadvertently, the tract closes very
quickly, usually without any complications and within 24 hr. However, persistent fistulas and peritonitis
have been reported.128(5554) If the tube is inadvertently removed and replacement is desired (in a
mature tract), it is often expedient to initially insert a Foley catheter in order to maintain tract patency
and thereby allow for a more permanent replacement tube to be inserted nonendoscopically. A mature
gastrocutaneous tract is probably present after the first few weeks. Should the tube be inadvertently
removed before this time, it should be replaced endoscopically.129(5555) If an existing tract is still
patent but too narrow, careful use of various dilators may be considered to gently dilate the mature
tract.119(5556)
Examples of replacement devices include gastrostomy tubes with either a soft deformable inner
bumper or a balloon tip; or skin-level devices with either an internal button or a balloon tip. All four
types are introduced externally via the mature tract into the stomach and do not require endoscopy.
The gastrostomy tube with deformable inner bumper and the skin-level gastrostomy button are
externally introduced by means of a stylet that deforms the tip of the device. After an initial conversion
to a skin-level gastrostomy tube, it is often prudent to obtain fluoroscopic verification of correct
intragastric placement.130(5557) In addition, some patients or their caretakers find the skin-level
devices difficult to manipulate and use for feedings. Furthermore, skin-level devices are significantly
more expensive than standard replacement tubes. Therefore, the relative merits and disadvantages of
each type should be discussed with the patient or caregiver before insertion of a replacement device. A
Foley catheter has also been successfully used as a replacement device,131,132(5558) although there
may be disadvantages to the use of a balloon-tipped catheter for long-term feeding.133(5559)
Surgical, Endoscopic, and Radiologic Gastrostomy

The present concept of surgical gastrostomy was proposed by Egeberg in 1837, but it was not until 39
years later, in 1876, that Verneuil performed the first successful gastrostomy in humans. The most
commonly performed surgical gastrostomy adheres to the technique proposed by Stamm in 1894 and
requires a laparotomy and general anesthesia.134(5560) Complication rates for surgical gastrostomies
have varied between 0.3% and 61%.135(5561) Many retrospective studies of surgical gastrostomy
versus PEG have led to a conclusion that PEG is the procedure of choice.134138(5562) This may be
explained in part by the fact that the surgical gastrostomies were performed under general anesthesia.
In an attempt to decrease the complication rate, the techniques for surgical gastrostomy have been
modified so that the procedure can be performed with local anesthesia. A prospective randomized trial
of PEG versus surgical gastrostomy (under local anesthesia) found no differences between the two
methods with respect to morbidity, mortality, or tube function.139(5563) Nevertheless, the endoscopic
technique had a definite economic advantage, costing $700 less than the surgical version.139(5564)
Percutaneous, radiologic (nonendoscopic) gastrostomy was introduced in 1983. Since then, technical
modifications of the original methods have been suggested.140142(5565) One retrospective study of
complication rates for surgical versus radiologic gastrostomy found that the latter technique had far
fewer complications.143(5566) A metaanalysis comparing radiologic, endoscopic, and surgical
gastrostomy favored the radiologic technique as having a better success rate and fewer
complications.144(5567) Another retrospective study found no significant difference among the three
techniques.145(5568) On the contrary, another study found the endoscopic method to be far superior,
with a lower rate of serious complications.146(5569) Laparoscopic gastrostomy has been described
but requires further evaluation, including prospective, comparative studies.147155(5570)
It is our opinion that, in general, the endoscopic method should be attempted first in most patients
unless there are specific contraindications to EGD. This technique has an inherent major advantage in
that the upper digestive tract can be examined thoroughly before performing the gastrostomy. This has
been shown to be beneficial because EGD detects previously unrecognized but significant
abnormalities and frequently leads to changes in patient management.81,82(5571) The endoscopic
method has been clearly shown to be more cost effective than surgical gastrostomy.139(5572) In
certain specific instances, however, a radiologic gastrostomy may be preferred.156(5573)
Percutaneous Endoscopic Jejunostomy

It has been known for some time that a surgically placed feeding jejunostomy has a much lower
incidence of aspiration of tube feeds than does a gastrostomy.157,158(5574) There have been
attempts to attain a similar low risk of aspiration by endoscopic placement of a jejunostomy via a
percutaneous route. As an adaptation of PEG, percutaneous endoscopic jejunostomy (PEJ) permits
placement of an enteral feeding tube in the jejunum of patients who are unable to tolerate gastric
feedings.159(5575)
The technique of PEJ is very similar to that of PEG, except that the tube design is modified to provide
for a gastric tube, which anchors the apparatus and drains the stomach, and a longer enteric tube,
which is passed beyond the pylorus and serves for administration of nutrient solutions (Figure 559).
The tubes may lie side by side or a small tube may be passed through the lumen of a larger one. None
of the present designs is ideal, and tube dysfunction is a routine occurrence.160(5576) Although the
technique was originally thought to be ideal for patients with frequent aspiration secondary to
gastroesophageal reflux, no benefit has been demonstrated in available studies.161,162(5577) The
major reason for the disappointing results appears to be the fact that the tube ends up in the
duodenum or even the stomach in the majority of attempts at jejunal insertion.163(5578) Despite poor
initial results, the PEJ technique continues to undergo modifications. These include: longer-length
feeding tubes,164(5579) an over-the-wire method,165167(5580) a laparoscope-assisted
technique,168,169(5581) and even a direct percutaneous endoscopic technique using a push
enteroscope.170,171(5582) Much further information and data from randomized, prospective,
comparative studies will be needed before any of these new techniques can be recommended.
Currently, the percutaneous placement of a jejunostomy feeding tube is probably best reserved for
those patients with poor gastric emptying as opposed to those with pathologic gastroesophageal reflux.
(5583)Figure 559. After the gastrostomy portion of the catheter is pulled up to the gastric
wall, the gastroscope is used to place the weighted portion of the feeding tube across the
duodenal bulb.(From Ponsky JL. Percutaneous endoscopic stomas. Surg Clin North Am 1989;
69[6].)
Complications
The incidence and variety of complications occurring during or after placement of a PEG have been
well documented.172177(5584)
Infectious complications may be minor, such as slight cellulitis around the exit of the tube from the
abdominal wall, or catastrophic, such as necrotizing fasciitis of the abdominal wall.178180(5585)
Careful preparation of the abdominal wall with antiseptic solution, swabbing of the oral cavity at the
commencement of the procedure with an antiseptic solution, careful adherence to sterile technique,
and the intravenous administration of antibiotics prophylactically will reduce the incidence of infectious
complications. Patient selection is also important. The procedure should be delayed in patients with
ongoing septic processes, to avoid seeding of the gastrostomy site. When examination of the
abdominal wall after the procedure reveals evidence of infection, antibiotic therapy should be instituted
immediately and any subcutaneous collections of pus should be drained. The latter may be
accomplished with a small incision, using local anesthesia, at the bedside. Drainage of a small
peritubal abscess will most often resolve the problem and may avert progression to necrotizing
infection of the abdominal wall. When the latter is suspected, prompt radical surgical dbridement is
necessary. The gastrostomy insertion site may be an unrecognized reservoir of methicillin-resistant
Staphylococcus aureus.181(5586)
Tube migration into the abdominal wall may occur after PEG ("buried bumper syndrome"). Patients
with this complication may present with peritubular leakage, abdominal pain, or resistance to the flow
of liquid through the tube.182(5587) Externally directed migration is usually due to the application of
excessive tension when the tube is originally seated. Clinical and laboratory studies have demonstrated
that a satisfactory tract will form even if no tension is applied to the tube. Tension causes ischemia and
necrosis of the intervening tissue, which result in tube migration and often subsequent
infection.101(5588) When a tube has migrated into the gastric submucosa, or into the abdominal wall,
it should be removed and replaced. A variety of surgical183(5589) and nonsurgical182,184187(5590)
techniques have been described for this purpose.
Separation of the stomach from the abdominal wall may occur after accidental dislodgment of the tube
before the formation of an adequate tract, as a result of excessive tension applied to the tube with
migration out of the stomach, or as a result of tube failure (as in premature deflation of a balloon-tip
catheter). When this complication is recognized shortly after its occurrence, it is acceptable to attempt
nonoperative therapy with nasogastric suction and intravenous administration of antibiotics. If the
patient improves, PEG may be performed again in 5 to 7 days. Should signs of peritonitis be evident
initially or after a trial of nonoperative management, the patient must undergo laparotomy with
cleansing of the abdominal cavity and operative replacement of the gastrostomy. Successful
immediate percutaneous endoscopic replacement of the gastrostomy tube has been reported in five
patients with premature tube dislodgment.129(5591)
Penetration of the colon and the consequent development of a gastrocolic fistula is a recognized
complication of PEG. Often this complication remains undetected for some time after the
procedure.188(5592) Rarely, the entire inner bumper may have migrated externally and be found in the
colon, resulting in a colocutaneous fistula (without gastric involvement).189(5593) The patient may
present with severe diarrhea,190(5594) or the problem may be discovered only when attempted
replacement of the original tube results in placement of the new tube within the colonic lumen. When
this is suspected, water-soluble contrast medium should be instilled through the gastrostomy tube;
opacification of the colon at fluoroscopy confirms the existence of the fistula. In most instances, this
complication is easily managed by removing the gastrostomy tube and allowing the tract and fistula to
close. In general, surgical repair is unnecessary.
Enlargement of the gastrostomy site, with subsequent leakage of gastric contents onto the skin, may
occur owing to tension, with resultant ischemia of the abdominal wall or pivoting of the tube as it
emerges from the abdomen. This situation may often be ameliorated by replacing the original tube with
a skin-level device that will reduce the pivoting action of the tube against the soft tissue (Figure 5510).
Alternatively, the original tube may be removed and the gastrostomy replaced at a new site, taking care
to avoid tension on the tube.
(5595)Figure 5510. Skin-level devices may be used to improve the cosmetic result of a
gastrostomy or to treat stomal problems.
Skin maceration or fungal infection around the gastrostomy site may be the result of a wet environment
caused by an occlusive dressing. In most instances, no dressing is required over the gastrostomy site.
Daily cleansing of the skin with exposure to air will help prevent maceration. Application of an
antifungal powder and steroid cream is often beneficial once this condition has developed. Fungal
colonization of the gastrostomy tube may also be a major contributing factor to eventual tube
deterioration and breakage.191193(5596)
Rare instances of implantation metastases at the gastrostomy tube insertion site have been reported
after PEG in patients with cancer of the head and neck.43,194198(5597) The exact mechanism of
seeding has not been determined, but it may be related to the passage of the gastrostomy apparatus
across obstructive malignant tumors in the upper aerodigestive tract. The other possibility is
hematogenous tumor spread.199(5598) Nevertheless, this complication is rare and, until further
studies elucidate the pathologic basis for metastasis, PEG continues to have an important role in the
management of patients with these types of cancers.38(5599)
Radiographic evidence of pneumoperitoneum is common after PEG tube placement and is found in
approximately 40% of cases.200(5600) In most instances, the patient is asymptomatic and the clinical
course is benign, although pneumoperitoneum can persist for up to 8 weeks.201(5601)
Other types of complications of PEG include small bowel fistula,202(5602) intestinal
volvulus,203(5603) colonic obstruction,204(5604) duodenal obstruction,205,206(5605) pyoderma
gangrenosum,207(5606) and acute gastric dilation.208(5607)
Ethical Considerations
PEG has become one of the most frequent indications for upper endoscopy in hospitalized patients.
Hence, it is not surprising that ethical issues abound. Once the decision has been made to proceed,
the method is technically straightforward. However, it is often much easier to start gastrostomy tube
feedings than to terminate them.209(5608) Extended care facilities, delighted with the ease of feeding
patients with gastrostomies, have often been adamant in their request that the procedure be performed
before they accept a patient. One study documented that health care professionals believed that
feeding via a PEG tube was indicated in geriatric patients with non-dysphagia-related
malnutrition.210(5609) This may be due to false or overly optimistic expectations regarding the benefits
of gastrostomy tube feedings. In certain subgroups of patients, there is no improvement in functional or
nutritional status.211(5610) Studies of survival after PEG have demonstrated that a significant number
of patients expire within weeks to months of tube placement as a consequence of underlying disease
processes.212216(5611)
Placement of a feeding gastrostomy is appropriate when it is anticipated that this will improve the
quality of a patient's existence. It is clear, however, that nasoenteric tube feedings may be equally
effective, less invasive, and more appropriate in patients with limited potential for survival beyond
several months. Certain predictive factors for early mortality after PEG have been identified that, if
validated in other studies, would assist in identifying the subgroup of patients who will not benefit from
such a tube.217(5612) These factors may include age greater than 75 years, previous aspiration, and
urinary tract infection. In one study, patients who had all three of these risk factors had a 1-month
probability of death approaching 70%.213(5613) Moreover, when tube feedings are anticipated to
prolong life in an otherwise unsalvageable situation, careful consultation with family members should
precede placement of a PEG.
Conclusion
PEG has become a well-accepted and widely used procedure. Several modifications of the original
method have met with good results, and advances in tube and accessory design have increased the
ease and safety of the procedure. Expanded applications of the method have been developed and
successfully implemented in a variety of clinical settings. Although complications are known to occur,
their incidence and outcomes may be minimized by careful attention to detail in the performance of the
procedure and close postprocedure follow-up. The wide acceptance of the method has led to its use in
some patients with a poor potential for recovery, and this has raised questions regarding the ethics of
patient selection. Careful, individual consideration of this issue will be in the best interest of each
patient and will enhance the overall value of the procedure.
REFERENCES
1.
2.
3.
Gauderer MWL, Ponsky JL, Izant RJ Jr. Gastrostomy without laparotomy: A percutaneous
endoscopic technique. J Pediatr Surg 1980;15:8725.
Gauderer MWL. An updated experience with percutaneous endoscopic gastrostomy in
children. Gastrointest Endosc Clin N Am 1992;2:195205.
Caulfield M. Percutaneous endoscopic gastrostomy placement in children. Gastrointest
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
Endosc Clin N Am 1994;4:17993.

Marin OE, Glassman MS, Schoen BT, Caplan DB. Safety and efficacy of percutaneous
endoscopic gastrostomy in children. Am J Gastroenterol 1994;89:35761.
Ni YH, Chang MH, Hsu HY, et al. Percutaneous endoscopic gastrostomy in infants. J Formos
Med Assoc 1995;94:6357.
Larson DE, Burton DD, Schroeder KW, DiMagno EP. Percutaneous endoscopic gastrostomy.
Indications, success, complications, and mortality in 314 consecutive patients.
Llaneza PP, Menendez AM, Roberts R, Dunn GD. Percutaneous endoscopic gastrostomy:
Clinical experience and follow-up. South Med J 1988;81:3214.
Pender SM, Courtney MG, Rajan E, et al. Percutaneous endoscopic gastrostomyResults of
an Irish single unit series. Ir J Med Sci 1993;162:4525.
Ponsky JL, Gauderer MWL, Stellato TA. Percutaneous endoscopic gastrostomy. Review of
150 cases. Arch Surg 1983;118:9134.
Rosenberg L, Fried GM. Percutaneous endoscopic gastrostomy: Indications and results. Can
J Surg 1986;29:3113.
Sangster W, Cuddington GD, Bachulis BL. Percutaneous endoscopic gastrostomy. Am J Surg
1988;155:6779.
Steffes C, Weaver DW, Bouwman DL. Percutaneous endoscopic gastrostomy. New
techniqueold complications. Am Surg 1989;55:2737.
Strodel WE, Lemmer J, Eckhauser F, et al. Early experience with endoscopic percutaneous
gastrostomy. Arch Surg 1983;118:44953.
Kohli H, Bloch R. Percutaneous endoscopic gastrostomy: A community hospital experience.
Am Surg 1995;61:1914.
Ganga UR, Ryan JJ, Schafer LW. Indications, complications, and long-term results of
percutaneous endoscopic gastrostomy: A retrospective study. S D J Med 1994;47:14952.
Joshi MA, Dewoolkar VV, Desai AN, et al. Gastrostomy without laparotomyA percutaneous
endoscopic technique: Review of 35 cases. Indian J Gastroenterol 1993;12:123.
Sali A, Wong PT, Read A, et al. Percutaneous endoscopic gastrostomy: The Heidelberg
Repatriation Hospital experience. Aust N Z J Surg 1993;63:54550.
Finucane P, Aslan SM, Duncan D. Percutaneous endoscopic gastrostomy in elderly patients.
Gauderer MWL, Stellato TA. Gastrostomies: Evolution, techniques, indications, and
complications. Curr Probl Surg 1986;23:657719.
Foutch PG, Haynes WC, Bellapravalu S, Sanowski RA. Percutaneous endoscopic
gastrostomy (PEG). A new procedure comes of age. J Clin Gastroenterol 1986;8:105.
Zakko W, Van Dam J. The technique of percutaneous endoscopic gastrostomy. J Crit Illness
1996;11:61723.
Mellinger JD. Percutaneous endoscopic gastrostomy: An evaluation after a decade.
Grant JP. Percutaneous endoscopic gastrostomy. Initial placement by single endoscopic
technique and long-term follow-up. Ann Surg 1993;217:16874.
Deitch EA, Winterton J, Li M, Berg R. The gut as a portal of entry for bacteremia: Role of
protein malnutrition. Ann Surg 1987;205:68192.
Deitch EA, Ma W-J, Ma L, et al. Protein malnutrition predisposes to inflammatory-induced
gut-origin septic states. Ann Surg 1990;211:5608.
Alverdy J, Chi HS, Sheldon GF. The effect of parenteral nutrition on gastrointestinal immunity:
The importance of enteral stimulation. Ann Surg 1985;202:6814.
Moore FA, Feliciano DV, Andrassy RJ, et al. Early enteral feeding, compared with parenteral,
reduces postoperative septic complications. Ann Surg 1992;216:17283.
American Gastroenterological Association. American Gastroenterological Association Medical
Position Statement: Guidelines for use of enteral nutrition. Gastroenterology
1995;108:12801.
29. Sant SM, Gilvarry J, Shannon R, O'Morain C. Percutaneous endoscopic gastrostomyIts

application in patients with neurological disease. Ir J Med Sci 1993;162:4501.
30. Mathus-Vliegen LM, Louwerse LS, Merkus MP, et al. Percutaneous endoscopic gastrostomy
in patients with amyotrophic lateral sclerosis and impaired pulmonary function. Gastrointest
Endosc 1994;40:4639.
31. Mazzini L, Corra T, Zaccala M, et al. Percutaneous endoscopic gastrostomy and enteral
nutrition in amyotrophic lateral sclerosis. J Neurol 1995;242:6958.
32. Frost RA, Rivers H, Tromans AM, Grundy DJ. The role of percutaneous endoscopic
gastrostomy in spinal cord injured patients. Paraplegia 1995;33:4168.
33. Norton B, Homer-Ward M, Donnelly MT, et al. A randomised prospective comparison of
percutaneous endoscopic gastrostomy and nasogastric tube feeding after acute dysphagic
stroke. BMJ 1996;312:136.
34. Mighell AJ, Carton AT, Boobis LH, Stassen LF. Acceptance and perception of percutaneous
endoscopic gastrostomy by patients with upper aerodigestive tract cancer. Br J Oral
Maxillofac Surg 1995;33:1922.
35. Koehler J, Buhl K. Percutaneous endoscopic gastrostomy for postoperative rehabilitation after
maxillofacial tumor surgery. Int J Oral Maxillofac Surg 1991;20:389.
36. Patton ML, Haith LR Jr, Germain TJ, et al. Use of percutaneous endoscopic gastrostomy
tubes in burn patients. Surg Endosc 1994;8:106771.
37. Fietkau R, Iro H, Sailer D, Sauer R. Percutaneous endoscopically guided gastrostomy in
patients with head and neck cancer. Recent Results Cancer Res 1991;121:26982.
38. Gibson S, Wenig BL. Percutaneous endoscopic gastrostomy in the management of head and
neck carcinoma. Laryngoscope 1992;102:97780.
39. Shike M, Berner YN, Gerdes H, et al. Percutaneous endoscopic gastrostomy and jejunostomy
for long-term feeding in patients with cancer of the head and neck. Otolaryngol Head Neck
Surg 1989;101:54954.
40. Wilson WR, Hariri SM. Experience with percutaneous endoscopic gastrostomy on an
otolaryngology service. Ear Nose Throat J 1995;74:7602.
41. Selz PA, Santos PM. Percutaneous endoscopic gastrostomy. A useful tool for the
otolaryngologist-head and neck surgeon. Arch Otolaryngol Head Neck Surg
1995;121:124952.
42. Sartori S, Trevisani L, Tassinari D, et al. Cost analysis of long-term feeding by percutaneous
endoscopic gastrostomy in cancer patients in an Italian health district. Support Care Cancer
1996;4:216.
43. Laccourreye O, Chabardes E, Merite-Drancy A, et al. Implantation metastasis following
percutaneous endoscopic gastrostomy. J Laryngol Otol 1993;107:9469.
44. Israel DM, Hassall E. Prolonged use of gastrostomy for enteral hyperalimentation in children
with Crohn's disease. Am J Gastroenterol 1995;90:10848.
45. Steinkamp G, von der Hardt H. Improvement of nutritional status and lung function after
long-term nocturnal gastrostomy feedings in cystic fibrosis. J Pediatr 1994;124:2449.
46. Slezak FA, Kofol WH. Combined tracheostomy and percutaneous endoscopic gastrostomy.
Am J Surg 1987;154:2713.
47. D'Amelio LF, Hammond JS, Spain DA, Sutyak JP. Tracheostomy and percutaneous
endoscopic gastrostomy in the management of the head-injured trauma patient. Am Surg
1994;60:1805.
48. Ockenga J, Suttmann U, Selberg O, et al. Percutaneous endoscopic gastrostomy in AIDS and
control patients: Risks and outcome. Am J Gastroenterol 1996;91:181722.
49. Cappell MS, Godil A. A multicenter case-controlled study of percutaneous endoscopic
gastrostomy in HIV-seropositive patients. Am J Gastroenterol 1993;88:205966.
50. Koh ML, Lipkin EW. Nutrition support of a pregnant comatose patient via percutaneous
endoscopic gastrostomy. JPEN J Parenter Enteral Nutr 1993;17:3847.
51. Ishizaki Y, Mitsusada M, Wakayama T. Percutaneous transhepatic biliary drainage combined
with percutaneous endoscopic gastrostomy for internal biliary drainage. J Am Coll Surg
1994;179:73840.
52. Ponsky JL, Aszodi A. External biliary-gastric fistula: A simple method for recycling bile. Am J
53. Adelson MD, Kasowitz MH. Percutaneous endoscopic drainage gastrostomy in the treatment
of gastrointestinal obstruction from intraperitoneal malignancy. Obstet Gynecol
1993;81:46771.
54. Herman LL, Hoskins WJ, Shike M. Percutaneous endoscopic gastrostomy for decompression
of the stomach and small bowel. Gastrointest Endosc 1992;38:3148.
55. Marks WH, Perkal MF, Schwartz PE. Percutaneous endoscopic gastrostomy for gastric
decompression in metastatic gynecologic malignancies. Surg Gynecol Obstet
1993;177:5736.
56. Campagnutta E, Cannizzaro R, Gallo A, et al. Palliative treatment of upper intestinal
obstruction by gynecological malignancy: The usefulness of percutaneous endoscopic
gastrostomy. Gynecol Oncol 1996;62:1035.
57. Cannizzaro R, Bortoluzzi F, Valentini M, et al. Percutaneous endoscopic gastrostomy as a
decompressive technique in bowel obstruction due to abdominal carcinomatosis. Endoscopy
1995;27:31720.
58. Moulis H, Vender RJ. Percutaneous endoscopic gastrostomy for treatment of gas-bloat
59. Foutch PG, Talbert G, Sanowski RA. Nonoperative traction method for placement of
esophageal stents: A new use for the percutaneous endoscopic gastrostomy. Gastrointest
Endosc 1988;34:25962.
60. Holderman WH, Etzkorn KP, Harig JM, Watkins JL. Endoscopic retrograde
cholangiopancreatography and stent placement via gastrostomy: Technical aspects and
clinical application. Endoscopy 1995;27:1357.
61. Ghosh S, Palmer KR. Double percutaneous endoscopic gastrostomy fixation: An effective
treatment for recurrent gastric volvulus. Am J Gastroenterol 1993;88:12712.
62. Zera RT, Nava HR, Fischer JI. Percutaneous endoscopic gastrostomy (PEG) in cancer
patients. Surg Endosc 1993;7:3047.
63. Mansberger JA, Kavouklis JH. Combined endoscopic and surgical gastrostomy when
percutaneous endoscopic gastrostomy is contraindicated. Am Surg 1990;56:13840.
64. Raaf JH, Manney M, Okafor E, et al. Laparoscopic placement of a percutaneous endoscopic
gastrostomy (PEG) feeding tube. J Laparoendosc Surg 1993;3:4114.
65. Vargo JJ, Germain MM, Swenson JA, Harrison CR. Ultrasound-assisted percutaneous
endoscopic gastrostomy in a patient with advanced ovarian carcinoma and recurrent intestinal
obstruction. Am J Gastroenterol 1993;88:19468.
66. Panzer S, Harris M, Berg W, et al. Endoscopic ultrasound in the placement of a percutaneous
endoscopic gastrostomy tube in the non-transilluminated abdominal wall. Gastrointest Endosc
1995;42:8890.
67. Vogt W, Messmann H, Lock G, et al. CT-guided PEG in patients with unsuccessful
endoscopic transillumination. Gastrointest Endosc 1996;43:13840.
68. Sawyer AM, Ghosh S, Eastwood MA. Satisfactory outcome of percutaneous endoscopic
gastrostomy in two patients with cirrhosis and portal hypertension. Am J Gastroenterol
1995;90:8268.
69. Sartori S, Trevisani L, Nielsen I, et al. Identification of a safe site for percutaneous endoscopic
gastrostomy placement in patients with marked bowel distension: May octreotide have a
role?. Endoscopy 1994;26:7101.
70. Stellato TA, Gauderer MWL, Ponsky JL. Percutaneous endoscopic gastrostomy following
previous abdominal surgery. Ann Surg 1984;200:4650.
71. Townsend MC, Flancbaum L, Cloutier CT, Arnold MW. Early post laparotomy percutaneous
endoscopic gastrostomy. Surg Gynecol Obstet 1992;174:468.
72. Foutch PG, Talbert GA, Waring JP, Sanowski RA. Percutaneous endoscopic gastrostomy in
patients with prior abdominal surgery: Virtues of the safe tract. Am J Gastroenterol
1988;83:14750.
73. Graham SM, Flowers JL, Scott TR, et al. Safety of percutaneous endoscopic gastrostomy in
patients with a ventriculoperitoneal shunt. Neurosurgery 1993;32:9324.
74. Cappell MS, Iacovone FM Jr. The safety and efficacy of percutaneous endoscopic
gastrostomy after recent myocardial infarction: A study of 28 patients and 40 controls at four
university teaching hospitals. Am J Gastroenterol 1996;91:1599603.
75. Jain NK, Larson DE, Schroeder KW, et al. Antibiotic prophylaxis for percutaneous endoscopic
gastrostomy. A prospective, randomized, double-blind clinical trial. Ann Intern Med
1987;107:8248.
76. Sturgis TM, Yancy W, Cole JC, et al. Antibiotic prophylaxis in percutaneous endoscopic
gastrostomy. Am J Gastroenterol 1996;91:23014.
77. Ponsky JL, Gauderer MWL. Percutaneous endoscopic gastrostomy: A nonoperative
technique for feeding gastrostomy. Gastrointest Endosc 1981;27:911.
78. Sacks BA, Vine HS, Palestrant AM, et al. A non-operative technique for establishment of a
gastrostomy in the dog. Invest Radiol 1983;18:4857.
79. Russell TR, Brotman M, Norris F. Percutaneous gastrostomy. A new simplified and
cost-effective technique. Am J Surg 1984;148:1327.
80. Ferguson DR, Harig JM, Kozarek RA, et al. Placement of a feeding button ("one-step button")
as the initial procedure. Am J Gastroenterol 1993;88:5014.
81. Scott JS, Edelman DS, Unger SW. Percutaneous endoscopic gastrostomy. A mandate for
complete diagnostic upper endoscopy. Am Surg 1989;55:857.
82. Wolfsen HC, Kozarek RA, Ball TJ, et al. Value of diagnostic upper endoscopy preceding
percutaneous gastrostomy. Am J Gastroenterol 1990;85:24951.
83. Mellinger JD, Simon IB, Schlecter B, et al. Tract formation following percutaneous endoscopic
gastrostomy in an animal model. Surg Endosc 1991;5:18991.
84. Aisenberg J, Cohen L, Lewis BS. Marked endoscopic gastrostomy tubes permit one-pass
Ponsky technique. Gastrointest Endosc 1991;37:5523.
85. Rogers BH, Kaminski MV Jr, All J. Stabilizing sutures for percutaneous endoscopic
gastrostomy. Gastrointest Endosc 1989;35:2413.
86. Brown AS, Mueller PR, Ferrucci JT Jr. Controlled percutaneous gastrostomy: Nylon
T-fastener for fixation of the anterior gastric wall. Radiology 1986;158:5435.
87.
Kelsey PB. Percutaneous endoscopic gastrostomy: The "push" technique. In Ponsky JL (ed).
Techniques of Percutaneous Gastrostomy. New York, Igaku-Shoin, 1988;338.
88. Hogan RB, DeMarco DC, Hamilton JK, et al. Percutaneous endoscopic gastrostomyTo
push or pull? A prospective randomized trial. Gastrointest Endosc 1986;32:2538.
89. Miller RE, Winkler WP, Kotler DP. The Russell percutaneous endoscopic gastrostomy: Key
technical steps. Gastrointest Endosc 1988;34:33942.
90. Robertson FM, Crombleholme TM, Latchaw LA, Jacir NN. Modification of the "push"
technique for percutaneous endoscopic gastrostomy in infants and children. J Am Coll Surg
1996;182:2158.
91. Miller RE, Castlemain B, Lacqua FJ, et al. Percutaneous endoscopic gastrostomy: Results in
316 patients and review of the literature. Surg Endosc 1989;3:18690.
92. Adler DJ, Nunez DC. Large bore percutaneous endoscopic gastrostomy using the Russell
technique. Gastrointest Endosc 1989;35:23941.
93. Kozarek RA, Ball TJ, Ryan JA Jr. When push comes to shove: A comparison between two
methods of percutaneous endoscopic gastrostomy. Am J Gastroenterol 1986;81:6426.
94. Gauderer MWL, Picha GJ, Izant RJ. The gastrostomy buttonA simple, skin-level,
nonrefluxing device for long-term enteral feedings. J Pediatr Surg 1984;19:8035.
95. Foutch PG, Talbert GA, Gaines JA, et al. The gastrostomy button: A prospective assessment
of safety, success, and spectrum of use. Gastrointest Endosc 1989;35:414.
96. Kozarek RA, Payne M, Barkin J, et al. Prospective multicenter evaluation of an initially placed
button gastrostomy. Gastrointest Endosc 1995;41:1058.
97. Marion MT, Zweng TN, Strodel WE. One-stage gastrostomy button: An assessment.
Endoscopy 1994;26:66670.
98. Treem WR, Etienne NL, Hyams JS. Percutaneous endoscopic placement of the "button"
gastrostomy tube as the initial procedure in infants and children. J Pediatr Gastroenterol Nutr
1993;17:3826.
99. Stylianos S, Flanigan LM. Primary button gastrostomy: A simplified percutaneous, open,
laparoscopy-guided technique. J Pediatr Surg 1995;30:21920.
100. Stylianos S, Flanigan LM. Laparoscopy-guided percutaneous button gastrostomy in children
after previous abdominal surgery. J Laparoendosc Surg 1995;5:199201.
101. Chung RS, Schertzer M. Pathogenesis of complications of percutaneous endoscopic
gastrostomy. A lesson in surgical principles. Am Surg 1990;56:1347.
102. Starkey JF, Jefferson PA, Kirby DF. Taking care of percutaneous endoscopic gastrostomy.
Am J Nurs 1988;88:425.
103.
Kundtz JE. Nursing implications for patients with a percutaneous endoscopic gastrostomy or
jejunostomy. Gastrointest Endosc Clin N Am 1992;2:20714.
104. Choudry U, Barde CJ, Markert R, Gopalswamy N. Percutaneous endoscopic gastrostomy: A
randomized prospective comparison of early and delayed feeding. Gastrointest Endosc
1996;44:1647.
105. Brown DN, Miedema BW, King PD, Marshall JB. Safety of early feeding after percutaneous
endoscopic gastrostomy. J Clin Gastroenterol 1995;21:3301.
106. Werlin S, Glicklich M, Cohen R. Early feeding after percutaneous endoscopic gastrostomy is
safe in children. Gastrointest Endosc 1994;40:6923.
107. Ciocon JO, Galindo-Ciocon DJ, Tiessen C, Galindo D. Continuous compared with intermittent
tube feeding in the elderly. JPEN J Parenter Enteral Nutr 1992;16:5258.
108. American Gastroenterological Association. American Gastroenterological Association
technical review on tube feeding for enteral nutrition. Gastroenterology 1995;108:1282301.
109. Powers DA, Brown RO, Cowan GSM Jr, et al. Nutrition support team vs nonteam
management of enteral nutrition support in a Veterans Administration medical center teaching
hospital. JPEN J Parenter Enteral Nutr 1986;10:6358.
110. Hussain A, Geddes A, Woolfrey S, et al. The team approach to percutaneous endoscopic
gastrostomy feeding. Br J Hosp Med 1995;54:30812.
111. Hull MA, Rawlings J, Murray FE, et al. Audit of outcome of long-term enteral nutrition by
percutaneous endoscopic gastrostomy. Lancet 1993;341:86972.
112. Launay V, Gottrand F, Turck D, et al. Percutaneous endoscopic gastrostomy in children:
Influence on gastroesophageal reflux. Pediatrics 1996;97:7268.
113. Johnson DA, Hacker JF, Benjamin SB, et al. Percutaneous endoscopic gastrostomy effects
on gastroesophageal reflux and lower esophageal sphincter. Am J Gastroenterol
1987;82:6224.
114. Cole MJ, Smith JT, Molnar C, Shaffer EA. Aspiration after percutaneous gastrostomy.
Assessment by Tc-99m labeling of the enteral feed. J Clin Gastroenterol 1987;9:905.
115. Heine RG, Reddihough DS, Catto-Smith AG. Gastrooesophageal reflux and feeding problems
after gastrostomy in children with severe neurological impairment. Dev Med Child Neurol
1995;37:3209.
116. Grunow JE, al-Hafidh A, Tunell WP. Gastroesophageal reflux following percutaneous
endoscopic gastrostomy in children. J Pediatr Surg 1989;24:424; discussion 445.
117. Heloury Y, Plattner V, Mirallie E, et al. Laparoscopic Nissen fundoplication with simultaneous
percutaneous endoscopic gastrostomy in children. Surg Endosc 1996;10:83741.
118. Sartori S, Trevisani L, Tassinari D, et al. Prevention of aspiration pneumonia during long-term
feeding by percutaneous endoscopic gastrostomy: Might cisapride play any role? An open
pilot study. Support Care Cancer 1994;2:18890.
119.
Ponsky JL. Percutaneous endoscopic gastrostomy: Techniques of removal and replacement.
120. Cullado MJ, Slezak FA, Porter JA. Repeat percutaneous endoscopic gastrostomy (PEG): An
outpatient procedure. Surg Endosc 1990;4:1734.
121. Coventry BJ, Karatassas A, Gower L, Wilson P. Intestinal passage of the PEG end-piece: Is it
safe?. J Gastroenterol Hepatol 1994;9:3113.
122. Saslow SB, Stark ME, DeVault KR, et al. Diarrhea due to a percutaneous gastrostomy
catheter head lodged in the terminal ileum. Am J Gastroenterol 1994;89:22556.
123. Perkins JA, Smith S, Lyons M. Bowel obstruction from retained inner bumper following
removal of gastrostomy tube: A case report. Mil Med 1993;158:1201.
124. Weston AP, Campbell DR. Distal small bowel obstruction by a severed PEG tube: Successful
endoscopic management by ileoscopic retrieval. Gastrointest Endosc 1995;42:26971.
125. Yaseen M, Steele MI, Grunow JE. Nonendoscopic removal of percutaneous endoscopic
gastrostomy tubes: Morbidity and mortality in children. Gastrointest Endosc 1996;44:2358.
126. Lambertz MM, Earnshaw PM, Short J, Cumming JG. Small bowel obstruction caused by a
retained percutaneous endoscopic gastrostomy gastric flange. Br J Surg 1995;82:951.
127. Mutabagani KH, Townsend MC, Arnold MW. PEG ileus. A preventable complication. Surg
Endosc 1994;8:6947.
128. Bender JS, Levison MA. Complications after percutaneous endoscopic gastrostomy removal.
Surg Laparosc Endosc 1991;1:1013.
129. Galat SA, Gerig KD, Porter JA, Slezak FA. Management of premature removal of the
percutaneous gastrostomy. Am Surg 1990;56:7336.
130. Romero R, Martinez FL, Robinson SYJ, et al. Complicated PEG-to-skin level gastrostomy
conversions: Analysis of risk factors for tract disruption. Gastrointest Endosc 1996;44:2304.
131. Kadakia SC, Cassaday M, Shaffer RT. Prospective evaluation of Foley catheter as a
replacement gastrostomy tube. Am J Gastroenterol 1992;87:15947.
132. Kadakia SC, Cassaday M, Shaffer RT. Comparison of Foley catheter as a replacement
gastrostomy tube with commercial replacement gastrostomy tube: A prospective randomized
trial. Gastrointest Endosc 1994;40:18893.
133. Pereira MC, Mersich K. Foley catheter gastrostomy tube migration: Small bowel obstruction
relieved by percutaneous balloon aspiration. Gastrointest Endosc 1991;37:3724.
134. Grant JP. Comparison of percutaneous endoscopic gastrostomy with Stamm gastrostomy.
Ann Surg 1988;207:598603.
135. Mamel JJ. Percutaneous endoscopic gastrostomy. Am J Gastroenterol 1989;84:70310.
136. Himal HS, Schumacher S. Endoscopic vs surgical gastrostomy for enteral nutrition. Surg
Endosc 1987;1:335.
137. Scott JS, de la Torre RA, Unger SW. Comparison of operative versus percutaneous
endoscopic gastrostomy tube placement in the elderly. Am Surg 1991;57:33840.
138. Cameron BH, Blair GK, Murphy JJ III, Fraser GC. Morbidity in neurologically impaired children
after percutaneous endoscopic versus Stamm gastrostomy. Gastrointest Endosc
1995;42:414.
139. Stiegmann GV, Goff JS, Silas D, et al. Endoscopic versus operative gastrostomy: Final results
of a prospective randomized trial. Gastrointest Endosc 1990;36:15.
140. Lorentzen T, Skjoldbye B, Nolsoe C, et al. Percutaneous gastrostomy guided by ultrasound
and fluoroscopy. Acta Radiol 1995;36:15962.
141. Rosenzweig TB, Palestrant AM, Esplin CA, Gilsdorf RB. A method for radiologic-assisted
gastrostomy when percutaneous endoscopic gastrostomy is contraindicated. Am J Surg
1994;168:58790; discussion 5901.
142. Bodley R, Banerjee S. Radiological percutaneous gastrostomy placement for enteral feeding.
Paraplegia 1995;33:1535.
143. Ho CS, Yee AC, McPherson R. Complications of surgical and percutaneous nonendoscopic
gastrostomy: Review of 233 patients. Gastroenterology 1988;95:120610.
144. Wollman B, D'Agostino HB, Walus-Wigle JR, et al. Radiologic, endoscopic, and surgical
gastrostomy: An institutional evaluation and meta-analysis of the literature. Radiology
1995;197:699704.
145. Goretsky MF, Johnson N, Farrell M, Ziegler MM. Alternative techniques of feeding
gastrostomy in children: A critical analysis. J Am Coll Surg 1996;182:23340.
146. Elliott LA, Sheridan MB, Denyer M, Chapman AH. PEGIs the E necessary? A comparison
of percutaneous and endoscopic gastrostomy. Clin Radiol 1996;51:3414.
147. Edelman DS, Arroyo PJ, Unger SW. Laparoscopic gastrostomy versus percutaneous
endoscopic gastrostomy. A comparison. Surg Endosc 1994;8:479.
148. Sylvester KG, Paskin DL, Schuricht AL. Combined laparoscopic-endoscopic gastrostomy.
Surg Endosc 1994;8:10725; discussion 1144.
149. Scheer MF, Miedema BW. Laparoscopic assisted percutaneous endoscopic gastrostomy.
150. Stringel G, Geller ER, Lowenheim MS. Laparoscopic-assisted percutaneous endoscopic
gastrostomy. J Pediatr Surg 1995;30:120910.
151. Viani MP, Poggi RV, Pinto A, et al. Gasless laparoscopic gastrostomy. J Laparoendosc Surg
1995;5:2459.
152. Lee WJ, Chao SH, Yu SC, Chen KM. Laparoscopic-guided gastrostomy. J Formos Med
Assoc 1993;92:9113.
153. Lee WJ, Chao SH, Yu SC, Chen KM. Laparoscopic assisted gastrostomy tube placement. J
Laparoendosc Surg 1994;4:2014.
154. Bessell JR, Stanley B, Maddern GJ. The emerging role for laparoscopic gastrostomy. Aust N
Z J Surg 1994;64:5157.
155. Modesto VL, Harkins B, Calton WC Jr, Martindale RG. Laparoscopic gastrostomy using
four-point fixation. Am J Surg 1994;167:2736.
156. Foutch PG, van Sonnenberg E, Casola G, D'Agostino H. Controversies, dilemmas, and
dialogues. Nonsurgical gastrostomy: X-ray or endoscopy?. Am J Gastroenterol
1990;85:15603.
157. Burtch GD, Shatney CH. Feeding gastrostomy: Assistant or assassin?. Am Surg
1985;51:2047.
158. Weltz CR, Morris JB, Mullen JL. Surgical jejunostomy in aspiration risk patients. Ann Surg
1992;215:1405.
159. Ponsky JL, Aszodi A. Percutaneous endoscopic jejunostomy. Am J Gastroenterol
1984;79:1136.
160. Wolfsen HC, Kozarek RA, Ball TJ, et al. Tube dysfunction following percutaneous endoscopic
gastrostomy and jejunostomy. Gastrointest Endosc 1990;36:2613.
161. DiSario JA, Foutch PG, Sanowski RA. Poor results with percutaneous endoscopic
jejunostomy. Gastrointest Endosc 1990;36:25760.
162.
Foutch GP. Complications of percutaneous endoscopic gastrostomy and jejunostomy:
Recognition, prevention, and treatment. Gastroenterol Clin North Am 1992;2:23148.
163. Lewis BS. Perform PEJ, not PED. Gastrointest Endosc 1990;36:3113.
164. Mathus-Vliegen EM, Tytgat GN, Merkus MP. Feeding tubes in endoscopic and clinical
practice: The longer the better?. Gastrointest Endosc 1993;39:53742.
165. MacFadyen BV Jr, Catalano MF, Raijman I, Ghobrial R. Percutaneous endoscopic
gastrostomy with jejunal extension: A new technique. Am J Gastroenterol 1992;87:7258.
166. Duckworth PF Jr, Kirby DF, McHenry L, et al. Percutaneous endoscopic gastrojejunostomy
made easy: A new over-the-wire technique. Gastrointest Endosc 1994;40:3503.
167. DeLegge MH, Patrick P, Gibbs R. Percutaneous endoscopic gastrojejunostomy with a
tapered tip, nonweighted jejunal feeding tube: Improved placement success. Am J
168. Etzkorn KP, Vitello J, Resnick D, et al. A new technique for jejunal tube placement: A
marriage of enteroscope and laparoscope. Gastrointest Endosc 1996;43:61014.
169. Edelman DS, Unger SW. Laparoscopic gastrostomy and jejunostomy: Review of 22 cases.
170. Shike M, Latkany L, Gerdes H, Bloch AS. Direct percutaneous endoscopic jejunostomies for
enteral feeding. Gastrointest Endosc 1996;44:53640.
171. Mellert J, Naruhn MB, Grund KE, Becker HD. Direct endoscopic percutaneous jejunostomy
(EPJ). Clinical results. Surg Endosc 1994;8:8679; discussion 86970.
172.
Strodel WE, Ponsky JL. Complications of percutaneous gastrostomy. In Ponsky JL (ed).

Techniques of Percutaneous Gastrostomy. New York, Igaku-Shoin, 1988;6378.
173. Schapiro GD, Edmundowicz SA. Complications of percutaneous endoscopic gastrostomy.
174. Davidson PM, Catto-Smith AG, Beasley SW. Technique and complications of percutaneous
endoscopic gastrostomy in children. Aust N Z J Surg 1995;65:1946.
175.
Levine CD, Handler B, Baker SR, et al. Imaging of percutaneous tube gastrostomies:
Spectrum of normal and abnormal findings. AJR 1995;164:34751.
176. Broscious SK. Preventing complications of PEG tubes. Dimens Crit Care Nurs
1995;14:3741.
177. Beasley SW, Catto-Smith AG, Davidson PM. How to avoid complications during percutaneous
endoscopic gastrostomy. J Pediatr Surg 1995;30:6713.
178. Cave DR, Robinson WR, Brotschi EA. Necrotizing fasciitis following percutaneous endoscopic
179. Greif JM, Ragland JJ, Ochsner MG, Riding R. Fatal necrotizing fasciitis complicating
percutaneous endoscopic gastrostomy. Gastrointest Endosc 1986;32:2924.
180. Evans DA, Bhandarkar DS, Taylor TV. Necrotising fasciitisA rare complication of
percutaneous endoscopic gastrostomy. Endoscopy 1995;27:627.
181. Nunley D, Berk SL. Percutaneous endoscopic gastrostomy as an unrecognized source of
methicillin-resistant Staphylococcus aureus colonization. Am J Gastroenterol 1992;87:5861.
182. Boyd JW, DeLegge MH, Shamburek RD, Kirby DF. The buried bumper syndrome: A new
technique for safe, endoscopic PEG removal. Gastrointest Endosc 1995;41:50811.
183. Zollinger RM Jr, Morris JB, Stellato TA. Migratory failed percutaneous endoscopic
gastrostomy. Surg Gynecol Obstet 1991;173:656.
184. Klein S, Heare BR, Soloway RD. The "buried bumper syndrome": A complication of
percutaneous endoscopic gastrostomy. Am J Gastroenterol 1990;85:44851.
185. van Bergeijk JD, Mulder CJ. Frequent submucosal migration of soft percutaneous endoscopic
gastrostomy catheters, and conservative management. Endoscopy 1995;27:406.
186. Rychter O. Percutaneous endoscopic gastrostomy feeding tube perforating the abdominal
wall: A technique for endoscopic removal. Endoscopy 1995;27:277.
187. Fay DE, Luther R, Gruber M. A single procedure endoscopic technique for replacing partially
extruded percutaneous endoscopic gastrotomy tubes. Gastrointest Endosc 1990;36:298300.
188. Minocha A, Rupp TH, Jaggers TL, Rahal PS. Silent cologastrocutaneous fistula as a
complication of percutaneous endoscopic gastrostomy. Am J Gastroenterol 1994;89:22434.
189. Siddique I, Krishnamurthy M, Choubey S, et al. Colocutaneous fistula: A rare and silent
complication of percutaneous endoscopic gastrostomy. Dig Dis Sci 1996;41:3014.
190. Bui HD, Dang CV, Schlater T, Nghiem CH. A new complication of percutaneous endoscopic
gastrostomy. Am J Gastroenterol 1988;83:44851.
191. Iber FL, Livak A, Patel M. Importance of fungus colonization in failure of silicone rubber
percutaneous gastrostomy tubes (PEGs). Dig Dis Sci 1996;41:22631.
192. Gottlieb K, Leya J, Kruss DM, et al. Intraluminal fungal colonization of gastrostomy tubes.
193. Gottlieb K, Iber FL, Livak A, et al. Oral Candida colonizes the stomach and gastrostomy
feeding tubes. JPEN J Parenter Enteral Nutr 1994;18:2647.
194. Becker G, Hess CF, Grund KE, et al. Abdominal wall metastasis following percutaneous
endoscopic gastrostomy. Support Care Cancer 1995;3:3136.
195. van Erpecum KJ, Akkersdijk WL, Warlam-Rodenhuis CC, et al. Metastasis of hypopharyngeal
carcinoma into the gastrostomy tract after placement of a percutaneous endoscopic
gastrostomy catheter. Endoscopy 1995;27:1247.
196. Lee DS, Mohit-Tabatabai MA, Rush BF Jr, Levine C. Stomal seeding of head and neck
cancer by percutaneous endoscopic gastrostomy tube placement. Ann Surg Oncol
1995;2:1703.
197. Sharma P, Berry SM, Wilson K, et al. Metastatic implantation of an oral squamous-cell
carcinoma at a percutaneous endoscopic gastrostomy site. Surg Endosc 1994;8:12325.

198. Schiano TD, Pfister D, Harrison L, Shike M. Neoplastic seeding as a complication of
199. Meurer MF, Kenady DE. Metastatic head and neck carcinoma in a percutaneous gastrostomy
site. Head Neck 1993;15:703.
200. Gottfried EB, Plumser AB, Clair MR. Pneumoperitoneum following percutaneous endoscopic
gastrostomy. A prospective study. Gastrointest Endosc 1986;32:3979.
201. Thirumahilmaran S, Patel N, Thomas E. Prolonged benign pneumoperitoneum after
202. Evans PM, Serpell JW. Small bowel fistula: A complication of percutaneous endoscopic
gastrostomy insertion. Aust N Z J Surg 1994;64:51820.
203. Ghahremani GG, Tsang TK, Vakil N. Complications of endoscopic gastrostomy:
Pneumoperitoneum and volvulus of the colon. Gastrointest Radiol 1987;12:1724.
204. Hennigan TW, Forbes A. Colonic obstruction caused by endoscopic percutaneous
gastrostomy. Eur J Surg 1992;158:435.
205. McGovern R, Barkin JS, Goldberg RI, Phillips RS. Duodenal obstruction: A complication of
percutaneous endoscopic gastrostomy tube migration. Am J Gastroenterol 1990;85:10378.
206. Neiblum DR, Gilman PB, Vaze MM. Duodenal obstruction by multiple percutaneous
endoscopic gastrostomy tube inner bumpers. Am J Gastroenterol 1996;91:80910.
207. Motoo Y, Watanabe H, Okai T, Sawabu N. Pyoderma gangrenosum as a complication of
percutaneous endoscopic gastrostomy tube insertion. J Clin Gastroenterol 1994;19:2678.
208. Ghosh S, Eastwood MA, Palmer KR. Acute gastric dilatationA delayed complication of
percutaneous endoscopic gastrostomy. Gut 1993;34:85960.
209. Fairman RP. Withdrawing life-sustaining treatment: Lessons from Nancy Cruzan. Arch Intern
Med 1992;152:257.
210. Hasan M, Meara RJ, Bhowmick BK, Woodhouse K. Percutaneous endoscopic gastrostomy in
geriatric patients: Attitudes of health care professionals. Gerontology 1995;41:32631.
211. Kaw M, Sekas G. Long-term follow-up of consequences of percutaneous endoscopic
gastrostomy (PEG) tubes in nursing home patients. Dig Dis Sci 1994;39:73843.
212. Raha SK, Woodhouse K. The use of percutaneous endoscopic gastrostomy (PEG) in 161
consecutive elderly patients. Age Ageing 1994;23:1623.
213. Light VL, Slezak FA, Porter JA, et al. Predictive factors for early mortality after percutaneous
endoscopic gastrostomy. Gastrointest Endosc 1995;42:3305.
214. Stuart SP, Tiley EH III, Boland JP. Feeding gastrostomy: A critical review of its indications and
mortality rate. South Med J 1993;86:16972.
215. Stellato TA, Gauderer MWL. Percutaneous endoscopic gastrostomy in the cancer patient. Am
Surg 1988;54:41922.
216. Wolfsen HC, Kozarek RA, Ball TJ, et al. Long-term survival in patients undergoing
percutaneous endoscopic gastrostomy and jejunostomy. Am J Gastroenterol
1990;85:11202.
217. Taylor CA, Larson DE, Ballard DJ, et al. Predictors of outcome after percutaneous endoscopic
gastrostomy: A community-based study. Mayo Clin Proc 1992;67:10429.
Chapter 56 Indications, Contraindications, and Complications of

Diagnostic Endoscopic Retrograde Cholangiopancreatography
(5614)
(5615)
ASSAAD M. SOWEID, M.D.
GREGORY ZUCCARO, JR., M.D.
D. ROY FERGUSON, M.D.

The introduction of endoscopic retrograde cholangiopancreatography (ERCP) was a major advance in
pancreaticobiliary imaging. Endoscopic cannulation of the papilla of Vater, introduced by McCune et
al.1(5616) in 1968, provided another means of obtaining a more detailed radiographic image of the
biliary and pancreatic ducts. ERCP evolved rapidly over the succeeding two decades with the addition
of a multitude of therapeutic maneuvers. Percutaneous transhepatic cholangiography (PTC) has
provided another, albeit more invasive, means of visualizing the biliary tree. Oral cholecystography and
intravenous cholangiography have been largely supplanted by ultrasonography (US), computed
tomography (CT), and magnetic resonance imaging (MRI). More recently, powerful tools such as
endoscopic ultrasonography (EUS) and magnetic resonance cholangiopancreatography (MRCP) have
been added to the armamentarium, thus further expanding the field of pancreaticobiliary imaging. This
imposing array of technology provides the clinician with choices in the effort to diagnose and to plan
therapy for patients with pancreaticobiliary diseases. In this modern era of cost-consciousness, the
clinician is also obligated not only to select a procedure(s) that will provide the most information with
the necessary degree of safety and comfort for the patient but also to select tests on the basis of
cost-effectiveness. Analyses of the cost and effectiveness of these different imaging modalities are
only now being conducted in an effort to obtain data that can be used to guide this complex process.
ERCP is widely accepted as the diagnostic/therapeutic procedure of choice for various
pancreaticobiliary disorders. It is important, therefore, that the endoscopist have full knowledge of the
indications and contraindications as well as possible complications of ERCP. This chapter focuses on
the general indications, contraindications, and complications of diagnostic ERCP, with further
discussion of the therapeutic implications left for other chapters in this book. A discussion of the
indications for ERCP in infants and children is presented in Chapter 79: Endoscopic Retrograde
Cholangiopancreatography in Infants and Children. The endoscopist must also recognize that there are
newer tests for the evaluation of pancreaticobiliary diseases that have comparative value to ERCP.
Thus, it is relevant to discuss these modalities in comparison with ERCP.
Indications for ERCP

The indications for ERCP (Table 561) are reviewed in relation to disorders of the hepatobiliary system
and pancreas.
TABLE 561
BILIARY TRACT DISEASES

Jaundice
Cholestasis
Acute cholangitis
Tissue diagnosis or confirmation of biliary tumors demonstrated
by other imaging modalities
Evaluation of ampullary neoplasms
Choledocholithiasis
Evaluation of suspected sphincter of Oddi dysfunction by
manometry
Suspected sclerosing cholangitis
Adjunct to laparoscopic cholecystectomy when stones are
present
Management of complications of biliary surgery including
laparoscopic cholecystectomy
To obtain bile for diagnostic studies
TABLE 561
Access for choledochoscopy or therapeutic applications (e.g.,

lithotripsy, biopsy)
PANCREATIC DISEASES
Biliary pancreatitis
Evaluation for pancreatic cancer if other imaging modalities are
inconclusive or normal, especially if management may be
altered
Tissue diagnosis or confirmation of pancreatic mass
demonstrated by other imaging modalities
Unexplained recurrent pancreatitis
Chronic pancreatitis-related abdominal pain
Chronic pancreatitis before surgery to define anatomic details
Evaluation for chronic pancreatitis (e.g., steatorrhea, diabetes
mellitus)
Evaluation of pancreatic pseudocyst
Pancreatic trauma
Unexplained elevated serum amylase or lipase
Obtain pancreatic juice for diagnostic studies
MISCELLANEOUS
Abdominal pain of unclear cause
Unexplained weight loss
Unexplained gastric varices
Unexplained ascites
ERCPendoscopic retrograde cholangiopancreatography.
Hepatobiliary System
Jaundice
The evaluation of patients with jaundice is one of the primary reasons for consultation for ERCP.
Jaundiced patients can be accurately separated into broad diagnostic categories of nonobstructive and
obstructive biliary tract disease on clinical grounds alone in about 90% of the cases.2(5617) Tests
performed in the jaundiced patient should therefore confirm the presence or absence of obstruction
and provide more specific information regarding the nature and location of the blockage when
present.3(5618)
US can have an accuracy as high as 90% in differentiating between biliary obstruction and
hepatocellular causes of jaundice, but it has limitations.4(5619) Common bile duct (CBD) imaging is
limited by body habitus and bowel gas. For example, between 24 and 50% of proven CBD stones are
missed by US.5,6(5620) In the postcholecystectomy state, slight dilation of the bile duct appears to
occur frequently.7,8(5621) This may lead to a false diagnosis of obstructive jaundice by US. Jaundice
without ductal dilation may also occur in several conditions including choledocholithiasis9(5622) and
the intrinsic duct disease, primary sclerosing cholangitis (PSC). A variable period of time is required for
ductal dilation to develop, so that widening of the duct may not be evident when biliary obstruction is of
relatively recent onset.10(5623) In such cases, US may erroneously indicate an absence of
obstruction. Despite the accuracy of US in differentiating obstructive from nonobstructive jaundice, it
often does not define the cause of obstruction. In a study by Dewbury et al.,11(5624) US correctly
defined the presence or absence of the obstruction in 97% of cases, but it was diagnostic as to the
cause of the obstruction in only 58%.
CT has approximately the same accuracy as US in differentiating biliary from hepatocellular causes of
jaundice.12(5625) Newer techniques, such as spiral CT after intravenous infusion of a contrast agent,
provide more detailed information as to the presence and cause of biliary obstruction.13(5626)
PTC has a success rate of up to 100% when the bile ducts are dilated,14(5627) but in the absence of
dilation, as in PSC, the success rate drops to 50% or less.15(5628)
ERCP, by virtue of direct opacification of the bile duct, not only demonstrates dilation of the bile duct
but also defines the site and nature of an obstruction. Although earlier studies, such as that of Bilbao et
al.,16(5629) found the overall success rate for ERCP in visualizing the biliary tree to be only 70%, more
recent studies have demonstrated success rates of up to 90% for retrograde cholangiography and in
differentiating obstructive from non-obstructive jaundice.17(5630) ERCP is thus indicated in the
evaluation of jaundice to confirm the presence of obstruction and to delineate the location and nature
of an obstructing lesion.
Cholestatic Disease
Jaundice is not necessarily present in patients with cholestatic disease. The initial manifestation of
diseases such as primary biliary cirrhosis or PSC is now more often a high alkaline phosphatase value
on a standard blood screening survey rather than clinical jaundice, which is in fact a late manifestation.
PSC is often suspected in certain clinical scenarios, as for example, a patient with cholestatic liver
enzyme abnormalities and a history of inflammatory bowel disease. ERCP is indicated when the
diagnosis of PSC is likely. Usually, the characteristic radiographic appearance of the biliary ducts is
confirmatory, and in some cases, therapy can be initiated if a dominant stricture is present (see
Chapter 65: Diagnosis and Management of Nonneoplastic Biliary Obstruction, Biliary Leakage, and
Disorders of the Liver Affecting the Bile Ducts). In this setting, ERCP also provides the opportunity to
obtain intraductal cytologic samples or biopsies in an effort to differentiate PSC from
cholangiocarcinoma. With the advent of liver transplantation, the early diagnosis of PSC and
elimination of the possibility of cholangiocarcinoma have added importance (see later).
Cholangitis
Since the introduction of endoscopic sphincterotomy in 1974, the overall indications for ERCP have
shifted from purely diagnostic to therapeutic.18,19(5631) This is illustrated by the former and current
approaches to cholangitis. Previously, ERCP was avoided because the addition of nonsterile contrast
material to bile, already infected, in an obstructed, confined space along with the associated increases
in intrabiliary pressure could result in septicemia. However, studies such as that of Leung et
al.20(5632) have de-monstrated that endoscopic drainage can be performed successfully and with a
mortality rate comparable with that of surgical control subjects. A study by Lai21(5633) indicated that
the hospital mortality rate was lower in an endoscopically treated group (10%) than in a surgically
treated group (32%). Data such as these, along with general practical experience, have converted
acute suppurative cholangitis from a relative contraindication to a specific indication for diagnostic and
therapeutic ERCP (see Chapter 62: Calculus Disease of the Bile Ducts).
Jaundice is relatively uncommon in patients with the acquired immunodeficiency syndrome (AIDS), but
the list of potential causes is long. In a study by Chalasani and Wilcox22(5634) of more than 500
patients, the most common causes were drug reactions and alcohol abuse. Other causes included
neoplasms and opportunistic infections. ERCP may be confirmatory or diagnostic in AIDS-related
cholangitis, a disorder that has the radiographic appearance of PSC and is closely associated with
infection by cryptosporidia, cytomegalovirus, and microsporidia (see Chapter 65: Diagnosis and
Management of Nonneoplastic Biliary Obstruction, Biliary Leakage, and Disorders of the Liver Affecting
the Bile Ducts).2327(5635) In the appropriate clinical setting, US was found by Daly and
Padley28(5636) to be highly accurate in predicting the presence or absence of abnormal findings on
retrograde cholangiography in patients with AIDS. These investigators suggested that ERCP adds little
relevant clinical information if the bile duct is normal by US. However, given the wide range of potential
causes of jaundice in patients with AIDS and the well-recognized difficulty of US diagnosis when the
bile ducts are not dilated, ERCP remains an appropriate study when the diagnosis is uncertain or
AIDS-related cholangitis is suspected.
Sphincter of Oddi Disorders
Disorders of the sphincter of Oddi, both functional and anatomic, provide another indication for ERCP.
Ampullary carcinoma may be suspected on clinical grounds, but the diagnosis is confirmed by
obtaining biopsy and cytology specimens at ERCP (see Chapter 69: Tumors of the Main Duodenal
Papilla). The procedure also provides an opportunity to evaluate the extent of disease within the
common duct and to perform therapeutic maneuvers for decompression or palliation. When sphincter
dysfunction is suspected, ERCP provides the means for placing a manometry catheter to measure
pressures within the sphincter, common duct, and pancreatic duct.29(5637) Sphincter of Oddi function
as well as dysfunction and its treatment are discussed in Chapter 67: Physiology and Pharmacology of
the Sphincter of Oddi, and Chapter 68: Sphincter of Oddi Stenosis and Dysfunction.
Stenosis of the main duodenal papilla has been described in patients with AIDS. Clinical features
include severe abdominal pain and marked elevations in serum alkaline phosphatase. In a study of 25
cases, Cello and Chan30(5638) found that endoscopic sphincterotomy provided symptomatic
improvement.
Laparoscopic Cholecystectomy
Laparoscopic cholecystectomy (LC) has become the procedure of choice for removal of the gallbladder
in most patients (see Chapter 97: Operative Laparoscopy).31(5639) The rapid development and
acceptance of this procedure have generated increased interest in the use of endoscopic (minimally
invasive) techniques and given rise to new indications for ERCP.32(5640)
Surgical Complications
As with most procedures, the complication rate for LC is directly related to experience; for the
individual surgeon, it is likely to be highest during his or her initial series of operations. Complications
include bile leak, bile duct injury, and postoperative stricture (see Chapter 65: Diagnosis and
Management of Nonneoplastic Biliary Obstruction, Biliary Leakage, and Disorders of the Liver Affecting
the Bile Ducts).3336(5641) A suspicion of a bile leak after cholecystectomy may be confirmed by a
variety of imaging techniques including scintigraphy and ERCP.37(5642) Although scintigraphy is
usually satisfactory for demonstrating the presence of a leak, ERCP has the further advantage of
therapeutic options. A number of studies have demonstrated that decompression and drainage by
means of endoscopic sphincterotomy with or without stent placement will result in closure of the
leak.34,38,39(5643) More serious bile duct injuries may also be a consequence of LC,40(5644)
including inadvertent transection, clipping, and fulguration. Although management of these problems is
often complex, ERCP and related therapeutic maneuvers may be of considerable assistance,
especially in the management of strictures.4143(5645) A suspicion during the postoperative course of
a stricture is usually an indication for ERCP with dilation and placement of one or more biliary stents.
Choledocholithiasis
The possibility of stones in the bile duct, whether pre- or post-LC, can be problematic. One point of
view, as advocated by Stiegmann et al.,44(5646) proposes that ERCP with stone removal before LC is
not superior to stone extraction either during LC or by ERCP immediately after LC. Others advocate
that ERCP and stone extraction be performed before LC for patients who meet certain clinical
criteria.45(5647) Unfortunately, there are no prospective studies or even uniformity of opinion as to the
exact criteria for selection of patients for ERCP. Although objective evidence of choledocholithiasis is
generally considered an indication for ERCP, the clinical findings, biochemical evidence, and
information derived from noninvasive imaging studies are frequently equivocal.
In general, the prevelance of bile duct stones is high in older patients with cholangitis, biliary
pancreatitis, jaundice, cholestatic liver enzyme abnormalities, or a US study that reveals biliary dilation
with or without a suspicion of ductal stones.4648(5648) However, patients with only mild clinical,
biochemical, or ultrasonographic abnormalities to suggest choledocholithiasis remain a clinical
dilemma. In a study of 1390 consecutive LC patients, Rieger and Wayand49(5649) found that the
combination of elevated liver function enzymes plus radiologic findings suggesting bile duct obstruction
had an accuracy of 69% in predicting the presence of ductal abnormalities. Either parameter alone was
less accurate; 69% for abnormal laboratory tests alone, 42% for radiographic findings alone. However,
Rijna et al.50(5650) found elevated liver function enzymes and bile duct dilation to be substantially less
helpful in a prospective study of 699 patients with symptomatic cholelithiasis. Intravenous infusion
cholangiography was found to be equivalent to endoscopic retrograde cholangiography in a
prospective comparison study of 111 nonicteric patients who were considered at risk for
choledocholithiasis.51(5651) However, in another study of 66 nonicteric patients with suspected bile
duct stones, intravenous cholangiography using meglumine iotroxate failed to demonstrate bile duct
stones in 14 of 27 cases confirmed by retrograde cholangiography.52(5652) Intravenous
cholangiography also gave one false-positive diagnosis. In the near future, EUS may become the
standard screening procedure for choledocholithiasis in this group of patients (see later and Chapter
78: Endoscopic Ultrasonography of the Retroperitoneal Organs).
Numerous studies have been published concerning the role of ERCP in patients about to undergo LC
because of symptomatic gallbladder disease.49,50,5362(5653) Unfortunately, most of the available
data are retrospective or derived from case series. Some authors favor the management of virtually all
patients by laparoscopic means, including the majority of those patients who are likely to have
choledocholithiasis.62(5654) However, the majority of reports describe selective use of ERCP before
LC when there is evidence to suggest the presence of stones in the bile duct.49,50,5361(5655)
At present, relatively few surgeons explore the bile duct laparoscopically. Where this is the case, it is
prudent that there be a reasonable degree of certainty before LC that choledocholithiasis is not
present. ERCP is therefore indicated before LC when the probability of choledocholithiasis is high as
predicted by clinical, biochemical, and radiologic factors. When the possibility of choledocholithiasis
appears to be remote, ERCP before LC will identify relatively few cases of clinically silent
stones.58(5656) This low yield must be considered together with the cost of the procedure and the
potential for a complication. Whereas the benefits of early preoperative ERCP in patients with severe
gallstone-induced pancreatitis or acute cholangitis are well established, routine preoperative ERCP to
"clear" the bile duct of stones cannot be justified except in patients who are at high risk for
choledocholithiasis.
Intraoperative cholangiography can be performed during laparoscopy via the cystic duct. Intraoperative
US may also be useful during laparoscopy to examine the bile duct, although experience is limited at
present (see Chapter 96: Laparoscopic Ultrasonography).63(5657) The technique of laparoscopic bile
duct exploration with stone extraction is difficult and therefore not widely practiced. When such
expertise is not available, ERCP and stone extraction with or without endoscopic sphincterotomy can
be performed post-LC when choledocholithiasis is demonstrated by laparoscopic
cholangiography.58,59(5658) ERCP post-LC minimizes cost and morbidity when the presence of
choledocholithiasis pre-LC is unlikely.64(5659)
The appropriate approach to the individual patient suspected of harboring bile duct stones depends in
the final analysis on the levels of experience of endoscopist and surgeon. With the further technical
development of laparoscopic bile duct exploration and increasing experience with this procedure, it is
likely that the indications for ERCP before LC will be modified.6567(5660) At present, however, ERCP
is indicated before LC when there is reason to suspect that choledocholithiasis is present and is
indicated after LC when it is either known or suspected that one or more stones have been left in the
bile duct. Choledocholithiasis is discussed in detail in Chapter 62: Calculus Disease of the Bile Ducts.
Orthotopic Liver Transplantation
Choledochostomy has become the preferred method of biliary reconstruction for orthotopic liver
transplantation (OLT). As with LC, OLT has widened the utilization of ERCP in patients with end-stage
hepatic and biliary diseases; indications include evaluation of the biliary tree before transplantation as
well as the diagnosis and management of complications of OLT.
ERCP is usually performed before OLT because of a suspicion that an additional but unrecognized
disorder of the bile ducts may be present, such as choledocholithiasis. Patients with PSC represent a
special category of candidates for OLT because of the increased incidence of cholangiocarcinoma
associated with this disorder and because of the difficulty of making the diagnosis of carcinoma
whether at ERCP or surgery. There are few reliable clinical indications that cholangiocarcinoma has
developed; sudden worsening of cholestasis, systemic symptoms such as weight loss, a rapidly
developing dominant stricture, especially with dilation of the proximal ducts, and the presence of a
polypoid mass are possible manifestations. However, all of these indications may be absent.
Furthermore, cholangiocarcinoma may be multifocal. Collection of specimens for cytopathologic
evaluation by brushing the extrahepatic bile duct has a relatively low sensitivity, and the presence of
cholangiocarcinoma may be overlooked.68(5661)
A wide range of biliary complications may occur after OLT, including fistula, anastomotic stricture,
multifocal strictures, choledocholithiasis, hemobilia, necrosis of the bile duct, and (presumably
transient) sphincter of Oddi dysfunction. Available data on ERCP in the diagnosis and management of
these problems are largely retrospective.6871(5662) Endoscopic intervention is likely to be beneficial
in patients with bile duct stones and many bile leaks. Its role in the management of anastomotic
strictures is less certain and may depend on unique factors in the individual case. For patients who
develop multiple biliary strictures after OLT, especially those with both extrahe-patic and intrahepatic
segments of narrowing, endoscopic therapy will probably prove to have little if any benefit.69,71(5663)
Pancreatic Diseases
Chronic pancreatitis may have features that can be detected by US or CT. Abdominal plain films may
be diagnostic if they show pancreatic calcifications. Because early ductal changes may not be apparent
on US, CT, or MRI, the diagnosis of chronic pancreatitis may be difficult or impossible using these
noninvasive imaging studies. Retrograde pancreatography is relatively sensitive for the diagnosis of
chronic pancreatitis (see also Chapter 73: Endoscopic Retrograde Cholangiopancreatography in Acute
and Chronic Pancreatitis). In one early study, the positive diagnostic accuracy of ERCP for duct
changes approached 100%.72(5664) When the presence of chronic pancreatitis can be confirmed by
CT or other imaging modalities, however, ERCP is not necessary for diagnosis.
At one time, it was believed that ERCP might be essential before surgery for chronic pancreatitis;
however, surgery can be accomplished successfully without ERCP.73,74(5665) Although not essential,
ERCP nevertheless remains useful in delineating the dimensions of the pancreatic ducts,
demonstrating communication with a pseudocyst, and determining the presence or absence of
strictures, obstruction, or pancreatic duct calculi before surgery. ERCP is therefore indicated for the
diagnosis of chronic pancreatitis when its presence cannot be confirmed by other less-invasive imaging
studies and for preoperative assessment of pathologic changes in pancreatic anatomy when the
diagnosis is already established. Although less by comparison with the biliary tree, there have been
considerable efforts to develop endoscopic approaches to the treatment of structural abnormalities
caused by pathologic processes involving the pancreas. Possibilities include dilation of strictures,
extraction of pancreatic stones, and placement of stents.75(5666) These evolving indications for ERCP
are discussed in Chapter 76: Endoscopic Management of Pancreatic Disease.
Acute Pancreatitis
Acute pancreatitis may also have features that can be detected by US or CT. These imaging modalities
can be used to detect enlargement of the pancreas or to follow the development or resolution of a
pancreatic phlegmon or pseudocyst.
The role of ERCP in acute pancreatitis is more problematic. Because pancreatitis is a recognized
potential complication of ERCP, the possibility exists that the procedure might aggravate or prolong an
episode of pancreatitis that is resolving. When the cause of acute pancreatitis is clear, ERCP is not
necessary. The vast majority of episodes of acute pancreatitis in the Western world are due to either
alcohol ingestion or gallstones; when alcohol can be excluded as a cause, the leading potential cause
is gallstone disease. Gallstone pancreatitis resolves in the majority of patients. However, an occasional
patient may have slowly resolving, persistent, or worsening pancreatitis; others may have recurrent
episodes of pancreatitis. ERCP has a role in the management of these patients.
Safrany and Cotton76(5667) reported their preliminary experience with urgent endoscopic
sphincterotomy for acute gallstone pancreatitis in 1981. A controlled trial reported in 1988 found that
endoscopic sphincterotomy reduced the morbidity and mortality in severe pancreatitis owing to
gallstone disease.77(5668) These findings have been subsequently confirmed by others.78(5669)
Therefore, a suspicion of gallstone-induced pancreatitis, especially in cases of severe acute
pancreatitis or unresolving pancreatitis, is an indication for at least diagnostic and probably therapeutic
ERCP during the acute episode (see also Chapter 73: Endoscopic Retrograde
Cholangiopancreatography in Acute and Chronic Pancreatitis).
The role of ERCP in the patient with a single episode of idiopathic acute pancreatitis is more
problematic. When the results of history and physical examination, noninvasive imaging studies such
as CT and US, and biochemical studies are negative, and the symptoms resolve in a matter of days
without complications, it is unlikely that ERCP will provide an incremental benefit.79(5670) When no
causative factors can be identified, the prognosis is good for the majority of patients.80(5671)
Pancreatic Trauma
ERCP will demonstrate the site of disruption of the main pancreatic duct in patients who have
sustained blunt trauma to the abdomen; sensitivity and specificity for this diagnosis approach
100%.81(5672) Disruption of the main pancreatic duct, when demonstrated, is usually an indication for
surgical drainage.82(5673) Improvements in CT and US may have made ERCP somewhat less
imperative, but it remains the most sensitive method of delineating pancreatic duct anatomy. The role
of ERCP in the diagnosis of pancreatic trauma is discussed in Chapter 75: Pancreatic Trauma,
Ascites, Fistula, and Pseudocyst.
Pancreatic Carcinoma
The pancreas remains the "hidden organ" in which disease processes, particularly carcinoma, may
remain undetected in their early stages. US, CT, and MRI have been variously reported as having
sensitivities as high as 94% for delineation of a mass in the pancreas but also failure rates as high as
10%.8386(5674) The major limitations of these methods of imaging relate to the size of the mass, its
location within the pancreas, and the presence of underlying pancreatic disease other than tumor. A
search for evidence of pancreatic carcinoma in a patient with a suggestive history but negative or
inconclusive results on other imaging modalities is one of the most common indications for ERCP.
Although most of the other noninvasive imaging techniques remain insensitive for identification of
smaller cancers, ERCP offers the best chance to detect subtle changes within the pancreatic ducts
suggestive of carcinoma (see Chapter 77: Endoscopic Diagnosis of Cancer of the
Pancreas).8789(5675) Collection from the pancreatic duct of specimens for cytologic evaluation may
also improve the positive diagnostic yield in carcinoma of the pancreas (see Chapter 74: Adjunct
Diagnosis of Pancreatic Disease and Pancreatic Physiology).90(5676) EUS also has a role in the
detection of small pancreatic tumors, although its accuracy in distinguishing carcinoma in the presence
of chronic pancreatitis remains uncertain (see Chapter 78: Endoscopic Ultrasonography of the
Retroperitoneal Organs).
Abdominal Pain
The role of ERCP in patients with abdominal pain that remains unexplained by less-invasive imaging
studies including upper gastrointestinal endoscopy is problematic. Nevertheless, the evaluation of
these patients is of-ten concluded with ERCP. Several studies have attempted to define the role of
ERCP in unexplained abdominal pain (see Chapter 73: Endoscopic Retrograde
Cholangiopancreatography in Acute and Chronic Pancreatitis).9193(5677) In the study of Ruddell et
al.,92(5678) 140 patients with undiagnosed severe chronic abdominal pain underwent ERCP. The
procedure was diagnostic in 24%, but there was a significant history of alcohol abuse in many patients
with abnormal pancreatograms. Although ERCP was diagnostic, chronic pancreatitis could have been
suspected on clinical grounds and, thus, the indication for ERCP could be considered as confirmatory
rather than the investigation of unexplained abdominal pain. Furthermore, 13 patients (9%) had either
peptic ulcer or gallstones, disorders that could have been diagnosed by other less-invasive means.
Chen et al.,93(5679) studied 86 patients with idiopathic chronic or recurrent abdominal pain.
Pancreatograms were normal in all patients, whereas cholangiograms were abnormal in 13 patients,
10 of whom had elevated levels of serum alkaline phosphatase or total bilirubin. Thus, in the absence
of clues to an organic cause, the usefulness of ERCP appears to be limited.
If, after careful history, physical examination, laboratory tests, and less-invasive diagnostic tests, a
complaint of abdominal pain remains perplexing, ERCP should be considered, especially if
laparoscopy or surgical exploration is also under consideration.
Miscellaneous Indications
Miscellaneous indications for ERCP include selective cannulation and aspiration for study of either
pancreatic juice or bile with or without hormonal stimulation. ERCP can also delineate anomalous
ductal anatomy (see Chapter 72: Congenital Anomalies of the Pancreas). With annular pancreas, the
pancreatic duct characteristically surrounds the duodenum.94(5680) Pancreas divisum is almost
exclusively diagnosed by ERCP. However, it is likely that the role of ERCP in pancreas divisum will be
altered by the further development of MRCP (see later).
ERCP is occasionally performed to remove stents left in the bile duct at surgery (e.g., liver
transplantation). A portion of a T-tube was removed in one reported case.95(5681)
ERCP may play a role in a variety of parasitic infestations of the biliary and pancreatic ductal
systems.96,97(5682) For example, ascaris worms can be removed from the bile and pancreatic
ducts.98(5683)
Contraindications to Diagnostic ERCP

In general, the same contraindications that apply to any endoscopic procedure also apply to ERCP.
Advanced age alone is not a contraindication.99(5684) Perhaps the only true absolute
contraindications to ERCP are refusal of the patient to undergo endoscopy or an acute unstable
cardiovascular or cardiopulmonary condition. ERCP with stone extraction has been performed
successfully during the recovery period after acute myocardial infarction; no complications occurred in
the four cases described by Cappell100(5685) in which the procedure was performed from 15 to 56
days after the acute event. In another study, tachycardia appeared to be a major factor in the
development of cardiac ischemia.101(5686)

As with any procedure, a relative contraindication must be carefully weighed in terms of risk versus
benefit. A ratio of risk:benefit that approaches 1 or becomes greater than 1 means that performance of
the procedure is unacceptable. When a patient is willing to undergo ERCP for a valid indication but is
unable to cooperate, general anesthesia or newer sedative drugs may be used to perform the
procedure. Structural abnormalities of the esophagus, stomach, or duodenum may be relative
contraindications to ERCP. This is especially true for esophageal obstructing lesions because passage
of the side-viewing duodenoscope is essentially a "blind" maneuver. A large esophageal diverticulum
and an unrecognized esophageal stricture are hazards that increase the risk of an esophageal tear or
perforation. In the former circumstance, ERCP can usually be accomplished after an adequate lumen
has been established by dilation; in the latter, it may be possible to avoid entering the diverticulum by
passing the duodenoscope under fluoroscopy. A large paraesophageal hiatus hernia also represents a
special hazard, particularly if it is not recognized before attempting to maneuver the endoscope into the
duodenum. Chronic gastric volvulus may be relatively asymptomatic and unsuspected before ERCP; it
may also be complex and associated with paraesophageal herniation of the stomach. Although very
infrequent, gastric volvulus represents a potential hazard if unrecognized.
Passage of a side-viewing instrument may be impossible in cases of pyloric stenosis or gastric outlet
obstruction owing to a variety of causes including carcinoma and peptic stricture. Extensive
retroperitoneal malignancy may result in deformity and narrowing of the duodenal bulb or descending
duodenum to the extent that ERCP cannot be performed. A previous partial gastrectomy with Billroth
type II anastomosis increases the technical difficulty of ERCP, especially when therapeutic maneuvers
are required (see Chapter 61: Special Papillotomy Techniques; see also Chapter 52: Endoscopy of the
Postoperative Upper Gastrointestinal Tract). The decision whether the procedure should be
undertaken in this situation should therefore be influenced by the levels of skill and experience of the
endoscopist. A Roux-en-Y jejunojejunostomy is not an absolute contraindication, although the success
rate of ERCP in this situation is low because of the long length of small intestine that must be traversed
and the acuity of the angle of the anastomosis.102(5687)
The presence of a pancreatic pseudocyst is not an absolute contraindication to ERCP. However, there
is always a possibility that a communicating pseudocyst will be converted to a pancreatic abscess, a
complication with a relatively high morbidity and mortality (see Chapter 75: Pancreatic Trauma,
Ascites, Fistula, and Pseudocyst). For this reason, it is advisable to limit the volume of contrast
material injected to the minimum needed to demonstrate the major features of the pathologic anatomy
of the pancreatic duct including the point of communication with the pseudocyst; it is not necessary to
fully opacify the pseudocyst. This potential for conversion of a pseudocyst to a pancreatic abscess
must always be weighed against the value of the information to be obtained (see later). Usually, this
information has its greatest relevance when surgery is to be performed for drainage of the pseudocyst.
Adverse Reactions to Contrast Media

Adverse reactions to the iodinated radiographic contrast media used during ERCP are not considered
to be contraindications to ERCP. As with all procedures, the risk:benefit ratio must be considered in
patients with a history of an adverse reaction.
Ionic media are the standard contrast agents used in radiologic procedures, including ERCP. Because
they dissociate in solution, ionic agents are characterized by high osmolality. Nonionic media may or
may not dissociate; those that do have greater numbers of iodine atoms compared with ionic agents.
Therefore, the osmolality of all nonionic contrast media is said to be low in comparison with that for
ionic agents.103(5688)
Absorption of iodinated contrast agents has been demonstrated in many clinical
studies;104107(5689) nevertheless, adverse reactions to contrast media during or after ERCP are
exceedingly rare, with only a few cases reported and even fewer of these were of a life-threatening
nature.108110(5690) Two categories of adverse reactions are recognized: anaphylactoid or

idiosyncratic and chemotoxic or nonidiosyncratic.103(5691) Anaphylactoid reactions can be
life-threatening with clinical features that include itching and hives, bronchospasm, laryngeal edema,
and shock. Chemotoxic reactions manifest as nausea and vomiting, cardiac arrhythmias, shock,
pulmonary edema, and renal failure. The risk of a chemotoxic reaction increases with increasing
dosage of the contrast agent.
Nonionic agents given intravenously cause fewer and less-severe adverse effects than their ionic
counterparts.111114(5692) However, they are much more expensive.103,115(5693) Unfortunately,
there are relatively few prospective data on the benefits of premedication with corticosteroids,
antihistamines, and other drugs. In an early prospective study, Lasser et al.116(5694) demonstrated
that the use of ionic media with corticosteroid pretreatment (methylprednisolone, 32 mg given orally at
12 and 2 hours before the procedure) served as a reasonable alternative to intravenous nonionic
contrast agent without loss of safety. However, in a more recent study by Greenberger and
Patterson,117(5695) patients with a history of generalized reactions to conventional media had
repeated reactions to the same media even with premedication. Thus, these investigators
recommended the use of a nonionic agent together with pretreatment with a combination of
corticosteroid and antihistamine drugs. Some patients in this study also received additional
pretreatment with ephedrine (25 mg by mouth), which these investigators had found to be beneficial in
an earlier study.118(5696)
In summary, adverse reactions to iodinated radiographic contrast agents administered during ERCP
are relatively rare and seldom life-threatening. In patients with documented severe reactions, however,
it is prudent to administer premedication with antihistamine and corticosteroid drugs and to use a
nonionic contrast agent.
ERCP in Pregnancy
Pancreaticobiliary diseases are relatively uncommon during pregnancy.119,120(5697) However,
pregnant women have an increased risk of gallstone formation and subsequent choledocholithiasis.
This may predispose them to cholangitis and pancreatitis that can occasionally become
life-threatening. Surgical intervention (cholecystectomy, CBD exploration) is preferably postponed until
after delivery because the rate of fetal loss is high.119,121(5698)
ERCP with endoscopic sphincterotomy and stone removal can be performed safely and effectively
during pregnancy.121(5699) Jamidar et al.120(5700) performed 29 ERCP procedures in 23 pregnant
women with symptoms of pancreaticobiliary diseases. Fifteen procedures were performed in the first
trimester, 8 in the second, and 6 in the third. Complications were pancreatitis (1 patient), spontaneous
abortion 3 months after ERCP (1 patient), and neonatal death (1 patient). The last two complications
had no apparent causal relationship with the procedure. There is a case report of successful
endoscopic management of a spontaneous perforation of the bile duct in a woman with
choledocholithiasis who was in the third trimester of pregnancy.122(5701)
When ERCP is performed in a pregnant woman, it is essential to keep the exposure of the fetus to
ionizing radiation to minimum levels. Measures to decrease the amount of fetal irradiation to nearly
undetectable levels include placement of lead shielding over the patient's abdomen, reduction of
fluoroscopy time to the minimum required to visualize the biliary tree and confirm sphincterotome
position, avoidance of spot radiographs, aspiration of bile to verify catheter position within the bile duct,
and capture of fluoroscopic images with a digital archival system rather than spot
radiography.119,121,123(5702)
Complications of ERCP
The complications of diagnostic ERCP can be divided into those that are potentially associated with
any type of endoscopy, such as perforation, complications related to the use of sedative and narcotic
analgesic drugs, and those particular to ERCP, such as pancreatitis and sepsis. With regard to general
complications, those related to conscious sedation and perforation require further comment whereas
the potential complications of endoscopy, such as aspiration and bleeding, are addressed elsewhere in
this text (see Chapter 37: Indications, Contraindications, and Complications of Upper Gastrointestinal
Endoscopy).
Complications Related to Conscious Sedation

Conscious sedation for diagnostic ERCP is similar to that for upper endoscopy or colonoscopy.
However, the potential for oversedation is greater during ERCP because patients are often elderly and
multiple doses of sedative and analgesic agents may be given during prolonged
procedures.124,125(5703) It should be assumed, therefore, that deep sedation with loss of protective
reflexes occurs during most if not all ERCP procedures. Arterial hypoxia is a common event during
ERCP.126,127(5704) This potential hazard is often compounded by the fact that the patient is lying
prone under the x-ray machine while the attention of nurse assistants and endoscopist is focused on
tasks related to the procedure. To reduce the risk of oversedation, rigorous monitoring of the patient is
mandatory. As a routine, at least two gastrointestinal assistants should be required for the procedure,
one assigned to the technical aspects of the procedure and the other to monitor the patient's vital
functions (see Chapter 57: Technique of Endoscopic Retrograde Cholangiopancreatography).
Additionally, arterial hypoxia can be prevented by presedation oxygenation.126(5705) Continuous
electrocardiographic monitoring is not routinely necessary, but it should be used for patients with
significant cardiopulmonary disease.
Perforation
Perforation of the gut wall with the tip of an endoscope is a very rare event.128(5706) Most
perforations related to ERCP occur during therapeutic manipulation such as papillotomy. For reasons
that are usually unclear, considerable difficulty may occasionally be encountered with passage of a
side-viewing instrument through the cricopharyngeus. Because passage of the duodenoscope is
essentially a blind maneuver, it is reasonable to believe that the risk of perforating a pyriform sinus is
probably increased in such cases. When this type of difficulty is encountered, the endoscopist should
proceed gently and with caution. Structural problems such as a Zenker diverticulum should be
excluded by endoscopy with a forward-viewing instrument or an x-ray contrast study of the pharynx and
proximal esophagus. If there is no structural defect, it is sometimes useful to pass a nasogastric tube
before the administration of sedative drugs. The tube, which serves as a guide, often facilitates
passage of the instrument. Because a patient-initiated swallow is usually helpful and frequently
necessary, difficulty with passage of the duodenoscope at this level can sometimes be traced to
excessive sedation or the overzealous use of topical agents for pharyngeal anesthesia.
Whenever there is a suspicion of upper gastrointestinal pathology that could complicate passage of a
side-viewing endoscope (e.g., dysphagia, even if mild), it is prudent to perform endoscopy with a
forward-viewing instrument or to obtain an upper gastrointestinal x-ray series before ERCP. This might
also be appropriate for the patient with prior upper gastrointestinal surgery, such as a Billroth type II
anastomosis. In the latter situation, the entire diagnostic procedure can sometimes be carried out using
a forward-viewing endoscope.
Perforation may occur in very rare instances as a result of trauma from the use of guidewires or other
accessories. Submucosal injection of contrast material is technically a limited form of perforation. This
usually resolves but may persist for an extended period of time around the main duodenal papilla such
that it is necessary to abort the procedure (see Chapter 57: Technique of Endoscopic Retrograde
Cholangiopancreatography). As a technical error, submucosal injection of contrast material relates to

the experience of the endoscopist.
Pancreatitis
Pancreatitis is the most common complication of diagnostic ERCP. It may be serious or even
fatal.129(5707) By comparison with retrograde injection of contrast material into the bile duct, where
clinical factors such as evidence of an obstructed biliary tree may alert the endoscopist to the
possibility of a complication, post-ERCP pancreatitis can be very difficult to predict and can occur
despite scrupulous adherence to correct technique during the instillation of contrast material into the
main pancreatic duct. Furthermore, in patients with cholangitis owing to infected bile, with or without an
obstructed biliary tree, the endoscopist can take some comfort in the availability of parenteral
antibiotics and endoscopic maneuvers to establish drainage of the biliary tree. In most cases, no form
of endoscopic intervention is available to prevent or ameliorate severe pancreatitis; rather, only
supportive care can be offered.
The exact mechanisms whereby postprocedure pancreatitis develops are not precisely defined.
However, certain risk factors have been delineated and some preventive measures may be helpful.
Incidence
According to a review by Sherman and Lehman,130(5708) the reported incidence of post-ERCP
pancreatitis ranges from 0 to 39.5%. In prospective series, the frequency is approximately 5%.
Variations in reported figures are due to multiple factors including differences in the operative definition
of pancreatitis used by the various investigators. The study of Nordback and Airo131(5709) pertained
to patients with acute necrotizing pancreatitis (incidence 0.5%), whereas Stanten and Frey132(5710)
defined post-ERCP pancreatitis as abdominal pain and nausea with a serum amylase greater than 300
U/L (denominator not reported). Freeman et al.133(5711) defined postprocedure pancreatitis as new or
worse abdominal pain that required more than one night of hospitalization plus a serum concentration
of pancreatic enzymes (amylase or lipase) that was two or more times the upper limit of normal.
Prospective studies usually employ more rigorous pre- and postprocedure testing and patient
monitoring, and this accounts in part for the higher reported incidence of post-ERCP pancreatitis in this
type of investigation. In retrospective series, underestimation of complications may occur because
patients may not report mild postprocedure symptoms. The reported frequency of post-ERCP
pancreatitis also varies depending on the vigor of postprocedural investigation. For example, Thoeni et
al.134(5712) performed CT scans on 31 patients after ERCP and found that 11 (35%) developed
radiologic signs of pancreatitis. Three patients with radiologically "severe" pancreatitis were
nonetheless asymptomatic.
Clinical assessment immediately after ERCP is unreliable for predicting which patients will develop
pancreatitis. However, Gottlieb et al.135(5713) found that serum levels of amylase (less than 276 U/L)
and lipase (less than 1000 U/L) at 2 hours postprocedure are useful for eliminating the possibility of
pancreatitis with negative predictive values of 0.97 and 0.98, respectively. Measurements of urinary
trypsinogen activation peptides at 4 hours after diagnostic or therapeutic ERCP in 107 patients did not
predict postprocedure pancreatitis in the study of Banks et al.136(5714)
Hyperamylasemia is frequent after ERCP and may occur even if retrograde pancreatography was not
performed. Skude et al.137(5715) measured serum amylase levels in 234 patients undergoing
endoscopy; 18 patients had endoscopy without catheter contact with the main duodenal papilla; in 17,
the papilla was traumatized but a pancreatogram was not obtained; in 25, the CBD alone was
opacified; and in 174, the main pancreatic duct was injected with or without retrograde injection of the
bile duct. Modest elevations in serum amylase occurred in patients who underwent endoscopy alone
and in patients in whom there was contact with the papilla without injection of the main pancreatic duct.
In a prospective study of 50 consecutive patients undergoing upper gastrointestinal endoscopy without
attempts to cannulate the main duodenal papilla, Pelletier et al.138(5716) found mild elevations of
serum amylase at 2 hours postprocedure in 9 patients (18%); hyperamylasemia was still present at 24
hours in 5 patients. However, the increase in serum amylase was found to be largely due to increased
S-type isoamylase, thereby indicating that the salivary glands were the source of the hyperamylasemia
in patients in whom endoscopy has been performed without contact with the papilla.
Contributing Factors
Important risk factors for the development of postprocedure pancreatitis include the volume of contrast
material injected, the injection pressure (including the related factor of opacification to the level of the
acinus or "acinarization"), and the degree of difficulty encountered during cannulation of the ampulla. It
is controversial as to whether the type of contrast material used to opacify the pancreatic ducts should
be considered a factor.
In the study of Skude et al.137(5717) noted earlier, the extent of opacification of the main pancreatic
duct correlated positively with serum amylase elevation. If the main pancreatic duct and primary side
branches were opacified, serum amylase elevations were, on average, five times the baseline value; if
secondary branches were opacified or acinarization occurred, the mean elevation was greater than
seven times baseline. The degree of hyperamylasemia did not correlate with the presence or absence
of pancreatic pathology. Hamilton et al.139(5718) found an overall frequency of post-ERCP
pancreatitis of 1.3%; the rate was higher in patients in whom acinarization occurred. In a study of 140
ERCP procedures, Roszler and Campbell140(5719) correlated the occurrence of pancreatitis with the
appearance of contrast material in the renal collecting system. In 26 patients with acinarization and no
urogram, pancreatitis developed in 4%. In 19 patients with acinarization plus urographic visualization,
the rate of pancreatitis was 53%. It has been suggested that the presence of a patent minor papilla
might decrease the risk of pancreatitis, as it provides a second avenue of decompression or drainage
of contrast material injected via the major papilla.141(5720) Although an interesting concept, data to
support this hypothesis are lacking. In a prospective comparative study, Chen et al.142(5721) found
that patients with a history of prior episodes of pancreatitis were less likely to develop severe
pancreatitis after ERCP compared with patients with no previous history of pancreatitis.
Mechanical trauma to the papilla during cannulation with resulting edema, sphincter spasm, and
restriction of flow is a well-recognized pathogenetic mechanism for pancreatitis.130(5722) A
commonplace scenario for the development of acute pancreatitis is the procedure that is prolonged by
unsuccessful attempts to achieve a difficult CBD cannulation but with repeated, albeit inadvertent,
cannulation of the main pancreatic duct. The risk:benefit ratio of further attempts to opacify the CBD
must be considered with the recognition that the risk of pancreatitis increases in this situation.
The addition of sphincter of Oddi manometry to diagnostic ERCP increases the frequency of
post-ERCP pancreatitis.143(5723) Two studies by Sherman et al.144,145(5724) have demonstrated
that this risk can be decreased from as high as 31% to approximately 4% by using a special aspirating
catheter for obtaining manometric measurements.
The standard contrast agents used for diagnostic ERCP (see earlier) are iodinated and of relatively
high osmolality (1400 to 1500 mOsm/kg H2O). There has been recent interest in the use of lower
osmolality or nonionic contrast media, agents with physical properties that might be less likely to cause
ERCP-related pancreatitis. These appear to cause less ductal inflammation in animal
models.146(5725) However, they may be up to 20 times more expensive than established contrast
agents.115(5726)
A relatively large number of clinical trials have been conducted in an effort to determine whether the
use of an alternative contrast agent reduces the risk of ERCP-related pancreatitis.147154(5727) The
numbers of patients included in many of these studies are small, although several authors have
reported that various agents decrease the risk of pancreatitis.147149(5728) However, when these
data are examined carefully, it is not conclusive that these agents provide a clear
advantage.148152(5729)
Cunliffe et al.149(5730) found that pancreatitis was more frequent when retrograde pancreatography
was performed with a high-osmolality contrast agent compared with one of lower osmolality. However,
a higher volume of contrast material was injected and the rate of acinarization was greater in those
patients who underwent pancreatography with the contrast media of higher osmolality. Barkin et
al.148(5731) also found higher serum amylase and lipase levels and more evidence of biochemical
pancreatitis when a high-osmolality agent was used as compared with low-osmolality contrast material;
however, in this study, the overall rate of secondary branch filling or acinarization was 49%. In addition,
studies of relatively large numbers of patients have found that the use of low-osmolarity contrast media
does not decrease the incidence or severity of post-ERCP pancreatitis.155,156(5732) Given the
increased cost and the lack of definitive proof that these alternative contrast agents are advantageous,
their routine use is not warranted.103(5733)
Prevention
There have been attempts to develop preventive measures, largely through the use of pharmacologic
agents, to reduce the risk of ERCP-related pancreatitis.
Somatostatin reduces pancreatic exocrine secretion. It has therefore been administered
prophylactically to patients undergoing ERCP in an attempt to decrease the incidence or severity of
postprocedural pancreatitis. However, several trials have failed to demonstrate any benefit from the
use of this agent;157160(5734) in one, the rate of pancreatitis in patients receiving somatostatin was
higher than that for a comparable control group.161(5735) A study of the somatostatin analog,
octreotide, found that this agent did not prevent post-ERCP pancreatitis.162(5736) Based on currently
available data, the administration of somatostatin before ERCP as prophylaxis against pancreatitis
cannot be recommended.
Nifedipine, a drug that relaxes the sphincter of Oddi, may reduce epigastric pain induced by ERCP, but
a prospective, randomized trial in patients undergoing ERCP with or without endoscopic
sphincterotomy failed to demonstrate any benefit with respect to prevention of acute
pancreatitis.163(5737)
A retrospective study of patients undergoing ERCP who received corticosteroid therapy because of a
history of iodine sensitivity showed a decrease in the overall incidence of post-ERCP
pancreatitis.164(5738) However, a prospective, randomized control trial demonstrating benefit is
needed before this form of prophylaxis can be recommended.
Some new pharmacologic agents, such as the protease inhibitor, gabexate (Foy, Lepetit, Lainate,
Italy), appear promising in relation to the prevention of postprocedure pancreatitis. The multicenter,
double-blind trial of Cavallini et al.,165(5739) in which 435 patients underwent randomization, found
that prophylactic administration of gabexate before ERCP and for 12 hours after the procedure
resulted in significant decreases in postprocedure hyperamylasemia, pancreatic pain, and clinical
pancreatitis.
Unfortunately, ERCP-induced pancreatitis is not completely unavoidable. However, the risk can be
reduced to a minimum if there is meticulous adherence to the principles of proper technique. Some of
these are listed in Table 562 (see also Chapter 57: Technique of Endoscopic Retrograde
Cholangiopancreatography).
Tactics to Decrease the Risk of

Postprocedural Pancreatitis
TABLE 562
Tactics to Decrease the Risk of

Postprocedural Pancreatitis
TABLE 562
Keep the indication for the procedure in mind and opacify the
main pancreatic duct only when clinically relevant. If the
objective is common bile duct opacification alone, selective
cannulation is preferable.
Instill only that volume of contrast material necessary to assess
the ductal system.
Limit the volume of contrast material injected into the pancreatic
duct if the duct is inadvertently opacified.
If opacification of the main pancreatic duct cannot be made
except forcefully, assess technical factors, especially catheter
position. The catheter may be impacted against a side wall or
placed in a primary branch of the main pancreatic duct.
Repositioning may avoid submucosal injection or overfilling of
distal side branches and allow better control of the injection.
When studying the main pancreatic duct in the tail
fluoroscopically during contrast material injection, repeatedly
and frequently assess the main pancreatic duct in the head
and body to be certain that overfilling is not occurring in these
areas.
If common bile duct injection or cannulation is desired but
difficult, avoid repeated filling of the main pancreatic duct.
Consider the risk:benefit ratio to the patient.
Pancreatic Sepsis
Serious infections of the pancreas may arise if contaminated equipment is used in the performance of
ERCP.130,166168(5740) A case has been reported of multiple pancreatic abscesses due to Candida
albicans that were attributed to ERCP.169(5741)
The role of ERCP in the preoperative management of patients with chronic pancreatitis or nonresolving
pseudocysts is well established.170(5742) However, the procedure is often undertaken with a certain
degree of trepidation for fear of converting a sterile pseudocyst to one that is infected. Kolars et
al.171(5743) retrospectively reviewed 51 cases of pancreatic pseudocysts; preoperative ERCP was
performed in 42 cases. A communication with the main pancreatic duct was present in 69%. Although
59 pseudocysts were infected when cultured, preoperative ERCP apparently did not increase the risk
of infection. However, pancreatograms were performed very carefully in this study; in particular, the
injection of contrast material was halted as soon as evidence of filling of the cavity was demonstrated.
As with any diagnostic study, the need for pancreatography should be considered carefully before the
procedure. In our practices, endoscopic pancreatography is performed in the setting of a nonresolving
pseudocyst to assess pancreatic ductal anatomy before surgery, preferably with the operation being
scheduled within 24 hours of the procedure. It may be prudent to administer broad-spectrum antibiotics
before ERCP to decrease the risk of bacteremia, with the recognition that there is unlikely to be
significant penetration of the pseudocyst and that antimicrobial agents are not a substitute for definitive
drainage. It may also be advisable to administer antibiotics intravenously before ERCP in cases of
suspected pancreatic ascites or pancreatic duct trauma.
Biliary Sepsis
Another serious, potentially fatal complication of ERCP is cholangitis or biliary sepsis. Factors
predisposing to postprocedural cholangitis include biliary obstruction and infected bile as well as
contaminated endoscopes and accessories.
Biliary Obstruction
An obstructed biliary tree was present in 90% of patients who developed post-ERCP cholangitis in the
large survey of Bilbao et al.16 Deviere et al.172 found that 52 of 55 patients with post-ERCP
septicemia had bile duct obstruction; sepsis was more frequent in patients with strictures owing to
malignancy as opposed to other causes, and those with hilar strictures were more likely to develop
cholangitis than those with more distal obstruction. In another study, factors that were more likely to
lead to postprocedural cholangitis included malignant hilar stricture, incomplete drainage after contrast
material injection, and difficult cases in which prolonged or multiple procedures were necessary to
establish drainage.173(5744)
Only the smallest volume of contrast material necessary to demonstrate pathologic abnormalities
should be injected when the biliary system is occluded. Injection of large volumes of contrast material
into the dilated or obstructed ducts proximal to a stricture increases the risk of a septic complication.
Once the essential features of a stricture, such as location and length, have been demonstrated, it is
advisable to establish drainage of the biliary system proximal to the obstruction as soon as possible,
preferably at the time of ERCP.
Infected Bile
Cholangitis can also occur after ERCP when obstruction of the bile duct is due to causes other than
neoplasms, including choledocholithiasis. When ductal instrumentation is being considered in patients
with bile duct stones, it should be assumed that the bile is already infected. Sand et al.174(5745) found
infected bile in 88% of patients with CBD stones before ERCP; two or more different bacteria were
identified in 67%. Leung et al.20(5746) reported a 97% rate of successful treatment with endoscopic
drainage in patients with cholangitis owing to choledocholithiasis. Escherichia coli, Klebsiella,
Enterococci, Enterobacter, and Bacteroides were identified in various patients in this series.
Colonized Endoscopes
The most common organisms found by Deviere et al.172(5747) in blood cultures from 55 patients with
ERCP-related sepsis were Pseudomonas aeruginosa and E. coli. These investigators attributed sepsis
due to the latter bacterium to faulty cleaning and disinfection of endoscopes and
accessories.172,173(5748) Consecutive infections with P. aeruginosa or Serratia marcescens have
been linked to con-taminated endoscopes or water bottles.175177(5749) Thus, when cholangitis,
sepsis, or pancreatic abscess develops secondary to organisms such as Pseudomonas or Serratia,
contamination of endoscopic equipment should be suspected. Proper techniques for mechanical
cleaning, disinfection, storage of endoscopes, and sterilization of accessories (especially water bottles)
is therefore mandatory (see Chapter 8: Disinfection of Endoscopes and Accessories).
Antibiotics
Most experts advise prophylactic administration of broad-spectrum antibiotics before ERCP in cases of
biliary obstruction; this should include coverage for P. aeruginosa if there has been a history of recent
ERCP, particularly if drainage was not established at a previous procedure. However, there is no
evidence that antibiotics alone effectively prevent postprocedure cholangitis.173,178,179(5750)
Furthermore, the addition of antibiotics to contrast media appears to offer little increased
benefit.180(5751)
Byl et al.181(5752) randomized 82 patients with biliary obstruction from various causes but no clinical
evidence of cholangitis to receive antibiotic prophylaxis (piperacillin) or placebo before diagnostic or
therapeutic ERCP. Treatment was maintained in both patient groups until biliary drainage was
achieved. Of the 68 evaluable patients who completed the trial, 32 of 34 patients (94%) who received
antibiotic prophylaxis had no clinical evidence of cholangitis or septicemia. In the placebo group, 24 of
34 patients (71%) remained asymptomatic.
In the double-blind study of van den Hazel et al.,182(5753) a total of 551 patients undergoing ERCP for
suspected biliary tract obstruction were randomized to receive antibiotic prophylaxis with piperacillin (4
g intravenously 30 minutes before the procedure) or placebo. Retrograde cholangiograms were normal
in 113; causes of obstruction included choledocholithiasis (147 patients) and malignant distal strictures
(203 patients). Among the 270 patients who received piperacillin, 12 developed acute cholangitis
compared with 17 of 281 patients who were given the placebo. Except for 1 case in the
antibiotic-treated group, all episodes of cholangitis were regarded as either mild or moderate in
severity. Data obtained in this study indicated that prophylaxis with a single intravenous dose of
piperacillin did not significantly reduce the incidence of clinical acute cholangitis in patients with
choledocholithiasis or distal CBD strictures.
Because it is unlikely or at best uncertain that therapeutic concentrations of antibiotics can be achieved
in the obstructed biliary tree, it is for practical purposes impossible to render the bile sterile. Therefore,
it cannot be assumed that prophylactic antibiotics will substantially reduce the risk of cholangitis.
Moreover, the use of antimicrobial agents is no substitute for prompt drainage of an infected enclosed
space. Whenever there is a possibility of biliary obstruction, the endoscopist must always be prepared
to establish drainage if obstruction is in fact demonstrated at ERCP. When there is any suspicion of
obstruction, it is advisable that a "diagnostic" ERCP not be performed if the endoscopist is not fully
trained and capable of establishing proper drainage of an obstructed system.
Miscellaneous Complications
Hemorrhage occurs very rarely in diagnostic ERCP. There are a few reports of splenic trauma during
ERCP;183185(5754) one case was complicated by the development of splenic abscess.186(5755)
Rare postdiagnostic ERCP complications include acute and emphysematous cholecystitis187(5756)
and hepatic subcapsular biloma.188(5757)
Radiation Exposure
When the fiber bundles of fiberoptic endoscopes are exposed to radiation, their optical density
increases, thereby resulting in a reduction in light transmission. The loss of light has a direct relation to
the total exposure dose of radiation. The degree of damage becomes severe enough to reduce the
level of light transmission to less than the minimum acceptable for performing ERCP after about 1200
cases, thereby defining one measure of the life expectancy of a duodenoscope.189(5758)
Radiation dosages in patients undergoing ERCP are comparable with other standard radiographic
procedures. Patient entrance dose after ERCP was less than that from abdominal angiography and
comparable with conventional gastrointestinal radiologic examinations.190,191(5759)
Radiation exposure to endoscopists and assistants during ERCP is related to duration of fluoroscopy
time.192,193(5760) This exposure is usually minimum190,193196(5761) and can be decreased to
negligible amounts by employing simple measures such as placement of protective shields around the
x-ray apparatus,192,194(5762) minimizing fluoroscopy time, and using higher fluoroscopy
voltages.193(5763) Eye (lens) and thyroid gland radiation dosages to personnel have been determined
to be extremely low, suggesting that protective measures such as lead-shielded glasses and thyroid
shields are apparently not necessary.196(5764)
An overview of radiation safety during endoscopy is summarized in a publication from the American
Society for Gastrointestinal Endoscopy.197(5765)
ERCP and New Imaging Procedures

The accuracy of EUS is comparable with that of ERCP for diagnosis of the cause of obstruction in
patients with obstructive jaundice or dilation of bile ducts (see Chapter 78: Endoscopic
Ultrasonography of the Retroperitoneal Organs). Both tests are superior to CT and US; the reported
accuracy ranges from 29 to 95% for US and 20 to 66% for CT compared with 97% for ERCP and 74 to
100% for EUS.198(5766) For detection of small biliary tumors, that is, those less than 3 cm in
diameter, EUS was equivalent to ERCP in one study.199(5767)
A number of studies using radial-type echoendoscopes have shown EUS to be highly accurate
(approaching 100%) for detection of bile duct stones.200203(5768) For this diagnosis, it is as
sensitive and specific as ERCP.204,205(5769) In a study involving 422 patients who were evaluated
for choledocholithiasis, all CBD stones found by ERCP were also seen by EUS.203(5770) ERCP is
obviously the procedure of choice in patients with known choledocholithiasis. However, even with
careful selection, about half to three quarters of patients with suspected choledocholithiasis referred for
ERCP do not have stones.200202,206(5771) EUS is less costly than ERCP and avoids the potential
morbidity of pancreatitis.
EUS compares favorably with ERCP in detecting small and large pancreatic cancers.207(5772) It is
valuable for pancreatic cancer staging208(5773) and could prove to have a role in the diagnosis of
chronic pancreatitis.209(5774)
EUS is an evolving imaging modality that is effective in the diagnosis of many benign and malignant
lesions of the pancreas and biliary tree. EUS might replace ERCP as the diagnostic test of choice for
screening patients with suspected choledocholithiasis and for detecting smaller pancreaticobiliary
tumors, given its sensitivity and lower rate of complications. ERCP would then be performed only for
therapy. With continued technical development and wider availability, EUS will likely transform the
indications for ERCP. Some clinical scenarios in which EUS may be preferable to ERCP are
summarized in Table 563.
Potential Roles for EUS Before

Therapeutic ERCP
TABLE 563
Patients with nondilated bile ducts on US who have marginally abnormal

liver function tests before laparoscopic cholecystectomy.
Patients with a relative contraindication to ERCP such as contrast
allergy or pregnancy.
Patients with a history of post-ERCP pancreatitis.
Patients with suspected bile duct stones in whom ERCP is
unsuccessful.
Patients who have undergone cholecystectomy who are considered to
be at low or medium risk (20%) for bile duct stones because of mildly
elevated liver function tests, minimum or no biliary dilation and no
remote history of jaundice or pancreatitis.
Potential Roles for EUS Before

Therapeutic ERCP
TABLE 563
Patients in whom ERCP is difficult, to define the level of invasiveness

that would be warranted. The finding of a normal biliary tree on EUS
may allow the endoscopist to curtail prolonged attempts at cannulation
and to stop short of aggressive measures to achieve bile duct
opacification including needle-knife and precut endoscopic
sphincterotomy.
EUSendoscopic ultrasonography; ERCPendoscopic retrograde
cholangiopancreatography; USultrasonography.
Magnetic Resonance Cholangiopancreatography

MRCP is a new, noninvasive method of imaging both the biliary tree and the pancreatic ductal system
without the use of contrast agents or radiation. Image processing can also provide three-dimensional
views of pancreaticobiliary anatomy.
Numerous studies have demonstrated that MRCP compares favorably with ERCP for detection of
choledocholithiasis as well as the evaluation of biliary obstruction and various pancreatic
disorders.210218(5775) MRCP clearly depicts the pancreatic duct in patients with chronic pancreatitis
and can demonstrate narrowing, dilation, and filling defects.219(5776) In addition to having a very high
sensitivity and specificity for the detection of various biliary disorders,220(5777) MRCP also accurately
displays anatomic variations of the biliary tree221(5778) and pancreas divisum.222(5779) However,
the limited spatial resolution of MRCP curtails its role in the characterization of bile duct stenosis, the
visualization of small intraampullary tumors, and the diagnosis of chronic pancreatitis.223(5780)
Various imaging sequences of MRCP are being used to enhance images and obtain them more
rapidly. Although the exact role of MRCP in the evaluation of pancreaticobiliary diseases has not been
fully defined, continued advances in technique and increasing levels of expertise are beginning to
clarify its potential. MRCP is likely to become especially useful in patients in whom ERCP is
contraindicated, unsuccessful, or incomplete.224(5781) MRCP may also narrow further the use of
ERCP to therapeutic indications.225(5782)
REFERENCES
1.
2.
3.
4.
5.
6.
7.
McCune WS, Shorb PE, Moscovitz H. Endoscopic cannulation of the ampulla of vater: A
preliminary report. Ann Surg 1968;167:7526.
Matzen P, Malchow-Moller A, Brun B, et al. Ultrasonography, computed tomography, and
cholescintigraphy in suspected obstructive jaundice: A prospective comparative study.
Schenker S, Balint J, Schiff L. Differential diagnosis of jaundice: Report of a prospective study
of 61 proved cases. Am J Dig Dis 1962;12:44963.
Frederic N, Deltenre M, d'Hondt M, et al. Comparative study of ultrasound and ERCP in the
diagnosis of hepatic, biliary and pancreatic diseases: A prospective study based on a
continuous series of 424 patients. Eur J Radiol 1983;3:20811.
Cronan JJ. US diagnosis of choledocholithiasis: A reappraisal. Radiology 1986;161:1334.
Laing FC, Jeffrey RB, Wing VW. Improved visualization of choledocholithiasis by sonography.
AJR 1984;143:94952.
Graham MF, Cooperberg PL, Cohen MM, et al. Ultrasonography screening of the common
hepatic duct in symptomatic patients after cholecystectomy. Radiology 1981;138:1379.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
Mueller PR, Ferrucci JT Jr, Simeone JF, et al. Postcholecystectomy bile duct dilation: Myth or
reality?. AJR 1981;136:3558.
Greenwald RA, Pereiras R Jr, Morris SJ, Schiff ER. Jaundice, choledocholithiasis, and a
nondilated common duct. JAMA 1978;240:19834.
Lapis JL, Orlando RC, Mittelstaedt A, et al. Ultrasonography in the diagnosis of obstructive
jaundice. Ann Intern Med 1978;89:613.
Dewbury KC, Joseph AEA, Hayes S, Murray C. Ultrasound in the evaluation and diagnosis of
jaundice. Br J Radiol 1979;52:27680.
Goldberg HI, Filly RA, Korobkin M, et al. Capability of CT body scanning and ultrasonography
to demonstrate the status of the biliary ductal system in patients with jaundice. Radiology
1978;129:7317.
Fleischmann D, Ringl H, Schofl R, et al. Three-dimensional spiral CT cholangiography in
patients with suspected obstructive biliary disease: Comparison with endoscopic retrograde
cholangiography. Radiology 1996;198:8618.
Choi SH, Pralzer FA Jr, Park HC. Percutaneous transhepatic cholangiography with thin
needle: Recent update. J Med 1981;12:14758.
Kreek MJ, Balint JA. "Skinny needle" cholangiographyResults of a pilot study of a voluntary
prospective method for gathering risk data on new procedures. Gastroenterology
1980;78:598604.
Bilbao MK, Dotter CT, Lee TG, Katon RM. Complications of endoscopic retrograde
cholangiopancreatography (ERCP). A study of 10,000 cases. Gastroenterology
1976;70:31420.
Pasanen PA, Partanen KP, Pikkarainen PH, et al. A comparison of ultrasound, computed
tomography and endoscopic retrograde cholangiopancreatography in the differential diagnosis
of benign and malignant jaundice and cholestasis. Eur J Surg 1993;159:239.
Kawai K, Akasaka Y, Murakami K, et al. Endoscopic sphincter-otomy of the ampulla of vater.
Classen M, Demling L. Endoskopische Sphinkterotomie der Papilla Vateri und Steinextraktion
aus dem Ductus Choledochus. Dtsch Med Wochenschr 1974;99:4967.
Leung JW, Chung SCS, Sung JJY, et al. Urgent endoscopic drainage for acute suppurative
cholangitis. Lancet 1989;1:13079.
Lai ECS, Mok FPT, Tan ESY, et al. Endoscopic biliary drainage for severe acute cholangitis.
N Engl J Med 1992;326:15826.
Chalasani N, Wilcox CM. Etiology, evaluation, and outcome of jaundice in patients with
acquired immunodeficiency syndrome. Hepatology 1996;23:72833.
Bird GL, Kennedy DH, Forrest JA. AIDS-related cholangitis: Diagnostic features and course in
four patients. Scott Med J 1995;40:534.
Dowsett JF, Miller R, Davidson R, et al. Sclerosing cholangitis in acquired immunodeficiency
syndrome. Scand J Gastroenterol 1988;23:126774.
Cappell MS. Hepatobiliary manifestations of the acquired immunodeficiency syndrome. Am J
Ducreux M, Buffet C, Lamy P, et al. Diagnosis and prognosis of AIDS-related cholangitis.
AIDS 1995;9:87580.
Willson R, Harrington R, Stewart B, Fritsche T. Human immu-nodeficiency virus 1-associated
necrotizing cholangitis caused by infection with Septata intestinalis. Gastroenterology
1995;108:24751.
Daly CA, Padley SP. Sonographic prediction of a normal or abnormal ERCP in suspected
AIDS-related sclerosing cholangitis. Clin Radiol 1996;51:61821.
Schmalz MJ, Geenen JE, Hogan WJ, et al. Pain on common bile duct injection during ERCP:
Does it indicate sphincter of Oddi dysfunction?. Gastrointest Endosc 1990;36:45861.
Cello JP, Chan MF. Long-term follow-up of endoscopic retrograde cholangiopancreatography
sphincterotomy for patients with acquired immune deficiency syndrome papillary stenosis. Am
J Med 1995;99:6003.
31. . NIH Consensus Conference on gallstones and laparoscopic cholecystectomy. Am J Surg

1993;165:387547.
32. Legorreta AP, Brooks RJ, Staroscik RN, et al. Increased ERCP rate following the introduction
of laparoscopic cholecystectomy. J Laparoendosc Surg 1995;5:2718.
33. Raijman I, Catalano MF, Hirsch GS, et al. Endoscopic treatment of biliary leakage after
laparoscopic cholecystectomy. Endoscopy 1994;26:7414.
34. Himal HS. The role of ERCP in laparoscopic cholecystectomy-related cystic duct stump leaks.
35. Pencev D, Brady PG, Pinkas H, Boulay J. The role of ERCP in patients after laparoscopic
cholecystectomy. Am J Gastroenterol 1994;89:15237.
36. Kent AL, Cox MR, Wilson TG, et al. Endoscopic retrograde cholangiopancreatography
following laparoscopic cholecystectomy. Aust N Z J Surg 1994;64:40712.
37. Cotton PB, Baillie J, Pappas TN, Meyers WS. Laparoscopic cholecystectomy and the biliary
endoscopist. Gastrointest Endosc 1991;37:947.
38. Azurin DJ, Go LS, Maslack M, Kirkland ML. Bile leak following laparoscopic cholecystectomy.
J Laparoendosc Surg 1995;5:2336.
39. Kupferschmidt H, Havelka J, Schwery S, et al. Endoscopic therapy of bile leakage following
laparoscopic cholecystectomy. Schweiz Med Wochenschr Suppl 1996;79:89S-93S.
40. Deziel DJ, Millikan KW, Economou SG, et al. Complications of laparoscopic cholecystectomy:
A national survey of 4,292 hospitals and an analysis of 77,604 cases. Am J Surg
1993;165:914.
41. Bergman JJ, van den Brink GR, Rauws EA, et al. Treatment of bile duct lesions after
laparoscopic cholecystectomy. Gut 1996;38:1417.
42. Bezzi M, Silecchia G, Orsi F, et al. Complications after laparoscopic cholecystectomy.
Coordinated radiologic, endoscopic, and surgical treatment. Surg Endosc 1995;9:2936.
43. Borowicz MR, Adams DB, Simpson JP, et al. Management of biliary strictures due to
laparoscopic cholecystectomy. J Surg Res 1995;58:869.
44. Stiegmann GV, Goff JS, Mansour A, et al. Precholecystectomy endoscopic cholangiography
and stone removal is not superior to cholecystectomy, cholangiography, and common duct
exploration. Am J Surg 1992;163:22730.
45. Lilemoe KD, Yeo CJ, Tallimini MA, et al. Selective cholangiography: Current role in
laparoscopic cholecystectomy. Ann Surg 1992;215:66974.
46. Chan AC, Chung SC, Wyman A, et al. Selective use of preoperative endoscopic retrograde
cholangiopancreatography in laparoscopic cholecystectomy. Gastrointest Endosc
1996;43:2125.
47. Robertson GS, Jagger C, Johnson PR, et al. Selection criteria for preoperative endoscopic
retrograde cholangiopancreatography in the laparoscopic era. Arch Surg 1996;131:8994.
48. Barkun AN, Barkun JS, Fried GM, et al. Useful predictors of bile duct stones in patients
undergoing laparoscopic cholecystectomy. McGill Gallstone Treatment Group. Ann Surg
1994;220:329.
49. Rieger R, Wayand W. Yield of prospective, noninvasive evaluation of the common bile duct
combined with selective ERCP/sphincterotomy in 1390 consecutive laparoscopic
cholecystectomy patients. Gastrointest Endosc 1995;42:612.
50. Rijna H, Borgstein PJ, Meuwissen SG, et al. Selective preoperative endoscopic retrograde
cholangiopancreatography in laparoscopic biliary surgery. Br J Surg 1995;82:11303.
51. Bloom IT, Gibbs SL, Keeling-Roberts CS, Brough WA. Intravenous infusion cholangiography
for investigation of the bile duct: A direct comparison with endoscopic retrograde
cholangiopancreatography. Br J Surg 1996;83:7557.
52. Tham TC, Collins JS, Watson RG, et al. Diagnosis of common bile duct stones by
intravenous cholangiography: Prediction by ultrasound and liver function tests compared with
endoscopic retrograde cholangiography. Gastrointest Endosc 1996;44:15863.
53. Welbourn CR, Mehta D, Armstrong CP, et al. Selective preoperative endoscopic retrograde
cholangiography with sphincterotomy avoids bile duct exploration during laparoscopic
cholecystectomy. Gut 1995;37:5769.

54. Chang KK, Mo LR, Yau MP, et al. Endoscopic sphincterotomy prior to laparoscopic
cholecystectomy for the treatment of cholelithiasis. Hepatogastroenterology 1996;43:2036.
55. Liu CL, Lai E CS, Lo CM, et al. Combined laparoscopic and endoscopic approach in patients
with cholelithiasis and choledocholithiasis. Surgery 1996;119:5347.
56. Balsara KP, Shah CR, Maydeo A. Endoscopic retrograde cholangiography in the era of
laparoscopic cholecystectomy. Indian J Gastroenterol 1996;15:101.
57. Bonatsos G, Leandros E, Polydorou A, et al. ERCP in association with laparoscopic
cholecystectomy. A strategy to minimize the number of unnecessary ERCPs. Surg Endosc
1996;10:3740.
58. Changchien CS, Chuah SK, Chiu KW. Is ERCP necessary for symptomatic gallbladder stone
patients before laparoscopic cholecystectomy?. Am J Gastroenterol 1995;90:21247.
59. Kum CK, Goh PM. Preoperative ERCP in the management of common bile duct stones
before laparoscopic cholecystectomy. Eur J Surg 1996;162:20510.
60. Miller RE, Kimmelstiel FM, Winkler WP. Management of common bile duct stones in the era
of laparoscopic cholecystectomy. Am J Surg 1995;169:2736.
61. Coppola R, D'Ugo D, Ciletti S, et al. ERCP in the era of laparoscopic biliary surgery.
Experience with 407 patients. Surg Endosc 1996;10:4036.
62. Phillips EH, Liberman M, Carroll BJ, et al. Bile duct stones in the laparoscopic era. Is
preoperative sphincterotomy necessary?. Arch Surg 1995;130:8805; discussion 8856.
63. Santambrogio R, Bianchi P, Opocher E, et al. Intraoperative ultrasonography (IOUS) during
laparoscopic cholecystectomy. Surg Endosc 1996;10:6227.
64. Erickson RA, Carlson B. The role of endoscopic retrograde cholangiopancreatography in
patients with laparoscopic cholecystectomies. Gastroenterology 1995;109:25263.
65. Berci G, Morgenstern L. Laparoscopic management of common bile duct stones. A
multi-institutional SAGES study. Society of American Gastrointestinal Endoscopic Surgeons.
66. Khoo DE, Walsh CJ, Cox MR, et al. Laparoscopic common bile duct exploration: Evolution of
a new technique. Br J Surg 1996;83:3416.
67. Lezoche E, Paganini AM. Single-stage laparoscopic treatment of gallstones and common bile
duct stones in 120 unselected, consecutive patients. Surg Endosc 1995;9:10705.
68. Catalano MF, Van Dam J, Sivak MV Jr. Endoscopic retrograde cholangiopancreatography in
the orthotopic liver transplant patient. Endoscopy 1995;27:5848.
69. Gholson CF, Zibari G, McDonald JC. Endoscopic diagnosis and management of biliary
complications following orthotopic liver transplantation. Dig Dis Sci 1996;41:104553.
70. Sherman S, Jamidar P, Shaked A, et al. Biliary tract complications after orthotopic liver
transplantation. Endoscopic approach to diagnosis and therapy. Transplantation
1995;60:46770.
71. Theilmann L, Kuppers B, Kadmon M, et al. Biliary tract strictures after orthotopic liver
transplantation: Diagnosis and management. Endoscopy 1994;26:51722.
72. Swobodnik W, Meyer W, Brecht-Kraus D, et al. Ultrasound, computed tomography and
endoscopic retrograde cholangiopancreatography in the morphologic diagnosis of pancreatic
disease. Klin Wochenschr 1983;61:2916.
73.
Grodinsky C, Schumann B, Brush BE. Endoscopic retrograde pancreatic duct cannulation
(ERCP) as an aid to pancreatic surgery. Bull Soc Int Chir 1975;34:60510.
74. Cooperman AM, Sivak MV Jr, Sullivan BH Jr, et al. Endoscopic pancreatography: Its value in
preoperative and postoperative assessment of pancreatic disease. Am J Surg
1975;129:3843.
75. Ammann RW. A critical appraisal of interventional therapy in chronic pancreatitis. Endoscopy
1991;23:1913.
76. Safrany L, Cotton PB. A preliminary report: Urgent duodenoscopic sphincterotomy for acute
gallstone pancreatitis. Surgery 1981;89:4248.
77. Neoptolemos JP, Carr-Locke DL, London NJ, et al. Controlled trial of urgent endoscopic
retrograde cholangiopancreatography and endoscopic sphincterotomy versus conservative

treatment for acute pancreatitis due to gallstones. Lancet 1988;2:97983.
78. Fan ST, Lai EC, Mok FP, et al. Early treatment of acute biliary pancreatitis by endoscopic
papillotomy. N Engl J Med 1993;328:22832.
79. Gregor JC, Ponich TP, Detsky AS. Should ERCP be routine after an episode of "iopathic"
pancreatitis? A cost-utility analysis [see comments]. Gastrointest Endosc 1996;44:11823.
80. Ballinger AB, Barnes E, Alstead EM, Fairclough PD. Is intervention necessary after a first
episode of acute idiopathic pancreatitis?. Gut 1996;38:2935.
81. Barkin JS, Ferstenberg RM, Panullo W, et al. Endoscopic retrograde
cholangiopancreatography in pancreatic trauma. Gastrointest Endosc 1988;34:1025.
82. Linos DA, King RM, Mucha P Jr, Farnell MB. Blunt pancreatic trauma. Minn Med
1983;66:15360.
83. Taylor KJW, Buchin PJ, Viscomi GN, Rosenfield AT. Ultrasonography scanning of the
pancreas. Prospective study of clinical results. Radiology 1981;138:2113.
84.
Pollock D, Taylor KJW. Ultrasound scanning in patients with clinical suspicion of pancreatic
cancer: A retrospective study. Radiology 1981;47:16625.
85. Hessel SJ, Siegelman SS, McNeil BJ, et al. A prospective evaluation of computed
tomography and ultrasound of the pancreas. Radiology 1982;143:12933.
86. Husband JE, Meire HB, Krell L. Comparison of ultrasound and computer-assisted tomography
in pancreatic diagnosis. Br J Radiol 1977;50:85562.
87. Moss AA, Federle M, Shapiro HA, et al. The combined use of computed tomography and
endoscopic retrograde cholangiopancreatography in the assessment of suspected pancreatic
neoplasm: A blind clinical evaluation. Radiology 1980;134:15963.
88. Savarino V, Mansi C, Bistolfi L, et al. Failure of new diagnostic aids in improving detection of
pancreatic cancer at a resectable stage. Dig Dis Sci 1983;28:107882.
89. Frick MP, Feinberg SB, Goodale RL. The value of endoscopic retrograde
cholangiopancreatography in patients with suspected carcinoma of the pancreas and
indeterminate computed tomographic results. Surg Gynecol Obstet 1982;155:17782.
90. Foutch PG, Kerr DM, Harlan JR, et al. Endoscopic retrograde wire-guided brush cytology for
diagnosis of patients with malignant obstruction of the bile duct. Am J Gastroenterol
1990;85:7915.
91.
Bull J, Keeling PW, Thompson RP. Endoscopic retrograde cholangiopancreatography for
unexplained upper abdominal pain. BMJ 1980;280:764.
92. Ruddell WS, Lintott DJ, Axon AT. The diagnostic yield of ERCP in the investigation of
unexplained abdominal pain. Br J Surg 1983;70:745.
93. Chen YK, McCarter TL, Santoro MJ, et al. Utility of endoscopic retrograde
cholangiopancreatography in the evaluation of idiopathic abdominal pain. Am J Gastroenterol
1993;88:13558.
94. Urayama S, Kozarek R, Ball T, et al. Presentation and treatment of annular pancreas in an
adult population. Am J Gastroenterol 1995;90:9959.
95. Benakis P, Nicolakis D, Triantafillidis JK. Successful endoscopic removal of part of a T-tube
from the common bile duct. Endoscopy 1994;26:756.
96. Danilewitz M, Kotfila R, Jensen P. Endoscopic diagnosis and management of Fasciola
hepatica causing biliary obstruction. Am J Gastroenterol 1996;91:26201.
97. Spiliadis C, Georgopoulos S, Dailianas A, et al. The use of ERCP in the study of patients with
hepatic echinococcosis before and after surgical intervention. Gastrointest Endosc
1996;43:5759.
98. Valgaeren G, Duysburgh I, Fierens H, et al. Endoscopic treatment of biliary ascariasis. Report
of a case. Acta Clin Belg 1996;51:97100.
99. Scapa E, Berant A, Leibovitz A, Habot B. Endoscopic retrograde cholangiopancreatography in
octogenarians. J Clin Gastroenterol 1996;22:212.
100. Cappell MS. Endoscopic retrograde cholangiopancreatography with endoscopic
sphincterotomy for symptomatic choledocholithiasis after recent myocardial infarction. Am J
101. Rosenberg J, Overgaard H, Andersen M, et al. Double blind randomised controlled trial of
effect of metoprolol on myocardial ischaemia during endoscopic cholangiopancreatography.
BMJ 1996;313:25861.
102. Gostout CJ, Bender CE. Cholangiopancreatography, sphincter-otomy, and common duct
stone removal via Roux-en-Y limb enteroscopy. Gastroenterology 1988;95:15663.
103.
American Society of Gastrointestinal Endoscopy Committee Report. Radiographic contrast
media used in ERCP. Gastrointest Endosc 1996;43:64751.
104. Carstensen F, Wurbs D. Contrast medium resorption in ERCP. Z Gastroenterol
1993;31:36975.
105. Hopper KD, Wegert SJ, Hallgren SE. Renal excretion of endoscopic retrograde
cholangiopancreatography injected contrast. A common phenomenon. Invest Radiol
1989;24:3946.
106. Mann K, Rendl J, Busley R, et al. Systemic iodine absorption during endoscopic application of
radiographic contrast agents for endoscopic retrograde cholangiopancreaticography. Eur J
Endocrinol 1994;130:498501.
107. Sable RA, Rosenthal WS, Siegel J, et al. Absorption of contrast medium during ERCP. Dig
Dis Sci 1983;28:8016.
108. Lorenz R. Allergic reaction to contrast medium after endoscopic retrograde pancreatography.
109. Moreira VF, Merono E, Larraona JL, et al. ERCP and allergic reactions to iodized contrast
media. Gastrointest Endosc 1985;31:293.
110. Ladas SD, Rokkas T, Kaskarelis J, et al. Absorption of iodized contrast media during ERCP.
111.
Wolf GL, Arenson RL, Cross AP. A prospective trial of ionic vs nonionic contrast agents in
routine clinical practice: Comparison of adverse effects. AJR 1989;152:93944.
112. Palmer FG. The RACR survey of intravenous contrast media reactions: Final report. Australas
Radiol 1988;32:4268.
113.
Wolf GL, Arenson RL, Cross AP. A prospective trial of ionic vs nonionic contrast agents in
routine clinical practice: Comparison of adverse effects. AJR 1989;152:93044.
114. Katayama H, Yamaguchi K, Kozuka T, et al. Adverse reaction to ionic and nonionic contrast
media. Report from the Japanese Committee on the Safety of Contrast Media. Radiology
1990;175:6218.
115. King BF, Hartman GW, Williamson B, et al. Low-osmolality contrast media: A current
perspective. Mayo Clin Proc 1989;64:97685.
116. Lasser EC, Berry CC, Talner LB, et al. Pretreatment with cortico-steroids to alleviate reactions
to intravenous contrast material. N Engl J Med 1987;317:8459.
117. Greenberger PA, Patterson R. The prevention of immediate generalized reactions to
radiocontrast media in high-risk patients. J Allergy Clin Immunol 1991;87:86772.
118. Greenberger PA, Patterson R, Radin RC. Two pretreatment regimens for high-risk patients
receiving radiographic contrast media. J Allergy Clin Immunol 1984;74:5403.
119. Axelrad AM, Fleischer DE, Strack LL, et al. Performance of ERCP for symptomatic
choledocholithiasis during pregnancy: Techniques to increase safety and improve patient
management. Am J Gastroenterol 1994;89:10912.
120. Jamidar PA, Beck GJ, Hoffman BJ, et al. Endoscopic retrograde cholangiopancreatography in
pregnancy. Am J Gastroenterol 1995;90:12637.
121. Baillie J, Cairns SR, Putman WS, Cotton PB. Endoscopic management of choledocholithiasis
during pregnancy. Surg Gynecol Obstet 1990;171:14.
122. Thaggard WG, Johnson PN, Baron TH. Endoscopic management of spontaneous bile duct
perforation and bile peritonitis complicating term pregnancy. Am J Gastroenterol
1995;90:20545.
123. Rahmin MG, Hitscherich R, Jacobson IM. ERCP for symptomatic choledocholithiasis in
pregnancy. Am J Gastroenterol 1994;89:16012.
124. Haines DJ, Bibbey D, Green JR. Does nasal oxygen reduce the cardiorespiratory problems
experienced by elderly patients undergoing endoscopic retrograde
cholangiopancreatography?. Gut 1992;33:9735.
125. Iber FL, Sutberry M, Gupta R, Kruss D. Evaluation of complications during and after
conscious sedation for endoscopy using pulse oximetry. Gastrointest Endosc 1993;39:6205.
126. Rigg JD, Watt TC, Tweedle DE, Martin DF. Oxygen saturation during endoscopic retrograde
cholangiopancreatography: A comparison of two protocols of oxygen administration. Gut
1994;35:40811.
127. al-Hadeedi S, Leaper DJ. Falls in hemoglobin saturation during ERCP and upper
gastrointestinal endoscopy. World J Surg 1991;15:8894.
128. Nebel OT, Silvis SE, Rogers G, et al. Complications associated with endoscopic retrograde
cholangiopancreatography: Results of 1974 A.S.G.E. survey. Gastrointest Endosc
1975;22:346.
129. de Beaux AC, Palmer KR, Carter DC. Factors influencing morbidity and mortality in acute
pancreatitis: An analysis of 279 cases. Gut 1995;37:1216.
130. Sherman S, Lehman GA. ERCP- and endoscopic sphincter-otomy-induced pancreatitis.
Pancreas 1991;6:35067.
131. Nordback I, Airo I. Post-ERCP acute necrotizing pancreatitis. Ann Chir Gynaecol
1988;77:1520.
132. Stanten R, Frey CF. Pancreatitis after endoscopic retrograde cholangiopancreatography. An
underreported disease whose severity is often unappreciated. Arch Surg 1990;125:10324;
discussion 1035.
133. Freeman ML, Nelson DB, Sherman S, et al. Complications of endoscopic biliary
sphincterotomy. N Engl J Med 1996;335:90918.
134. Thoeni RF, Fell SC, Goldberg HI. CT detection of asymptomatic pancreatitis following ERCP.
135. Gottlieb K, Sherman S, Pezzi J, et al. Early recognition of post-ERCP pancreatitis by clinical
assessment and serum pancreatic enzymes. Am J Gastroenterol 1996;91:15537.
136. Banks PA, Carr-Locke DL, Slivka A, et al. Urinary trypsinogen activation peptides (TAP) are
not increased in mild ERCP-induced pancreatitis. Pancreas 1996;12:2947.
137. Skude G, Wehlin L, Maruyama T, Ariyama J. Hyperamylasaemia after duodenoscopy and
retrograde cholangiopancreatography. Gut 1976;17:12732.
138. Pelletier G, Nee N, Brivet M, et al. Upper gastrointestinal endoscopy. An unrecognized cause
of hyperamylasaemia. Dig Dis Sci 1987;32:2546.
139. Hamilton I, Lintott DJ, Rothwell J, Axon AT. Acute pancreatitis following endoscopic
retrograde cholangiopancreatography. Clin Radiol 1983;34:5436.
140. Roszler MH, Campbell WL. Post-ERCP pancreatitis: Association with urographic visualization
during ERCP. Radiology 1985;157:5958.
141. Nairose UB, Wurbs D, Classen M. Santorini's ductAn insignificant variant from normal or an
important overflow valve?. Endoscopy 1978;10:249.
142. Chen YK, Abdulian JD, Escalante-Glorsky S, et al. Clinical outcome of post-ERCP
pancreatitis: Relationship to history of previous pancreatitis. Am J Gastroenterol
1995;90:21203.
143. Steinberg WM. Sphincter of Oddi dysfunction: A clinical controversy. Gastroenterology
1988;95:140915.
144. Sherman S, Troiano FP, Hawes RH, Lehman GA. Sphincter of Oddi manometry: Decreased
risk of clinical pancreatitis with use of a modified aspirating catheter. Gastrointest Endosc
1990;36:4626.
145. Sherman S, Hawes RH, Troiano FP, Lehman GA. Pancreatitis following bile duct sphincter of
Oddi manometry: Utility of the aspirating catheter. Gastrointest Endosc 1992;38:34750.
146. Bub H, Burner W, Riemann JF, Stolte M. Morphology of the pancreatic ductal epithelium after
traumatization of the papilla of Vater or endoscopic retrograde pancreatography with various
contrast media in cats. Scand J Gastroenterol 1983;18:58192.
147. Banerjee AK, Grainger SL, Thompson RP. Trial of low versus high osmolar contrast media in
endoscopic retrograde cholangiopancreatography. Br J Clin Pract 1990;44:4457.
148. Barkin JS, Casal GL, Reiner DK, et al. A comparative study of contrast agents for endoscopic
retrograde pancreatography. Am J Gastroenterol 1991;86:143741.
149. Cunliffe WJ, Cobden I, Lavelle MI, et al. A randomised, prospective study comparing two
contrast media in ERCP. Endoscopy 1987;19:2012.
150. Hannigan BF, Keeling PW, Slavin B, Thompson RP. Hyperamylasemia after ERCP with ionic
and non-ionic contrast media. Gastrointest Endosc 1985;31:10910.
151. O'Connor HJ, Ellis WR, Manning AP, et al. Iopamidol as contrast medium in endoscopic
retrograde pancreatography: A prospective randomised comparison with diatrizoate [see
comments]. Endoscopy 1988;20:2447.
152. Hamilton I, Lintott DJ, Rothwell J, Axon AT. Metrizamide as contrast medium in endoscopic
retrograde cholangiopancreatography. Clin Radiol 1982;33:2935.
153. Osnes M, Skjennald A, Larsen S. A comparison of a new nonionic (metrizamide) and a
dissociable (metrizoate) contrast medium in endoscopic retrograde pancreatography (ERP).
154. Reimer Jensen A, Malchow-Moller A, Matzen P, et al. A randomized trial of iohexol versus
amidotrizoate in endoscopic retrograde pancreatography. Scand J Gastroenterol
1985;20:836.
155. Johnson GK, Geenen JE, Bedford RA, et al. A comparison of nonionic versus ionic contrast
media: Results of a prospective, multicenter study. Midwest Pancreaticobiliary Study Group.
156. Sherman S, Hawes RH, Rathgaber SW, et al. Post-ERCP pancreatitis: Randomized,
prospective study comparing a low- and high-osmolality contrast agent. Gastrointest Endosc
1994;40:4227.
157. Binmoeller KF, Harris AG, Dumas R, et al. Does the somatostatin analogue octreotide protect
against ERCP induced pancreatitis?. Gut 1992;33:112933.
158. Persson B, Slezak P, Efendic S, Haggmark A. Can somatostatin prevent injection pancreatitis
after ERCP?. Hepatogastroenterology 1992;39:25961.
159.
Tyden G, Nyberg B, Sonnenfeld T, Thulin L. Effect of somatostatin on hyperamylasemia
following endoscopic pancreatography. Acta Chir Scand 1986;530(suppl):435.
160. Bordas JM, Toledo V, Mondelo F, Rodes J. Prevention of pancreatic reactions by bolus
somatostatin administration in patients undergoing endoscopic retrograde
cholangiopancreatography and endoscopic sphincterotomy. Horm Res 1988;29:1068.
161. Sternlieb JM, Aronchick CA, Retig JN, et al. A multicenter, randomized, controlled trial to
evaluate the effect of prophylactic octreotide on ERCP-induced pancreatitis [see comments].
162. Arcidiacono R, Gambitta P, Rossi A, et al. The use of a long-acting somatostatin analogue
(octreotide) for prophylaxis of acute pancreatitis after endoscopic sphincterotomy. Endoscopy
1994;26:7158.
163. Sand J, Nordback I. Prospective randomized trial of the effect of nifedipine on pancreatic
irritation after endoscopic retrograde cholangiopancreatography. Digestion 1993;54:10511.
164. Weiner GR, Geenen JE, Hogan WJ, Catalano MF. Use of cortico-steroids in the prevention of
post-ERCP pancreatitis. Gastrointest Endosc 1995;42:57983.
165. Cavallini G, Tittobello A, Frulloni L, et al. Gabexate for the prevention of pancreatic damage
related to endoscopic retrograde cholangiopancreatography. Gabexate in Digestive
EndoscopyItalian Group. N Engl J Med 1996;335:91923.
166. Allen JI, Allen MO, Olson MM, et al. Pseudomonas infection of the biliary system resulting
from use of a contaminated endoscope. Gastroenterology 1987;92:75963.
167. Cotton PB. Outcomes of endoscopy procedures: Struggling towards definitions. Gastrointest
Endosc 1994;40:5148.
168. Tseng A, Sales DJ, Simonowitz DA, Enker WE. Pancreas abscess: A fatal complication of
endoscopic cholangiopancreatography (ERCP). Endoscopy 1977;9:2503.
169. Jalan R, Jones HL, Walker RJ. Multiple pancreatic abscesses due to Candida albicans
following ERCP. Scott Med J 1994;39:178.
170. O'Connor M, Kolars J, Ansel H, et al. Preoperative endoscopic retrograde
cholangiopancreatography in the surgical management of pancreatic pseudocysts. Am J Surg
1986;151:1824.
171. Kolars JC, Allen MO, Ansel H, et al. Pancreatic pseudocysts: Clinical and endoscopic
experience. Am J Gastroenterol 1989;84:25964.
173. Motte S, Deviere J, Dumonceau JM, et al. Risk factors for septicemia following endoscopic
174. Sand J, Airo I, Hiltunen KM, et al. Changes in biliary bacteria after endoscopic
cholangiography and sphincterotomy. Am Surg 1992;58:3248.
175. Classen DC, Jacobson JA, Burke JP, et al. Serious Pseudomonas infections associated with
endoscopic retrograde cholangiopancreatography. Am J Med 1988;84:5906.
176. Bass DH, Oliver S, Bornman PC. Pseudomonas septicaemia after endoscopic retrograde
cholangiopancreatographyAn unresolved problem. S Afr Med J 1990;77:50911.
177. Godiwala T, Andry M, Agrawal N, Ertan A. Consecutive Serratia marcescens infections
following endoscopic retrograde cholangiopancreatography. Gastrointest Endosc
1988;34:3457.
178. Sauter G, Grabein B, Huber G, et al. Antibiotic prophylaxis of infectious complications with
endoscopic retrograde cholangiopancreatography. A randomized controlled study. Endoscopy
1990;22:1647.
179. Brandes JW, Scheffer B, Lorenz-Meyer H, et al. ERCP: Complications and prophylaxisA
controlled study. Endoscopy 1981;13:2730.
180. Collen MJ, Hanan MR, Maher JA, Stubrin SE. Modification of endoscopic retrograde
cholangiopancreatography (ERCP) septic complications by the addition of an antibiotic to the
contrast media. Randomized controlled investigation. Am J Gastroenterol 1980;74:4936.
181. Byl B, Deviere J, Struelens MJ, et al. Antibiotic prophylaxis for infectious complications after
therapeutic endoscopic retrograde cholangiopancreatography: A randomized, double-blind,
placebo-controlled study. Clin Infect Dis 1995;20:123640.
182. van den Hazel SJ, Speelman P, Dankert J, et al. Piperacillin to prevent cholangitis after
endoscopic retrograde cholangiopancreatography. A randomized, controlled trial. Ann Intern
Med 1996;125:4427.
183. Lewis FW, Moloo N, Stiegmann GV, Goff JS. Splenic injury complicating therapeutic upper
gastrointestinal endoscopy and ERCP. Gastrointest Endosc 1991;37:6323.
184. Lo AY, Washington M, Fischer MG. Splenic trauma following endoscopic retrograde
cholangiopancreatography (ERCP). Surg Endosc 1994;8:6923.
185. Ong E, Bohmler U, Wurbs D. Splenic injury as a complication of endoscopy: Two case reports
and a literature review. Endoscopy 1991;23:3024.
186. Furman G, Morgenstern L. Splenic injury and abscess complicating endoscopic retrograde
cholangiopancreatography. Surg Endosc 1993;7:3434.
187. Alvarez C, Hunt K, Ashley SW, Reber HA. Emphysematous cholecystitis after ERCP. Dig Dis
Sci 1994;39:171923.
188. Dupas JL, Mancheron H, Sevenet F, et al. Hepatic subcapsular biloma. An unusual
complication of endoscopic retrograde cholangiopancreatography. Gastroenterology
1988;94:12257.
189. Ashby BS, Goodman DA. Effects of x-radiation on fibreoptic endoscopes. Gut
1980;21:71724.
190. Cohen G, Brodmerkel GJ Jr, Lynn S. Absorbed doses to patients and personnel from
endoscopic retrograde cholangiopancreatographic (ERCP) examinations. Radiology
1979;130:7735.
191. van Husen N, Hogemann B, Egen V, Mehnert C. Radiation exposure in endoscopic
retrograde cholangiopancreatography. Endoscopy 1984;16:1124.

192. Chen MY, Van Swearingen FL, Mitchell R, Ott DJ. Radiation exposure during ERCP: Effect of
a protective shield. Gastrointest Endosc 1996;43:15.
193. Heyd RL, Kopecky KK, Sherman S, et al. Radiation exposure to patients and personnel during
interventional ERCP at a teaching institution. Gastrointest Endosc 1996;44:28792.
194. Sander D, Brunner G. Studies on radiation exposure of personnel in endoscopy. Z
195. Germanaud J, Legoux JL, Sabattier R, et al. Radiation protection of operators during
endoscopic retrograde cholangiopancreatography. Gastroenterol Clin Biol 1993;17:25963.
196. Selmaier M, Stillkrieg W, Muller RG, et al. Radiation burden in diagnostic and therapeutic
endoscopic retrograde cholangiopancreatography (ERCP). Z Gastroenterol 1994;32:6714.
197.
American Society of Gastrointestinal Endoscopy Committee Report. Radiation safety during
198. Buscail L. Endoscopic ultrasonography in pancreatobiliary disease using radial instruments.
199.
Yasuda K, Nakajima M, Kawai K. Diseases of the biliary tract and the papilla of Vater. In
Kawai K (ed). Endoscopic Ultrasonography in Gastroenterology. Tokyo, New York: Igaku
Shoin, 1988;96105.
200. Amouyal P, Amouyal G, Levy P, et al. Diagnosis of choledocholithiasis by endoscopic
201. Canto M. Endoscopic ultrasonography and gallstone disease. Gastrointest Endosc
1996;43:S3743.
202. Edmundowicz SA, Aliperti G, Middleton WD. Preliminary experience using endoscopic
ultrasonography in the diagnosis of choledocholithiasis. Endoscopy 1992;24:7748.
203. Palazzo L, Girollet PP, Salmeron M, et al. Value of endoscopic ultrasonography in the
diagnosis of common bile duct stones: Comparison with surgical exploration and ERCP.
204. Edmundowicz SA. Common bile duct stones. Gastrointest Endosc Clin N Am 1995;5:81724.
205. Prat F, Amouyal G, Amouyal P, et al. Prospective controlled study of endoscopic
ultrasonography and endoscopic retrograde cholangiography in patients with suspected
common-bile duct lithiasis. Lancet 1996;347:759.
206. Amouyal P, Amouyal G, Mompoint D, et al. Endosonography: Promising method for diagnosis
of extrahepatic cholestasis. Lancet 1989;2:11958.
207. Yasuda K, Mukai H, Nakajima M. Endoscopic ultrasonography diagnosis of pancreatic
cancer. Gastrointest Endosc Clin N Am 1995;5:699712.
208. Brugge WR. Pancreatic cancer staging. Endoscopic ultrasonography criteria for vascular
invasion. Gastrointest Endosc Clin N Am 1995;5:74153.
209. Dancygier H. Endoscopic ultrasonography in chronic pancreatitis. Gastrointest Endosc Clin N
Am 1995;5:795804.
210. Meakem TJ III, Schnall MD. Magnetic resonance cholangiography. Gastroenterol Clin North
Am 1995;24:22138.
211. Guibaud L, Bret PM, Reinhold C, et al. Bile duct obstruction and choledocholithiasis:
Diagnosis with MR cholangiography. Radiology 1995;197:10915.
212. Pavone P, Laghi A, Catalano C, et al. Magnetic resonance cholangiopancreatography. A new
method of noninvasive biliopancreatic diagnosis. Radiol Med (Torino) 1995;90:43843.
213. Laubenberger J, Buchert M, Schneider B, et al. Breath-hold projection magnetic
resonance-cholangiopancreaticography (MRCP): A new method for the examination of the
bile and pancreatic ducts. Magn Reson Med 1995;33:1823.
214.
Regan F, Fradin J, Khazan R, et al. Choledocholithiasis: Evaluation with MR cholangiography.
AJR 1996;167:14415.
215. Chan YL, Chan AC, Lam WW, et al. Choledocholithiasis: Comparison of MR cholangiography
and endoscopic retrograde cholangiography. Radiology 1996;200:859.
216. Lee MG, Lee HJ, Kim MH, et al. Extrahepatic biliary diseases: 3D MR
cholangiopancreatography compared with endoscopic retrograde cholangiopancreatography.

Radiology 1997;202:6639.
217. Soto JA, Barish MA, Yucel EK, et al. Pancreatic duct: MR cholangiopancreatography with a
three-dimensional fast spin-echo technique. Radiology 1995;196:45964.
218. Pavone P, Laghi A, Catalano C, et al. Non-invasive evaluation of the biliary tree with magnetic
resonance cholangiopancreatography: Initial clinical experience. Ital J Gastroenterol
1996;28:639.
219. Takehara Y, Ichijo K, Tooyama N, et al. Breath-hold MR cholangiopancreatography with a
long-echo-train fast spin-echo sequence and a surface coil in chronic pancreatitis. Radiology
1994;192:738.
220. Soto JA, Barish MA, Yucel EK, et al. Magnetic resonance cholangiography: Comparison with
endoscopic retrograde cholangiopancreatography. Gastroenterology 1996;110:589597.
221. Taourel P, Bret PM, Reinhold C, et al. Anatomic variants of the biliary tree: Diagnosis with MR
cholangiopancreatography. Radiology 1996;199:5217.
222. Bret PM, Reinhold C, Taourel P, et al. Pancreas divisum: Evaluation with MR
cholangiopancreatography. Radiology 1996;199:99103.
223.
Reinhold C, Bret PM. Current status of MR cholangiopancreatography. AJR
1996;166:128595.
224. Soto JA, Yucel EK, Barish MA, et al. MR cholangiopancreatography after unsuccessful or
incomplete ERCP. Radiology 1996;199:918.
225. Lawson JM. Magnetic resonance cholangiography: Requiem for routine diagnostic ERCP?.
Chapter 57 Technique of Endoscopic Retrograde

Cholangiopancreatography
(5783)
PAUL N. YAKSHE, M.D.
JACK A. VENNES, M.D.
Endoscopic retrograde cholangiopancreatography (ERCP) is a sophisticated technique that can be
mastered by a skilled endoscopist with a reasonable commitment of training time. When diagnostic
ERCP is skillfully performed, the results are both informative and accurate, and the therapeutic options
of endoscopic sphincterotomy, stone extraction, and stent placement may obviate the need for surgery.
However, when performed infrequently, or with modest success, ERCP is yet another expensive and
often frustrating road to diagnostic equivocation. It is therefore advisable to determine early whether
the number of patients for whom ERCP is indicated will provide sufficient experience for the
development and maintenance of proficiency. Because ERCP is more complex than standard
endoscopic procedures, the technique of ERCP requires advanced hand-eye coordination and a large
prerequisite endoscopic experience. When the elements of innate skill, adequate training, patience,
and practice are combined, the technique of ERCP is certain to provide the desired clinical information
and the opportunity for therapeutic intervention in a rewarding and expedient fashion.
The time required to learn the procedure varies greatly from person to person, and it is clearly best that
training be under the guidance of an experienced endoscopist who has the ability and patience to
teach others to perform the procedure. Prospective data from one study suggest that 180 supervised
procedures are required to achieve a basic level of proficiency.1(5784) Although expert guidance
greatly facilitates learning, specific written instructions on the technique of ERCP can be useful,
particularly if they can be referred to as training progresses. We recognize that there are often several
methods for doing things properly; the method described here is only one of these.
Equipment
Certain basic pieces of equipment are necessary to the proper performance of ERCP.
Fluoroscopic Equipment
Although the endoscope enables access to the duodenal papilla, ERCP is substantially a radiographic
study of the ductal systems of the pancreas and biliary tree. This requires an x-ray examining room
equipped with a high-resolution fluoroscopic monitor and spot film capability. A radiographic tube with a
small focal spot (0.3 to 0.6 mm) and a large milliamperage generator is best for producing sharply
detailed x-ray images with short exposure times. Because the orientation of the patient on the x-ray
table is the reverse of that with routine barium radiographic procedures, the ability to reverse the
fluoroscope image in both the horizontal and the vertical axes is preferable. The x-ray table should be
tiltable to at least 30 degrees in both head-up and head-down positions. This option helps differentiate
filling defects as air bubbles or stones; by moving the table into a head-up position, gravity will pull
stones down toward the papilla whereas air bubbles will rise upward toward the intrahepatic ducts.
Because additional endoscopic equipment and personnel are required during ERCP, ample space in
the fluoroscopy room is necessary.
ERCP is being performed in some endoscopy units with mobile image intensifiers that employ a C-arm
design. In addition to being portable, these machines are relatively inexpensive and offer the added
advantage that the imaging tube can be rotated in the transverse plane. The latter function (also
available in large, fixed x-ray machines) can be very useful in that it provides oblique x-ray views
without changing the position of the patient. Furthermore, the x-ray images can be saved in digital
format, although the capacity for storage with most systems is relatively limited. Unfortunately, there
are very little comparative data with regard to the quality of the fluoroscopic images compared with that
of larger machines. It appears that the images are somewhat inferior, but available information
suggests that they are satisfactory for most ERCP procedures.2(5785) Depending on the process by
which they are made, the "hard-copy" images are usually of lesser quality than standard x-ray films.
With less expensive systems, the x-ray pictures are usually generated by a photographic process that
tends to produce images of lesser resolution that are excessively grainy compared with the digital
images from which they are derived.
Endoscopic Equipment
Technologic advances in endoscopic equipment are responsible for the evolution of ERCP. The
procedure became possible with the development of side-viewing fiberoptic instruments. These
provided an en face view of the papilla, greater maneuverability, and a mechanism (elevator) to control
the catheter during cannulation. Although video endoscopes are slightly less maneuverable, they
greatly reduce endoscopist fatigue and facilitate involvement and coordination of the ERCP team.
Regardless of whether fiberoptic or video technology is used, currently available duodenoscopes are
120 cm in working length, and the techniques used to intubate the upper digestive tract are similar.
Side-viewing duodenoscopes are available with a variety of accessory channel sizes (2.8 to 4.0 mm);
the larger-diameter channels are useful and often necessary to accommodate the specialized
accessories associated with therapeutic biliary endoscopy.
Cannulas
A spectrum of cannulas and specialized catheters is now commercially available for both diagnostic
and therapeutic ERCP. In general, cannulas should have an indwelling stylus wire to stiffen the cannula
outside of the duodenoscope and that portion residing within the proximal portion of the accessory
channel of the instrument; this permits a firm finger grip and authoritative advancement of the cannula
without crimping. We prefer a short tapered-tip cannula with a radiopaque tip marker that
accommodates a 0.035-inch-diameter guidewire. Its advantages include ease of cannulation and the
ability to see the cannula tip fluoroscopically. The latter feature is desirable because it makes it
possible to identify the desired duct without repetitive contrast material injections. In addition, it
facilitates "free-duct" cannulation and assists in over-the-wire exchanges of the cannula for specialized
catheters.
The ERCP Team

As part of the effort to establish ERCP as a viable procedure, it is important from the outset to
gradually combine endoscopic and radiographic techniques into a smooth, disciplined routine. A team
approach with a radiologist is essential to initial success because ERCP requires expertise that
crosses two training disciplines. With skill, experience, and teamwork, detailed information will be
available from a few well-positioned, coned x-rays; thus there will be less radiation exposure to the
patient, the team, and the fiber bundles of the endoscope. After the procedure is well established and
as the endoscopist gains experience, an x-ray technician will suffice and the presence of a radiologist
is required only in unusual or complicated cases and for later interpretation of x-ray films.
Specially trained gastrointestinal nurses or technicians are vitally important members of the ERCP
team. Because the procedure demands the complete attention of the endoscopist, often requires
deeper patient sedation, and frequently takes longer than routine upper and lower endoscopic
procedures, it is preferable, indeed advisable, to have two assistants available during ERCP
procedures. In addition to monitoring the patient's vital signs and level of comfort, they assist with the
preparation of equipment, administration of medications, and selection and operation of a variety of
catheters, guidewires, and devices throughout the procedure. For successful performance of ERCP, a
knowledgeable, interested, and dedicated assistant is a required member of the team.
Preprocedure Preparation
The likelihood that ERCP will be well tolerated, expeditious, and successful is related to the
thoroughness of the preparation by the team beforehand. Preparation begins with a conversation
between physician and patient in which ample time is allowed to discuss the indications for, and
potential risks of, ERCP. In general, the rapport established during the informed consent process
inspires confidence and improves the emotional state of the patient. When coupled with an
atmosphere of quiet, unhurried competence in the x-ray room and appropriate conscious sedation,
ERCP can be performed safely as an outpatient procedure in the majority of patients.
Preprocedure orders are fairly standard. The patient fasts after midnight except for her or his usual
cardiac or blood pressure medications. If the procedure is to be performed in the afternoon, a liquid
breakfast may be permitted. It is usually preferable to establish intravenous access in the right arm
because it will be more accessible during the procedure than the left arm. We routinely infuse 5%
dextrose in one half normal saline solution at a rate of 50 ml/hr. It is preferable that diabetic patients
undergo ERCP during the morning hours. Type I and insulin-requiring type II diabetic patients are
instructed to take half of their neutral protamine Hagedorn (NPH) or lente insulin and none of their
regular-acting insulin. Usually, the patient will recover from the effects of the sedation in time to take
the normal midday meal.
Patients reporting an allergic reaction to iodine or iodinated radiographic contrast agents fall into two
groups (see also Chapter 56: Indications, Contraindications, and Complications of Diagnostic
Endoscopic Retrograde Cholangiopancreatography). Most frequently, they have experienced urticaria,
flushing, or nausea. The risk of serious reaction to contrast material instilled into the bile or pancreatic
ducts is very small, and blood iodide levels are low. After further discussion with the patient, it is
prudent to commence the procedure with appropriate medications nearby. During the procedure itself,
contrast material should be injected in only the clinically relevant ductal system and with minimum
required filling. A second small group of patients report a serious prior reaction to intravenous contrast
material, such as laryngeal edema, anaphylaxis, or vascular collapse; we treat these patients with
prednisone 20 mg by mouth every 12 hours for three doses before the procedure as well as cimetidine
300 mg together with benadryl 25 mg, both being administered intravenously immediately before the
procedure.
It is important to recognize the potential for infection associated with ERCP and to take precautions to
prevent its occurrence. Routine cleansing and disinfection of endoscopes are important immediately
after each use but are not sufficient to prevent procedure-related infection. Water-loving organisms like
Pseudomonas may survive disinfection and thrive during storage of the endoscope within moist
confines and channels. To prevent the introduction of these exogenous bacteria into the
pancreaticobiliary system and subsequent infection, the following safeguards are necessary. First, the
air/water channel should be cleaned and disinfected as part of total care. Second, the growth of
water-loving organisms should be prevented by air-drying all channels before storage. Third, water
bottles and tubing should be disinfected or gas-sterilized after each day's use; a newly disinfected or
sterilized bottle with connecting tube should be used each day. Fourth, sterile water should be used in
the water bottle to avoid the possibility of introducing Pseudomonas from the public water supply. And
finally, the instrument should be soaked in glutaraldehyde (or another suitable disinfecting solution) and
rinsed again just before the procedure (see also Chapter 8: Disinfection of Endoscopes and
Accessories).
Additional safeguards are necessary to prevent ERCP-related infection in certain clinical situations.
Because the endoscope is passed transorally, translocation of endogenous bacteria from the
oropharynx into the pancreaticobiliary system may occur. Usually, this will not cause a problem.
However, if an obstructing lesion of either ductal system is present, the introduction of organisms into
this "semiclosed space" may result in serious cholangitis or pancreatitis. In selected patients with
dilated ducts, pseudocysts, suspected obstruction, sclerosing cholangitis, or other conditions that may
place the patient at increased risk for infection, we administer ampicillin 2.0 g intravenously plus
gentamicin 1.5 mg/kg intravenously preprocedure. In patients with concomitant neutropenia, an
attempt can be made to reduce the oropharyngeal flora by having the patient gargle and rinse with an
antiseptic mouthwash. When ductal obstruction and poor emptying of one or both ducts are confirmed
during ERCP, immediate drainage is indicated via stent placement or nasobiliary catheter drainage
with continuing antibiotic coverage. If all of these precautions are taken, ERCP-related infections can
be virtually eliminated.
The mechanical functions, illumination, and image systems of the endoscopic equipment should be
routinely checked before the procedure. Proper function of the air/water and suction channels is
necessary for adequate visualization and identification of endoscopic landmarks. Over time, wear and
tear may stretch the cables controlling the endoscope's mechanical functions and compromise the
success of the procedure. Under these conditions, the manner with which the control knobs flex the
distal tip may be imprecise or the elevator may raise the cannula tip to only the middle or lower half of
the visual field. Because optimum vertical and lateral flexion and maximum cannula elevation are
sometimes required for successful cannulation, these functions should be checked periodically and
kept in working order according to the specifications of the instrument manufacturer.
It is advisable to check the operation of the fluoroscopic equipment and to obtain a scout film of the
abdomen before the procedure. In addition to ensuring that the equipment is operating properly, this
practice allows for the evaluation of soft tissue densities, documentation of the location of surgical
clips, and screening for residual contrast within the abdomen from previous radiographic studies that
may interfere with either obtaining or interpreting ERCP films.
Although electronic monitoring devices are no substitute for close patient observation, we recommend
that pulse oximetry be used routinely. ERCP often requires the complete attention of the endoscopist
and nurse assistant, and the procedures are usually performed in a darkened room, which makes
clinical evaluation of the patient more difficult. The procedure is generally longer and requires more
sedation than either upper or lower endoscopic procedures. Finally, several studies have shown that
clinical observation is less sensitive than electronic monitoring devices in detecting certain variables
that may predict a bad outcome.3,4(5786) Consequently, we believe pulse oximetry facilitates patient
observation and utilize it for all patients undergoing ERCP.
Although ERCP is generally regarded as safe, up to 44% of patients undergoing the procedure have
been shown to encounter some degree of oxygen desaturation.5(5787) Administration of continuous
oxygen may prevent hypoxia during ERCP. Crantock et al.6(5788) recorded oxygen saturation using
pulse oximetry in 50 consecutive patients undergoing ERCP sedated using a combination of
midazolam and fentanyl. Oxygen saturation fell below 90% in 47% of patients not receiving
supplemental oxygen compared with 0% in those randomly assigned to receive low-flow oxygen (2
L/min) via nasal cannula.6(5789)
Frothy gastric and intestinal secretions may obscure visibility during endoscopy. This nuisance can be
greatly reduced by having the patient swallow 30 ml of water mixed with a few drops of simethicone
before the procedure or by the addition of simethicone to the water in the water bottle.
We place the patient in a modified left lateral decubitus position on the x-ray table. This is achieved by
first placing the patient in the prone position with the head facing right and both arms at the side. The
torso is then rotated by bringing the right leg and right arm into a flexed position and supporting the
patient by placement of a triangular wedge or cylindrical pillow underneath the right side. The left arm
remains parallel to the torso behind the patient's back. This position approximates the left lateral
decubitus position, which is optimum for intubation of the esophagus and advancement of the
endoscope into the duodenum. When the duodenum is entered, the supporting wedge or pillow is
removed and the right leg is straightened. This brings the patient back into the prone position, which is
optimum for cannulating the papilla and obtaining quality radiographic images of the pancreas and
biliary tree.
Final preparations commence with the administration of pharmacologic agents. The oropharynx is
anesthetized with a topical anesthetic (we prefer lidocaine [Xylocaine] spray) to diminish the gag reflex
and increase patient comfort. Atropine (0.6 mg intravenously) or its equivalent will usually control
duodenal motility, decrease troubling oral secretions, and provide some relaxation of the smooth
muscle of the sphincter of Oddi. Glucagon 0.2 to 0.4 mg IV is also used to obtain duodenal atony.
When titrated in small increments, we find a combination of midazolam (1 to 6 mg) and meperidine (25
to 100 mg) or fentanyl (0.075 to 0.3 mg) provides suitable conscious sedation (see also Chapter 38:
Technique of Upper Gastrointestinal Endoscopy). In a randomized trial comparing patient-controlled
versus physician-directed conscious sedation for Ercp, Jowell et al.7(5790) found no significant
difference in patient satisfaction. Interestingly, the physicians participating in this study believed patient
satisfaction was greater when conscious sedation was directed by the physician.
Premedication should be unhurried, precise, and tailored to the patient's age and any comorbid
conditions, that is, smaller aliquots are used initially (particularly in the elderly or very ill patient) to
observe the clinical effect before additional medication is given. Within wide limits, the correct amount
of drug administered is the least amount the patient requires for comfort. We talk to the patient before
and during the administration of medication; when there is evidence of ptosis and dysarthria or slowed
speech, the patient is usually sedated enough to begin the procedure. The total amount of required
sedative or narcotic medication is less if the patient is relaxed and the endoscopist is gentle. As with
any procedure that requires these medications, equipment and drugs for cardiopulmonary resuscitation
must be readily available and each member of the ERCP team must be familiar with basic cardiac life
support; the physician should be familiar with advanced cardiac life support.
Technical Aspects of ERCP

Operation of the Duodenoscope
During the transition from a forward-viewing to a side-viewing instrument, a mental image of the
mechanical design and function of the duodenoscope helps to reduce confusion. By convention, the
field of view of a duodenoscope has the same radial orientation as the projection of the control valves
for air/water and suction/ accessory channels (Figure 571A). Therefore, when the control handle is
held in the usual manner and a clock face is mentally superimposed on its axis, both the projection of
the control valves and the endoscopic view field are vertical or toward 12 o'clock; hence, this is referred
to as the 12 o'clock position (Figure 571B). With the patient in the left lateral decubitus position and
the endoscope tip in a neutral (unflexed) position, the twelve o'clock position will view vertically or
toward the right side of the digestive tract. For example, with the tip of the side-viewing endoscope in
the stomach (patient in left decubitus position), the 12 o'clock position will view the lesser curvature.
(5791)Figure 571. A, By convention, duodenoscopes view in the same radial orientation as

the projection of the valves on the control handle. Therefore, with the patient in the left lateral
decubitus position and the tip of the endoscope in the stomach, the endoscopic image can be
predicted based on the orientation of the control handle. B, The conventional 12 o'clock
position; the endoscope views the lesser curvature. C, The endoscope is rotated into the 3
o'clock position; the posterior wall is viewed. D, The 9 o'clock position; the view is of the
anterior wall. E, The 6 o'clock position; the greater curvature is viewed.
If one supports the insertion tube with the right hand and stands a proper distance away from the
patient, rotational movements of the control handle with the left arm and wrist will be transmitted along
the insertion tube to the optical components at the distal tip in a nearly one-to-one axial rotation. Thus,
rotating the control handle clockwise 90 degrees so the valve projections are pointing to 3 o'clock (as if
"winning a left-handed arm wrestle") will result, when the tip is in the stomach, in a view of the posterior
wall (see Figure 571C); rotation counterclockwise so the valve projections point to 9 o'clock ("losing a
left-handed arm wrestle") presents a view of the anterior wall (see Figure 571D). Further rotation to
the 6 o'clock position brings the greater curvature into view (see Figure 571E). In contrast to
colonoscopy, grasping and torquing the insertion tube with the right hand is unnecessary with ERCP.
As with upper endoscopy, if one stands too close to the patient, the insertion tube remaining outside of
the patient may become slack and allow rotational movements to be absorbed by a loop rather than
being transmitted down the axis of the endoscope. Reviewing this mental imagery of the side-viewing
endoscope will facilitate learning; in time, the appropriate motor response to a given visual stimulus will
become automatic.
The use of a clock face as a system of reference in the previous discussion refers to the direction in
which the optical view field of the side-viewing instrument is oriented. The clock face is also used as a
reference system to specify position within the view field. Thus, a lesion located at 3 o'clock will be
found in the right side of the image field.
The dictum "keep the lumen in view" is as fundamental to ERCP as it is to all endoscopic procedures.
However, because the duodenoscope is a side-viewing instrument, a slight adjustment in technique is
required. A lumenal (forward) view is usually afforded by downward deflection of the tip of the
endoscope (Figure 572A). Conversely, the left thumb is primarily responsible for making this
adjustment. Because the tip of the endoscope is tethered to the vertical control knob by deflection
cables, pushing the knob up with the thumb will result in downward deflection (and vice versa). Once
one adjusts to this slight variation, adjustments of the visual field are made using the vertical (up,
down) and horizontal (right, left) control knobs analogous to routine endoscopy.
(5792)Figure 572. A, When attempting to "pursue the lumen," downward flexion of the
endoscope tip permits a forward view; this is accomplished by pushing upward on the inner
(larger) control knob with the left thumb. B, Conversely, when attempting to pass through the
pylorus or move closer to a point in the endoscopic image, the required upward flexion of the
endoscope tip is accomplished by pushing down on the inner control knob with the left thumb.
A detailed discussion of the terminology used to describe the operation of the control knobs may be
found in Chapter 38: Technique of Upper Gastrointestinal Endoscopy. Terms used to describe the
directions in which the control knobs must be turned to deflect the tip are relative. For the purpose of
description, the control section can be viewed on the side that has the deflection knobs. Each knob
may then be considered as moving in either a clockwise or a counterclockwise direction. Thus,
counterclockwise rotation of the innermost control knob deflects the tip upward; a counterclockwise
turn of the outer knob deflects the tip to the left. Clockwise rotation of the inner and outer knobs results
in downward and right tip deflection, respectively.
Passing the Endoscope

Intubation of the Esophagus
For the patient, perhaps the most anxiety-provoking aspect of endoscopy occurs as the procedure is
begun. Intubating the esophagus in a gentle, expedient fashion is often as important as topical
anesthesia and conscious sedation in establishing the tone for the entire procedure. The endoscopist
should have an appreciation of the distance from the incisors to the hypopharynx and the usual
resistance offered by the cricopharyngeus muscle based on prior experience with forward-viewing
endoscopes.
With the flexion knobs unbraked, the endoscope is introduced into the oropharynx and advanced
gently into the hypopharynx (approximately 15 cm from the incisors). The patient is then asked to
swallow, which relaxes the cricopharyngeus muscle and permits easy intubation of the esophagus.
Occasionally, the endoscope tip will remain in the pyriform sinus (usually the left) despite repeated
attempts to intubate the esophagus. A small degree of left lateral flexion toward the midline is
sometimes helpful in aligning the tip of the endoscope with the lumen of the upper esophageal
sphincter and facilitating intubation. One should be aware of the potential for perforation of a pyriform
sinus or a Zenker diverticulum during attempts to intubate the esophagus. If resistance encountered
with a side-viewing instrument is unusual, the instrument should be withdrawn and a forward-viewing
instrument substituted to investigate the problem. Obviously, the watchword is gentleness.
The endoscope is advanced through the esophagus blindly. If there is resistance to advancement of
the instrument, or if some information about the esophagus is desired, gentle downward flexion of the
tip usually provides a luminal view of the esophagus.
Traversing the Stomach
At approximately 40 cm from the incisors, the tip of the endoscope enters the stomach and the
procedure changes from the realm of "feel" to that of eye-hand coordination. Successful advancement
of the endoscope through the stomach and into proper position in the duodenum depends on adequate
visualization of the gastrointestinal lumen, recognition of anatomic landmarks, and appropriate motor
responses. With the patient in the left lateral decubitus position and the control section in the 12 o'clock
position, the usual endoscopic view is that of the proximal lesser curvature of the stomach. The initial
image may be "redded out," in which case the tip of the endoscope is on, or very close to, the mucosa.
Insufflation of air into the stomach, extension of the left wrist (counterclockwise rotation of the control
section), and downward tip deflection will usually result in an image of the gastric lumen. Often,
downward deflection reveals a pool of gastric secretions (which also identifies the greater curvature of
the body). As with routine upper endoscopy, it is advisable to aspirate a large gastric pool to reduce the
risk of serious aspiration should vomiting occur during the procedure. Repetitive aspiration and
insufflation, however, are time-consuming and should be avoided. The luminal view in the proximal
stomach is usually that of fundic rugal folds converging toward the stomach body (Figure 573A and
B). At this point, upward deflection to a neutral position, in conjunction with gentle advancement of the
insertion tube and clockwise axial rotation, will produce a smooth spiraling motion downward into the
antrum. The endoscope will hug the greater curvature of the stomach and provide visual images of the
lesser curvature and lumen. Notable landmarks along the way include the disappearance of rugal
folds, a view of the angulus, and the appearance of the pyloric ring (Figure 573C and D).
(5793)Figure 573. Passage of the duodenoscope through the stomach is facilitated by

recognition of visual landmarks. On passage of the instrument into the stomach, the initial
image is usually that of the lesser curvature or anterior wall. Diagram (A) and endoscopic view
(B) of the route from the fundus to the antrum, marked by converging rugal folds. Diagram (C)
and endoscopic view (D) as the instrument enters the antrum. The rugal folds disappear, the
angulus can be identified superiorly, and the pylorus is seen in the distance.
It may be desirable in certain clinical situations to study the stomach and duodenum on the way to the
papilla. As familiarity with side-viewing duodenoscopes is achieved, this can be accomplished
expeditiously and may be included in the examination. Downward flexion of the endoscope tip will
usually afford a tubular view of the gastric lumen (see Figure 572A); the top of the visual field is the
lesser curvature, the bottom the greater curvature, the posterior wall is to the right, and the anterior
wall to the left. If the tip of the endoscope is brought back into the neutral position, one is frequently
looking at a wall of the stomach. Using the clock-face reference system for axial rotation, the 12 o'clock
position views the lesser curvature, the 3 o'clock position the posterior wall, and the 9 o'clock position
the anterior wall. Full upward flexion (pushing the large wheel down with the thumb) will provide a
retroflexed view of the proximal stomach. As with other gastrointestinal endoscopic procedures, the
image may become redded out or obscure; when lost, withdraw the insertion tube a bit, insufflate,
downflex, and search the image for visual landmarks.
Anatomic variants of the normal stomach can be confusing and can give the impression that the
stomach has no outlet. Early recognition of the problem is important; otherwise the procedure can
become prolonged or uncomfortable for the patient. In a "cascade" stomach, a significant portion of the
fundus is lying posterior and to the left of the spine; the upper body of the stomach is virtually folded on
itself. Insufflation of air preferentially dilates the fundus, and consequently, more time is spent looking
for an outlet in the wrong area. A "horizontal" stomach presents a similar dilemma. In either situation, it
is effective to bring the tip of the endoscope back to the esophagogastric junction and rotate into the 9
o'clock position. With full downward deflection and gradual advancement of the insertion tube, the
radial convergence of rugal folds is then recognized and indicates the lumen to be entered.
Entering the Duodenum
Advancing the endoscope through the pylorus and into the duodenum is usually easily accomplished if
a few rules are observed. Slight downward flexion should provide a luminal view of the antrum. Next,
the visual image is adjusted using axial rotation, so that the angulus is seen superiorly. As the
endoscope is advanced toward the pylorus, the angulus will disappear from view superiorly and the
pylorus will be brought closer and its channel will become more apparent. The goal with advancement
of the insertion tube is to position the pylorus inferiorly in the visual field; the so-called setting-sun
position (Figure 574A and B). Remember that the tip of the side-viewing endoscope is pointing to an
area below the endoscopic image. Therefore, objects disappearing from view in the lower portion of
the image field are brought into close proximity with the tip. Restated, the area that is just disappearing
from view at the bottom of the view field is the area that the tip of the endoscope is about to traverse
next. With continued slow advancement and slight upward deflection (see Figure 572B), the tip of the
endoscope will intubate the pyloric channel as it disappears from view; typically, the mucosa will
suddenly rush by and the villous mucosal texture of the duodenal bulb will appear (Figure 574C and
D). When viewed close up with the magnification of a side-viewing instrument, the villous mucosal
pattern of the normal bulb has an appearance found nowhere else in the gut.
(5794)Figure 574. Entry into the duodenum requires proper alignment of the duodenoscope
tip with the pyloric channel. Diagram (A) and endoscopic view (B) of the so-called setting-sun
position. This is accomplished by advancing the insertion tube of the instrument until the
pylorus reaches the lower portion of the endoscopic field. At this point, the left thumb is used
to flex the tip of the duodenoscope up to align it with the channel of the pylorus, whereupon
the tip usually will suddenly advance into the duodenal bulb. Diagram (C) and close-up
endoscopic view (D) of the mucosa of the duodenal bulb as the tip of the duodenoscope passes
through the pylorus.
As the pylorus is approached, it may deviate from the inferior midline position, causing failure of the
endoscope tip to engage the pyloric ring. There are two strategies for rectifying this problem. First, lock
the lateral flexion control knob and then turn it to bring the pylorus into the proper inferior midline
position; then proceed as usual. If this is unsuccessful, something may be causing the tip of the
endoscope to stray from the desired course. As a second strategy, withdraw the insertion tube and
down-deflect the tip to provide a tubular view of the antrum with the pylorus centered or superior in the
visual field; then examine the terrain of the greater curvature of the antrum for extrinsic compression or
another obstacle that may be preventing the tip from sliding toward the pyloric ring in the usual fashion.
This information can then be used to plan an alternate route to the pylorus that will be more suitable for
intubation of its channel.
If difficulty entering the duodenum is still encountered, variations in stomach motility or anatomy may
be the cause. Hypermotility, particularly in the early stages of learning, may require pharmacologic
control by administration of atropine 0.6 mg or glucagon 0.2 mg intravenously. When the stomach is
horizontal, the approach to the pylorus will be oblique rather than en face and the endoscope will enter
the duodenum with the tip in the neutral position; hence, upward deflection is not required. With
J-shaped stomachs, the endoscope prefers to wrap around the greater curvature of the stomach while
the tip resists advancement toward the pylorus. Overinsufflation of the stomach will exacerbate this
problem; it often helps to deflate the stomach to reduce the arc and use a moderate amount of effort to
advance the tip toward the pylorus. Whatever the gastric configuration, it may be helpful to flex the tip
upward and identify the retroflexed endoscope at the cardia. Dropping straight down along a 6 o'clock
axis from there will frequently result in visualization of the pylorus.
Examination of the duodenal bulb is easily accomplished and can be useful in plotting a course into the
descending duodenum. If the endoscope tip is deflected downward and the insertion tube withdrawn
slightly, the entire length of the duodenal bulb can be seen. With the patient still in the modified left
lateral decubitus position and the control section in the 12 o'clock axial position, the superior, anterior,
and posterior walls of the bulb can be identified in the top, left, and right sides of the endoscopic field,
respectively. The apical fold (superior duodenal angle) is usually visible, and the duodenal lumen may
be seen to angle in a posteroinferior direction. Entering the descending duodenum, therefore, requires
gentle advancement of the endoscope as flexion of the left wrist and elbow rotate the control section
clockwise into the 3 o'clock position and the left thumb is flexed, bringing the endoscope tip into an
upward vertical flexion. The view of the lumen is usually lost during this maneuver, and mucosa is seen
to slide across the visual field.
Parenthetically, it is possible to overorganize the maneuvers through the bulb. Because the descending
duodenum generally turns in a posterior direction (from 3 to 6 o'clock in the visual field), one can do
whatever is necessary with the tip controls to achieve a tubular view of the duodenum and "pursue the
lumen" by gently advancing the insertion tube. Leaving the bulb may be as simple as downward
deflection and right lateral deflection. Whichever method is used, a little blind mucosal "slide by"
frequently occurs; gentleness is therefore required. A constant tactile appreciation of the degree of
resistance encountered is necessary to optimize patient comfort and identify obstructing duodenal
lesions that can be overlooked with the side-viewing duodenoscope.
After the descending duodenum is entered, the patient should be turned into the prone position as
described previously. This maneuver is equivalent to further clockwise axial rotation of the insertion
tube (except that, in this case, the patient is rotated), and it facilitates a tubular view of the duodenum
that can be maintained with minimal insufflation and affords the best patient position for subsequent
imaging of both ductal systems.
Identifying the Duodenal Papilla

Once the descending duodenum has been entered and the patient turned into the prone position, the
goal is to locate the main duodenal papilla. If there is peristaltic activity in the duodenum, it should be
controlled with intravenously administered glucagon (0.2 to 0.4 mg); an immobile duodenum facilitates
early discovery of the papilla. Hyoscine N-butyl bromide has also been used to induce duodenal atony
and has been found to be essentially equivalent to glucagon for this purpose.8,9(5795) Chang et
al.9(5796) found that intravenously administered glucagon and hyoscine N-butyl bromide were
equivalent when used for control of duodenal motility during ERCP. In a prospective, double-blind trial,
Van Dam et al.10(5797) found that the administration of somatostatin analog (octreotide) before ERCP
did not reduce the dosage of glucagon required for maintenance of duodenal atony during the
procedure and did not reduce the incidence of postprocedure nausea.
With the patient prone and the control section held in the 12 o'clock position, the endoscopic view
should be that of the medial wall of the duodenum. If the papilla is not in view, usually the tip of the
endoscope has reached the mid-descending duodenum and the papilla is located more proximally.
Slight downward deflection will usually afford a tubular view of the descending duodenum. This is a
quick way to determine that further advancement of the insertion tube is not necessary. If the
concentric plicae circularis are visible and there is no suggestion of a longitudinal fold or papilla, return
the vertical flexion knob to the neutral position so the posteromedial wall of the duodenum (from 10 to
2 o'clock in the visual field) can be scrutinized as the insertion tube is slowly withdrawn. Visual clues to
the location of the papilla include (1) a vertical (longitudinal) fold that follows the axis of the lumen,
often crossing the duodenal folds at right angles (Figure 575) and (2) a trickle of bile on the mucosa.
Following either in a proximal direction will usually lead to the major papilla.
(5798)Figure 575. Schematic diagram of the main duodenal papilla. Although the main
papilla is variable in its appearance, certain common anatomic features aid in its recognition.
Familiarity with these anatomic landmarks is essential to successful cannulation and
therapeutic intervention.
Generally, the papilla can be recognized as a protuberant mound (see Figure 575); its color is slightly
more pink or red and the surface texture more papillary than the surrounding duodenal mucosa. It may
exhibit a central dimple within a concentric ring. Typically, a cylindrical bulge deforms the lumen
cephalad to the papilla and merges with the duodenal wall after crossing one or two of the
circumferential folds. This longitudinal bulge is produced by the intramural segment of the distal
common bile duct as it traverses obliquely through the duodenal wall (see Figure 575).
There are various reasons for failure to locate the papilla. The most common is premature termination
of the proximally moving search pattern. This occurs largely due to a fear of falling back into the
antrum and concern over reintubation of the pylorus, a more difficult task with the patient in the prone
position. Because the descending duodenum often courses posteriorly and then horizontally, the fold at
this angle may be mistaken for the junction of the bulb and descending duodenum, in which case the
search may be stopped short of the papilla. Locking the horizontal control knob in the extreme right
position may help to secure the endoscope in the duodenum and prevent the tip from falling back into
the antrum. Occasionally, the longitudinal fold or horizontal circular folds are quite prominent and
obscure the papilla; it may be necessary to use the cannula as a probe to lift the horizontal folds or
move the longitudinal fold to the left in an attempt to locate the papilla. Rarely, the papilla may be
diminutive or absent, or inflammation, edema, and extrinsic deformity of the duodenum may obscure
the conventional landmarks. Under these circumstances, the papilla can be located only by close
scrutiny of the medial wall of the duodenum during a period of meticulous and patient searching. A
fluoroscopic image of the endoscope position within the duodenal sweep may help to determine
whether the visual search is being directed toward the right location.
The minor papilla has a relationship to the major papilla that is rather consistent (Figure 576). It lies
cephalad and slightly to the right or anteriorly. Occasionally, the minor papilla is relatively large and
may be confused with the major ampulla. If a papilla large enough to qualify as a major structure has
been identified, but it is situated between the confluence of two horizontal folds that merge to the left,
or if it is unusually firm and unyielding, it may be rewarding to quickly move distally in the duodenum a
few centimeters to be certain that the real major papilla is not hidden from view.
(5799)Figure 576. Endoscopic view of the minor papilla. It is typically located proximal and
anterior to the major papilla. It may be so prominent as to be confused with the major papilla
or so indistinct that it can be seen only as an evanescent black hole during secretin injection. It
is usually located between two duodenal folds.
Preparing for Cannulation
Positioning the papilla in the field of view is the single most important component of a successful
procedure. The orifice is usually evident and cannulated with the most authority and success when it is
viewed up close and directly en face from precisely angulated vantage points. Attempts to cannulate
from a distance are often haphazard and unsuccessful. Taking the time to position the papilla properly
in the visual field, to observe the detail of the papillary structure, to define the relationship of the papilla
to the intramural segment of the common bile duct, and to plan the approach to the pancreatic and
common bile ducts will increase the likelihood of successful cannulation.
There are two standard positions for the insertion tube during attempts to cannulate the papilla. The
greater curve position, or long position, refers to the position of the insertion tube along the greater
curvature of the stomach such that an intragastric loop is formed (Figure 577A). The endoscope is
often in this position immediately after entering the descending duodenum. Occasionally, this position
is necessary to gain access to the papilla or to provide the best approach for cannulation of the bile
duct. In general, however, the greater curve position is undesirable because it is more uncomfortable
for the patient and subtle movements of the control handle are absorbed by the intragastric loop rather
than being transmitted to the tip.
(5800)Figure 577. A, Radiograph of a duodenoscope that has been advanced into the greater
curvature position. B, Radiograph showing the duodenoscope in a straightened position along
the lesser curvature of the stomach. Optimum positions for cannulation frequently lie
somewhere between these two extremes. The main duodenal papilla typically appears on the
left side of the endoscopic view in the greater curvature position and on the right side in the
straightened position.
The lesser curve position, or short position, refers to the position of the insertion tube along the lesser
curvature of the stomach (see Figure 577B). It is achieved by locking the lateral deflection knob in the
extreme right position, applying a clockwise torque to the axis of the insertion tube, and withdrawing
slowly, a sequence of steps sometimes referred to as the straightening maneuver. The image of the
papilla will usually shift from left to right in the visual field, and an oblique, right-to-left, close-up position
usually results. Frequently, paradoxical motion occurs during the straightening maneuver; the
endoscope tip will initially advance further into the descending duodenum while the insertion tube is
withdrawn because the intragastric loop is being reduced. With the duodenoscope in the lesser curve
position, the length of the insertion tube within the adult patient is typically only 60 to 70 cm. This
position provides greater patient comfort, allows rotational movements of the control handle to be
transmitted down the axis of the insertion tube to the endoscope tip in a one-to-one fashion, and often,
with left lateral flexion, provides the best approach to the pancreatic duct. It is important to use the
straightened position when it works but not to become committed exclusively to its use.
Intermediate endoscope positions are sometimes necessary to provide the best approach to the
papilla. Starting from the lesser curvature position, adjust the vertical flexion to bring the endoscope tip
close to the papilla, and then advance or withdraw the insertion tube as necessary to center the papilla
in the endoscopic view. Remember to stand a proper distance away from the patient; this
approximates a straight insertion tube outside the patient and allows subtle wrist, arm, or body
movements to be transmitted down the length of the insertion tube to the tip of the endoscope. At this
point, the control knobs should be locked and only small, fine motions or adjustments of the flexion
knobs will be necessary to properly orient the duodenal papilla.
Care should now be taken to optimize the conditions that favor successful cannulation; be certain that
the patient is comfortable and the vital signs are stable, and adjust sedation accordingly. If the
duodenum is filled with foamy bilious secretions, instill approximately 10 ml of water mixed with a few
drops of simethicone to improve visibility. This can be accomplished via the accessory channel port
using a blunt-ended needle and syringe. Duodenal peristalsis should be inhibited with intravenous
administration of 0.2 to 0.4 mg of glucagon every 10 to 15 minutes as needed to provide a stable
target. On occasion, significant respiratory motion will be transmitted to the duodenum and result in a
moving target, in which case it is sometimes helpful to bring the patient's right leg into a flexed position;
this brings the patient out of the prone position enough to reduce respiratory interference. Similar
changes in position are sometimes useful when duodenal motion is caused by vascular structures.
With the papilla identified and brought into proper position, it is tempting to attack immediately. Not
quite yet! Additional time spent in observation and planning will increase the likelihood of expedient
cannulation. Observe the papilla and look for evidence of a duct orifice (even two orifices). The orifice
is usually evident as a punctate central depression amidst the papillomatous fronds of the papilla.
Careful observation may reveal transient sphincter opening, manifested as an evanescent black hole
or a trickle of bile from the orifice. Occasionally, a small cluster of prominent papillomatous fronds or a
spot that is redder or darker than the rest of the papillary mound provides clues to the location of the
orifice. Separate orifices are present in less than 10% of patients; the best opportunity for this
discovery occurs before the papilla has been traumatized by the cannula. If the two orifices lie in a
vertical plane, the more cephalad one belongs to the biliary tree. If the two orifices lie in a horizontal
plane, usually the left (posterior) orifice is in continuity with the biliary tract. The orifices may be
separated by 2 mm or less. Now, memorize the details of the precisely mapped papilla; edema,
friability, or worse (submucosal injection) may occur in the course of attempted cannulation and
obscure its topography. Next, move the tip of the endoscope away from the papilla to observe the axis
of the intramural bulge proximal to the papillary mound. Successful cannulation of the bile duct will
require alignment of the axis of the cannula with the axis of the intramural bulge. Finally, with the
cannula advanced a short distance into the duodenum, observe how tiny movements of the wrist, arm,
and body influence the orientation of the cannula with respect to the papilla. This information is useful
in making fine adjustments while cannulating because the right hand is occupied with grasping and
advancing the cannula and is not readily available for operating the control knobs.
Cannulation
Once the cannula and desired duct are in proper alignment, a coordinated series of moves should
result in successful cannulation. The cannula is grasped by the thumb and forefingers of the right hand
and advanced smoothly toward the papilla. During this process, refinements in cannula position and
trajectory are made by positional changes of the left thumb, left wrist and arm, or body. The left thumb
controls the cannula elevator, which allows for adjustments of the cannula tip in the vertical plane.
Subtle changes in the position of the left wrist and arm or body induce rotational movements of the
insertion tube, which may enable tiny adjustments of the cannula tip in the horizontal plane. When the
tip of the cannula engages the papillary orifice, gentle but authoritative advancement of the cannula
with the right hand should be matched by maintenance of position and proximity to the papilla with the
left hand.
Cannulation of the Desired Duct
Because both the common bile duct and the pancreatic duct usually empty via the same orifice,
selective cannulation of the desired duct is part of the challenge of ERCP. Early success requires both
familiarity with the anatomy of the papillary structures and a strategy for approaching the desired duct
(see Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas). In general, the main
duodenal papilla has a common channel of variable length that then bifurcates into the two ductal
systems. At the bifurcation, the ducts are separated by a mucosal septum; from this point, the common
bile duct tends to run in a cephalad direction whereas the pancreatic duct courses more medially
(Figure 578A and C). Although the precise spatial orientation of the septum varies from person to
person, it is useful to envision the septum as a diagonal membrane running through the circular lumen,
analogous to the red international road sign that signifies that entry is prohibited, that is, "Do Not
Enter."
(5801)Figure 578. A, Diagram of the usual directions of the bile duct and pancreatic duct in
their terminal courses through the papilla. B, Endoscopic view of cannulation of the pancreatic
duct. The duct courses straight in or angles anteriorly to the right. C, Endoscopic view of
cannulation of the bile duct by orienting the cannula in an immediately cephalad and obliquely
posterior (left) direction.
Cannulation of the pancreatic duct tends to be easier than cannulation of the bile duct. With the
endoscope in a straightened position, orient the papilla in the visual field so that it is directly en face,
but with a slight left-to-right approach (i.e., a smidgen more of the posterior and slightly less of the
anterior surface of the papilla are visible) (see Figure 578B). As the cannula is advanced toward the
orifice, its axis should be nearly perpendicular to the medial duodenal wall. Once the orifice is engaged,
the elevator is dropped down and the cannula is directed (using the left wrist and arm) slightly to the
right, this being the most favorable trajectory for sliding the cannula underneath the septum and
selectively into the pancreatic duct. If the cannula encounters resistance to advancement and begins to
push the tip of the endoscope away or deform the papilla, recheck to see if the trajectory of the cannula
matches the desired approach and make tiny adjustments as needed by changing the position of the
control handle with the left arm and wrist and the elevator position with the left thumb.
A few admonitions regarding pancreatic duct cannulation: If free cannulation is achieved, cease
advancement soon after the rings etched on the cannula disappear into the papilla; excessively deep
cannulation can result in injury to a side branch of the main pancreatic duct or the genu of the main
duct. In addition, the pancreatic duct is easily deformed and the segments of the duct wall on either
side of the tip become coapted and therefore occlude the duct if the cannula tip approaches from the
side rather than straight on. In this situation, slight withdrawal of the cannula will correct this error and
result in successful pancreatography.
Cannulation of the common bile duct is often the more challenging, particularly if the pancreatic duct
has been cannulated first, and requires a different strategy. The tip of the endoscope is brought into a
"tucked under" position; this is usually accomplished by gently advancing the insertion tube and turning
the lateral angulation (right-left) knob to the right. In this position, the papilla is viewed from below
rather than directly en face, and a smidgen more of the anterior and less of the posterior surface are
visible (see Figure 578C). As the cannula is advanced toward the orifice, its axis should be almost
aligned with the axis of the cylindrical bulge caused by the intraduodenal segment of the distal common
bile duct; a slight posterior direction (to the left or toward 11 o'clock in the visual field) is desirable
initially. Once the orifice is engaged, the elevator is raised slightly to lift the papilla, this maneuver being
favorable for sliding the cannula above the septum and selectively into the bile duct. Occasionally,
resistance to advancement of the cannula is encountered and the papilla becomes deformed in a
cephalad direction. This situation is caused by misalignment of the cannula axis with the takeoff of the
bile duct, and the tip of the catheter becomes trapped in false channels within the papilla. In this
instance, it is often helpful to pull back slightly on the insertion tube while the cannula is still engaged in
the papilla. This helps to realign the axis of the catheter with that of the bile duct and may result in a
sudden free cannulation.
Common Problems During Cannulation
Frequently, the cannula enters the papillary orifice but stops advancing after 3 to 10 mm. When this
occurs, the cannula tip may have become entrapped in blind passages en route to the main channel.
While maintaining the endoscope's proximity to the papilla and the cannula tip firmly but gently
engaged in the orifice, make small corrections in the trajectory of the cannula by adjusting the vertical
orientation with the cannula elevator, insertion tube, or the horizontal orientation using axial torque
generated by movements of the left wrist, arm, and body. During these maneuvers, authoritative but
not forceful advancement of the cannula tip may cause it to slide into a free position in the ampulla or
one of the ducts. If this does not succeed, reorientation of the tip of the endoscope and redirection of
the cannula may be necessary. This may require repositioning of the insertion tube according to the
steps described previously, beginning with reorientation of the papilla in the field of view and
realignment of the axis of the cannula with the assumed position of the desired duct. Note that the
cannula may have been directed as calculated but the assessment of the position of the duct may have
been incorrect.
If the cannula tip is engaged in the papilla but there is resistance to further advancement despite the
maneuvers described previously, injection of a small volume of contrast material under fluoroscopic
guidance can provide useful information. However, this should be done cautiously. If the catheter is
impacted, there is an increased risk of submucosal injection, particularly if a long tapered cannula is
being used. Although this is seldom harmful, radiographic contrast media are hypertonic and draw fluid
into the area. The resulting edema can distort landmarks to the point that the procedure must be
aborted. Several days are required for this edema to resolve before another attempt at ERCP is
feasible. Therefore, it is essential that the ERCP team guard against submucosal injection and
recognize the onset of this complication as quickly as possible. To do so, a few caveats must be
observed: (1) the endoscopist should not request an injection of contrast while the cannula is impacted
under great pressure; (2) the person responsible for injecting contrast material should be trained both
to inject slowly and to refrain from injecting against resistance; and (3) members of the team should
have separate responsibilities for monitoring the results of injectionthat is, one person should focus
attention on the endoscopic image; the other on the fluoroscopic image. The signs of a submucosal
injection include endoscopic visualization of a clear grayish swelling of the mucosa and a focal
nonmoving amorphous collection of contrast material on fluoroscopy.
If the catheter tip is in the ampulla, simultaneous filling of both ductal systems may occur. This is
suboptimum for a number of reasons: (1) contrast is likely to reflux back into the duodenum during the
injection and obscure radiographic detail of the ducts; (2) the pancreatic duct generally fills before the
biliary tree, and sustained injection can therefore lead to overinjection of the pancreatic duct, resulting
in opacification to the level of the acinus ("acinarization") and an increased likelihood of post-ERCP
pancreatitis; and (3) therapeutic interventions require free duct cannulation. Nevertheless, partial
opacification of the ducts will usually provide fluoroscopic images with enough information about the
relationships of the ducts and the papilla to obtain selective cannulation. When repeated attempts at
cannulation fail, radiographs should be obtained. These provide greater anatomic detail and may
reveal the nature of the problem. For example, a radiograph may demonstrate a small ventral
pancreatic duct (pancreas divisum) that was not visible on fluoroscopy and thereby shift emphasis to
cannulation of the minor papilla. In a very few patients, the ventral pancreatic duct may be absent (see
Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas).
The Difficult Cannulation
In perhaps half of ERCP procedures, the anatomy is straightforward, the desired ducts are identified,
and a successful study is completed. In perhaps another 25% of attempts, the procedure is completed
but requires more time and repeated efforts at cannulation, with the standard techniques being
ultimately successful. In about 15% to 25% of attempts, successful cannulation requires extraordinary
concentration and patience and adherence to a systematic algorithm of procedural skills and
techniques.
When standard attempts at cannulation are unsuccessful, it is often wise to consider whether to quit for
now. Often, considerable time has been spent before coming to the realization that this is a difficult
cannulation. Lengthy procedures result in fatigue for the endoscopist and frustration for the nurses and
are hard on the patient. Know when to quit. Decide this ahead of time. The better part of valor may be
to recognize one's limitations and terminate the procedure. For patients with compelling indications for
ERCP, a second attempt may succeed after several days' recovery of papillary topography. In
technically challenging cases, it is often prudent that subsequent procedures be attempted in centers
where there is a high level of expertise.11
When difficulty is encountered, a simple rule of thumb is "change something." Continued efforts at the
same maneuvers using the same cannula and even the same endoscopist are pointless. Variation
keeps the team interested and focused, and if one approach does not work within a reasonable
amount of time, it probably will not work.
The simplest change is to opt for a different cannula. This is an appropriate response when no
correctable problems with trajectory or alignment can be identified. A metal-tipped cannula, which may
offer less friction to the mucosal surface, may work where a standard Teflon catheter fails. These are
available in a variety of shapes (e.g., cone tip, bullet tip, long nose) that may have advantages in some
situations.
The guidewire technique is another alternative. This entails the use of a guidewire with a radiopaque tip
and a cannula with a special tip that is visible fluoroscopically. The guidewire is advanced so that it
protrudes a few millimeters beyond the end of the cannula, and selective cannulation of the desired
duct is attempted. Gentleness is mandatory because the guidewire tends to be more traumatic to the
papilla. Once the alignment is correct, the guidewire is advanced under fluoroscopic guidance into the
desired duct; the catheter is then advanced over the guidewire. After free duct cannulation is achieved,
the guidewire is removed, fluid is aspirated from the duct to remove air bubbles from the cannula, and
contrast material is then injected. A side-arm adapter allows injection of contrast material while the
guidewire is still in place. Injection is easier if the lumen of the cannula is occupied by a 0.021-inch
rather than a 0.035-inch guidewire.
When a patient's particular anatomy prohibits a suitable "tucked-under" position, the sphincterotome
technique is useful. The cannula is removed and replaced with a standard double-lumen
sphincterotome. Changes in the trajectory of the tip can be made by bowing the sphincterotome within
the duodenum until the papilla is successfully engaged and cannulated (Figure 579). A
radiopaque-tipped guidewire can also be used with this technique as described earlier.
(5802)Figure 579. Diagram of the use of a papillotome to align the tip of the cannula with
the axis of the bile duct to facilitate cannulation. This technique is useful when it is not
possible to obtain an optimum tucked-under position for cannulation of the bile duct.
Occasionally, it is necessary to advance the insertion tube and bring the endoscope into the long
(greater curve) position. This may provide the best cephalad views of the papilla for selective
cannulation of the common bile duct. If cannulation attempts are unsuccessful, turning the patient
obliquely or almost laterally may place the cannula and intramural segment in better alignment.
Radiographic Technique
Radiographs of high quality can be obtained safely and efficiently if one adheres to the following
technical suggestions. In general, injection of contrast material should be performed slowly and with
careful attention to the fluoroscopic image as the duct fills with contrast material.
Pancreatography
The pancreatic duct is easily recognized fluoroscopically. After flowing medially for a short distance, the
column of contrast material courses cephalad for a short distance, then turns medially again and
crosses over the spine, typically at the level of the second lumbar vertebral body (see Chapter 59: The
Normal Retrograde Pancreatogram and Cholangiogram). The normal duct is of small caliber and the
total volume of contrast material required to fill it is small (only 1 to 4 ml). As the contrast material
reaches the left upper quadrant, the duct in the tail of the gland frequently lies in the splenic hilum,
where it may bifurcate. Radiographs should be taken when the contrast material reaches the tail of the
pancreatic duct. It is unnecessary and unwise to wait for fluoroscopic recognition of lateral filling of the
main duct branches; these are usually apparent on the radiograph. In the absence of disease,
pancreatic secretions rapidly clear the contrast material from the duct. Hence, unless overdistention
and pain or acinarization occur, contrast material injection should be sustained until immediately after
the radiograph is taken.
Poor technique can lead to complications. Rapid, high-pressure injections, overdistention of the duct,
acinarization of the gland, and repeated injections of contrast material into the pancreatic duct are
associated with an increased incidence of post-ERCP pancreatitis. Care should be taken to avoid
these breakdowns in proper technique. Once the duct has been entered with a cannula that has a
radiopaque tip, it is usually possible to discern from the fluoroscopic image, without further contrast
material injection, whether the same duct has been recannulated; this technique is helpful in reducing
the number of injections into the pancreatic duct.
Cholangiography
The bile duct is a relatively large volume system and is recognized fluoroscopically by its larger
diameter and cephalad course toward the right upper quadrant. Although its relatively large capacity
has little importance in terms of the absolute amount of contrast material used, it has several
implications with regard to the technique of contrast material injection.
First, a large-volume syringe (e.g., 20 to 30 ml) should be used for contrast material injection. This will
avoid the introduction of air into the cannula and duct as syringes are attached and detached from the
cannula. When it is necessary to exchange syringes or when cannula exchanges are made over a
guidewire, the syringe should be inverted and bile aspirated to remove air from the cannula before
injecting the contrast agent.
Second, a radiograph should be taken during the early phases of injection of the common bile duct. On
occasion, small stones will flow proximally on the meniscus of the bile-contrast material interface and
later become obscured as the duct becomes more densely opacified with contrast material. A habit of
taking radiographs during the initial phase of injection will prevent missing these small stones. For
similar reasons, when dilated ducts are present or when small stones may be present, dilute contrast
material (20% mixture) should be used to avoid obscuring the filling defects within densely opacified
ducts. The fluid dynamics of flow within a tubular structure are such that flow is most rapid in the center
and slower at the margins. This is of relatively little consequence if the duct is of normal caliber. If the
duct is dilated, however, a relatively large volume of contrast material must be injected before the duct
becomes fully opacified. The duct is then densely opacified, so that small filling defects may be very
difficult to detect. When initial filling of the bile duct reveals a stricture, it is advisable to instill only a
minimum volume of contrast material proximal to the stricture. The stricture is visualized by minimum
injection pressure, just enough to keep the contrast material column from collapsing.
Third, the presence or absence of a gallbladder has implications for adequate visualization of the
intrahepatic bile ducts as well as for patient tolerance for the procedure. In patients who have not
undergone cholecystectomy, contrast material may preferentially flow via the cystic duct into the
compliant gallbladder, which often results in poor filling and inadequate opacification of the intrahepatic
ducts. To overcome this problem, the tip of the catheter must be advanced into the bile duct to a level
proximal to the cystic duct insertion. When it is necessary to demonstrate the intrahepatic ducts in still
greater detail, the best images are obtained by exchanging the cannula over a guidewire for a
balloon-tipped catheter; the balloon is used to occlude the common hepatic duct during further
injection. This "occlusion cholangiogram" technique is also useful in patients who have undergone
sphincterotomy. In patients who have undergone cholecystectomy, injection of contrast material into
the relatively noncompliant biliary tree can result in high pressures, distention, and pain. In this
instance, it is necessary to aspirate bile and then reinject the bile-contrast material mixture, repeating
this cycle until adequate filling of the ductal system is achieved. Syringe exchanges during this process
are generally not necessary.
Exceptional Circumstances
Periampullary Diverticulum
Periampullary diverticula are common and come in a variety of shapes and sizes. Smaller diverticula
are usually located to the upper left and less often to the upper right of the papilla. Larger varieties may
drape over the intraduodenal segment of the common bile duct like saddlebags or may totally engulf
the papilla. Although a diverticulum may initially present a strategic obstacle, the papilla is usually easy
to cannulate once the proper approach is achieved. The main risk associated with these structures is
perforation, and this can be prevented by avoiding haphazard probing of the diverticulum.
On occasion, the major papilla will be located within a large diverticulum. Sometimes it is possible to
coax it out into the duodenal lumen by probing the mucosa on the crest of the diverticulum or by
applying suction when the tip of the endoscope is in close proximity to the diverticular orifice.
Experienced endoscopists can enter the diverticulum, identify the papilla, cannulate its orifice with a
balloon-tipped catheter, and eventrate the papilla into the duodenal lumen.
Pancreas Divisum
Under certain circumstances, it is necessary to attempt cannulation of the minor ampulla. Either the
pancreatogram obtained via the major papilla shows a ventral pancreas or repeated attempts at
opacification fail to provide a pancreatogram, thus suggesting the possibility of a dorsal pancreas.
Cannulation of the minor ampulla requires a high degree of manual dexterity and extensive ERCP
experience (see also Chapter 72: Congenital Anomalies of the Pancreas).
It is first necessary to find the minor papilla. It is usually located approximately 2 cm above and to the
right of the major papilla. It may be situated between two horizontal folds that merge to the left of the
minor papilla. The appearance of the minor papilla is quite variable; it may be so large as to be
confused with the major papilla or it may require secretin stimulation to detect it as a small, evanescent
black hole spewing clear fluid (see Figure 576).
Successful cannulation of the minor papilla requires special cannulas and an extremely close-up
position. In general, an intermediate endoscope position with left lateral flexion is necessary for both
visualization and cannulation. To achieve this position, it is often necessary to withdraw the insertion
tube until the minor papilla is located. Then, while the insertion tube is slowly advanced, deflect the tip
upward and add left lateral flexion. If the orifice of the minor papilla is not identifiable, intravenous
administration of secretin will help to make it more prominent and may cause it to spew clear fluid. For
simple pancreatography, a metal-point cannula (Cremer's catheter) will usually suffice. However, we
prefer a Soehendra dilating catheter with a 0.018-inch-diameter Roadrunner guidewire (Wilson-Cook
Medical Inc., Winston-Salem, NC) protruding from its orifice. The wire is used to first engage the orifice
of the minor papilla, which tends to prevent slipping off the mark until the orifice can be intubated with
the cannula tip. The guidewire can then be removed and contrast material injected. If it is necessary to
dilate the duct, the guidewire can be reinserted through the dilating catheter and into the ductal system
without having to exchange catheters.
ERCP After Abdominal Surgery

Technical problems with upper gastrointestinal endoscopy after abdominal surgery are discussed in
Chapter 52: Endoscopy in the Postoperative Upper Gastrointestinal Tract. However, problems unique
to ERCP in the postoperative upper gastrointestinal tract are encountered often enough to merit a brief
discussion.
A simple but often vexing problem may occur in the patient who has had prior upper abdominal surgery
in and around the duodenum but whose duodenum remains in continuity. The normal rotational
mobility of the duodenum is lost, and frequently, the papilla is not situated in its usual posteromedial
position. Furthermore, it may be difficult to bring the insertion tube into a fully shortened configuration
because of fibrosis around the proximal descending duodenum and bulb. As a result of duodenal
fixation, it is difficult or impossible to align the axes of cannula and duct and failure is frequent.
Billroth Type II Gastrojejunostomy
The patient with a prior subtotal gastrectomy and Billroth type II anastomosis presents several different
technical challenges.
The first problem is to identify and enter the afferent loop. This is best accomplished by beginning the
examination with a forward-viewing endoscope. The afferent loop will be found most often just beyond
the stoma, lying at either the 3 o'clock or the 9 o'clock position in the visual field, depending on the type
of surgical anastomosis created when the gastrojejunostomy was constructed (see Chapter 52:
Endoscopy in the Postoperative Upper Gastrointestinal Tract). Typically, the efferent loop is the one
most easily entered and, once entered, the lumen courses in a rather straight caudal direction. Two
stomas may be relatively far apart or may be separated by only a raphe, again depending on the
surgical technique used to construct the stoma. The afferent limb may be seen as a slit that, when
entered, is identified as the afferent loop by the presence of bile.
Occasionally, an enteroenterostomy will have been constructed at surgery by creating an anastomosis
between the afferent and the efferent limbs of the small bowel at a point several centimeters proximal
and distal to the stoma. This is usually done in an effort to divert bile from the stomach. The
enteroenterostomy will usually be appreciated as three lumens. Identification of the continuation of the
afferent limb may be difficult. The presence of larger amounts of bile in one limb and fluoroscopic
observation of the course of the instrument may be helpful, but a "trial-and-error" approach with
exploration of each limb is often necessary to reach the proximal afferent limb and the main duodenal
papilla.
The next problem is to traverse the afferent loop safely and successfully, arriving at the papilla in a
"cannulatable" position. Here again the instrument of choice is a forward-viewing endoscope. The
afferent loop may be somewhat irregular in its course, being tented by external adhesions. Thus,
abrupt turns are best and are most safely traversed with the end-viewing instrument. The length of the
afferent loop varies considerably, occasionally being so long that the papilla is beyond the reach of the
fully inserted endoscope. The closed end of the duodenum can usually be reached, however, and the
papilla is identified by withdrawing the instrument a few centimeters. Because the bile duct courses in a
cephalad direction, often parallel to the long axis of the duodenum, it is often easier to cannulate this
duct because the endoscope automatically assumes a "tucked-under" position.
It is important to remember that the conventional 12 o'clock visual image position for cannulation is
now at 6 o'clock, with downward flexion therefore being necessary for successful bile duct entry.
Precise manipulation of the cannula tip must be accomplished by deflecting the tip of the
forward-viewing endoscope because of the absence of a cannula elevator. It may be feasible to impale
the tip of the cannula on the papillary orifice and then rotate the cannula and the endoscope until the
axes of the cannula and duct are lined up and cannulation occurs. If cannulation of one or both ducts
fails with the forward-viewing endoscope, the process must be repeated using a side-viewing
instrument. The afferent loop may be marked for subsequent identification by taking (and discarding)
several biopsies with a forceps.
Roux-en-Y Gastrojejunostomy
Roux-en-Y reconstruction surgery serves a number of important functions (see Chapter 52: Endoscopy
in the Postoperative Upper Gastrointestinal Tract). Partial gastrectomy with Billroth type II anastomosis
may result in symptomatic alkaline bile reflux, requiring interposition of a length of gut between the
gastric pouch and the biliary tree. A Roux-en-Y gastrojejunostomy is used in several types of gastric
bypass operation. Choledochojejunostomy by means of a Roux-en-Y limb of small bowel is used for
permanent decompression of a biliary tree obstructed by stones or cancer. Choledochojejunostomy is
one of the several procedures incorporated in the Whipple operation for pancreatic malignancy. The
affect of each of these procedures is to separate the stomach from the biliary tree by a variable
distance and route. Traversing such routes is a rather brief but frequently unsuccessful task. It is
gratifying, however, to occasionally arrive at the papilla; usually this requires a pediatric colonoscope.
When successful, the approach to the papilla will be the same as that described previously for the
patient with a standard Billroth type II anastomosis.12(5803)
A Roux-en-Y loop is used for retrograde drainage of the pancreatic duct in patients with chronic
pancreatitis in the Puestow or DuVall pancreatojejunostomy procedure. In the former case, the
pancreatic duct is opened along its length and, for practical purposes, anastomosed in side-to-side
fashion to the jejunum. In the latter procedure, the proximal duct in the body or tail of the gland is
anastomosed to the jejunum end-to-side or end-to-end. Features of the Roux-en-Y reconstruction
noted previously with regard to the approach to the main duodenal papilla obviously do not pertain in
this situation; insertion of the instrument to the descending duodenum is performed in standard
fashion. Clinical events may raise the question of continued patency of the Roux-en-Y anastomosis in
patients with chronic pancreatitis. Patency is easily assessed by pancreatography. If the anastomosis
is patent, contrast material will be seen fluoroscopically to flow into the jejunum, either in the region of
the tail of the gland or along its length. Careful observation during injection is essential because rapid
flow of relatively large volumes of contrast material into the jejunum may quickly obscure important
structural features of the duct and the anastomosis.
When permanent drainage of the biliary tree is required, a choledochoduodenostomy may be
preferable to the aforementioned choledochojejunostomy and Roux-en-Y loop. The anastomosis is
usually side-to-side between the superior wall of the duodenal bulb and the duct as it courses posterior
to the bulb. When viewed endoscopically, the anastomosis may appear as a single small round hole or
as a pair of apertures, one directed proximally into the intrahepatic ducts and the other distally to the
common bile duct. Radiographic evaluation of the bile duct is usually best performed with an extraction
balloon catheter. With the balloon inflated to occlude the anastomotic opening, contrast material can
be injected proximally and distally. Air is frequently present in the biliary system, and this may make it
difficult to determine whether stones are present or absent. Examination of a
choledochoduodenostomy is nearly always best accomplished by the use of a forward-viewing
instrument. If visualization of the common bile duct or "sump" is inadequate by the occlusive method, it
is usually feasible to cannulate the main papilla in the standard fashion using a side-viewing
instrument. The most common indication for evaluation of the distal bile duct is the so-called sump
syndrome in which stones, food material, and debris may be present in the distal bile duct.
Miscellaneous
Although technically difficult, ERCP can be performed successfully in patients with situs
inversus.13(5804)
Various diseases and conditions may preclude passage of an endoscope via the mouth or the
esophagus. In such cases, ERCP including therapeutic maneuvers has been performed via a mature
gastrostomy.1416(5805)
Postprocedure Aspects
Although patient recovery after ERCP is similar to that of other endoscopic procedures, a few
differences are noteworthy. First, because of the risk of post-ERCP pancreatitis, patients should be
cautioned against eating large meals for approximately 12 to 24 hours after the procedure. We usually
enjoin patients from ingesting anything for 4 to 6 hours after a routine diagnostic procedure. In difficult
cases or therapeutic procedures, we ask patients not to eat anything until the following morning and
then to begin with a clear liquid diet and thereafter advance their diet as tolerated. Second, patients
with sclerosing cholangitis are particularly susceptible to procedure-related bacterial superinfection. In
addition to administration of periprocedure intravenous antibiotics, we also prescribe oral antibiotics for
1 week postprocedure.
Summary
At all levels of experience, it is worthwhile to constantly review the essential aspects of ERCP with the
intent of improving one's technique. The following axioms are repeated here for that purpose.
Optimally prepare the patient. A well-informed patient is more likely to be trusting and cooperative
throughout the procedure. Take time to discover the patient's expectations and compare them with
your own; address serious differences beforehand lest they lead to future discord.
Check to be sure that the mechanical functions of the endoscopic and fluoroscopic equipment are in
working order before subjecting the patient to the risks of the procedure.
Create optimum conditions for a successful procedure. Ensure patient comfort. Control duodenal
motility. Duodenal atony facilitates early discovery and cannulation of the papilla. Bring the patient into
the prone position.
Bring the papilla into a close-up en face position for detailed inspection, and above all resist the urge
for immediate attempts at cannulation. Exercise patience while precisely delineating orifice location,
and concentrate on the desired trajectory for selective cannulation of the most clinically important duct.
Resist attempts to cannulate from afar, as these are almost always unsuccessful.
The first 10 minutes of a cannulation attempt are the most important. Everyone involved should be
watching the papilla quietly, patiently, and up close. The orifice on the papilla often opens transiently,
thus defining for all precisely where the cannula must be placed. From this point on, all observers must
concentrate on precise orifice location, although the orifice may never again display itself.
Be gentle ("stroke, don't poke"). Edema, sphincter spasm, and submucosal injections are the enemies
of a successful procedure.
Probably the most frequent cause of unsuccessful cholangiography is the failure to appreciate or
cannulate in an immediately cephalad direction, that is, the direction taken by the duct as it courses
through the intramural segment.
If at first you don't succeed, change something ... don't just try and try and try again. Develop a
systematic approach to deal with the difficult cannulation and avoid repeating the same unsuccessful
moves over and over again.
Although successful ERCP requires skill and eye-hand coordination, the learning process can be
facilitated by paying attention to specific details. Early in the process, there is considerable random or
haphazard motion, sometimes even successful, without knowing why something worked or whether it
will apply to the next patient. Learn from your mistakes. With each unsuccessful attempt at
cannulation, try to identify why the attempt failed, then adjust the approach accordingly. In this manner,
specific maneuvers and a finer understanding of anatomic points and endoscopic techniques will
evolve from this experience.
It is impossible to learn ERCP from a book. Although the general pattern of this process has been
described in this chapter, it should serve only as a reference. Individual guidance by an experienced
endoscopist in the actual "hands-on" supervised performance of the procedure is clearly the best, in
fact, the only reasonable method of attaining proficiency. Even as one gains experience, it is useful to
consult with a more experienced endoscopist when difficult cases are encountered.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Jowell PS, Baillie J, Branch MS, et al. Quantitative assessment of procedural competence. A
prospective study of training in endoscopic retrograde cholangiopancreatography [see
comments]. Ann Intern Med 1996;125:9839.
Vellacott KD, Ogunbiyi OA. Results of endoscopic retrograde cholangiopancreatography
using a mobile image intensifier. J R Coll Surg Edinb 1993;38:2169.
O'Connor KW, Jones S. Oxygen desaturation is common and clinically underappreciated
during elective endoscopic procedures. Gastrointest Endosc 1990;36:S2-S4.
Nelson DB, Silvis SE, Yakshe PN, et al. Inaccuracy of clinical assessment in detecting
hypoxemia and hypoventilation during ERCP. Gastrointest Endosc 1992;38:227.
Woods SD, Chung SC, Leung JW, et al. Hypoxia and tachycardia during endoscopic
retrograde cholangiopancreatography: Detection by pulse oximetry. Gastrointest Endosc
1989;35:5235.
Crantock L, Cowen AE, Ward M, Roberts RK. Supplemental low flow oxygen prevents
hypoxia during endoscopic cholangiopancreatography. Gastrointest Endosc 1992;38:41820.
Jowell PS, Eisen G, Onken J, et al. Patient-controlled analgesia for conscious sedation during
endoscopic retrograde cholangiopancreatography: A randomized controlled trial. Gastrointest
Endosc 1996;43:4904.
Hannigan BF, Axon AT, Avery S, Thompson RP. Buscopan or glucagon for endoscopic
cannulation of ampulla of Vater?. J R Soc Med 1982;75:212.
Chang FY, Guo WS, Liao TM, Lee SD. A randomized study comparing glucagon and
hyoscine N-butyl bromide before endoscopic retrograde cholangiopancreatography. Scand J
Van Dam J, Catalano MF, Ferguson DR, et al. A prospective, double-blind trial of
somatostatin analog (octreotide) versus glucagon for the inhibition of small intestinal motility
during endoscopic retrograde cholangiopancreatography. Gastrointest Endosc
1995;42:3214.
Kumar S, Sherman S, Hawes RH, Lehman GA. Success and yield of second attempt ERCP.
Gostout CJ, Bender CE. Cholangiopancreatography, sphincterotomy, and common duct stone
removal via Roux-en-Y limb enteroscopy. Gastroenterology 1988;95:15663.
McDermott JP, Caushaj PF. ERCP and laparoscopic cholecystectomy for cholangitis in a
66-year-old male with situs inversus. Surg Endosc 1994;8:12279.
Schapira L, Falkenstein DB, Zimmon DS. Endoscopy and retrograde cholangiography via
Gray R, Leong S, Marcon N, Haber G. Endoscopic retrograde cholangiography,
sphincterotomy, and gallstone extraction via gastrostomy [Letter]. Gastrointest Endosc
1992;38:7312.
Holderman WH, Etzkorn KP, Harig JM, Watkins JL. Endoscopic retrograde
cholangiopancreatography and stent placement via gastrostomy: Technical aspects and
clinical application. Endoscopy 1995;27:1357.
Chapter 58 Anatomy and Embryology of the Biliary Tract and

Pancreas
(5806)
(5807)
BERNARD H. HAND, M.S., F.R.C.S.
In the past, it was rightly said that variation in the biliary tree is rampant or even that variation is
constant. It has been estimated that fewer than 50% of subjects possess what is referred to as normal
anatomy when considering vascular variations,1(5808) whereas ductal variations are less common and
occur in about 13% of cases.2(5809) The word anomalous has been used to describe the differences,
which implies abnormality; but the differences must surely be regarded as normal variationsdistinct
from true congenital abnormalities. Bearing in mind that the liver, biliary tract, and pancreas develop
close to each other from small buddings of the gut tube, these variations must be expected. With the
advent of preoperative x-rays, endoscopic retrograde cholangiopancreatography (ERCP),
transabdominal ultrasound, computed tomography (CT), endoscopic ultrasonography, and
arteriography, a more realistic appreciation has become possible than was obtained from earlier
postmortem studies.
Embryology
The liver, biliary tract, and pancreas all develop from diverticula or cylinders, which appear at the
junction of foregut and midgut as early as the 19th somite embryoon or about the 25th day, although
some anatomists put this a little earlier. These endodermal buds produce the epithelium and secretory
cells, whereas the musculature, connective tissue, and blood vessels develop from the splanchnic
mesoderm and the septum transversum, the latter being invaded by a plexus of vessels derived from
the vitelline and umbilical veins.
The diverticulum for the liver and biliary tract is the first to be seen at the beginning of the fourth week,
projecting into the ventral mesentery. A further bulge develops on it, resulting in its division into a larger
cranial portion, destined to become the liver and bile ducts, and a smaller caudal portion that will
produce the gallbladder and cystic duct. The cranial portion divides into two solid cords to ultimately
produce the right hepatic duct (RHD) and the left hepatic ducts (LHD). These cords push into the
adjacent septum transversum and anastomose round clusters of mesenchyme cells that form small
blood vessels. In turn, these blood vessels become confluent with the vitelline and umbilical veins and
eventually produce the liver sinusoids.
Major blood vessels are seen in the liver by the 32nd day, namely the right and left umbilical veins, the
transverse portal sinus, and the hepatic part of the inferior vena cava; later, the right umbilical vein
atrophies. In utero, the blood is shunted from the left umbilical vein directly into the inferior vena cava
by the ductus venosus. At birth, this atrophies and remains as the ligamentum venosum.
The septum transversum becomes one of the main sites for hematopoiesis until nearly full term. It also
produces the Kupffer cells, the liver capsule, and all the ligaments that suspend the liver, as well as the
central tendon of the diaphragm.
The hepatic cords eventually acquire a lumen, and with the rapid growth of the liver parenchyma, far in
excess of other tissues, the gallbladder, cystic duct, and bile ducts increase in size and length.
Ultimately, the liver is about 10% of the total body weight; whereas at birth this proportion falls to 5%.
Bile is secreted from the fourth or fifth month.
The pancreas develops from two diverticula of the duodenal endoderm when the embryo is about 30
days old (Figure 581). The first to appear lies dorsal and slightly cranial to the hepatic diverticulum
and grows out into the dorsal mesentery to the left of the vitelline veins, reaching eventually as far as
the developing spleen. The ventral bud, paired initially, arises a little later in the angle below the hepatic
diverticulum. The left bud of the pair atrophies.
(5810)Figure 581. Embryologic development of the pancreas. 1, Bile duct; 2, gallbladder; 3,

ventral bud; 4, dorsal bud; 5, superior mesenteric and portal veins.
As the common bile duct (CBD) bud elongates and the ventral bud extends away from the duodenum,
it appears that the ventral bud is actually a branch of the CBD. Soon the duodenum rotates 90 degrees
on its long axis, and there is a more rapid growth of the left side of the duodenum. This converts the
original ventral surface into the right side and brings the ventral pancreas into the mesoduodenum in
front of the bile duct and caudal to the dorsal pancreas. The left vitelline vein, destined to become the
portal vein, thus lies behind the dorsal pancreas and in front of the ventral pancreas (viewing from
anterior to posterior). After this rotation, the mesoduodenum fuses with the posterior peritoneum,
making most of the duodenum and the whole pancreas, except for the tail, retroperitoneal in position.
At about 6 to 7 weeks, the two lobes fuse into a single organ. The ventral pancreas becomes the
uncinate lobe and the lower part of the head, and the dorsal pancreas forms the rest of the head and
the neck, body, and tail.
The duct systems of these two portions usually fuse in the head, the dorsal duct (Santorini) entry being
cranial and dorsal to the ventral duct (Wirsung). The former frequently seals completely, so that all
pancreatic juice flows via the ventral duct. The ventral duct may drain directly into the duodenum or
acquire a common opening with the bile duct. Occasionally, the two ducts do not fuse, and therefore
their respective parts of the pancreas drain separately into the duodenum. Least commonly of all, the
ventral duct obliterates and all drainage takes place through the dorsal duct. These variations should
be expected because any of them may be the normal mode of entry in higher mammalian species.
Within the now-formed gland, the major pancreatic ducts lengthen into the surrounding mesenchyme,
and by sprouting and subsequent canalization, the collecting ducts and acini are produced by the 12th
week. At about the 10th week, cells bud off from the pancreatic endoderm, become disconnected from
the duct system, and are destined to differentiate into the endocrine cells. By the 4th month, secretory
granules appear in the exocrine pancreas.
General Description
Liver
The liver is the largest of the viscera and fills most of the abdominal cavity enclosed between the
diaphragm and the ribs. It normally weighs about 1500 g, about 1/40 of total body weight. The
anterosuperior, posterior, and right lateral surfaces are in curved continuity and are referred to as the
diaphragmatic surface. Classically, the liver is divided into a larger right and a smaller left lobe by the
falciform ligament, which contains the ligamentum teres as the obliterated left umbilical vein (Figure
582).
(5811)Figure 582. Diaphragmatic surface of the liver.

The under, or visceral, surface is further divided into two additional smaller lobes by four vertical
fissures and one transverse fissure. The latter is the porta hepatis, which provides entry of vessels and
bile ducts into the liver. Anterior to this and lying between the ligamentum teres and the gallbladder
fossa is the quadrate lobe, whereas posterior to it and lying between the inferior vena cava and the
ligamentum venosum lies the caudate lobe and its caudate process, which bridges into the right lobe
(Figure 583).
(5812)Figure 583. Visceral surface of the liver and its relations.

The previous standard description of the liver lobes is no longer acceptable for clinical purposes. From
studies of human liver corrosion casts and x-rays, a segmental division has been demonstrated. This is
based on the distribution of the intrahepatic bile ducts and associated branches of the hepatic artery
and portal vein, which closely follow one another to the finest branches. There are no cross
connections between segments that are sufficient in size to maintain an area in which the ducts or
arteries have been occluded. Furthermore, although these segments are easily discernible in the casts,
they have no anatomic landmark that can be seen on the surface of the liver or within its
substance.39(5813) Extensive research has shown that the right and left lobes are demarcated by a
line, not quite sagittal in section, that runs from the gallbladder fossa to the left of the inferior vena
cava. The quadrate lobe and most of the caudate lobe, with the exception of the caudate process,
therefore lie within the left lobe.
These two main lobes have been further subdivided into segments. There have been some slight
differences in nomenclature, but it is now generally accepted that the right lobe is divided into
anteroinferior, anterosuperior, posteroinferior, and posterosuperior segments; the left lobe is divided
into mediosuperior, medioinferior, laterosuperior, and lateroinferior segments; subsegmental ducts are
also described (Figures 584 and 585). A branch of the anteroinferior segment of the right lobe lies
close to the surface under the gallbladder fossa, and this has been named the subvesical duct.
(5814)Figure 584. Segmental pattern of the intrahepatic ductsdiaphragmatic surface.
(5815)Figure 585. Segmental pattern of the intrahepatic ductsvisceral surface.

The caudate process drains into the RHD, and sometimes the right side of the caudate lobe does also.
Most of the caudate lobe usually drains into the LHD. The quadrate lobe drains into the LHD.
It must be further appreciated that the hepatic veins do not follow the same segmental drainage but
pass in an oblique posterior direction to reach the inferior vena cava crossing these segments.
As might be anticipated, there is some variation in the segmental and subsegmental ducts. Aberrant
drainage of subsegmental ducts within the right lobe occurs in 20% of cases; the anterior and posterior
segmental ducts do not join to form an RHD in 28%; the posterior segmental duct joins the LHD in
22%; in 6%, the right anterior segmental duct joins the LHD. The medial and lateral segmental ducts of
the left lobe unite to form an LHD in 67% of cases; in 25%, the medial segmental duct drains into the
lateral inferior duct. Other variations are much less common (Figure 586). These and even more
minor variations, which are really of significance only in hepatic surgery, are described further in the
literature.10,11(5816)
(5817)Figure 586. Variations of the segmental intrahepatic ducts. A-C, Variations related to
the right lobe. D-J, Variations related to the left lobe. ASDanterior segmental duct;
PSDposterior segmental duct; RHDright hepatic duct; LHDleft hepatic duct;
CHDcommon hepatic duct; LSlateral segmental duct; MSmedial segmental duct;
LSAlateral superior area duct; LIAlateral inferior area duct. (From Linder RM, Cady B.
Hepatic resection. Surg Clin North Am 1980; 60:34967.)
The liver is suspended from the diaphragm and the anterior abdominal wall by ligaments. The anterior
falciform ligament divides, proceeding over the superior surface of the liver, into right and left coronary
ligaments. These are oriented to the left and right on the superior surface. The width of each coronary
ligament narrows as the anterior and posterior layers come closer together toward their farthest
attachment to the left and right, the terminal edges being called the left and right triangular ligaments.
The liver is also held by the inferior vena cava, which passes in a groove posteriorly. It is covered by
peritoneum, except on these ligaments and a small bare area in contact with the diaphragm to the right
of the inferior vena cava, which is skirted by the separated layers of the right coronary ligament.
Beneath the peritoneum lies a thin connective tissue covering, Glisson's capsule.
From the visceral surface, the lesser omentum extends downward from the porta hepatis and the fossa
for the ligmentum venosum to attachments to the lesser curvature of the stomach and the first part of
the duodenum. The bile ducts, portal vein, and hepatic arteries are found within the free edge of the
lesser omentum. They therefore lie anterior to the foramen of Winslow, which leads from the main
peritoneal cavity into the lesser sac.
The first three surfaces (see previously) of the liver are in contact with the diaphragm and the anterior
abdominal wall. The visceral surface on the left is related to the esophagus and the lesser curvature of
the stomach; to the right, it is in contact with the right kidney, the right adrenal gland, the hepatic
flexure of the colon, the pylorus, and the duodenum.
Hepatic Ducts
The RHD and the LHD join just outside the porta hepatis to form the common hepatic duct (CHD),
which has a diameter similar to that of the CBD. Although on average the CHD is 2.5 to 3.5 cm in
length, the union may be lower, and these have even been described as opening separately into the
duodenum.12(5818) This, combined with the variable site of entry of the cystic duct, determines the
lengths of the CBD and the CHD. The CHD passes down the free edge of the lesser omentum with the
hepatic artery, or with one of its two main branches, the artery lying to the left of the bile duct and the
portal vein lying behind. Although these relationships are regarded as relatively constant, a postmortem
CT study suggests that variations may occur in the relative positions of these three structures.13(5819)
The cystic artery usually crosses the CHD posteriorly, although in 25% of cases it passes anterior to
the duct. Also, in 15% of cases, the right hepatic artery lies anterior to the duct.1,12,14(5820)
Gallbladder
The gallbladder, with a capacity of about 50 ml, is a pear-shaped sac lying in a fossa on the right
visceral surface of the liver lateral to the quadrate lobe and separated from the caudate lobe by the
structures leading to the porta hepatis. It is covered by peritoneum to a variable extent and may even
be on a mesentery.15,16(5821)
The gallbladder is composed of columnar epithelium thrown into folds, giving it a honeycombed
appearance, outside of which lies a fibromuscular layer mainly longitudinal in direction. There are no
mucous glands, but the neck of the gallbladder contains goblet cells. The gallbladder consists of a
fundus that may or may not project beyond the liver to come in contact with the anterior abdominal wall
at the level of the ninth costal cartilage.17,18(5822) Posteriorly, the gallbladder is in contact with the
transverse colon. With the normal tilt of the liver, the body passes upward, backward, and to the left
and narrows to form the neck. A dilation here, called Hartmann's pouch, is a pathologic and not an
anatomic entity. The body is in contact with the liver superiorly, into which it may be embedded to a
variable extent, with the duodenum and transverse colon below; the neck narrows into the cystic duct.
Cystic Duct
The cystic duct, usually 3 to 4 cm long, runs backward and to the left in a series of S-shaped bends to
join the CHD, thereby forming the CBD. It contains the valves of Heister, which have a splinting effect.
Common Bile Duct

This duct continues downward and slightly to the left in the free edge of the lesser omentum, with the
hepatic artery proper on its left and the portal vein behind. It passes behind the first part of the
duodenum, where now the gastroduodenal artery lies to its left and the retroduodenal artery lies in
front. It then passes into a groove in the head of the pancreas and runs to the right in front of the
inferior vena cava. Its entry into the pancreas is recognized, radiologically, as the apex of a curve or
even a sharp angle, and here the CBD may narrow slightly. Behind the pancreas, it is approached by
the pancreatic duct, and the two ducts become enveloped in a common muscular and connective
tissue sheath and pass together through the choledochal window in the duodenal muscle to open on
the main duodenal papilla about 8 to 10 cm from the pylorus.
In some 15% of cases, there is no pancreatic tissue behind the CBD as it lies in the pancreatic groove.
In the remaining 85%, it is partly or wholly covered by pancreatic parenchyma. Even if the duct is
completely covered, there is a dissectible plane between the two organs.1921(5823) It may lie quite
close to the duodenum, or it may be as much as 2 mm away. In this region, it is usually crossed by the
superior pancreaticoduodenal vessels. Measurements of the diameter have been made radiologically,
on postmortem specimens, at intravenous or operative cholangiography, and by ultrasonography. In
comparing these measurements, magnification factors must be taken into account. Table 581 shows
some of the reported series.2228,30(5824) It is generally accepted that the upper normal limit is 10
mm, or at most 12 mm, measured radiologically. This measurement is taken conventionally at point of
entry of the duct into the pancreas (see Chapter 59: The Normal Retrograde Pancreatogram and
Cholangiogram).
Measurements of the Internal Diameter

of the Normal Common Bile Duct
TABLE 581
AUTHORS
DATE
DIAMETER*
REFERENCE
Partington and Sachs

1951
511
22
Poppel et al.
1956
46
23
Sullens and Sexton
1955
Up to 7
24
Cole and Harridge
1956
Up to 15
25
Wise and O'Brien
1956
315
26
Hutchinson and Blake
1957
Up to 8
27
Le Quesne et al.
1959
Up to 10
28
Mahour et al.
1967
412
30
From Hand B. Anatomy and function of the extrahepatic system. Br J
Hosp Med 1968; 1; 822.
*Diameter in millimeters on entry into the pancreas measured on x-ray
films.
Choledochoduodenal Area
Although the existence of a sphincter at the lower end of the CBD was postulated by Glisson in
1659,31(5825) serious study of the anatomy of the choledochoduodenal area actually began in the
19th century, when the existence of a sphincter mechanism at the lower end of the CBD appears to
have been established.32,33(5826)
Early in the 20th century, further studies were carried out, mainly to determine the cause of
pancreatitis.34,35(5827) Later, the microscopic appearance of this area was extensively investigated,
and the more important works have been reviewed and acknowledged.3638(5828) Unfortunately, the
initial researches tended to confuse the issue as to the extent of the musculature and whether the
sphincter muscle was a functional entity separate from the duodenal muscle.
Let us consider first the functional independence: In the early part of this century, several authorities
believed that the sphincter was only partly responsible for the control of the flow of bile.3941(5829) A
little later it was considered to be fully independent of the duodenal muscle.4244(5830) In more
recent times, with the advent of the image intensifier, it was concluded that the sphincter is fully
independent.4548(5831) Subsequently, this was confirmed in electromyographic studies.49(5832)
Consider next the anatomic studies: Comprehensive study of fetal sections established the existence
of a sphincter that was anatomically separate from the duodenal wall and involved both the CBD and
the pancreatic duct. Initially, it was thought that the sphincter around the pancreatic duct was
incomplete and that the common channel was not encompassed by the sphincter.50,51(5833) Later
studies on adult tissue, however, demonstrated that the sphincter encircles the pancreatic duct and the
common channel.52(5834)
It was not appreciated until the 1950s, when cholangiography was being more widely practiced, that
radiologically the CBD becomes abruptly narrowed as it approaches the
duodenum.22,23,36,5357(5835)
Even with these extensive investigations, no clear picture emerged of the relationship between the
macroscopic, microscopic, and radiologic appearances of the choledochoduodenal area. A study of
autopsy material using cholangiography and pancreatography, followed by the preparation of ductal
casts and subsequent dissection of the specimen, made it possible to correlate the configuration of the
cast with the gross anatomy and radiologic appearance.36,37(5836) By microscopic study of these
specimens, it was possible to correlate the sphincter mechanism with the radiologic appearance.
This work demonstrated that the terminal narrowing of the CBD, not detectable on its external surface,
was due to the marked increase of muscle fibers in the wall. This produced an abrupt reduction in the
width of the lumen that correlated with a radiologically recognizable notch. The notch was found to lie
some 2 mm proximal to the duodenal wall and corresponded with the upper edge of the superior
choledochal sphincter. The muscular narrow-lumened portion of the common duct, referred to as the
thickened segment, varies in length from 11 to 27 mm, the average being 16 mm. The length of the
intraduodenal part of the thickened segment is thus 2 mm less, that is, 9 to 25 mm (Figure 587).
Using corrosion cast, radiographic, and histologic techniques in 64 postmortem specimens, Flati et
al.58(5837) found the superior limit of the choledochal sphincter to be, respectively, 14.48 4.85 mm
(range 9 to 27 mm), 13.44 4.01 mm (range 8 to 28 mm), and 13.6 0.4 mm (range 13 to 25 mm).
(5838)Figure 587. A diaphragmatic representation of the common bile duct shows a notch
that divides it into two distinct parts: an upper, wide-lumened, thin-walled portion and a lower,
narrow-lumened, thick-walled portion. The lower portion is seen to lie mainly in the
submucosal layer of the duodenum. (From Hand BH. An anatomical study of the
choledochoduodenal junction. Br J Surg 1963; 50:48694. By permission of Blackwell
Science Ltd.)
Based on the findings described earlier, it was suggested that the previous subdivision of the common
duct into four portionssupraduodenal, retroduodenal, retropancreatic, and intraduodenalshould be
abandoned in favor of only two portions: namely, an upper, longer, wide-lumened, thin-walled portion
and a lower, shorter, narrow-lumened, thick-walled portion demarcated by a notch. This division relates
not only to the radiologic appearance but also to the microscopic appearance and function. The upper
part is merely a conduit composed of a mucosa and a lot of elastic tissue and connective tissue, but
very little muscle, the latter having a predominantly longitudinal orientation. In this section, numerous
mucous glands can be seen opening as depressed pores. The second, lower part of the CBD is the
sphincteric mechanism.
The pancreatic duct, lying posteriorly or posteromedially, approaches the CBD just above the level of
the notch. Here the two ducts become surrounded by a common muscular sheath but remain
separated by a muscular septum.37,59,60(5839) This muscular sheath is predominantly circular in
direction and surrounds both the CBD and the pancreatic duct separately and also passes across the
septum between them in a figure-of-8 fashion. This muscle extends to the apex of the papilla and
therefore envelops the common channel when present or surrounds both ducts individually if there are
separate openings. An outer layer of the longitudinal muscle is present and thought to produce
shortening and lengthening and thus erection of the papilla.

In the submucous layer of the duodenum, the two ducts may fuse by a gradual thinning and breakdown
of the septum to form a common channel or pancreaticobiliary canal (Figure 588). Alternatively, the
two ducts may open separately on the major papilla (Figure 589). A great deal of work has been done
on the frequency of these two modes of entry and the length of the common channel.61(5840) Based
on a comprehensive review of the literature involving some 3000 specimens, it would appear that some
85% have a common channel and 13% have separate openings; in 2%, the duct of Wirsung is not
patent and all the pancreatic drainage passes via the duct of Santorini.36,37(5841) Traditionally, the
common channel is referred to as the ampulla of Vater,62(5842) but studies have failed to demonstrate
any consistent dilation in this region.29,36,37,58,60(5843) The length of the common channel has
been reported as measuring between 1 and 17 mm, the majority being between 2 and 7 mm.37(5844)
(5845)Figure 588. Autopsy cast and cholangiogram show a well-marked notch on the
common bile duct. The narrow-lumened portion, called the thickened segment, and the
common channel are well visualized. The pancreatic duct has a similar appearance. (From
Hand BH. An anatomical study of the choledochoduodenal junction. Br J Surg 1963;
50:48694. By permission of Blackwell Science Ltd.)
(5846)Figure 589. Autopsy cast and cholangiogram of a specimen with separate openings of
the bile and pancreatic ducts into the duodenal lumen. The notch of the common bile duct is
not prominent but that of the pancreatic duct is well defined. (From Hand BH. An anatomical
study of the choledochoduodenal junction. Br J Surg 1963; 50:48694. By permission of
Blackwell Science Ltd.)
The orifice of the major papilla is round or slitlike and between 2 and 5 mm in diameter. Sometimes
minute fronds of mucosa prolapse through the orifice. These are called valvules, and they arise from
the common channel mucosa. They contain a central core of muscle that permits lengthening and
shortening.63(5847)
Pancreas
The pancreas, which is 12 to 25 cm in length and weighs 70 to 100 g, depending on gender, lies
retroperitoneally on the posterior abdominal wall and possesses an ill-defined connective tissue
capsule. It is prismatic in cross section and therefore has three borderssuperior, inferior, and
anterior.
The pancreas is divided anatomically into head, neck, body, and tail. Its form is hooklike, the hook
being the head and body and the handle being the tail. It is pale yellow and has a slightly lobulated
surface. In the past, the pancreas was regarded as a fixed structure, but evidence from CT indicates
that some movement occurs during respiration, with the tail being the more mobile segment.
Anatomically, the pancreas passes from right to left, usually in a cephalad direction. The shape of the
gland has been described as oblique (36% of cases), sigmoid (19%), transverse (3%), horseshoe
(8%), an L (33%), and inverted (1%).64,65(5848) Where it crosses the spine and aorta, there is a
forward bulge, and it is here that the pancreas is liable to injury (see Chapter 75: Pancreatic Trauma,
Ascites, Fistula, and Pseudocyst). Its relationship to the spine is variable. Whereas the head may be as
low as S1 and the tail as high as T10, usually the whole gland lies between L1 and T12.
The mean thickness of the pancreas is 2.3 cm in the head, 1.9 cm in the neck, 2 cm in the body, and
1.5 cm in the tail. The mean width is 4.4 cm, 3.4 cm, 3.5 cm, and 3 cm, respectively, at these anatomic
divisions. Thus, the vertical span is twice that of the anteroposterior dimension. The head lies within
the duodenal loop and is overlapped by it superiorly but overlaps it elsewhere. In 98% of patients, it lies
between T12 and L2.66(5849) It lies in front of the CBD, the medial edge of the right kidney and its
vessels, the inferior vena cava, the right gonadal vein, the end of the left renal vein, and the right crus
of the diaphragm. The hepatic artery lies at the upper border of the head, whereas the transverse
colon, part of the greater omentum, and the root of the transverse mesocolon lie anteriorly. In the lower
part of the head lies the uncinate process, which extends to the left behind the superior mesenteric
vein, separating it from the aorta behind.
The neck, which is more triangular than the head and is slightly constricted, lies approximately level
with the pylorus and the gastroduodenal artery. It is in front of the superior mesenteric vein, the
beginning of the portal vein, and the cisterna chyli. The celiac artery hooks over its top edge. The body
passes obliquely upward, slightly backward, and to the left across the spine at levels that vary between
T12 and L2, in front of the superior mesenteric artery, the aorta, the left suprarenal gland and the left
kidney and its vessels, and finally the left crus. The splenic vein runs behind its upper border and may
be partly embedded in its substance; the splenic artery lies above, running a tortuous course. The
inferior mesenteric vein passes posteriorly to join the splenic vein. The transverse mesocolon is
attached to its inferior edge. Anteriorly, it forms the stomach bed, separated by the lesser sac. The left
end of the body is related to the duodenojejunal junction and the splenic flexure of the colon.
The tail moves away from the posterior abdominal wall to enter the lienorenal ligament with its splenic
vessels and usually reaches the hilum of the spleen. It normally lies at the level of T12 or L1, but it may
be as low as L2 and as high as T11.
Pancreatic Ducts
The main pancreatic duct begins at the tail and runs in an undulant fashion through the gland nearer to
its upper edge and posterior surface and receives its branches at right angles in a herringbone pattern.
At the point of junction of the ducts of Santorini and Wirsung, there is frequently a slight constriction or
a loop segment. Occasionally, the main duct becomes bifurcate in the body or begins as a confluence
of several small ducts in the tail. The length of the pancreatic duct varies between 16.4 and 24.2
cm.66(5850) As it approaches the duodenum, it turns sharply downward and slightly backward and to
the right. This segment may have a slightly wider diameter relative to the other segments of the duct
and may appear somewhat dilated. In the head, the herringbone pattern of main duct branches is not
so obvious, although several branches are found. In particular, it receives a branch from the uncinate
lobe. Close to the duodenal wall, it comes to lie alongside the CBD, and just outside the muscular wall
of the duodenum, it becomes enveloped with it in a common muscle sheath. A notch identical to that
occurring on the CBD can also be seen at the same level on the pancreatic duct, so that the most
distal segment of the pancreatic duct is markedly narrowed relative to the diameter of the adjacent
segment of the duct (see Figure 588).37(5851) The main pancreatic duct, or duct of Wirsung, drains
directly into the duodenum in 13% of cases. In 2%, the duct is not patent, whereas in the remaining
85%, it forms a common channel with the CBD before duodenal entry. The findings in a study of 64
autopsy specimens using silicone corrosion cast and radiographic techniques by Flati et al.58(5852)
were similar. Analysis of contrast radiographs of the ducts demonstrated a common channel in 87.6%
of cases. However, there was considerable variation in the length of the channel. In 34 specimens
(69.3%), the length was in the range of 1.5 to 10 mm, but in 3 instances (6.1%), it was more than 10
mm, and in 6 specimens (12.2%), it was less than 1.5 mm. Radiographs of the postmortem specimens
demonstrated separate openings for the CBD and pancreatic duct on the same papilla in 10.2% of
specimens; separate openings for the two ducts were discovered in 1 specimen (2%). Flati et
al.58(5853) found somewhat different results when the ductal arrangement at the main duodenal
papilla was studied using corrosion casts. A common channel was found in 79.5% of specimens, but it
was in general shorter. Also, the percentage of specimens with separate openings at the same papilla
more than doubled, to 22.4%. These findings suggest that a certain degree of variability is inherent in
stated anatomic measurements owing to differences in methods. When the results obtained by the
corrosion cast method were compared with radiographs, Flati et al.58(5854) found agreement in only
60% of cases. These investigators felt that the radiographic method provided inaccurate data in the
remaining 40%.
The course of the main pancreatic duct has been reported at ERCP as ascending in 48.5% of cases,
horizontal in 26.5%, sigmoid in 16.2%, and descending in 8.8%.67(5855)
The duct of Santorini lies anterior and cranial to the main duct and drains the anterosuperior part of the
gland. It passes anterior to the CBD and enters the duodenum on the minor papilla. The duct usually
widens proximal to the orifice (Figure 5810).
(5856)Figure 5810. Autopsy cast and pancreatogram of a specimen in which the duct of
Wirsung could not be found. The duct of Santorini can be seen to drain into the duodenum.
There is marked dilation before entry. The minor papilla is outlined by a halo of contrast. The
common bile duct is well notched, but the thickened segment is short. (From Hand BH. An
anatomical study of the choledochoduodenal junction. Br J Surg 1963; 50:48694. By
permission of Blackwell Science Ltd.)
The normal range of diameter of the main pancreatic duct is 3 to 7 mm in the head, the usual diameter
being between 4 and 5 mm. Narrowing occurs in its terminal portion, which is about 2 mm in diameter.
The diameter in the body ranges from 2.5 to 5 mm, with most ducts being between 3 and 3.5 mm, and
from 1.5 to 3 mm in the tail, with 2 to 2.5 mm being the usual dimension. Maximum normal diameters
for the head, body, and tail have been quoted as high as 6.5 mm, 5 mm, and 3 mm,
respectively,68(5857) but higher measurements have been reported.69(5858) There is evidence that,
with increasing age, the diameter of the duct increases and the level of the papilla moves lower in the
body.64(5859) This was not confirmed in similar studies of the site of papilla, the course and the
caliber of duct, and the like.70(5860) The normal retrograde pancreatogram and cholangiogram are
discussed in Chapter 59: The Normal Retrograde Pancreatogram and Cholangiogram.
Papillae
The major duodenal papilla is a smooth elevation that almost always lies in the second part of the
duodenum on the posterior or posteromedial wall. A transverse mucosal fold may obscure the main
papilla, and a characteristic longitudinal fold extends distally from its lower edgethe plica
longitudinalis. The major papilla usually lies about 8 cm from the pylorus, but some investigators have
found it can be as close as 5 cm or as far away as 14 cm, which means that it can be located at any
point from the junction of the first and second parts of the duodenum to within the third
part.17,23,71(5861) Schwartz and Birnbaum72(5862) found 74% in the second part of the duodenum,
18% at the junction of the second and third parts, and 8% in the third part. Flati et al.58(5863) found
the main papilla to be within the second portion in all 64 postmortem specimens studied; distance from
the pylorus ranged from 7 to 13 cm, with a mean of 10.56 1.3 cm. Thus, both vertical and horizontal
levels of the papilla in relation to the spine are variable, the structure lying between L1 and S2.
However, 90% lie between L2 and L3 in a vertical direction; 49% are to the right of the spine; 46% are
over the spine; and 5% are to the left side of the spine in the horizontal plane.64(5864)
The major papilla is difficult to visualize endoscopically when located in or at the margin of a
diverticulum. Its external appearance has been described as flat, papillary, or hemispherical.73(5865)
In the autopsy study of Flati et al.,58(5866) the papilla was hemispherical in 14.8%, cylindrical or
conical in 77.8%, and flat in 8.2% of postmortem specimens. Whereas cylindrical or conical would
appear to be the most common form, other shapes namely uniformed, swollen, villous, nipple-shaped,
and sharply pointed, have been described.74(5867) There appears to be considerable variation in size.
The average papilla is about 1 cm in diameter, but a diameter as large as 3 cm has been
reported.23(5868)
The minor papilla lies either ventral to or level with the major papilla in a transverse plane and either
proximal to or level with it in a vertical plane, but it is never found dorsal or distal to the major papilla. It
lies on a transverse mucosal fold that may be bifurcate. Its distance from the major papilla is usually
1.0 to 3.5 cm, its size variable, and in dissected specimens, it is nearly always identifiable. It may be as
large as or even larger than the major papilla, but surprisingly its size bears no direct relationship to
patency. The minor papilla may be too small to be identified endoscopically, or even if it is large
enough to be identified, its orifice may be too narrow to permit cannulation. Details of its incidence and
microscopic structure were reported long ago.21(5869)
In microscopic anatomic studies, minute accessory pancreatic ducts have been found entering the
CBD in its intrapancreatic and intraduodenal portions and also entering the duodenum
directly.75(5870)
Anomalies and Congenital Abnormalities

Liver
Biliary atresia, as the name suggests, was originally thought to be due to a failure of canalization
involving variable amounts of the intrahepatic and extrahepatic duct systems. This is not the case,
however.76(5871) Although its cause is still unknown, it is now considered an infective or ischemic
condition and not truly a congenital abnormality.77(5872) Therefore, it will not be considered in this
chapter (see Chapter 79: Endoscopic Retrograde Cholangiopancreatography in Infants and Children).
Cystic disease affecting the liver is not rare.78,79(5873) In the polycystic form, it appears to be
associated with renal cysts in 50% of cases. Solitary cysts also occur. None of these mucus-containing
cysts communicates with the bile ducts. Rarely, cystic disease is associated with hepatic
fibrosis.80(5874) A rare diffuse dilation of the intrahepatic ducts has been reported.81(5875)
Intrahepatic choledochal cystic disease is described subsequently.
Right, Left, and Common Hepatic Ducts

It has been stated that the RHD and the LHD may fuse within the liver. This has rightly been denied,
and it is now established that fusion takes place 7 to 15 mm below the hilum.10,18(5876) The LHD
follows a more oblique course and is therefore longer. In some cases, major tributaries of the RHD, but
not the LHD, may also be outside the liver, specifically the right anterior and right posterior segmental
ducts. In other cases, the right anterior and right posterior segmental ducts may join the LHD as a
confluence. Less commonly, the right posterior segmental duct joins the LHD. In these last two
examples, no true RHD exists (Figure 5811). Rarely, the RHD and the LHD fuse far down in the free
edge of the lesser omentum, in which case, the cystic duct is more liable to join the RHD. Naturally,
when the cystic duct joins a confluence, as it does in about 2% of cases, there is no CHD. These
variations have been well documented.29,82,83(5877) The RHD has been reported as draining directly
into the gallbladder, directly into the cystic duct, and directly into the duodenum.84,85(5878) Rarely,
both hepatic ducts may drain into the gallbladder, a condition sometimes referred to as gallbladder
interposition.86,87(5879) Drainage of a segment or subsegment, a sector, or rarely, the entire right
lobe of the liver into the cystic duct or gallbladder occurs in about 1% of individuals.88(5880)
Depending on the extent of liver drained by these aberrant ducts, failure to recognize their presence at
cholecystectomy may have significant adverse consequences.
(5881)Figure 5811. Modes of junction of the hepatic ducts. The upper drawings show the
findings in 75% of cases in which a true right hepatic duct is formed. In the lower drawings,
no true right hepatic duct is present, either because the three ducts join as a confluence (left)
or because the right posterior segmental duct joins the left hepatic duct (right). RHDright
hepatic duct; LHDleft hepatic duct; RASDright anterior segmental duct; RPSDright
posterior segmental duct.
With the recognition of the segmental structure of the liver, the term accessory bile duct is no longer
tenable. These ducts are not additional in any sense, but merely ducts draining a specific part of the
liver that run an aberrant extrahepatic course. A wide range of incidence has been reported, with this
aberrant finding occurring in up to 28% of cases. Differences in incidence are related to the method of
demonstration. Many of these ducts are quite small, they may possibly result from infection, and they
are of no clinical significance. They have been reported as draining into the RHD, CHD, cystic duct,
gallbladder, CBD, and main pancreatic duct.89(5882) Some reported series are listed in Table
582.1,12,17,29,9092(5883) Few of these would be demonstrable radiologically or at operation. The
most important variation, from an operative point of view, is an aberrant right segmental hepatic duct
that joins the CHD low down in 16% of cases.1,2,4,5,12,17,18,29,9092(5884)
TABLE 582
AUTHORS
DATE
Incidence of Mode of Entry of So-Called Accessory Hepatic Ducts, in Some Serie

METHOD OF
DETECTION
NUMBER OF
SPECIMENS
ACCESSORY
DUCTS
RHD
CHD
CD/CHD
Flint
1923
Dissection
200
15
9
9
10
Lurje
1937
Dissection
194
11
10
4
Michaels
1951
Dissection
200
18
7
20
Mooseman
1951
Dissection
250
16
5
15
5
and Coller
Hayes et al.
1958
Operation
400
18
Dowdy et al.
1962
Dissection
100
15
12
Grant
1962
Dissection
95
7
4
From Hand B. Anatomy and function of the extrahepatic system. Br J Hosp Med 1968: 1:822.
RHDright hepatic duct; CHDcommon hepatic duct; CD/CHDjunction of cystic duct and common hepatic duct; CD
CBDcommon bile duct; Anastanastomosing or double ducts.
Gallbladder
Unlike in other higher mammals, especially the cat, congenital abnormalities of the gallbladder are rare
in humans.93,94(5885) However, phrygian cap deformity (folded fundus) is found in 18% of
cholecystograms.95(5886) Diverticulum of the gallbladder, which may occur anywhere on the wall, is
rare96(5887) and must be distinguished together with hourglass constriction from an inflammatory
condition known as cholecystitis glandularis proliferans.97(5888)
Bilobed gallbladder and a multiseptate form have been reported and usually look normal on the
external surface.98(5889) Very rarely, the gallbladder may be rudimentary or absent without being part
of congenital biliary atresia. Absence is usually associated with other congenital abnormalities, and
over 150 cases have been reported.99104(5890)
Double gallbladder may be expected in 1 in 3000 to 4000 cases.105(5891) It may be associated with
two separately opening cystic ducts, one of which may join the RHD, or the two ducts may join in a Y
fashion before entry (Figure 5812). There is some question as to whether this condition carries an
increased incidence of gallstone formation.106108(5892)
(5893)Figure 5812. Congenital variations of the gallbladder. 1, Phrygian cap deformity; 2,

diverticulum; 3, hourglass; 4, septum; 5, double gallbladder with single cystic duct; 6, double
gallbladder with two cystic ducts draining into common bile duct; 7, double gallbladder with
two cystic ducts, one of which drains into the right hepatic duct.
The gallbladder may be ectopic in position. It can lie intrahepatically, a condition that must be
distinguished from absence, or it may lie under the left lobe of the liver without transposition, where its
cystic duct may drain into the LHD. It has been recorded as being accompanied by a normal
right-sided gallbladder. The gallbladder has also been found retrodisplaced onto the inferior and
posterior part of the liver and also lying transversely, in which case both hepatic ducts may drain into it.
Cystic Duct
The cystic duct is extremely variable in length and course, as well as in its site and mode of
termination. It has been reported as absent,109(5894) but it is likely that this is most often the result of
pathologic processes. Although it normally joins the CBD, it may join the RHD, an aberrant right
segmental duct, the confluence of the RHD and the LHD, or even the LHD. Its length depends on its
site of junction with the CHD, which may be high and close to the origin of the CHD or low and close to
the entry of the CBD into the duodenum.
Its mode of entry has been described as angular, long, short, parallel, and spiral. In the latter two types,
the cystic duct may be separated from the CBD by loose connective tissue, or they may be fused in a
nondissectible plane. In these two types, the site of entry will be nearer to the duodenal terminus of the
bile duct, lying retroduodenally or even within the pancreas, in which case, it may be as long as 6 cm.
In spiral entry, the cystic duct usually passes posterior to the CHD, but it may pass anteriorly, and a
double spiral has been described.18,29,90,110(5895) These variations and their approximate
incidence, based on a review of the literature, are depicted schematically in Figure 5813.
(5896)Figure 5813. Modes of entry of the cystic duct and the approximate percentage of
incidence obtained from a review of the literature. POSTposterior; ANTanterior.
Common Bile Duct

Anomalies of the CBD are rare. Both absence111(5897) and duplication112,113(5898) are reported.
Drainage into the stomach and double entry into the duodenum are described, but there is some
suspicion that these were the result of pathologic processes.114(5899)
Choledochal cyst, which is uncommon although encountered more often in Asia, takes various forms
(see also Chapter 63: Cystic Disorders of the Bile Ducts). The widely accepted classification115(5900)
into Types 1, 2, and 3 is shown in Figure 5814. The capacity of a choledochal cyst varies from a few
milliliters to 8 L. When intrahepatic cystic dilations were recognized, Types 4 and 5, also shown in
Figure 5814, were added to the classification.116118(5901) All of these cysts are part of, or
communi-cate with, the bile ducts and are not associated with polycystic disease, nor are they visible
on the surface of the liver.
(5902)Figure 5814. Types of choledochal cysts.

The intrahepatic variety has been divided into two groups:1(5903) a simple type with a tendency to
stone formation and subsequent sepsis, now known as Caroli's disease,119(5904) and a type
associated with periportal fibrosis that develops cirrhosis and portal hypertension.2(5905)
Pancreas
Anomalies of the pancreas are also discussed in Chapter 72: Congenital Anomalies of the Pancreas,
with emphasis on endoscopic, radiologic, and clinical aspects.
Arrested development resulting in absence of the pancreas, failure of formation of either dorsal or
ventral buds, and failure of the two buds to fuse are exceedingly rare and are associated with other
major congenital abnormalities that are usually incompatible with life. Pancreatic hypoplasia is known
to occur. This familial condition, known as the Shwachman-Diamond syndrome, is associated with
hematologic disorders. Inherited cystic fibrosis affecting the pancreas and lungs is a functional and not
an anatomic abnormality. True cysts of the pancreas are relatively common. They are usually
unilocular and associated with cystic disease of the liver and kidneys. They contain serous fluid and
are not connected with the pancreatic duct system.
A bifid tail of the pancreas, each limb having its own duct, has been described, as has suspension of
the pancreas on a short mesentery from the posterior abdominal wall.
Ectopic Pancreas
Ectopic pancreatic tissue is readily identified in postmortem studies. When a specific search was
made, an incidence of over 13% was recorded. These islands may be as small as 1 mm or as large as
5 cm in diameter. They may be multiple and may be found almost anywhere in the abdomen. About
80% are located in the stomach or duodenum, but they have also been found in the spleen, in the
gallbladder and bile ducts, and in the midgut loop to its apex in a Meckel diverticulum.120(5906) It is
not clear whether these islands arise from displaced bud tissue (which would seem likely because a
high proportion are close to the area where the buds develop) or whether the gut epithelium
spontaneously produces them heterotopically.
Annular Pancreas
Annular pancreas is uncommon, and its incidence is difficult to assess. One authority reports 3 cases
in 7000 autopsies.121(5907) Another found only 1 adult case in 20,000 autopsies.10(5908) It may
present soon after birth or later in life or may be found incidentally at autopsy.
Early presentation usually occurs within the first few days of life and appears to be associated with
duodenal stenosis or atresia in over 40% of cases.122(5909) It is frequently associated with other
congenital abnormalities such as esophageal atresia, imperforate anus, congenital heart disease,
malrotation of the gut, and Down's syndrome. Annular pancreas causes no more than 1% of neonatal
intestinal obstructions and less than 5% of duodenal obstructions.123,124(5910) The site of the
annular pancreas in 85% of cases is in the second part of the duodenum. It has a variable relationship
with the major papilla but is usually proximal to it, even though bile can frequently be aspirated. Its site
has also been reported in the first and third parts of the duodenum. Radiologically, the dilated
duodenum proximal to the narrowing and the duodenal bulb produce a "double-bubble" appearance.
The pancreatic tissue is usually embedded in the duodenal muscle and has its own duct.125(5911)
These two facts render local excision or division highly undesirable.
When annular pancreas presents later in life, the symptoms are similar to those of peptic ulcer, and
patients usually have duodenal obstruction. Peptic ulcer has been documented in a third of the cases.
Acute pancreatitis and jaundice have also been reported.126130(5912) Duodenal obstruction
produced by kinking due to general visceroptosis has been suggested as the explanation for the
presentation in later life.
It has been postulated that annular pancreas is derived embryologically from persistence of the left part
of the ventral bud or, alternatively, that it is due to fixation of a portion of the right part of the ventral bud
to the duodenal wall before rotation occurs. It has also been suggested that it is merely a variety of
ectopic pancreas.131(5913)
Pancreas Divisum
As a result of increasing use of endoscopic pancreatography, greater attention has been given to the
anatomy of the pancreatic duct. Although referred to earlier in this chapter, pancreas divisum requires
elaboration in the light of reports that patients with pancreatitis appear to have a higher incidence of
this anomaly than the general population (see also Chapter 72: Congenital Anomalies of the
Pancreas).132(5914)
There is still some confusion concerning nomenclature. Some authors refer to the ducts as those of
Wirsung and Santorini, whereas others refer to them as the main and accessory pancreatic ducts,
respectively. If the main drainage occurs via the duct of Santorini and the minor papilla, clearly it is
illogical to refer to the former as an accessory duct. It is better, therefore, to use classic terminology
when referring to the ducts, particularly in relation to their embryologic development.
The dorsal pancreasthat is, the tail, body, neck, and the upper anterior part of the headrelates to
the duct of Santorini. The rest of the head relates to the duct of Wirsung. In humans, these two ducts
most often fuse in the head, and variable degrees of drainage take place through both ducts, although
most often the greater part occurs via the duct of Wirsung. In some cases in which fusion of the ducts
has occurred, the duct of Santorini ends blindly in the duodenal wall.58(5915) Occasionally, fusion of
the two ducts does not occur. Least common of all is obliteration of the duct of Wirsung in its terminal
portion. In a radiographic postmortem study of 330 human pancreases, Sigfusson et al.133(5916)
found the latter anomaly in 8 cases (2.4%). There are, therefore, five possible pancreatic drainage
patterns (Figure 5815).
(5917)Figure 5815. Pancreatic duct drainage patterns.

The earlier studies of pancreatic anatomy, which were based on autopsy material, were attempts to
explain the occurrence of acute pancreatitis when a stone became impacted in the distal CBD and
obstructed the duct of Wirsung.34,35(5918) Notable among these studies was one in which the orifice
of the major papilla was obstructed by a fine suture in 200 autopsy specimens and pancreatograms
were performed.134,135(5919) The duct of Wirsung was shown to be the sole or main pancreatic
outlet in 91% of cases. The duct of Santorini was found in 50% of cases but was the sole or main
outlet in only 9% of cases, usually without communication with the duct of Wirsung. In 2.5% of cases,
the duct of Santorini was the sole outlet because the terminal part of the duct of Wirsung was
obliterated. From these figures, it was calculated that the duct of Santorini could substitute for the duct
of Wirsung in 21% and relieve the obstruction in 12%. In 10% of cases, it could not relieve the
obstruction at all, and it ended blindly in 8%.
In another investigation, injection of air under water, dye injections, and dissections showed that the
duct of Santorini was always present, communicated with the duct of Wirsung in 89% of cases, and
drained into the duodenum in 73% of cases.136(5920) Although this was not actually mentioned,
presumably the duct of Santorini merely drained part of the head and drained separately into the
duodenum in 11% of specimens. In 4%, the duct of Santorini was larger than the duct of Wirsung.
In a study using bismuth or barium gelatin injections, it was found that the duct of Santorini was
present in 68%, but ended blindly in 13%, whereas in 4% of cases, it proved to be the main drainage
route of the pancreas.69(5921)
Another study using dye injections and pancreatograms showed that fusion between the ducts of
Santorini and Wirsung occurred in 57.5% of cases and drainage occurred via both papillae. In 35%,
the duct of Santorini was blind, and in 7.5%, the ducts did not fuse.71(5922) It was concluded that in
over 50% of cases, the duct of Santorini could decompress the duct of Wirsung. These results are in
contrast to those of another autopsy study that found that in many cases, perhaps the majority, the
duct of Santorini does not drain into the duodenum or it may not be present.58(5923)
Because these autopsy studies were carried out by different techniques and the statistics were
reported in different ways, unfortunately it is not possible to form an accurate idea of the relative
incidence of the five patterns of drainage shown in Figure 5815. Indeed, most authors do not
specifically mention Type 5, although by subtraction of their figures, it must be presumed to exist. The
percentages shown in Figure 5815 are an attempt to rationalize the figures quoted here, together with
others found in the literature.
In a review of the literature on the incidence of pancreas divisum, an analysis of the findings in 750
autopsy specimens gave figures ranging from 4.7 to 14%, whereas the incidence was reported as
being between 0.3 and 5.8% in over 9000 ERCP studies.132(5924) This latter work, of course, was
carried out in patients with symptoms who were thought to have a normal pancreas. No ERCP studies
have been carried out on normal volunteers. Earlier autopsy work was not designed to study pancreas
divisum specifically, but it would not be difficult to carry out further duct studies with this in view.
Acknowledgments
I am deeply indebted to Mrs. Janice Henney, M.A., A.L.A., and the staff of the Ipswich Medical Library;
to Mr. Philip Dove and the staff of the Medical Illustration Department of the Ipswich Hospital; and to
Mrs. Diane Robertson, who typed the script.
REFERENCES
1.
Michaels NA. The hepatic, cystic and retroduodenal arteries and their relations to the biliary
ducts. Ann Surg 1951;133:50324.
2. Benson EA, Page RE. A practical reappraisal of the anatomy of the extrahepatic bile ducts
and arteries. Br J Surg 1976;63:85360.
3.
Hjortsjo CH. Topography of the intrahepatic duct systems. Acta Anat (Basel)
1951;1415(Suppl 11):699715.
4.
Healey JE, Schroy PC. Anatomy of biliary ducts within the human liver: Analysis of prevailing
pattern of branchings and major variants of biliary ducts. AMA Arch Surg 1953;66:599616.
5.
Hobsley M. Intrahepatic anatomy: A surgical evaluation. Br J Surg 1958;45:63544.
6.
Goldsmith NA, Woodburne RT. Surgical anatomy pertaining to liver resection. Surg Gynecol
Obstet 1957;105:3108.
7. Gupta CD, Mittal VK, Gupta SC. Intrahepatic pattern of biliary ducts and their major variations.
Indian J Med Res 1975;63:11307.
8.
Couinaud C. Etudes des voies biliares intrahepatiques. J Chir (Paris) 1954;70:31028.
9.
Chevrell JP, Duchene P, Salama A, et al. Anatomical bases of intrahepatic biliodigestive
anastomoses. Anat Clin 1980;2:15967.
10. Linder RM, Cady B. Hepatic resection. Surg Clin North Am 1980;60:34967.
11. Gazelle GS, Lee MJ, Mueller PR. Cholangiographic segmental anatomy of the liver.
Radiographics 1994;14:100513.
12.
Flint ER. Abnormalities of the right hepatic, cystic, and gastroduodenal arteries and of bile
ducts. Br J Surg 1923;10:50919.

13. Bret PM, de Stempel JV, Atri M, et al. Intrahepatic bile duct and portal vein anatomy revisited.
Radiology 1988;169:4057.
14.
Daseler EH, Anson BJ, Hambley WC, et al. Cystic artery and constituents of the hepatic
pedicle. Study of 500 specimens. Surg Gynecol Obstet 1947;85:4763.
15.
Carter R, Thompson RJ, Brenan LP. Volvulus of the gallbladder. Surg Gynecol Obstet
1963;116:1058.
16.
Ashby BS. Acute and recurrent torsion of the gallbladder. Br J Surg 1965;52:1824.
17.
Lurje A. Topography of extrahepatic biliary passages with reference to dangers of surgical
technique. Ann Surg 1937;105:1618.
18.
Johnson EV, Anson BJ. Variations in the formation of vascular relationship of the bile ducts.
19.
Lytle WJ. The common bile duct groove in the pancreas. Br J Surg 1959;47:20912.
20.
Smanio T. Varying relations of common bile duct with posterior face of pancreatic tissue in
negroes and white persons. J Int Coll Surg 1954;22:15073.
21.
Baldwin WM. The pancreatic ducts in man together with a study of the microscopic structure
of the minor duodenal papilla. Anat Rec 1911;5:197228.
22.
Partington PF, Sachs MD. Monographs in Surgery. New York: Nelson. 1951.
23.
Poppel MH, Jacobson HG, Smith RW. Roentgen aspects of papilla of Vater. Am J Dig Dis
1956;1:4958.
24.
Sullens WE, Sexton GA. Indications for use of operative cholangiography. Ann Surg
1955;141:499503.
25.
Cole WH, Harridge WH. Diagnostic use of cholangiography in biliary tract disease. Surg Clin
North Am 1956;36:14959.
26.
Wise RE, O'Brien RG. Interpretation of the intravenous cholangiogram. JAMA
1956;160:81927.
27.
Hutchinson WB, Blake T. Operative cholangiography. Surgery 1957;41:60512.
28.
Le Quesne LP, Whiteside CG, Hand BH. The common bile duct after cholecystectomy. BMJ
1959;1:32932.
29.
Dowdy GS, Waldron GW, Brown WG. Surgical anatomy of the pancreatico-biliary system
observations. Arch Surg 1962;84:22946.
30. Mahour GH, Wakin KG, Ferris DO. Common bile duct in man, its diameter and
circumference. Ann Surg 1967;165:4159.
31.
Glisson F. Anatomia hepatis. London: Typis Du-Gardianis, impensis Octavian Pullein. 1659.
32.
Oddi R. D'une disposition sphincter speciale de l'ouverture du canal cholidoque. Arch Ital
Biol 1887;8:31722.
33.
Hendrickson WF. A study of the musculature of the entire extrahepatic biliary system
including that of the duodenal portion of the common bile duct and sphincter. John Hopkins
Hosp Bull 1898;9:22132.
34.
Opie EL. The aetiology of acute haemorrhagic pancreatitis. Johns Hopkins Hosp Bull
1901;12:1828.
35.
Opie EL. The anatomy of the pancreas. Johns Hopkins Hosp Bull 1903;14:22932.
36.
Hand BH. M.S. thesis. University of London. 1958.
37.
Hand BH. An anatomical study of the choledochoduodenal junction. Br J Surg
1963;50:48694.
38.
Hand BH. Anatomy and function of the extrahepatic biliary system. Rev Gastroenterol
1973;2:329.
39.
Buget GE. The regulation of the flow of bile. Am J Physiol 1925;74:5839.
40.
Elman R, McMaster PD. The physiological variations in resistance to bile flow to the intestine.
J Exp Med 1926;44:15171.
41.
Potter JG, Mann FC. Pressure changes in the biliary tract. Am J Med Sci 1926;171:20217.
42.
Lueth HC. Studies on the flow of bile into the duodenum and the existence of a sphincter of
Oddi. Am J Physiol 1931;99:23752.
43.
Bergh GS, Lane JA. A demonstration of the independent contraction of the sphincter of the
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
common bile duct in human subjects. Am J Physiol 1940;128:6904.

Long H. Observations on choledochoduodenal mechanics and their bearing on physiology
and pathology of biliary tract. Br J Surg 1942;29:42237.
Smith JL, Walters RL, Beal JM. A study of choledochal sphincter action. Gastroenterology
1952;20:12937.
Caroli J, Porcher P, Peqvignot G, et al. Contributions of cineradiography to the study of
function of the human biliary tract. Am J Dig Dis 1960;5:67796.
Daniels BT, McGlone FB, Job H, et al. Changing concepts in common bile duct anatomy and
embryology. JAMA 1961;178:3947.
Nebesar RA, Pollard TJ, Potsaid MS. Cinecholangiography: Some physiological observations.
Radiology 1966;86:4759.
Ono K, Watanabe N, Suzuki K, et al. Bile flow mechanisms in man. Arch Surg
1968;96:86974.
Schwegler RA, Boyden EA. Development of pars intestinalis of the common bile duct in the
human foetus with special reference to origin of ampulla of Vater and sphincter of Oddi:
Composition of musculus proprius. Anat Rec 1937;68:193219.
Kreilcamp BL, Boyden EA. Variability in composition of sphincter of Oddi: Possible factor in
pathologic physiology of biliary tract. Anat Rec 1940;76:48697.
Boyden EA. Anatomy of the choledochoduodenal junction in man. Surg Gynecol Obstet
1957;104:64152.
Sterling JA. Termination of common bile duct. Rev Gastroenterol 1949;16:821845.
Sterling JA, Freidmann RS, Ravel VP, et al. Analysis of post-operative cholangiograms. Surg
Hicken NF, McAlister AT, Cole DW. Residual choledochal stones: Etiology and complications.
Arch Surg 1954;68:64356.
Wapshaw H. Radiographic and other studies of the biliary and pancreatic ducts. Br J Surg
1955;43:13241.
Kantor HG, Evans JA, Glen FG. Cholangiographic: A critical analysis. Arch Surg
1955;70:23752.
Flati G, Flati D, Porowska B, et al. Surgical anatomy of the papilla of Vater and biliopancreatic
ducts. Am Surg 1994;60:7128.
Hughes ESR, Kernutt RH. Terminal portions of the common bile duct and pancreatic duct of
Wirsung. Aust N Z J Surg 1954;23:22335.
Kune GA. Surgical anatomy of the common bile duct. Arch Surg 1964;89:9951004.
Frierson HF, Jr. The gross anatomy and histology of the gallbladder, extrahepatic bile ducts,
vaterian system, and minor papilla. Am J Surg Pathol 1989;13:14662.
Vater A. Dissertatio Anatomica qua Novum Bilis Diverticulum Circa Orificum Ductus
Cholidochi. Disputationum Anatomica Selectarum 1748;3:25973.
Kirk J. Observations on histology of choledochoduodenal junction and papilla duodeni with
particular reference to ampulla of Vater and sphincter of Oddi. J Anat 1944;78:11820.
Kreel L, Sandin B. Changes in pancreatic morphology associated with aging. Gut
1973;14:96270.
Kreel L, Sandin B, Slavin G. Pancreatic morphology. A combined radiological and pathological
study. Clin Radiol 1973;24:15461.
Classen M, Hellwig H, Rsch W. Anatomy of the pancreatic duct: A
duodenoscopic-radiological study. Endoscopy 1973;5:147.
Kasugai T, Kuno N, Kobayashi S, Hattori K. Endoscopic pancreatocholangiography. I. The
normal endoscopic pancreatocholangiogram. Gastroenterology 1972;63:21726.
Cotton PB. The normal endoscopic pancreaticograms. Endoscopy 1974;6:6570.
Birnstingl M. A study of pancreatography. Br J Surg 1959;47:12839.
Varley PF, Rohrmann CA, Silvis SE, et al. The normal endoscopic pancreaticograms.
Radiology 1976;118:295300.
Dawson W, Langman J. An anatomical-radiological study of the pancreatic duct in man. Anat
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
Rec 1961;139:5968.
Schwartz A, Birnbaum D. Roentgenological study of the topography of the
choledochoduodenal junction. AJR Am J Roentgenol 1962;87:7726.
Oi J, Takemoto T, Kondo T. Fibreduodenoscopy: Direct observations on the papilla of Vater.
Phillip J, Koch H, Classen M. Variations and anomalies of the papilla of Vater, pancreas and
biliary duct system. Endoscopy 1974;6:707.
Cross KR. Accessory pancreatic ducts. Special reference to the intrapancreatic course of the
common bile duct. Arch Pathol 1956;61:43440.
Tan CE, Moscoso GJ. The developing human biliary system at the porta hepatis level
between 29 days and 8 weeks of gestation: A way to understanding biliary atresia. Part
1.<\f>Pathol Int. 44:587991994.
Tan CE, Moscoso GJ. The developing human biliary system at the porta hepatis level
between 11 and 25 weeks of gestation: A way to understanding biliary atresia. Part 2. Pathol
Int 1994;44:60010.
Melnick PJ. Polycystic liver. Arch Pathol 1955;59:16272.
Rashed A, May RE, Williamson RCN. The management of large congenital liver cysts.
Kerr DNS, Harrison CV, Sherlock S, et al. Congenital hepatic fibrosis. Q J Med
1961;30:91117.
Foulk WT. Congenital malformation of the intrahepatic tree in the adult. Gastroenterology
1970;58:2536.
Balasegaram M. Hepatic surgery past and present. Ann R Coll Surg Engl 1970;47:13958.
Koga A, Watanabe K, Takiguchi S, et al. Etiologic significance of anatomic variations in the
main intrahepatic bile ducts in hepatolithiasis. Acta Radiol 1987;28:2858.
Paul M. An important anomaly of the right hepatic duct and its bearing on the operation of
cholecystectomy. Br J Surg 1948;35:3835.
Rabinovitch P, Zisk HJ. Rare anomalies of the extrahepatic bile ducts. Ann Surg
1956;144:938.
Walia HS, Abraham TK, Baraka A. Gall-bladder interposition: A rare anomaly of the
extrahepatic ducts. Int Surg 1986;71:11721.
Stringer DA, Dobranowski J, Ein SH, et al. Interposition of the gallbladderOr the absent
common hepatic duct and cystic duct. Pediatr Radiol 1987;17:1513.
Champetier J, Letoublon C, Alnaasan I, Charvin B. The cystohepatic ducts: Surgical
implications. Surg Radiol Anat 1991;13:20311.
Wakabayashi T, Morimoto H, Suzuki K, et al. Aberrant hepatic duct connected with the main
pancreatic duct by anomalous pancreato-biliary ductal union: Case report. J Gastroenterol
1994;29:65660.
Mooseman DA, Coller FA. Prevention of traumatic injury to the bile ducts. Study of structures
of the cystohepatic angle encountered in cholecystectomy and supraduodenal
choledochotomy. Am J Surg 1951;82:13243.
Grant JCB. Atlas of Anatomy. 5th ed. Baltimore: Williams & Wilkins. 1962.
Hayes MA, Goldbert IS, Bishop CC. The developmental basis for bile duct anomalies. Surg
Gross RE. Congenital anomalies of the gallbladder: Review of 146 cases with a report of a
double gallbladder. Arch Surg 1936;32:13162.
Lockwood BC. Congenital abnormalities of the gallbladder. JAMA 1948;136:6789.
Boyden EA. Phrygian cap in cholecystography: Congenital anomaly of the gallbladder. Am J
Roentgenol Radium Ther Nucl Med 1935;33:589602.
Weisel W, Walters W. Diverticulosis of the gall bladder. Proc Staff Meet, Mayo Clinic
1941;16:7537.
Lequesne LP, Ranger I. Cholecystitis glandularis proliferans. Br J Surg 1957;44:44758.
Hobby JAE. Bilobed gallbladder. Br J Surg 1970;57:8702.
99.
Latimer EO, Mendez FL, Hage WJ. Congenital absence of the gallbladder. Ann Surg
1947;126:22942.
100.
Morezas G, Wilson AK. Congenital absence of the gallbladder with stone in the common bile
duct. Lancet 1953;1:6289.
101.
Gerwig WL, Countryman LK, Gomez AC. Congenital absence of gallbladder and cystic duct.
Ann Surg 1961;153:11325.
102.
Rogers AI. Congenital absence of gallbladder with choledocholithiasis. Gastroenterology
1965;48:5249.
103.
Ferris DO, Glazer IM. Congenital absence of gallbladder. Arch Surg 1965;91:35961.
104. Richards RJ, Taubin H, Wasson D. Agenesis of the gallbladder in symptomatic adults. A case
and review of the literature. J Clin Gastroenterol 1993;16:2313.
105.
Boyden EA. Accessory gallbladder: Embryological and comparative study of aberrant biliary
vesicles occurring in man and domestic animals. Am J Anat 1926;38:177231.
106.
Stolkind E. Double gallbladder: A report of a case and a review of 38 cases. Br J Surg
1940;27:7606.
107.
Corcoran DB, Wallace KK. Congenital anomalies of the gallbladder. Am J Surg
1954;20:70925.
108. Harlaftis N, Gray SW, Skandalakis JE. Multiple gallbladders. Surg Gynecol Obstet
1977;145:92834.
109. Adam Y, Metcalf W. Absence of cystic duct. A report of a case, the embryology and a review
of the literature. Ann Surg 1966;164:10568.
110.
Hess N. Surgery of the Biliary Passages and Pancreas. Princeton, NJ, New York: D. van
Nostrand, 1965;241301.
111. Markle GB. Agenesis of the common bile duct. Arch Surg 1981;116:3502.
112.
Swartley WB, Weeder SD. Choledochus cyst with double common bile duct. Ann Surg
1940;141:499503.
113. Teilum D. Double common bile duct. Case report and review. Endoscopy 1986;18:15961.
114. Quintana EV, Labat R. Ectopic drainage of the common bile duct. Ann Surg
1974;180:11923.
115.
Alonso-Lej F, Rever WB, Pessagno DJ. Collective review: Congenital choledochal cyst, with
report of two, and an analysis of 94 cases. Surg Gynecol Obstet 1959;108:130.
116. Longmire WR, Mandiola SA, Gordon HE. Congenital cystic disease of the liver and biliary
system. Ann Surg 1971;174:71126.
117. Watts DR, Lorenzo GA, Beal JM. Congenital dilatation of the intrahepatic bile ducts. Arch
Surg 1974;108:5928.
118. Todani T, Watanabe Y, Narusue M, et al. Congenital bile duct cysts: Classification, operative
procedures and review of thirty-seven cases including cancer arising in a choledochal cyst.
Am J Surg 1977;134:2639.
119. Caroli J. Diseases of the intrahepatic ducts. Clin Gastroenterol 1973;2:14761.
120.
Feldman M, Weinberg T. Aberrant pancreas: A cause of duodenal syndrome. JAMA
1952;148:8938.
121.
Stofer BE. Annular pancreas: A tabulation of the recent literature and report of a case. Am J
Med Sci 1944;207:4305.
122.
Whelan IJ, Hamilton GB. Annular pancreas. Ann Surg 1957;146:25262.
123.
Hays DM, Greaney EM, Hill JT. Annular pancreas as a cause of acute neonatal duodenal
obstruction. Ann Surg 1961;153:10312.
124. Ravitch MM. The pancreas in infants and children. Surg Clin North Am 1975;55:37785.
125. Heyman RL, Whelan TG. Annular pancreas, demonstration of annular duct on
cholangiography. Ann Surg 1967;165:4702.
126. Lloyd Jones W, Mountain JC, Warren KW. Annular pancreas in the adult. Ann Surg
1972;176:16370.
127.
Swynnerton BF, Tanner NC. Annular pancreas. BMJ 1953;1:10289.
128. McGregor AMC, Green BJ, Stein MA. Symptomatic annular pancreas in the adult. Br J Surg
1969;56:7135.
Huebner GD, Reed PA. Annular pancreas. Am J Surg 1962;104:86973.
Anderson JR, Wapshaw H. Annular pancreas. Br J Surg 1951;39:439.
O'Rahilly R, Muller F. A model of the pancreas to illustrate development. Acta Anat
1978;100:3805.
132. Cotton PB. Congenital anomaly of pancreas divisum can cause obstructive pain and
pancreatitis. Gut 1980;21:10514.
133. Sigfusson BF, Wehlin L, Lindstrom CG. Variants of pancreatic duct system of importance in
endoscopic retrograde cholangiopancreatography. Observations on autopsy specimens. Acta
Radiol (Diagn) (Stockh) 1983;24:11328.
134.
Millbourn E. Calibre and appearance of the pancreatic ducts and relevant clinical problems:
Roentgenological and anatomical study. Acta Chir Scand 1960;118:286303.
135.
Millbourn E. On excretory ducts of pancreas in man, with special relations to each other, to
the common bile duct and to duodenum. Radiological and anatomical. Acta Anat (Basel)
1950;9:134.
136.
Reinhoff WR, Pickerell KL. Pancreatitis: An anatomical study of the pancreatic and
extrahepatic systems. Arch Surg 1945;51:20519.
129.
130.
131.
Chapter 59 The Normal Retrograde Pancreatogram and

Cholangiogram
(5925)
(5926)
Postmortem radiographic contrast study of the biliary system was reported in 1921,1(5927) and
percutaneous cholangiography was introduced into clinical practice by Huard and Do-Xuan-Hop in
1937,2(5928) although the procedure was not widely utilized until after the report of Carter and
Saypol.3(5929)
The main pancreatic duct was described in 1642 by Wirsung; 100 years later, Santorini demonstrated
the existence of the accessory pancreatic duct. There are numerous subsequent descriptions of
pancreatic anatomy, some dealing with the morphology of the ductal system. These had little clinical
relevance, however, until the development of intraoperative pancreatography in the early 1950s by
Doubilet and Mulholland.4(5930) With the introduction of this technique, the fine points of pancreatic
ductal anatomy assumed greater clinical relevance. Furthermore, operative pancreatography
demonstrated the feasibility and safety of radiographic contrast material imaging of the pancreatic duct,
although clinical application was restricted by the need for laparotomy.
Because of the earlier availability of radiologic techniques for visualization of the bile ducts, familiarity
with the radiographic anatomy of the biliary system became commonplace while pancreatography
remained relatively obscure. The inception in 1968 of endoscopic retrograde cannulation of the main
duodenal papilla resulted in detailed radiographs of the entire pancreaticobiliary system; thus, a greater
knowledge of the radiographic anatomy of these structures became of paramount importance to
endoscopists. Knowledge of the essential features of the embryologic development of the pancreas
contributes to an understanding of pancreatobiliary ductal anatomy (see Chapter 58: Anatomy and
Embryology of the Biliary Tract and Pancreas).
The Normal Retrograde Pancreatogram

The endoscopic retrograde pancreatogram is a representation of a limited portion of the anatomy of
the pancreas; only the main pancreatic duct and its branches are demonstrated, whereas the bulk of
the parenchyma is not visualized. The pancreatic ducts have been studied by different methods,
including standard dissection, casts of the ductal system, and postmortem radiographs, either in
cadavers or after removal of the gland. It is likely that the results of each are not strictly comparable
with respect to the dimensions of the ductal system. The possibility that autolysis and other
postmortem changes in the pancreas may alter the results of studies of autopsy specimens must also
be considered. However, Kano et al.5(5931) found that endoscopic retrograde and postmortem
pancreatograms were almost identical in six dogs.
The Radiographic Configuration and Location of the Main Pancreatic Duct

The general shape of the main pancreatic duct as demonstrated by retrograde pancreatography is
angulated or "pistol-shaped" (Figure 591). An abrupt bend or genu, sometimes greater than 90
degrees, is characteristically found to the right of the spine. This frequently corresponds to the point of
embryologic fusion of the dorsal and ventral ducts. If the accessory pancreatic duct is opacified by
retrograde contrast material injection, its junction with the main pancreatic duct is usually found near
this bend (Figure 592). However, this relatively abrupt angulation in the course of the duct need not
be present.
(5932)Figure 591. Normal "pistol-shaped" retrograde pancreatogram demonstrates the

characteristic angulation or genu.
(5933)Figure 592. Normal retrograde pancreatogram demonstrates the accessory duct

(arrow) and the normal pattern of the main pancreatic duct branches in the tail.
The main pancreatic duct often follows a generally cephalad course within the retroperitoneum,
proceeding from the medial duodenal wall upward and to the left. In one autopsy series, the position of
the main pancreatic duct in the head remained within 5 mm of the common bile duct over a mean duct
length of 3.9 cm in 68% of specimens studied.6(5934) However, this ascending course was found by
Kasugai et al.7(5935) in only 48.5% of patients. In another study, the mean inclines in degrees
(measured from a line perpendicular to the spine) were 74, 19, and 21 degrees for the main pancreatic
duct in the head, body, and tail, respectively.8(5936) Kang et al.,9(5937) in a study from Korea of 286
normal retrograde pancreatograms, found the course of the duct to be ascending in 51.7% of cases,
horizontal in 22.4%, and descending in 0.7%. In 25.2% of cases, the pancreatogram had a variable
("sigmoid") course. In general, the course of the main duct may be approximately parallel to the spine
in the head of the gland, a relatively constant relationship, whereas it tends to be horizontal in the
body.10,11(5938) The greatest variation in course is usually found in the tail of the pancreas.
Varley et al.12(5939) found that the origin of the pancreatic duct at the main duodenal papilla was
separated from the spine by a mean distance of 22.8 mm (10.0 mm); the maximum value found for
this distance was 49.0 mm. Karak et al.13(5940) found the origin at the papilla to be within 30 to 39
mm of the right lateral margin of the corresponding vertebrae in slightly more than half of 101 normal
pancreatograms.
Retrograde pancreatograms most often present the ductal anatomy in an anteroposterior view of the
gland, a perspective that is not three-dimensional. The course of the duct is actually not flat within the
retroperitoneal space. Rather, it is U-shaped because it lays over the spine (Figure 593). The main
pancreatic duct in the body may be several centimeters anterior to the ductal segments in the head
and tail. This also means that slight rotation of the patient from the prone position to an oblique one will
markedly alter the apparent radiographic shape of the main duct (this can also be accomplished by
rotation of the x-ray tube). The anteroposterior separation of the main duct from the spine was
determined by Varley et al.12(5941) to be 56.7 mm (17.4 mm). Unfortunately, the wide range of
normal (7.9 to 96.1 mm) for this measurement negates its value for determination of pathologic
displacement of the gland.
(5942)Figure 593. A, Anteroposterior retrograde pancreatogram. B, Left lateral view of the

pancreatogram demonstrating the U-shaped configuration. Note the position of the midportion
of the ductal system in the body of the pancreas posterior to the contrast medium-filled
stomach (long arrow) and the posterior position of the pancreatic tail relative to the position
of the pancreatic body (short arrow). C, Right lateral view of the duct in the head of the
pancreas. Note the posterior position of the duct in the head of the gland relative to that of the
duct in the body.
Radiographically, the entire main pancreatic duct is usually located within the span of T12-L2 when
viewed with the patient in a perfectly prone position. There is, however, a large variation in normal
location.
Classifications of the radiographic shape as well as the course and position of the main pancreatic duct
have been proposed by many investigators. However, its configuration and location are so highly
variable that assessment of these characteristics is of limited practical value.11(5943)
Radiographic Characteristics of the Pancreatic Ductal System

General Appearance
The main pancreatic duct has a smooth or somewhat undulating contour and progressively decreases
in caliber, proceeding from its duodenal terminus to the tail of the gland.*(5944) Minor variations in
width are usually present, although the variations in caliber from point to point along the course of the
duct are only slight and never abrupt.
Normal Narrowing of the Main Pancreatic Duct
Narrowing of the radiographic outline of the main pancreatic duct may be normal in two locations. The
first is near the genu in proximity to the junction with the accessory pancreatic duct. Birnsting14(5945)
found this relative stenosis to be present in 3% of cases in an autopsy study but found no associated
histologic abnormality (Figure 594). He attributed this narrowing to anatomic "folding" of the duct at
the junction of the head and body. Berman et al.15(5946) also found this narrowing in a postmortem
study using vinyl acetate and latex casts of the ductal system, and it has been noted in other autopsy
and radiographic studies16,17(5947) and operative pancreatograms as well.18(5948)
(5949)Figure 594. Retrograde pancreatogram shows slight narrowing (arrow) of the main
duct at the genu. (From Sivak MV Jr, Sullivan BH Jr. Endoscopic retrograde pancreatography.
Analysis of the normal pancreatogram. Dig Dis 1976; 21:2639. By permission of Karger,
Basel.)
The second normal narrowing of the main pancreatic duct occurs within the body of the gland where
the main duct overlies the superior mesenteric vessels. Berman et al.15(5950) found this narrowing to
be consistently present, and distortion of the pancreatographic anatomy by the superior mesenteric
vessels has been noted in other postmortem studies.19(5951) Although narrowing has been noted at
this location in endoscopic retrograde cholangiopancreatography (ERCP) series,17(5952) it is not
described in most pancreatographic studies nor in other postmortem investigations.
Good-quality pancreatograms with complete opacification of the pancreatic ductal system often
demonstrate the terminus of the main duct at the duodenal wall and main duodenal papilla (Figure
595). The diameter of the duct decreases as it comes into relationship with these structures. Kasugai
et al.7(5953) were able to measure ductal length and diameter within these areas on a few retrograde
pancreatograms. The section of the main pancreatic duct in the head of the gland immediately
adjacent to the duodenal wall and papilla may appear widened in a fusiform fashion. This is also a
normal finding and does not indicate obstruction to flow at the main duodenal papilla. This appearance
has been attributed to the decrease in diameter of the ductal system as it passes through the duodenal
wall and papilla.20(5954)
(5955)Figure 595. Retrograde pancreatogram shows narrowing of the duct in the region of
the duodenal wall and papilla (long arrow) with "fusiform" widening of the main duct in its
retropapillary portion (short arrow). The main duct branch to the uncinate process is also
present.
Main Pancreatic Duct Branches
Branching of the main pancreatic duct occurs in a orderly fashion in the body and tail of the pancreas.
Branches join the main duct at right angles and usually alternate points of attachment with adjacent
ducts on the opposite side of the main duct (Figure 596A). The branches in the midbody are said to
be fewer in number.11(5956) By contrast, the branching pattern in the pancreatic head is disorderly,
reflecting the extensive embryologic transposition of the ventral anlagen (Figure 596B).
(5957)Figure 596. A, Normal retrograde pancreatogram demonstrates the normal branching

pattern in the body and tail. Note the orderly arrangement of the branches. B, Normal
branching pattern in the tail of the pancreas.
With good technique, it is usually possible to safely opacify the first- and some second-order main
pancreatic duct branches during retrograde pancreatography. Ishibashi and Matsubara,21(5958) in
their study of 33 cadaver pancreases, found the mean number of main pancreatic duct first-order
branches to be 56, with a range of 52 to 66. The number of main pancreatic duct branches is thought
by some investigators to decrease with increasing age of the patient, although this was not found by
Ishibashi and Matsubara.21(5959) When demonstrated by retrograde contrast material injection, the
branch ducts taper smoothly and gradually, proceeding from the main duct. Marked variations in
caliber, especially a wider peripheral diameter relative to that near the point of origin, are usually
considered pathologic, especially in younger patients. This contrasts, however, with the necropsy study
of Kreel and Sandin,22(5960) in which cystic changes and other irregularities of the branch ducts were
found, especially in older but apparently normal subjects. Similar distortions were found in association
with increasing age in the necropsy study by Ishibashi and Matsubara.21(5961) In a study of
pancreatograms from 101 patients aged 75 years or more, Jones et al.23(5962) found minor variations
in diameter of the main pancreatic duct to be common along with filling defects, abnormalities in main
duct branches, cavities less than 5 mm in diameter, and intraductal or acinar calcification. Only 4
patients in this study were subsequently shown to have chronic pancreatitis. Although it was assumed
that the remaining patients did not have pancreatic disease, there is the possibility that the gland was
not absolutely normal in some additional patients.
Alterations in the main pancreatic duct branches are recognized more frequently in postmortem studies
and less often in endoscopic investigations. Although clinical data are limited, it appears that cystic
dilation, tortuosity, and narrowing of branch ducts are not due to postmortem changes in the pancreas.
Microscopic study of the abnormal areas generally reveals some evidence of mild fibrosis but no
inflammatory cells. It is possible that the postmortem radiographs are of better quality than those
obtained in vivo or that distortion is caused by postmortem methods of investigation. It is more likely,
however, that minor degrees of cystic change, tortuosity, and narrowing of the main pancreatic duct
branches may be a normal accompaniment of aging (Figure 597). As discussed in Chapter 73:
Endoscopic Retrograde Cholangiopancreatography in Acute and Chronic Pancreatitis, this
radiographic appearance may also be interpreted as evidence of minimum pancreatitis. It is possible
that some patients with minor ductal abnormalities were excluded from endoscopic series as being
examples of a mild degree of pancreatitis. In the study of Schmitz-Moormann et al.,24(5963) 69
postmortem pancreatograms from subjects with no clinical or histologic evidence of chronic
pancreatitis were interpreted by a group of six experts as demonstrating chronic pancreatitis in 81% of
cases. The degree of pancreatitis was interpreted as minimum in 37%, moderate in 33%, and even
severe in 11% of cases. The presence of ductular changes correlated with the presence of perilobular
fibrosis histologically. This problem is not considered or is unrecognized in most studies of normal
retrograde pancreatograms, and some latitude within the normal range is allowable with respect to the
radiographic appearance of the main pancreatic duct branches in elderly patients.
(5964)Figure 597. Normal pancreatogram in an 80-year-old patient. On close inspection,

minor irregularities can be seen in the main pancreatic duct branches.
Other than the accessory pancreatic duct, the only relatively consistent major branch of the main duct
is a large unpaired one in the head of the gland that is directed inferiorly to the uncinate process
(Figure 598). Birnstingl14(5965) credited Cordier with the first description of this branch in 1952. It
has been noted by numerous investigators in both operative and endoscopic pancreatographic
studies.7,25(5966) In the endoscopic study of Sivak and Sullivan,17(5967) it was found in 55.5% of
pancreatograms; in the autopsy study by Rienhoff and Pickrell,26(5968) it was noted in 61% of cases.
This branch to the uncinate process may be quite long, and there is a description of its termination in a
third papilla in the third portion of the duodenum.26(5969)
(5970)Figure 598. Retrograde pancreatogram demonstrates a large unpaired main pancreatic

duct branch to the uncinate process of the pancreas (long arrow). Note the accessory duct also
(short arrow). (From Sivak MV Jr, Sullivan BH Jr. Endoscopic retrograde pancreatography.
Analysis of the normal pancreatogram. Dig Dis 1976; 21:2639. By permission of Karger,
Basel.)
Accessory Pancreatic Duct
Opacification of the accessory pancreatic duct by retrograde pancreatography is unpredictable. In the
study of Sivak and Sullivan,17(5971) it was found in only 14% of pancreatograms. Kasugai et
al.7(5972) noted the accessory duct in 32.4% of normal pancreatograms; Roberts-Thompson27(5973)
in 40%; Varley et al.12(5974) in 62%. However, the accessory duct is found in most specimens studied
by postmortem methods. Rienhoff and Pickrell26(5975) noted it in 100% of autopsy specimens.
Birnstingl14(5976) found it in 68%, although it was patent in 55.3%. The accessory pancreatic duct
may not communicate with the main duct. In a postmortem study by Hand,20(5977) the two ducts
communicated in 88% of specimens; Reinhoff and Pickrell26(5978) found a communication in 89%.
Generally, visualization or lack of opacification at retrograde pancreatography is not a reliable indicator
of accessory duct patency, nor does caliber relate to functional status. Occasionally, careful
observation at the time of retrograde injection at the main papilla may reveal flow of contrast material
into the duodenum via the accessory duct and papilla, but this is unusual (Figure 599).
(5979)Figure 599. Normal retrograde pancreatogram is demonstrated by retrograde injection

via the minor duodenal papilla. The configuration of the duct in the head of the gland is
unusual but can be regarded as a variant of normal. Contrast medium flows through the main
duodenal papilla into the duodenum.
Dimensions of the Main Pancreatic Duct
The dimensions of the main pancreatic duct have been obtained in a number of ways. Radiograms
have been made by contrast material injection into the duct in cadavers;16(5980) they have also been
obtained after the pancreas has been removed from cadavers and placed flat on an x-ray
cassette.14(5981) Differences in radiologic technique undoubtedly produce variations in the ductal
measurements. In the postmortem study by Kreel and Sandin,22(5982) the length of the main duct
was determined by contrast radiography in cadavers and after the pancreases were removed. The
dimensions obtained by the two techniques differed by 1 to 2 cm, the measurements made with the
pancreas in situ being of lesser magnitude in most cases. Another method utilizes injection casts of the
ductal system. The latter technique, probably the most accurate, also produces a three-dimensional
representation of the ductal system, although postmortem tissue changes may distort normal features.
Measurements may also be obtained from endoscopic retrograde pancreatograms with the aid of
calipers and a magnifying lens. The length of the main pancreatic duct can be approximated with an
instrument designed for measurement of distance on road maps. Although opacification of
progressively finer branches of the ductal system and ultimately the pancreatic acini is a function of the
injected volume of contrast material, the radiographic ductal diameters probably do not change
appreciably in response to fluctuations or differences in injection pressure. Different radiologic
techniques produce variations from one study to another. In particular, it is necessary to correct for the
magnification that occurs in making the x-rays. The magnification factor associated with the usual
ERCP procedure with the patient prone on the x-ray table is about 10 to 20%. This may be estimated
by comparing the known diameter of the tip of the duodenoscope with its apparent diameter on the
x-ray films. Because the three-dimensional shape of the gland is not completely flat within the
retroperitoneum, the use of a correction factor will be more accurate with measurements in the head of
the pancreas than with those in the body and the tail. The reported dimensions of the pancreatic duct
are corrected for magnification in some ERCP reports of normal pancreatographic
anatomy,11,12,17,28(5983) but not in others.8(5984)
Despite the technical limitations of the various methods, an approximation of the normal dimensions of
the pancreatic ductal system can be obtained by collective reference to published reports. The
dimensions obtained by in vivo endoscopic studies are shown in Table 591, and those from
postmortem investigations are given in Table 592. Howard and Short18(5985) found 4 mm to be the
upper limit of normal for the diameter of the main duct by intraoperative pancreatography.
Cotton11(5986) accepts 6 mm as the upper limit of normal for ductal diameter by retrograde
pancreatography. Birnstingl,14(5987) in an autopsy-radiographic study of 150 pancreases, found that
the diameter of the duct in the head of the gland ranged from 1.8 to 9.2 mm, and he recommended
that diameters up to 8 mm be considered normal unless there is other evidence of pancreatic disease.
As a general rule, the dimensions found in autopsy studies are somewhat greater than those obtained
from retrograde pancreatograms. The greatest variability in duct diameter occurs in the head of the
pancreas.
TABLE 591
Radiographic Dimensions of the Pancreatic Duct by Retrograde
Pancreatography
that diameters up to 8 mm be considered normal unless there is other evidence of pancreatic disease.
As a general rule, the dimensions found in autopsy studies are somewhat greater than those obtained
from retrograde pancreatograms. The greatest variability in duct diameter occurs in the head of the
pancreas.
Radiographic Dimensions of the Pancreatic Duct by Retrograde
TABLE 591
Pancreatography
REFERENCE
NUMBER
OF DUCTS
HEAD
(MM)
BODY
(MM)
TAIL
(MM)
LENGTH
(CM)
RANGE
(CM)
Ogoshi et al.28
Classen et al.8
25
3.4
2.9
2.0
48
4.8
3.5
2.4
20.1
16.424.2
Kasugai et al.7
Cotton11
68
3.5
2.7
1.7
16.2
57
3.3
2.5
14.5
2.5
9.518
Seifert*
Okuda et al.43
28
20
4.0
3.9
3.0
3.4
2.3
2.1
18.6
15.7
1721
Sivak and Sullivan17
35
3.2
2.3
1.2
15.4
12.219
Moshal and Engelbrecht44

Varley et al.12
27
3.7
2.7
1.7
16.3
102
3.1
2.0
0.9
16.9
10.722.3
Roberts-Thompson27
Anand et al.30
30
3.1
2.4
1.4
55
3.3
2.56
1.2
16.55
10.5223.6
Kang et al.9
126
3.3
2.4
1.5
17.5
* Seifert's data quoted by Cotton.11
Dimensions of the Main Pancreatic Duct

in Postmortem Studies
TABLE 592
REFERENCE
NUMBER OF
DUCTS
HEAD
(MM)
BODY
(MM)
TAIL
(MM)
Hand20
50
3.04.0
Birnstingl14
150
4.1
Trapnell et al.16
Milbourn29
45
4.0
2.0
1.0
146
4.4
2.1
In some postmortem studies, the diameter of the main pancreatic duct has been found to be greater in
specimens obtained from older as opposed to younger individuals.21,29(5988) The findings from two
such studies are tabulated in Table 593. Kreel and Sandin22(5989) in studying 120 necropsy cases
found that the width of the main pancreatic duct increased by about 8% with each decade of increase
in the age of the subjects. The lengths of the main duct when separated according to age category did
not differ significantly in this study. Kasugai et al.7(5990) found a tendency for the diameter of the main
duct on retrograde pancreatograms to be greater in older age groups, but this increase was not
significant statistically. Similar findings were noted by Kang et al.9(5991) In a study from India of
retrograde pancreatograms from 55 patients with no evidence of pancreatic disease, Anand et

al.30(5992) found a statistically significant difference in the diameter of the main pancreatic duct in the
head and body but not the tail for individuals 40 years of age or older compared with those of lesser
age. Statistically significant differences were also found for duct segment diameters in patients 40
years or less in age compared with subjects of greater age by Karak et al.13(5993) in a study of 101
normal pancreatograms. In the study of Anand et al.,30(5994) there was also a significant difference in
the diameter of the accessory pancreatic duct, but there was no difference in length of the main
pancreatic duct. Other studies of retrograde pancreatograms fail to demonstrate a relation between
age and ductal diameters.11,17,27(5995)
TABLE 593
Dimensions of the Main Pancreatic Duct by Age
REFERENCE
AGE
(YR)
HEAD
(MM)
BODY
(MM)
TAIL
(MM)
LENGTH
(CM)
Postmortem Studies
Kreel and Sandin22
<50
3.2
2.4
1.4
17.0
4.3
5.0
4.6
5.3
3.3
3.5
3.5
3.5
4.0
1.8
2.0
2.1
2.1
19.0
19.6
19.8
19.4
MacCarty et al.31
5059
6069
7079
80>
3150
5170
7190
3.5
4.6
<40
2.97
2.36
1.23
16.06
40
1019
3.78
2.6
2.86
1.7
1.15
0.8
17.31
13.2
2029
2.7
2.1
1.5
15.0
3039
3.0
2.3
1.5
16.8
4049
3.2
2.3
1.6
17.9
5059
3.5
2.5
1.7
18.0
60>
4.0
3.0
2.1
18.6
40
>40
2.63
1.95
0.99
16.1
3.31
2.34
1.23
16.58
Retrograde Pancreatography
Anand et al.30
Kang et al.9
Karak et al.13
Attempts have been made to relate ductal dimensions to sex and body habitus. Cotton11(5996) found
the main duct to be significantly longer in men than in women in a review of retrograde
pancreatographic data from several sources. Others have not found any relation between main
pancreatic duct length and gender.7,17,27,30(5997) Milbourn29(5998) did not find any correlation
between length and gender in a postmortem study. No difference in ductal diameter with respect to sex
was found in the necropsy study by MacCarty et al.31(5999) Sivak and Sullivan17(6000) attempted to
correlate the length of the main pancreatic duct with body surface area, but a relation was not found.
Although studies of the pancreatic ductal system have been reported from many countries, there are
no readily apparent differences in the dimensions of the pancreatic ductal system with respect to race
or ethnic background.
The length of the main pancreatic duct visualized by retrograde pancreatography may be reduced by
virtually any disease that affects the pancreas. In particular, pancreatic carcinoma (which usually
originates within the ductal system) may produce the radiographic pattern of ductal obstruction. When
this occurs in the head or body, there is little difficulty in recognizing that the pancreatogram is
abnormal. However, obstruction nearer or within the tail of the pancreas may be more problematic; it is
more difficult to be certain whether some portion of the main duct remains unopacified.
It would be desirable from the standpoint of diagnosis to be able to categorize pancreatograms as
normal or abnormal by obtaining measurements of the pancreatic duct from x-rays. Unfortunately, this
is not possible in practice. For example, there is relatively close agreement among the various reports
cited in Table 591 with respect to the mean length of the main duct. This is approximately 16 to 17
cm. However, the range for the normal length of the main duct (see Table 591) is too great for the
determination of ductal length to be of value in the diagnosis of obstructing carcinoma in the tail of the
pancreas. The normal radiographic length of the main duct is never less than 9 cm in any report. A
relatively short duct may still be within normal limits, although ductal length of less than 9 cm will
usually be pathologic.
The diameter of the main duct is also altered by certain diseases, for example, chronic pancreatitis.
For practical purposes, the values 3 to 4 mm, 2 to 3 mm, and 1 to 2 mm (corrected for magnification)
may be used as normal diameters for the main duct in the head, body, and tail, respectively. However,
a diameter of up to 6 mm in the head, 5 mm in the body, and 3 mm in the tail of the pancreas can be
normal.11(6001) By measurement alone, this upper limit of normal probably overlaps the dimensions
of the duct in patients with chronic pancreatitis.
Normal Retrograde Cholangiogram

The subdivisions and terminology applied to the radiographic anatomy of the biliary system are the
same as those used for surgical and gross anatomic description. Gross anatomic features of the bile
ducts are discussed in Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas. The
many normal variations in the biliary ducts and anomalies are also discussed in Chapter 58. There are
relatively few reports of the normal radiographic anatomy of the biliary system as demonstrated by
retrograde cholangiography. On cannulation of the bile duct, it is usually possible to opacify the
common bile duct, cystic duct, common hepatic duct, and at least some parts of the right and left
hepatic ducts and intrahepatic system. Certain technical problems with opacification and radiographic
demonstration of the gallbladder and intrahepatic ducts are discussed in Chapter 66: The Gallbladder
and Intrahepatic Bile Ducts.
General Appearance
The radiographic appearance of the biliary ducts is similar to that of a leafless tree. Branching into
progressively smaller tributaries occurs within the liver, whereas the extrahepatic ducts are without
branches (Figure 5910). The length, course, and origin of the cystic duct are highly variable (see
Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas). The origin may be in a very
distal location on the medial aspect of the bile duct where it is in close relation to the pancreas (Figure
5911). When in this position, the cystic duct may be confused with the common duct during the initial
injection of contrast material.
(6002)Figure 5910. Normal retrograde cholangiogram.
(6003)Figure 5911. Normal retrograde cholangiogram with "low" distal-medial insertion of

the cystic duct.

When viewed from anterior to posterior, the bile duct is usually aligned in an inferosuperior direction
approximately parallel to the spine. However, it may be angulated in the elderly individual or its course
may be distorted as a result of prior surgery. The point of contact with the superior margin of the
pancreas is sometimes marked radiographically by a slight curve or bend toward the left. Viewed
laterally, the course of the duct is directed somewhat anteriorly (Figure 5912).
(6004)Figure 5912. Retrograde pancreatocholangiogram in a patient who has had

cholecystectomy. A, Anteroposterior view demonstrates the relatively straight course of the
extrahepatic bile duct that is approximately parallel to the spine except for a slight medial
angulation of the common hepatic duct. B, Anteriorly directed plane of the extrahepatic
system is demonstrated by a lateral view.
For practical purposes, the extrahepatic bile ducts have no muscular component except at the most
distal aspect of the common duct. The muscular component of the distal common duct can produce an
abrupt narrowing in the radiographic contour of the duct; this may be noted as occurring within a few
millimeters of the duodenal wall itself. This narrowing may be abrupt in some cases and has been
described as a "notch" by Hand.20(6005) The ERCP cannula may also distort and exaggerate the
normal radiographic anatomy of the distal bile duct (Figure 5913). A radiographic shelflike deformity of
the bile duct just above the papilla can be misinterpreted as a constricting lesion. This can often be
resolved by removing the cannula after opacification of the duct and observing the
choledochoduodenal area with sequential x-ray views over the course of time (Figure 5914). As
pharmacologically induced duodenal atony wanes, the outline of the distal muscular portion of the bile
duct and choledochoduodenal area will change in appearance as the bile duct empties.
(6006)Figure 5913. Notching of the distal common duct (arrow) in the choledocho-duodenal
area. The appearance suggests a constricting lesion. Note that the bile duct is not dilated.
Impaction of the cannula may alter the appearance.
(6007)Figure 5914. A, Retrograde cholangiogram demonstrates an apparent stricture of the

distal common bile duct. Note the associated duodenal contraction. B, Cholangiogram (same
patient) obtained 1 minute later shows the free flow of contrast medium from the bile duct and
pancreatic duct into the duodenum and no evidence of stricture.
There are no normal points of narrowing within the biliary system, with the exception of that which
occurs in the region of the choledochoduodenal junction, as described previously.
Dimensions of the Biliary System

The accuracy of measurements of the bile duct, as for the pancreatic duct, is no doubt influenced by
the method of measurement. Chang et al.32(6008) measured the maximum diameter of the
extrahepatic bile duct by transabdominal ultrasonography in 24 patients immediately before and after
endoscopic retrograde opacification. The sonographic diameter increased by 3 mm or more in 9
patients. An increase in diameter in response to retrograde injection of radiographic contrast medium
was found more often in patients who had undergone cholecystectomy. These investigators
suggested, therefore, that simple measurements from x-ray films probably do not give the true ductal
diameter. Davies et al.33(6009) also obtained ultrasonographic measurements before and after
retrograde cholangiography in a prospective study of 19 patients. Measurements of the diameter of the
bile duct obtained cholangiographically, corrected for magnification, were more than twice those
determined ultrasonographically. Distention of the duct due to retrograde injection of contrast material
was not thought to be the explanation for the discrepancy. Based on retrospective comparisons of
paired imaging studies, these investigators suggested that most often the ultrasonographic
measurement was of the right hepatic duct, in which case there was no significant difference between
the two types of measurements.
Niederau et al.34(6010) performed ultrasonography of the bile ducts in 101 patients with a variety of
hepatic and biliary diseases within 1 day before contrast radiographic studies that included intravenous
cholangiography (n = 37), endoscopic retrograde cholangiography (n = 47), and percutaneous
transhepatic cholangiography (n = 33). As measured by all three cholangiographic methods, the mean
diameter of the bile duct was significantly greater than that determined by ultrasonography. The
differences were greater for the retrograde and percutaneous as compared with the intravenous
method of cholangiography. When cholangiograms were compared for 22 patients who underwent
both intravenous and retrograde cholangiography, the diameter as determined by the latter imaging
method was significantly greater. Belsito et al.35(6011) determined that measurements of the bile
ducts on ERCP x-rays exceeded corresponding determinations from intravenous cholangiogram x-rays
in almost all instances, the average difference being about 3 mm. Niederau et al.34(6012) believed that
sonographic measurement of the bile duct at varying sites did not account for the differences in their
study. They considered the following as factors that could account for the discrepancies: radiographic
magnification, choleretic effect of the intravenous contrast agent, sonographic minification, and
distention of the bile duct due to injection of a contrast agent.
There are only a few reports of the dimensions of the bile ducts by retrograde cholangiography. No
reports compare radiographic dimensions taken from retrograde cholangiograms with ones from
percutaneous transhepatic cholangiograms.
As with pancreatography, variations in technique will produce different values with respect to
measurement of the width of the bile duct at any given point. O'Dwyer et al.,36(6013) for example,
measured the mean maximum diameter of the common hepatic and common bile ducts on
radiographs taken in both the supine and the prone positions in 12 patients. The measured widths of
the common hepatic duct with patients supine and prone were 10.1 mm (0.9) and 11.0 mm (1.0),
respectively. The diame-ters of the common duct, measured supine and prone, were 13.0 (1.1) and
11.6 mm (1.0), respectively. Measurements from retrograde cholangiogram x-rays must be corrected
for magnification. Usually this is accomplished by comparing the known diameter of the tip of the
endoscope with its apparent diameter on the x-ray films. O'Dwyer et al.36(6014) found that the average
magnification of the common bile and common hepatic ducts was about 23%. However, there was a
marked variation from one patient to the next; the magnification factor ranged from about 8% to almost
50%. Because such a marked degree of magnification is possible, every retrograde cholangiogram
should be corrected for magnification before considering the possibility of dilation of the biliary system.
The normal dimensions for the bile ducts from retrograde cholangiographic studies are summarized in
Table 594. The range for the reported normal diameters in Table 594 is given in Table 595.
Hamilton et al.37(6015) also give normal values for three generations of intrahepatic ducts after the
bifurcation into the left and right hepatic ducts; these are 2.5 mm (1.5 to 4.5), 2 mm (1 to 3), 1 mm (1 to
2). These investigators also found that the diameters of the ducts of the left and right intrahepatic
system did not differ significantly.
Normal Bile Duct Diameters by Retrograde
TABLE 594
Cholangiography
REFERENCE
NUMBER OF
SUBJECTS
CBD*
CHD*
LHD*
RHD*
Normal Bile Duct Diameters by Retrograde
TABLE 594
Cholangiography
REFERENCE
O'Dwyer et al.36
Hamilton et al.37
Lasser et al.38
NUMBER OF
SUBJECTS
CBD*
CHD*
30
6.1
6.1
50
6.5(p)
6.0
49
4.5(i)
4.9(p)
4.6
LHD*
4.5
RHD*
4.5
4.3(i)
CBDcommon bile duct; CHDcommon hepatic duct; LHDleft hepatic duct;
RHDright hepatic duct; (p) prepancreatic; (i) intrapancreatic.
* Mean maximum values in millimeters, corrected for magnification.
TABLE 595
REFERENCE
O'Dwyer et al.36
Hamilton et al.37
Lasser et al.38
Range of Normal Bile Duct Diameters by Retrograde Cholangiography

NUMBER OF
SUBJECTS
CBD*
CHD*
30
3.510
3.513.3
50
313(p)
313
49
310(i)
2.38.5(p)
LHD*
2.59
RHD*
2.58.5
2.19.2
2.36.9(i)
* Mean maximum values in millimeters, corrected for magnification.
CBDcommon bile duct; CHDcommon hepatic duct; LHDleft hepatic duct; RHDright hepatic duct;
(p)prepancreatic; (i)intrapancreatic.
Lasser et al.38(6016) found that the diameters of the prepancreatic portion of the common bile duct
and the common hepatic duct (but not the intrapancreatic portion of the common duct) were
significantly greater in older than in younger individuals. This conclusion is supported by data from
other investigators.39,40(6017) However, O'Dwyer et al.36(6018) found no correlation between
diameter and age. This view has also been supported by data from other investigators.41(6019)
Extrahepatic duct diameters are also not influenced by the presence of parenchymal liver
disease.38(6020)
Chung et al.42(6021) measured bile duct diameters on retrograde cholangiograms obtained in 43
patients before cholecystectomy and on cholangiograms obtained from 4 to 14 months after removal of
the gallbladder. All patients had a normal bile duct at operation and were asymptomatic after surgery.
With reference to 35 sets of cholangiograms that could be evaluated, the largest mean diameter of the
common duct, corrected for magnification, was 0.96 cm (0.27) before and 1.16 (0.29) after
operation. An increase in ductal diameter was demonstrated in 31 patients; the increase was 1 mm or
more in 26 patients. The magnitude of the increase in diameter correlated with length of time elapsed
after surgery.
It is not possible to determine that an abnormality of the biliary system is present or absent by simply
measuring the caliber of the ducts. O'Dwyer et al.36(6022) determined bile duct diameters in 14
patients who had a cholecystectomy and found that the mean maximum diameters of the common bile
and common hepatic ducts were significantly greater than the corresponding values in normal patients.
However, the measured duct diameter was beyond the upper limit of the normal range in only 2 of the
14 patients. These same authors also measured duct diameters in 46 patients with biliary obstruction.
The mean maximum diameters were again significantly greater than the normal values, but for 25
patients they were still within the range of normal for the common duct and for 13 patients within the
range of normal for the common hepatic duct. Hamilton et al.37(6023) performed a similar study and
also found that the degree of overlap between the measurements in normal, postcholecystectomy
ducts without obstruction and those in obstructed ducts was so marked that an abnormality of the
ductal system could not be established by reference to ductal diameters alone.
It is usual to think of the biliary system as tapering in diameter from proximal to distal. Tapering is true
of the intrahepatic ducts (distal to proximal), but the extrahepatic system is relatively uniform in
diameter. However, Lasser et al.38(6024) found that although the diameter of the extrahepatic biliary
segments did not vary by more than 1 to 2 mm, a given segment was wider in about one-third of cases.
This was most often the prepancreatic portion of the duct, followed closely by the common hepatic
duct, and less frequently, by the intrapancreatic section.
REFERENCES
1.
Burckhardt H, Muller W. Versuche ber die Punction der Gallenblase und ihre
Rontgendarstellung. Dtsch Z Chir 1921;162:16897.
2.
Huard P, Do-Xuan-Hop. La ponction transhepatique des canaux biliaires. Bull Soc Med Chir l
Indochine 1937;15:1090100.
3.
Carter RF, Saypol GM. Transabdominal cholangiography. JAMA 1952;148:2535.
4.
Doubilet H, Mulholland JH. Intubation of the pancreatic duct in the human. Proc Soc Exp Biol
Med 1951;76:11344.
5. Kano T, Kurimoto K, Kasugai T, et al. Correlation between endoscopic retrograde
pancreatogram and postmortem pancreatogram. Endoscopy 1984;16:534.
6.
Dawson PM, Allen-Mersh TG. The anatomical relationship between the retropancreatic part of
the bile duct and the main pancreatic duct. Ann R Col Surg Engl 1983;65:18890.
7. Kasugai T, Kuno N, Kobayashi S, Hattori K. Endoscopic pancreatocholangiography. I. The
normal endoscopic pancreatocholangiogram. Gastroenterology 1972;63:21726.
8.
Classen M, Hellwig H, Rsch W. Anatomy of the pancreatic duct: A
duodenoscopic-radiological study. Endoscopy 1973;5:147.
9. Kang JK, Chung JB, Moon YM, Choi HJ. The normal endoscopic pancreatogram in Koreans.
Korean J Intern Med 1989;4:749.
10. Porter A, Warren G. The morphology of the main pancreatic duct at E.R.C.P. as a guide to its
demonstration by ultrasound. Australas Radiol 1982;26:14955.
11.
Cotton PB. The normal endoscopic pancreaticograms. Endoscopy 1974;6:6570.
12. Varley PF, Rohrmann CA, Silvis SE, et al. The normal endoscopic pancreaticograms.
Radiology 1976;118:295300.
13. Karak PK, Vashisht S, Tandon RK, Berry M. Normal endoscopic pancreatogram in an Indian
referral hospital. Indian J Med Res 1991;94:4269.
14.
Birnstingl M. A study of pancreatography. Br J Surg 1959;47:12839.
15.
Berman LG, Prior JT, Abramow SM, Ziegler DD. A study of the pancreatic duct system in man
by the use of vinyl acetate casts of postmortem preparations. Surg Gynecol Obstet
1960;110:391403.
16. Trapnell JE, Howard JM, Brewster J. Transduodenal pancreatography: An improved
technique. Surgery 1966;60:11129.
17.
Sivak MV Jr, Sullivan BH Jr. Endoscopic retrograde pancreatography. Analysis of the normal
pancreatogram. Dig Dis 1976;21:2639.
18. Howard JM, Short WF. An evaluation of pancreatography in suspected pancreatic disease.

19. Kreel L, Sandin B, Slavin G. Pancreatic morphology. A combined radiological and pathological
study. Clin Radiol 1973;24:15461.
20.
1963;50:48694.
21. Ishibashi T, Matsubara O. Studies on the retrograde pancreatography in autopsy specimens.
Bull Tokyo Med Dent Univ 1977;24:4351.
22. Kreel L, Sandin B. Changes in pancreatic morphology associated with aging. Gut
1973;14:96270.
23. Jones SN, McNeil NI, Lees WR. The interpretation of retrograde pancreatography in the
elderly. Clin Radiol 1989;40:3936.
24. Schmitz-Moormann P, Himmelmann GW, Brandes JW, et al. Comparative radiological and
morphological study of human pancreas. Pancreatitis like changes in postmortem ductograms
and their morphological pattern. Possible implication for ERCP. Gut 1985;26:40614.
25.
Newman HF, Weinberg SB, Newman EB, Northup JD. The papilla of Vater and distal portions
of the common bile duct and duct of Wirsung. Surg Gynecol Obstet 1958;106:68794.
26.
Rienhoff WF Jr, Pickrell KL. Pancreatitis: An anatomic study of the pancreatic and
extrahepatic biliary systems. Arch Surg 1945;51:20519.
27. Roberts-Thompson IC. Endoscopic retrograde pancreatography. Analysis of the normal
pancreatogram, and changes which are associated with chronic pancreatitis and pancreatic
cancer. Med J Aust 1977;2:7936.
28. Ogoshi K, Niwa M, Hara Y, Nebel OT. Endoscopic pancreatocholangiography in the
evaluation of pancreatic and biliary disease. Gastroenterology 1973;64:2106.
29.
Milbourn E. Calibre and appearance of the pancreatic ducts and relevant clinical problems. A
roentgenographic and anatomical study. Acta Chir Scand 1960;118:286303.
30. Anand BS, Vij JC, Mac HS, et al. Effect of aging on the pancreatic ducts: A study based on
endoscopic retrograde pancreatography. Gastrointest Endosc 1989;35:2103.
31.
MacCarty RL, Stephens DH, Brown AL Jr, Carlson HC. Retrograde pancreatography in
autopsy specimens. AJR 1975;123:35966.
32.
Chang VH, Cunningham JJ, Fromkes JJ. Sonographic measurement of the extrahepatic bile
duct before and after retrograde cholangiography. AJR 1985;144:7535.
33. Davies RP, Downey PR, Moore WR, et al. Contrast cholangiography versus ultrasonographic
measurement of the "extrahepatic" bile duct: A two-fold discrepancy revisited. J Ultrasound
Med 1991;10:6537.
34. Niederau C, Sonnenberg A, Mueller J. Comparison of the extrahepatic bile duct size
measured by ultrasound and by different radiographic methods. Gastroenterology
1984;87:61521.
35. Belsito AA, Marta JB, Cramer GG, Dickinson PB. Measurement of biliary tract size and
drainage time. Comparison of endoscopic and intravenous cholangiography. Radiology
1977;122:659.
36. O'Dwyer JA, Pemberton J, Thompson RP. Measurement of the endoscopic retrograde
cholangiogram. Dig Dis Sci 1981;26:5614.
37. Hamilton I, Ruddell WS, Mitchell CJ, et al. Endoscopic retrograde cholangiograms of the
normal and post-cholecystectomy biliary tree. Br J Surg 1982;69:3435.
38.
Lasser RB, Silvis SE, Vennes JA. The normal cholangiogram. Dig Dis 1978;23:58690.
39. Faris I, Thomson JPS, Grundy DJ, LeQuesne LP. Operative cholangiography: A reappraisal
based on a review of 400 cholangiograms. Br J Surg 1975;62:96672.
40. Mahour GH, Wakin KG, Ferris DO. Common bile duct in man, its diameter and
circumference. Ann Surg 1967;165:4159.
41.
Sachs MD. Routine cholangiography operative and post-operative. Radiol Clin North Am
1966;4:54770.
42. Chung SC, Leung JW, Li AK. Bile duct size after cholecystectomy: An endoscopic retrograde
cholangiopancreatographic study. Br J Surg 1990;77:5345.
43.
Okuda K, Someya N, Goto A, et al. Endoscopic pancreato-cholangiography. A preliminary

report on technique and diagnostic significance. AJR 1973;117:43745.
44. Moshal MG, Engelbrecht H. Technique and results of endoscopic retrograde
pancreaticocholangiography. A preliminary report on 140 patients. S Afr Med J
1975;49:21824.
Chapter 60 Endoscopic Papillotomy

(6025)
(6026)
MEINHARD CLASSEN, M.D.
PETER BORN, M.D.
Not many years after Langenbuch performed the first cholecystectomy, McBurney introduced surgical
sphincterotomy. Seventy-five years later, the first endoscopic papillotomies of the main duodenal
papilla were carried out simultaneously in Japan and Germany, and shortly thereafter, successful
endoscopic papillotomy (EPT) procedures were reported in the United States by Zimmon et al.1(6027)
and Geenen.2(6028) EPT*(6029) is now an established therapeutic procedure for various disorders of
the main papilla, the biliary tract, and the pancreas. Since the late 1980s, it has, in fact, become an
indispensable procedure of choice for many of these disorders throughout the world.
The number of applications of EPT has greatly increased since its introduction because of various
technical extensions of the procedure. Not only does EPT widen the normal or, in some cases, the
abnormally constricted opening of the bile and pancreatic ducts, but it also facilitates access to both
duct systems for further endoscopic treatment procedures such as biliary decompression by means of
nasobiliary tubes, insertion of an endoprosthesis, hydrostatic or pneumatic dilation of ductal stenoses,
and local irradiation of inoperable tumors by implantation of a radioactive source such as an
iridium-191 wire. Stones in the bile duct and in the pancreatic duct can be extracted with balloon
catheters or Dormia baskets. Chemical, mechanical, and hydraulic methods as well as lasers are used
to reduce the size of stones that are too large to be extracted by the usual means.
Indications for EPT

As a result of the continuing development of EPT, a shift has occurred in the pattern of endoscopic
retrograde cholangiopancreatography (ERCP) procedures. Although ERCP was used exclusively for
diagnosis between 1970 and 1974, this indication has been supplanted to a significant degree by EPT
and related procedures. Since 1979, there has been a continuing increase in the number of therapeutic
procedures, so that in many centers, these now account for 60% or more of the indications for ERCP.
A further shift in the reasons for EPT itself is also occurring: Whereas choledocholithiasis had hitherto
been the indication for EPT in 80% of cases, endoscopic management of bile duct stenoses by
insertion of a biliary endoprosthesis has become increasingly more common.
The indications for surgical sphincterotomy at the main duodenal papilla were standardized by
surgeons and pathologists years ago and still apply to EPT. However, the range of indications has
broadened (Table 601), so that the spectrum now includes virtually all causes of obstruction of the
biliary system as well as occlusion of the pancreatic ducts in some situations, such as a stone near the
papilla or pancreas divisum.
General Indications for

Endoscopic Papillotomy
TABLE 601
MAIN DUODENAL PAPILLA

Juxta-ampullary cancer
Benign papillary stenosis
General Indications for

TABLE 601
Sphincter of Oddi dysfunction

PANCREAS
Acute (biliary) pancreatitis
Pancreatic cancer
Chronic pancreatitis
Pancreas divisum
BILE DUCT
Cancer of the bile duct
Choledocholithiasis
Choledochocele
Suppurative cholangitis (sump syndrome)
Facilitation of endoscopic therapy (e.g., fistulas, strictures)
MINOR PAPILLA
Pancreas divisum
Chronic pancreatitis (better access to main pancreatic duct
via minor papilla in selected cases)
Choledocholithiasis is an indication for EPT and stone extraction. Our previous view, that EPT for this
indication should be restricted to patients over age 50 years in whom the surgical risk is high, has
proved to be overprotective because experience of more than 20 years since the introduction of EPT
has not revealed any negative long-term effects as a result of the loss of sphincter function. Even if
EPT causes a shift in the composition of bile and loss of the barrier function of the sphincter to
bacterial infection from the intestine, long-term observations indicate that these changes do not have
any clinical significance.
The presence or absence of the gallbladder is irrelevant when the indication for EPT is emergency
decompression of the bile duct. In this instance, removal of a stone obstructing the bile duct is of
primary clinical importance and substantially improves the condition of the patient. Whether
cholecystectomy is necessary after EPT is a controversial issue. Gallbladder symptoms sufficient to
indicate cholecystectomy occur in 10 to 20% of patients in whom the gallbladder is left in place during
follow-up periods of 5 to 10 years (see later).
The place of preoperative ERCP and EPT with stone extraction before cholecystectomy has been
reassessed in two prospective randomized trials.3,4(6030) Neoptolemos et al.4(6031) concluded that
patients fit for surgery should undergo open cholecystectomy and bile duct exploration, whereas those
who are elderly or at increased risk for surgical intervention should undergo EPT with stone extraction.
Because of problems with the design of these studies, these conclusions have been
challenged.5(6032) However, the role of ERCP/EPT in relation to surgical removal of the gallbladder
has changed substantially and will continue to evolve as laparoscopic cholecystectomy largely replaces
open cholecystectomy as the procedure of choice for removal of the gallbladder (see Chapter 56:
Indications, Contraindications, and Complications of Diagnostic Endoscopic Retrograde
Foreign material, such as vegetable matter and debris, and parasites may occasionally be found in the
biliary ducts. EPT with endoscopic extraction has been performed in many instances. The so-called
sump syndrome is a recognized complication, albeit rare, of side-to-side choledochoduodenostomy or
choledochojejunostomy when the distal nonfunctioning portion of the common bile duct becomes a
sump or well in which lithogenic bile, gastrointestinal contents, and debris accumulate. This may
produce obstruction of the enterobiliary anastomosis that can result in cholestasis and cholangitis. A
variety of endoscopic maneuvers including EPT have been utilized in the nonsurgical management of
the sump syndrome.
External and internal biliary fistulas have been treated successfully by EPT with or without ductal
stenting.6(6033) The number of patients undergoing the procedure for this indication appears to have
increased in relation to the growing use of laparoscopic methods for removal of the gallbladder (see
Disorders of the Liver Affecting the Bile Ducts).
Patients who are symptomatic as a result of sphincter of Oddi dysfunction are treated by EPT.7(6034)
Papillary stenosis is a long-established indication for papillotomy. This includes not only benign
stenosis of a short, circumscribed type but also obstructing papillary carcinoma.
The indication of EPT in acute pancreatitis depends on the demonstration of biliary ductal stones or
strictures. The absence of alcohol intake or other factors predisposing to acute pancreatitis is not
sufficient to presume stone impaction or passage as the cause of pancreatitis. However, the
demonstration of stones can be difficult because microlithiasis, a condition undetectable by
transcutaneous ultrasound, occurs frequently.8,9(6035) A motility disorder of the sphincter of Oddi is
found in approximately 50% of patients with relapsing pancreatitis; some of these patients respond to
EPT.10(6036)
An association between the congenital anomaly of pancreas divisum and recurrent episodes of
pancreatitis has been proposed (see Chapter 72: Congenital Anomalies of the Pancreas). One of the
possible explanations for such an association is that the minor papilla may be anatomically or
functionally inadequate to handle secretion from the larger portion of the pancreatic parenchyma.
Cannulation of the minor papilla can now be accomplished in at least 90% of patients with pancreas
divisum and may demonstrate pathologic findings in 20 to 70%. A careful study by Bernard et
al.11(6037) suggests that a subset of patients with pancreas divisum has an increased incidence of
acute pancreatitis. Despite some uncertainties concerning the relation between pancreas divisum and
recurrent pancreatitis, EPT of the minor papilla has been performed in patients with this anomaly.
EPT has been utilized in efforts to remove concrements from the main pancreatic duct in patients with
chronic pancreatitis and to facilitate placement of an endoprosthesis in main pancreatic duct stenoses
near the papilla (see Chapter 76: Endoscopic Management of Pancreatic Disease).12,13(6038)
Acute obstructive cholangitis is also an indication for EPT and EPT-related maneuvers such as stone
extraction, nasobiliary drainage, and insertion of a biliary endoprosthesis. The objective of endoscopic
drainage is rapid decompression of the biliary system in combination with resuscitation and antibiotic
treatment.14(6039) The technical approach will differ according to the cause of obstruction. The most
common causes are choledocholithiasis, malignancy, postoperative strictures, and parasites.
Techniques and Methods

Equipment
All duodenoscopes with lateral optics are, in principle, suitable for EPT. Over the years, however,
developments in the instrument have substantially improved its capabilities, including increases in the
viewing angle, in the range and tightness of the deflection angle of the instrument tip, and in the
diameter of the accessory channel. The most recent models have instrument channels up to 4.2 mm in
diameter; these are particularly well suited for implantation of an endoprosthesis.
The Demling-Classen papillotome (pull-type) is now in general use (Figure 601A). The basic
construction of this device is as follows: The outer Teflon catheter contains a thin steel wire that exits
the catheter about 3 cm before its distal end and reenters the catheter about 3 to 5 mm from its tip.
Tension applied to the wire produces a bowing effect in the catheter, the wire forming the bowstring.
When the device has been properly placed in the papilla and bile duct, the exposed wire functions as a
knife when high-frequency electrosurgical current is applied. The tip of the catheter, which has no wire,
serves as a pathfinder during cannulation of the bile duct. Occasionally, a papillotome, when inserted
through the duodenoscope, will not be in proper or desired orientation in relation to the papilla. This is
sometimes encountered when the device has not been used previously, although it is now less of a
problem with better quality control by manufacturers. When this occurs, however, it is usually possible
to reshape the cannula to achieve a more desirable orientation to the papilla. This can be
accomplished by repeated twisting, bending, and pulling the end of the cannula through the fingers, a
process sometimes referred to as grooming.15(6040)
(6041)Figure 601. Diagram of different papillotomes. A, Standard Erlangen papillotome of

Classen and Demling. B, Needle-knife (Mori). C, "Billroth II" papillotome. D, Push
papillotome. E, Shark-fin papillotome.
There are numerous variations of the basic Demling-Classen papillotome. One modification is to
shorten the length of the exposed portion of the wire to 15 or 20 mm. This papillotome is useful for
making short incisions ("minicutting") and for "precutting"; the latter technique is used when deep
cannulation and standard placement of the papillotome cannot be achieved. A special piercing knife or
"needle-knife" papillotome is also available (see Figure 601B). This instrument consists simply of a
catheter and wire; the end of the wire can be extended 2 or 3 mm from the distal tip of the catheter.
The needle-knife can be used to puncture and incise the roof of the duodenal papilla, a procedure
termed fistulotomy. "Push-type" papillotomes (see Figure 601D; including the so-called shark-fin
variant [see Figure 601E]) are those in which the wire is pushed out to form a bow. Push-type
papillotomes are used uncommonly because the direction of the incision produced with this device is
unpredictable. All papillotome catheters should have a side port to allow for injection of contrast
material.
Most currently available high-frequency electrosurgical generators with cutting and coagulation modes
are suitable for EPT. The machine should have a "rapid-start" function that helps to initiate the incision.
The output of an electrosurgical generator can fluctuate according to the electric resistance it
encounters. Machines with the rapid-start function have electronic circuitry that provides an initial
high-power current for a fraction of a second to ensure that the electrosurgical incision will begin no
matter what electric resistance is encountered (see Chapter 9: Principles of Electrosurgery). If a
machine without this capability is used, the onset of cutting may be delayed and unnecessary heating
of the papillary area may occur.
Dormia baskets (Figures 602 and 603) and balloon catheters (Figure 604) are used to extract
stones from the bile duct. Additional instruments for lithotripsy and litholysis are described in a following
Methods section.
(6042)Figure 602. Retrograde cholangiogram shows the introduction of a Dormia basket

into the bile duct. A stone has been captured in the basket.
(6043)Figure 603. A, Dormia basket catheter. B, Close-up of Dormia basket.
(6044)Figure 604. A, Balloon catheter (American Edwards Co.) for extraction of bile duct
stones. B, Fully inflated 1-cm-diameter balloon of balloon extraction catheter.
Preparation
As with every invasive procedure, the patient, and when appropriate, his or her relatives, must be
informed beforehand of the goal(s) and nature of the procedure, the premedication required, the risks
and potential complications, and when appropriate, the alternatives to endoscopic treatment. Because
serious complications may occur, the endoscopist must have the close cooperation of an experienced
abdominal surgeon. A coagulation profile should be obtained, and any coagulopathy should be
corrected before the procedure. Methods of preparation and sedation are broadly similar to those used
for diagnostic ERCP (see Chapter 57: Technique of Endoscopic Retrograde
Although hemorrhage is a recognized and potentially life-threatening complication of EPT, units of
blood are crossmatched before the procedure only in isolated instances. Endoscopic Doppler
ultrasonography has been used to locate the retroduodenal artery in the wall of the duodenum along
with any aberrant branches of this vessel in the region of the papilla.16,17(6045) Theoretically, this
technique should help the endoscopist to avoid arterial vessels as the EPT incision is made, and it is
believed to reduce the risk of life-threatening hemorrhage. However, in our prospective randomized
trial of suprapapillary Doppler ultrasound in patients undergoing EPT, the results obtained with this
device were disappointing.18(6046)
We administer antibiotic prophylaxis before the procedure only in isolated instances, although there are
differences of opinion as to the need for this measure (see Chapter 56: Indications, Contraindications,
and Complications of Diagnostic Endoscopic Retrograde Cholangiopancreatography).
Because of the risk and potential for serious complications, EPT is performed as an inpatient
procedurein an institution capable of dealing with any and all potential complications. There are
reports in which selected individuals have been treated by EPT on an outpatient basis.19,20(6047)
However, data on the safety of this practice are extremely limited (see later). As with all endoscopic
procedures, high-level disinfection of endoscopes and all accessories is imperative (see Chapter 8:
Disinfection of Endoscopes and Accessories).
Methods
Except for minor modifications, the technique of EPT has remained virtually unchanged since it was
first described. After endoscopic retrograde cholangiography, the standard papillotome is introduced
into the bile duct (Figure 605) while its position at the papilla and within the duct is monitored
endoscopically and fluoroscopically (Figure 606). If there is any doubt whether the papillotome has
been placed in the bile duct, a small amount of contrast medium can be instilled through the
papillotome for confirmation. The proximal aspect or roof of the papilla should "ride" up and down on
the cutting wire of the papillotome when tension is exerted on the wire because this ensures that the
cut will correspond to the course of the bile duct and that the tension on the wire is not excessive.
About one third of the length of the tensed cutting wire should be visible outside the papilla. This
ensures that the cutting process will be controlled and that cutting will proceed by increments (Figure
607). If the papillotome is introduced too deeply into the bile duct, the cut cannot be controlled and an
abrupt incision of excessive length may occur, with the attendant danger of perforation or hemorrhage,
or both. The commonest error is to have too much wire inside the papilla and to apply too much bowing
tension. In this circumstance, it is likely that cutting will not occur promptly, and there is then a
temptation to increase the current settings and tension on the wire. Eventually and suddenly, this will
result in a cut through the coagulated area at alarming speed, with an attendant risk of significant
bleeding.21(6048)
(6049)Figure 605. Endoscopic view of a papillotome in the correct position in the bile duct
with proper tension applied to the wire. The roof of the papilla "rides" on the tensed wire. The
cutting procedure can be initiated in this position.
(6050)Figure 606. Radiographic view of a papillotome correctly introduced into the bile
duct; one third of the length of the exposed wire is in the duodenum and outside the bile duct.
(6051)Figure 607. Endoscopic view of the endoscopic papillotomy (EPT) procedure. A,

Papilla before papillotome placement. B, Papillotome in place. C, Appearance immediately
after EPT.
The correct length of the incision varies. Usually, this will be about 10 to 15 mm when the papilla itself
is normal. With short pulses of current, the roof of the papilla is incised as far as the point at which it
enters the wall of the duodenum. In most cases, this transition point can be readily seen endoscopically
(see Figure 607). Dilation of the distal aspect of the bile duct may produce a bulge in the duodenal
wall proximal to the papilla. When observable, this bile duct infundibulum usually defines the course of
the bile duct endoscopically. If, however, the papilla is small and flat and the bile duct is not dilated, it
may be more difficult to determine the length to which the cut can be made safely. The position of the
bile duct in relation to the duodenal wall as observed fluoroscopically is also of assistance in gauging
the length of the incision. If the duct lies close beneath the wall and is dilated, a relatively long cut can
be made safely as a general rule. If the duct is not dilated and, for the most part, is positioned some
distance from the duodenal wall on fluoroscopy, then the length of the incision must be reduced and
there is less margin for error.
In up to 10% of cases, a small incision (precut, minicut) must be made first, so that the standard
papillotome can be selectively introduced into the bile duct to a sufficient depth. The modified version
of the papillotome with a shorter exposed cutting wire section is first introduced as far as possible into
the papilla. Then the roof of the ampulla is incised as far as possible. If, as a result, the bile-colored
mucosa of the biliary duct can be seen, the precut papillotome can be exchanged for the standard
papillotome, whereupon selective cannulation of the duct is usually successful. We frequently use the
needle-knife as another method of minicutting. The papilla is incised from its meatus along its roof in
the direction of the bile duct. The papilla must be opened cautiously, using small incisions. When the
ampulla is reached, the papillotomy is completed using the standard papillotome. When there is
anatomic distortion involving the main duodenal papilla and orientation of the cutting device is difficult,
it may be possible to insert a small stent that can then be used to guide the incision.22(6052) These
precut maneuvers require considerable expertise and experience if a satisfactory degree of safety is to
be maintained. Precutting is more dangerous than the standard papillotomy. In general, it should not
be used to gain access to the bile duct purely for diagnostic purposes. Special papillotomy techniques
are discussed in Chapter 61.
EPT has also been performed using lasers. The first report was that of Sander and Poesl,23(6053) in
which they describe two patients in whom successful neodymium:yttrium-aluminum-garnet laser
papillotomies were carried out when other treatments were deemed not to be effective.
Another endoscopic technique for incision into the common bile duct is endoscopic fistulotomy
(choledochoduodenostomy).24,25(6054) This method is sometimes suitable when there is obstruction
at the orifice of the main papilla. It is necessary that the distal bile duct produce a prominent impression
on the wall of the duodenum that can be seen endoscopically proximal to the papilla. The prominence
of this intramural segment is usually the result of dilation of the distal bile duct proximal to the point of
obstruction. This dilated segment is also a necessary factor if the procedure is to be accomplished
safely. Furthermore, it is advisable to inject contrast medium into the bile duct to demonstrate dilation
and proximity of the bile duct to the wall of the duodenum. If these prerequisites are present, the
intramural segment is simply punctured using an electrosurgical device such as a needle-knife.
Depending on the anatomic circumstances in the individual case, the incision can be lengthened using
a standard or precut papillotome. As with the precut techniques, endoscopic choledochoduodenostomy
should be undertaken only by the most expert endoscopists.
In contrast to previous opinion, there is no increased risk of perforation in patients with juxtapapillary
diverticula.26(6055) Usually a diverticulum does not present any particular technical difficulty.
Occasionally, the papilla will be entirely within a diverticulum and cannulation may be difficult or
impossible. Even if the papillotome is placed in the bile duct, the deformed anatomy may make it
difficult to control the incision. If the papilla is at the distal margin of a shallow diverticulum, a ridgelike
structure may bisect the diverticulum from proximal to distal to terminate at the papilla. This is usually
the bile duct. EPT may actually be easier in this case because the course of the bile duct is clearly
visible and the duct is close to the duodenal lumen. The possible relationship between juxtapapillary
diverticula and bile duct stones is under continuous discussion (see Chapter 62: Calculus Disease of
the Bile Ducts).
The papilla is occasionally difficult to find in patients who have undergone Billroth II operations.
Moreover, once the papilla is located, the papillotome frequently cannot be properly placed for cutting
because of the reverse orientation. Even the use of an endoscope with forward-viewing optics does not
always resolve this problem. Papillotomes with a sigmoid shape (reverse bow) are designed for use in
these patients (see Figure 601C).27(6056) These devices are believed to reduce the risk of injury to
the pancreas.
The guidewire papillotomy technique is helpful when difficulty has been encountered in cannulation of
the bile duct with the standard diagnostic catheter. Once selective cannulation of the bile duct has been
achieved, a 0.035-inch-diameter guidewire is passed through the cannula, which is then removed
leaving the guidewire in the duct, and the papillotome is passed over the wire. A properly insulated
guidewire must be used if the papillotomy is to be performed with the wire in place.28,29(6057)
Double-lumen papillotomes have two insulated channels that make it possible to maintain deep biliary
cannulation with the guidewire during EPT without the risk of the current being conducted through the
guidewire. These papillotomes are larger and less flexible, and available data indicate that use of a
single-lumen papillotome together with a plastic-coated (insulated) guidewire is safe and
effective.30(6058)
Endoscopic Treatment of Gallstones
Most gallstones are discharged spontaneously from the bile duct after EPT.7(6059) Nevertheless, it is
advisable to extract as many calculi as possible using a Dormia basket or a balloon extraction catheter
(see Chapter 62: Calculus Disease of the Bile Ducts). Before papillotomy, the size of the stone should
be estimated and related to the possible length of the incision. On occasion, extraction will not be
successful until some days have passed and the edema produced by the papillotomy has regressed. If
a stone is larger than the papillotomy incision, distal impaction of the stone followed by septic
cholangitis may occur either spontaneously or during attempts to remove the stone endoscopically.
Spontaneous stone impaction with biliary obstruction can be prevented by insertion of a nasobiliary
tube (drain) or an endoprosthesis.31(6060) The nasobiliary drain can also be used for cholangiography
or to infuse saline solution or solvents for stone dissolution. When the condition of the patient is stable
and a few days have elapsed, further attempts can be made to remove residual stones. Particularly in
older or frail patients, stent placement guarantees the adequacy of long-term treatment. The EPT
incision can often be extended, but this should not be done before 1 week has elapsed. EPT results in
hypervascularization of the papilla that increases the risk of hemorrhage if an attempt is made to
extend the incision shortly after the first procedure.32(6061) If the stone has already become impacted
during attempted extraction with a Dormia basket, then a nasobiliary tube is advanced beyond the
entrapped basket into the bile duct to ensure outflow of bile. Another attempt at extraction after a few
days is frequently successful. Lithotripsy or chemical litholysis can reduce the size of some stones that
are unsuitable for extraction because of their large size or because of limits on the maximum possible
length of the incision. Once such stones are rendered smaller by dissolution or fracturing, the
fragments can usually be removed by Dormia basket or balloon catheter extraction. Otherwise,
surgical removal of the stone is necessary.
Electrohydraulic and mechanical methods are available for lithotripsy. In mechanical lithotripsy, the
stone is captured in an appropriate Dormia basket and is then broken by forcefully closing the basket.
This requires that the basket be made of particularly strong wires. When the stone is trapped, the
endoscope is removed from the patient. This requires that the handle or proximal mechanism of the
basket be removed by cutting the wire. The outer Teflon sheath of the Dormia basket catheter is also
removed, so that only the wire remains. The basket now remains around the stone in the bile duct, and
the opposite end of the wire exits via the patient's mouth. A flexible metal rod is then passed over the
wire and is advanced to the basket and stone under fluoroscopic guidance (Figure 608). Once the rod
is in proper position, the Dormia basket wire is attached to a knurled wheel drive. This allows the
operator to forcefully close the basket by winding the wire onto the cylinder of the lithotripsy device.
Essentially, the wire is pulled into the flexible metal rod, and as the basket is pulled into the rod, it is
forced to close (Figure 609). This causes the stone to disintegrate into two or usually several
pieces.33(6062) Even stones that are trapped in a Dormia basket at the papilla can be crushed in this
manner.34(6063) Occasionally, it is possible to introduce a mechanical lithotriptor into the bile duct to
crush a stone without previous EPT. Fragments must then be extracted through the intact
papilla.35(6064) Lithotripsy devices for insertion through the endoscope are available; these obviate
the need to cut the basket wire and remove the endoscope.
(6065)Figure 608. Mechanical lithotriptor with a flexible wire sheath to contain Dormia
basket wire.
(6066)Figure 609. Diagram of the steps involved in mechanical lithotripsy. Left to right, A
large stone is captured in a strong Dormia basket; the endoscope is removed, leaving the
basket in place; a flexible wire sheath (Figure 608) is advanced over the basket; and the
basket is forcefully withdrawn into the sheath, thereby fracturing the stone.
There are a number of reports of lithotripsy using an electrohydraulic probe.3639(6067)
Electrohydraulic lithotripsy (EHL) employs an electrode device that is placed in the bile duct near the
stone to be broken. A high-voltage electric discharge from the electrode produces a steep
high-pressure hydraulic wave in a fluid medium that fractures the stone.40(6068) In our experience,
EHL applied under careful endoscopic control has proved to be safe. Nevertheless, EHL should still be
considered investigational at the present time.
A flash lamp excited dye laser with a wavelength of 504 mm (MDL 2000, Candela, Boston, MA) has
been used with a high rate of success in patients with large bile duct stones. The laser light was
applied through a quartz fiber under direct vision with choledochoscopes passed percutaneously,
through a "baby" cholangioscope, or under fluoroscopic guidance at standard endoscopy.41,42(6069)
The newest pulsed rhodamine laser (Telemit Electronic, Munich, Germany), wavelength 594 nm,
interrupts the pulses within nanoseconds when the back-scattered laser light indicates that the quartz
fiber (300 nm in diameter) is not in contact with the stone.43(6070)
Other methods for fragmenting stones in the bile duct, such as water-jet cutting44(6071) and
ultrasound,45(6072) are in preclinical stages of development.
A variety of chemical agents have been used in efforts to dissolve biliary stones (chemical litholysis).
Because an ideal agent has not yet been discovered, this method has not gained wide acceptance.
Bile salt solutions have been used for the chemical dissolution of cholesterol stones in the bile duct.
However, 40% of recurrent stones after cholecystectomy are calcium bilirubinate and not cholesterol
stones. Methyl tert-butyl ether (MTBE) has also been used to dissolve stones in the bile duct.46(6073)
Although this technique might have a role in the treatment of large retained stones, it has some
significant technical problems and serious potential complications (see later).
Results
Success Rates
Success rates for EPT, as determined by survey, have been reported from several countries (Table
602).4758(6074) These reports pertain mainly to experience through the end of the 1970s, and
choledocholithiasis was the indication for EPT in the majority of cases. Success rates for EPT from
selected single centers or individuals are shown in Table 603.4,19,5366(6075) When performed by
an experienced operator, EPT is successful in 90 to 95% of patients; in 80 to 95%, the bile duct can be
cleared of stones.52(6076)
TABLE 602
Surveys of Endoscopic Papillotomy Success Rates, Complications, and Mortality
During the 1970s

COUNTRY AND REFERENCE
Centers (n)
Year reported
Patients (n)
Success (%)
Complications (%)
Mortality (%)
GERMANY49
JAPAN47
ITALY50
ENGLAND48
9
1978
955
92.1
7.3
1.7
25
1979
468
96.5
8.5
0.4
8
1979
239
81
6.7
0.5
14
1981
679
87
8.5
1.0
UNITED STATE
Success Rates, Complications, and Mortality of Endoscopic Papillotomy

in Selected Series
TABLE 603
PROCEDURES
(n)
SUCCESS
(%)
COMPLICATIONS
(%)
Safrany, 197754
265
92.0
10.0
1.2
Koch et al., 197755
267
95.0
7.1
0.8
Viceconte et al., 198153

Siegel, 198156
296
86.1
7.0
0.8
267
96.6
5.0
0.77
Wurbs, 198258
808
9499*
7.3
1.4
Escourrou et al., 198457

Leese et al., 198559
443
92.0
7.0
1.5
394
98.0
10.4
0.8
STUDY
55
96.4
9.1
Neoptolemos et al., 19874
DEATH
(%)
Success Rates, Complications, and Mortality of Endoscopic Papillotomy

in Selected Series
TABLE 603
STUDY
PROCEDURES
(n)
Ikeda et al., 198860

Vaira et al., 198961
SUCCESS
(%)
COMPLICATIONS
(%)
DEATH
(%)
469
99
6.3
0.4
1000
97.5
6.9
0.4
Podolsky et al., 198919

Lambert et al., 199163
137
100
6.6
602
91.5
10.5
2.2
Hill et al., 199164
218
96.8
7.6
0.9
Moreira et al., 199165

Sherman et al., 199166
18
94.4
13.3
6.6
423
6.9
1.7
9.8
0.43
5.4
0.3
Freeman et al., 1996125

Coppola et al., 1997129
2347
546
98
* Varies with experience.
Includes only patients with cirrhosis.
Successful EPT in 25 of 35 patients who had previously undergone a Billroth II gastrectomy was
reported by Safrany et al.67(6077) in 1980. A complication occurred in each of 2 patients. The authors
described the technical difficulties of the procedure and suggested that the complication rate might be
higher than that associated with EPT in patients who had not undergone this type of surgery.
Subsequent reports of experience with ERCP and EPT after Billroth II surgery were consistent with that
of Safrany et al.6769(6078) However, more recent reports indicate that success rates in the range of
80 to 92% are possible with experience, together with the use of improved instruments and technical
methods.70(6079) The technical aspects of EPT in Billroth II patients as well as success rates and
complications of the procedure in these patients are discussed in Chapter 61: Special Papillotomy
Techniques.
Choledocholithiasis
Treatment of choledocholithiasis is discussed in detail in Chapter 62: Calculus Disease of the Bile
Ducts. Also, a discussion of endoscopic approaches other than the usual peroral route for the
management of biliary stones may be found in Chapter 71: Choledochofiberoscopy and Endoscopic
Lithotripsy.
EPT is successful in 90 to 98% of cases of choledocholithiasis. A second session is necessary for
about 8% of patients; a third session in less than 1%. With respect to clearing the bile duct of stones,
the success rates published by groups working in Great Britain, Japan, and Germany are about 85%.
For example, in series of 602 patients reported from a single center in England, EPT was performed
successfully in 91.5% and stone clearance was achieved in 81.6% of patients.63(6080) A mean of 1.9
procedures per patient was necessary to attain these results; 30-day mortality was 2.2%. Failure to
clear the duct can be accounted for by factors such as anatomic abnormalities of the duodenum, prior
surgery such as Billroth II resection, abnormalities of the papilla as, for example, extreme stenosis, and
prior biliary surgery. In addition, the large bile duct stone still presents a technical problem.
Gallstones are impossible to extract, as a rule, only if they are too large, are in the intrahepatic ducts,
or are located proximal to a biliary stricture. It is virtually always possible to extract stones with a
diameter less than 10 to 15 mm. The difficulties increase with stones above 15 mm in diameter,
although every experienced endoscopist has encountered spontaneous discharge of concrements

larger than 25 mm.71(6081) Notwithstanding the possibility of spontaneous clearance, attempted
extraction of stones wider than 15 mm in diameter can be difficult, and success is less predictable.
Naturally, there is a limit to the length of a sphincterotomy that can safely be made, and larger stones
must be reduced before they can be removed.
A variety of methods have been devised to deal with large biliary stones (see earlier). Probably the
most commonly used and readily available technique is mechanical lithotripsy. Usually, the ducts can
be cleared with this method in about 80% or more of patients with stones too large for extraction by
conventional means, although high success rates have been reported.72(6082) Mechanical lithotripsy
fails in some patients in whom large stones cannot be captured in the basket because there is no
space within the duct to open the basket.73(6083)
EHL has been employed endoscopically to clear the biliary system of stones (see earlier). For
example, Siegel et al.74(6084) utilized an EHL device with a balloon incorporated proximal to the
electrodes for proper positioning within the bile duct. EHL was attempted using this electrohydraulic
lithotriptor in 21 patients; stones ranging in diameter from 1.0 to 1.5 cm were successfully fractured
and cleared from the ducts of 18 patients (86%) with only 1 minor complication (pancreatitis). The
success rate for stone clearance by transhepatic cholangioscopic EHL was high in the series of Bonnel
et al.,38(6085) but severe complications (e.g., hemobilia requiring transfusions) were encountered.
However, this may have been due to use of a rapid technique for formation of the percutaneous track.
Biliary stones can be fragmented using lasers, although the types of lasers suitable for this purpose are
relatively expensive and not widely available. We performed laser lithotripsy under cholangioscopic
control in 35 patients in whom conventional ERCP with extraction or mechanical fragmentation had
failed. Whenever possible, peroral cholangioscopy employing a mother-baby endoscope system was
used; this was possible in 12 patients. In the remaining 23 patients, percutaneous cholangioscopy was
performed. A pulsed-dye laser was used in 32 patients, an alexandrite laser (HMT Alexantriptor, High
Medical Technologies, Kreuzlingen, Switzerland) in 2 patients. Complete disintegration of stones was
achieved in 33 patients. Peroral cholangioscopy failed in 2 patients; percutaneous cholangioscopy was
performed in 1 and the other underwent surgery. No complication after laser application was
encountered, but the establishment of the cutaneobiliary fistula resulted in a complication in 8 of 23
patients (fever, transient hemobilia, and subcapsular liver hematoma).42(6086) In another study, the
pulsed rhodamine laser (wavelength 594 nm), proved to be rapid, efficient, and safe when used under
cholangioscopic control (n = 26) and even when applied under fluoroscopic control (n = 3).44(6087)
Chemical litholysis offers the prospect of dissolution or reduction of large stones by a relatively less
invasive and less traumatic means. Leuschner et al.75(6088) reviewed the experience up to about
1982 with chemical dissolution (litholysis) of stones in the bile duct. They concluded that the overall
success rate was about 50% and that side effects occurred in the majority of patients.
Because 40% of recurrent stones after cholecystectomy are calcium bilirubinate and not cholesterol
stones, our group investigated the use of alternating infusions of bile acid-ethylenediaminetetraacetic
acid (EDTA) and glycerol monoctanoate solutions.75(6089) During the period of study, most stones
(87%) were extracted by EPT and the usual methods. Surgery was required in 4.2% of cases.
Chemical litholysis was undertaken in 17 patients (8.8%). In 8 (47%), this resulted in dissolution of the
stones or reduction in size with subsequent extraction. Nausea and diarrhea occurred in 60% of the
patients. Histologic examination of the gallbladders of patients subjected to operation showed acute
ulceration and mild chronic inflammation. Better results are reported by Swobodnik et al.,76(6090) but
at the onset of treatment, the stones were smaller in this series than in ours.
Another cholesterol solvent, MTBE, has been used to dissolve stones in the common duct. MTBE was
instilled through a nasobiliary or percutaneous catheter directly into the common bile duct with
subsequent clearance of stones in 10 of 12 patients in the study of Kaye et al.46(6091) There are,
however, some practical difficulties with administration and tolerance because the extrahepatic bile
duct is not a closed compartment and the contact time of the solvent with the stone is short. Moreover,
the instillation and absorption of higher volumes of the ether may lead to general anesthesia.
Preliminary work with butyl ether also indicates that the search for an appropriate agent for chemical
litholysis of common bile duct stones still continues.
EPT is effective in the treatment of patients with retained ductal stones and a T-tube in situ. In those
whose ductal stones are above or astride the T-tube, the latter may have to be removed to achieve bile
duct clearance.77(6092) Cotton78(6093) reviewed available publications pertaining to this question
and recommended percutaneous extraction for stable patients who are willing to wait some weeks until
the percutaneous track is mature and stable.
Foreign Bodies
A foreign body in the bile duct is a rarity. After choledochoduodenostomy, retention of food in the biliary
tract, particularly in the distal stump, may occur (sump syndrome, cholangiophytiasis). EPT and
endoscopic methods are effective in eliminating this condition.7981(6094)
Parasites in the bile duct or pancreatic duct (e.g., Ascaris lumbricoides and Fasciola hepatica) are rare
in the Western world but are more frequent in Middle and Far Eastern countries.82(6095) They may
cause obstructive jaundice or pancreatitis and should be extracted with a Dormia basket after
EPT.83,84(6096)
Gallstone-Induced (Biliary) Pancreatitis
The passage of a stone through the main duodenal papilla with temporary blockade of the outflow of
secretions from the pancreas is the most frequent cause of acute pancreatitis.85(6097) Persistence of
the stone in the papilla (Figure 6010) or the discharge of many stones in a short time may lead to
life-threatening pancreatitis with its attendant complications. Gallstone-induced pancreatitis is
discussed in Chapter 62: Calculus Disease of the Bile Ducts.
(6098)Figure 6010. Endoscopic view of the main duodenal papilla in biliary pancreatitis.
The papilla in distended and inflamed. Two ulcers caused by pressure from the stone are seen
in the orifice and the roof of the papilla. The surface of an intrapapillary stone can be seen in
the ulcerated areas.
Pancreas Divisum
Since the first report by Cotton86(6099) in 1980 that pancreas divisum may be causally related to
recurrent pancreatitis or recurrent pain, over 120 articles have appeared supporting or refuting this
contention.87(6100) The major issues are whether pancreas divisum is associated with an increased
likelihood of recurrent episodes of acute pancreatitis, and whether EPT of the minor duodenal papilla,
with or without stenting of the dorsal pancreatic duct, reduces the risk of pancreatitis or episodes of
pancreatic-like pain. Some authorities share the opinion of DiMagno and Holtmann88(6101) that
pancreas divisum does not cause pancreatitis and that endoscopic attacks on the minor papilla or duct
are not warranted. Cotton89(6102) concluded that dorsal duct sphincterotomy alone is successful in
achieving long-term freedom from recurrence of acute pancreatitis associated with pancreas divisum.
Probably the majority share the opinion of Carr-Locke:87(6103) "Despite the somewhat confusing and
opinionated messages from the literature, there seems little doubt that in certain patients there is a
causal relationship between pancreas divisum and pancreatitis symptoms and many carefully selected
for therapy do benefit, at least in the short-term. Hopefully in time we shall have well defined
parameters to characterize such patients and offer them the most appropriate treatment." Pancreas
divisum is discussed in detail in Chapter 72: Congenital Anomalies of the Pancreas.
Lans et al.90(6104) performed a prospective randomized controlled clinical trial and placed stents
across the minor papilla in patients with pancreas divisum and at least two attacks of acute
pancreatitis. Dilation of the minor papilla allowed full insertion of the endoprosthesis into the dorsal
pancreatic duct in 83% of the patients. Patients in the treatment group had no hospitalizations or
emergency room visits for abdominal pain. Pancreatitis, documented as twice normal elevation of
serum amylase, occurred in 1 of 10 patients in the stent group and 7 of 8 in the untreated control
group. Improvement in symptoms was also significantly superior in the stent group. In addition, 4 of 9
patients in the control group did poorly and subsequently had stent placement; no further episodes of
pancreatitis were observed and none required hospitalization. Thus, endoscopic treatment resulted in a
significant subjective and objective clinical improvement compared with controls in this short-term
follow-up study lasting approximately 30 months. The results suggest that impairment of pancreatic
flow through the minor papilla may be an important causative factor for recurrent pancreatitis in
patients with pancreas divisum.
In general, endoscopic therapy in patients with pancreas divisum has been most successful in those
with recurrent episodes of acute pancreatitis. The results of the procedure have been less compelling
in patients with pain in the absence of documented episodes of pancreatitis. Long term-results have
been least satisfactory in patients with chronic pancreatitis and pancreas divisum (see Chapter 72:
Congenital Anomalies of the Pancreas).
A variety of endoscopic treatment methods including papillotomy, stone extraction, stent placement,
and pseudocyst drainage have been utilized in selected patients with chronic calcifying pancreatitis for
the treatment of pancreatic pain or restoration of pancreatic exocrine function.91(6105) Pancreatic duct
stones have been fragmented by extracorporeal shockwave lithotripsy or laser energy and have been
subsequently removed from the ducts with balloons or baskets.92,93(6106) Data concerning
endoscopic treatment of pancreatic duct stones and other structural abnormalities commonly
associated with chronic pancreatitis are limited, and additional controlled clinical studies are needed in
this area (see Chapter 76: Endoscopic Management of Pancreatic Disease).
Obstruction of the Main Duodenal Papilla
Benign Papillary Stenosis
Benign papillary stenosis is a logical indication for EPT (see also Chapter 68: Sphincter of Oddi
Stenosis and Dysfunction). There is a wide range in the reported incidence of benign papillary
stenoses in endoscopic series, varying from 0.04 to 28.4%; the incidence in postmortem studies
ranges from 0.04 to 0.12%.94(6107) Papillary stenosis was found 363 times (2.7%) in 13,300
diagnostic ERCP examinations at five gastroenterologic centers in Germany. Calculus disease of the
bile ducts occurred as an associated pathologic finding in 61% and was considered to be the most
important associated or causative factor. Papillary stenosis was attributed to surgical manipulation at
the main duodenal papilla in 7%, to papillary carcinoma in 14%, and to a juxtapapillary diverticulum in
1%.
According to the collected statistics reported in 1982 by Seifert et al.,95(6108) circumscribed short
papillary stenosis is the indication for papillotomy in 10.6% of all cases. EPT may be technically difficult
in true papillary stenosis. It may be difficult to place the standard papillotome properly, and the fibrotic
papilla may resist electrosurgical cutting. Moreover, recurrent stenosis after EPT for papillary stenosis
is much more common (11.5% of cases) than after EPT for other indications. In addition, the
complication and mortality rates for EPT in papillary stenosis without choledocholithiasis are higher
than with EPT for other indications (see later).66(6109)
Neoplastic Papillary Obstruction
Benign papillary tumors are rare, being found in 1.5% of our patients who underwent EPT. However,
the frequency of this diagnosis may increase as endoscopic methods become more widespread (see
also Chapter 69: Tumors of the Main Duodenal Papilla).
Papillary carcinoma is the third most common tumor causing obstructive jaundice (9 to 13% of cases),
after carcinoma of the head of the pancreas and of the bile duct. However, it must be given special
emphasis among tumors that obstruct the bile duct because it is often resectable, with a low operative
mortality and a favorable survival of about 2 years on average.
In advanced cases, carcinoma of the main papilla may be obvious endoscopically as an exophytic
growth in the duodenum. Biopsy specimens from such a lesion will usually be positive. However, the
diagnosis of small tumors originating from inside the ampulla (ampullomas) is particularly difficult. The
duodenal aspect of the papilla may be normal or enlarged. Careful fluoroscopic study of the emptying
of contrast medium from the bile duct and pancreatic duct may show defective, irregular filling of the
ampulla. A small papillary tumor, in addition to passage of small stones, should be considered in the
differential diagnosis of every case of papillary obstruction.
Malignant papillary tumors are sometimes difficult to diagnose even with available endoscopic
methods, including biopsy (see Chapter 69: Tumors of the Main Duodenal Papilla). Inability to obtain
an adequate specimen (or specimens) for pathologic study is a significant problem. From the
pathologic standpoint, the difficulties associated with the histologic diagnosis of papillary carcinoma
derive from the juxtaposition of benign and malignant elements that is typical of this tumor. Thus,
benign hyperplasia and adenomas may be found in association with carcinoma. Although a large
particle biopsy by means of an electrosurgical snare wire improves diagnosis, in our experience it is
more important to open the papilla by EPT and then obtain a large number of specimens from all
exposed areas in the inner aspect. In 55 cases of periampullary carcinoma reported by Bourgeois et
al.,96(6110) the diagnosis was confirmed histologically in biopsies of the papilla before EPT in only
50% of the cases. When biopsy specimens were obtained after EPT, the diagnosis was confirmed in
all cases. These investigators also obtained biopsy specimens within 48 hr of EPT in 22 cases of
benign biliary tract disease and found cellular atypism in these specimens.
Thus, EPT has both diagnostic and therapeutic roles in the management of carcinoma of the main
duodenal papilla (see Chapter 69). EPT is frequently necessary for accurate diagnosis by biopsy when
the tumor is small and confined within the papilla. It has not been shown that preoperative EPT with
biliary decompression improves operative results in terms of morbidity and mortality. However, EPT in
patients with papillary carcinoma gains time for both patient and physician during which the operability
of the tumor and surgical risk can be determined. The patient can be prepared for surgery by
correcting abnormalities such as coagulation defects and protein deficiency. If the operation is
contraindicated or the tumor is unresectable, EPT with or without implantation of an endoprosthesis will
improve the quality of the patient's remaining life. Recurrent stenosis occurs in about half of patients
but can be eliminated by repeat EPT with or without implantation of a biliary endoprosthesis. EPT for
tumor carries a somewhat increased risk of bleeding, although this is rarely severe.
Electrosurgical snare papillectomy of the main duodenal papilla has been performed when cannulation
has proved to be impossible. Reported experience with this technique, which is similar in most
respects to snare polypectomy, is extremely limited and available data are inadequate to establish the
efficacy, the nature of potential complications, and morbidity and mortality rates. Nevertheless, Farrell
et al.97(6111) performed this procedure in 10 patients, 5 of whom had exophytic ampullary cancers.
Another patient had cancer arising in the head of the pancreas, 2 had biliary stones lodged in the
papilla, and in 2 patients, the papilla was "protuberant." Subsequent cannulation was successful in all
cases, resected ampullary cancer was retrieved in 4, and significant bleeding ensued in 1 patient.
Acute Obstructive Cholangitis
We undertook a retrospective study of the efficacy of endoscopic treatment of 112 patients with acute
obstructive cholangitis. Sixty-one patients had benign obstruction; the causes were gallstones in 51
patients and benign stenoses in 10. Malignancy was the cause of obstruction in the remaining 51
patients, and in these cases cholangitis occurred after endoscopic manipulation of the bile duct
("postendoscopic cholangitis"). The relatively high percentage of patients with malignant obstruction
and postendoscopic cholangitis emphasizes the importance of carefully cleansing and disinfecting the
endoscopic equipment.98(6112)
Cholangitis was eliminated in all patients in whom the obstruction to bile flow was relieved by EPT and,
when appropriate, the EPT-related maneuvers of stone extraction, drainage by nasobiliary tube or
endoprosthesis, and litholysis. All of these patients survived. The endoscopic procedure failed in 7
patients. Five patients died, 4 of whom were in the group in which endoscopic extraction was
unsuccessful. The clinical condition of these patients deteriorated after endoscopic treatment failed
and was complicated by shock and kidney failure to such an extent that surgical intervention could not
be considered. Changes in laboratory values after successful endoscopic decompression of the bile
ducts in patients with choledocholithiasis are shown in Figure 6011.
(6113)Figure 6011. Changes in serum bilirubin (circles), serum alkaline phosphatase

(triangles), and leukocyte count (boxes) after EPT in 51 patients with acute obstructive
cholangitis due to choledocholithiasis.
The 51 patients with acute cholangitis due to malignant obstruction of the bile duct were treated by
EPT and related maneuvers, as shown in Table 604. One patient with malignant obstruction died
before EPT could be attempted. Endoscopic management was unsuccessful in 14 patients, all of
whom died. Six of the remaining patients with effective biliary drainage established by EPT died in the
hospital, a mortality rate of 16%. Two of these 6 patients died as a result of cholangitis; 3 died as a
result of the primary disease while still hospitalized; and 1 died after elective surgery, presumably as a
result of the primary disease.
Experience with Endoscopic Methods of

Treatment for Acute Obstructive Cholangitis Owing to Malignant
Bile Duct Obstruction
TABLE 604
TREATMENT
Endoscopic papillotomy
Nasobiliary drainage
Bilioduodenal prosthesis
Nasobiliary drainage, subsequent prosthesis
Endoscopic drainage, examination technically impossible
Death before endoscopic examination
Total
PATIENTS (n)
2
13
10
22
3
1
51
There is general agreement that EPT is the treatment of first choice for patients with acute cholangitis
due to choledocholithiasis. The aim of EPT, nasobiliary catheter, or stent placement is rapid
decompression. The high mortality rate associated with acute cholangitis (10 to 65%) can be reduced
significantly by endoscopic means.
Leese et al.99(6114) retrospectively studied the outcome for 94 patients with acute cholangitis, 87% of
whom had gallstones as the cause. There were 12 deaths (14.6%) among patients with gallstones and
3 (25%) in the group without biliary calculi, 1 after percutaneous transhepatic cholangiography. Factors
associated with increased mortality were a low serum albumin, high blood urea nitrogen, high bilirubin,
and a higher score for medical risk factors (Table 605). Only 7 patients could be stabilized clinically
with antibiotics alone. Four patients were moribund on admission and died before they could be
stabilized for biliary decompression. Of the remainder, 28 underwent urgent or early surgery with 6
deaths (21.4%), and 43 had endoscopic biliary decompression with only 2 deaths (4.7%) (Table 606).
The number of complications after EPT was also significantly less (27.9%) compared with that after
surgery (57.1%) (Table 607). Although this was a retrospective study, the results in favor of EPT are
striking.99(6115)
Comparison of Risk Factors in Relation to Mortality in 94 Patients With

Acute Cholangitis*
TABLE 605
RISK FACTOR
SURVIVED (n = 79)
DIED (n = 15)
CORRELATION
Age (years)
Medical risk factors
Systolic BP <100 mm Hg
Blood culture
Prothrombin ratio
70 (2588)
2 (05)
11/79 (4%)
28/55 (51%)\(0.94.0)
1.11 (0.944.0)
12.2 (1.238.5)
78 (6084)
3 (14)
4/15 (27%)
5/12 (42%)\(0.681.35)
1.2 (0.681.35)
15.0 (5.031.0)
NS
p <.05
NS
NS
NS
WBC (x 109/L)
110 (20520)
171 (40660)
p =.06
Serum bilirubin (mol/L)
Serum alkaline
474 (1052400)
631 (2512178)
NS
phosphatase (units/L)
Serum y-GT (units/L)
379 (221170)
445 (1221926)
NS
Serum ALT (units/L)
151 (29895)
114 (30520)
NS
Serum albumin (g/L)
34 (2444)
29 (2535)
p <.005
Serum urea (mmol/L)
6.1 (2.134.9)
10.3 (2.256.5)
p <.05
90 (30500)
137 (60600)
NS
Serum creatinine (mol/L)
Serum glucose (mmol/L)
6.8 (4.219.8)
6.6 (3.510.7)
NS
Modified from Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and the
impact of endoscopic sphincterotomy. Br J Surg 1986; 73:98892. By permission of Blackwell Science Ltd.
NSnot significant; BPblood pressure; WBCwhite blood cell count; y-GTgamma-glutamyl transferase;
ALTalanine transaminase.
* Values are median (range).
Outcome in Relation to Treatment in 82 Patients With Acute Cholangitis Associate
TABLE 606
Stones (n = 82)
EARLY SURGERY DURING
ACUTE ADMISSION
(n = 28)
30-day mortality
NEITHER SURGERY
NOR ES DURING
ACUTE ADMISSION
(n = 11)
EARLY ES DURING
Previous
Cholecystectomy
Gallbladder
Present
No Surgery; Previous
Cholecystectomy
(n = 4)
(n = 24)
(n = 7)
2/4
4/24
4/1
(50%)
(16.7%)
(36.4%)
0/7
(0%)
(50%)
(16.7%)
(36.4%)
(0%)
Modified from Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and the impact of en
1986; 73:98892. By permission of Blackwell Science Ltd.
ESendoscopic sphincterotomy.
Comparison of Complications After Early Endoscopic Sphinctertomy

Decompression of the Biliary Tree in Gallstone Cholangitis
TABLE 607
ENDOSCOPIC SPHINCTEROTOMY
(n = 43)
Complication
Hemorrhage
Recurrent cholangitis
Acute pancreatitis
Gallstone ileus
Death
Number of patients with
complications
Number
5
4 (1*)
1
1 (1*)
2
12 (27.9%)
SURGERY
(n = 28)
Complication
Multisystem failure
Respiratory failure
Blood loss (transfused)
Wound infection
Biliary leak
Burst abdomen/incisional hernia
Number
8
1
7
7 (2*)
4
2 (1*)
Retained common bile duct calculi

3 (1*)
Death
6
Number of patients with complications
16 (57.1%)
Modified from Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and the
impact of endoscopic sphincterotomy. Br J Surg 1986; 73:98892. By permission of Blackwell Science Ltd.
* Requiring surgery.
When stone extraction is unsuccessful, it is advisable to place a nasobiliary catheter or stent or to

decompress the biliary system by transhepatic drainage.100(6116) As already mentioned,
decompressive measures must be combined with antibiotic treatment. Based on a study of the
pathophysiology of bacterial infection in association with biliary obstruction, Leung and
Venezuela14(6117) recommended systemic administration of broad-spectrum antibiotics as an initial
treatment measure before the isolation of a specific organism(s).
Acute suppurative cholangitis is more common in Asia than in other parts of the world. Lam101(6118)
attempted EPT in 134 patients with this condition, most of whom had stones in the common bile duct.
The procedure was successful in 88%, and complete clearing of stones from the bile duct was
achieved in 91.7% of the cases of successful EPT. Major complications occurred in 7.5%, about half of
whom required surgery. More than 85% of the patients were over age 60 years, and over one third had
other significant illnesses. The mortality was 1.5%. This compared favorably with an operative mortality
of 7% in a similar group of patients who underwent surgical sphincteroplasty. Five of 109 patients
without residual stones after EPT developed mild to moderate symptoms of cholangitis during
follow-up of from 6 months to 6 years (mean 2.3 years).
Long-Term Results
EPT was introduced in 1973. Despite the fact that this is a well-established procedure, there have been
few carefully performed follow-up studies of sizable groups of patients. In general, it can be stated that
no long-term complications involving the liver, pancreas, stomach, or colon after EPT have yet been
recognized.
In 1980, we undertook follow-up studies on 51 patients by endoscopic, radiologic, manometric, and

laboratory methods; an additional 66 patients received questionnaires.102(6119) The mean time lapse
after EPT was 21.6 months. Ninety percent of the patients were symptom-free; one third of the
remaining 10% had minor changes in laboratory values that could be explained by other disorders such
as excessive consumption of alcohol. The biliary-duodenal pressure gradient had been completely
abolished in 75% of the patients; duodenobiliary reflux of contrast medium was detected in 25%, and
aerobilia in 65%. None of the patients had signs of cholangitis or any detectable adverse sequelae.
Follow-up information was obtained by Hawes et al.103(6120) on 148 of 163 patients whose common
duct was cleared of stones by EPT between 1975 and 1980. Information on the status of the 148
patients was collected in 1982, and again on 115 of these in 1986. The mean follow-up period was 8
years. Further biliary problems were found in 15 patients. However, sphincter stenosis or stones were
documented in only 5 cases; further endoscopic or conservative treatment was effective in all but 3
patients. Bacterial contamination of the bile was demonstrated in almost two thirds of 44 patients who
underwent endoscopy to check the status of the bile ducts, but the presence of bacteria did not
correlate with symptoms. Hawes et al.103(6121) concluded that their long-term results were
comparable with those of surgical procedures and justify the continued use of EPT for patients with bile
duct stones and prior cholecystectomy.
In 1982, Seifert et al.95(6122) reported results of a study of 9041 patients who had undergone EPT in
25 centers in Germany. Follow-up investigations in the form of ERCP, percutaneous transhepatic
cholangiography, or intravenous cholangiography were obtained in 1050 of these patients. Recurrent
bile duct stones were found in 5.77%; post-EPT stenosis occurred in 3.4%. Most of the patients were
symptom free (93.55%); symptoms were unchanged in 4.9% and worse in 1.63% of patients.
An interesting observation reported by Seifert et al.95(6123) was that recurrent stenosis occurred in
11.5% of patients who underwent EPT for papillary stenosis versus a post-EPT stenosis rate of only
2.9% in patients in whom the indication was choledocholithiasis. Stones recurred in about 5.9% of
patients who had choledocholithiasis, but none was found in those who underwent EPT for papillary
stenosis. Recurrent stone formation is not necessarily related to the presence of a stricture and may
occur even when the orifice produced by EPT is widely patent (Figure 6012). It would appear that
abnormalities in the composition of bile still have an important role in the pathogenesis of these stones.
(6124)Figure 6012. Endoscopic view of a papilla after EPT. A healed, inflammation-free

orifice is seen with free communication between the bile duct and the duodenum.
Seifert et al.104(6125) conducted a second long-term follow-up study of 10,177 patients from 20
German centers. These patients underwent EPT between 1981 and 1986. Choledocholithiasis
remained the main indication, but EPT was performed with increasing frequency in patients having
their gallbladder in situ. The complication rate in this group was low (0.61% required emergency
cholecystectomy). Benign papillary stenosis was the indication for the procedure in 5.85% of patients
compared with 10.64% in the earlier study. The complication rate decreased from 7.55 to 5.04%, the
pattern of complications remaining unchanged. The mortality rate dropped from 1.12 to 0.60%, with
complications occurring twice as frequently among patients who underwent EPT for papillary stenosis
compared with those in whom the procedure was performed for bile duct stones. Recurrent stones and
restenosis of the papilla were encountered more frequently than in the first study. These latter
observations might be explained by more precise control studies by differing distributions of patients
during the two observation periods. Seifert et al.104(6126) concluded that while the therapeutic benefit
of EPT had not changed during the two study periods, the safety of the procedure had increased
(Table 608).
restenosis of the papilla were encountered more frequently than in the first study. These latter
observations might be explained by more precise control studies by differing distributions of patients
during the two observation periods. Seifert et al.104(6126) concluded that while the therapeutic benefit
of EPT had not changed during the two study periods, the safety of the procedure had increased
(Table 608).
TABLE 608
Comparison of Mortality for Two Study Periods

YEARS
PATIENTS
(n)
DEATHS
(n)
MORTALITY
RATE
Choledocholithiasis
197480
7585
77
1.02
198186
7161
36
0.5
Papillary stenosis
197480
813
18
2.21
198186
595
7
1.18
Papillary tumor
197480
187
1
0.53
198186
298
7
2.35
Endoprosthesis
197480
198186
1688
11
0.65
Other
197480
198186
435
Total
197480
8585
96
1.12
198186
10,177
61
0.60
Modified from Seifert E, Schulte F, Chalybans C. Quo vadis endoskopische
Spinkterotomie. Z Gastroenterol 1989; 27:7781.
There is a difference of opinion regarding the need for cholecystectomy in patients who have
undergone EPT with clearance of stones from the bile duct. Numerous studies addressing this issue
have been published.57,62,105118(6127) Most of these are retrospective in design and antedate the
development and widespread use of laparoscopic cholecystectomy. In many studies, a certain
percentage of patients underwent cholecystectomy, either electively or for symptoms, after endoscopic
clearance of the bile duct. This is a considerable source of bias, as the patients who did not undergo
cholecystectomy represent a highly selected group.
Hagenmller et al.105(6128) and Escourrou et al.57(6129) grouped patients in their individual series
according to those who retained and those who did not retain their gallbladders. The mean age of
patients who did not undergo cholecystectomy was 75 years in both studies, 10 years more than the
mean age of the patients in each study who had cholecystectomy before EPT. The long-range survival
of this older group of patients did not seem to be adversely affected by the intact gallbladder. However,
4 to 15% of patients had long-term biliary complications. In the series reported by Escourrou et
al.,57(6130) EPT was performed in 234 patients with intact gallbladders because of advanced age or
poor condition for surgery. Late complications occurred in 16 patients (12%) who retained their
gallbladders after EPT. Follow-up in 130 patients ranged from 6 to 66 months (mean 22 months). In 8
patients, acute cholecystitis developed from 1 to 9 months (mean 4 months) after EPT; 7 of these
underwent cholecystectomy.
EPT was performed by Neoptolemos et al.107(6131) in 100 patients with intact gallbladders. Fifty-nine
patients were considered unfit for surgery. Five of these patients eventually required surgery, 3
because of technical failure and 2 because of empyema of the gallbladder. One patient in this group
died after a large stone could not be extracted. Follow-up ranged from 4 to 50 months, and during this
time 16 patients died, although only 1 death was attributable to sepsis related to cholecystitis. One
patient underwent cholecystectomy because of persistent pain. In a second group of 38 patients, the
EPT procedure preceded cholecystectomy. Choledochotomy was avoided in 29 of these patients. In a
third group of 3 patients, EPT was performed after emergency cholecystostomy. No further surgery
was required in these cases.
Cotton and Vallon108(6132) attempted EPT in 71 elderly patients with intact gallbladders with acute
symptoms due to bile duct stones. The procedure was successful in 70 patients, and the duct was
cleared of stones in 61 (86%). Two patients underwent cholecystectomy for acute cholecystitis within 1
week of the endoscopic procedure. One patient with a retained stone who was considered a poor
operative candidate died 6 weeks after EPT. Eleven patients underwent elective cholecystectomy.
Follow-up (mean 19 months) was obtained in 44 of 48 patients (mean age 75 years) who did not
undergo cholecystectomy. Five of these patients required cholecystectomy because of persistent pain,
but neither cholangitis nor jaundice developed in this group. The authors recommended EPT for all
patients who were acutely ill because of bile duct stones regardless of whether the gallbladder was
present. Although longer follow-up was believed necessary to determine the indications for
cholecystectomy subsequent to EPT, they also recommended that indefinite postponement of
cholecystectomy be considered after successful endoscopic treatment of choledocholithiasis in elderly
and poor-surgical-risk patients.
Martin and Tweedle110(6133) attempted EPT in 81 patients with gallbladders and symptomatic
choledocholithiasis. This was successful in 80 patients, with procedure-related complications in 6 (7%),
1 of whom died. The bile duct was cleared of stones in 70 patients (86%). Subsequent
cholecystectomy, with no operative deaths, was performed in 5 patients electively and in 4 patients
because of biliary symptoms. The remaining 61 patients were followed for 12 to 44 months (mean 24
months), during which time there were 18 deaths with none of these related to biliary tract disease.
These investigators concluded that routine cholecystectomy is not indicated if the bile duct can be
cleared of stones endoscopically.
Forty of 162 patients with gallbladders underwent cholecystectomy after EPT because of cholelithiasis
in the study of Tanaka et al.,116(6134) and the remaining 122 patients were followed for from 6
months to 9 years (mean 3 years). In the latter group, 25 patients with gallbladder stones did not
undergo cholecystectomy because of advanced age, high operative risk, or refusal to have surgery.
There were 19 deaths from causes other than biliary tract disease. Of the surviving 103 patients, only 3
had biliary symptoms. Five of 122 patients followed developed acute cholecystitis from 9 months to 2.4
years (mean 1.8 years) after EPT; 3 had gallbladder stones, and in 2 there was nonvisualization of the
gallbladder at ERCP. Gallbladder stones were found to have developed in 2 patients who did not have
cholelithiasis at the time of EPT. In addition, 5 patients had recurrent bile duct stones, although it was
believed that these had not arisen in the gallbladder with subsequent passage into the bile duct.
Tanaka et al.116(6135) concluded that cholecystectomy should be recommended after EPT when
possible for patients with gallbladder stones or nonvisualization of the gallbladder.
Cholecystectomy for acute cholecystitis was required in 5 of 31 patients (16%) who had gallbladders
with stones after EPT, whereas none of 91 patients without gallbladder stones required surgery in the
series of Ikeda et al.60(6136)
In the series of Siegel et al.,115(6137) 1272 patients with gallbladders underwent EPT. Stones were
present in the gallbladder in 1208 of these patients. After EPT, acute cholecystitis developed in 109
patients (8.6%), occurring within 2 days of the procedure in 23 and within 10 days of EPT in 86 cases.
The majority of these patients were managed by nonsurgical means, but 25 underwent emergency
surgery with a mortality rate of 0.15% (2 deaths). Elective cholecystectomy for biliary symptoms was
performed within 3 years of EPT in 108 patients. Unfortunately, follow-up in the remaining 337 patients
was not complete. However, 2 patients died as a result of recurrent cholecystitis.
Ingoldby et al.109(6138) obtained follow-up data on 186 elderly patients (mean age 79.7 years), many
considered unfit for surgery, with the gallbladder in situ after EPT for choledocholithiasis. Bile duct
clearance was achieved in 172 (92.5%), in some cases after additional endoscopic procedures, with
overall procedure-related complication and mortality rates of 4.8% and 1.6%, respectively. Follow-up
ranged from 6 to 72 months (mean 32 months), during which time cholecystectomy was required in 18
patients (9.6%), with 6 major surgical complications but no deaths. There were 27 deaths from causes
unrelated to biliary tract disease, leaving 156 patients who remained free of biliary symptoms during
the period of the study. These investigators concluded that endoscopic treatment alone is adequate in
a highly selected group of elderly patients who are considered to be unfit for surgery.
Hammarstrom et al.119(6139) retrospectively reviewed 265 cases of patients who had been referred
for EPT and endoscopic removal of bile duct stones. Nineteen cases were eliminated from analysis
because endoscopic treatment was not performed (n = 15) or patient records were unavailable (n = 4).
Bile duct clearance was not achieved in 27 patients (11%), and 35 patients (16%) underwent elective
cholecystectomy despite the absence of biliary symptoms. Follow-up in the remaining 184 patients
ranged from 14 to 150 months (median 69 months), during which time 35 patients (19%) underwent
cholecystectomy, for either acute cholecystitis (n = 23) or biliary symptoms (n = 12). The majority of
patients (86%) underwent operation within 2 years of endoscopic therapy. In this study, the frequency
of cholecystectomy was significantly higher in younger patients, those with complete opacification of
the gallbladder by retrograde cholangiography, and those with gallbladder stones. During follow-up, 76
patients died of causes unrelated to the biliary tract, 17 of whom had mild to moderate symptoms
referable to the biliary tract. Of the remaining 73 patients followed for 16 to 146 months (median 40
months), 59 remained free of biliary symptoms and 14 did not. Hammarstrom et al.119(6140)
concluded from these data that endoscopic treatment alone was satisfactory in most of these selected
patients, although certain factors, such as cholelithiasis, indicated an increased risk for cholecystitis
and could be used to identify patients in whom cholecystectomy would be advisable.
In a prospective study, Holbrook et al.120(6141) performed hepato-iminodiacetic acid (HIDA) scans in
62 patients with the gallbladder in situ after EPT for various pancreaticobiliary diseases. In all cases,
cystic duct patency had been demonstrated at ERCP by occlusion cholangiography. In 44 patients
(70.9%), gallbladder visualization was either delayed or normal; in 18 patients (29.1%), there was
nonvisualization. Only 1 of 21 patients with gallbladder stones and visualization of the gallbladder
subsequently required cholecystectomy, whereas 5 of 13 patients (38.5%) with stones and
nonvisualization required surgery. Holbrook et al.120(6142) concluded that although nonvisualization of
the gallbladder was relatively frequent, this was not of clinical significance except in patients with
cholelithiasis, in which group it was predictive of the need for cholecystectomy in more than one third of
patients.
Worthley and Toouli,121(6143) in a follow-up study of 18 aged, high-surgical-risk patients with
gallbladders in situ after endoscopic extraction of bile duct stones, found that 10 patients with a patent
cystic duct, as demonstrated by retrograde cholangiography at the time of EPT, remained
asymptomatic during follow-up that ranged from 2 to 24 months (median 9 months). However, 6 of the
8 patients with an obstructed cystic duct at EPT subsequently developed recurrent gallbladder
symptoms, and 3 died as a result of gallbladder disease.
The only prospective, randomized trial of surgical versus endoscopic treatment of elderly,
high-surgical-risk patients with choledocholithiasis and gallbladder in situ is that of Targarona et
al.122(6144) Fifty patients were randomized to endoscopic treatment alone and 48 to open
cholecystectomy with intraoperative cholangiography and bile duct exploration if indicated. Surgery was
successful in 92% of patients; endoscopic treatment succeeded in 88%. There were no significant
differences in immediate morbidity or mortality. During follow-up (mean 17 months), 3 patients in the
surgical group had biliary symptoms but none underwent further surgery, whereas 10 endoscopically
treated patients had symptoms and 7 required operation. Multivariate regression analysis disclosed
that endoscopic sphincterotomy was an independent predictor of recurrent biliary symptoms (odds ratio
6.9; 95% confidence interval 1.46 to 32.54).

The results of many of these studies, that of Escourrou et al.,57(6145) in particular, and of our group,
show that patients without a gallbladder have fewer long-term complications after EPT than patients
with the gallbladder in situ. Surgery is usually necessary in these patients, whereas long-term biliary
complications in patients without gallbladders can be managed endoscopically in most cases. For this
reason, we are continuing to recommend elective cholecystectomy after endoscopic treatment of bile
duct stones in younger, surgically fit patients. We assume that the advent of laparoscopic
cholecystectomy will provide added support of this recommendation.
The mortality resulting from biliary complications that occur in patients with intact gallbladders after
EPT is about 1%.57,105,106(6146) This low mortality has been adduced as a reason against elective
cholecystectomy after EPT. However, among almost 900 cases of elective cholecystectomy after EPT,
there were no fatalities.105,106,123,124(6147) It is possible that this is a chance result, but it is more
likely that it supports our recommendation of elective cholecystectomy after EPT in good-surgical-risk
patients.
Complications
The overall complication rate for EPT as determined in large surveys ranges from about 5 to 13% (see
Tables 602 and 603); the mortality rate ranges from 0 to 1% (see Tables 602 and 603). In a large
study from Germany comparing late results and complications of EPT for the period from 1974 to 1980
with the period from 1981 to 1986 demonstrated a general reduction in rates for EPT-related
complications and mortality, although within certain subgroups no improvement was evident (see Table
608).104(6148) A procedure-related mortality rate of 6.6% was reported in a series of 18 patients with
cirrhosis who underwent EPT for extraction of bile duct stones.65(6149) This suggests that the
morbidity and mortality rates for EPT may be higher when the indications for the procedure are
broadened to include certain high-risk patients. Death after EPT is probably caused in most cases by
underlying disorders, such as recurrent cholangitis and sepsis, rather than as a direct consequence of
the EPT, with the possible exception of severe bleeding. Morbidity and mortality rates after EPT are
higher in patients with papillary stenosis than in other patients. In fact, the mortality rate is about two to
four times that in patients with choledocholithiasis without papillary stenosis (see Table 608).
One of the most important studies to date of complications of EPT is that of Freeman et al.125(6150)
Over a period of 2 years, these investigators prospectively collected data from 17 institutions in the
United States and Canada on 2347 patients who underwent biliary EPT. A complication occurred in
229 cases (9.8%); 10 patients (0.43%) died within 30 days as a direct result of a complication of the
procedure. Five factors were identified as indicating a significant increased risk of a complication:
suspected sphincter of Oddi dysfunction, liver cirrhosis, difficult cannulation of the bile duct, "precut"
sphincterotomy, and the combination of percutaneous and endoscopic procedures to achieve access
to the bile duct. With respect to indications for EPT, choledocholithiasis after laparoscopic
cholecystectomy had the lowest associated complication rate (4.9%).
The most frequent complications of EPT are pancreatitis, hemorrhage, cholangitis, and perforation of
the duodenal wall. Perforations and hemorrhages usually become manifest within 12 to 24 hr, whereas
pancreatitis and cholangitis may appear later.
Risk Factors
There does not appear to be a significant relationship between patient age and the risk of a
complication of EPT. In the series of 71 patients over age 70 years who underwent EPT for
choledocholithiasis reported by Mee et al.,126(6151) a complication occurred in 9 patients (13%),
although there were no deaths. However, in the prospective study of Freeman et al.,125(6152) there
was no relationship between patient age and the overall risk of EPT-related complications. Deans et
al.127(6153) prospectively audited 1000 EPT procedures performed in several centers in 958 patients
with an age range of 14 to 97 years (mean 73 years). There were 677 patients greater than 65 years
old and 281 patients of lesser age. The complication rate in the former group was 2.6% versus 2.2% in
the latter, a difference that was not significant. Complications in the older age group occurred in 18
patients and included cholangitis (7 patients), pancreatitis (5), bleeding (3), and perforation (3). By
comparison, 6 patients in the younger age group experienced a complication, including pancreatitis in
5 and bleeding in 1. Two deaths occurred among the older patients, both related to pancreatitis,
whereas there were no deaths among the patients under 65 years of age.
Some authorities regard precut techniques for EPT as having an increased risk of complications,
especially when performed by inexperienced operators.128(6154) In the large prospective study of
Freeman et al.,125(6155) the precut EPT was a significant risk factor for complications. This topic is
discussed extensively in Chapter 61: Special Papillotomy Techniques.
Lambert et al.,63(6156) in a study of 602 patients who underwent EPT for bile duct stones, found that
complication rates decreased with increasing experience. In other large series, a decrease in
complications with increasing experience has been noted.129(6157) Freeman et al.125(6158) found
that endoscopists who performed more than one EPT per week had a significantly lower overall
complication rate (8.4%) compared with those who performed less than one such procedure per week
(11.1%). Furthermore, severe complications were significantly less frequent when EPT was performed
by endoscopists with greater experience (0.9% vs. 2.3%). Factors influencing the frequency of
EPT-related complications in the study of Sherman et al.66(6159) were long sphincterotomy, extraction
of large stones, sphincter of Oddi dysfunction, and a small-diameter common bile duct (5 mm or
less).66(6160) A precut sphincterotomy was performed more frequently in the group of 166 patients
who underwent EPT for sphincter of Oddi dysfunction compared with those with other indications, but it
was not associated with an increase in complications. However, sphincter of Oddi dysfunction as an
indication for EPT together with a small-diameter common duct was associated with a high
complication rate (37.5%). These results contrast with those of Wilson et al.130(6161) In this review of
655 patients who underwent EPT, there was a 10-fold increase in the relative risk of a complication if
the distal bile duct diameter was greater than 8 mm. In this study, the most common complication of
EPT when performed for extraction of bile duct stones was hemorrhage. In the large prospective
multicenter study of Freeman et al.,125(6162) bile duct diameter was not found to be a risk factor for
complications of EPT. As in the study of Sherman et al.,66(6163) however, the complication rate was
significantly higher (21.7%) when the indication for the procedure was suspected sphincter of Oddi
dysfunction.
The periampullary diverticulum as a risk factor for complications of EPT was studied by Vaira et
al.26(6164) EPT was performed in 227 patients in whom the main papilla was located adjacent to or
within a diverticulum. The papilla was situated deeply within the diverticulum in 21 cases. Data from
these cases were compared to data from 1223 EPT cases in which no diverticulum was present.
Complications occurred in 5.2% of patients with periampullary diverticula versus 4.0% of those without
a diverticulum, a difference that was not statistically significant. The respective mortality rates were
0.9% and 0.7%, this difference also being not significant.
Endoscopic sphincterotomy of the minor papilla (Figure 6013) and dorsal pancreatic duct in patients
with pancreas divisum does not appear to be associated with a higher complication rate compared with
EPT of the major papilla (see Chapter 72: Congenital Anomalies of the Pancreas).91,131(6165)
However, there are relatively little data available on safety. In some patients, EPT of the main
pancreatic duct at the major papilla is performed for the treatment of various pancreatic disorders.
Again, relatively little information is available on complications of this procedure (see Chapter 76:
Endoscopic Management of Pancreatic Disease).
(6166)Figure 6013. Needle-knife papillotomy of a minor papilla in a patient with pancreas

divisum and chronic pancreatitis.
There are several reports, generally favorable, of EPT per-formed on an outpatient basis.19,20(6167)
One rationale for this practice is that the majority of complications of EPT will be recognized during the
procedure or become clinically evident within a few hours. However, it is well established that certain
serious complications, bleeding in particular, may be delayed in onset. The data currently available on
EPT and other therapeutic ERCP procedures are largely retrospective in nature.
In the retrospective study of Podolsky et al.,19(6168) the records of 137 patients who underwent EPT
on an outpatient basis in a single center were compared with those of an equal number in whom the
procedure was performed with hospitalization. Papillary stenosis, which as an indication for EPT has
been shown in some studies to be associated with a higher rate of complications, was the indication for
the procedure in 35% of outpatient cases versus 15% of EPT procedures performed on an inpatient
basis. The complication rates for outpatient and inpatient procedures were 6.6% and 7.3%,
respectively. Complications were recognized within 2 to 4 hr of EPT in patients who underwent the
procedure on an ambulatory basis; no delayed complications were identified, and there were no
deaths. Mehta et al.20(6169) collected data on 209 consecutive patients who underwent 262 ERCP
procedures; EPT was performed in 181 patients (69%). The major indication for the procedure was
suspected or documented choledocholithiasis (50%), followed by malignant obstruction (29%) and
sphincter of Oddi dysfunction (14%). Overall, a complication occurred in 15 patients (5.7%). Of these,
9 (3.7%) required hospitalization, all of whom had undergone a therapeutic procedure, although this
was not an EPT in all cases; in 7, the complication was recognized within 4 hr of the procedure,
whereas diagnosis was delayed 3 and 12 days in the other 2 patients.
As a rule, EPT-related complications can be avoided if careful attention is given to two important
principles: (1) the position of the papillotome in the bile duct should be checked carefully before starting
the incision, and (2) drainage of bile must be ensured after the EPT. Strict adherence to these
principles has reduced the EPT-related complication rate in our patients from 7.4 to 4.8%; the
necessity for surgical treatment has been reduced by one third.
Hemorrhage
A retrospective cohort study by Nelson and Freeman132(6170) of patients undergoing EPT revealed,
by univariate and multivariate analysis, that three factors were predictive of post-EPT hemorrhage:
hemodialysis, a prothrombin time prolonged 2 or more sec above the control value, and observed
bleeding at the time of the procedure. The onset of major hemorrhage was delayed in the majority of
patients in this study. Using this methodology, these investigators found that sphincter of Oddi
dysfunction as an indication for EPT, the use of nonsteroidal antiinflammatory drugs, and
sphincterotomy length were not independently associated with hemorrhage. In the retrospective
analysis of Leung et al.,133(6171) bleeding occurred in 119 (12.1%) of 983 patients who underwent
EPT for choledocholithiasis. The following were found to be independent risk factors for bleeding:
coagulopathy, stone impaction, extension of previous sphincterotomy, and precut sphincterotomy with
a needle-knife.
The length of the EPT incision is generally believed to be related to the risk of hemorrhage, that is,
longer incisions are associated with an increased risk. Although this was not identified as a risk factor
in the study of Freeman et al.,125(6172) it is possible and perhaps likely that many if not most of the
endoscopists at the 17 institutions participating in this multicenter study may have been cognizant of
the relationship between incision length and bleeding. A small volume of self-limited blood loss occurs
almost always during or immediately after EPT, but this is clinically insignificant unless transfusions are
required. The mean diameter of the vessels in the papillary arterial plexus is about 0.98 mm; the mean
diameter of vessels in the roof (proximal aspect) of the papilla is only about 0.43 mm. However, in
about 4% of patients, the retroduodenal artery is in the region of the papillotomy. Thus, severe
hemorrhage may occur in a minority of patients undergoing EPT.
Goodall134(6173) found evidence of post-EPT bleeding in 21 of 194 patients who underwent 235 EPT
procedures. However, bleeding was occult in 7 cases. Six of the 14 patients with overt bleeding
required operation, and 2 patients died as a result of hemorrhage. There was an increased risk of
bleeding with repeated procedures. Goodall134(6174) did not indicate the time interval between the
first and the subsequent EPT procedures and did not indicate whether the bleeding could be related to
excessive length of the incision or the hypervascularity that occurs at the papillotomy site after EPT. In
almost all cases, bleeding began within 24 hr of EPT, although in 2 patients without evidence of
bleeding during EPT, the onset of bleeding was delayed for 5 days. Finnie et al.135(6175) encountered
delayed bleeding (3 to 21 days after EPT) in 5 of 200 patients who underwent EPT for bile duct stones.
No explanation for this phenomenon was offered, but these investigators advised endoscopists to warn
patients about the potential for delayed onset of this complication. In a review of 75 EPT procedures
performed in 68 patients, Mellinger and Ponsky136(6176) found that delayed bleeding had occurred in
3 patients. The series reported by Gholson et al.137(6177) included 250 patients who underwent EPT.
Procedure-related bleeding occurred in 5 patients (2%); in 3, it was delayed in onset by 20 to 48 hr with
the main symptom being melena. Transfusion of from 2 to 6 units of blood was necessary, and 1
patient underwent surgery for control of the bleeding. Post-EPT bleeding was delayed in onset in 5 of
119 patients (4.2%) in whom this complication occurred in the study of Leung et al.133(6178)
Operation is necessary to control bleeding in only about 11% of cases of hemorrhage after EPT.
Endoscopic tamponade of the incision using special balloon catheters has been performed to control
hemorrhage in a few patients with some success.138(6179) Various methods of endoscopic
hemostasis have been recommended to arrest the hemorrhage: injections of epinephrine,51(6180)
polidocanol,139(6181) radiographic contrast agent,140(6182) and application of coagulation current by
means of the sphincterotome.59(6183) None of these has proved to be entirely satisfactory.
Leung et al.133(6184) classified EPT-induced bleeding on the basis of severity into mild and severe.
The former was treated by epinephrine irrigation, whereas severe bleeding was managed by injection
of a 1:10,000 solution of epinephrine. Of the 119 patients with bleeding, 46% responded to irrigation
alone; epinephrine injection was required in the remainder. Initial hemostasis was achieved in all
cases, although bleeding was recurrent in 5 patients who nevertheless responded to further
conservative measures. There were no deaths as a result of bleeding, and no patient required surgery.
Electrocoagulation using a multipolar probe was employed by Sherman et al.141(6185) in nine patients
with EPT-induced hemorrhage that was unresponsive to injection therapy or balloon tamponade.
Electrocoagulation was not associated with any complications. Bleeding was delayed in onset 12 hr or
more in eight patients and was classified as moderate or severe and required transfusion on average
of 4.8 units of blood. Hemostasis was achieved in eight patients, two of whom had recurrent bleeding
that was controlled in one case by further multipolar electrocoagulation.
There is a wide range in severity of bleeding. In some cases, bleeding stops spontaneously, whereas
in others arterial bleeding can be so brisk that the duodenum is filled with blood before any endoscopic
measures can be applied. When bleeding is severe, selective angiographic embolization of the
appropriate artery might prove to be the hemostatic method of choice, although experience with this
approach is very limited. Of the five patients with severe post-EPT bleeding reported by Saeed et
al.,142(6186) arteriography demonstrated bleeding from branches of the gastroduodenal artery in four
cases. Although embolization was not possible in one patient because of arterial spasm and stenosis,
it controlled the bleeding in three others. Bleeding in the fifth patient had stopped at the time of
angiography.
It was a reasonable expectation that EPT-related hemorrhage might be avoided by use of endoscopic
Doppler ultrasonography, a procedure not yet widely available. With this technique, it is possible to
measure reasonably intense arterial signals within the wall of the duodenum. Our results, however, in a
first controlled clinical trial were disappointing.18(6187) Thus, at present, the best way to avoid
hemorrhage is to restrict the length of the incision to the visible roof of the papilla.
Extrahepatic portal venous obstruction can involve the extrahepatic bile duct. Retrograde
cholangiography may demonstrate narrowing of the bile ducts in various locations including the left and
right hepatic ducts as well as the common hepatic and common bile ducts.143145(6188) Various
mechanisms have been proposed to account for these narrowed segments, including fibrous scarring,
poor venous drainage of bile duct venules with resultant ischemia, and a simple mass effect of the
portal cavernoma. If portal venous obstruction is not recognized as the cause of bile duct narrowing,
therapeutic endoscopic maneuvers may result in bleeding.146(6189)
Perforation
Perforations may occur even when the length of the incision is properly restricted to the roof of the
papilla. This may be accounted for in part by anatomic variations in the distal retropapillary aspect of
the papilla, although the degree to which such variations are of importance is not certain and can be
determined only by a careful prospective study. After every EPT incision, the choledochoduodenal
junction should be checked radiographically for leakage. Extravasation of contrast medium or air into
the retroperitoneum confirms the presence of a perforation. Perforation may result in
pneumomediastinum,147(6190) pneumothorax, and subcutaneous emphysema.148,149(6191) There
is a case report of pneumomediastinum after EPT in which there was no radiologic evidence of a
perforation of the duodenal wall.150(6192) The presence of air in the mediastinum or chest does not
correlate with the severity of perforation.
Perforation is an infrequent complication. Characteristic symptoms include pain that begins in the right
hypochondrium and radiates throughout the upper abdomen and nausea with or without vomiting.
Bowel sounds are present but hypoactive. Abdominal tenderness, leukocytosis, and hyperamylasemia
may be encountered.151(6193) Fever, peritonitis, and septic shock may occur. Cases of submucosal
choledochal emphysema152(6194) and hepatic portal venous gas have been reported.153,154(6195)
Herman et al.154(6196) speculated that gas entered the portal system via small vessels that are cut
during EPT. In reported cases, gas in the portal venous system resolved spontaneously and without
evidence of adverse clinical consequences.
Detection of a perforation does not necessarily mean that surgery must be performed
immediately.155(6197) Many perforations heal spontaneously with or without nasobiliary drainage.
Byrne et al.156(6198) encountered 5 perforations in approximately 500 procedures. They found that
most patients responded satisfactorily to conservative management, provided that biliary drainage was
adequate. Evidence of perforation is usually present on plain x-rays of the abdomen as free air in the
retroperitoneal space that sometimes outlines retroperitoneal structures such as the kidney. Computed
tomography was recommended by Kuhlman et al.157(6199) as the "study of choice" for evaluating the
patient in whom a perforation is suspected. The patient in whom a perforation has occurred must be
monitored clinically at close intervals, and if there are any signs of deterioration such as increasing
pain, abdominal tenderness, or fever, surgery should be performed immediately. In severe cases, the
surgeon will usually establish drainage rather than locate and close the perforation site.
Cholangitis
The bile ducts in patients with choledocholithiasis harbor bacteria in almost all cases. If biliary blockage
is not eliminated or is worsened by endoscopic intervention, cholangitis may occur. This may result if a
stone(s) becomes impacted because the incision is too short to allow removal or passage of the stone
or the stone is too large. Impeded flow of contrast medium caused by benign or malignant biliary
stenoses may also provoke cholangitis. Insertion of a nasobiliary tube will greatly reduce the likelihood
of stone impaction or stasis of contrast and cholangitis.
The risk of bacteremia after EPT was found in the study of Low et al.158(6200) to be not significant,
and antibiotic prophylaxis was therefore considered unnecessary by these investigators. In the
randomized controlled trial of Sauter et al.,159(6201) pre-ERCP antibiotic prophylaxis resulted in a
significant decrease in the rate of bacteremia, although there was no difference in the frequency of
cholangitis after ERCP. Ninety-six patients undergoing 100 procedures were randomized to receive
antibiotic prophylaxis (cefotaxime 2 g intravenously 15 min before the procedure) or a control group.
The rate of bacteremia was 2% in the former group and 16% (p < .02) in the latter. In another study,
biliary sepsis had the highest mortality rate of the EPT-related complications.63(6202) Because the
presence of biliary stones or obstruction cannot always be anticipated before ERCP, it is advisable to
obtain informed consent for EPT from every patient, so that endoscopic maneuvers can be performed
as needed to establish biliary drainage.
Pancreatitis
The most common complication of EPT in the study of Freeman et al.125(6203) was pancreatitis.
Based on a survey of published data that included 7168 cases in which EPT was performed, Sherman
and Lehman160(6204) calculated the rate of EPT-related pancreatitis at 1.7%. Surgery was required in
the management of 8% of these cases. Pancreatitis is believed to result from too many inadvertent
injections into the pancreatic duct during attempts at placement of the papillotome into the bile duct,
direct trauma to the pancreatic duct with the papillotome by repeated or forceful cannulation, and
excessive use of coagulation current during the EPT with resultant occlusion of the orifice of the
pancreatic duct because of edema or direct tissue damage. Efforts to prevent ERCP- and
EPT-induced pancreatitis by administration of somatostatin are discussed in Chapter 56: Indications,
Contraindications, and Complications of Diagnostic Endoscopic Retrograde
Cholangiopancreatography.
Miscellaneous
Gallstone ileus may rarely occur after EPT for large stones in the bile duct.161163(6205) In the case
reported by Halter et al.,161(6206) a 3.5-cm-diameter calculus lodged in the jejunum about 50 cm
distal to the ligament of Treitz 3 days after EPT, at which time stone extraction from the bile duct was
unsuccessful. The stone was removed at laparotomy 9 days after the endoscopic procedure. A stone,
2.8 cm in smallest dimension, lodged 5 cm proximal to the ileocecal valve in the case reported by
Despland et al.163(6207)
Functional Consequences of EPT

Changes in Bile Composition
The composition of the bile is influenced by changes in the anatomy of the extrahepatic biliary system.
Cholecystectomy leads to a decrease in bile acid pool size, especially in patients with a large pool size
preoperatively.164(6208) At the same time, the fractional turnover of primary bile acids increases and
there is a rise in the proportion of secondary bile acids in the bile.
EPT may accentuate the effects of cholecystectomy or induce changes that resemble those associated
with cholecystectomy in patients who have retained their gallbladders. Sauerbruch et al.165(6209)
studied the effect of EPT on bile acid pool and lipid composition in three patients with gallbladders and
seven patients who had undergone cholecystectomy. Determinations were made within several days
and at several months after EPT. Total bile acid pool size was markedly reduced by EPT in patients
with gallbladders, whereas there was no significant change in pool size in those without gallbladders.
We have found that cholesterol saturation of bile obtained from the bile duct decreases after EPT in
patients with intact gallbladders.166(6210) However, Sauerbruch et al.165(6211) found no increase in
the degree of cholesterol saturation in hepatic bile after EPT in patients with intact gallbladders.
Stellaard et al.167(6212) studied the effect of EPT on bile acid composition and cholesterol saturation
of bile in 13 women who had undergone cholecystectomy at least 9 months before the endoscopic
procedure. Twelve women who had cholecystectomy and did not undergo EPT were a control group.
The EPT group exhibited a significantly higher percentage of chenodeoxycholic acid in bile, but no
difference in the proportion of cholic acid. The percentages of the secondary bile acids were also lower
in the EPT group, but the differences were not statistically significant. The biliary lipid composition in
the EPT group did not differ from that in controls, so that the cholesterol saturation index in the patients
undergoing EPT was similar to that of the control patients. Stellaard et al.167(6213) suggested that the
changes demonstrated in bile composition would not be deleterious to the biliary or gastrointestinal
systems. The significance of these findings remains unclear.
The normal diurnal fluctuations in biliary lipid concentrations are unchanged after cholecystectomy or
EPT.
Chijiiwa et al.168(6214) studied seven patients who underwent EPT because of common bile duct
stones. Six of these patients had undergone cholecystectomy 1 month to 23 years before EPT. They
found that there was little effect on the biliary lipids and bile acid composition over an observation
period of 1 year. Nevertheless, longer studies appear to be necessary because at the end of the
observation period there was a tendency, albeit not significant, for primary bile acids to decrease
whereas there was a slight increase of secondary bile acids.
Manometric Investigations
The duodenobiliary pressure gradient is normally 8 mm Hg.169(6215) This is abolished in 80% of
patients and reduced to 1 to 5 mm Hg in 20% of patients after EPT for bile duct stones in which the
incision produced is relatively long. When the EPT incision is shorter, such as in small papillotomies for
the purpose of endoprosthesis insertion, the duodenobiliary pressure gradient usually remains
unchanged or is reduced to an insignificant degree. In 11 patients who underwent this type of EPT and
had manometry before and after the procedure, the pressure difference was eliminated in only 3
cases. Thus, this more selective type of papillotomy probably leaves remnants of the sphincter
mechanism substantially intact. These observations have been confirmed by Staritz et
al.170,171(6216)
Geenen et al.7(6217) studied the length of the incision after EPT using the papillotome itself and an
inflated Fogarty balloon for reference. At 12 and 24 months after EPT, the duodenobiliary pressure
gradient and the basal sphincter of Oddi pressure as determined by sphincter of Oddi manometry had
been virtually abolished, and the amplitude of the phasic sphincter contractions had decreased
significantly. However, an increase in the amplitude of the phasic contractions was found 24 months
after EPT to the point that the difference between the determinations at 24 months and the baseline
determinations before EPT no longer reached statistical significance. Over the long term, the size of
the papillary opening (length of the incision) decreased from 11.6 0.8 mm to 7.5 0.7 mm after 1
year and to 6.5 0.7 mm after 2 years. These morphologic and manometric investigations confirm that
the papillary orifice widened by EPT remains open to a significant degree for at least 2 years.
An intact sphincter of Oddi is probably an important barrier to spread of bacteria from the intestines. In
39 patients undergoing follow-up several months to some years after EPT, we found that there was
considerable bacterial colonization in bile aspirated selectively from the bile duct (Table 609).
However, none of these patients had symptoms or biochemical evidence of cholangitis. For this
reason, we regard bacteriocholia as innocuous as long as bile flow remains unhindered.
year and to 6.5 0.7 mm after 2 years. These morphologic and manometric investigations confirm that
the papillary orifice widened by EPT remains open to a significant degree for at least 2 years.
An intact sphincter of Oddi is probably an important barrier to spread of bacteria from the intestines. In
39 patients undergoing follow-up several months to some years after EPT, we found that there was
considerable bacterial colonization in bile aspirated selectively from the bile duct (Table 609).
However, none of these patients had symptoms or biochemical evidence of cholangitis. For this
reason, we regard bacteriocholia as innocuous as long as bile flow remains unhindered.
Bacteria Strains in the Bile Duct After

Endoscopic Papillotomy in 39 Patients*
TABLE 609
STRAIN OF BACTERIA
PATIENTS IN WHOM
STRAIN WAS DETECTED
(n)
Escherichia coli
34
Klebsiella sp.
7
Enterobacter sp.
3
Proteus sp.
13
Providencia
1
Pseudomonas sp.
6
Enterococci
18
Streptococcus sp.
1
Staphylococcus aureus hemolytica 2
Candida sp.
1
* Cultures of bile obtained by endoscopic retrograde aspiration.
Gregg et al.172(6218) studied the bacterial content of bile before and after either EPT or surgical
sphincteroplasty in 45 patients with sphincter of Oddi stenosis. In all cases, the bile was sterile before
the procedure. Bile recultured 6 to 36 months after EPT contained bacteria in 70% of cases. However,
there was no associated clinical evidence of cholangitis. More than one type of bacteria was found in
most patients, and the majority were enteric organisms.
Greenfield et al.173(6219) studied 25 patients after EPT (mean 36 months) for choledocholithiasis.
Although symptomatic improvement occurred in all and there were no cases clinically of acute
cholangitis, 20% of the patients had experienced mild episodes of abdominal pain, and a similar
number had elevated serum gamma glutamyl transpeptidase activity as high as three times the normal
level. In about half of the patients, air in the biliary system was shown radiographically. In these
patients, liver biopsy revealed mild portal tract fibrosis and inflammation. There was a statistical
correlation between episodes of mild upper abdominal pain and air in the biliary system
radiographically, which these investigators accepted as evidence of biliary reflux of gastrointestinal
contents. The long-term significance of observations such as these is unclear.
Alternatives to EPT
Although it is clear that EPT can result in changes in gastrointestinal tract physiology, the degree to
which these are of practical importance remains unclear. There are thus some reservations about EPT
in relatively young individuals in whom normal life expectancy would allow adequate time for any
adverse effect to arise as a result of these alterations in anatomy and physiology. Because of this,
approaches other than EPT to endoscopic therapy that produce a more physiologic result are being
studied.
The most important indication for EPT of the major duodenal papilla is choledocholithiasis. There have
been a number of reports of balloon and pharmacologic dilation of the sphincter of Oddi for the
purpose of endoscopic removal of bile duct stones. The endoscopic use of dilating balloons has been
reviewed by Tytgat et al.174(6220)
Staritz et al.175(6221) demonstrated that the extraction of small stones from the bile duct may be
possible after pharmacologic dilation of the sphincter of Oddi using glycerol trinitrate. The extraction of
32 bile duct stones 6 to 12 mm in diameter from 21 patients was attempted after administration of 1.2
to 3.6 mg of glycerol trinitrate. Thirty stones were extracted without difficulty. The remaining 2 stones
were reduced in size by endoscopic mechanical lithotripsy and then extracted. Subsequent
manometric investigations showed that papillary function was retained. With the induction of sphincter
of Oddi dilation by use of sublingual nitroglycerin (0.3 to 0.6 mg), Uchida et al.176(6222) were able to
clear the bile duct of stones in 18 of 21 patients. By fluoroscopic assessment, stones ranged in
maximum diameter from 4 to 11 mm, with an average diameter of about 6 mm. Single stones were
present in 13 patients; mechanical lithotripsy was required in 2 cases. Complications included mild
pancreatitis in 1 patient and transient hypotension in another. Despite these favorable reports,
pharmacologic dilation of the sphincter of Oddi for stone extraction and other purposes has not been
accepted widely or applied by other groups of investigators.
Staritz et al.177(6223) also proposed balloon dilation of the sphincter of Oddi, using a special
banana-shaped balloon 4 cm in length and 1.5 cm in diameter, as an alternative to EPT. This method
also appears to be suitable for the removal of smaller stones from the bile duct.
Mac Mathuna et al.178(6224) studied the immediate and long-term morphologic effects of balloon
sphincteroplasty in pigs. Some animals were sacrificed between 15 min and 2 hr after balloon dilation
of the sphincter of the major duodenal papilla to 8 mm diameter at 10 atm pressure, and others were
sacrificed from 6 to 12 weeks after the procedure. The major morphologic change shortly after the
procedure was acute transmural inflammation. Chronic follicular hyperplasia was evident at 6 to 12
weeks, but there was no fibrosis or change in the papillary structure.
Minami et al.179(6225) compared sphincter dilation to EPT in patients with bile duct stones less than
12 mm in diameter. Although the mean duration of the dilation procedure was 63 min compared with
42 min for EPT, this difference was not significant. Additional maneuvers, such as mechanical
lithotripsy, were more frequently required in the group of patients who underwent dilation. Decreases in
parameters of sphincter function were noted by manometry in patients who underwent dilation,
whereas function was completely abolished in those who had EPT.
Endoscopic sphincter dilation was performed by Mac Mathuna et al.180(6226) in a series of 100
patients with bile duct stones up to 20 mm in diameter. The ducts were completely cleared of stones in
78% and incompletely in 4%. Mechanical lithotripsy was necessary to achieve ductal clearance in 10%
of patients. In all cases in which clearance was initially incomplete, the ducts were ultimately cleared of
stones in all cases when the procedure was repeated. The major reason for the 18% failure rate was
stones greater than 15 mm in diameter. Pancreatitis occurred in 5% of the patients. Follow-up ranged
from 6 to 30 months (median 16 months), during which time no evidence of papillary stenosis was
encountered.
Endoscopic sphincter of Oddi manometry was performed by Sato et al.181(6227) before and 1 week
after balloon dilation of the sphincter in 10 patients with bile duct stones. Manometry was also obtained
in 7 of these patients at 1 month after dilation. There were no complications of the procedure. At 1
week after dilation, significant decreases were recorded for common duct pressure and sphincter of
Oddi peak and basal pressures as well as frequency of contractions. At 1 month, however, most
parameters had returned toward normal with the exception of basal sphincter and bile duct pressures,
and there were no significant differences in any parameter between the values obtained before dilation
and those at 1 month after procedure.
In the randomized, controlled trial of Bergman et al.182(6228) of EPT versus balloon dilation in the
endoscopic treatment of patients with bile duct stones, there was no significant difference in rates of
success for clearance of the bile duct, although mechanical lithotripsy was required significantly more
often in the dilation group. In addition, there were no significant differences with respect to
complications including pancreatitis, which occurred with equal frequency in the two groups of patients.
One dilation-treated patient died as a result of a retroperitoneal perforation. EPT-induced bleeding
occurred in 4 patients.
Serial dilation of the minor papilla with 4- to 7-French catheters was used by Lans et al.90(6229)
preparatory to insertion of stents into the dorsal pancreatic duct in an effort to avoid problems with
restenosis and severe procedure-related pancreatitis.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Zimmon DS, Falkenstein DB, Kessler RE. Endoscopic papillotomy for choledocholithiasis. N
Engl J Med 1975;293:11812.
Geenen JE. Endoscopic papillotomy. In Demling L, Classen M (eds). Endoscopic
Sphincterotomy of the Papilla of Vater. Stuttgart: Georg Thieme, 1978;812.
Stain SC, Cohen H, Tsuishoysha M, Donovan AJ. Choledocholithiasis. Endoscopic
sphincterotomy or common bile duct exploration. Ann Surg 1991;213:62733; discussion
6334.
Neoptolemos JP, Carr-Locke DI, Fossard DP. A prospective randomised study of
preoperative endoscopic sphincterotomy versus surgery alone on common bile duct stones.
BMJ 1987;294:4704.
Cotton PB, Williams CB. In Practical Gastrointestinal Endoscopy. 3rd ed. Cambridge, MA:
Blackwell Scientific. 1990.
Del Olmo L, Merono E, Moreira VF, et al. Successful treatment of postoperative external
biliary fistulas by endoscopic sphincterotomy. Gastrointest Endosc 1988;34:3079.
Geenen JE, Toouli J, Hogan WJ, et al. Endoscopic sphincterotomy: Follow-up evaluation of
effects on the sphincter of Oddi. Gastroenterology 1984;87:7548.
Oria A, Alveraz J, Chiappetta L, et al. Choledocholithiasis in acute gallstone pancreatitis.
Incidence and clinical significance. Arch Surg 1991;126:5668.
Ros E, Navarro S, Bru C, et al. Occult microlithiasis in idiopathic acute pancreatitis:
Prevention of relapses by cholecystectomy or ursodeoxycholic acid therapy. Gastroenterology
1991;101:17019.
Geenen JE, Rolny P. Endoscopic therapy of acute and chronic pancreatitis. Gastrointest
Endosc 1991;37:37782.
Bernard JP, Sahel J, Giovannini M, Sarles H. Pancreas divisum is a probable cause of acute
pancreatitis: A report of 137 cases. Pancreas 1990;5:24854.
Schneider MU, Lux G. Floating pancreatic duct concrements in chronic pancreatitis. Pain
relief by endoscopic removal. Endoscopy 1985;17:810.
Fuji T, Amano H, Harima K, et al. Pancreatic sphincterotomy and pancreatic endoprosthesis.
Endoscopy 1985;17:6972.
Leung JWC, Venezuela RR. Cholangiosepsis: Endoscopic drainage and antibiotic therapy.
Endoscopy 1991;23:2203.
Seibert DG, Matulis SR. Consistent improvement in sphincterotome orientation with manual
grooming. Gastrointest Endosc 1995;42:3259.
Silverstein FE, Deltenre M, Tytgat G, et al. An endoscopic Doppler probe: Preliminary clinical
evaluation. Ultrasound Med Biol 1985;11:34753.
Martin RW, Gilbert DA, Silverstein FE, et al. An endoscopic Doppler probe for assessing
intestinal vasculature. Ultrasound Med Biol 1985;11:619.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
Neuhaus H, Hagenmller F, Lauer R, Classen M. A prospective randomized trial of the

influence of suprapapillary Doppler ultrasound on endoscopic sphincterotomy [Abstract].
Podolsky I, Kortan P, Haber GB. Endoscopic sphincterotomy in outpatients. Gastrointest
Endosc 1989;35:3726.
Mehta SN, Pavone E, Barkun AN. Outpatient therapeutic ERCP: A series of 262 consecutive
cases. Gastrointest Endosc 1996;44:4439.
Cotton PB, Williams CB. Therapeutic endoscopic retrograde cholangiopancreatography
(ERCP). In Practical Gastrointestinal Endoscopy. 3rd ed. Cambridge, MA: Blackwell Scientific,
1990;11859.
Siegel JH, Cohen SA, Kasmin FE, Veerappan A. Stent-guided sphincterotomy. Gastrointest
Endosc 1994;40:56772.
Sander R, Poesl H. Endoscopic papillotomy with the Nd YAG laser. Endoscopy
1985;17:1156.
Osnes M. Endoscopic choledochoduodenostomy for common bile duct obstructions. Lancet
1979;1:105960.
Schapira L, Khawaja FI. Endoscopic fistulo-sphincterotomy: An alternative method of
sphincterotomy using a new sphincterotome. Endoscopy 1982;14:5860.
Vaira D, Dowsett JF, Hatfield AR, et al. Is duodenal diverticulum a risk factor for
sphincterotomy?. Gut 1989;30:93942.
Cremer M, Gulbis A, Toussaint J, et al. Techniques of endoscopic papillotomy. In Delmont J
(ed). The Sphincter of Oddi. Basel: S. Karger, 1977;21927.
O'Brien JW, Chen SL, Connolly R, Libby ED. Current induction in a fiberglass guidewire
compared to conventional wires during simulated papillotomy. Gastrointest Endosc
1997;45:4937.
Johlin FC, Tucker RD, Ferguson S. The effect of guidewires during electrosurgical
Schoenfeld PS, Jones DM, Lawson JM. Conducted current on guidewires in single lumen
papillotomes. Gastrointest Endosc 1991;37:3446.
Wurbs D, Phillip J, Classen M. Experiences with the long-standing nasobiliary tube in biliary
diseases. Endoscopy 1980;12:21923.
Stolte M, Wiessner V, Rsch W. Todesursachen nach endoskopischer Papillotomie. Z
Riemann JF, Seuberth K, Demling L. Clinical application of a new mechanical lithotripter for
smashing common bile duct stones. Endoscopy 1982;14:22630.
Frimberger E, Weingart J, Khner W, et al. Eingeklemmter Papillenstein: Mechanische
Lithotripsie mglichkeit. Dtsch Med Wochenschr 1983;108:38.
Riemann JF, Seuberth K, Demling L. Mechanical lithotripsy through the intact papilla of vater.
Endoscopy 1983;15:1113.
Frimberger E, Khner W, Weingart J, et al. Eine neue Methode der elektrohydraulischen
Cholelithotripsie (Lithoklasie). Dtsch Med Wochenschr 1982;107:2135.
Tanaka M, Yoshimoto H, Ikeda S, et al. Two approaches for electrohydraulic lithotripsy in the
common bile duct. Surgery 1985;98:3138.
Bonnel DH, Liguory CE, Cornud FE, Lefebvre JFP. Common bile duct and intrahepatic
stones: Results of transhepatic electrohydraulic lithotripsy in 50 patients. Radiology
1991;180:3458.
Classen M, Hagenmller F, Knyrim K, Frimberger E. Giant bile duct stonesNon-surgical
treatment. Endoscopy 1988;20:216.
Martin EC, Wolff M, Neff RA, et al. Use of the electrohydraulic lithotriptor in the biliary tree in
dogs. Radiology 1981;139:2157.
Cotton PB, Kozarek RA, Schapiro RH, et al. Endoscopic laser lithotripsy of large bile duct
stones. Gastroenterology 1990;99:112833.
Neuhaus H, Hoffmann W, Zillinger C, Classen M. Laser lithotripsy of difficult bile duct stones
under direct visual control. Gut 1993;34:41521.

43. Neuhaus H. Fragmentation of pancreatic stones by extracorporeal shock wave lithotripsy.
Endoscopy 1991;23:1615.
44.
Jessen K, Phillip J, Classen M. Endoscopic jet cuttingA new method for stone destruction in
the common bile duct. 6th International Symposium on Jet-Cutting Technology. BH RA Fluid
Engineering, Cranfield, Bedford MK 43 OAJ, England, 1982;39.
45. Demling L, Ermert H, Riemann JF, et al. Lithotripsy in the common bile duct using ultrasound.
Preliminary in vitro experiments. Endoscopy 1984;16:2268.
46. Kaye GL, Summerfield JA, McIntyre N, Dooley JS. Methyl tert butyl ether dissolution therapy
for common bile duct stones. J Hepatol 1990;10:33740.
47. Nakajima M, Kizu M, Akasaka Y, Kawai K. Five years experience of endoscopic
sphincterotomy in Japan: A collective study from 25 centres. Endoscopy 1979;11:13841.
48. Cotton PB, Vallon AG. British experience with duodenoscopic sphincterotomy for removal of
bile duct stones. Br J Surg 1981;68:3735.
49. Seifert E. Endoscopic papillotomy and removal of gallstones. Am J Gastroenterol
1978;69:1549.
50. Montori A, Viceconte G, Viceconte GW, Bogliolo G. ERCP and EPT: Italian experience.
Endoscopy 1979;11:1425.
51. Geenen JE, Vennes JA, Silvis SE. Resum of a seminar on endoscopic retrograde
sphincterotomy (ERS). Gastrointest Endosc 1981;27:318.
52.
Chung SCS. Biliary endoscopy. Curr Opin Gastroenterol 1992;8:7708.
53. Viceconte G, Viceconte GW, Pietropaolo V, Montori A. Endoscopic sphincterotomy:
Indications and results. Br J Surg 1981;68:37680.
54. Safrany L. Duodenoscopic sphincterotomy and gallstone removal. Gastroenterology
1977;72:33843.
55. Koch H, Rsch W, Schaffner O, Demling L. Endoscopic papillotomy. Gastroenterology
1977;73:13936.
56. Siegel JH. Endoscopic papillotomy in the treatment of biliary tract disease. 258 procedures
and results. Dig Dis Sci 1981;26:105764.
57. Escourrou J, Cordova JA, Lazorthes F, et al. Early and late complications after endoscopic
sphincterotomy for biliary lithiasis with and without the gall bladder ": situ". Gut
1984;25:598602.
58.
Wurbs D. Endoscopic papillotomy. Scand J Gastroenterol 1982;77(suppl):10715.
59. Leese T, Neoptolemos JP, Carr-Locke DL. Successes, failures, early complications and their
management following endoscopic sphincterotomy: Results in 394 consecutive patients from
a single centre. Br J Surg 1985;72:2159.
60. Ikeda S, Tanaka M, Matsumoto S, et al. Endoscopic sphincter-otomy: Long-term results in
408 patients with complete follow-up. Endoscopy 1988;20:137.
61. Vaira D, Ainley C, Williams S, et al. Endoscopic sphincterotomy in 1000 consecutive patients.
Lancet 1989;2:4314.
62. Hansell DT, Millar MA, Murray WR, et al. Endoscopic sphincter-otomy for bile duct stones in
patients with intact gallbladders. Br J Surg 1989;76:8568.
63. Lambert ME, Betts CD, Hill J, et al. Endoscopic sphincterotomy: The whole truth. Br J Surg
1991;78:4736.
64. Hill J, Martin DF, Tweedle DEF. Risks of leaving the gallbladder in situ after endoscopic
sphincterotomy for bile duct stones. Br J Surg 1991;78:5547.
65. Moreira VF, Arribas R, Sanroman AL, et al. Choledocholithiasis in cirrhotic patients: Is
endoscopic sphincterotomy the safest choice?. Am J Gastroenterol 1991;86:100610.
66. Sherman S, Ruffolo TA, Hawes RH, Lehman GA. Complications of endoscopic
sphincterotomy. A prospective series with emphasis on the increased risk associated with
sphincter of Oddi dysfunction and nondilated bile ducts. Gastroenterology 1991;101:106875.
67. Safrany L, Neuhaus B, Portocarrero G, Krause S. Endoscopic sphincterotomy in patients with
Billroth II gastrectomy. Endoscopy 1980;12:1622.
68. Rosseland AR, Osnes M, Kruse A. Endoscopic sphincterotomy (EST) in patients with Billroth
II gastrectomy. Endoscopy 1981;13:1924.
69. Forbes A, Cotton PB. ERCP and sphincterotomy after Billroth II gastrectomy. Gut
1984;25:9714.
70. Rosseland AR. Endoscopic papillotomy. Scand J Gastroenterol 1987;22:11539.
71. Weizel A, Stiehl A, Raedsch R. Passage of a large bilirubin stone through a narrow
papillotomy. Endoscopy 1980;12:1913.
72. Siegel JH, Ben-Zvi JS, Pullano WE. Mechanical lithotripsy of common duct stones.
73. Chung SCS, Leung JWC, Leong HT, Li AKC. Mechanical lithotripsy of large common duct
stones using a basket. Br J Surg 1991;78:144850.
74. Siegel JH, Ben-Zvi JS, Pullano WE. Endoscopic electrohydraulic lithotripsy. Gastrointest
Endosc 1990;36:1346.
75.
Leuschner U, Baumgrtel H, Klempa J. Chemical treatment of choledocholithiasis. In Classen
M, Geenen J, Kawi K (eds). Non-surgical Biliary Drainage. Berlin, Heidelberg:
76. Swobodnik W, Wechsler JG, Kluppelberg U, et al. Auflsung von Rezidivsteinen im
Choledochus durch modifizierte Splbehandlung ber eine nasobilire Verweilsonde. Dtsch
77. Tandon RK, Nijhawan S, Arora A. Management of retained common bile duct stones in
patients with T-tube in situ: Role of endoscopic sphincterotomy. Am J Gastroenterol
1990;85:112631.
78. Cotton PB. Retained bile duct stones. T-tube in place, percutaneous or endoscopic
management [Editorial]. Am J Gastroenterol 1990;85:10758.
79. Siegel JH. Biliary bezoar: The sump syndrome and choledochoenterostomy. Endoscopy
1982;14:23840.
80. Baker AR, Neoptolemos JP, Carr-Locke DL, et al. Sump syndrome following
choledochoduodenostomy and its endoscopic treatment. Br J Surg 1985;72:4335.
81. Tanaka M, Ikeda S, Yoshimoto H. Endoscopic sphincterotomy for the treatment of biliary
sump syndrome. Surgery 1983;93:2647.
82. Al Karawi MA, El Sheikh Mohamed AR, Sultan Khurro M, Neuhaus H. Bedeutung der
Endoskopie in der Diagnostik und Therapie gastrointestinaler und biliarer Parasiten. Internist
1988;29:80714.
83.
Phillip J, Classen M. Interventions endoscopiques sur la papille et les voies biliaires. In Hess
W, Rohner A, Cirenei A, Akovbiantz A (eds). Maladies des voies biliaires et du pancreas.
Padua, Piccin, 1992;pp 15811618.
84. Houang MT, Franklin A, Arozena X. Ascaris lumbricoides detected in common bile duct
following routine cholecystectomy and common bile duct sphincterotomy. Br J Radiol
1980;53:8045.
85.
Acosta JM, Ledesma CL. Gallstone migration as a cause of acute pancreatitis. N Engl J Med
1984;190:4847.
86. Cotton PB. Congenital anomaly of pancreas divisum can cause obstructive pain and
87. Carr-Locke DL. Pancreas divisum: The controversy goes on. Endoscopy 1991;23:8890.
88.
DiMagno EP, Holtmann G. Chronic pancreatitis. Curr Opin Gastroenterol 1992;8:8249.
89. Cotton PB. Pancreas divisum: Curiosity or culprit?. Gastroenterology 1985;89:14315.
90. Lans JI, Geenen JE, Johanson JF, Hogan WJ. Endoscopic therapy in patients with pancreas
divisum and acute pancreatitis: A prospective, randomized, controlled clinical trial.
91. Kozarek RA, Ball TJ, Patterson DJ. Endoscopic approach to pancreatic duct calculi and
obstructive pancreatitis. Am J Gastroenterol 1992;87:6003.
92. Delhaye M, Vandermeeren A, Baize M, Cremer M. Extracorporeal shock-wave lithotripsy of
pancreatic calculi. Gastroenterology 1992;102:61020.
93. Neuhaus H, Hoffmann W, Classen M. Laser lithotripsy of pancreatic and biliary stones via 3.4
mm and 3.7 mm miniscopes: First clinical results. Endoscopy 1992;24:20814.
94. Classen M, Leuschner U, Schreiber HW. Stenosis of papilla vateri and common duct calculi.
95. Seifert E, Gail K, Weismller J. Langzeitresultate nach endoskopischer Sphinkterotomie.
Dtsch med Wochenschr 1982;107:6104.
96. Bourgeois N, Dunham F, Verhest A, Cremer M. Endoscopic biopsies of the papilla of Vater at
the time of endoscopic sphincterotomy: Difficulties in interpretation. Gastrointest Endosc
1984;30:1636.
97. Farrell RJ, Khan MI, Noonan N, et al. Endoscopic papillectomy: A novel approach to difficult
cannulation. Gut 1996;39:368.
98. Helm EB, Bauernfeind F, Frech K, Hagenmller F. Pseudomonas-Septikmie nach
endoskopischen Eingriffen am Gallengangsystem. Dtsch Med Wochenschr
1984;109:697701.
99. Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and
the impact of endoscopic sphincter-otomy. Br J Surg 1986;73:98892.
100. Simon DM, Brooks S, Hersh T. Endoscopic sphincterotomy: A reappraisal. Am J
101. Lam SK. A study of endoscopic sphincterotomy in recurrent pyogenic cholangitis. Br J Surg
1984;71:2626.
102. Burmeister W, Wurbs D, Hagenmller F, Classen M. Langzeituntersuchungen nach
endoskopischer Papillotomie. Z Gastroenterol 1980;18:52731.
103. Hawes RH, Cotton PB, Vallon AG. Follow-up 6 to 11 years after duodenoscopic
sphincterotomy for stones in patients with prior cholecystectomy. Gastroenterology
1990;98:100812.
104. Seifert E, Schulte F, Chalybans C. Quo vadis endoskopische Spinkterotomie. Z Gastroenterol
1989;27:7781.
105.
Hagenmller F, Wurbs D, Classen M. Long-term complications after endoscopic papillotomy
(EPT) in patients with gallbladder in situ [Abstract]. Endoscopy 1973;4:2834.
106. Riemann JF, Gierth K, Lux G, et al. Die belassene SteingallenblaseEin Risikofaktor nach
endoskopischer Papillotomie?. Z Gastroenterol 1984;22:18893.
107. Neoptolemos JP, Carr-Locke DL, Fraser I, Fossard DP. The management of common bile
duct calculi by endoscopic sphincter-otomy in patients with gallbladders in situ. Br J Surg
1984;71:6971.
108. Cotton PB, Vallon AG. Duodenoscopic sphincterotomy for removal of bile duct stones in
patients with gallbladders. Surgery 1982;91:62830.
109. Ingoldby CJH, El Saadi J, Hall RI, Denyer ME. Late results of endoscopic sphincterotomy for
bile duct stones in elderly patients with gallbladders in situ. Gut 1989;30:112931.
110. Martin DF, Tweedle DE. Endoscopic management of common duct stones without
111. Moss JG, Saunders JH, Wild SR. Endoscopic papillotomy for removal of common bile duct
stones without cholecystectomy. J R Coll Surg Edinb 1985;30:1124.
112. Olaison G, Kald B, Karlqvist PA, et al. Endoscopic removal of common bile duct stones
without subsequent cholecystectomy. Acta Chir Scand 1987;153:5413.
113. Rosseland AR, Solhaug JH. Primary endoscopic papillotomy (EPT) in patients with stones in
the common bile duct and the gallbladder in situ: A 58-year follow-up study. World J Surg
1988;12:1116.
114. Saraswat VA, Kapur BML, Vashisht S, Tandon RK. Duodenoscopic sphincterotomy for
common bile duct stones in patients with gallbladder in situ. Int Surg 1991;76:1425.
115. Siegel JH, Safrany L, Ben-Zvi JS, et al. Duodenoscopic sphincterotomy in patients with
gallbladders in situ: Report of a series of 1272 patients. Am J Gastroenterol 1988;83:12558.
116. Tanaka M, Ikeda S, Yoshimoto H, Matsumoto S. The long-term fate of the gallbladder after
endoscopic sphincterotomy. Am J Surg 1987;154:5058.
117. Jacobson O, Matzen P. Long-term follow-up study of patients after endoscopic

sphincterotomy for choledocholithiasis. Scand J Gastroenterol 1987;22:9036.
118. Shemesh E, Klein E, Czerniak A, et al. Endoscopic sphincterotomy in patients with gallbladder
in situ: The influence of periampullary duodenal diverticula [see comments]. Surgery
1990;107:1636.
119. Hammarstrom LE, Holmin T, Stridbeck H. Endoscopic treatment of bile duct calculi in patients
with gallbladder in situ: Long-term outcome and factors. Scand J Gastroenterol
1996;31:294301.
120. Holbrook RF, Jacobson FL, Pezzuti RT, Howell DA. Biliary patency imaging after endoscopic
retrograde sphincterotomy with gallbladder in situ. Clinical impact of nonvisualization. Arch
Surg 1991;126:73841; discussion 7412.
121. Worthley CS, Toouli J. Gallbladder non-filling: An indication for cholecystectomy after
endoscopic sphincterotomy [see comments]. Br J Surg 1988;75:7968.
122. Targarona EM, Ayuso RM, Bordas JM, et al. Randomised trial of endoscopic sphincterotomy
with gallbladder left in situ versus open surgery for common bile duct calculi in high-risk
patients. Lancet 1996;347:9269.
123.
Cremer M, Toussaint J, Dunham F, Jeanmart J. Endoscopic sphincterotomy (ES) with
gallbladder in situ [Abstract]. Gastrointest Endosc 1981;27:141.
124.
Tulassay Z, Papp J. Endoskopische Sphinkterotomie bei Patienten mit Gallenblase in situ. Z
sphincterotomy [see comments]. N Engl J Med 1996;335:90918.
126.
Mee AS, Vallon AG, Croker JR, Cotton PB. Non-operative removal of bile duct stones by
duodenoscopic sphincterotomy in the elderly. BMJ 1981;283:5213.
127. Deans GT, Sedman P, Martin DF, et al. Are complications of endoscopic sphincterotomy
age-related?. Gut 1997;41:5458.
128. Cotton PB. Precut papillotomyA risky technique for experts only. Gastrointest Endosc
1989;35:5789.
129. Coppola R, Riccioni ME, Ciletti S, et al. Analysis of complications of endoscopic
sphincterotomy for biliary stones in a consecutive series of 546 patients. Surg Endosc
1997;11:12932.
130. Wilson MS, Tweedle DEF, Martin DF. Common bile duct diameter and complications of
endoscopic sphincterotomy. Br J Surg 1992;79:13467.
131. Keith RG, Shapero TF, Saibil FG, Moore TL. Dorsal duct sphincterotomy is effective
long-term treatment of acute pancreatitis associated with pancreas divisum [see comments].
Surgery 1989;106:6606; discussion 6667.
132. Nelson DB, Freeman ML. Major hemorrhage from endoscopic sphincterotomy: Risk factor
analysis. J Clin Gastroenterol 1994;19:2837.
133. Leung JW, Chan FK, Sung JJ, Chung S. Endoscopic sphincterotomy-induced hemorrhage: A
study of risk factors and the role of epinephrine injection. Gastrointest Endosc 1995;42:5504.
134. Goodall RJ. Bleeding after endoscopic sphincterotomy. Ann R Coll Surg Engl 1985;67:878.
135. Finnie IA, Tobin MV, Morris AI, et al. Late bleeding after endoscopic sphincterotomy for bile
duct calculi. BMJ 1991;302:1144.
136. Mellinger JD, Ponsky JL. Bleeding after endoscopic sphincterotomy as an underestimated
entity. Surg Gynecol Obstet 1991;172:4659.
137. Gholson CF, Favrot D, Vickers B, et al. Delayed hemorrhage following endoscopic retrograde
sphincterotomy for choledocholithiasis. Dig Dis Sci 1996;41:8314.
138. Staritz M, Ewe K, Goerg K, Meyer zum Bschenfelde KH. Endoscopic balloon tamponade for
conservative management of severe hemorrhage following endoscopic sphincterotomy. Z
139. Grimm H, Soehendra N. Unterspritzung zur Behandlung der Papillotomieblutung. Dtsch Med
Wochenschr 1983;108:15124.
140. Lo SK, Patel A. Treatment of endoscopic sphincterotomy-induced hemorrhage: Injection of
bleeding site with ERCP contrast solution using a minor papilla diagnostic catheter.
141. Sherman S, Hawes RH, Nisi R, Lehman GA. Endoscopic sphincterotomy-induced
hemorrhage: Treatment with multipolar electrocoagulation. Gastrointest Endosc
1992;38:1236.
142. Saeed M, Kadir S, Kaufmann SL, et al. Bleeding following endoscopic sphincterotomy:
Angiographic management by transcatheter embolization. Gastrointest Endosc
1989;35:3003.
143. Khuroo MS, Yattoo GN, Zargar SA, et al. Biliary abnormalities associated with extrahepatic
portal venous obstruction. Hepatology 1993;17:80713.
144. Bayraktar Y, Balkanci F, Kayhan B, et al. Bile duct varices or pseudo-cholangiocarcinoma
sign in portal hypertension due to cavernous transformation of the portal vein. Am J
145. Dilawari JB, Chawla YK. Pseudosclerosing cholangitis in extrahepatic portal venous
obstruction. Gut 1992;33:2726.
146. Tighe M, Jacobson I. Bleeding from bile duct varices: An unexpected hazard during
therapeutic ERCP. Gastrointest Endosc 1996;43:2502.
147. Evrard S, Mendoza L, Mutter D, et al. Massive gas spread through a duodenal perforation
after endoscopic sphincterotomy. Gastrointest Endosc 1993;39:8178.
148. Savides T, Sherman S, Kadell B, et al. Bilateral pneumothoraces and subcutaneous
emphysema after endoscopic sphincterotomy. Gastrointest Endosc 1993;39:8147.
149. Tam F, Prindiville T, Wolfe B. Subcutaneous emphysema as a complication of endoscopic
sphincterotomy of the ampulla of Vater. Gastrointest Endosc 1989;35:4479.
150. Ciaccia D, Branch MS, Baillie J. Pneumomediastinum after endoscopic sphincterotomy. Am J
151. Dunham F, Bourgeois N, Gelin M, et al. Retroperitoneal perforation following endoscopic
sphincterotomy: Clinical course and management. Endoscopy 1982;14:926.
152. Kim T, Messersmith R. Submucosal choledochal emphysema after transduodenal
sphincteroplasty. Am J Gastroenterol 1988;83:1736.
153. Simmons TC. Hepatic portal venous gas due to endoscopic sphincterotomy. Am J
154. Herman JB, Levine MS, Long WB. Portal venous gas as a complication of ERCP and
endoscopic sphincterotomy. Am J Gastroenterol 1995;90:8289.
155. Martin DF, Tweedle DE. Retroperitoneal perforation during ERCP and endoscopic
sphincterotomy: Causes, clinical features and management. Endoscopy 1990;22:1745.
156. Byrne P, Leung JW, Cotton PB. Retroperitoneal perforation during duodenoscopic
sphincterotomy. Radiology 1984;150:3834.
157. Kuhlman JE, Fishman EK, Milligan FD, Siegelman SS. Complications of endoscopic
retrograde sphincterotomy: Computed tomographic evaluation. Gastrointest Radiol
1989;14:12732.
158. Low DE, Shoenut JP, Kennedy JK, et al. Risk of bacteremia with endoscopic sphincterotomy.
Can J Surg 1987;30:4213.
159. Sauter G, Grabein B, Huber G, et al. Antibiotic prophylaxis of infectious complications with
endoscopic retrograde cholangiopancreatography. A randomized controlled study. Endoscopy
1990;22:1647.
160. Sherman S, Lehman GA. ERCP- and endoscopic sphincterotomy-induced pancreatitis.
Pancreas 1991;6:35067.
161. Halter F, Bangerter U, Gigon JP, Pusterla C. Gallstone ileus after endoscopic sphincterotomy.
162. Gandhi A, Maxwell AJ, Wells S, Hobbiss JH. Gallstone ileus following endoscopic
sphincterotomy. Br J Hosp Med 1995;54:22930.
163. Despland M, Clavien PA, Mentha G, Rohner A. Gallstones ileus and bowel perforation after
endoscopic sphincterotomy. Am J Gastroenterol 1992;87:8868.
164. Paumgartner G, Sauerbruch T. Secretion, composition and flow of bile. Clin Gastroenterol
1983;12:323.
165. Sauerbruch T, Stellaard F, Paumgartner G. Effect of endoscopic sphincterotomy on bile acid
pool size and bile lipid composition in man. Digestion 1983;27:8792.
166.
Classen M, Kurtz W, Hagenmller F. The gallbladder after endoscopic papillotomy. In Bianchi
L, Gerok W, Popper H, et al. (eds). Trends in Hepatology. Lancaster: MTP, 1985;pp 739.
167. Stellaard F, Sauerbruch T, Brunholzl C, et al. Bile acid pattern and cholesterol saturation of
bile after cholecystectomy and endoscopic sphincterotomy. Digestion 1983;26:1538.
168.
Chijiiwa K, Ogawa Y, Hirota I, Nakayama F. Biliary lipids and bile acid composition before and
after endoscopic sphincterotomy. Hepatogastroenterology 1990;5:5102.
169.
Rsch W, Koch H, Demling L. Manometric studies during ERCP and endoscopic papillotomy.
170.
Staritz M, Schmidt HD, Meyer zum Bschenfelde KH. Einflu("Symbol")b der Cholecystektomie
auf die Funktion der Papilla Vateri: Eine pr- und postoperative endoskopisch manometrische
Studie [Abstract]. Z Gastroenterol 1983;21:423.
171. Staritz M, Ewe K, Meyer zum Bschenfelde KH. Investigation of the sphincter of Oddi before,
immediately after and six weeks after endoscopic papillotomy. Endoscopy 1986;18:146.
172. Gregg JA, De Girolami P, Carr-Locke DL. Effects of sphincteroplasty and endoscopic
sphincterotomy on the bacteriologic characteristics of the common bile duct. Am J Surg
1985;149:66871.
173. Greenfield C, Cleland P, Dick R, et al. Biliary sequelae of endoscopic sphincterotomy.
174. Tytgat GN, Meenan JK, Rauws EA, Huibregtse K. Endoscopic biliopancreatic balloon dilation.
Endoscopy 1996;28:36771.
175. Staritz M, Poralla T, Dormeyer HH, Meyer zum Bschenfelde KH. Endoscopic removal of
common bile duct stones through the intact papilla after medical sphincter dilation.
176. Uchida N, Ezaki T, Hirabayashi S, et al. Endoscopic lithotomy of common bile duct stones
with sublingual nitroglycerin and guidewire [see comments]. Am J Gastroenterol
1997;92:14403.
177. Staritz M, Ewe K, Meyer zum Bschenfelde KH. Endoscopic papillary dilation (EPD) for the
treatment of common bile duct stones and papillary stenosis. Endoscopy 1983;15:1978.
178. Mac Mathuna P, Siegenberg D, Gibbons D, et al. The acute and long-term effect of balloon
sphincteroplasty on papillary structure in pigs. Gastrointest Endosc 1996;44:6505.
179. Minami A, Nakatsu T, Uchida N, et al. Papillary dilation vs sphincterotomy in endoscopic
removal of bile duct stones. A randomized trial with manometric function. Dig Dis Sci
1995;40:25504.
180. Mac Mathuna PM, White P, Clarke E, et al. Endoscopic balloon sphincteroplasty (papillary
dilation) for bile duct stones: Efficacy, safety, and follow-up in 100 patients. Gastrointest
Endosc 1995;42:46874.
181. Sato H, Kodama T, Takaaki J, et al. Endoscopic papillary balloon dilation may preserve
sphincter of Oddi function after common bile duct stone management: Evaluation from the
viewpoint of endoscopic manometry. Gut 1997;41:5414.
182. Bergman JJ, Rauws EA, Fockens P, et al. Randomised trial of endoscopic balloon dilation
versus endoscopic sphincterotomy for removal of bile duct stones. Lancet 1997;349:11249.
Chapter 61 Special Papillotomy Techniques

(6230)
(6231)
KEES HUIBREGTSE, M.D., PH.D.
JOHN MEENAN, M.D., PH.D., M.R.C.P.
VINOOL K. PARASHER, M.D.

E. A. J. RAUWS, M.D.
VINAY DHIR, M.D., D.N.D.
Free cannulation of the common bile duct (CBD) cannot be achieved in from 1 to 20% of cases. This
may result from an acute angulation of the distal duct, from altered anatomy due to tumor or congenital
anomalies, from an impacted calculus, or from edema of the papillary region. The number of options
for dealing with such obstacles has increased in recent years.
Developments in catheter, guidewire, and specialized sphincterotome technology have reduced
reliance on more aggressive methods for deep cannulation. However, use of such instruments does
not always result in success. It is in these cases that needle-knife papillotomy has a useful role. This
procedure should be performed only after careful review of the clinical indication, and only by, or under
the supervision of, a highly experienced endoscopist.
Failure of standard maneuvers to achieve free cannulation of the CBD obliges the endoscopist to
consider the following options: repeating the procedure in several days' time, referring the patient to a
more experienced endoscopist, intervening radiologically, or using needle-knife papillotomy. The
decision to perform a papillotomy requires a strong indication based on radiologic, biochemical, or
clinical evidence.
Papillotomy Techniques
The term papillotomy, as used in this chapter, refers to the cutting of the superficial papillary sphincter
of the main duodenal papilla. The term sphincterotomy indicates a breach of the deeper sphincter.
However, in continental Europe and parts of Asia, these terms are used interchangeably. Papillotomy
can be achieved with a precut papillotome or a needle-knife.
These two devices should never be used to compensate for the lack of experience and ability. It must
be emphasized that in cases where cannulation is initially unsuccessful, exasperation is best
channeled toward ensuring that the major papilla has been correctly identified, improving the position
of the endoscope relative to the papilla, correctly predicting the course of the CBD by endoscopic
observation, and properly directing the cannula tip based on fluoroscopic observation (see Chapter 57:
Technique of Endoscopic Retrograde Cholangiopancreatography; see also Chapter 60: Endoscopic
Papillotomy). Hydrophilic-coated guidewires are useful if there is a distal CBD stricture. The
double-lumen, short-nosed Cotton cannulotome (Wilson Cook Medical, Inc., Winston-Salem, NC) is
helpful should endoscope position not permit adequate angulation of the standard catheter with respect
to the papilla.
Free cannulation of the CBD is achievable in 80 to 90% of cases through the employment of improved
guidewires and specialized cannulotomes. In the remaining cases, the judicious application of
needle-knife papillotomy can significantly increase the success of therapeutic endoscopic retrograde
cholangiopancreatography (ERCP). Such success depends on the correct identification of anatomic
landmarks, adequate control of the knife tip, and an unhurried approach.
Precut Papillotomy
Early attempts at papillotomy utilized the precut papillotome.14(6232) This consists of a modified
sphincterotome in which the cutting wire extends to the cannula tip (Figure 611). After insertion into
the papillary orifice, the cautery wire is tensed slightly in an 11 to 12 o'clock direction to approximate
the intramural course of the CBD. In cases of impacted calculi, multiquadrant precut papillotomy has
been performed.5(6233)
(6234)Figure 611. Left, Precut sphincterotome. Note that the cutting wire extends to the tip
of the instrument. Right, Needle-knife papillotome with cautery wire extended. (Left and right,
Courtesy of Wilson Cook Medical, Inc., Winston-Salem, NC.)
The safety of precut papillotomy is compromised by the instability of the papillotome position within the
papilla. This reduces control over the direction and depth of cutting. In addition, there is an increased
risk of pancreatic tissue damage when there is a short common channel (see also Chapter 58:
Anatomy and Embryology of the Biliary Tract and Pancreas).6(6235) For these reasons, use of the
needle-knife may be preferable.7(6236)
Needle-Knife Papillotomy
Needle-knife for papillotomy was a natural progression from the initial fashioning of an endoscopic
choledochoduodenostomy by Osnes et al.8,9(6237) The needle-knife consists of a 0.2-mm diameter
straight, retractable wire that can be extended up to 5 mm from the tip of a 5-French catheter (see
Figure 611). This is connected to an electrosurgery generator, which should be set for blended
current. The cutting characteristics of the needle-knife are closely related to the thickness of wire: A
fine wire will have a greater cutting effect and induces less tissue damage than one of greater diameter
(see Chapter 9: Principles of Electrosurgery).
Before employing the needle-knife, the endoscopist must ensure a stable endoscope position with a
clear view of the papilla and of the intramural segment of the CBD (as described in Chapter 57:
Technique of Endoscopic Retrograde Cholangiopancreatography). Overlying folds need to be pushed
away, and administration of antimuscarinic drugs may be required to reduce gut motility. Several
practice "golf-swings" should be made while the needle is still within the catheter, starting at the
papillary orifice, in an 11 o'clock direction (Figure 612). This allows the endoscopist to assess
sideways movement of the cannula within the endoscope channel and to judge the best combination of
endoscope movements to permit optimum directing of the knife. Successful use of the needle-knife is
completely dependent on control of movement. No single, fixed set of movements can be applied in all
cases, but the combination of gentle upward pressure on the elevator with slow retraction of the
endoscope may be helpful.
(6238)Figure 612. Idealized diagram demonstrates the 11 o'clock cutting direction for
needle-knife papillotomy.
The diathermy wire is extended 3 to 5 mm. Papillotomy is achieved through successive shallow cuts
starting at the orifice and extending proximally in the 11 o'clock direction toward the visible elevation
caused by the CBD. This incision should be extended no farther than the junction of the papillary
complex with the duodenal wall, that is, the cut should be confined to the visibly elevated papilla.
Current is applied only when the knife is moving in order to avoid excessive burning that may result in
edema or damage to the pancreatic duct orifice.
After each cut, the incision is inspected for direction by identifying the bulge of the sphincteric muscle.
Incision into this muscle results in the visible spreading of muscle fibers. The base of the cut is
inspected for a show of bile or the red-yellow spot of CBD mucosal prolapse. Gentle probing of the
incision should be performed with a ball-tip cannula because using the tip of the needle-knife sheath
for this purpose can result in blunt tissue injury. It is important to realize that the orifice of the CBD
does not always lie at the apex of the cut. While probing, injection of contrast medium is to be avoided
because the probability of an intramucosal injection is greater than that for opacification of the CBD.
Successive incisions should start at the site of the original papillary opening. This makes it easier to
maintain the correct anatomic orientation of the papillotomy. Siegel et al.10(6239) advocate an
opposite CBD-to-papilla direction for cutting and report that the complication rate associated with this
approach is low. However, Thornton and Axon7(6240) point out that such downward movements may
reduce the ability to control the knife tip.
Again, it is important to emphasize that adequate time be given to inspection between cuts. If required,
the sphincterotomy should be completed with a standard sphincterotome at the same session. Delayed
sphincterotomy to allow reepithelialization, as has been recommended,3(6241) confers no benefit. If
edema or bleeding makes it difficult to see the papillotomy site, it may be best to postpone completion
of the procedure for 3 to 5 days.
Indications for Needle-Knife Papillotomy

Impacted Calculi
Use of needle-knife papillotomy in cases of impacted ampullary calculi seems especially
appropriate.11(6242) The pancreatic duct is in a protected position behind the stone and the risk of
perforation is less owing to the large surface presented by the stretched CBD.12(6243) The impacted
stone is usually easy to remove, although occasionally the CBD proximal to the point of impaction may
collapse, and thereby, it may become difficult to cannulate. The distal papilla remnant can be opened
in a retrograde direction with the needle-knife, or a standard sphincterotome can be passed to
complete the incision.
Ampullary Lesions
Ampullary lesions are often associated with bulging of the intraduodenal segment of the CBD. The
needle-knife can be used to fashion a proximal choledochoduodenostomy, permitting placement of an
endoprosthesis.13(6244) It is important to bear in mind that what appears to be a bulging CBD may in
fact be submucosal tumor containing a compressed duct.
Pancreatic Duct Sphincterotomy

Endoscopic treatment of chronic pancreatitis provides an alternative to surgery in the short term (see
Chapter 76: Endoscopic Management of Pancreatic Disease).1416(6245) Fuji et al.17(6246)
described a technique for pancreatic sphincterotomy using a sphincterotome with a short (10-mm)
cutting wire. This is introduced into the papilla and a cut (5 mm) extending to the covering fold is made
in a 2 o'clock direction. After this, a biliary sphincterotomy is performed. Alternatively, the sequence
can be reversed, with the biliary sphincterotomy being performed first. Endoprostheses were not
routinely placed as part of the procedure described by Fuji et al.17(6247) The needle-knife may also be
employed in the performance of this procedure. The initial placement of a 7-French endoprosthesis
aids the endoscopist in guiding the direction of the incision.
Minor Papilla
Cannulation of the minor papilla is achievable in up to 90% of cases with pancreas divisum,18(6248)
with successful endoprosthesis placement in a similar percentage.19(6249)
Papillotomy of the minor papilla may be required in the treatment of pancreas divisum. After
cannulation with a taper-tip catheter, a stent (5 to 7 French) can be placed to act as a guide for a
subsequent needle-knife incision in a 9 to 11 o'clock direction. The stent is removed some days later.
Reported rates of papillary restenosis vary, possibly because of differences in the size of the
papillotomy.2022(6250) The reported incidence of pancreatitis after this procedure ranges from 13 to
70%.21,22(6251)
Complications
The success rate of therapeutic ERCP can be increased by the judicious use of needle-knife
papillotomy.6,23(6252) Although the complications of this procedure (hemorrhage, perforation, and
pancreatitis) are similar to those attending sphincterotomy, these risks must be weighed against the
fact that there is no guarantee of duct cannulation or benefit to the patient. In experienced hands,
however, needle-knife papillotomy carries an acceptable rate of complications and is comparable in
this respect to standard sphincterotomy.
The reported frequency of use of needle-knife papillotomy varies significantly between centers, with
figures ranging from 3.8 to 19.2% (Table 611). Complication rates for these groups are in the range of
2.6 to 20%, with hemorrhage, perforation, and pancreatitis being the most common complications. The
discrepancies in these figures are a reflection of the variable numbers of patients in each series,
differences in the types of cases treated, the experience of the personnel, and use of dissimilar
definitions for complications.24(6253)
TABLE 611
Needle-Knife Papillotomy
FREQUENCY
OF USE (%)
ES COMPLICATION/
MORTALITY (%)
NK COMPLICATION/
MORTALITY (%)
19.2
2.1/0
2.6/0
<1
6.9/0.6
/0
6.7/0
12.8
5.2/2
Tweedle and Martin,

199154
Leung et al., 199011
4.2
7/0
12/0
3.9
20/0
Booth et al., 199025
9.7
9/0
16/1
Sherman et al., 199126

Shakoor et al., 199228
15.4
7/0
6.2/0
3.8
6/0.1
11/0
6.3
11
9.7/0
5.7/0
REFERENCE
Huibregtse et al., 198612
Vaira et al., 198923
Siegel et al., 198910
Dowsett et al., 19896
Martin and Tweedle,

199227
Foutch, 199555
ESendoscopic sphincterotomy; NKneedle-knife papillotomy.
The morbidity rates reported by Leung et al.11(6254) and Booth et al.25(6255) (20% and 16%,
respectively) are sobering for those contemplating the use of needle-knife papillotomy. However, the
similar complication rates for standard sphincterotomy and needle-knife papillotomy reported by
others12,26(6256) indicate that needle-knife papillotomy can be a relatively safe procedure when the
circumstances at endoscopy are favorable. This is reflected in the complication rates reported by
Siegel et al.10(6257) (6.7%), Dowsett et al.6(6258) (5.2%), and Martin and Tweedle27(6259) (6.3%).
Why then do the various reports differ with respect to complication rates?
The comparison of percentages without qualification may not be appropriate. Most series pertain to
use of the needle-knife in tens of cases, as compared with series in which the standard
sphincterotome has been used in hundreds or thousands of cases.21,28(6260) Those series in which
a needle-knife was used in over 12% of cases have lower complication rates (2.6 to
6.7%).6,12,26(6261) Although these results may be explained partly in terms of experience, the
vagaries of statistical comparison must also be taken into account.
The types of cases included in each trial must influence complication rates. The 20% morbidity in the
series of Leung et al.11(6262) is put into perspective when it is noted that all these patients had
undergone a previously unsuccessful sphincterotomy. Sherman et al.26(6263) found that patients
undergoing precut papillotomy for sphincter of Oddi dysfunction have a greater complication rate
(8.6%) compared with patients who undergo the procedure for other indications (3.3%).
Probably the single most important factor that influences the complication rate associated with
needle-knife papillotomy is the experience of the endoscopist.28(6264) The results of a large
multicenter trial from the United States and Canada confirm that the frequency of therapeutic ERCP
complications correlates with the experience of the endoscopist.29(6265) In this study, use of precut
technique to gain access to the CBD was identified specifically as a risk factor.
The value of obtaining a coagulation screen before standard sphincterotomy has been
questioned.30(6266) At worst, however, the knowledge that there is no coagulopathy is one less worry
for the endoscopist. Bleeding, should it occur, is usually minor and stops spontaneously. Injection of
adrenalin (1:10,000 solution) has achieved hemostasis when oozing of blood does not stop
spontaneously.11(6267) If the procedure has been interrupted because of bleeding and a second
session is scheduled 3 to 5 days later, ample time should be allotted to inspect the papillotomy site for
the CBD orifice. Bleeding can recur with even minor trauma such as probing with a diagnostic cannula.
Perforation secondary to needle-knife papillotomy is retroperitoneal (Figure 613). The endoscopist
should be wary of an abnormal contrast pattern including a "tramline" effect along the CBD. A cannula
that follows an unusual course when seemingly inserted into the CBD is also an indication that a
perforation has occurred. In the majority of cases, the perforation is identified during the endoscopy
and can be treated conservatively with fasting, intravenous fluids, and antibiotics.31(6268) Some
authors recommend that a plain x-ray of the right upper quadrant be obtained with the patient supine
as a routine procedure after a needle-knife sphincterotomy.27(6269) Others advise more rigorous
short-term follow-up including anteroposterior and oblique x-ray films with regular evaluations of the
patient for symptoms and physical findings of perforation.25(6270) If a perforation has occurred, biliary
drainage must be achieved either endoscopically or by the percutaneous route. If a perforation is
identified during ERCP, the procedure should be continued to completion with achievement of drainage
by insertion of a prosthesis or nasobiliary drain.
(6271)Figure 613. Retroperitoneal perforation after needle-knife papillotomy. Contrast and

air can be seen outside the borders of the common bile duct. The procedure was completed
with placement of a 10-French endoprosthesis for biliary drainage. Subsequent management
was conservative.
By extrapolation from a study of bacteremia related to diagnostic and therapeutic ERCP,32(6272) the
routine administration of prophylactic antibiotics to patients specifically undergoing needle-knife
papillotomy does not seem warranted. Placement of a pancreatic duct stent for several days after
papillotomy has been advocated,26(6273) but there are no data to support this practice.
Sphincterotomy After Billroth II Gastrectomy

An increased prevalence of cholelithiasis in patients who have undergone gastrectomy is well
recognized.33(6274) The Billroth II gastrectomy, although having certain physiologic advantages for
patients, places two potential constraints on therapeutic ERCP: (1) the type of anastomosis (retrocolic,
antecolic, or Roux-en-Y) may prohibit entry into the afferent loop or localization of the papilla (see also
Chapter 52: Endoscopy in the Postoperative Upper Gastrointestinal Tract),34(6275) and (2) the
inverted position of the endoscope relative to the intramural segment of the CBD makes
sphincterotomy technically more difficult.
Many investigators have attempted to overcome these obstacles by developing specialized instruments
and techniques. These include graduated endoscope stiffeners,35(6276) steerable catheters,36(6277)
specialized sphincterotomes,37,38(6278) descending sphincterotomy through a T-drain,39(6279) and
instillation of monooctanoin through a nasobiliary drain.40(6280) However, the single most important
technical development remains the introduction of duodenoscopes with distal tips that are short.
Early reports of ERCP after Billroth II gastrectomy presented uniformly disappointing results with
success rates of 55 to 69%.34,4143(6281) More recent studies indicate that it is possible to reach the
papilla using either a side-viewing or a forward-viewing endoscope in over 90% of
cases.36,44,45(6282) Cannulation of the required duct and biliary sphincterotomy is achievable in a
similar proportion. A combined percutaneous-endoscopic approach to biliary stenting may be
successful in cases where the use of endoscopic methods alone has failed.46(6283) Despite improved
success rates with endoscopic techniques alone, some investigators have recommended the
percutaneous approach as an initial step.47(6284)
The standard side-viewing duodenoscope remains the instrument of choice for the initial attempt at
ERCP in patients who have previously undergone a Billroth II gastrectomy. The afferent loop is usually
seen as the uppermost orifice when looking at the anastomosis head-on. Unfortunately, it is also the
most difficult to intubate. Fluoroscopy provides the quickest way to identify which loop the endoscope
has entered. Rapid transit of the endoscope in a straight configuration toward the pelvis indicates that
the instrument is in the efferent loop. The use of a biopsy forceps to remove several pieces of mucosa
and thereby tattoo the entrance of the efferent loop has been suggested as a way to avoid repeated
cannulation, but this adds little information to that available from fluoroscopy.
There is no standard set of maneuvers that can be applied to intubate the afferent loop in all patients.
For the most part, maneuvers should be performed with the control wheels and the suction valve. The
endoscope should be advanced gently, keeping it in as short a configuration as possible. Although it is
tempting to persist in pushing the instrument to force it into the afferent loop, this will never succeed
and risks anastomotic rupture. This last point is particularly relevant when ERCP must be attempted in
a patient who has recently undergone gastric resection. Stretching the gut wall may make it liable to
tear with even moderate pressure.
Having negotiated the anastomosis, the performance of a biliary sphincterotomy adds yet another level
of complexity to ERCP after Billroth II surgery. As the papillary complex is in an inverted position, the
standard sphincterotomes will almost always align with the pancreatic duct. There have been
numerous attempts to overcome this problem through the employment of custom-designed
accessories or by adaptation of techniques that are intended for other purposes. Success and
complication rates in published series for sphincterotomy using various techniques in patients who
have undergone gastrectomy with a Billroth II anastomosis are presented in Table 612.
TABLE 612
REFERENCE
Endoscopic Sphincterotomy After Billroth II Gastrectomy

ES FREQUENCY
OF USE (%)
ES COMPLICATION/
MORTALITY
ES
SUCCESS (%)
OVERALL
SUCCESS OF
CANNULATION
TABLE 612
Endoscopic Sphincterotomy After Billroth II Gastrectomy
REFERENCE
ES FREQUENCY
OF USE (%)
ES COMPLICATION/
MORTALITY
ES
SUCCESS (%)
Safrany et al., 198048

Osnes et al., 198637
4.3/0
66
34
4/2
92
Ricci et al., 198953
51
4.7/0
89
Costamagna et al., 199344

Meenan et al., 199445
61
5.9/0
97
48
5.8/0*
95
OVERALL
SUCCESS OF
CANNULATION
ESendoscopic sphincterotomy.
* Immediate complications.
The technical difficulty of performing sphincterotomy in the presence of a Billroth II anatomy must not
be underestimated. This is illustrated by a report from Safrany et al.,48(6285) who performed
sphincterotomy in 23 such patients. However, an adequate sphincterotomy was achieved in only 3
patients at the first attempt, in 15 at the second, in 3 at the third, and in 2 at the fourth.
Specialized sphincterotomes have been used successfully by a number of investigators. Osnes et
al.37(6286) described the application of a long-nosed sphincterotome with an angulation midway
between the insertions of the cutting wire (the so-called 3030 papillotome). The papillotome is tensed
after deep cannulation of the CBD while the tip of the endoscope remains at some distance from the
papilla. Subsequently, the endoscope is advanced farther into the duodenum, and the sphincterotome
is pulled back to the papilla. Leach49(6287) used the Sohma-type sphincterotome through a
forward-viewing instrument. Others have used a reverse, or sigmoidal, sphincterotome44,50(6288)
(Figure 614). These sphincterotomes usually have a long leading tip to aid in cannulation of the
correct duct.
(6289)Figure 614. Custom-designed sphincterotomes for use in patients after Billroth II

gastrectomy. Left, Soehendra-type. Right, Sigmoid-type. (Left and right, Courtesy of Wilson
Cook Medical, Inc., Winston-Salem, NC.)
Rosseland et al.51(6290) introduced a technique in which a choledochoduodenostomy is fashioned
first, and subsequently, a papillotomy is performed by distal passage of a Classen-Demling-type
papillotome through the newly created opening. Bedogni et al.52(6291) suggested the use of the
needle-knife. To increase the margin of safety, this method was later modified to include prior
placement of a nasobiliary drain as a guide to ensure that the incision is made in the correct
direction.53(6292) One disadvantage of this modified method is that the endoscope must be removed
and reintroduced to allow passage of the needle-knife. This problem can be obviated if an
endoprosthesis is used as a cutting guide (Figure 615).
(6293)Figure 615. Idealized diagram shows use of a biliary endoprosthesis to guide

needle-knife papillotomy in a patient after Billroth II gastrectomy. The direction of cutting as
seen with a side-viewing instrument is downward. Inset: Endoscope position.

It is standard practice in our unit to insert either a 7- or 10-French endoprosthesis into the CBD. The
standard diagnostic ERCP catheter made of Teflon becomes curved at its distal end after it has been
used several times. In some cases, it is desirable for the purposes of cannulation that this curve not be
present. Thus, use of a new diagnostic catheter, one that has not been previously inserted through a
duodenoscope, may be advantageous. The guide catheter (7 French) used for insertion of a 10-French
endoprosthesis is stiffer than the standard Teflon catheter and has a greater tendency to maintain a
straight configuration. It may therefore facilitate correct alignment with the papilla.
The side-viewing endoscope is positioned so that the intramural CBD runs downward from 12 o'clock.
The needle-knife is moved downward from the stent toward the midportion of the papilla. As in patients
who have not had prior gastric surgery, it is advisable to take a few practice swings, keeping the cutting
wire in its catheter. If a suitable endoscope position cannot be achieved to allow proper movement of
the needle-knife, the incision can be made in an antegrade direction, that is, upward toward the point
where the endoprosthesis exits the papilla. The cutting maneuver is repeated using the endoprosthesis
as a template. There are no guidelines as to the best set of maneuvers to employ in moving the knife;
this is dictated entirely by the suitability and stability of the endoscope position. To permit extraction of
calculi, the stent is removed and left in the duodenal stump. It is retrieved at the end of the procedure.
Needle-knife papillotomy, without the presence of an endoprosthesis, in patients with Billroth II
gastrectomies is particularly dangerous because there is usually poor control over the tip of the knife.
Despite daunting obstacles, sphincterotomy in patients with a Billroth II gastrectomy is both safe and
effective. A large array of instruments has been developed to aid sphincterotomy in such cases, but
these are perhaps superfluous. Standard equipment can be used with equal success. As with
needle-knife papillotomy, proper placement of the endoscope with respect to the papilla and stability of
position are of paramount importance.
REFERENCES
1.
Siegel JH. Endoscopic papillotomy: Sphincterotomy or sphincteroplasty?. Am J Gastroenterol

1979;72:5116.
2. Sloof M, Baker R, Lavelle MI, et al. What is involved in endoscopic sphincterotomy for
gallstones?. Br J Surg 1980;67:1821.
3. Siegel JH. Precut papillotomy: A method to improve success of ERCP and papillotomy.
Endoscopy 1980;12:1303.
4. Passi RB, Raval B. Endoscopic papillotomy. Surgery 1982;92:5818.
5. Marshall JB, Stassen WN. Multiquadrant precut papillotomy for extraction of large impacted
common bile duct stone. Gastrointest Endosc 1985;31:3368.
6.
Dowsett JF, Polydorou AA, Vaira D, et al. Needle knife papillotomy: How safe and how
effective?. Gut 1989;31:9058.
7. Thornton J, Axon A. Towards safer endoscopic retrograde cholangiopancreatography. Gut
1993;34:7214.
8. Osnes M, Kahrs T. Endoscopic choledochoduodenostomy for choledocholithiasis through
choledochoduodenal fistula. Endoscopy 1977;9:1625.
9. Aabakken L, Osnes M. Endoscopic choledocho-duodenostomy (ECDT) as palliative treatment
of malignant periampullary obstructions of the common bile duct: A follow-up study.
10. Siegel JH, Ben-Zvi JS, Pullano W. The needle knife: A valuable tool in diagnostic and
therapeutic ERCP. Gastrointest Endosc 1989;35:499503.
11. Leung JW, Banez VP, Chung SC. Precut (needle knife) papillotomy for impacted common bile
duct stone at the ampulla. Am J Gastroenterol 1990;85:9913.
12. Huibregtse K, Katon RM, Tytgat GNJ. Precut papillotomy via the needle-knife papillotome: A
safe and effective technique. Gastrointest Endosc 1986;32:4035.
13. Osnes M. Endoscopic choledochoduodenostomy for common bile duct obstructions. Lancet
1979;1:105960.
14. Guelrud M, Siegel JH. Hypertensive pancreatic duct sphincter as a cause of pancreatitis.
Successful treatment with hydrostatic balloon dilatation. Dig Dis Sci 1984;29:22531.
15. Huibregtse K, Schneider B, Vrij AA, Tytgat GN. Endoscopic pancreatic drainage in chronic
pancreatitis. Gastrointest Endosc 1988;34:915.
16. Grimm H, Meyer WH, Nam VC, Soehendra N. New modalities for treating chronic
pancreatitis. Endoscopy 1989;21:704.
17. Fuji T, Amano H, Ohmura R, et al. Endoscopic pancreatic sphincterotomyTechnique and
evaluation. Endoscopy 1989;21:2730.
18. Watanapa P, Williamson RC. Pancreatic sphincterotomy and sphincteroplasty. Gut
1992;33:8657.
19. Siegel JH, Ben-Zvi JS, Pullano W, Cooperman A. Effectiveness of endoscopic drainage for
pancreas divisum: Endoscopic and surgical results in 31 patients. Endoscopy
1990;22:12933.
20. Galdermans D, Michielsen P, Pelkmans P, et al. Post-endoscopic sphincterotomy stenosis.
Endoscopy 1989;21:2379.
21. Lehman GA, Sherman S, Nisi R, Hawes RH. Pancreas divisum: Results of minor papilla
22. Cohen SA, Kasmin FE, Siegel JH. Minor papilla sphincterotomy in pancreas divisum.
23. Vaira D, Ainley C, Williams S, et al. Endoscopic sphincterotomy in 1000 consecutive patients.
Lancet 1989;2:4314.
24.
Lee JG, Leung JW. Endoscopy and gallstones. In Cotton PB, Tytgat GNJ, Williams CB (eds).
Annual of Gastrointestinal Endoscopy. 6th ed. London: Current Science Ltd, 1993;97106.
25. Booth FVM, Doerr RJ, Khalafi RS, et al. Surgical management of complications of endoscopic
sphincterotomy with precut papillotomy. Am J Surg 1990;159:1326.
27. Martin DF, Tweedle DE. Risks of precut papillotomy and the management of patients with
duodenal perforation [Letter; Comment]. Am J Surg 1992;163:2734.
28.
Shakoor T, Hogan WJ, Geenen JE. Needle-knife papillotomy efficacy and risks [Abstract].
sphincterotomy. N Engl J Med 1996;335:90918.
30.
Hill J, Maxwell AJ, Tweedle DE, Martin DR. Do measured parameters of blood clotting help
predict haemorrhage after endoscopic sphincterotomy? [Abstract]. Gut 1990;31:A608.
31. Martin DF, Tweedle DE. Retroperitoneal perforation during ERCP and endoscopic
sphincterotomy: Causes, clinical features and management. Endoscopy 1990;22:1745.
32. Kullman E, Borch K, Lindstrom E, et al. Bacteremia following diagnostic and therapeutic
33.
Horwitz A, Kirson SM. Cholecystitis and cholelithiasis as a sequel to gastric surgeryA
clinical impression. Am J Surg 1965;109:7602.
34. Thon HJ, Loffler A, Buess G, Gheorghiu T. Is ERCP a reasonable diagnostic method for
excluding pancreatic and hepatobiliary disease in patients with a Billroth II resection?.
35.
Osnes M, Myren J. Endoscopic retrograde cholangiopancreatography (ERCP) in patients with
Billroth II gastrectomies. Endoscopy 1975;7:22732.
36. Staritz M, Baas U, Ewe K, Meyer zum Buschenfelde KH. ERCP using a special catheter with
external steering. A reliable aid in typical ERCP problems. Endoscopy 1985;17:268.
37. Osnes M, Rosseland AR, Aabakken L. Endoscopic retrograde cholangiography and
endoscopic papillotomy in patients with a previous Billroth II resection. Gut 1986;27:11938.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
Liebe S, Arendt R, Jentzen H. Initial experiences using endoscopic papillotomy and bile duct
drainage in patients with Billroth II gastrectomy. Dtsch Verdau Stoffwechselkr 1986;46:8995.
Wurbs D, Dammermann R, Ossenberg FW, Classen M. Descending sphincterotomy of the
papilla of Vater through the T-drain under endoscopic view. Variants of endoscopic
papillotomy (EPT). Endoscopy 1978;10:199203.
Zillessen E, Louis B, Durr HC. Dissolution of biliary duct stones without papillotomy in a
patient with Billroth II resection and Crohn's disease. Endoscopy 1984;16:1223.
Forbes A, Cotton PB. ERCP and sphincterotomy after Billroth II gastrectomy. Gut
1984;25:9714.
Farkas I, Patko A. Pathologic relations and diagnostic difficulties in pancreatic and gallbladder
diseases following gastric surgery. Rntgenbltter 1983;36:10913.
Safrany L. Endoscopic and retrograde cholangiopancreatography after Billroth II operation.
Endoscopy 1972;4:198202.
Costamagna G, Mutignani M, Gabbrielli A, et al. Endoscopic sphincterotomy in patients with
Billroth II gastrectomy: Improvements with a new inverted sphincterotome. Gastroenterology
1993;104:A357.
Meenan J, van Dullemen H, Tytgat GN, Huibregtse K. Endoprosthesis guided needle-knife
biliary sphincterotomy is effective in cases of Billroth II gastrectomy [Abstract]. Gastrointest
Endosc 1994;40:29, #59.
Tsang T-K, Crampton AR, Meiselman M, et al. Percutaneous endoscopic biliary stent
placement for Billroth II and total gastrectomy with Roux-en-Y enteroenterostomy.
Nakamura H, Yasuhara, Kimura K, Iwadare M. Non-surgical biliary drainage for cholelithiasis.
Nippon Rinsho 1993;51:184650.
Safrany L, Neuhaus B, Portocarrero G, Krause S. Endoscopic sphincterotomy in patients with
Billroth II gastrectomy. Endoscopy 1980;12:1622.
Leach RE. Endoscopic sphincterotomy. Proceedings of World Congress of Gastroenterology,
Madrid. 1978.
Cremer M, Gulbis A, Toussaint J, et al. Endoscopic papillotomy in Belgium. In Demling L,
Classen M (eds). Endoscopic Sphincterotomy of the Papilla of Vater. Stuttgart: Thieme,
1978;8390.
Rosseland AR, Osnes M, Kruse A. Endoscopic sphincterotomy (EST) in patients with Billroth
II gastrectomy. Endoscopy 1981;13:1924.
Bedogni G, Bertoni G, Contini S, et al. Endoscopic sphincterotomy in patients with Billroth II
partial gastrectomy: Comparison of three different techniques. Gastrointest Endosc
1984;30:3004.
Ricci E, Bertoni G, Conigliaro R, et al. Endoscopic sphincterotomy in Billroth II patients: An
improved method using a diathermic needle as sphincterotome and a naso-biliary drain as
guide. Gastrointest Endosc 1989;35:4750.
Tweedle DE, Martin DF. Needle knife papillotomy for sphincterotomy and cholangiography.
Foutch PG. A prospective assessment of results for needle-knife papillotomy and standard
endoscopic sphincterotomy. Gastrointest Endosc 1995;41:2532.
Chapter 62 Calculus Disease of the Bile Ducts

(6294)
(6295)
DIETMAR F. W. WURBS, M.D.
Research has substantially advanced the understanding of stone formation in the gallbladder, and the
surgical treatment of cholelithiasis and cholecystitis has been revolutionized by laparoscopic
cholecystectomy. Unfortunately, knowledge is less advanced with respect to bile duct stone formation,
including intrahepatic duct stones. In particular, the development of black pigment stones is poorly
understood, although the role of bacteria in the formation of brown pigment stones is reasonably
established. The range of therapeutic options for the treatment of bile duct stones has exploded into a
spectrum that encompasses endoscopic, radiologic, and laparoscopic methods and the incorporation
of a variety of new lithotripsy techniques that are especially beneficial for patients with prior surgery and
those with intrahepatic duct stones.
Diagnosis
Bilirubin and standard serum enzyme test of liver function are usually elevated in a pattern suggestive
of biliary obstruction in patients with choledocholithiasis. Occasionally, however, these values can be
within normal limits despite the presence of choledocholithiasis.1(6296) Fluctuations in the levels of
serum bilirubin and liver function tests are frequent in patients with clinical findings that indicate the
presence of stones in the bile duct. Roston and Jacobson2(6297) found a correlation between the
changing pattern of these tests and the presence or absence of stones in the duct at retrograde
cholangiography. In this study of 94 patients with elevated liver function tests and clinical findings
suggesting bile duct stones at presentation, stones were much more likely to be present in the duct at
endoscopic retrograde cholangiopancreatography (ERCP) if levels of the biochemical parameters of
liver function were either stable or increasing, whereas a decrease or return to normal limits was
associated with a decreased likelihood of finding stones. For example, stones were found in 94% of
patients with one or more rising liver function tests versus only 13% of patients in whom the liver
function parameters had returned to normal at the time of ERCP.
Ultrasonography (US) has become the method of choice for demonstration of stones in the
gallbladder.3,4(6298) However, the diagnostic approach with respect to the bile ducts is more complex:
US can demonstrate the level of obstruction,5(6299) but identification of the cause is often impossible.
This is especially true for stones in the preampullary portion of the duct. Reported success rates for
identification of stones range from 20 to 75%,6,7(6300) although US is excellent for the demonstration
of intrahepatic stones.
The conventional intravenous cholangiogram (IVC) is error-prone, with incorrect conclusions in as
many as 40% of examinations, even with technically adequate studies in nonjaundiced patients.
Although there are some advocates for IVC as a screening procedure for bile duct stones before
laparoscopic cholecystectomy,8(6301) the value of this procedure is open to question.6,9,10(6302)
Before ERCP, Tham et al.11(6303) obtained IVCs in 60 nonjaundiced patients in whom there was a
suspicion of bile duct stones. Of the 27 patients with bile duct stones at cholangiography, IVC made a
correct diagnosis in 13 cases. The IVC was negative in 29 of 30 patients with no stones at ERCP. By
comparison with direct cholangiography, IVC had a sensitivity of 48%, specificity 97%, positive
predictive value 93%, negative predictive value 67%, and accuracy 73%.
Bile duct stones can also be detected by computed tomography (CT), although the sensitivity of this
imaging method for the diagnosis of choledocholithiasis is less than that of direct cholangiography.
Neitlich et al.12(6304) performed unenhanced helical CT before ERCP in 51 patients with a clinical
suspicion of choledocholithiasis, followed by ERCP. CT correctly diagnosed choledocholithiasis in 15 of
17 patients found to have choledocholithiasis by retrograde cholangiography, with one false-positive
diagnosis (sensitivity 88%, specificity 97%, accuracy 94%). CT is especially useful for estimation of the
cholesterol content of stones by densitometry.13(6305) There is a correlation between the ability to
fragment stones with laser energy and their CT appearance: Dense stones require fewer pulses and
lower energies compared with faint or rimmed stones.14(6306)
Magnetic resonance cholangiography (MRC) may prove to be more accurate than US and CT for the
diagnosis of bile duct stones.15,16(6307) However, MRC is costly and comparative data on efficacy
are lacking, including prospective comparative studies with direct cholangiography. Chan et
al.17(6308) compared MRC findings in 45 patients with suspected bile duct stones with findings by
retrograde cholangiography performed within 5 hr of the MRC procedure. MRC correctly diagnosed
choledocholithiasis in 18 of 19 patients with stones demonstrated by cholangiography and correctly
identified the absence of stones in 22 of 26 patients (sensitivity 95%, specificity 85%, positive
predictive value 82%, negative predictive value 96%). Two false-positive diagnoses were attributed to
pneumobilia.
Endoscopic ultrasonography (EUS) appears to be virtually equivalent to retrograde cholangiography for
the diagnosis of bile duct stones and offers certain advantages.18,19(6309) For example, it can
eliminate the need for direct cholangiography in cases where the clinical suspicion for bile duct stones
is low (see Chapter 78: Endoscopic Ultrasonography of the Retroperitoneal Organs).
Despite these many imaging studies, direct cholangiography remains the diagnostic reference
standard for stones in the bile ducts.
Direct Cholangiography
A complete and detailed cholangiogram is a necessity in cholestasis for a correct diagnosis and to
decide whether operative therapy is indicated. A complete cholangiogram, one that includes the
intrahepatic ducts, may be made either by percutaneous cholangiography (PTC) or by endoscopic
retrograde cholangiography (ERC). Both are effective, but in a sense, ERC is performed via a natural
route and therefore is less invasive than PTC. The technique of PTC is less difficult to master, and the
PTC needle is less expensive than a duodenoscope. However, in experienced hands, ERC may be
performed in a short time with little patient discomfort and with a success rate of 80 to 90%. ERCP also
provides information about the main duodenal papilla, duodenum, and pancreas. Both PTC and ERCP
have potential complications, and the mortality rate with both methods is about 0.1%.2023(6310)
Technique of Direct Cholangiography in Stone Disease

The technique of ERCP is described in Chapter 57. Some special problems, such as a juxtapapillary
diverticulum, may be encountered in the course of this procedure that may make cannulation of the
papilla more difficult. Normally, a 60% concentration of a water-soluble contrast medium will produce
excellent radiographs. However, with dilated bile ducts, small stones may be obscured by this high
concentration of radiopaque material. Therefore, a 30% concentration is preferable when
choledocholithiasis is suspected; the less dense column of contrast medium will be more transparent
and less likely to conceal small calculi. X-ray films made during the early stages of injection of the
contrast medium are helpful, as are delayed films at 30 min after ERCP. On initial injection, the high
specific weight of the contrast medium leads to preferential opacification of the dependent parts of the
biliary system. With the patient in the usual semiprone position for ERCP with the right side raised
slightly, the left lobe of the liver will be in the more dependent position relative to the right lobe and,
therefore, the former will fill in preference. A nearly equal distribution of contrast medium in the biliary
tree can be attained by turning the patient on the table or by tilting the table. Only filling defects that are
mobile and completely surrounded by contrast medium can be classified as stones. Tumors (Figure
621), cysts (Figure 622), and even a normal papilla24(6311) (Figure 623) can mimic stones.
Conversely, a stone may mimic a tumor (Figure 624).
(6312)Figure 621. Round filling defect in the ampulla was initially interpreted as a stone.
After endoscopic papillotomy, a polypoid carcinoma protruded into the duodenum.
(6313)Figure 622. Lesion that mimicked a stone was shown at operation to be a bile duct
cyst.
(6314)Figure 623. Asymmetric shape of the papilla and long choledochoduodenal sphincter
were initially interpreted as a stone impacted in the papilla.
(6315)Figure 624. Stone simulates a tumor at the confluence of the hepatic ducts. The stone
completely obstructs the right hepatic duct.
Therapeutic Access to the Biliary Tree

Many complications of biliary diseases relate to bile stasis and cholangitis. An operator must therefore
be able to decompress the biliary system, either by percutaneous transhepatic drainage or by
endoscopic papillotomy (EPT) with drainage by nasobiliary tube or insertion of a biliary stent (biliary
endoprosthesis). For emergent decompression, both approaches to drainage achieve good
results.2527(6316) To remove stones from the bile ducts, the endoscopic retrograde route is used in
over 90% of patients. In difficult cases, it is possible to extend the capability of this route by means of
retrograde cholangioscopy (see Chapter 70: Peroral Cholangioscopy and Pancreatoscopy).28(6317)
Removal of stones via percutaneous routes, with or without cholangioscopy, has proved to be an
effective, low-risk technique.29(6318) Employed primarily in the Far East, its use has increased in
Western countries in recent years (see Chapter 71: Choledochofiberoscopy and Endoscopic
Lithotripsy). Another route for removal of retained bile duct stones after surgery is the T-tube tract (see
Chapter 71) (Figure 625).30,31(6319) During laparoscopic cholecystectomy, a basket or even a
cholangioscope can be introduced into the common bile duct for the purpose of removing stones
(Table 621).32(6320)
TABLE 621
Biliary Access Routes and Techniques
ACCESS (INCLUDING CHOLANGIOSCOPY)
TECHNIQUES
Endoscopic retrograde
Extraction139
Lithotripsy
Percutaneous transhepatic27,29
Percutaneous transcystic
Percutaneous transjejunal36
T-tube channel30,31
Hidden loop3335
Surgery
Mechanical38,39
Electrohydraulic41,42
Laser43,44
Chemical
dissolution4547
ESWL40
Open
Laparoscopic32
Jejunal interposition207
ESWLextracorporeal shockwave lithotripsy.
(6321)Figure 625. A, Retained stone in the prepapillary region, T-tube in situ. B, After
bougienage of the T-tube tract with a 5-mm Savary-Gilliard bougie, a 4.4-mm
choledochoscope is introduced into the bile duct followed by complete electrohydraulic
destruction of the stone. The resulting debris was pushed through the papilla into the
duodenum.
In general, surgery as the primary treatment for bile duct stones has become a rare exception
nowadays. An unequivocal indication for surgery continues to be the resection of parts of the liver,
especially the left lobe, if there is a circumscribed malformation of the bile ducts with stones and
cholangitis. If a surgical hepaticojejunostomy is to be performed, especially after resection of a stricture
in a patient with a high risk of stone recurrence, subcutaneous placement of the anastomosed loop of
jejunum ("hidden loop") or creation of a "permanent access" hepaticojejunostomy provides a relatively
accessible route to the biliary system for repeated cholangioscopy and endoscopic treatment of stones
(Figure 626).3335(6322) Even puncture of a Roux-en-Y loop has been used to access the bile
ducts.36(6323) Another "endoscopy-friendly" surgical procedure is the interposition of a jejunal
segment between the liver hilum and the duodenum.37(6324)
(6325)Figure 626. A straight-viewing gastroscope reaches a hepaticojejunostomy via an

opened "hidden loop." Stones are extracted with a Dormia basket.
Technical Principles for Stone Removal

After EPT, many if not most bile duct stones will pass spontaneously into the duodenum over the
course of time. However, to prevent stone impaction with cholangitis, and to obviate the need for
repeated procedures to verify duct clearance, complete extraction by basket or by balloon, or both, at a
single procedure has became the rule. If extraction using these simple methods fails, which may occur
for various reasons including a disproportion between the stone diameter and the smallest diameter of
the access tract (e.g., papillotomy), stone fragmentation or lysis is necessary. A range of techniques is
available for this purpose, including mechanical lithotripsy;38,39(6326) destruction by shock waves,
external40(6327) or internal electrohydraulic;41,42(6328) and fragmentation by application of laser
energy.43,44(6329) All of these techniques are well established, have advantages and disadvantages,
and can be employed via any of the access routes (see Table 621).
Chemical dissolution of cholesterol stones can be achieved with alpha-1-glyceromonooctanoin or with
methyl tert-butyl ether (MTBE); alkaline ethylenediaminetetraacetic acid (EDTA) preparations have
been used to dissolve brown pigment stones.4547(6330) Chemical dissolution requires placement of
a delivery catheter near the stones. Often, a combination of procedures, such as extracorporeal
shockwave lithotripsy followed by endoscopic extraction of fragments, is necessary (Figure 627).
(6331)Figure 627. A, "Stone-street" in the right hepatic lobe that cannot be reached by a
retrograde approach. B, After puncture of the intrahepatic bile duct and bougienage of the
percutaneous tract, a stable access has been created that allows percutaneous transhepatic
endoscopy and electrohydraulic lithotripsy. C, Persisting stones and debris are dissolved via a
small tube (arrowheads).
The fundamental advantage of endoscopic therapy of stones, in contrast to surgery, is that treatment is
accomplished step by step in several sessions according to the unique circumstances in each case,
rather than attempting a single operation that may be prolonged or unsuccessful.
The Papilla and Peripapillary Findings in Stone Disease

The most impressive endoscopic finding is a stone impacted within the main duodenal papilla. It may
be visible in the meatus of the papilla (Figure 628), or even if it is not visible, the stone may cause the
whole papilla to protrude into the duodenum to an extreme degree (Figure 629). Redness around the
meatus indicates inflammation, sometimes referred to as papillitis. An enlarged papillary meatus with a
red and fissured margin together with recent and usually severe colic may be interpreted as evidence
of passage of a stone (Figure 6210).
(6332)Figure 628. Large stone (17 mm) impacted in the papilla causes recurrent acute
pancreatitis.
(6333)Figure 629. Large, prominent papilla harbors a 20-mm stone. The meatus of the
papilla is directed caudally and cannot be seen. On the protruding papilla, there is a second
opening, probably a spontaneous perforation due to the presence of the stone.
(6334)Figure 6210. Papilla atypically located in a diverticulum. The meatus is large and red
as a result of passage of a stone.
There is strong evidence that bacteria play a significant role in the formation of brown pigment
stones.48(6335) This is the usual type of stone present in patients with recurrence of calculi within the
bile duct. Beta-glucuronidase is produced by some bacteria, especially Escherichia coli.49(6336) This
enzyme splits conjugated bilirubin, which in turn combines with calcium to form insoluble calcium
bilirubinate.50(6337) In one study, glucuronidase-producing bacteria were present in the bile of 70% of
patients who underwent EPT for gallstone disease.51(6338) There is also some evidence that
beta-glucuronidase in the bile ducts may be the result of a metabolic defect in the liver,52(6339)
although this is controversial and confirmatory data are lacking.53(6340) Normally, glucuronic acid is
metabolized to glucaro-1,4-lactone, which is excreted in bile and which inhibits beta-glucuronidase. It
has been suggested that this metabolic pathway prevents stone formation,54(6341) but this has not
been established.
The incidence of duodenal diverticula increases with age, as does that of gallstones, and a relation
between the peripapillary diverticulum and the development of common duct and gallbladder stones
has been postulated. Usually, diverticula are located on the medial wall of the descending duodenum
close to the papilla (Figure 6211). Duodenoscopy with ERCP permits accurate localization, definition
of their relation to the papilla, and correct diagnosis of associated bile duct stones (Figures 6212 and
6213).5558(6342) Chandy et al.59(6343) studied patients with bile duct stones in the absence of
gallbladder stones (n = 44), patients with duct stones together with gallbladder stones (n = 750), and
two age-matched control groups comprising, respectively, patients who underwent upper
gastrointestinal endoscopy with a side-viewing duodenoscope (n = 100) and patients who underwent
ERCP because of pancreatic cancer (n = 100). Juxtapapillary diverticula were found in 70% of patients
with primary bile duct stones (i.e., those without gallbladder stones) and 25% of patients with
choledocholithiasis and cholelithiasis (current or by history) versus 7% and 8% in the respective control
groups. Lotveit et al.56(6344) reported on ERCP studies of patients in whom symptoms of biliary tract
or pancreatic disease developed after cholecystectomy and who were asymptomatic for intervals of 2
years or more. The incidence of recurrent calculi in patients with diverticula was 87.5%; without
diverticula, the incidence was 39.1%. Studies such as this suggest a causal relationship between
duodenal diverticula and gallstone disease. Although the relationship of diverticula to the pathogenesis
of biliary calculi is speculative, it has been suggested that the diverticulum causes a kind of stasis by
impeding the flow of bile into the duodenum. In a study of a small number of patients, there were no
significant differences in common bile duct pressure in relation to the presence or absence of
diverticula.60(6345) Phasic activity and peak sphincter of Oddi pressure were similar in both groups.
(6346)Figure 6211. Two large juxtapapillary diverticula. The papilla is usually located at the
brim of the diverticulum in a dorsocaudal or ventrocaudal position. Here the diverticulum
surrounds the papilla.
(6347)Figure 6212. Retrograde cholangiogram in the same patient as in Figure 6211

demonstrates diverticulum surrounding the distal end of the bile duct.
(6348)Figure 6213. Normal pancreatic duct. The common bile duct contains a 10- 20-mm
stone, and there is a diverticulum between bile duct and duodenum.
There is a very high incidence (70 to 92%) of E. coli colonization of the bile ducts in association with
papillary or peripapillary diverticula. In contrast, there is a lower incidence (45%) when diverticula are
located at some distance from the papilla.61(6349) In a study by Lotveit of stones from 32 patients with
juxtapapillary diverticula,62(6350) 22 patients had stones classified by quantitative infrared
spectroscopy as pigment stones in which the main component was calcium bilirubinate. At
cholecystectomy, patients with cholelithiasis but not choledocholithiasis have a significantly higher
incidence of positive bacterial cultures from the common bile duct if there is a juxtapapillary
diverticulum compared with patients who do not have a diverticulum.63(6351)
Another interesting finding at duodenoscopy and ERCP is the choledochoduodenal fistula (Figure
6214; see also Figure 629).64(6352) Typically, these fistulas are located in the roof of the papilla or
a little bit cephalad to this in the infundibular portion of the bile duct. This has been explained as the
result of spontaneous stone-induced perforation;65(6353) however, similar fistulas were found in 18 of
36 patients after surgical sphincterotomy and after surgery during which the distal bile duct was probed
blindly.66(6354) In my (D.F.W.W.) experience, this type of fistula was discovered in 26 of 3000
patients undergoing ERCP; 5 patients with fistulas had no prior surgery, and operative reports were
available in 16 of the remaining 21 patients. In all of these patients, probing in a blind fashion or
somewhat forceful passage of dilators "through the papilla" because of suspected stenosis had been
carried out. It therefore seems probable that these fistulas represent unintended perforations in a
significant number of cases (unpublished data). Endoscopic fistulotomy has been suggested as
treatment for these fistulas.67(6355) Therapeutic intervention is probably necessary only if there are
stones in the bile duct. In my (D.F.W.W.) series, this was the case in 11 of the 26 patients.
(6356)Figure 6214. Normal papilla (arrow) with cannulation of a second opening slightly
cephalad to the papilla. Surgery with "bougienage of the papilla" had been performed
previously.
Stones in the Biliary Tract: Typical and Atypical Findings

The most common finding is the small or medium-sized mobile stone up to 15 mm in diameter in
normal-caliber or slightly dilated bile ducts with or without a juxtapapillary diverticulum (see Figure
6213); but multiple stones in dilated ducts are also relatively common (Figure 6215). These may be
cholesterol stones even if they occur after resection of the gallbladder. It has been suggested that the
dilated bile duct to some extent assumes the function of the gallbladder when the latter organ has been
resected or is not functioning.68(6357) Patients with very large bile duct stones may have a long
history, as long as 30 years, of mild symptoms together with long asymptomatic intervals (Figure
6216).69(6358) Ultimately, some will have evidence of secondary biliary cirrhosis.
(6359)Figure 6215. Multiple stones in the common bile duct. These proved on analysis to be
cholesterol stones.
(6360)Figure 6216. Multiple calculi in the gallbladder. Contrast material does not enter the
gallbladder, and therefore, cystic duct occlusion can be inferred. A large stone can be seen in
the bile duct. This 83-year-old patient had right upper quadrant colicky pain about 50 years
earlier, a symptom-free interval of more than 40 years, and recurrent episodes of mild jaundice
during the previous 4 years. Now she has acute suppurative cholangitis, and a nasobiliary tube
has been inserted to decompress the biliary tree.
Hundreds of intrahepatic stones are sometimes found in association with Caroli's disease, a relative
congenital stenosis, or after repeated operations for removal of bile duct stones (Figure 6217).
Jaundice in a patient who has had biliary surgery, especially for calculus disease, should always raise a
suspicion for recurrent stones. Bile duct stricture, with or without associated stones, must also be
considered, as this is a relatively frequent occurrence after bile duct surgery (Figure 6218). Even
patients with choledochoduodenal anastomoses may have recurrent stones (Figures 6219 and
6220).
(6361)Figure 6217. Hundreds of stones of various sizes in dilated intrahepatic bile ducts.
There is free reflux of the radiographic contrast material into the duodenum. Three
choledochotomies were performed in this patient; after each operation, the number of stones
increased. A nasobiliary tube has been inserted. An attempt at stone dissolution was
unsuccessful.
(6362)Figure 6218. Large stone in the bile duct proximal to a stricture. Cholecystectomy
and choledochotomy had been performed. Endoscopic removal of this stone is not possible.
(6363)Figure 6219. Recurrent stone after choledochoduodenostomy. The bile duct has been
cannulated through the small-diameter anastomosis. Relative to the diameter of the stone, the
diameter of the bile duct is narrow distal to the position of the stone. This stone and others that
later recurred were composed of soft material and were extracted endoscopically.
(6364)Figure 6220. This 93-year-old patient has had a Billroth II procedure,

cholecystectomy, and choledochoduodenostomy. Cannulation and retrograde opacification
through the anastomosis demonstrate complete filling of the bile duct with stones. The stones
were composed of a soft, yellow material and thus extractable.
The Mirizzi syndrome consists of a triad of cystic duct stone, chronic cholecystitis with shrinkage of the
gallbladder, and benign stenosis of the hepatic duct that leads to cholestasis and cholangitis.70(6365)
A few reports have suggested an association between Mirizzi's syndrome and gallbladder
cancer.71,72(6366) In my (D.F.W.W.) unit, 26 cases of Mirizzi's syndrome were encountered in 5434
ERCPs (0.43%) performed over a 6-year period.73(6367) Direct cholangiography typically
demonstrates an asymmetric stenosis with a smooth configuration, an appearance that can usually be
interpreted as due to compression. However, interpretation of the cholangiographic findings may be
difficult and the correct diagnosis may be made only at operation (Figures 6221 and 6222).
(6368)Figure 6221. Mirizzi's syndrome. A 7duct.
3-cm stone compresses the common bile
(6369)Figure 6222. Mirizzi's syndrome. The bile duct contains multiple stones. The
bow-shaped impression at the bifurcation directed from the patient's right is caused by a stone.
Preoperatively, a tumor was suspected.
Sometimes there may be more than one reason for a patient to be jaundiced. For example, a stone
and carcinoma may both be present (Figure 6223), or a long cystic duct may harbor symptomatic
stones (Figure 6224).
(6370)Figure 6223. Stone and carcinoma of the bile duct with prestenotic dilation of the
intrahepatic ducts.
(6371)Figure 6224. Long cystic duct with stone; whether it is residual or recurrent is not
known.
The incidence of bile duct stones in patients who undergo operation for acute cholecystitis is 9 to
10%.74(6372) Residual or recurrent stones after cholecystectomy performed without duct exploration
occur in 1 to 4% of patients.75(6373) The frequency of retained or recurrent stones after removal of
bile duct stones at operation is 10%.74(6374) About 20% of patients who have undergone reoperation
for bile duct stones will develop choledocholithiasis again.76(6375)
Gallstone Pancreatitis
One of the most common identifiable causes of acute pancreatitis is migration of bile duct stones
through the papilla (see Figure 628). Delivery of stones into the duodenum is often accompanied by
relief of symptoms and a rapid decline in serum amylase and bilirubin levels. With spontaneous
passage of stones into the duodenum, gallstone pancreatitis in most cases becomes a self-limiting
disease with a benign natural course. However, some patients develop severe and life-threatening
pancreatitis because of persistent ampullary obstruction by a stone. Mortality rates as high as 10%
have been quoted in older, unselected series.7781(6376) In a study of 279 patients with acute
pancreatitis from various causes, the mortality rate for those with gallstone-induced pancreatitis was
3%. In general, the mortality rate was higher for patients over 55 years of age.82(6377) Local
complications of acute pancreatitis include pancreatic necrosis, pancreatic ascites, pseudocyst
formation, and pancreatic and peripancreatic abscess. Systemic complications include fat necrosis
(subcutaneous and osseous intramedullary), aseptic epiphyseal necrosis, arthritis and polyserositis
(inflammation and effusions involving the pleura, pericardium, and articular synovium), coagulopathies,
sepsis, and multiorgan failure. Other patients will have recurrent episodes of pancreatitis, ranging from
mild to severe, because of repeated passage of stones.
Although fatal pancreatitis due to lodgment of a small stone at the main duodenal papilla was first
described by Opie83(6378) in 1901, the precise mechanisms whereby gallstones induce acute
pancreatitis is still a matter of some debate. Opie postulated that the acute inflammatory reaction was
induced by reflux of bile into the pancreatic duct as a result of obstruction at the papilla. This
explanation is supported by the observation that contrast medium injected into the bile duct is much
more likely to reflux into the pancreatic duct in patients with a history of pancreatitis compared with
those in whom acute pancreatitis has not occurred.84,85(6379) Furthermore, a common channel (see
Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas) is much more likely to be
found in patients who have had acute pancreatitis.86(6380) In the great majority of patients with known
gallstones, calculi are found in the stool within 10 days after an attack.87(6381) When the bile duct is
explored surgically soon after the onset of pancreatitis, choledocholithiasis will be found in about 70%
of patients.8890(6382) If operation is delayed, stones will be found in substantially less than one third
of patients.77,86,8992(6383) Others have suggested that sphincter of Oddi dysfunction, especially
laxity, is a major factor in the pathogenesis of acute pancreatitisthat the cause is not just a simple
matter of obstruction but that the dysfunctional sphincter, perhaps the result of repeated passage of
stones, permits duodenal contents to reflux into the pancreatic duct.93(6384) The relative importance
of these possible mechanisms and whether they work in consort has not been clarified. Furthermore,
the importance of other factors such as bacterial infection in the bile, a frequent occurrence in patients
with choledocholithiasis, and microlithiasis remains uncertain.
Duration of obstruction may be a factor in the severity of pancreatitis. Acosta et al.94(6385) compared
the severity of pancreatitis observed at surgery in 97 patients with a stone impacted at the papilla and
49 patients with evidence of spontaneous stone passage within 24 hr of the onset of symptoms. When
the duration of obstruction was greater than 48 hr, 84.6% of patients had evidence of severe
pancreatitis, including pancreatic necrosis, at operation. Some data also suggest that smaller stones in
the gallbladder, those less than 5 mm in diameter, are more likely to be associated with episodes of
pancreatitis.95(6386)
The role of cholesterol microcrystals and calcium bilirubinate granules in the pathogenesis of acute
pancreatitis is uncertain. These are the major components of biliary sludge, and passage of such
material from the bile duct has been postulated as a cause of apparently idiopathic pancreatitis. Lee et
al.96(6387) found biliary sludge in bile aspirated from the duodenum in 23 of 36 patients with idiopathic
pancreatitis. Transcutaneous US demonstrated biliary sludge in only 11 patients; in all but 1 case this
was composed of calcium bilirubinate granules. Patients treated by cholecystectomy or EPT had
significantly fewer recurrent episodes of pancreatitis during follow-up compared with those who were
not treated.
In addition to the usual measures for treatment, direct intervention in the form of surgery or ERCP
often becomes a consideration in patients with severe pancreatitis, especially in those in whom there is
evidence of cholelithiasis and a suspicion of choledocholithiasis. A variety of methods have been
devised to estimate the severity of acute pancreatitis, including systems based on age and clinical
parameters,97101(6388) peritoneal lavage (color of the fluid),102(6389) and imaging studies, CT in
particular.103(6390) Unfortunately, noninvasive imaging studies are not very helpful for detection of
stones in the bile duct during the acute stages of pancreatitis.104,105(6391) Although surgery to resect
the gallbladder, which is acting as a stone reservoir, or to remove a stone from the papilla or bile duct
is generally accepted, the timing of surgical intervention after the onset of pancreatitis is controversial.
There is a lack of well-designed, prospective, and randomized trials that focus on this issue. Acosta et
al.106(6392) reported that results are most favorable if surgery is performed within 6 to 48 (average
28) hr. Others report that the most satisfactory outcome occurs when surgery is delayed for 5 to 7
days.89,107(6393) Delayed surgery carries the potential risk of further episodes of
pancreatitis.89,107(6394) The earlier the operation, the more often stones are found impacted at the
papilla.106,107(6395) Duration of obstruction may also be a factor in the severity of pancreatitis.
Acosta et al.94(6396) compared the severity of pancreatitis observed at surgery in 97 patients with a
stone impacted at the papilla and 49 patients with evidence of spontaneous stone passage within 24 hr
of the onset of symptoms. When the duration of obstruction was greater than 48 hr, 84.6% of patients
had evidence of severe pancreatitis at operation, including pancreatic necrosis.
Endoscopic techniques have profoundly altered the therapeutic approach to biliary pancreatitis. ERCP,
formerly regarded as contraindicated in acute pancreatitis, is reasonably risk-free when the pancreatic
duct is carefully opacified.108(6397)
Endoscopic removal of a stone lodged in the papilla by enlarging the orifice by taking multiple biopsies
with a biopsy forceps was described in 1974.109(6398) The systematic use of EPT in the management
of gallstone pancreatitis was first reported in 1978 by Classen et al.110(6399) in a series of 58
patients. This was followed by several other reports of studies with successful results.111115(6400)
The first prospective, randomized, controlled trial of urgent ERCP with EPT compared with
conservative treatment for acute gallstone-induced pancreatitis was that of Neoptolemos et
al.116(6401) For this trial, US and biochemical parameters were used to select 121 patients with
suspected gallstone pancreatitis from a larger group of 223 consecutive patients with acute
pancreatitis. The patients were randomized to undergo conventional treatment (n = 62) or ERCP (n =
59) within 72 hr of onset (with EPT if stones were found in the bile duct). Patients were also stratified
into mild and severe groups on the basis of clinical criteria. Injection of the pancreatic duct was avoided
as much as possible. In the group undergoing cholangiography, bile duct stones were found in 63% of
patients in whom a severe attack of pancreatitis was predicted on the basis of clinical criteria versus in
23% of those in whom a mild episode was predicted. Choledocholithiasis was found in only 13% of
patients with mild pancreatitis who received conventional therapy. This relatively low incidence is
probably accounted for by passage of stones, as cholangiography was delayed in this group and in
some cases was not performed. Although there were no deaths in the ERCP/EPT group versus a
mortality rate of 18% for patients with severe pancreatitis in the conventional therapy group, the
difference did not reach statistical significance. This might be due to the short duration and relatively
small numbers of patients enrolled in the trial. However, there was a significant difference in morbidity
due to pancreatitis; 12% of endoscopically managed patients versus 61% of those in the conventionally
treated group experienced complications, most of which occurred in patients who were stratified as
having severe pancreatitis. Furthermore, the median number of days in hospital for endoscopically
treated patients was 9.5 compared with 17 days for patients treated conventionally, a statistically
significant difference.
Fan et al.117(6402) randomized 195 patients with acute pancreatitis to undergo either ERCP within 24
hr of hospital admission with EPT if stones were found (n = 97) or initial conservative treatment (n =
98). ERCP with EPT as needed was performed in patients in the latter group if their condition
worsened. Biliary sepsis did not occur in any of the endoscopically managed patients but was noted in
12 patients who initially received conservative treatmenta significant difference. This difference was
noted in patients classified as having mild pancreatitis as well as those who were considered to have
severe inflammation, although the difference was more pronounced in the latter group. There were no
significant differences between the two groups with regard to local or systemic complications of
pancreatitis. Five endoscopically treated patients died versus 9 treated conservatively (p = .4).
The results in the trials of Neoptolemos et al.116(6403) and Fan et al.117(6404) must be contrasted
with those obtained by Flsch et al.118(6405) In this multicenter study, patients with acute gallstone
pancreatitis were randomized to undergo either ERCP with EPT as needed (n = 126) or conservative
treatment (n = 112) to include endoscopic intervention if evidence of biliary obstruction or sepsis
developed (n = 22). ERCP was successful in 121 patients randomized to receive endoscopic therapy.
EPT with stone extraction was successful in 57 of 58 cases in which stones were recognized. Among
the 22 conservatively treated patients who underwent ERCP, stone removal was successful in 13.
There was no difference in complications, although these were more severe in endoscopically treated
patients. Death occurred in 10 endoscopically treated patients as a result of biliary pancreatitis versus
4 patients treated conservatively, a difference that was not significant (p = .16). At 30 days, 14 in the
endoscopic treatment group had died compared with 7 in the group that underwent conservative
treatment, also a difference that was not significant (p = .10). Although patients enrolled in this study
were thought to have gallstoneinduced pancreatitis, relatively few patients in the conservative
treatment group were found to have choledocholithiasis. This might be explained by passage of
stones. However, the criteria for inclusion of patients in this study could also have selected patients
who did not actually have choledocholithiasis.119(6406)
Duodenoscopy and EPT on an emergency basis for acute gallstone pancreatitis is now widely
practiced by experienced endoscopists.116(6407) The procedure is not contraindicated or limited by
the severity of the pancreatitis, the age of the patient, or accompanying diseases. Improvement of the
patient's general clinical condition and in laboratory parameters is prompt and sometimes dramatic.
Although some investigators attempt to avoid injection of contrast medium into the pancreatic duct,
biliary stones have even been extracted from the main pancreatic duct (Figure 6225).111,120(6408) I
(D.F.W.W.) therefore recommend cautious opacification of the main pancreatic duct in patients with
biliary pancreatitis.
(6409)Figure 6225. Endoscopic retrograde cholangiopancreatography was performed

because of recurrent acute pancreatitis. A, Simultaneous opacification of both pancreatic and
bile ducts through a common channel. There is a 2- to 3-mm stone in the pancreatic duct.
Stones of the same size were present in the gallbladder (not illustrated). B, A Dormia basket
is introduced into the pancreatic duct to extract the stone, which proved to be of yellow
material.
The point at which ERCP should be performed in relation to the patient's course remains a matter of
considerable discussion. When pancreatitis is clinically severe or prolonged, and if the patient's clinical
status is worsening, especially if there appears to be some doubt about the outcome, I (D.F.W.W.)
recommend immediate intervention. ERCP provides a certain diagnosis with respect to the presence
or absence of stones, avoids further loss of time, and represents a definitive mode of therapy.
Choledocholithiasis and Cholangitis

Cholangitis is often associated with choledocholithiasis. Its pathogenesis relates to the presence of
bacteria in an obstructed duct. Bacterial colonization is more frequent with obstruction caused by
stones as opposed to tumor.121(6410) Additional factors may be the length of time during which the
obstruction has been present, the number of episodes of colic, and the age of the patient.122(6411)
With repeated operations on the bile ducts, bacterial colonization approaches 100% of
cases.121124(6412) It is mandatory that these facts be kept in mind when performing diagnostic or
therapeutic procedures on the bile ducts. ERCP, PTC, and even T-tube cholangiography have septic
complications that range from uncomplicated bacteremia to fatal endotoxic shock.20,124126(6413) It
seems possible that bacterial spreading during PTC results from iatrogenic biliovenous fistulas, but
bacterial propagation may also occur with retrograde cannulation (see also Chapter 56: Indications,
Contraindications, and Complications of Diagnostic Endoscopic Retrograde
Cholangiopancreatography).20,126,127(6414) It appears that an important factor, especially with
respect to life-threatening endotoxic shock, is the pressure increase that occurs in an obstructed and
infected duct. It is well established that the occurrence of cholangiovenous bile reflux depends on the
degree to which intraductal pressure is increased.128130(6415) The critical pressure level is about 25
cm H2O in animal studies129,130(6416) as well as in human beings,131(6417) a level that is easily
exceeded with every method of direct cholangiography. At pressures above 30 cm H2O, there is
evidence that bile flow ceases.131(6418) Cholangiovenous reflux occurs directly from the bile ductules
through the spaces of Mall and Disse into the hepatic sinusoids.132(6419)
From these data, it must be assumed that bacterial colonization of the bile ducts is present in cases of
biliary obstruction, especially that caused by calculi. ERCP must be performed by cautiously injecting
only as much contrast medium as is necessary to determine the cause of obstruction.
Antibiotic prophylaxis can be recommended for all types of direct cholangiography when obstruction is
likely to be present (see Chapter 56: Indications, Contraindications, and Complications of Diagnostic
Endoscopic Retrograde Cholangiopancreatography). This has been accomplished by addition of
aminoglycoside antibiotics, which are not antagonized by iodinated agents or bile, to the contrast
agent.133(6420) However, there is little evidence that this is efficacious.134(6421) A presumably better
approach is the systematic administration of beta-lactam antibiotics, which achieve high serum levels,
are excreted into bile, and are not antagonized by bile.135,136(6422) Selection of antibiotics should be
based on the expected or identified bacterial contaminants of the ducts. These will be mainly E. coli
and Klebsiella, Enterobacter, and Pseudomonas spp. If bacteremia or sepsis occurs, it is not unusual
to culture several different organisms from the blood. The spectrum of bacteria in the bile ducts as
determined by ERC studies differs from that found in samples collected at surgery.126(6423) When
improperly cleaned, disinfected, or stored, endoscopes are notorious for harboring Pseudomonas
aeruginosa (see Chapter 8: Disinfection of Endoscopes and Accessories). This species presents a
more significant problem in patients with tumors than in those with calculus disease,126(6424)
especially when effective drainage of the biliary tree cannot be obtained.127(6425) It should be
emphasized that biliary excretion of antibiotics into bile is reduced when there is obstruction and
stasis,135(6426) and therefore, antibiotic activity within obstructed bile ducts is likely to be substantially
reduced with incomplete elimination of bacteria.137(6427)
Although systemic antibiotic administration will reduce the incidence of septic complications, it is logical
that free bile flow must be established immediately after the cause of obstruction is determined. In the
past, this has been possible only by surgery. However, since the late 1970s, temporary or definitive
endoscopic therapy has become established as an effective approach.138(6428) This includes EPT
with stone removal,139(6429) biliary stenting using an indwelling prosthesis, and the use of nasobiliary
tubes for drainage, decompression,25,140(6430) and dissolution of stones45,47(6431) (Figures 6226
and 6227).
(6432)Figure 6226. Nasobiliary tube. Stones have been mobilized from the prepapillary bile
duct and pushed proximally in the duct. Solvents have been perfused through the tube, and
both stones have lacunae as signs of beginning dissolution.
(6433)Figure 6227. Diagram shows distal configuration of a Wurbs-type nasobiliary tube.
Acute Obstructive Suppurative Cholangitis

Acute obstructive suppurative cholangitis is the most dangerous form of bacterial cholangitis. Nearly all
cases are caused by common duct stones, and almost all patients will die if treated conservatively.
Immediate mechanical decompression is mandatory. This principle was formulated by Rogers in
1903,141(6434) and it has proved valid over many years.142,143(6435)
Endoscopists became more familiar with acute cholangitis with the introduction of EPT. During the
early development of EPT, cholangitis was a serious problem, especially suppurative cholangitis
associated with stone impaction after the procedure. The mortality rate due to this complication was
16% during this early developmental period.144(6436) However, the problem has been largely
resolved by the development and endoscopic use of nasobiliary drainage tubes that may be left in the
bile ducts for long periods of time (Figure 6228; see also Figure 6226).140(6437) Endoscopic
insertion of a nasobiliary tube fulfills the requirements for decompression by mobilizing the impacted
stone, preventing reimpaction, and ensuring free flow of bile into the duodenum. If there is concern that
the bile flow around the nasobiliary tube and into the duodenum is inadequate, suction can be applied
to the tube to drain the ducts externally. In my (D.F.W.W.) early experience, the nasobiliary tube was
used to treat septic complications arising from stone impaction after EPT that otherwise would have
required urgent surgery.25(6438) I (D.F.W.W.) now use nasobiliary drainage in all cases of acute
obstructive suppurative cholangitis, with good results in comparison with surgical or other nonsurgical
methods of management. After an interval of satisfactory endoscopic decompression, further treatment
of choledocholithiasis may be endoscopic or surgical. When the clinical situation is stable, the
papillotomy can be enlarged, if necessary, and stones removed from the bile duct according to the
previously described principles (Table 622).
Acute Obstructive Suppurative

Cholangitis: Therapy After Nasobiliary Drainage*
TABLE 622
TYPE OF THERAPY
Endoscopic therapy
Spontaneous delivery/extraction
Stone dissolution
Surgery
No treatment
Died before further treatment
Unresolved
Total
*Author's series, partly published.140
PATIENTS (n)
10
3
2
5
2
22
Not extractable or dissolvable; surgery impossible. One patient

underwent stent drainage.
(6439)Figure 6228. Nasobiliary tube. The prepapillary stone has not been mobilized. There
is no free reflux of contrast medium around the stone or the tube into the duodenum;
decompression is incomplete. Slight suction on the nasobiliary tube will evacuate the bile tree.
The endoscopic method of decompression of the bile ducts via the upper gastrointestinal tract and
main duodenal papilla is physiologic in its approach. Only a small EPT incision is required for
introduction of the nasobiliary tube. Endoscopic decompression may be performed in virtually any
patient. The use of sedative drugs can be minimized, general anesthesia is not required, and the
procedure can be performed by an experienced endoscopist even if there are the additional serious
problems of advanced septic shock, chills, mental confusion, defects in blood clogging, advanced age,
or the presence of other diseases. Percutaneous transhepatic drainage is also an effective, albeit more
invasive, method of decompression.27,145(6440)
Before the era of EPT, the mortality rate for patients with septic shock treated by surgical
decompression was 33%; for patients who did not undergo surgery because of advanced septic shock,
the mortality rate was 100% (Table 623). In my (D.F.W.W.) experience, the mortality rate for patients
undergoing endoscopic decompression by means of nasobiliary tube insertion was 23%. Furthermore,
endoscopically treated patients were on average 10 years older than surgically treated patients. These
data underscore the fact that immediate biliary decompression in acute suppurative cholangitis is the
only effective treatment. Although there are difficulties with the use of collected data that are not fully
comparable, this type of analysis indicates a favorable trend that supports the logical basis for the
endoscopic management of acute suppurative cholangitis in association with biliary obstruction.
Mortality of Acute Suppurative Cholangitis: Comparison of Surgical and

Endoscopic Therapy
TABLE 623
Surgical decompression
(19591976)*
Endoscopic data (19761983)
Total
Percent
Total
WHOLE
GROUP
OPERATED
GROUP
NONOPERATED
GROUP
51/110
29/88
22/22
46
5/22
33
5/22
100
0/0
23
Percent
23
* Data collected from 9 reports: references 142 and 208215.
Personal (D.F.W.W.) series.
These data have been confirmed by others146150(6441) and have been summarized in a review by
Leung and Venezuela in 1991.151(6442) Suppurative cholangitis can be recurrent in certain
populations, particularly Asians with so-called Oriental cholangiohepatitis. Endoscopic therapy can be
especially efficacious in this group.152,153(6443) In the retrospective study of Leese et al.,146(6444)
the 30-day mortality for early surgical decompression was significantly higher at 21% compared with
4.6% for endoscopic therapy, despite the fact that surgical patients were significantly younger and had
fewer associated risk factors than those who underwent endoscopic therapy. These investigators also
found that low levels of serum albumin and elevated serum urea nitrogen levels at presentation were
associated with a higher mortality rate. A prospective study by Boender et al.154(6445) found that
delay in establishing biliary drainage was associated with an increased risk of complications in patients
who were not immediately responsive to antibiotics and conservative treatment. Although these
investigators found that delayed drainage did not increase the morbidity associated with suppurative
cholangitis in patients who responded to antibiotics, this approach can be considered venturesome
inasmuch as it is usually difficult to predict whether a patient will respond to antibiotics. For patients
who are seriously ill with septic shock, especially those of advanced age and with additional medical
problems, rapid establishment of biliary drainage is the treatment of choice. In expert hands,
endoscopic drainage can be accomplished rapidly and safely in the majority of patients. However,
when endoscopic therapy is not available or fails, percutaneous transhepatic drainage or surgical
drainage must be performed.155(6446)
Lai et al.156(6447) randomized patients with severe acute cholangitis due to choledocholithiasis to
undergo either endoscopic biliary drainage or surgical decompression. There were 41 patients in each
group. The diagnosis of choledocholithiasis was established by retrograde cholangiography in all
cases, at which point patients were randomized to undergo endoscopic therapy by means of a small
EPT with insertion of a nasobiliary tube or emergency surgical decompression. Patients in the
endoscopically treated group subsequently underwent either endoscopic stone extraction or surgery
based on the presence or absence of the gallbladder and their general medical condition. The groups
were evenly matched with regard to age, gender, clinical manifestations, biochemical parameters, and
risk factors. Three patients treated endoscopically and 5 who underwent surgery had concomitant
biliary strictures. Treatment-related complications occurred in 34% and 66% (p > .05) of the
endoscopically and surgically treated patients, respectively. Although the length of time required for
clinical manifestations of cholangitis to resolve was about the same in each group, significantly more
surgically treated patients required ventilatory support. Four deaths (10%) occurred among
endoscopically treated patients versus 13 deaths (32%) in the surgical group, a statistically significant
difference. Factors noted before decompression that were found to be independently associated with a
greater risk of death included thrombocytopenia, hypoalbuminemia, elevated serum urea nitrogen, and
associated medical problems.
Stone Dissolution
Any solvent that can be perfused through a T-tube can also be given through a nasobiliary drainage
tube or any tube that reaches the stones. Alpha-1-glyceromonooctanoin has been in use for many
years as a solvent for biliary stones.45,46(6448) It has been largely superseded by MTBE as an
effective solvent for cholesterol stones in the gallbladder.157(6449) Although MTBE has been used
successfully to dissolve stones in the bile duct,158(6450) it has limitations for the treatment of
choledocholithiasis because of potential toxicity if absorbed from the duodenum. Moreover, it has been
recognized that the chemical composition of residual stones after surgery differs from that of stones
that recur in the bile ducts after a long period of time. The former likely contain large amounts of
cholesterol, whereas the latter usually have many components, including bile pigment (30 to 80%),
organic and inorganic calcium salts (3 to 10%), cholesterol (20%), and an organic matrix (10 to 30%)
that is formed by glycoproteins, mucin, and carbohydrates.47(6451) Pigment stones often develop a
calcareous rim. Best results using MTBE perfusion can be expected with predominantly cholesterol
stones detected in the early postoperative period. For stones present in the ducts for long periods of
time, or especially for recurrent stones after surgery and for stones in malformed bile ducts, the use of
different solvents with respect to their differing chemical composition has been proposed.47(6452)
Solvents include EDTA, ursodeoxycholate, and detergents, which act best in an alkaline milieu of pH
9.5. However, none of the presently available agents is entirely satisfactory and further development of
new agents is needed. In my (D.F.W.W.) experience and that of others, about two thirds of stones can
be dissolved.47(6453) Intrahepatic stones can also be treated by chemical dissolution (Figure
6229).159(6454)
(6455)Figure 6229. A, Multiple stones in the common bile duct and in the left hepatic duct
can be dissolved (B), via an individually designed nasobiliary tube with a small loop for
fixation and a long tip for placement above the last stone.
Before the dissolution procedure is begun, it must be established radiographically that there is free
reflux of perfused fluid from the bile ducts into the duodenum. If not, obstructive cholangitis will be
induced. If it is suspected on clinical grounds that pressure is rising in the duct, the perfusion must be
stopped and the nasobiliary or the percutaneous tube used to drain the duct externally. Aliquots of bile
may also be taken via the tube for culture.
There are some significant drawbacks to stone dissolution. As a result, this type of therapy is not used
extensively as a primary method for clearing the bile duct of stones. Complete dissolution may require
perfusion of the solvent for more than 1 week up to several weeks. However, it can be performed on
an outpatient basis. An adequate rate of flow must be maintained, and this may not be tolerated by all
patients. Side effects include abdominal cramps and diarrhea. It is often not necessary to completely
dissolve the calculi. The solvent may reduce the size of the stone or soften it to the point that
endoscopic extraction is possible. In summary, stone dissolution is a supportive therapy for rare and
difficult cases in which bile duct stones are not treatable by other means.
Management of Choledocholithiasis
The thoroughgoing change in the treatment of bile duct stones began on June 6, 1973, with the first
EPT and stone extraction by Classen and Demling.139(6456) Two years later, Berk and
Kaplan74(6457) cautiously stated, "It is entirely possible that as more endoscopists become skilled in
their [EPT] performance, these procedures may well assume the status of standard methods for
removing low-lying stones from the common bile duct." More than enough data are now available to
verify this prophecy, even though there are caveats. It is impossible to compare the basically different
surgical and endoscopic methods of stone removal with respect to all possible circumstances. Further,
comparison must be based on reports that have usually been published some years apart. Collected
statistics from numerous investigators and centers where the procedures are performed in great
numbers are most reliable. Pooled data are often obtained by questionnaire with the knowledge that
success may be overstated, and risks and complications may be underestimated. In collected series,
the diverse results of different investigators make interpretation of the composite data more difficult.
Although reservations are inherent in this method, and prospective comparative controlled studies are
more desirable in principle, these are the only data available. As a result of the impressive success of
EPT and endoscopic extraction of bile duct stones, there is a lack of recent data concerning surgical
success and complication rates, which nowadays are likely to be better than older but available data
suggest. Nevertheless, the established success of the endoscopic procedure probably prevents the
performance of controlled studies comparing stone extraction using EPT with the open surgical
procedure.
The impact and potential role of laparoscopic methods of bile duct exploration and stone extraction are
not yet known. Laparoscopic instruments and techniques for clearance of stones from the bile duct are
under development.160(6458) It is clear that the technical level of skill required for this type of
procedure is considerably higher than that needed for laparoscopic cholecystectomy alone. Success
rates are therefore operator-dependent at present. This being the case, it is logical that complication
rates, especially various forms of bile duct injury, will be higher when laparoscopic bile duct surgery is
performed by an inexperienced surgeon. A certain number of cases will be converted to an open
procedure. Nevertheless, there has been a steady increase in the number of favorable reports of
laparoscopic clearance of bile duct stones.161164(6459)
The role of ERCP in relation to laparoscopic cholecystectomy in patients suspected of having bile duct
stones is discussed in Chapter 56: Indications, Contraindications, and Complications of Diagnostic

Endoscopic Retrograde Cholangiopancreatography.
Data on mortality rates for endoscopic therapy of choledocholithiasis compare favorably with those for
operative therapy (Tables 624, 625, 626). The surgical mortality rate for choledocholithotomy is not
precisely known, but it is probably similar to that for cholecystectomy with choledochotomy and to
reoperation for retained stones. The surgical mortality rate in acute suppurative cholangitis is about
46% (see Table 623).
TABLE 624
Mortality: Biliary Surgery
TYPE OF SURGERY
Cholecystectomy
Chronic Cholecystitis
1.50% (n = 24,415)
1.02% (n = 3885)
0.18% age < 60 years
REFERENCE
Acute Cholecystitis
3.50% (n = 4206)
3.15% (n = 564)
0.39% age < 60
years
4.85% age > 60
years
3.00% age > 60 years
216
217
1.30% (n = 4327)
Cholecystectomy and Choledochotomy
4.26% (n = 844)
5.36% (n = 1358)
12.80% age > 60 years
47.68% emergency surgery
Reoperation for Retained Stones
34%
Transduodenal Sphincterotomy
4.3% (n = 13,225)
* Collected data.
165
217
165
218,219
166*
Mortality: Biliary Surgery for

Gallstone-Induced Pancreatitis*
TABLE 625
OVERALL
EARLY
DELAYED
11.0% (132)
2% (46)
16% (86)
2.3% (172)
12% (24)
0% (134)
5.3% (114)
31% (17)
2% (98)
* Number of patients in parentheses.
LATE
REFERENCE
7% (14)
41
42
43
Gallstone-Induced Pancreatitis*
OVERALL
EARLY
TABLE 626
DELAYED
LATE
REFERENCE
Mortality: Endoscopic Treatment of Gallstones in the
Bile Duct
Choledocholithiasis
(including cholestasis
and cholangitis)
Acute obstructive
suppurative cholangitis
Biliary pancreatitis
MORTALITY (%)
PATIENTS (n)
1.10
0.87
1.3
1989
20,000
1060
23
22
0
18
58
11
REFERENCE
144*
37*
Author
(D.F.W.W.)
140
220*
111
* Collected data.
Varied with level of experience from 0.3 to 5.1%.
Mean age 68 years.
Retained, overlooked, or forgotten calculi are discovered by postoperative T-tube cholangiography at a

minimum in 0.3 to 0.7% of patients.76,165(6460) The frequency usually reported is substantially higher
at 1 to 4%.75(6461) The late complications of common bile duct surgery include stricture, papillary
stenosis, and retention of suture material or clips in the bile duct. The exact incidence of these various
complications is not known.
Patients who have had calculi removed from the common bile duct at cholecystectomy seem more
likely to develop recurrent calculi months or years later than those who did not have choledocholithiasis
at a primary operation.68,166(6462) There are a number of reports in which surgical clips are
described as the nidus for stone formation.167170(6463) The relative importance of this mechanism
to the overall clinical problem of bile duct stone formation after surgery is uncertain.
The success rate for the initial attempt at EPT and stone extraction is reduced 6 to 9% by the presence
of stones that cannot be removed using the standard Dormia basket or extraction balloon catheter.
The technique of mechanical destruction, referred to as mechanical lithotripsy, decreases the number
of instances in which all stones cannot be cleared from the bile duct.38,39(6464) In my (D.F.W.W.)
experience during a 1-year period, this simple technique reduces the failure rate for complete stone
removal to 1.3% of those patients in whom EPT was successful (Table 627). Numerous reports now
confirm the value of mechanical lithotripsy.171173(6465) Although success rates decrease somewhat
with increasing stone diameter, the introduction of baskets made of high-tensile-strength steel wire has
made it possible to fracture even the largest of stones.174(6466)
TABLE 627
Stone Extraction After Endoscopic Papillotomy,
1983*
EPT and stone extraction
With mechanical lithotripsy
197
0
TABLE 627
Stone Extraction After Endoscopic Papillotomy,
1983*
Removal unsuccessful
1
Removal not attempted
3
Total patients
224
EPTendoscopic papillotomy.
*In patients in whom extraction was unsuccessful at initial EPT.
Technically, there are several approaches to mechanical fragmentation of bile duct stones. In the first
method to be described, a stone is captured in a basket, the handle of the basket wire is detached
using a wire cutter, the endoscope is removed, and a steel spiral sheath is passed over the basket wire
until it is adjacent to the basketed stone. By means of a mechanical crank, the basket and stone are
pulled against the spiral sheath until the stone breaks. Newer devices operate on the same principle
but do not require removal of the endoscope.
Intraductal shockwave lithotripsy, whether electrohydraulic or by laser, is an effective alternative or
supplement to mechanical lithotripsy.
Electrohydraulic lithotripsy is based on the generation of an abrupt high-pressure shock wave within a
fluid medium by means of a high-voltage electric discharge. This method of fracturing stones has been
used endoscopically175(6467) as well as transhepatically.176(6468) An important technical aspect of
this technique is that the stones to be fractured must be kept within a fluid medium. This has been
accomplished endoscopically by attachment of an occlusion balloon to the electrode catheter just
proximal to the tip of the device. Bile duct stones can be destroyed by laser energy. Such energy has
been delivered to bile duct stones via a number of approaches including direct visualization using
peroral and percutaneous cholangioscopy,177,178(6469) placement of the laser wave guide against
the stone using fluoroscopic guidance, centering of the wave guide on the stone by means of a balloon
catheter, and entrapment of the stone in a special basket through which a wave guide can be passed
to the stone.179(6470) Newer laser systems for stone fragmentation are able to recognize when the
laser wave guide is not in contact with a stone by analysis of backscattered light. Because the laser
output is interrupted when the wave guide is not in contact with a stone, these systems are easier to
use and lithotripsy can often be performed under fluoroscopic control.180,181(6471)
When the standard methods of extraction fail to clear the bile duct of stones, selection of one of the
methods of lithotripsy just described can be problematic because relatively high success rates have
been reported with all techniques. Unfortunately, there are no comparative studies. Binmoeller et
al.182(6472) reported high rates of success for both mechanical and electrohydraulic lithotripsy.
Because it is relatively inexpensive, is readily available, and has a high success rate, mechanical
lithotripsy is an excellent first alternative when stones cannot be extracted using basic methods.
Although there are some reports of endoscopic electrohydraulic lithotripsy, there are relatively few data
concerning the efficacy and safety of the technique. Laser lithotripsy is expensive and therefore not
readily available. Because of the high success rates for less costly techniques, laser destruction of bile
duct stones is seldom indicated. However, it may be successful in cases where all other methods of
stone fragmentation have failed.
Insertion of a biliary endoprosthesis has been used as temporary, intermediate-term, and permanent
therapy for choledocholithiasis.183185(6473) As a temporary measure in patients with cholangitis in
whom there is an intention of further therapy once the acute phase of the illness has subsided, a
nasobiliary drain is preferable to a stent because patency can be maintained and bile samples can be
obtained for culture. Peters et al.186(6474) used biliary stents as a temporary measure in 13 patients
and as definitive treatment in 27 patients with choledocholithiasis. Six patients (15%) developed early
complications. Late complications occurred in 8 patients (20%), including 3 with cholangitis and 1 with
cholecystitis, 2 of whom died as a result of the complication. In the study of Maxton et al.,187(6475)
there were 26 elderly patients with additional medical problems in whom biliary stents were left in place
for 1 year or longer. Four episodes of cholangitis occurred in this group, all of which were treated
successfully. Insertion of a biliary stent was used as a permanent treatment measure in 58 patients in
the study of Bergman et al.188(6476) Follow-up ranged from 1 to 117 months (median 36 months),
during which time 23 patients (40%) experienced a total of 34 complications, with cholangitis being the
most common untoward event. Biliary complications were noted to increase in relation to the length of
time stents were in place. Forty-four patients died, with death being due to biliary-related causes in 9.
A few patients with Mirizzi's syndrome have been treated endoscopically with relatively good
results.189,190(6477) In some cases, this was intended to improve the patient's clinical status before
surgery; in others, endoscopic treatment was performed without subsequent surgery. Biliary stenting
was employed in the majority of cases. Of the 25 patients diagnosed as having Mirizzi's syndrome by
retrograde cholangiography in the study of England and Martin,190(6478) 11 had endoscopic therapy
before surgery and 9 were managed by long-term stent insertion.
Intraductal stone destruction by means of one of the methods described previously, especially
mechanical lithotripsy, with extraction only of the resulting fragments and debris, even for
medium-sized stones, has some evident advantages. Only a medium-length, "tailored" papillotomy is
necessary, in which case the risk of bleeding and perforation is lower. Furthermore, an incomplete
sphincterotomy may preserve some degree of papillary function. For these reasons, I (D.F.W.W.) no
longer perform large papillotomies, especially in younger patients. Others have proposed that
endoscopic balloon dilation of the sphincter of Oddi is a more physiologic and safer method of gaining
access to the bile duct for removal of stones (see Chapter 60: Endoscopic Papillotomy).
When it is not possible to reach the papilla because of anatomic problems such as a juxtapapillary
diverticulum, upside-down stomach, or prior surgery such as a Billroth II resection or Roux-en-Y
hepaticojejunostomy, percutaneous transhepatic access to the bile ducts offers an alternative
approach, one that may be technically difficult but very effective with a relatively low risk (see Chapter
57: Technique of Endoscopic Retrograde Cholangiopancreatography).
Despite the fact that more than two decades have elapsed since the first EPT, there are relatively few
long-term, longitudinal follow-up data on patients who have undergone the procedure, especially
information based on objective studies such as cholangiography. In an early study, 51 patients
underwent follow-up evaluation including ERCP at a mean of 21.6 months after EPT.191(6479) None
had papillary stenosis, recurrent stones, or cholangitis, although massive bacterial colonization of the
bile ducts was present in nearly all cases. These results agree with findings after surgical
sphincterotomy, specifically that duodenobiliary reflux is common but without adverse
consequences.192(6480) In a collective series, recurrent stones were found in 6% of 954 patients who
had undergone EPT and stone extraction.193(6481) Such stones are soft and can be extracted easily.
A dilated bile duct after EPT seems to predispose to recurrent stones. This has been described in
patients who have undergone surgery.194(6482) Stenosis at the papillotomy site occurred after EPT in
2.3% of patients.193(6483)
Bergman et al.195(6484) carried out a retrospective follow-up evaluation of 100 patients who
underwent EPT for bile duct stones between 1976 and 1980. Inclusion criteria in this study, reported in
1996, were complete clearance of stones from the bile duct, prior cholecystectomy or elective removal
of the gallbladder within 2 months of EPT, and age of 60 years or less. Data were obtained for 94
patients (26 men, 68 women) who ranged in age from 23 to 60 years (mean 51 years) at the time of
EPT. Fourteen patients (15%) experienced complications within 30 days of the procedure, 1 of whom
died as a result of a procedure-related retroperitoneal perforation. The median elapsed time between
EPT and follow-up evaluation was 15 years, with a range of 3 to 18 years. Twelve deaths had occurred
from causes unrelated to the biliary tract or pancreas. During this time, however, 36 biliary
complications occurred in 22 patients (24%), the most common being symptoms of recurrent
choledocholithiasis (21 patients). Twenty patients with late complications underwent ERCP; findings
included recurrent bile duct stones in 13 and stenosis at the papillotomy site in 9 cases. Most of these
late complications were managed by endoscopic methods, but 1 patient underwent surgery after
endoscopic therapy failed and 1 patient died before ERCP as a result of cholangitis. These results
indicate the need for further long-term studies of patients who have undergone endoscopic treatment
of choledocholithiasis.
Another, more theoretical problem is the potential for carcinoma of the colon as a result of changes in
bile acid metabolism, such as occurs after cholecystectomy,196,197(6485) with resultant accelerated
turnover in the enterohepatic circulation of the bile acid pool and a consequent increase of secondary
bile acids. It has been shown that after EPT in patients who have undergone cholecystectomy, there is
no further augmentation in the turnover of bile acids,198,199(6486) and diurnal rhythm of biliary
lithogenicity persists after cholecystectomy and papillotomy.200(6487) But epidemiologic studies of this
issue are not available because EPT has been performed in large numbers of patients only for a
relatively short period of time.
Another problem after EPT and successful clearance of stones from the bile ducts is the effect on the
intact gallbladder. This is discussed in Chapter 60: Endoscopic Papillotomy.
The aim of surgical and endoscopic approaches to calculi of the bile ducts is the same: to cure as
many patients as possible of bile duct stones with the lowest possible risk. Endoscopy is safer, and the
convalescent period after the procedure is minimal. However, not all stones can be removed and
problems with the gallbladder cannot be resolved. Surgery has a higher risk and requires a prolonged
recovery period, but bile duct problems and gallbladder disease can be treated by one operation. In
some situations, it is of benefit or even necessary to combine different methods (see Table 621) to
achieve definitive treatment while minimizing risk to the patient. Suggested approaches to
choledocholithiasis are outlined in Figures 6230 and 6231. In actual practice, the approach depends
substantially on circumstances, especially available expertise.
(6488)Figure 6230. Suggested approach to treatment of choledocholithiasis with a history of

previous cholecystectomy. EPTendoscopic papillotomy; ESWLextracorporeal
shockwave lithotripsy.
(6489)Figure 6231. Suggested approach to treatment of choledocholithiasis with no history

of previous surgery. EPTendoscopic papillotomy; ESWLextracorporeal shockwave
lithotripsy.
Complications
The major risks associated with endoscopic treatment of choledocholithiasis relate to the EPT
procedure rather than the actual extraction of stones from the duct. Complications of EPT are
discussed in Chapter 60: Endoscopic Papillotomy. Relatively few complications are associated with
attempts to remove stones from the bile duct. The most common problem related to endoscopic
maneuvers to extract stones is precipitation of cholangitis. The risk of this complication is markedly
reduced by duct clearance of stones. The risk increases if stones remain impacted in the duct without
adequate decompression. Basket entrapment may occur if a stone too large to be removed is captured
in the basket. As discussed previously, there are a number of effective techniques for dealing with this
problem. Other rare complications related to endoscopic extraction of bile duct stones include gallstone
ileus201(6490) and cystic duct perforation.202(6491) Complications may occasionally be related to
malfunction of endoscopic devices.203(6492)
Cappell204(6493) reported four cases in which ERCP was performed for symptomatic
choledocholithiasis in four patients from 15 to 56 days after acute myocardial infarction. There were no
complications, including cardiovascular complications, and ductal clearance of stones was successful
in all cases. Friedman and Friedman205(6494) reported a case of ERCP and EPT for bile duct stones
in a pregnant patient.
The results of a small but prospective study of patients who underwent endoscopic treatment for bile
duct stones on an outpatient basis suggest that this may be a safe practice.206(6495) However, only
97 patients were included in this study, and confirmatory data based on larger, prospective studies are
needed before this practice can be recommended. The incidence of pancreatitis was 2.1% and
postpapillotomy bleeding 3.2%. There were no cases of perforation or sepsis. One patient was
admitted to the hospital during the postprocedure observation period, and 2 were hospitalized within 24
hr of discharge.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Goldman DE, Gholson CF. Choledocholithiasis in patients with normal serum liver enzymes.
Dig Dis Sci 1995;40:10658.
Roston AD, Jacobson IM. Evaluation of the pattern of liver tests and yield of cholangiography
in symptomatic choledocholithiasis: A prospective study. Gastrointest Endosc 1997;45:3949.
Cooperberg PL, Burhenne HJ. Real-time ultrasonography. Diagnostic technique of choice in a
calculus gallbladder disease. N Engl J Med 1980;302:12779.
Hessler PC, Hill DS, Deforie FM, Rocco AF. High accuracy sonographic recognition of
gallstones. AJR 1981;136:51720.
Rettenmaier G. Sonografischer Oberbauchstatus. Aussagefhigkeit und Indikationen der
Ultraschall-Schnittbilduntersuchung des Oberbauchs. Internist 1976;17:54964.
Myllyla V, Paivansalo M, Pyhtinen J, et al. Sensitivity of ultrasonography in the demonstration
of common bile duct stones and its ranking in comparison with intravenous cholangiography
and endoscopic retrograde cholangiopancreatography. Rfo Fortschr Geb Rntgenstr
Nuklearmed 1984;141:1924.
Seitz K. Ultraschalldiagnostik bei Erkrankungen der Gallenwege. Dtsch Med Wochenschr
1985;110:153942.
Berggren P, Farago I, Gabrielsson N, Thor K. Intravenous cholangiography before 1000
consecutive laparoscopic cholecystectomies. Br J Surg 1997;84:4726.
Osnes M, Grondseth K, Larsen S, et al. Comparison of endoscopic retrograde and
intravenous cholangiography in diagnosis of biliary calculi. Lancet 1978;2:230.
Goodman MW, Ansel HJ, Vennes JA, et al. Is intravenous cholangiography still useful?.
Tham TC, Collins JS, Watson RG, et al. Diagnosis of common bile duct stones by
intravenous cholangiography: Prediction by ultrasound and liver function tests compared with
endoscopic retrograde cholangiography. Gastrointest Endosc 1996;44:15863.
Neitlich JD, Topazian M, Smith RC, et al. Detection of choledocholithiasis: Comparison of
unenhanced helical CT and endoscopic retrograde cholangiopancreatography. Radiology
1997;203:7537.
Brakel A, Lameris JS, Nijs HGT, et al. Predicting gallstone composition with CT: In vivo and in
vitro analysis. Radiology 1990;174:33741.
Lehmann CD, Goldman ML, Baron RL, et al. Gallstone susceptibility to in vitro fragmentation
by a 480-nm pulsed dye laser: Correlation with computed tomography characteristics. Invest
Radiol 1991;26:799803.
15. Chan YL, Lam WW, Metreweli C, Chung SC. Detectability and appearance of bile duct
calculus on MR imaging of the abdomen using axial T1- and T2-weighted sequences. Clin
Radiol 1997;52:3515.
16.
Regan F, Fradin J, Khazan R, et al. Choledocholithiasis: Evaluation with MR cholangiography.
AJR 1996;167:14415.
17. Chan YL, Chan AC, Lam WW, et al. Choledocholithiasis: Comparison of MR cholangiography
and endoscopic retrograde cholangiography. Radiology 1996;200:859.
18. Prat F, Amouyal G, Amouyal P, et al. Prospective controlled study of endoscopic
ultrasonography and endoscopic retrograde cholangiography in patients with suspected
common-bile duct lithiasis. Lancet 1996;347:759.
19. Sugiyama M, Atomi Y. Endoscopic ultrasonography for diagnosing choledocholithiasis: A
prospective comparative study with ultrasonography and computed tomography. Gastrointest
Endosc 1997;45:1436.
20. Bilbao MK, Dotter CT, Lee TG, Katon RM. Complications of endoscopic retrograde
cholangiopancreatography (ERCP). A study of 10,000 cases. Gastroenterology
1976;70:31420.
21. Harbin WP, Mueller PR, Ferrucci JT Jr. Transhepatic cholangiography: Complications and
use patterns of the fine-needle technique. Radiology 1980;135:1522.
22. Kreek MJ, Balint JA. "Skinny needle" cholangiographyResults of a pilot study of a voluntary
prospective method for gathering risk data on new procedures. Gastroenterology
1980;78:598604.
23. Nebel OT, Silvis SE, Rogers G, et al. Complications associated with endoscopic retrograde
cholangiopancreatography: Results of 1974 A.S.G.E. survey. Gastrointest Endosc
1975;22:346.
24. Beneventano TC, Schein CJ. The pseudocalculus sign in cholangiography. Arch Surg
1969;98:7313.
25.
Wurbs D, Classen M. Transpapillary long-standing tube for hepatobiliary drainage. Endoscopy
1977;9:1923.
26. Soehendra N, Reynders-Frederix V. Palliative Gallengangsdrainage. Dtsch Med Wochenschr
1979;104:2067.
27. Nakayama T, Ikeda A, Okuda K. Percutaneous transhepatic drainage of the biliary tract.
28. Neuhaus H, Hoffmann W, Classen M. Laser lithotripsy of pancreatic and biliary stones via 3.4
mm and 3.7 mm miniscopes: First clinical results. Endoscopy 1992;24:20814.
29. Yamakawa T, Komaki F, Shikata J. Biliary tract endoscopy with an improved
choledochofiberscope. Gastrointest Endosc 1978;24:1103.
30. Burhenne HJ. Complications of nonoperative extraction of retained common duct stones. Am
J Surg 1976;131:2602.
31. Mazzariello RM. A fourteen-year experience with nonoperative instrument extraction of
retained bile duct stones. World J Surg 1978;2:44755.
32. Ido K, Kimura K, Kawamoto C, et al. Preliminary experience using laparoscopic transcystic
cholangioscopy for treatment of common bile duct stones. Endoscopy 1992;24:7503.
33. Peng SY, Yue MK, Qi VJ, Chen J. An operation for recurrent pigment
gallstonesSubcutaneous choledocho-jejunostomy. Ann R Coll Surg Engl 1983;65:501.
34. Russell E, Yrizzary JM, Huber JS, et al. Percutaneous transjejunal biliary dilatation: Alternate
management for benign strictures. Radiology 1986;159:20914.
35. Barker EM, Winkler M. Permanent access hepaticojejunostomy. Br J Surg 1984;71:18891.
36. Martin EC, Laffey KJ, Bixon R. Percutaneous transjejunal approaches to the biliary system.
Radiology 1989;172:10314.
37.
Machado G. Acute complications of endoscopic papillotomy. Endoscopic Papillotomy Now 10
Years Old. International Symposium. Erlangen. 1983.
38. Riemann JF, Demling L. Lithotripsy of bile duct stones. Endoscopy 1983;15:1916.
39. Frimberger E, Weingart J, Kuhner W, Ottenjann R. Eingeklemmter Papillenstein:
Mechanische Lithotripsie mglich. Dtsch Med Wochenschr 1983;108:38.

40. Sauerbruch T, Stern M. Fragmentation of bile duct stones by extracorporeal shock waves.
41. Koch H, Stolte M, Walz V. Endoscopic lithotripsy in the common bile duct. Endoscopy
1977;9:958.
42.
Leung JWC, Chung SSC. Electrohydraulic lithotripsy with peroral choledochoscopy. BMJ
1989;299:2958.
43. Ell C, Lux G, Hochberger J, et al. Laserlithotripsy of common bile duct stones. Gut
1988;29:74651.
44. Brambs HJ, Rieber A, Huppert PE, et al. Perkutan-transhepatische cholangioskopische
Farbstofflaser-Lithotripsie von Gallengangsteinen. Z Gastroenterol 1992;30:74851.
45. Thistle JL, Carlson GL, Hofmann AF, et al. Monooctanoin, a dissolution agent for retained
cholesterol bile duct stones: Physical properties and clinical application. Gastroenterology
1980;78:101622.
46. Leuschner U, Wurbs D, Baumgartel H, et al. Alternating treatment of common bile duct
stones with a modified glyceryl-1-monooctanoate preparation and a bile acid-EDTA solution
by nasobiliary tube. Scand J Gastroenterol 1981;16:497503.
47. Leuschner U, Gldutuna S, Hellstern A. Etiology, pathogenesis, and therapy of pigment
gallstones. Dig Dis 1991;9:28293.
48. Kaufmann HS, Magnuson TH, Lillemoe KD, et al. The role of bacteria in gallbladder and
common duct stone formation. Ann Surg 1989;209:58492.
49. Skar V, Skar AG, Stromme JH. Beta-glucuronidase activity related to bacterial growth in
common bile duct bile in gallstone patients. Scand J Gastroenterol 1988;23:8390.
50. Maki T. Pathogenesis of calcium bilirubinate gallstone: Role of E. coli, beta glucuronidase and
coagulation by inorganic ions, polyelectrolytes and agitation. Ann Surg 1966;164:90100.
51. Skar V, Skar AG, Midtvedt T, et al. Beta-glucuronidase-producing bacteria in bile from the
common bile duct in patients treated with endoscopic papillotomy for gallstone disease.
52.
Tritapepe R, DiPadova C, Rovagnati P. Are pigmented gallstones caused by a "metabolic"
liver defect?. BMJ 1980;280:832.
53. Ho JC, Ho KJ. Human a-glucuronidase: Measurement of its activity in gallbladder bile devoid
of intrinsic interference. Dig Dis Sci 1988;33:43542.
54. Yamaguchi I, Sato T, Matsushiro T, Sato H. Quantitative determination of D-glucaric acid in
bile in relation to inhibitory effect of bile on bacterial beta-glucuronidase. Tohoku J Exp Med
1965;87:12332.
55.
Hoffmann L, Weiss W, Classen M. Untersuchungen zum juxtapapillren Duodenaldivertikel.
Inn Med 1978;5:226.
56. Lotveit T, Osnes M, Larsen S. Recurrent biliary calculi. Duodenal diverticula as a predisposing
factor. Ann Surg 1982;196:302.
57. van der Spuy S. The relationship between juxtapapillary diverticula and biliary calculi. An
endoscopic study. Endoscopy 1979;11:197202.
58. Hall RI, Ingoldby CJ, Denyer ME. Periampullary diverticula predispose to primary rather than
secondary stones in the common bile duct. Endoscopy 1990;22:1278.
59. Chandy G, Hart WJ, Roberts-Thomson IC. An analysis of the relationship between bile duct
stones and periampullary duodenal diverticula. J Gastroenterol Hepatol 1997;12:2933.
60. Funch-Jensen P, Csendes A, Kruse A, Oster MJ. Common bile duct pressure and Oddi
sphincter pressure in patients with common bile duct stones with and without juxta-ampullar
diverticula of the duodenum. Scand J Gastroenterol 1979;14:2536.
61. Eggert A, Kirschner H, Teichmann W. Bakteriologische Befunde bei Patienten mit
Cholelithiasis und Duodenaldivertikeln. Chirurg 1979;50:4414.
62. Lotveit T. The composition of biliary calculi in patients with juxtapapillary duodenal diverticula.
63. Lotveit T, Osnes M, Aune S. Bacteriological studies of common duct bile in patients with
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
gallstone disease and juxtapapillary duodenal diverticula. Scand J Gastroenterol

1978;13:935.
Sheu BS, Shin JS, Lin XZ, et al. Clinical analysis of choledochoduodenal fistula with
cholelithiasis in Taiwan: Assessment by endoscopic retrograde cholangiopancreatography.
Ikeda S, Okada Y. Classification of choledochoduodenal fistula diagnosed by duodenal
fiberscopy and its etiological significance. Gastroenterology 1975;69:1307.
Soehendra N. Endoskopische und radiologische KontrollUntersuchungen nach
transduodenaler Papillotomie. Langenbecks Arch Chir 1976;34:3949.
Urakami Y, Kishi S. Endoscopic fistulotomy (EFT) for parapapillary choledochoduodenal
fistula. Endoscopy 1979;10:28994.
Glenn F. Postcholecystectomy choledocholithiasis. Surg Gynecol Obstet 1972;134:24952.
Marinella MA. Choledocholithiasis causing obstructive jaundice 52 years after
cholecystectomy. J Am Geriatr Soc 1995;43:13189.
Mirizzi PL. Sindrome del conducto hepatico. J Int Chir 1948;8:73177.
Miller FH, Sica GT. Mirizzi syndrome associated with gallbladder cancer and biliary-enteric
fistulas. AJR 1996;167:957.
Redaelli CA, Buchler MW, Schilling MK, et al. High coincidence of Mirizzi syndrome and
gallbladder carcinoma. Surgery 1997;121:5863.
Meyer GJ, Runge D, Gebhardt J. Das Mirizzi-Syndrom und Varianten. Z Gastroenterol
1990;28:2025.
Berk JE, Kaplan AA. Choledocholithiasis. In Bockus HS (ed). Gastro-enterology. Vol III.
Philadelphia: WB Saunders, 1976;pp. 84354.
Girard RM, Legros G. Retained and recurrent bile duct stones. Surgical or nonsurgical
removal?. Ann Surg 1981;193:1504.
Way LW. Retained common duct stones. Surg Clin North Am 1973;53:113047.
Ranson JHC. The timing of biliary surgery in acute pancreatitis. Ann Surg 1979;189:65462.
DeBolla AR, Obeid ML. Mortality in acute pancreatitis. Ann R Coll Surg Engl 1984;66:1846.
Goodman AJ, Neoptolemos JP, Carr-Locke DL, et al. Detection of gallstones after acute
Mayer AD, McMahon MJ, Benson EA, et al. Operation upon the biliary tract in patients with
acute pancreatitis: Aims, indications and timing. Ann R Coll Surg Engl 1984;66:17983.
. MRC Multicentre Trial: Death from acute pancreatitis. Lancet 1977;2:6325.
de Beaux AC, Palmer KR, Carter DC. Factors influencing morbidity and mortality in acute
pancreatitis: An analysis of 279 cases. Gut 1995;37:1216.
Opie EL. The etiology of acute hemorrhagic pancreatitis. Johns Hopkins Hosp Bull
1901;121:1828.
Kelly TR. Gallstone pancreatitis: Pathophysiology. Surgery 1976;80:48892.
Cuschieri A, Hughes JH. Pancreatic reflux during operative choledochography. Br J Surg
1973;60:9336.
Armstrong CP, Taylor TV, Jeacock J, et al. The biliary tract in patients with acute gallstone
pancreatitis. Br J Surg 1985;72:5515.
Acosta JM, Ledesma CL. Gallstone migration as a cause of acute pancreatitis. N Engl J Med
1984;190:4847.
Acosta JM, Peligrini CA, Skinner DB. Etiology and pathogenesis of acute biliary pancreatitis.
Surgery 1980;88:11825.
Kelly TR. Gallstone pancreatitis: The timing of surgery. Surgery 1980;88:34550.
Stone HH, Fabian TC, Dunlop WE. Gallstone pancreatitis: Biliary tract pathology in relation to
time of operation. Ann Surg 1981;194:30510.
Dixon JA, Hillam JD. Surgery treatment of biliary tract disease associated with acute
pancreatitis. Am J Surg 1970;120:3715.
Paloyan D, Simonowitz D, Skinner DB. The timing of biliary tract operations in patients with
pancreatitis associated with gallstones. Surg Gynecol Obstet 1975;141:7379.
93. Cuschieri A, Cumming JG, Wood RA, Baker PR. Evidence for sphincter dysfunction in
patients with gallstone associated pancreatitis: Effect of ceruletide in patients undergoing
cholecystectomy for gallbladder disease and gallstone associated pancreatitis. Br J Surg
1984;71:8858.
94. Acosta JM, Rubio Galli OM, Rossi R, et al. Effect of duration of ampullary gallstone
obstruction on severity of lesions of acute pancreatitis. J Am Coll Surg 1997;184:499505.
95. Diehl AK, Holleman DR Jr, Chapman JB, et al. Gallstone size and risk of pancreatitis. Arch
Intern Med 1997;157:16748.
96. Lee SP, Nicholls JF, Park HZ. Biliary sludge as a cause of acute pancreatitis. N Engl J Med
1992;326:58993.
97. Ranson JHC. Etiologic and prognostic factors in human acute pancreatitis: A review. Am J
98. Imrie CW, Benjamin IS, Ferguson JC, et al. A single centre double blind trial of Trasylol
therapy in primary acute pancreatitis. Br J Surg 1978;65:33741.
99. Obsorne DH, Imrie CW, Carter DC. Biliary surgery in the same admission for
gallstone-associated acute pancreatitis. Br J Surg 1981;68:75861.
100. Blamey SL, Imrie CW, O'Neill J, et al. Prognostic factors in acute pancreatitis. Gut
1984;25:13406.
101. Leese T, Shaw D. Comparison of three Glasgow multifactor prognostic scoring systems in
acute pancreatitis. Br J Surg 1988;75:4602.
102. McMahon MJ, Pickford A, Playforth MJ. Early prediction of severity in acute pancreatitis using
peritoneal lavage. Acta Chir Scand 1980;146:1716.
103. Nordestgaard AG, Wilson SE, Williams RA. Early computerized tomography as a predictor of
outcome in acute pancreatitis. Am J Surg 1986;152:12732.
104. Neoptolemos JP, Fossard DP, Berry JM. A prospective study of radionuclide biliary scanning
in acute pancreatitis. Ann R Coll Surg Engl 1983;65:1802.
105. Neoptolemos JP, Hall AW, Finlay DF, et al. The urgent diagnosis of gallstones in acute
pancreatitis: A prospective study of three methods. Br J Surg 1984;71:2303.
106. Acosta JM, Rossi R, Galli OMR, et al. Early surgery for acute gallstone pancreatitis:
Evaluation of systematic approach. Surgery 1978;83:36770.
107. Tondelli P, Stutz K, Harder F, et al. Acute gallstone pancreatitis: Best timing for biliary surgery.
Br J Surg 1982;69:70910.
108.
Lux G, Riemann JF, Demling L. Biliare PankreatitisDiagnostische und therapeutische
Mglichkeiten durch ERCP und endoskopische Papillotomie. Z Gastroenterol
1984;22:24656.
109.
Treske U, Schroeder U. Endoskopische Entfernung eines Konkrements aus der Vaterschen
Papille. In Lindner H (ed). Fortschritte der gastroenterologischen Endoskopie. Bd 5.
Baden-Baden, Kln, New York: Witzstrock, 1974;p 117.
110.
Classen M, Ossenberg FW, Wurbs D, Hagenmuller F. PancreatitisAn indication for
endoscopic papillotomy?. Endoscopy 1978;10:223.
111. Safrany L, Cotton PB. A preliminary report: Urgent duodenoscopic sphincterotomy for acute
gallstone pancreatitis. Surgery 1981;89:4248.
112. van der Spuy S. Endoscopic sphincterotomy in the management of gallstone pancreatitis.
113. Rosseland AR, Solhaug JH. Early or delayed endoscopic papillotomy (EPT) in gallstone
pancreatitis. Ann Surg 1984;199:1657.
114. Liu CL, Lo CM, Fan ST. Acute biliary pancreatitis: Diagnosis and management. World J Surg
1997;21:14954.
115. Shemesh E, Czerniak A, Schneabaum S, Nass S. Early endoscopic sphincterotomy in the
management of acute gallstone pancreatitis in elderly patients. J Am Geriatr Soc
1990;38:8936.
116. Neoptolemos JP, Carr-Locke DL, London NJ, et al. Controlled trial of urgent endoscopic
retrograde cholangiopancreatography and endoscopic sphincterotomy versus conservative
treatment for acute pancreatitis due to gallstones. Lancet 1988;2:97983.

117. Fan ST, Lai EC, Mok FP, et al. Early treatment of acute biliary pancreatitis by endoscopic
papillotomy. N Engl J Med 1993;328:22832.
118. Flsch UR, Nitsche R, Ldtke R, et al. Early ERCP and papillotomy compared with
conservative treatment for acute biliary pancreatitis. The German Study Group on Acute
Biliary Pancreatitis [see comments]. N Engl J Med 1997;336:23742.
119. Baillie J. Treatment of acute biliary pancreatitis [Editorial; comment]. N Engl J Med
1997;336:2867.
120.
Wurbs D. Gallensteinextraktion aus dem Ductus Wirsungianus. In Henning H (ed).
Fortschritte der gastroenterologischen Endoskopie. Bd 10. Baden-Baden, Kln, New York:
Witzstrock, 1979;p 182.
121. Scott AJ. Bacteria and disease of the biliary tract. Gut 1971;12:48792.
122. Gross E. Bakteriologische und histologische Untersuchungen an Leber und Gallenwegen vor
und nach Eingriffen am Sphinkterapparat. Med Welt 1977;28:10246.
123.
Eggert A, Wittmann DH, Schrder HJ, Schimmel G. Die Bedeutung intraoperativer bakterieller
Befunde bei Gallenblasen- und Gallenwegsoperationen. Mnch Med Wochenschr
1977;119:9558.
124.
Keighley MRB, Wilson G, Kelly JP. Fatal endotoxic shock of biliary tract origin complicating
transhepatic cholangiography. BMJ 1973;3:1478.
125. Blumgart LH, Cotton PB, Burwood R, et al. Endoscopy and retrograde
choledochopancreatography in the diagnosis of the jaundiced patient. Lancet
1972;2:126973.
126.
Helm EB, Hagenmller F, Stille W. Septikamie nach endoskopischen Eingriffen am
Gallengangs-system. Mnch Med Wochenschr 1983;125:2102.
128. Huang T, Bass JA, Williams RD. The significance of biliary pressure in cholangitis. Arch Surg
1969;98:62932.
129.
Jacobsson B, Kjellander J, Rosengreen B. Cholangiovenous reflux. An experimental study.
Acta Chir Scand 1962;123:31621.
130. Williams RD, Fish JC, Williams DD. The significance of biliary pressure. Arch Surg
1967;95:3749.
131. Csendes A, Sepulveda A, Burdiles P, et al. Common bile duct pressure in patients with
common bile duct stones with or without acute suppurative cholangitis. Arch Surg
1988;123:6979.
132. Stewart L, Pellegrini CA, Way LW. Cholangiovenous reflux pathways as defined by corrosion
casting and scanning electron microscopy. Am J Surg 1988;155:238.
133. Jendrzejewski JW, McAnally T, Jones SR, Katon RM. Antibiotics and ERCP: In vitro activity of
aminoglycosides when added to ionidated contrast agents. Gastroenterology 1980;78:7458.
134. Collen MJ, Hanan MR, Maher JA, Stubrin SE. Modification of endoscopic retrograde
cholangiopancreatography (ERCP) septic complications by the addition of an antibiotic to the
contrast media. Randomized controlled investigation. Am J Gastroenterol 1980;74:4936.
135. Helm EB, Wurbs D, Gundlach H, et al. Keimelimination und Konzentrationsbestimmungen
unter Mezlocillin in der Galle bei Gallenwegsinfektionen. Dtsch Med Wochenschr
1981;106:108790.
136. Helm EB, Wurbs D, Haag R, et al. Elimination of bacteria in biliary tract infections during
ceftizoxime therapy. Infection 1982;10:6770.
137. Williams R. The management of ascending cholangitis. I. Bacteriology and choice of
antibiotic. Proc R Soc Med 1969;62:2434.
138. Summerfield JA. Biliary obstruction is best managed by endoscopists. Gut 1988;29:7415.
139.
aus dem Ductus Choledochus. Dtsch Med Wochenschr 1974;99:4967.
140.
Wurbs D. Endoscopic papillotomy. Scand J Gastroenterol 1982;77(suppl):10715.
141.
142.
Rogers L. Biliary abscesses of the liver with operation. BMJ 1903;2:706.

Reynolds BM, Dargan EL. Acute obstructive cholangitis: A distinct clinical syndrome. Ann
Surg 1959;150:299303.
143.
Wurbs D. Calculus disease of the bile ducts. In Sivak MV (ed). Gastroenterologic Endoscopy.
Philadelphia: WB Saunders, 1987;p 663.
144.
Cremer M. Complications of endoscopic sphincterotomy. In Classen M, Geenen J, Kawai K
(eds). The Papilla Vateri and Its Diseases. Baden-Baden, Kln, New York: Witzstrock, 1979;p
134.
145. Ferruci JT Jr, Mueller PR, Harbin WP. Percutaneous transhepatic biliary drainage. Technique,
results, and applications. Radiology 1980;135:113.
146. Leese T, Neoptolemos JP, Baker AR, Carr-Locke DL. Management of acute cholangitis and
the impact of endoscopic sphincter-otomy. Br J Surg 1986;73:98892.
147. Gogel HK, Runyon BA, Volpicelli NA, et al. Acute suppurative cholangitis due to stones:
Treatment by urgent endoscopic sphincterotomy. Gastrointest Endosc 1987;33:2103.
148. Leung JW, Chung SCS, Sung JJY, et al. Urgent endoscopic drainage for acute suppurative
cholangitis. Lancet 1989;1:1307- 9.
149. Worthley CS, Toouli J. Endoscopic decompression for acute cholangitis due to stones. Aust N
Z J Surg 1990;60:3559.
150. Ditzel H, Schaffalitzky de Muckadell OB. Endoscopic sphincter-otomy in acute cholangitis due
to choledocholithiasis. Hepatogastroenterology 1990;37:2047.
151. Leung JWC, Venezuela RR. Cholangiosepsis: Endoscopic drainage and antibiotic therapy.
Endoscopy 1991;23:2203.
152. Sperling RM, Koch J, Sandhu JS, Cello JP. Recurrent pyogenic cholangitis in Asian
immigrants to the United States: Natural history and role of therapeutic ERCP. Dig Dis Sci
1997;42:86571.
153. Lam SK. A study of endoscopic sphincterotomy in recurrent pyogenic cholangitis. Br J Surg
1984;71:2626.
154. Boender J, Nix GA, de Ridder MA, et al. Endoscopic sphincterotomy and biliary drainage in
patients with cholangitis due to common bile duct stones. Am J Gastroenterol 1995;90:2338.
155. Himal HS, Lindsay T. Ascending cholangitis: Surgery versus endoscopic or percutaneous
drainage. Surgery 1990;108:62934.
156. Lai ECS, Mok FPT, Tan ESY, et al. Endoscopic biliary drainage for severe acute cholangitis.
N Engl J Med 1992;326:15826.
157. Allen MJ, Borody TJ, Bugliosi TF, et al. Rapid dissolution of gallstones by methyl tert-butyl
ether. N Engl J Med 1985;312:21720.
158. LaFerla G, McCulloch A, Murray WR. In-vivo chemical choledocholitholysis using MTBE. Br J
Hosp Med 1987;37:1634.
159. Leuschner U, Wurbs D, Landgraf H. Dissolution of biliary duct stones with mono-octanoin
[Letter]. Lancet 1979;2:1034.
160. Cuschieri A, Croce E, Faggioni A, et al. EAES ductal stone study. Preliminary findings of
multi-center prospective randomized trial comparing two-stage vs single-stage management.
Surg Endosc 1996;10:11305.
161. Berci G, Morgenstern L. Laparoscopic management of common bile duct stones. A
multi-institutional SAGES study. Society of American Gastrointestinal Endoscopic Surgeons.
162. Dion YM, Ratelle R, Morin J, Gravel D. Common bile duct exploration: The place of
laparoscopic choledochotomy. Surg Laparosc Endosc 1994;4:41924.
163. Fallahzadeh H. Common duct exploration during laparoscopic cholecystectomy. Am Surg
1997;63:1214.
164. Rhodes M, Nathanson L. Laparoscopic choledochoduodenostomy. Surg Laparosc Endosc
1996;6:31821.
165.
Spohn K, Fux HD, Mller-Kluge M, et al. GallenwegserkrankungenChirurgische Therapie
und intraoperative Diagnostik. Therapiewoche 1975;25:103347.
166. Way LW, Admirand WH, Dunphy JE. Management of choledocholithiasis. Ann Surg
1972;176:34759.
167. Bradfield H, Granke D. Surgical clip as a nidus for a common bile duct stone: Radiographic
demonstration. Abdom Imaging 1997;22:2934.
168. Martinez J, Combs W, Brady PG. Surgical clips as a nidus for biliary stone formation:
Diagnosis and therapy. Am J Gastroenterol 1995;90:15214.
169. Rizzo J, Tripodi J, Gold B, Opper F. Surgical clips as a nidus for stone formation in the
common bile duct. J Clin Gastroenterol 1995;21:16971.
170. Shibata S, Okumichi T, Kimura A, et al. A case of choledocholithiasis with an endoclip nidus,
6 months after laparoscopic cholecystectomy. Surg Endosc 1996;10:10978.
171. Chung SCS, Leung JWC, Leong HT, Li AKC. Mechanical lithotripsy of large common duct
stones using a basket. Br J Surg 1991;78:144850.
172. Siegel JH, Ben-Zvi JS, Pullano WE. Mechanical lithotripsy of common duct stones.
173. Van Dam J, Sivak MV Jr. Mechanical lithotripsy of large common bile duct stones. Cleve Clin
J Med 1993;60:3842.
174. Schneider MU, Matek W, Bauer R, Domschke W. Mechanical lithotripsy of bile duct stones in
209 patientsEffect of technical advances. Endoscopy 1988;20:24853.
175. Siegel JH, Ben-Zvi JS, Pullano WE. Endoscopic electrohydraulic lithotripsy. Gastrointest
Endosc 1990;36:1346.
176. Bonnel DH, Liguory CE, Cornud FE, Lefebvre JFP. Common bile duct and intrahepatic
stones: Results of transhepatic electrohydraulic lithotripsy in 50 patients. Radiology
1991;180:3458.
177. Prat F, Fritsch J, Choury AD, et al. Laser lithotripsy of difficult biliary stones. Gastrointest
Endosc 1994;40:2905.
178. Brambs HJ, Duda SH, Rieber A, et al. Treatment of bile duct stones: Value of laser lithotripsy
delivered via percutaneous endoscopy. Eur Radiol 1996;6:73440.
179. Sander R, Poesl M, Frank F. An Nd:YAG laser with a water guided laser beam. Gastrointest
Endosc 1988;34:3369.
180. Neuhaus H, Hoffmann W, Gottlieb K, Classen M. Endoscopic lithotripsy of bile duct stones
using a new laser with automatic stone recognition. Gastrointest Endosc 1994;40:70815.
181. Schreiber F, Gurakuqi GC, Trauner M. Endoscopic intracorporeal laser lithotripsy of difficult
common bile duct stones with a stone-recognition pulsed dye laser system. Gastrointest
Endosc 1995;42:4169.
182. Binmoeller KF, Bruckner M, Thonke F, Soehendra N. Treatment of difficult bile duct stones
using mechanical, electrohydraulic and extracorporeal shock wave lithotripsy. Endoscopy
1993;25:2016.
183. Navicharern P, Rhodes M, Flook D, Lawrie B. Endoscopic retrograde
cholangiopancreatography (ERCP) and stent placement in the management of large common
bile duct stones. Aust N Z J Surg 1994;64:8402.
184. Law NM, Lim CC, Yap CK, et al. Endoscopic stenting in the management of biliary stones.
Singapore Med J 1996;37:4758.
185. Kubota Y, Takaoka M, Fujimura K, et al. Endoscopic endoprosthesis for large stones in the
common bile duct. Intern Med 1994;33:597601.
186. Peters R, Macmathuna P, Lombard M, et al. Management of common bile duct stones with a
biliary endoprosthesis. Report on 40 cases. Gut 1992;33:14125.
187. Maxton DG, Tweedle DE, Martin DF. Retained common bile duct stones after endoscopic
sphincterotomy: Temporary and longterm treatment with biliary stenting. Gut 1995;36:4469.
188. Bergman JJ, Rauws EA, Tijssen JG, et al. Biliary endoprostheses in elderly patients with
endoscopically irretrievable common bile duct stones: Report on 117 patients. Gastrointest
Endosc 1995;42:195201.
189. Baron TH, Schroeder PL, Schwartzberg MS, Carabasi MH. Resolution of Mirizzi's syndrome
using endoscopic therapy. Gastrointest Endosc 1996;44:3435.
190. England RE, Martin DF. Endoscopic management of Mirizzi's syndrome. Gut 1997;40:2726.
191. Burmeister W, Wurbs D, Hagenmuller F, Classen M. Langzeituntersuchungen nach
endoskopischer Papillotomie. Z Gastroenterol 1980;18:52731.
192.
Bhmig HJ, Zeidler G, Tuchmann A. Sphincterotomy for benign papillary stenosis. In Classen
M, Geenen J, Kawai K (eds). The Papilla Vateri and Its Diseases. Baden-Baden, Kln, New
York: Witzstrock, 1979;p 103.
193. Seifert E, Gail K, Weismuller J. Langzeitresultate nach endoskopischer Sphinkterotomie.
194. Nagase M, Setoyama M, Hikasa Y. Recurrent common duct stones, with special reference to
primary common duct stones. Gastroenterol Jpn 1978;13:2906.
195. Bergman JJ, van der Mey S, Rauws EA, et al. Long-term follow-up after endoscopic
sphincterotomy for bile duct stones in patients younger than 60 years of age. Gastrointest
Endosc 1996;44:6439.
196. Linos DA, Beard CM, O'Fallon WM, et al. Cholecystectomy and carcinoma of the colon.
Lancet 1981;2:37981.
197. Peters H, Keimes AM. Die Cholezystektomie als pradisponierender Faktor in der Genese des
Kolorektalen Karzinoms?. Dtsch Med Wochenschr 1979;104:15813.
198. Sauerbruch T, Stellaard F, Paumgartner G. Effect of endoscopic sphincterotomy on bile acid
pool size and bile lipid composition in man. Digestion 1983;27:8792.
199. Stellaard F, Sauerbruch T, Brunholzl C, et al. Bile acid pattern and cholesterol saturation of
bile after cholecystectomy and endoscopic sphincterotomy. Digestion 1983;26:1538.
200.
Kurtz W, Leuschner U, Schneider S, et al. Diurnal rhythm of biliary lithogenicity persists after
cholecystectomy and papillotomy [Abstract]. Gastroenterology 1982;82:1107.
201. Gandhi A, Maxwell AJ, Wells S, Hobbiss JH. Gallstone ileus following endoscopic
sphincterotomy. Br J Hosp Med 1995;54:22930.
202. Manbeck MA, Chang AC, McCracken JD, Chen YK. Cystic duct perforation as a complication
of endoscopic stone extraction. Am J Gastroenterol 1996;91:5924.
203. Schutz SM, Chinea C, Friedrichs P. Successful endoscopic removal of a severed, impacted
Dormia basket. Am J Gastroenterol 1997;92:67981.
204. Cappell MS. Endoscopic retrograde cholangiopancreatography with endoscopic
sphincterotomy for symptomatic choledocholithiasis after recent myocardial infarction. Am J
205. Friedman RL, Friedman IH. Acute cholecystitis with calculus biliary duct obstruction in the
gravid patient. Management by ERCP, papillotomy, stone extraction, and laparoscopic
cholecystectomy. Surg Endosc 1995;9:9103.
206. Elfant AB, Bourke MJ, Alhalel R, et al. A prospective study of the safety of endoscopic therapy
for choledocholithiasis in an outpatient population. Am J Gastroenterol 1996;91:1499502.
207. Henne-Bruns D, Kremer B, Thonke F, et al. "Endoscopy friendly" resection technique of
choledochal cysts. Endoscopy 1993;25:1768.
208.
Glenn F, Moody FG. Acute obstructive suppurative cholangitis. Surg Gynecol Obstet
1961;113:26573.
209.
Ostermiller W Jr, Thompson RJ Jr, Carter R, Hinshaw DB. Acute obstructive cholangitis. Arch
Surg 1965;90:3925.
210. Wadell GF. Acute obstructive cholangitis. Scot Med J 1966;11:13742.
211. Haupert AP, Carey LC, Evans WE, Ellison HE. Acute suppurative cholangitis. Experience with
15 consecutive cases. Arch Surg 1967;94:4608.
212. Dow RW, Lindenauer SM. Acute obstructive suppurative cholangitis. Ann Surg
1969;169:2726.
213. Hinchey EJ, Couper CE. Acute obstructive suppurative cholangitis. Am J Surg
1969;117:628.
214. Andrew DJ, Johnson SE. Acute suppurative cholangitis: A medical and surgical emergency.
215. Welch JP, Donaldson GA. The urgency of diagnosis and surgical treatment of acute
suppurative cholangitis. Am J Surg 1976;131:52732.

216. American College of Surgeons, Ohio Chapter. 28,621 Cholecystectomies in Ohio. Results of
a survey in Ohio hospitals by the Gallbladder Survey Committee. Am J Surg 1970;119:7147.
217.
Chigot JP. Le risque operatoire dans la lithiase biliare. A propos de 5433 interventions. Sem
Hop Paris 1981;57:13119.
218.
Hess W. Nachoperationen an den Gallenwegen. Stuttgart: Enke. 1977.
219. Way LW, Motson RW. Dissolution of retained common duct stones. Adv Surg
1976;10:99119.
220.
Classen M, Ossenberg FW. Indications for endoscopic papillotomy. In Classen M, Geenen J,
Kawai K (eds). The Papilla Vateri and Its Diseases. Baden-Baden, Kln, New York:
Witzstrock, 1979;p 143.
Chapter 63 Cystic Disorders of the Bile Ducts

(6496)
DAVID S. BARNES, M.D.
Cysts can occur at any level of the biliary tract from the intrahepatic bile ducts to the intraduodenal
portion of the bile duct. Commonly called choledochal cysts or cystic disorders of the bile ducts, this
condition was first described by Vater in 1723.1(6497) To date, about 3000 cases have been reported
in the medical literature. In the United States, the incidence is about 1 in 13,000 hospital admissions,
but the incidence is higher in Japan, the country from which the majority of cases have been
reported.2(6498) The overall incidence in whites is believed to be less than 1 in 200,000.3(6499)
Pathogenesis
The cause of choledochal cysts is unclear, and numerous theories abound as to how they form. In
1959, Alonso-Lej et al.4(6500) stated that, "There are very few subjects in medicine submitted to more
speculation than the etiology of this condition." Cysts are generally believed to be congenital because
they have been found in stillborn infants and in children.5(6501) However, the clinical presentation of
choledochal cysts can be as late as the fifth decade of life, implying an acquired rather than a
congenital condition. Yotuyanagi6(6502) first proposed that during the early stages of embryologic
development, when the developing bile ducts are solid, an uneven proliferation of the duct epithelial
cells results in subsequent bile duct dilation in the area of more active cell proliferation.
Babbitt et al.7,8(6503) were the first to note an abnormality of the junction of the pancreatic and bile
ducts in which the bile duct inserts into the pancreatic duct at a long distance from the ampulla of
Vater, thereby forming a long common channel (Figure 631). They proposed that the long common
channel and a consequent lack of an adequate bile duct sphincter, together with the naturally occurring
higher pressure in the pancreatic duct compared with that in the bile system, result in chronic reflux of
pancreatic secretions into the bile duct. The onset of symptoms has been attributed to the
development of protein plugs in relation to the long common channel. In the study of Kaneko et
al.,9(6504) radiolucent filling defects were demonstrated by cholangiography in 22 of 55 (40%)
symptomatic patients with choledochal cysts. For the diagnosis of anomalous pancreatobiliary junction,
the length of the common channel proximal to the duodenal wall should be greater than 6 mm, or the
common channel should have a length of more than 15 mm.10,11(6505) Numerous reports have
noted the presence of this anomalous connection in patients with choledochal cysts. Anomalous
pancreatobiliary junction in the absence of choledochal cyst has also been described in symptomatic
pediatric patients.12(6506)
(6507)Figure 631. A, Schematic drawing of choledochal cyst shows the long common
channel. B, Retrograde cholangiogram of a Type I choledochal cyst. C, Close-up view of
anomalous junction of the bile and pancreatic ducts. Note the long common channel. (B and C,
From Thatcher BS, Sivak MV Jr, Hermann RE, Esselstyn CB Jr. ERCP in evaluation and
diagnosis of choledochal cyst: Report of five cases. Gastrointest Endosc 1986; 32:2731.)
Amylase levels in the bile within the choledochal cysts have been measured and are frequently but not
invariably elevated.13,14(6508) This lends credence to the hypothesis that reflux of pancreatic juice
into the choledochus plays a role in the formation of choledochal cysts. Davenport et al.15(6509)
studied biliary amylase concentrations in 55 children with choledochal cyst. Elevated amylase levels
were encountered significantly more often in older children who presented with pancreatitis compared
with younger children whose initial manifestation was painless jaundice. Furthermore, elevated levels
of the enzyme were not found in 5 infants in whom choledochal cyst was diagnosed antenatally by
ultrasonography. The consequences of this chronic reflux of pancreatic juice into the bile duct are
believed to be recurrent episodes of cholangitis, thickening of the bile duct wall, and eventual cystic
dilation of the proximal bile duct, along with inflammation and stenosis of the distal common bile duct
(Figure 632). Other authors have confirmed the association between this pancreaticoductal anomaly
and cystic dilation of the bile duct; they coexist in from 47.5 to 97% of cases in large
series.10,13,16,17(6510)
(6511)Figure 632. Retrograde opacification of a choledochal cyst. A stricture (arrow) of the

distal bile duct is seen at the anomalous connection with the pancreatic duct. (From Thatcher
BS, Sivak MV Jr, Hermann RE, Esselstyn CB Jr. ERCP in evaluation and diagnosis of
choledochal cyst: Report of five cases. Gastrointest Endosc 1986; 32:2731.)
Embryologic studies of fetal tissue have shown that before 8 weeks' gestation, the common bile duct
and ventral pancreatic duct form at or outside of the muscular wall of the duodenum.18(6512) As the
fetus develops, the pancreatic duct-common bile duct junction moves toward the duodenal lumen and
continues to migrate even after birth to form a common channel about 5 mm long. Formation of the
anomalous pancreatic duct-common bile duct junction can be seen as the consequence of arrested
migration of the choledochopancreaticoductal junction whereby the junction remains outside the
duodenal muscular wall.13,18(6513)
Intraoperative biliary manometry of the sphincter of Oddi and the choledochopancreaticoductal junction
has been performed before and after gastrin stimulation.14(6514) There is a high-pressure zone at the
sphincter of Oddi, which increases significantly with gastrin infusion. There is no high-pressure zone in
the common channel or choledochopancreaticoductal junction, even after gastrin infusion. This is
further evidence that free reflux of pancreatic juice into the biliary tree can occur, that this is not under
sphincter control, and that reflux may therefore play a role in the pathophysiology of choledochal cyst
formation.
Other studies of the pathogenesis of choledochal cysts have focused on the neuronal properties of the
cyst itself. Kusunoki et al.19(6515) studied the contractile properties of strips of bile duct taken from the
dilated and narrowed portions of choledochal cysts and compared them with the contractile property of
strips from normal bile ducts. There were low concentrations of acetylcholine and nicotine-induced
contractions in the normal strips and the strips from the dilated portion of the cyst but not from the
strips from the narrowed segment. This suggests reduced postganglionic cholinergic activity in the
narrowed part of the cyst. Similar experiments with gamma-aminobutyric acid (GABA) also suggested
that there are GABA receptors on the postganglionic cholinergic neurons in the normal bile duct and
dilated portion of the choledochal cyst that are absent in the narrow segment of the cyst. In another
study, Kusunoki et al.20(6516) counted ganglion cell bodies in five patients with choledochal cysts and
five with normal bile ducts and found a significant reduction in the distal narrow segment of duct in the
patients with choledochal cysts. These structural and functional experimental studies suggest that
postcholinergic cell autonomic dysfunction occurs in the narrow portion of choledochal cysts. Whether
this is a primary cause of cyst formation or a secondary phenomenon is not clear at present.
Two experimental animal models of choledochal cysts have been developed. The first involves
curettage of the bile duct in dogs with subsequent ligation of the main duodenal papilla, thereby
causing biliary obstruction.21(6517) The resultant choledochal cysts were grossly and histologically
similar to cysts seen in humans. These results suggest that both structural weakness of the biliary wall
and high pressure within the bile duct are necessary for cyst formation. A second and perhaps more
physiologic method involves surgical creation of a pancreaticocholecystostomy and results in drainage
of pancreatic juice into the common bile duct, thereby simulating the anomalous
choledochopancreaticoductal junction found in the majority of cases of choledochal cysts in
humans.22(6518) This surgical alteration leads to inflammation and loss of structural integrity of the
bile duct, stenosis of the lower portion of the duct, and cystic dilation of the proximal duct. In a study
using peroxidase as a tracer for permeability, Kato et al.22(6519) found that exposure of the bile duct
to pancreatic enzymes resulted in disruption of the intracellular junctions within the epithelium, leading
to leakage of pancreatic enzymes into the submucosal layer and further injury.
Cystic dilation of the intraduodenal portion of the bile duct (choledochocele) is generally grouped with
the other cystic disorders of the bile duct, but its pathogenesis probably differs from that of the
choledochal cyst. The anomalous pancreaticobiliary junction does not occur in association with
choledochocele. Based on autopsy studies, Sterling23(6520) hypothesized that cystic dilation of the
intraduodenal portion of the bile duct (choledochocele) is an acquired evagination of the mucosa of the
terminal common bile duct into the duodenum caused by an obstruction to bile outflow resulting from
an impacted stone, fibrosis, or papillitis. More recently, Venu et al.24(6521) and Kagiyama et
al.25(6522) suggested that in some patients abnormal sphincter of Oddi motility could be a contributing
factor in the pathogenesis of choledochocele.
Classification
Choledochal cysts were first classified by Alonso-Lej et al.4(6523) in 1959 in an attempt to bring order
to an otherwise confusing array of descriptive terms. This classification included three types of cysts:
(1) congenital cystic dilation of the common bile duct, (2) diverticula of the common bile duct, and (3)
choledochocele. This classification scheme was expanded by Todani et al.26(6524) who included
intrahepatic cysts and further subdivided extrahepatic cysts. The classification system of Todani et
al.26(6525) is the one currently used most frequently in the medical literature (Figure 633).
(6526)Figure 633. Classification of choledochal cysts. (Reprinted by permission from

Todani T, Watanabe Y, Narusue M, et al. Congenital bile duct cysts: Classification, operative
procedures, and review of thirty-seven cases including cancer arising in a choledochal cyst.
Am J Surg, Vol 134, pp 2639. Copyright 1977 by Excerpta Medica Inc.)
The Type I cyst is the most common and accounts for 80 to 90% of cases.2,27(6527) It is subclassified
into three types: Type IA, cystic dilation of the entire extrahepatic bile duct; Type IB, focal segmental
dilation of only the common bile duct; and Type IC, diffuse fusiform dilation of the extrahepatic biliary
tree (Figure 634). Type II, which accounts for only about 2% of cases, is a diverticulum of the bile
duct. Type III, choledochocele, is a cystic dilation of the intraduodenal portion of the bile duct and
accounts for less than 5% of cases of choledochal cysts (Figure 635). However, Type III cysts have
been found to be relatively more common in some endoscopic retrograde cholangiopancreatography
(ERCP) series.28(6528) Type IV is subdivided into Type IVA, multiple cysts of the intrahepatic and
extrahepatic ducts, and Type IVB, multiple cysts of the extrahepatic bile duct only. Type IV cysts
account for 15 to 20% of cases. Type V consists of intrahepatic bile duct cysts, either single or multiple.
Caroli's disease and Type V choledochal cysts are said to be synonymous, but patients with Caroli's
disease can also have Type IVA cysts. Caroli's disease is not a single entity and includes patients with
choledochal cysts associated with renal cystic disease and those with cysts together with associated
congenital hepatic fibrosis, portal hypertension, and cirrhosis.29(6529) A minority (20%) of patients
with Caroli's disease can have monolobular involvement, usually the left hepatic lobe.30(6530)
(6531)Figure 634. Type IC choledochal cyst. The intrahepatic ducts not seen on this
cholangiogram were normal.
(6532)Figure 635. A and B, Upper gastrointestinal x-ray films show smooth, round filling
defect (arrows) in the descending duodenum that changes shape with motion of the duodenum.
C, Endoscopic view shows a smooth round defect in the region of the main duodenal papilla.
D, Retrograde pancreatocholangiogram, made after cannulation and injection, of a pinpoint
opening on the surface of the lesion shows a Type III choledochal cyst (choledochocele).
(A-D, From the collection of Dr. M. V. Sivak, Jr.)
Clinical Presentation
Signs and symptoms of a choledochal cyst can appear at any age, but 40 to 80% of patients are
younger than 10 years and about 20% are less than 1 year old.31,32(6533) However, there are
numerous instances in which choledochal cysts have been discovered during adult life.33,34(6534)
Clinical symptoms at presentation vary and are often intermittent and nonspecific, but the classic
clinical presentation is the triad of abdominal pain, jaundice, and a palpable abdominal mass. In one
large review of 1433 patients with choledochal cysts, 51% had abdominal pain, 45% were jaundiced,
and 37% had a palpable abdominal mass.31(6535) Only 13% of patients had all three findings. Other
series have confirmed the infrequent occurrence of the classic triad.32,3537(6536) Other
less-common findings at presentation include fever, nausea and vomiting, acholic stool, anorexia,
pruritus, hepatomegaly, splenomegaly, fatigue, back pain, and weight loss.31,32,35(6537)
Cholangiocarcinoma may be responsible for the initial presentation in adult patients.38(6538)
There appears to be a relationship between the configuration of the cyst and the clinical
presentation.13,39,40(6539) Infants less than 1 year of age virtually all have cystic-type choledochal
cysts and tend to present with a palpable abdominal mass and jaundice. Infants do not seem to have
abdominal pain as a presenting symptom, although this is difficult to assess.39(6540) In patients with
cylindrical dilation of the bile duct, abdominal pain and vomiting and fever are virtually always present.
Jaundice may or may not be present, and the palpable mass is absent. Patients older than 1 year with
cystic-type dilation often, but not invariably, present with abdominal pain, jaundice, and palpable
mass.39,40(6541) In a retrospective comparative study of 22 children and 27 adults with choledochal
cysts, Chaudhary et al.41(6542) found the Type I lesion to be most common in children, Type IV cysts
being more frequent in adults. Cholangitis was more frequent in adults; anomalous pancreaticobiliary
junction was encountered exclusively in adults. In a review that compared 11 pediatric and 32 adult
patients, Lipsett et al.38(6543) found that the Type I cyst was the most common in both groups,
followed by Type IVA. In this study, adult patients were more likely to have abdominal pain without a
palpable abdominal mass or jaundice.
In addition to the hyperbilirubinemia found at presentation in icteric patients, those who present with
abdominal pain are often found to have hyperamylasemia. It is unclear whether this represents true
pancreatitis or whether the amylase in the bile may reach the systemic circulation through the denuded
biliary epithelium or the hepatic sinusoids. In one series, no patients with hyperamylasemia had
evidence of acute pancreatitis at the time of surgery. There are, however, numerous case reports of
pancreatitis associated with choledochal cysts.40,4245(6544) In the retrospective review of Swisher
et al.,46(6545) 18 of 32 (56%) adult patients with choledochal cysts had documented episodes of
pancreatitis. Cysts were significantly larger in patients who had pancreatitis. Among the 14 patients
who underwent ERCP, episodic pancreatitis was a clinical feature in all 8 patients with an anomalous
pancreatobiliary junction versus only 2 of 6 patients in whom the junction was normal.
Other less-common presentations of choledochal cysts include cholangitis, hepatic abscess, and cyst
rupture resulting in peritonitis.32,45,47(6546) Intestinal intussusception in a patient with a
choledochocele and hematobilia from a pseudoaneurysm associated with a choledochal cyst have also
been reported.48,49(6547) Gallstones and biliary sludge may be found within a choledochocele (Figure
636).50(6548) Biliary cirrhosis and malignant degeneration are well known sequelae of choledochal
cysts but are rarely presenting features.51(6549)
(6550)Figure 636. Retrograde cholangiogram demonstrates a large choledochal cyst

containing sludge. The previous choledochalcystojejunostomy (short arrow) drains poorly.
The anomalous connection between the biliary and the pancreatic ducts (long arrow) occurs
about 15 to 20 mm from the orifice at the main papilla (From Thatcher BS, Sivak MV Jr,
Hermann RE, Esselstyn CB Jr. ERCP in evaluation and diagnosis of choledochal cyst: Report
of five cases. Gastrointest Endosc 1986; 32:2731.)
Choledochocele can give rise to symptoms, and the diagnosis is often overlooked unless ERCP is
performed. Choledochocele has been reported in association with recurrent obstructive
jaundice52(6551) and chronic relapsing pancreatitis.53(6552)
Diagnosis
The preoperative diagnosis of a choledochal cyst requires a high index of suspicion, as it is a rare
condition in many countries and is often not considered in the differential diagnosis. At one time,
preoperative diagnosis of choledochal cystic disease was believed to affect operative mortality, which
was greater than 50% if the cyst was not diagnosed before surgery versus 23% if correctly diagnosed
preoperatively.43(6553) Operative mortality is negligible in more recent surgical series. However,
preoperative knowledge of the extent of involvement of the biliary tree (intrahepatic or extrahepatic)
and the complexity of the anatomic connection of the pancreatic and bile ducts is important in planning
the type of surgery.5456(6554) Various imaging techniques are available for visualization of the biliary
tree, including noninvasive studies such as ultrasonography, computed tomography (CT), magnetic
resonance imaging (MRI), and nuclear medicine scintigraphy. More invasive tests include
percutaneous transhepatic cholangiography (PTC) and ERCP.
Noninvasive Imaging Studies

A plain x-ray of the abdomen may show a right upper quadrant mass. A barium upper gastrointestinal
x-ray series may reveal anterior and inferior displacement of the antrum and duodenal bulb as well as
effacement of the duodenal mucosa. Ultrasonography is very sensitive for the presence of
abnormalities of the biliary tree including choledochal cyst; these may appear as a large cystic mass in
the porta hepatis.57,58(6555) High-resolution, real-time ultrasonography can provide very detailed
anatomic information about the intrahepatic or extrahepatic ducts, and some investigators suggest that
ultrasonography is sufficient for preoperative imaging in uncomplicated cases.59,60(6556) However,
ultrasonography can be misleading and other investigators recommend that PTC or ERCP be
performed in all cases of suspected choledochal cyst.61,62(6557)
CT has also been useful in the diagnosis of choledochal cysts and choledochoceles.63,64(6558)
Intrahepatic and extrahepatic cystic structures are easily detected. Contrast material layering may be
evident within the cyst, and this confirms that the cyst is in continuity with the biliary tree.65(6559) In
patients with choledochoceles (Type III cysts), CT may also reveal a fluid-filled lesion protruding into
the duodenal lumen. The continuity of the cystic lesion with the biliary tree has been confirmed by CT
after ingestion of an oral contrast agent that initially opacifies the gallbladder; injection of sincalide,
which stimulates gallbladder emptying, results in opacification of the cystic mass (choledochocele) on a
subsequent CT.62(6560) MRI has also been found to be helpful in the diagnosis of choledochal
cysts.66(6561)
Biliary scintigraphy, also useful in the diagnosis of choledochal cysts, has been used mainly to confirm
that cystic lesions seen on ultrasonography or CT are part of the biliary tract and are not other types of
cystic lesions such as simple hepatic cysts, duodenal duplication cysts, pancreatic pseudocysts, or
hepatic artery aneurysms (Figure 637).67,68(6562) By scintigraphy, a choledochal cyst appears as an
area of photon deficiency that fills in late in the course of the study.69,70(6563) Other patterns include
the early appearance of photon activity in dilated extrahepatic ducts and complete nonvisualization of
the cyst owing to high-grade biliary obstruction.67(6564)
(6565)Figure 637. Biliary scintigraphy of a choledochal cyst shows accumulation (arrows)

in the dilated biliary tree.
PTC and ERCP

Although ultrasonography and CT are useful in the diagnosis of choledochal cysts, substantially more
detailed anatomic information can be obtained by cholangiography. Features such as cyst anatomy,
site of biliary origin, extent of intrahepatic and extrahepatic involvement, and the presence of stones
can readily be ascertained by cholangiography.62(6566) PTC can be performed safely and is generally
well tolerated, although there is a small risk of hemorrhage or infection. Biliary stents can be placed
percutaneously and may be of assistance to the surgeon at the time of operative
reconstruction.62(6567) ERCP is also useful in the diagnosis of choledochal cysts and has the added
advantage of visualization of the pancreatic duct.61,7074(6568) Many but not all investigators believe
that cholangiography and pancreatography are both essential in the preoperative
evaluation.5961,73(6569)
ERCP can be performed in infants and children using a pediatric duodenoscope (see Chapter 79:
Endoscopic Retrograde Cholangiopancreatography in Infants and Children).73,7580(6570) Because
the differential diagnosis of neonatal cholestasis includes choledochal cysts, ERCP has proved to be
useful in distinguishing anomalies of the biliary tree such as choledochal cysts or biliary atresia from
neonatal hepatitis, thus avoiding surgery if the biliary tree is normal.
Complications
Numerous complications of choledochal cysts have been reported, including carcinoma of the bile duct
and gallbladder, cyst rupture, cholangitis, pancreatitis, formation of stones within the cyst, and
secondary biliary cirrhosis.
The most ominous complication of a choledochal cyst is carcinoma of the biliary tree. The incidence of
carcinoma ranges from 2 to 14.5%, this being 5 to 35 times greater than the expected incidence of
0.012 to 0.048% in the general population.81,82(6571) The incidence of carcinoma appears to
increase with age. The risk is minimum (0.7%) in patients less than 10 years of age; it increases to
14.3% in patients over age 20 years.83(6572) It has been postulated that carcinoma develops in
choledochal cysts owing to recurrent infection, chronic inflammation, or the mutagenic effects of the
products of bile stasis. Examination of the cyst lining in cysts that have been surgically removed often
reveals proliferative changes in the epithelium, including cellular crowding, poor orientation, argyrophil
and goblet cell metaplasia, and dysplasia.84,85(6573) Epithelial metaplasia is found most often in
older patients (more than 15 years of age), which is consistent with the age-related increase in risk of
adenocarcinoma in choledochal cysts.84(6574) It has been hypothesized that these dysplastic and
metaplastic epithelial changes are the precursors of cancer. Analysis of bile from a patient who
underwent surgical excision of a choledochal cyst revealed a predominance of unconjugated
secondary bile acids (deoxycholic and lithocholic acid).81(6575) Secondary bile acids, probably
generated by bacterial overgrowth, are believed to be mutagenic and to cause metaplasia and
carcinoma in the cyst wall. However, Chijiiwa et al.86(6576) found no differences in the relative
composition or absolute concentration of secondary bile acids in bile aspirated from choledochal cysts
in 11 patients compared with concentrations in bile from 7 patients with biliary tract carcinoma and 5
patients with cholecystolithiasis.
There is also an association between the presence of a choledochal cyst and adenocarcinoma of the
gallbladder. This association, however, may not be a direct one but is probably due to the occurrence
in both conditions of the anomalous pancreaticobiliary ductal union (long common channel). This
anomaly is found frequently in cases of gallbladder cancer, being present in 10 to 25% of
patients.11,87,88(6577) Gallstones are usually absent in patients with gallbladder cancer and
anomalous pancreaticobiliary ductal junction.89(6578) It is hypothesized that cancer of the gallbladder
is due to reflux and stasis of pancreatic juice within the gallbladder leading to metaplasia and ultimately
to adenocarcinoma.90,91(6579)
Spontaneous perforation of the bile duct in five children has been described by Ando et al.92(6580)
Initial management was by placement of a T-tube at the site of perforation. Subsequent T-tube
cholangiography demonstrated the presence of an anomalous pancreaticobiliary junction in all cases
as well as obstruction of the long common channel that was believed to be due to the presence of
proteinaceous plugs. These investigators speculated that the spontaneous perforations were due to
distal obstruction that led to sudden increases in pressure within the cysts. Nine instances of
perforation were encountered in the series of Hsu et al.93(6581) of 74 patients ranging in age from 6 to
16 years.
Management
Treatment of choledochal cysts is generally surgical with the possible exception of the choledochocele
(Todani's Type III), which can be managed endoscopically. Medical therapy alone is associated with a
high mortalityup to 96.6% in early series.43,51(6582) Before the late 1960s, a variety of operations
were attempted: external drainage, especially in the critically ill patient; excision of part of the cyst wall
with reconstruction of the bile duct; internal drainage procedures through the gallbladder (e.g.,
cholecystoduodenostomy); and direct internal drainage of the cyst into the duodenum or jejunum
(choledochocystoduodenostomy or choledochocystojejunostomy).94(6583) The morbidity associated

with these procedures was high and included problems such as recurrent cholangitis,
choledocholithiasis, secondary biliary cirrhosis, acute pancreatitis, anastomotic stenosis, and a 20-fold
increase in the incidence of malignancy of the extrahepatic biliary tree when compared with the normal
population.37,9498(6584) Reoperation for complications was frequently necessary. Since the
mid-1970s, cyst excision with hepaticojejunostomy or hepaticoduodenostomy has become the
preferred treatment for choledochal cysts and is associated with fewer complications.37,95103(6585)
Complications of hepaticoenterostomy have included cholangitis, formation of stones within the
intrahepatic ducts, anastomotic stenosis, and retrograde intussusception after Roux-en-Y
anastomosis.97,104(6586) Among 28 patients followed up by Chijiiwa and Tanaka105(6587) for a
mean of 8 years (range 1 to 18 years) after Roux-en-Y hepaticojejunostomy, 3 patients were found to
have intrahepatic stones. One patient had a Type IVA cyst. An anastomotic stricture was not present in
any patient, but bile was found to be infected in all patients. Reoperation because of the development
of intrahepatic gallstones and recurrent cholangitis was necessary in 10 of 97 patients who underwent
either hepaticoduodenostomy (67) or hepaticojejunostomy (30) in the series of Todani et al.106(6588)
An anastomotic stricture was present in 9 of the 10 reoperated patients.
Patients with Type IVA cysts (cystic dilation of the extrahepatic cysts and dilation of the intrahepatic
ducts) with cylindrical intrahepatic dilation should undergo excision of the extrahepatic cysts, after
which the cylindrical intrahepatic dilation will regress.107,108(6589) The cystic-type intrahepatic cyst in
Type IVA choledochal cysts does not regress after extrahepatic cyst removal; these have been
managed by hemihepatectomy when confined to the left lobe of the liver.108(6590) Although
malignancy has been reported to occur in the intrahepatic cysts,109(6591) the surgical excision of the
extrahepatic biliary tree should eliminate the free reflux of pancreatic juice into the biliary tree caused
by the anomalous choledochopancreaticoductal junction and reduce the risk of malignant
transformation in the intrahepatic cysts.110,111(6592) In the study of Chijiiwa et al.,105(6593) 13
patients with Type IVA cysts who underwent excision of the extrahepatic cyst and hepaticojejunostomy
were followed up for periods that ranged from 2 months to 16 years. Although hepatolithiasis
developed in 2 patients and cholangitis in a third, none of the patients developed cholangiocarcinoma.
Hewitt et al.112(6594) reported three cases in which the diagnosis of choledochal cyst occurred during
pregnancy. Because clinical signs and symptoms were confusing in relation to the physiologic changes
of pregnancy, diagnosis proved to be difficult and was initially overlooked in two patients. Hewitt et
al.112(6595) recommended that surgery be delayed until the patient reaches the second trimester or
until after delivery if possible. However, spontaneous rupture of the cyst occurred in one patient and
resulted in fetal loss and prolonged hospitalization.
Type III choledochal cysts (choledochoceles) occur much less commonly than other type of cysts. In
series of selected patients, however, they can be relatively common. For example, Ladas et
al.28(6596) encountered 15 patients with symptoms due to choledochocele among 1019 referred for
ERCP over a period of 7 years. There were 10 men and 5 women who ranged in age from 47 to 93
years; choledocholithiasis was present in 10 patients. In another review of 9850 patients who
underwent ERCP, Schmidt et al.113(6597) found 10 in whom a choledochocele was demonstrated.
Only 3 had definite symptoms referable to the choledochocele, whereas 4 were asymptomatic.
Choledochoceles can be treated surgically with transduodenal cyst excision and
sphincteroplasty.96,114(6598) Endoscopic sphincterotomy of small choledochoceles has been
performed.24,103,113,115,116(6599) Dohmoto et al.117(6600) reported successful treatment of a
2-year-old girl by endoscopic sphincterotomy and temporary placement of a biliary prosthesis. Ladas et
al.,28(6601) treated 12 patients by endoscopic sphincterotomy; complications included pancreatitis in 1
patient and cholangitis in another. Follow-up ranged from 12 to 97 months (median 19 months).
Among the 14 patients for whom follow-up was complete, 11 (76%) remained asymptomatic whereas
episodes of cholangitis occurred in 2 and 1 patient died as a result of a misdiagnosed ampullary
carcinoma. There were no complications of endoscopic sphincterotomy in the 10 adult patients with
choledochoceles treated by endoscopic sphincterotomy by Martin et al.;118(6602) during follow-up,
which ranged from 2 to 20 months, all patients remained asymptomatic.
There has been no documentation that choledochocele has a long-term predisposition toward bile duct
carcinoma. Sphincterotomy, when possible, is probably adequate therapy.115,119(6603) A case of
associated ampullary carcinoma has been reported.28(6604) One case associated with pancreatic
carcinoma has also been reported.120(6605) However, the relationship if any between the
choledochocele and the malignant tumor is unclear.
Treatment of patients with Type V choledochal cysts (Caroli's disease) can be difficult and depends in
part on the extent of involvement. When the disease is localized, hepatic resection may be suitable. If
there is diffuse involvement, creation of a hepaticojejunostomy that allows for subsequent jejunal tract
choledochoscopy may be beneficial (see Chapter 71: Choledochofiberoscopy and Endoscopic
Lithotripsy). There are a few reports of cases in which various forms of retrograde endoscopic
intervention have been beneficial.121,122(6606)
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
Vater A, Ezler CS. Dissertatio de scirrhis viscerum occasione sectionis Viri tymparite defunte.
Witten burgae 1723;4 Pamphlets, 881:22.
Crittenden S, McKinley M. Choledochal cystClinical features and classification. Am J
Deeg HJ, Rominger JM, Shah AN. Choledochal cyst and pancreatic carcinoma demonstrated
simultaneously by endoscopic retrograde cholangiopancreatography. South Med J
1980;73:16789.
Alonso-Lej F, Rever WB, Pessagno DJ. Collective review: Congenital choledochal cyst, with
report of 2, and an analysis of 94 cases. Int Abstr Surg 1959;108:130.
Powell CS, Sawyers JL, Reynolds VH. Management of adult choledochal cysts. Ann Surg
1981;193:6667.
Yotuyanagi S. Contributions to the aetiology and pathogeny of idiopathic cystic dilatation of the
common bile-duct with report of three cases: A new aetiological theory based on supposed
unequal epithelial proliferation at the stage of the physiological epithelial occlusion of the
primitive choledochus. Jpn J Cancer Res 1936;30:60151.
Babbitt DP. Congenital choledochal cysts: New etiological concept based on anomalous
relationships of the common bile duct and pancreatic bulb. Ann Radiol (Paris)
1969;12:23140.
Babbitt DP, Starshak R, Clemett A. Choledochal cyst: A concept of etiology. AJR
1973;119:5762.
Kaneko K, Ando H, Ito T, et al. Protein plugs cause symptoms in patients with choledochal
cysts. Am J Gastroenterol 1997;92:101821.
Ono J, Sakoda K, Akita H. Surgical aspects of cystic dilatation of the bile ductAn anomalous
junction of the pancreaticobiliary tract in adults. Ann Surg 1982;195:2038.
Kimura K, Ohto M, Saisho H, et al. Association of gallbladder carcinoma and anomalous
pancreaticobiliary ductal union. Gastroenterology 1985;89:125865.
Ando H, Ito T, Nagaya M, et al. Pancreaticobiliary maljunction without choledochal cysts in
infants and children: Clinical features and surgical therapy. J Pediatr Surg 1995;30:165862.
Todani T, Watanabe Y, Fujii T, et al. Anomalous arrangement of the pancreaticobiliary ductal
system in patients with a choledochal cyst. Am J Surg 1984;147:6726.
Iwai N, Yanagihara J, Tokiwa K, et al. Congenital choledochal dilatation with emphasis on
pathophysiology of the biliary tract. Ann Surg 1992;215:2730.
15. Davenport M, Stringer MD, Howard ER. Biliary amylase and congenital choledochal dilatation.
J Pediatr Surg 1995;30:4747.
16. Nagata E, Sakai K, Kinoshita H, Hirohashi K. Choledochal cyst: Complications of anomalous
connection between the choledochus and pancreatic duct and carcinoma of the biliary tract.
World J Surg 1986;10:10210.
17. Sameshima Y, Uchimura M, Muto Y, et al. Coexistent carcinoma in congenital dilatation of the
bile duct and anomalous arrangement of the pancreatico-bile duct. Carcinogenesis of
coexistent gall bladder carcinoma. Cancer 1987;60:188390.
18. Wong KC, Lister J. Human fetal development of the hepatopancreatic duct junctionA
possible explanation of congenital dilatation of the biliary tract. J Pediatr Surg
1981;16:13945.
19. Kusunoki M, Yamamura T, Takahashi T, et al. Choledochal cyst. Its possible autonomic
involvement in the bile duct. Arch Surg 1987;122:9971000.
20. Kusunoki M, Saitoh N, Yamamura T, et al. Choledochal cysts. Oligoganglionosis in the narrow
portion of the choledochus. Arch Surg 1988;123:9846.
21. Kato T, Asakura Y, Kasai M. An attempt to produce choledochal cyst in puppies. J Pediatr
Surg 1974;9:50913.
22. Kato T, Hebiguchi T, Matsuda K, Yoshino H. Action of pancreatic juice on the bile duct:
Pathogenesis of congenital choledochal cyst. J Pediatr Surg 1981;16:14651.
23.
Sterling JA. Diverticula in the terminal portion of the common bile duct. Am J Pathol
1949;25:32535.
24. Venu RP, Geenen JE, Hogan WJ, et al. Role of endoscopic retrograde
cholangiopancreatography in the diagnosis and treatment of choledochocele.
25. Kagiyama S, Okazaki K, Yamamoto Y, Yamamoto Y. Anatomic variants of choledochocele
and manometric measurements of pressure in the cele and the orifice zone. Am J
26. Todani T, Watanabe Y, Narusue M, et al. Congenital bile duct cysts: Classification, operative
procedures, and review of thirty-seven cases including cancer arising in a choledochal cyst.
Am J Surg 1977;134:2639.
27.
Savader SJ, Benenati JF, Venbrux AC, et al. Choledochal cysts: Classification and
cholangiographic appearance. AJR 1991;156:32731.
28. Ladas SD, Katsogridakis I, Tassios P, et al. Choledochocele, an overlooked diagnosis: Report
of 15 cases and review of 56 published reports from 1984 to 1992. Endoscopy
1995;27:2339.
29. Tandon RK, Grewal H, Anand AC, Vashisht S. Caroli's syndrome: A heterogeneous entity. Am
30. Boyle MJ, Doyle GD, McNulty JG. Monolobar Caroli's disease. Am J Gastroenterol
1989;84:143744.
31. Yamaguchi M. Congenital choledochal cyst: Analysis of 1433 patients in the Japanese
literature. Am J Surg 1980;140:6537.
32. Chen W, Chang C, Hung W. Congenital choledochal cyst: With observations on rupture of the
cyst and intrahepatic ductal dilatation. J Pediatr Surg 1973;8:52938.
33. Bose SM, Lobo DN, Singh G, Wig JD. Bile duct cysts: Presentation in adults. Aust N Z J Surg
1993;63:8537.
34. Hewitt PM, Krige JE, Bornman PC, Terblanche J. Choledochal cysts in adults. Br J Surg
1995;82:3825.
35. Sherman P, Kolster E, Davies C, et al. Choledochal cysts: Heterogeneity of clinical
presentation. J Pediatr Gastroenterol Nutr 1986;5:86772.
36. Flanigan PD. Biliary cysts. Ann Surg 1975;182:63543.
37. Chijiiwa K, Koga A. Surgical management and long-term follow-up of patients with
choledochal cysts. Am J Surg 1993;165:23842.
38. Lipsett PA, Pitt HA, Colombani PM, et al. Choledochal cyst disease. A changing pattern of
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
presentation. Ann Surg 1994;220:64452.

Okada A, Nakamura T, Higaki J, et al. Congenital dilatation of the bile duct in 100 instances
and its relationship with anomalous junction [published erratum appears in Surg Gynecol
Obstet 1991; 172:246]. Surg Gynecol Obstet 1990;171:2918.
Todani T, Watanabe Y, Fujii T, et al. Cylindrical dilatation of the choledochus: A special type
of congenital bile duct dilatation. Surgery 1985;98:9649.
Chaudhary A, Dhar P, Sachdev A, et al. Choledochal cystsDifferences in children and
adults. Br J Surg 1996;83:1868.
Agrawal RM, Brodmerkel GJ. Jr. Choledochal cyst presenting as pancreatitis. Am J
Tsardakas E, Robnett AH. Congenital cystic dilatation of the common bile duct. A report of
three cases, analysis of fifty-seven cases and review of the literature. Arch Surg
1956;72:31127.
Altman MS, Halls JM, Douglas AP, Renner IG. Choledochal cyst presenting as acute
pancreatitis: Evaluation with endoscopic retrograde cholangiopancreatography. Am J
Hopkins NF, Benjamin IS, Thompson MH, Williamson RC. Complications of choledochal
cysts in adulthood. Ann R Coll Surg Engl 1990;72:22935.
Swisher SG, Cates JA, Hunt KK, et al. Pancreatitis associated with adult choledochal cysts.
Pancreas 1994;9:6337.
Sanders P, Jackson BT. Rupture of choledochus cyst in pregnancy. BMJ 1969;3:5734.
Ramos A, Castello J, Pinto I. Intestinal intussusception as a presenting feature of
choledochocele. Gastrointest Radiol 1990;15:2114.
Eliscu EH, Weiss GM. Hematobilia due to a pseudoaneurysm complicating a choledochal
cyst. AJR 1988;151:7834.
Tajiri H. Choledochocele-containing stones. Am J Gastroenterol 1996;91:10468.
Robertson JF, Raine PA. Choledochal cyst: A 33-year review. Br J Surg 1988;75:799801.
Marshall JB, Halpin TC. Choledochocele as the cause of recurrent obstructive jaundice in
childhood: Diagnosis by ERCP. Gastrointest Endosc 1982;28:8890.
Greene FL, Brown JJ, Rubinstein P, et al. Choledochocele and recurrent pancreatitis.
Diagnosis and surgical management. Am J Surg 1985;149:3069.
Oweida SW, Ricketts RR. Hepatico-jejuno-duodenostomy reconstruction following excision of
choledochal cysts in children. Am Surg 1989;55:26.
Lopez RR, Pinson CW, Campbell JR, et al. Variation in management based on type of
choledochal cyst. Am J Surg 1991;161:6125.
Todani T, Watanabe Y, Mizuguchi T, et al. Hepaticoduodenostomy at the hepatic hilum after
excision of choledochal cyst. Am J Surg 1981;142:5847.
Kangarloo H, Sarti DA, Sample WF, Amundson G. Ultrasonographic spectrum of choledochal
cysts in children. Pediatr Radiol 1980;9:158.
Reuter K, Raptopoulos VD, Cantelmo N, et al. The diagnosis of a choledochal cyst by
ultrasound. Radiology 1980;136:4378.
Jones SN, Lees WR, Russell RG. Preoperative ultrasound assessment of choledochal cysts.
Report of three cases. Acta Radiol 1989;30:357.
Young W, Blane C, White SJ, Polley TZ. Congenital biliary dilatation: A spectrum of disease
detailed by ultrasound. Br J Radiol 1990;63:3336.
Shemesh E, Czerniak A, Klein E, Avigad I. The role of endoscopic retrograde
cholangiopancreatography in the diagnosis and treatment of adult choledochal cyst. Surg
Savader SJ, Venbrux AC, Benenati JF, et al. Choledochal cysts: Role of noninvasive imaging,
percutaneous transhepatic cholangiography, and percutaneous biliary drainage in diagnosis
and treatment. J Vasc Interv Radiol 1991;2:37985.
Cory DA, Don S, West KW. CT cholangiography of a choledochocele. Pediatr Radiol
1990;21:734.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
Araki T, Itai Y, Tasaka A. CT of choledochal cyst. AJR 1980;135:72934.

Wolfe B. Contrast layering in a choledochal cyst: A new CT observation. J Can Radiol Assoc
1988;39:513.
Alexander MC, Haaga JR. MR imaging of a choledochal cyst. J Comput Assist Tomogr
1985;9:3579.
Camponovo E, Buck J, Drane W. Scintigraphic features of choledochal cyst. J Nucl Med
1989;30:6228.
Papanicolaou N, Abramson SJ, Teele RL, et al. Specific preoperative diagnosis of
choledochal cysts by combined sonography and hepatobiliary scintigraphy. Ann Radiol
1985;28:27682.
Huang M, Liaw Y. Intravenous cholescintigraphy using Tc-99m-labeled agents in the
diagnosis of choledochal cyst. J Nucl Med 1982;23:1136.
Paramsothy M, Somasundram K. Technetium 99m-diethyl-IDA hepatobiliary scintigraphy in
the pre-operative diagnosis of choledochal cyst. Br J Radiol 1981;54:11047.
Saowaros V. ERCP in diagnosis of adult choledochal cyst: Report of five cases. J Med Assoc
Thai 1990;73:4248.
Thatcher BS, Sivak MV Jr, Hermann RE, Esselstyn CB Jr. ERCP in evaluation and diagnosis
of choledochal cyst: Report of five cases. Gastrointest Endosc 1986;32:2731.
Hagen G, Kolmannskog F. Radiologic approach to bile duct cysts in adults. Acta Radiol
1992;33:2404.
Sharma AK, Wakhlu A, Sharma SS. The role of endoscopic retrograde
cholangiopancreatography in the management of choledochal cysts in children. J Pediatr Surg
1995;30:657.
Heyman MB, Shapiro HA, Thaler MM. Endoscopic retrograde cholangiography in the
diagnosis of biliary malformations in infants. Gastrointest Endosc 1988;34:44953.
Mauer K, Waye JD. A new pediatric duodenoscope: Successful cannulation without a cannula
elevator. Gastrointest Endosc 1989;35:4379.
Guelrud M, Jaen D, Torres P, et al. Endoscopic cholangiopancreatography in the infant:
Evaluation of a new prototype pediatric duodenoscope. Gastrointest Endosc 1987;33:48.
Wilkinson ML, Mieli-Vergani G, Ball C, et al. Endoscopic retrograde
cholangiopancreatography in infantile cholestasis. Arch Dis Child 1991;66:1213.
Putnam PE, Kocoshis SA, Orenstein SR, Schade RR. Pediatric endoscopic retrograde
cholangiopancreatography. Am J Gastroenterol 1991;86:82430.
Allendorph M, Werlin SL, Geenen JE, et al. Endoscopic retrograde cholangiopancreatography
in children. J Pediatr 1987;110:20611.
Reveille R, Stiegmann G, Everson G. Increased secondary bile acids in a choledochal cyst.
Kagawa Y, Kashihara S, Kuramoto S, Maetani S. Carcinoma arising in a congenitally dilated
biliary tract. Report of a case and review of the literature. Gastroenterology 1978;74:128694.
Voyles CR, Smadja C, Shands WC, Blumgart LH. Carcinoma in choledochal cysts.
Age-related incidence. Arch Surg 1983;118:9868.
Komi N, Tamura T, Tsuge S, et al. Relation of patient age to premalignant alterations in
choledochal cyst epithelium: Histochemical and immunohistochemical studies. J Pediatr Surg
1986;21:4303.
Rossi RL, Silverman ML, Braasch JW, et al. Carcinomas arising in cystic conditions of the bile
ducts. A clinical and pathologic study. Ann Surg 1987;205:37784.
Chijiiwa K, Nagai E, Makino I, Shimada K. Are secondary bile acids in choledochal cysts
important as a risk factor in biliary tract carcinoma?. Aust N Z J Surg 1993;63:10912.
Komi N, Takehara H, Kunitomo K. Choledochal cyst: Anomalous arrangement of the
pancreaticobiliary ductal system and biliary malignancy. J Gastroenterol Hepatol
1989;4:6374.
Chijiiwa K, Tanaka M, Nakayama F. Adenocarcinoma of the gallbladder associated with
anomalous pancreaticobiliary ductal junction. Am Surg 1993;59:4304.
89. Chijiiwa K, Kimura H, Tanaka M. Malignant potential of the gallbladder in patients with
anomalous pancreaticobiliary ductal junction. The difference in risk between patients with and
without choledochal cyst. Int Surg 1995;80:614.
90. Kinoshita H, Nagata E, Hirohashi K, et al. Carcinoma of the gallbladder with an anomalous
connection between the choledochus and the pancreatic duct. Report of 10 cases and review
of the literature in Japan. Cancer 1984;54:7629.
91. Ozmen V, Martin PC, Igci A, et al. Adenocarcinoma of the gallbladder associated with
congenital choledochal cyst and anomalous pancreaticobiliary ductal junction. Case report.
Eur J Surg 1991;157:54951.
92. Ando H, Ito T, Watanabe Y, et al. Spontaneous perforation of choledochal cyst. J Am Coll
Surg 1995;181:1258.
93. Hsu YC, Lin TY, Lee HC, et al. Congenital biliary tract dilatation in infancy and childhood74
cases experience. Chang Keng I Hsueh 1994;17:3018.
94. Trout H, Longmire W. Long-term follow-up study of patients with congenital cystic dilatation of
the common bile duct. Am J Surg 1971;121:6886.
95. Lilly JR. The surgical treatment of choledochal cysts. Surg Gynecol Obstet 1979;149:3642.
96. Nagorney D, McIlrath D, Adson M. Choledochal cysts in adults: Clinical management. Surgery
1984;96:65663.
97. Chijiiwa K. Hazard and outcome of retreated choledochal cyst patients. Int Surg
1993;78:2047.
98. Stain SC, Guthrie CR, Yellin AE, Donovan AJ. Choledochal cyst in the adult. Ann Surg
1995;222:12833.
99. Kasai M, Asakura Y, Taira Y. Surgical treatment of choledochal cyst. Ann Surg
1970;172:84451.
100. Joseph VT. Surgical techniques and long-term results in the treatment of choledochal cyst. J
Pediatr Surg 1990;25:7827.
101. Tan KC, Howard ER. Choledochal cyst: A 14-year surgical experience with 36 patients. Br J
Surg 1988;75:8925.
102. Todani T, Watanabe Y, Taki A, et al. Reoperation for congenital choledochal cyst. Ann Surg
1988;207:1427.
103. Benhidjeb T, Munster B, Ridwelski K, et al. Cystic dilatation of the common bile duct: Surgical
treatment and long-term results. Br J Surg 1994;81:4336.
104. Chen CC, Huang SC, Huang SC, Shieh CS. Retrograde intussusception as a complication of
Roux-en-Y anastomosis for choledochal cystReport of one case. Acta Paediatr Sin
1994;35:54651.
105. Chijiiwa K, Tanaka M. Late complications after excisional operation in patients with
choledochal cyst. J Am Coll Surg 1994;179:13944.
106. Todani T, Watanabe Y, Urushihara N, et al. Biliary complications after excisional procedure
for choledochal cyst. J Pediatr Surg 1995;30:47881.
107. Thambi Dorai CR, Visvanathan R, McAll GLG. Type IVa choledochal cysts: Surgical
management and literature review. Aust N Z J Surg 1991;61:50510.
108. Todani T, Watanabe Y, Fujii T, et al. Congenital choledochal cyst with intrahepatic
involvement. Arch Surg 1984;119:103843.
109. Tajiri K, Takenawa H, Yamaoka K, et al. Choledochal cyst with adenocarcinoma in the
cystically dilated intrahepatic bile duct. Abdom Imaging 1997;22:1903.
110. Okada A, Nakamura T, Okumura K, et al. Surgical treatment of congenital dilatation of bile
duct (choledochal cyst) with technical considerations. Surgery 1987;101:23843.
111. Deziel DJ, Rossi RL, Munson JL, et al. Management of bile duct cysts in adults. Arch Surg
1986;121:4105.
112. Hewitt PM, Krige JE, Bornman PC, Terblanche J. Choledochal cyst in pregnancy: A
therapeutic dilemma. J Am Coll Surg 1995;181:23740.
113. Schmidt HG, Bauer J, Wiessner V, Schonekas H. Endoscopic aspects of choledochoceles.
114. Sarris GE, Tsang D. Choledochocele: Case report, literature review, and a proposed
classification. Surgery 1989;105:40814.
115. Gerritsen J, Janssens AR, Kroon M. Choledochocele: Treatment by endoscopic
sphincterotomy. Br J Surg 1988;75:4956.
116. Siegel JH, Harding GT, Chateau F. Endoscopic incision of choledochal cysts
(choledochocele). Endoscopy 1981;13:2002.
117. Dohmoto M, Kamiya T, Hunerbein M, et al. Endoscopic treatment of a choledochocele in a
2-year-old child. Surg Endosc 1996;10:10168.
118. Martin RF, Biber BP, Bosco JJ, Howell DA. Symptomatic choledochoceles in adults.
Endoscopic retrograde cholangiopancreatography recognition and management. Arch Surg
1992;127:5368; discussion 5389.
119. Donald JJ, Coral A, Lees WR. Choledochocele complicated by carcinoma. Clin Radiol
1989;40:1013.
120. Silas DN. Association of a choledochocele and pancreatic carcinoma. Gastrointest Endosc
1994;40:946.
121. Ciambotti GF, Ravi J, Abrol RP, Arya V. Right-sided monolobar Caroli's disease with
intrahepatic stones: Nonsurgical management with ERCP. Gastrointest Endosc
1994;40:7614.
122. Gold DM, Stark B, Pettei MJ, Levine JJ. Successful use of an internal biliary stent in Caroli's
Chapter 64 Diagnosis and Management of Malignant Biliary

Obstruction
(6607)
(6608)
PETER B. COTTON, M.D., F.R.C.P.
Biliary obstruction may be caused by tumors arising in the bile ducts, adjacent organs (e.g., pancreas),
or metastases. Tumors are traditionally grouped by the level of obstruction, that is, periampullary
(arising in and around the main duodenal papilla), pancreatic head, bile duct, gallbladder, and liver
hilum. This wide variety of lesions and patients poses complex management questions. The task can
be separated into:
1. Diagnosis of the presence of a tumor (as opposed to a benign cause of obstruction such as
choledocholithiasis)
2. Confirmation of the malignant nature of the lesion by histopathologic diagnosis
3. Assessment of the indications for attempted surgical cure (resectability and operability)
4. Palliation of biliary obstruction
5. Selection of adjuvant therapies, including chemo-therapy, radiotherapy, and supportive measures
Initial Diagnosis and Differential Diagnosis

The clinical features of a patient's illness are helpful in the process of differential diagnosis. Those
patients with malignant obstruction usually present with progressive cholestasis and pruritus, but
without fever (in the absence of prior intervention). However, patients with tumors of the main duodenal
papilla may have intermittent attacks of jaundice and fever (resembling stones), presumably due to
tumor necrosis; tumors arising in the papilla may also manifest as anemia or overt bleeding (see
Chapter 69: Tumors of the Main Duodenal Papilla). Drug-induced cholestasis is common in the aging
population, whereas chronic hepatitis and primary liver diseases (e.g., primary biliary cirrhosis) are
more common in younger age groups. The patient with obstruction distal to the cystic duct junction
(i.e., involving the common bile duct) may have a palpable gallbladder. Determinations of standard liver
function tests will typically show dominant elevation of bilirubin and alkaline phosphatase, with less
disturbance of transaminases. Currently available tumor markers are of little clinical value in this
context,1(6609) although in one study the combination of serum levels of carcinoembryonic antigen
and carbohydrate antigen 199 was found to be useful in the diagnosis of cholangiocarcinoma in
patients with primary sclerosing cholangitis.2(6610)
Obstruction of the biliary tract is usually characterized by bile duct dilation that can be easily
demonstrated by ultrasound scanning or computed tomography (CT). However, recognition of
obstruction may be delayed in tumors of the hilum because the extrahepatic biliary tree is of normal
diameter. Patients with tumors arising in the main duodenal papilla usually have dilation of both
pancreatic and biliary ducts but no detectable mass by CT or ultrasound. When a mass is present, this
type of "double-duct" sign is pathognomonic of a lesion in the pancreatic head (Figure 641). Tumors
arising at any level in the bile duct are often not detectable by noninvasive imaging; their presence can
be inferred by the site and nature of the obstruction.
(6611)Figure 641. Computed tomography (CT) scan shows a tumor in the head of the
pancreas with dilation of both pancreatic and biliary ductal systems.
More detailed views of the area of stricturing or obstruction can be obtained by invasive
cholangiogra-phy, either endoscopic retrograde cholangiopancretography (ERCP) (Figure 642) or
percutaneous transhepatic cholangiography (PTC) (Figure 643). Where both methods are available,
ERCP is usually preferred over PTC because it is more comprehensive. With ERCP, a tumor of the
duodenum and papilla is easily diagnosed (with biopsy or cytologic confirmation), the status of both the
biliary and the pancreatic ductal systems are documented, other diseases (e.g., impacted stones) are
excluded, intraductal specimens can be obtained for cytologic evaluation, and a stent can be inserted
for temporary drainage or permanent palliation.3(6612) Cholangiography should be possible in more
than 90% of cases, but success rates depend on individual expertise and on the level of the lesion.
Technically, the most difficult are tumors that involve the papilla and obscure the bile duct orifice. PTC
has particular value when ERCP is unavailable or unsuccessful and in hilar tumors where it may be
necessary to access more than one duct system.
(6613)Figure 642. Endoscopic retrograde cholangiopancreatography (ERCP) demonstrates

the typical "double-duct" sign with main pancreatic duct obstruction and a common bile duct
stricture due to tumor in the pancreatic neck.
(6614)Figure 643. Percutaneous transhepatic cholangiogram (PTC) shows complete

obstruction of the bile duct in a patient with cholangiocarcinoma.
Operating for Cure

Surgical resection is the only treatment that currently offers any hope of cure for patients with
malignant obstructive jaundice. This goal must remain prominent despite the fact that success is
distressingly rare. For example, the overall 5-year survival for patients with pancreatic cancer remains
less than 1%.46(6615) In terms of the overall prognosis for patients with malignant biliary obstruction,
it is easy to become confused by the results of most surgical series. These show increasing rates of
resectability and survival along with reduced operative morbidity. There is no doubt that surgical
techniques and perioperative care have advanced in recent years.58(6616) However, the main
reason that surgical results have improved is that the patients are now (appropriately) carefully
selected. Until about the mid-1980s, almost all patients with malignant obstructive jaundice underwent
laparotomy for diagnosis and treatment (usually limited to palliative bypass); operative morbidity and
mortality were high.7(6617) The availability nowadays of nonoperative methods for diagnosis and
palliation means that frail patients with extensive tumors need not undergo operation. Results therefore
appear better in those patients who do.8(6618)
The decision for or against surgery is crucial and complex. It is as unfortunate to miss an opportunity
for surgical cure in one patient as it is to operate unnecessarily in another. The two factors that must be
assessed carefully are operability (the patient's fitness for surgery) and resectability (the stage of the
tumor).
Tumor Staging
Sophisticated imaging techniques now allow documentation of tumor size and extension.911(6619)
CT is a popular method for tumor staging, although ultrasound, in the hands of an expert, can provide
equivalent or even better information. Both imaging methods have been enhanced by a variety of
techniques such as CT portography, spiral reconstruction, and Doppler imaging of blood vessels.
Biopsy-proven metastases and invasion or occlusion of major vessels provide compelling evidence of
unresectability. Nodal involvement is more difficult to define, but loss of the retropancreatic fat planes
on CT is good presumptive evidence of invasion. A pragmatic approach to staging is to classify tumors
broadly into five grades after initial imaging: (1) small (no mass visible, presence inferred only from
duct obstruction); (2) moderate (mass detectable but less than 2 cm in diameter); (3) large (more than
2 cm, but no definite invasion); (4) locally invasive; and (5) metastatic. More detailed information can
be obtained, when it may truly influence the decision for or against surgery, through endoscopic
ultrasonography (see Chapter 78: Endoscopic Ultrasonography of the Retroperitoneal Organs),
angiography, and laparoscopy.1216(6620) A combination of PTC and cholangioscopy has been
advocated for accurate localization and staging of bile duct tumors.17(6621)
Patient Fitness (Risk Factors for Surgery)

Decisions are influenced strongly by the patient's general condition, that is, the perceived risk of
intervention. Age is often emphasized but should not be the sole criterion. Operative morbidity and
mortality are mainly influenced by tumor burden, comorbid illnesses, nutrition, renal function, and
sepsis.1820(6622) Several risk factor scoring systems have been proposed, but none is generally
accepted in this specific clinical context.
A simple scale for fitness is that used by the American Society for Anesthesiology: (1) completely
healthy, (2) minor problem not requiring continuous treatment, (3) significant comorbid illness
controlled by medication, (4) severe and life-threatening comorbid illness, and (5) moribund (unlikely to
survive 24 hours).
The Decision (Combining Staging and Risk)

The decision whether to recommend operation for attempted cure can be approached using the
previously discussed numerical scales of tumor stage and patient fitness (Figure 644). Surgery will
usually be appropriate in patients whose fitness and staging scores each do not exceed 2. Surgery
may be offered to good-risk patients despite the presence of a large or even invasive tumor (i.e.,
fitness 1 and tumor stage 3 or 4), in the hope that imaging studies have overestimated tumor invasion
(which they rarely do). Conversely, an aging patient with significant comorbid illness may be
considered for resection if the tumor is small, for example, involving only the main duodenal papilla
(e.g., patient fitness 3 and tumor stage 1). In either circumstance, a palliative bypass can be performed
if resection appears impossible or too hazardous, as is often the case. Patients with combinations of
(un)fitness and tumor size beyond these numerical limits are probably best served by nonoperative
methods for palliation (see later).
(6623)Figure 644. Correlations of patient fitness and tumor stage in decisions for surgical
treatment. ILLNESS (patient fitness) indicates the approximate American Society for
Anesthesiology (ASA) grade: 1 = healthy; 2 = mild clinical problem; 3 = clinical problem
under therapeutic control; 4 = unstable clinical problem; and 5 = moribund. TUMOR
BURDEN pertains to staging grade: 1 = localized, less than 2 cm diameter; 2 = localized,
greater than 2 cm; 3 = possibly locally invasive; 4 = invasive; 5 = metastatic. Hatched squares
identify patients who usually undergo surgery in the hope of cure.
An important task for the future is to add precision to the assessment of patient fitness and tumor
staging, to provide better predictors of good and bad outcomes. Patients and their relatives are looking
for a greater part in the decision process and need objective data that they can understand and trust.
Palliative Management
Palliation means dealing effectively with present or likely symptoms. Jaundice and pruritus are often
dominant, but pain and emotional distress may demand equal attention. Duodenal obstruction is
usually managed surgically, by open operation (or laparoscopy). Unfortunately, gastroenterostomy is
not always completely effective owing to tumor size and neural invasion.
Before considering nonoperative methods for palliation of jaundice, it is important to confirm the
diagnosis of malignancy histologically wherever possible.
Tissue Diagnosis
Malignancy can be mimicked by many other pathologic conditions, including benign tumors,
lymphoma, spontaneous benign strictures, and even chronic infection (e.g., tuberculosis) (see also
Disorders of the Liver Affecting the Bile Ducts). Methods for nonoperative tissue diagnosis have
become more important as fewer patients undergo open surgery.
Scan-Guided Needle Cytology and Biopsy

Skilled radiologists can sample any tumor detected by ultrasound or CT.21(6624) Fine-needle
aspiration cytology under CT guidance is probably the most popular technique in the United States,
with a sensitivity of approximately 70% in patients with mass lesions (e.g., pancreatic cancer) (Figure
645). Targeting is more difficult and the sensitivity lower when no mass is seen, but appropriate tissue
may be obtained by aiming directly at an area of stricturing (or a stent if present). Cytology results are
dependent on the expertise of the cytologist, a fact that has led to increasing interest in techniques of
core biopsy with larger (and spring-loaded) needles.
(6625)Figure 645. CT-guided biopsy of a pancreatic mass.

Percutaneous biopsy techniques are rarely appropriate if surgery is planned. The result is unlikely to
alter the approach, and there is concern about tumor seeding along needle tracks.22,23(6626)
Tissue Diagnosis at ERCP

Tissue diagnosis may also be attempted at ERCP (or PTC).2427(6627) Standard biopsy forceps are
used to sample lesions involving the papilla or duodenum, and brushes can be inserted into the ductal
systems to obtain cytologic specimens. Various types of cytologic sampling devices are available,
including brushes for insertion over a guidewire28(6628) and brushes with a spring wire tip (Figure
646).29(6629) The sensitivity of brush cytology can approach 80%, although it is highest for primary
bile duct tumors, intermediate for pancreatic cancers, and lowest when biliary obstruction is caused by
metastatic malignancy.28,30,31(6630) Mansfield et al.30(6631) prospectively studied cytologic
sampling methods in 54 patients, 52 of whom ultimately proved to have malignant biliary obstruction
(various causes). Overall, cytologic sampling detected 28 malignancies. Samples were obtained by
brushing, from the proximal end of occluded stents, by inserting a threaded stent retriever device
through the stricture, and by centrifuging 20-ml samples of bile; adequate samples were obtained by
these methods in 96%, 84%, 70%, and 44% of cases, respectively. Of the 28 malignant tumors
detected, brush cytology was positive in 22, whereas the other methods accounted for an additional 6
positive diagnoses.
(6632)Figure 646. Brushing of a biliary tumor at ERCP to obtain a sample for cytologic
evaluation.
Smaller biopsy forceps can be passed into either the bile or the pancreatic ductal systems (Figure
647), but tissue specimens are often small and inadequate for diagnosis. A sphincterotomy is often
necessary to allow insertion of a biopsy forceps into the bile duct, although the use of newer types of
forceps obviates the need for this procedure.31(6633) A prospective study by Pugliese et al.32(6634)
found that the yield of positive diagnoses increased only marginally if fluoroscopically guided
endobiliary biopsy was performed in addition to brush cytology. Perforation of the bile duct was thought
to have occurred as a result of the biopsy procedure in one patient in this study. These results are in
contrast to the study of Schoefl et al.33(6635) Brush cytology was performed in 63 patients, forceps
biopsy in 106, and both in 48 patients; sensitivities were 46.7%, 64.9%, and 70.4%, respectively, and
these investigators recommended that bile duct strictures be sampled with both brush and forceps. In
the study of Sugiyama et al.,31(6636) biopsies obtained from bile duct strictures had a sensitivity of
81% compared with 32% for brushing and 48% for study of aspirated bile.
(6637)Figure 647. Insertion of biopsy forceps into the bile duct to obtain a tissue specimen.
Needle aspiration biopsy through the duct wall has been advocated,34(6638) with apparent good
success and safety. Newer devices for tissue sampling (e.g., tissue "scrapers") are being
evaluated.35(6639) Aspirated bile can be studied for tumor markers,36,37(6640) but it has not been
established that these more costly analyses offer an incremental benefit compared with standard
methods of tissue sampling.
Similar cytology, biopsy, and scraping techniques are used by radiologists via transhepatic biliary
drainage tracts. Cytologic examination of cellular material in bile itself (obtained at ERCP or PTC) has
a low sensitivity but may be worthwhile. In some cases, confirmation of the diagnosis can be made by
examination of ascitic fluid or by sampling distant metastases.
An important practical question is how long to persist in attempts to prove the diagnosis of malignancy
in a patient whose radiologic findings are characteristic. The discomfort and cost (and risk, albeit
minimum) of repeated procedures must be balanced against the small chance of diagnosing an
unexpected pathologic condition, particularly a lesion that might be amenable to alternative therapy
(e.g., lymphoma). A tissue diagnosis of malignancy will be needed if radiotherapy and chemotherapy
methods are to be applied.
Relief of Biliary Obstruction

Biliary obstruction can be relieved by surgery and by stenting (percutaneous or
endoscopic).38,39(6641) Before offering either approach to the patient, it is essential to establish that
the jaundice is indeed mainly due to obstruction and not the result of intrahepatic metastases (or
parenchymal failure) (Figure 648). This question is usually easily resolved by reviewing imaging
studies for evidence of duct dilation. A related question is whether the tumor load is so great that the
patient's illness is truly terminal; in some patients, intervention may not be worthwhile. However, this
judgment is often difficult to make, as many patients with extensive malignancy appear to improve
quickly once bile starts to flow.
(6642)Figure 648. CT shows widespread metastatic disease involving the liver. There is no
role for stenting in this jaundiced patient.
The belief that relieving jaundice is worthwhile is to some degree an act of faith; there has been no
randomized controlled study to indicate a benefit to patients. It is assumed that persistent obstruction
not only shortens life but also diminishes the quality of life, particularly through its effects on digestion
and hepatic and renal function. A survey of 19 patients with malignant bile duct obstruction found that
stent insertion not only provided complete relief of jaundice and itching but also reduced dyspeptic
symptoms and improved appetite.40(6643) Relief of jaundice certainly provides an important
psychologic boost to the patient and his or her family.
The specific techniques, merits, and drawbacks of different approaches are discussed separately
followed by comparisons of their roles in specific clinical situations.
Endoscopic Drainage Techniques

Sphincterotomy and Debulking
Biliary obstruction due to tumors of the main duodenal papilla can be relieved by sphincterotomy.
Tumor debulking can be achieved using diathermy, injection (e.g., alcohol), or with the
neodymium:yttrium-aluminum-garnet laser.4143(6644) Endoscopic diagnosis and treatment of these
tumors is discussed in Chapter 69: Tumors of the Main Duodenal Papilla.
Brachytherapy
Brachytherapy involves the intracavitary placement of a radioactive source within a malignant stricture.
For biliary strictures, this can be accomplished either endoscopically or via the percutaneous
transhepatic route.44,45(6645) Although reported results appear favorable, these data should be
considered preliminary in nature, as they are based on the treatment of relatively few patients in
uncontrolled studies. Furthermore, the endoscopic procedure is technically intricate and the handling of
radioactive materials represents an additional potential hazard for the endoscopist and supporting
personnel. This approach to palliation of malignant biliary strictures should therefore be regarded as
investigational.
Endoscopic Stenting
Endoscopic biliary stenting is relatively straightforward, once deep cannulation has been achieved with
a catheter and guidewire.46(6646) Technically, the most difficult cases are those in which the tumor
distorts the duodenum or actually involves the bile duct orifice. In the latter circumstance, it is easy to
make false passages and to cause bleeding sufficient to obscure the endoscopic view.
Some experts recommend needle knife "precut" sphincterotomy if cannulation cannot be achieved
rapidly with standard catheters. We prefer to use a bowed double-lumen sphincterotome to facilitate
cannulation in difficult cases. It is not necessary to perform an actual sphincterotomy to place a
standard stent of 10-French gauge. Sphincterotomy is appropriate if larger or multiple stents are to be
placed, and it may facilitate subsequent stent exchanges. A retrospective study of patients with biliary
strictures found a significantly higher incidence of postprocedure pancreatitis in patients with proximal
strictures. Tarnasky et al.47(6647) speculated that this could be explained by mechanical forces that
pressed the distal end of the stent against the orifice of the pancreatic duct and suggested that this
could be obviated by sphincterotomy before stent insertion.
The stents most commonly used in patients with malignant biliary obstruction are made of
polyethylene; they are slightly curved, available in several different lengths, and have an outside
diameter 10 or 11.5 French. They also have "Amsterdam" flaps at both ends to prevent proximal or
distal migration. It is known from in vitro studies that flow rates are greatest through "straight" or slightly
curved stents as opposed to those having a "pigtail" design.48(6648)
The length of stent chosen is such that the distance between the flaps is approximately 1 cm greater
than the distance between the proximal margin of the stricture and the papilla. Stents are available in
5, 7, 9, and 12 cm lengths; in practice, one of these lengths will usually be suitable in most cases. With
experience, it is usually possible to estimate the length of stent required from the fluoroscopic images
with acceptable accuracy, although more precise methods have been described.49(6649)
The stent is placed through a "therapeutic duodenoscope" that has an accessory channel diameter of
at least 4 mm, employing a "three-layer" system (Figure 649). This consists of a 400-cm-long
guidewire of 0.035 inch diameter, carrying a 6-French guiding catheter. These are placed sequentially
through the stricture, and the stent is railroaded into place using a "pusher tube." A single assembly
stent delivery system, the so-called One-Action Stent Insertion Set (OASIS; Wilson Cook,
Winston-Salem, NC) is also available. In a randomized trial involving 20 patients, stent insertion
appeared to be easier and faster with the OASIS system compared with the standard three-layer
system.50(6650) Most experts do not dilate malignant biliary strictures before stenting, but this may be
necessary (using stepped dilators) in long, tight, and hilar strictures. Hydrophilic-coated guidewires can
expedite passage through tight and tortuous strictures.
(6651)Figure 649. Standard "three-layer" stent system for use at ERCP. Note the "pusher
tube" (curved arrow), guide catheter (long arrows), stent (solid arrowhead), and guidewire
(open arrowhead).
Endoscopic stenting is technically demanding in tumors that involve the bifurcation (Bismuth types II to
IV). They are often rigid, and the distance between the tumor and the fulcrum of pressure (i.e., the
endoscope elevator) is greater. Some experts have recommended that attempts be made to drain all
(or most) obstructed ducts, as for example, by placing stents in both the left and the right hepatic ducts
when both systems are obstructed by a hilar tumor. They quote very high rates of cholangitis as the
basis for this recommendation.51(6652) However, most prefer to place a single stent and to await
events because placing two or more stents endoscopically is formidably difficult (Figure
6410).52(6653) The goal should be to drain the most dilated duct system (in a nonatrophic liver lobe,
as indicated by ultrasound or CT). Aiming for a specific segment can prove difficult despite skillful
manipulation of special devices, such as steerable guidewires. Some endoscopists tailor the stent
shape to the lesion and cut additional side holes. Further intervention is necessary if primary
endoscopic drainage fails or when sepsis develops. One option is the so-called combined
procedure.5355(6654)
(6655)Figure 6410. Two polyethylene stents placed at ERCP across an obstructing hilar
lesion.
The combined endoscopic-radiologic procedure was developed to assist endoscopists in gaining

access to bile ducts when endoscopic methods alone have failed to do so for technical reasons or to
gain access to a duct that drains a specific liver segment. An interventional radiologist passes a
guidewire transhepatically down the bile duct and into the duodenum (through a small catheter to
protect the liver). By means of any of several different devices (e.g., snare), the endoscopist then
grasps the wire, pulls it up through the instrument channel, and uses it for stent placement. A
particularly long wire (usually at least 400 cm) is required. The combined procedure became popular
because it was assumed (although never proved) to be safer than placing a large-bore polyethylene
stent directly through the liver substance. Serious complications can occur with the combined
procedure.53,56(6656) It is therefore probable that primary transhepatic placement of an expandable
metal stent is now preferable in this context.
Short-Term Results and Risks

The results of endoscopic biliary stenting depend on the expertise of the endoscopist and the site of
the tumor.3,38(6657) Success rates of over 90% are claimed for stenting of distal and midduct lesions,
sometimes after more than one attempt. Effective drainage is achieved much less frequently in
patients with hilar tumors.52,57,58(6658)
Sepsis is the most common complication.51,52,59,60(6659) The risk may be reduced by assiduous
disinfection of instruments and accessories, and most experts recommend prophylactic administration
of antibiotics before stent placement. Infection is usually due to inadequate drainage, which must be
rectified quickly. Any ERCP procedure can result in pancreatitis, especially when cannulation has been
difficult (see Chapter 56: Indications, Contraindications and Complications of Diagnostic Endoscopic
Retrograde Cholangiopancreatography). Sphincterotomy is associated with a variety of complications
including bleeding (see Chapter 60: Endoscopic Papillotomy). Perforation as a result of biliary stenting
is extremely rare. False passages can be caused by the use of guidewires, but this event rarely causes
clinical problems. Acute cholecystitis and gallbladder sepsis have occasionally followed stenting of
distally located tumors, presumably due to cystic duct compromise.6163(6660) The cases of
gallbladder sepsis reported by Ainley et al.63(6661) were successfully managed by percutaneous
cholecystostomy. If the length of the stent outside the papilla is too great, irritation of the duodenal wall
may occur; this has also led to ulceration and bleeding.
Procedure-related complications (especially sepsis and pancreatitis) can be severe and occasionally
fatal. However, most deaths within 30 days of stenting are due to the underlying disease.3,38,39(6662)
Long-Term Complications
Stent Migration
Stent migration (proximally or distally) is unusual, but it may impair drainage or result in injury to the
duodenum.64(6663) Two cases of impaction of biliary stents in colonic diverticula have been
reported.65(6664) In both instances, the stent was removed endoscopically and the patients recovered
with no apparent sequela. In the majority of cases, proximally migrated plastic stents can be recovered
by endoscopic methods.66(6665) In the study of Tarnasky et al.,66(6666) the most effective method
was insertion of a guidewire through the stent with removal by means of an over-the-wire extraction
device. Other methods included grasping with a forceps or basket and extraction with a catheter
balloon placed next to the stent. Endoscopic extraction of a proximally migrated stent was facilitated by
percutaneous transhepatic insertion of a guidewire in one case.67(6667) Surgery may rarely be
necessary.66(6668)
Stent Fracture
Fracture of indwelling plastic biliary stents is a rare complication. It has occurred when stents were left
in place for long periods of time, as for example, in patients with bile duct stones that cannot be
extracted.68(6669) However, there are also a few reports of stent fracture in patients with malignant
bile duct obstruction.69,70(6670)
Clogging of Plastic Stents
The main problem with polyethylene stents is their tendency to clog after a few months. Bacteria play a
pivotal role in this process. Bile from normal individuals is usually sterile.71,72(6671) Factors that are
believed to maintain sterility of bile include continuous flow, production of mucus by the biliary
epithelium, and the sphincter of Oddi, which acts as a mechanical barrier to bacteria in the
duodenum.73(6672) Bacteria may reach the bile ducts from the systemic circulation or portal venous
system.74(6673) Infection of bile is thought to be facilitated by elevations in intrabiliary
pressure.74(6674) With loss of the barrier offered by the sphincter of Oddi, as occurs with
sphincterotomy or insertion of a stent, bacterial colonization of the bile ducts occurs
rapidly.7577(6675)
Various types of studies, including electron microscopy of clogged plastic stents and the sludge that
accumulates within stents, reveal that the clogging material comprises protein, deconjugated bilirubin,
microcolonies of bacteria, and amorphous debris (Figure 6411).7880(6676) Formation of a protein
biofilm is probably the initial event.81(6677) Although little is known about the chemical nature and
origin of this biofilm, it appears to enhance the adherence of bacteria.81(6678) Once formed, it is
difficult to eliminate this biofilm from an indwelling stent; once present, bacterial adherence is certain to
occur.82(6679) Accumulation of the sludge that eventually occludes a stent occurs through the action
of enzymes produced by certain bacteria, specifically beta-glucuronidase and phospholipase enzymes.
These enzymatic proteins deconjugate bilirubin and other substances present in bile with the resultant
formation of calcium bilirubinate and calcium salts of fatty acids. Microscopic surface irregularities and
imperfections that occur during the manufacture of stents probably facilitate bacterial
colonization;83,84(6680) sludge accumulates mainly around side holes and at the ends of
stents.83,84(6681)
(6682)Figure 6411. Scanning electron microscopy of the surface of an obstructed

polyethylene stent shows amorphous debris, crystals, and bacteria.
Prevention of Clogging
No dependable method has been devised to prevent stent clogging despite an extensive knowledge of
the process of stent occlusion. There are several approaches to this problem, including reduction or
elimination of bacterial colonization of the bile ducts, prevention of biofilm formation or adherence of
bacteria, increasing stent diameter, designs for stents that resist bacterial adherence, and maneuvers
to unclog stents.
There are three possibilities for preventing colonization of bile ducts: systemic administration of
antibiotics, pharmacologic manipulation of the composition of bile, and preservation of the mechanical
barrier function of the sphincter of Oddi. The first approach has limitations in terms of cost, patient
compliance, emergence of resistant bacterial strains, and the potential effects of long-term alterations
in the gut flora. Effective pharmacologic agents are lacking with respect to the second approach.
Preservation of the sphincteric barrier has associated technical problems.
Smit et al.85(6683) found no benefit when 60 patients who had undergone insertion of a biliary stent
were treated with aspirin (to reduce mucin secretion) and doxycycline. However, doxycycline is
probably not the best choice of antimicrobial agent. In the study of Ghosh and Palmer,86(6684) no
benefit was found for the sequential administration of ampicillin, metronidazole, and ciprofloxacin plus
ursodeoxycholic acid (to increase bile flow). In the randomized trial of Barrioz et al.,87(6685) the
combination of norfloxacin and ursodeoxycholic acid significantly increased stent patency, although the
median duration of patency in the control group was unusually short. Ursodeoxycholic acid has not
been used alone in any clinical trial, and the efficacy of this agent for the prevention of stent clogging is
therefore unknown.
Bile salts have variable bactericidal effects. Hydrophobic bile salts (e.g., taurodeoxycholate and
deoxycholate) inhibit the adherence of biliary pathogenic bacteria in vitro,88,89(6686) but these agents
are not well tolerated by patients. By contrast, hydrophilic bile salts (e.g., ursodeoxycholate) are better
tolerated, but they do not prevent bacterial adherence.86,87(6687)
In patients with proximal biliary strictures, the sphincter of Oddi may be preserved as a barrier to
bacterial colonization by inserting a stent, without sphincterotomy, completely into the bile duct.
Although there are no clinical data concerning the efficacy of this approach, one study in animals found
that full insertion prolongs stent patency.90(6688) Full insertion is impossible when there is obstruction
of the distal bile duct and full insertion will make stent exchange especially difficult.
Biofilm formation or the adherence of bacteria may in theory be prevented by using ultrasmooth
materials in making stents, impregnating stents with bactericidal agents, and designing stents that
resist adherence.
Among the various materials used to fabricate stents, Teflon has been shown in vitro to be the most
resistant to bacterial adherence.91(6689) Stents made with new chemical substances including
cross-linked polymers of urea (Vivathane) and poly-N-vinylpyrrolidone (PVP) (Hydromer) have been
shown to resist bacterial adherence in vitro.9294(6690) In vitro studies have demonstrated that
coating polyurethane stents with silver reduces bacterial adherence.95(6691) Unfortunately, none of
these approaches has been conclusively shown to increase the duration of stent patency in humans.
Stents have been designed with elimination of the rough edges and side holes that enhance adherence
of bacteria. Elimination of side holes was not advantageous in one clinical trial and caused structural
weakening that increased the likelihood of migration.96(6692) Newer designs may prove to be more
efficacious.91(6693)
In summary, construction of stents of different materials and designs as well as the use of various
medical therapies (e.g., antibiotics, bile salts) has not yet been shown to prolong stent function
significantly.85,97,98(6694) It is possible to unclog an indwelling biliary stent. One proposed innovative
approach uses low-power repetitive laser pulses.99(6695) There are no published data on the efficacy
of maneuvers to unclog stents, but it can be assumed that clogging will recur rapidly.
The most direct approach to prolonging patency is to increase the diameter of the stent. Clinical
studies have demonstrated that larger-diameter stents remain patent for longer periods.100102(6696)
Currently available duodenoscopes accept stents up to a maximum diameter of 12 French. It is
debatable whether an 11.5-French-diameter stent remains patent substantially longer than one of
10-French diameter.103(6697)
Stent Exchange
The risk of occlusion of standard 10- and 11.5-French-gauge polyethylene stents appears to increase
progressively after about 3 months. Whether an individual patient will require a stent exchange
depends on her or his length of survival. Patients, relatives, and primary doctors must be adequately
informed about the first symptoms of stent occlusion (essentially cholangitis); stents should be
changed urgently if these occur because life-threatening sepsis can develop quickly. Recently (without
proof of cost benefit), many patients are being advised to undergo elective stent change after 3 to 4
months. Intervals between stent changes of up to 6 months have also been recommended.104(6698)
Elective stent exchange appears reasonable if the patient's general condition remains good.
The standard technique for changing a stent is to grasp the distal tip in the duodenum (using a basket
or snare) and to pull it out through the mouth by removing the duodenoscope, which is then reinserted
for cannulation and stent replacement. Occasionally, difficulty in recannulating the papilla (or stricture)
may be encountered. This has led to the development of techniques for exchanging stents over a
guidewire, so that access is not lost (Figure 6412). These include the use of a threaded device over a
guidewire that has been placed through the stent,105(6699) extraction of the stent over a guidewire
and through the accessory channel of the endoscope by means of a small snare,106(6700) and
placement of a guidewire next to the stent with subsequent extraction of the stent through the
endoscope channel by means of a small snare while leaving the guidewire in place.107(6701) These
methods should be considered in patients where initial stent insertion had proved difficult, especially in
hilar tumors.
(6702)Figure 6412. Use of the threaded "screw-extractor" device to remove a stent at ERCP
over a guidewire.
Endoscopic Placement of Expandable Metal Stents

Several expandable metal mesh stents are now available for biliary use, and this approach to relief of
biliary obstruction is evolving rapidly.108111(6703) The principle is to place a large-diameter stent
through a relatively small catheter to minimize the trauma of insertion and to prolong stent survival.
Thus far, most experience has been obtained using the stainless steel Wallstent (Schneider, Inc.,
Plymouth, MN), which is delivered from a catheter system of 9 French gauge over a standard
guidewire (Figures 6413 and 6414). Once released, it expands to 10 mm diameter (over several
hours) and contracts in length from 100 to 68 mm; shorter versions are available. For distal bile duct
lesions (e.g., pancreatic head cancer), the tip is usually left just outside the papilla. Sphincterotomy is
not necessary. Wallstents cannot be moved once they have been fully implanted, but additional stents
can be placed through the lumen if necessary. Success rates for insertion of over 95% have been
reported, with few immediate complications (Table 641).108110,112(6704) Erosion of biliary metal
stents through the duodenal wall with resultant hemorrhage has been reported.113,114(6705)
Follow-up studies have shown a low incidence of stent clogging by the usual bacteria/bile salt process,
but stents have become obstructed by tumor (or tissue) ingrowth or overgrowth (Figure 6415; see
also Table 641).115117(6706) For example, in the study of O'Brien et al.,118(6707) 13 of 22
patients with a variety of malignant tumors developed stent obstruction due to tumor ingrowth (median
follow-up 14.6 months). The retrospective review of Van Berkel et al.119(6708) found that 13 of 28
patients with metastatic malignant biliary obstruction developed symptoms of recurrent obstruction.
Retrograde cholangiography in 10 patients demonstrated tumor overgrowth in 3 and ingrowth in 7. The
number of patients treated thus far has not been sufficient to identify specific risk factors for
obstruction. Blocked stents can be debulked by using diathermic devices,117(6709) by
brachytherapy,120(6710) or more simply by dragging an extraction balloon through the obstructed
segment.118(6711) However, most endoscopists prefer to insert a standard polyethylene stent (Figure
6416) or in some cases a second metal expandable stent.118,119,121(6712) This method seems to
work well, possibly because of friction between the two stents.109(6713)
TABLE 641
Reports of Endoscopic Insertion of Metal Expandable Stents for Malignant Biliary
segment.118(6711) However, most endoscopists prefer to insert a standard polyethylene stent (Figure
6416) or in some cases a second metal expandable stent.118,119,121(6712) This method seems to
work well, possibly because of friction between the two stents.109(6713)
TABLE 641
Reports of Endoscopic Insertion of Metal Expandable Stents for Malignant Biliary

YEAR
NUMBER OF
PATIENTS
TECHNICAL
SUCCESS (%)
COMPLICATIONS
(%)
DYSFUNCTION
(%)
Neuhaus et al. 111
1991
39
100
13
Huibregtse et al. 108
1992
103
97
17
Wagner et al. 164
1993
11
100
18
Goldin et al. 165
1993
10
100
20
20
Hoepffner et al. 166*
1994
118
96
48
14
Schofl et al. 167
1994
52
92
15
19
Van Berkel et al. 119
1996
28
100
48
REFERENCE
* 102/118 inserted endoscopically.

Biliary obstruction due to metastatic malignancy.
(6714)Figure 6413. Wallstent partially released from its catheter.
(6715)Figure 6414. Wallstent deployed in a distal biliary tumor in a patient with a

temporary nasobiliary drain.
(6716)Figure 6415. Tumor ingrowth through a Wallstent.
(6717)Figure 6416. Placement of a polyethylene stent through an obstructed Wallstent.

The problem of tissue and tumor ingrowth is being addressed with the development of expandable
metal stents with plastic sleeves. A comparative study in animals suggested that membrane-covered
metal stents induce less mucosal hyperplasia.122(6718) Clinical results are not yet available; there are
concerns that the sleeve may compromise drainage of other ducts (e.g., pancreatic, cystic, or
intrahepatic ducts), or result in migration of the stent.
Metal or Polyethylene?
Expandable metal stents are as easy, or easier, to place than polyethylene stents, and the incidence of
stent obstruction was lower in two randomized studies. Their major disadvantage is substantially higher
cost compared with plastic stents.123(6719)
The randomized trail of Davids et al.109(6720) comparing polyethylene and metal expandable stents
found that the latter were cost-effective in patients who survived beyond a few months because the
overall number of ERCP procedures was reduced by 28%. In this study of distal bile duct obstruction
due to malignancy, 105 patients were randomized to undergo placement of either a metal expandable
(n = 49) or polyethylene stent (n = 56). The median duration of patency after initial placement was
significantly longer for the metal stent.
In the randomized trial of Knyrim et al.,124(6721) 62 patients with malignant bile duct obstruction were
randomized to undergo placement of either a metal expandable stent (Wallstent) (n = 31) or an
11.5-French polyethylene stent (n = 31). The distal bile duct was the site of obstruction in the majority
of patients. Stents were placed successfully in all cases. The 30-day mortality rate was similar for both
groups; median follow-up was 5 months for both. There was no significant difference between the two
groups with regard to the frequency of stent dysfunction, either occlusion or migration of plastic stents.
However, there was a significantly higher incidence of cholangitis in the plastic stent group (36% vs.
15%). Furthermore, a significantly greater number of endoscopic interventions subsequent to initial
stent placement was required in the plastic stent group. There were also significant differences in days
in hospital as well as cost of treatment of stent-related complications. Although the total overall cost of
treatment was higher for plastic stent-treated patients, the difference was not significant.
Percutaneous Drainage Techniques

Transhepatic radiologic techniques have been refined over several decades, and skilled interventional
radiologists can access and drain obstructed (dilated) bile ducts in virtually every patient.125(6722)
Relative contraindications to the transhepatic approach include gross ascites, major hepatic
metastases, and coagulopathy. Simple external biliary drainage is useful only for short-term
decompression. Permanent palliation demands crossing the obstruction, which should be possible in
most cases using appropriate guidewire techniques.
External-internal drainage catheters were used for many years (Figure 6417).126(6723) This system
allowed the radiologist to check, manipulate, and change the catheter when necessary. However, the
presence of an external catheter for extended periods of time is uncomfortable and should be avoided
if possible. Radiologists can also insert indwelling polyethylene stents of 8 to 14 French gauge into the
bile ducts. However, transhepatic placement is probably more hazardous compared with the
endoscopic transpapillary approach because of trauma to the liver and capsule.127(6724) Bile
leakage, pulmonary complications, and hemobilia (and false aneurysms) are encountered.
Polyethylene stents placed transhepatically have the same tendency to occlude as those placed
endoscopically. They can be exchanged at ERCP if the tip has been left in the duodenum.
Interventional radiologists can also exchange or place new stents, but this requires a further
percutaneous procedure.
(6725)Figure 6417. External-internal percutaneous transhepatic drainage.

There is considerable enthusiasm among radiologists for placement of expandable metal mesh
stents.128136(6726) They are easier and less traumatic to insert than polyethylene stents, and the
risk of clogging may be lower. When appropriate, patients with hilar tumors can have stents placed
percutaneously into more than one liver segment via bilateral approaches or through a single liver
puncture (one stent into the duodenum and another connecting left and right intrahepatic ducts) (Figure
6418).137(6727)
(6728)Figure 6418. Bilateral metal expandable stents (Wallstent) placed percutaneously

across a hilar tumor.
Surgical Biliary Drainage

Surgical biliary bypass, a standard operation for many decades, was the only method available for
biliary decompression before the development and widespread application of stenting in the early
1980s. Cholecystojejunostomy and choledochoduodenostomy have limited roles because of the risk of
recurrent obstruction to the stoma or cystic duct. Most authorities recommend Roux-en-Y anastomosis
of the jejunum to the common hepatic duct.
Success rates for surgical bypass are rarely presented on an intention-to-treat basis. Many patients
with malignant obstructive jaundice have simply undergone "open-and-shut" laparotomy, especially
those with extensive or proximally located tumors; biliary bypass is performed for distal tumors when
there is no hope of resection.7,138(6729)
Smith et al.139(6730) randomized patients with malignant distal bile duct strictures to undergo
endoscopic stent placement (n = 101) or surgical biliary bypass (n = 103). Technical success rates
were high in both groups. There were significant differences in favor of endoscopy with regard to
procedure-related mortality and morbidity as well as total days in hospital. However, only 2 surgical
patients developed recurrent jaundice, whereas 36 stented patients became icteric subsequent to the
initial successful decompression. Also, 17% of endoscopically treated patients developed late gastric
outlet obstruction compared with only 2 patients who had undergone surgery. Survival rates were
similar for both groups.
The early mortality of surgical bypass has been reviewed extensively; figures vary enormously, largely
due to patient selection, but the average was 14% in studies published in the 1980s.7(6731)
Recurrence of jaundice after surgical bypass is very unusual (at least in patients with distal lesions).
The need to perform gastroenterostomy electively at the time of biliary bypass is a topic of
considerable debate.7(6732)
Gastroenterostomy and cholecystojejunostomy procedures are now being performed
laparoscopically.140142(6733) Anastomosis of the bile duct to the jejunum is technically much more
difficult, and appears unlikely to become a routine procedure.
Palliative surgical drainage of hilar tumors is complex and technically demanding.143(6734) New
dissection techniques and a greater understanding of hilar anatomy permit more extensive operations.
When feasible, Roux-en-Y jejunal anastomosis may be the best approach.
Palliative Drainage Choices in Context

As has been emphasized already, the spectrum of patients with malignant obstructive jaundice is
broad; tumors present in a variety of sites and sizes, in patients of (virtually) any age and health status.
Few clinicians see the entire spectrum. Surgeons, endoscopists, and radiologists alike have differing
viewpoints on the treatment of malignant biliary obstruction that are biased by their experience with
patients who have been more or less preselected by other physicians. Failure to appreciate this
problem of "apples and oranges" has led to some unnecessary and ill-informed debate. Much of the
available literature is not helpful because patients are not categorized sufficiently to indicate where they
fit on the checkerboard of fitness and tumor stage (see Figure 644). This problem of differences in
case mix is classically eliminated by performing randomized studies; there are rather few of
these.127,139,144146(6735) Furthermore, the results obtained with the various techniques (stenting
in particular) are somewhat operator-dependent; conclusions drawn based on data from one institution
may not be applicable elsewhere.127(6736)
Another problem in the individual case is defining the criteria by which the choice of treatment should
be made.147(6737) Short-term procedural morbidity and mortality are certainly important factors.
However, the quality and duration of palliation must also be considered. Who should make this
assessment? Physicians have concentrated on measuring the need for further intervention and length
of survival outside the hospital. Society and insurance carriers will look at relative costs. Patients are
interested also in quality of life issues, although individuals may define this differently.
All of these caveats make it difficult to provide authoritative guidance (based on firm scientific
evidence) in any given clinical situation. At best, it is possible to discuss the perceived advantages and
disadvantages of different approaches to the management of tumors in different sites.
Tumors of the Main Duodenal Papilla

Carcinomas of the papilla (and lesions showing only dysplastic changes on endoscopic biopsy) should
be resected in the hope of cure.148150(6738) This has an acceptably low mortality, even in selected
elderly patients. Local resection is a reasonable alternative in patients judged unfit for
pancreatoduodenectomy, although recurrence after this operation is relatively common. Surgery is
inappropriate when the risk is high because major comorbid illness or other diseases severely worsen
long-term prognosis. Rarely, tumors of the papilla may have metastasized by the time of presentation;
some papillary tumors are themselves metastases. Others occur in conjunction with multiple duodenal
tumors (e.g., Gardner's syndrome), in which case total surgical extirpation is impossible.
A variety of endoscopic techniques are available for palliative biliary drainage in patients who are not
operative candidates and those with tumor recurrence after local resection (see Chapter 69: Tumors of
the Main Duodenal Papilla).4143(6739) Repeated endoscopic treatment is often necessary, and
further complications such as bleeding and duodenal obstruction may arise. Percutaneous
transhepatic procedures are used only in high-risk patients whose biliary obstruction cannot be relieved
by endoscopic methods (e.g., previous major gastric surgery). Some poor-risk patients might be
candidates for laparoscopic cholecystoenterostomy.
Distal Obstruction Due to Pancreatic and Biliary Tumors

Enthusiasm for attempting resection of distal (nonpapillary) tumors varies among centers.5,6,20(6740)
In the individual patient, the decision for surgery should be influenced by general health status and
tumor staging (see Figure 644). Surgical exploration in the hope of resection (with bypass as a
fallback) is certainly appropriate in relatively young and fit patients when standard imaging studies
show no definite evidence of unresectability. At the other end of the disease spectrum are those
patients (unfortunately the majority) with a large tumor load and substantial comorbid illness. These
patients should be treated by stenting, unless operation is required for duodenal obstruction. Whether
the approach will be endoscopic or percutaneous will be determined by the relative levels of expertise
available. The endoscopic method is generally preferred, since it is equally effective, quicker, and
probably safer. Long-term results are likely to be similar. The only randomized study comparing these
two approaches is now obsolete.127(6741)
Clinical decisions will be more difficult for those patients who fall into the middle of the spectrum of risk
and tumor burden. Randomized studies have compared endoscopic stenting and surgery in
medium-risk patients with distal unresectable tumors not causing duodenal
obstruction.139,144,145(6742) Predictably, the short-term results favor endoscopy because of lower
morbidity, mortality, and shorter hospital stay. Unfortunately, these gains erode with time because
significantly more stented patients require further interventions for recurrent jaundice (stent blockage),
and a few will require surgery for duodenal obstruction.139(6743) The incidence of duodenal
obstruction has been lower than would be anticipated from general surgical series,7(6744) presumably
because patients with impending obstruction were excluded for these studies. Thus, in this context, the
patient may choose a somewhat more dangerous, prolonged, and painful initial (surgical) approach
that is more effective long-term, or a simpler, quicker, safer method (endoscopic stenting) that may
need to be repeated. Conclusions from these randomized studies are already outdated by the
development of expandable metal stents, which tend to function longer. Laparoscopic biliary bypass is
unlikely to have a significant impact, at least for several years.
Hilar Tumors
Primary tumors at the liver hilum are often relatively small, but their location makes resection difficult
and hazardous.143,147(6745) Newer surgical techniques justify exploration in those patients in whom
the risk of surgery is relatively low.151156(6746) However, most patients are either poor candidates
for operation or have extensive or metastatic lesions.157(6747) Bismuth et al.151(6748) state that
attempted resection can provide good palliation, but this applies only to a fraction of the total patient
spectrum; most patients are treated by stenting. Results of percutaneous and endoscopic intervention
in Bismuth type I lesions are reasonable, but many problems occur in the management of patients with
types II, III, and IV involvement. Opinions vary considerably as to whether it is sufficient to drain only
one segment of the liver (provided it is not atrophic) or whether it is necessary to attempt bilateral
treatment, which is more difficult with a primary endoscopic approach. Most endoscopists place a
single stent, adding a second via a combined or primary percutaneous procedure if drainage is not
sufficient or if sepsis occurs subsequently.51,52,57(6749) Most interventional radiologists believe that
hilar tumors are best dealt with by the percutaneous route, draining one or both sides.125(6750) This
has become much easier with expandable metal stents.128135(6751) However, a randomized study
comparing percutaneous and endoscopic stenting of hilar tumors is needed. At present, the choice
usually depends on locally available expertise.
Preoperative Biliary Drainage

The knowledge that deeply jaundiced patients may fare badly after surgery led to some enthusiasm for
preoperative external biliary drainage. Eventually this practice was shown to confer no advantage and
was furthermore associated with many complications.158(6752) Preoperative internal drainage by
stenting is more attractive; it can continue for several weeks if necessary, without catheter-induced
discomfort or risk of significant complication. Karsten et al.159(6753) compared outcomes for patients
with biliary obstruction due to pancreatic tumors who underwent preoperative biliary drainage (n = 184)
and those who did not undergo a drainage procedure before surgery (n = 57); there was no significant
difference in postoperative complications in this consecutive case series. A randomized study would be
necessary to demonstrate whether the presumed benefits of preoperative biliary drainage are actually
realized; however, operative mortality is now so low in many centers that large numbers of patients
would be required for a comparative trial.
In practice, many patients come to ERCP for diagnosis of malignant obstructive jaundice and will have
a stent placed even if surgical resection may be undertaken later. Stenting initiates treatment and
allows a period of further evaluation for operability and resectability. It is assumed that placement of a
stent reduces the risk of post-ERCP cholangitis, but this has not been proved.
Adjuvant Therapy
Radiotherapy and chemotherapy have thus far proved disappointing in the management of patients
with malignant biliary obstruction.160(6754) The combination of external beam radiotherapy and
5-fluorouracil has been shown to provide some benefit in randomized studies; improved regimens are
needed. Intrabiliary irradiation can be performed with iridium-192 wires placed via the percutaneous or
endoscopic route.161(6755) There are no controlled data to indicate significant benefit, and
prolongation of life is inherently unlikely. Some studies have suggested that some irradiated patients
can eventually be managed without stents. Whether this fact alone justifies the cost and complexity of
treatment requires clarification.
There have been attempts to "core-out" malignant biliary obstruction with lasers, diathermy, and heat
probes.162(6756) These methods should be considered investigational, as their efficacy and safety
remains unproved.
ConclusionTeam Management
Malignant biliary obstruction is a complex problem. The sophistication of the techniques now available
for the management of patients with these types of cancer demands a multispeciality team
approach.163(6757) The management team should include gastroenterologists, surgeons, and
interventional radiologists, as well as imaging radiologists, histopathologists, cytopathologists,
oncologists, support nurses, and others. However, few centers have developed integrated facilities for
management of patients with hepatobiliary and pancreatic diseases. Teamwork is effected mainly at a
personal level and is facilitated by regular meetings to discuss policies and outcomes. It will be
impossible to compare and improve the results of treatment unless all specialties dealing with these
patients collect comparable data and communicate using a common language.
A review of this topic would be incomplete without reference to the emerging imaging technology of
magnetic resonance cholangiography and pancreatography (MRCP). MRCP can provide good-quality
cholangiograms without any contrast, medication, or risk. It has been shown to have high sensitivity in
defining the site and cause of biliary obstruction. The role of MRCP in the investigation of the jaundiced
patient remains to be defined precisely (see also Chapter 56: Indications, Contraindications, and
Complications of Diagnostic Endoscopic Retrograde Cholangiopancreatography), but it is likely to have
considerable impact in the efficient triage of patients, especially those with hilar lesions.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
Niederau C, Grendell JH. Diagnosis of pancreatic carcinoma: Imaging techniques and tumor
markers. Pancreas 1992;7:668.
Ramage JK, Donaghy A, Farrant JM, et al. Serum tumor markers for the diagnosis of
cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1995;108:8659.
Slivka A, Carr-Locke D. Therapeutic biliary endoscopy. Endoscopy 1992;24:10019.
Gudjonsson B. Cancer of the pancreas: 50 years of surgery. Cancer 1987;60:2284303.
Crist DW, Sitzmann JV, Cameron JL. Improved hospital morbidity, mortality, and survival after
the Whipple procedure. Ann Surg 1987;206:35865.
Warshaw AL, Fernandez-del Castillo C. Pancreatic carcinoma. N Engl J Med
1992;326:45565.
Watanapa P, Williamson RC. Surgical palliation for pancreatic cancer: Developments during
the past two decades. Br J Surg 1992;79:820.
Pellegrini CA, Heck CF, Raper S, Way LW. An analysis of the reduced morbidity and mortality
rates after pancreaticoduodenectomy. Arch Surg 1989;124:77881.
Freeny PC. Radiologic diagnosis and staging of pancreatic ductal adenocarcinoma. Radiol
Warshaw AL, Gu ZY, Wittenberg J, Waltman AC. Preoperative staging and assessment of
resectability of pancreatic cancer. Arch Surg 1990;125:2303.
Adam A, Benjamin IS. The staging of cholangiocarcinoma [Review]. Clin Radiol
1992;46:299303.
Rsch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by
Tio TL, Cheng J, Wijers OB, et al. Endosonographic TNM staging of extrahepatic bile duct
cancer. Comparison with pathological staging. Gastroenterology 1991;100:135161.
Lightdale CJ. Laparoscopy for cancer staging. Endoscopy 1992;24:6826.
Mukai H, Nakajima M, Yasuda K, et al. Evaluation of endoscopic ultrasonography in the
pre-operative staging of carcinoma of the ampulla of Vater and common bile duct.
Warshaw AL, Tepper JE, Shipley WU. Laparoscopy in the staging and planning of therapy for
pancreatic cancer. Am J Surg 1986;151:7680.

17. Nimura Y. Staging of biliary carcinoma: Cholangiography and cholangioscopy. Endoscopy
1993;25:7680.
18. Pedrazzoli S, Bonadimani B, Sperti C, et al. Forecast of surgical risk in pancreatic cancer. Am
J Surg 1987;153:3747.
19. Pitt HA, Cameron JL, Postier RG, Gadacz TR. Factors affecting mortality in biliary tract
surgery. Am J Surg 1981;141:6672.
20. Cameron JL, Crist DW, Sitzmann JV, et al. Factors influencing survival after
pancreaticoduodenectomy for pancreatic cancer. Am J Surg 1991;161:1205.
21.
Hall-Craggs MA, Lees WR. Fine-needle aspiration biopsy: Pancreatic and biliary tumors. AJR
1986;147:359403.
22. Rashleigh-Belcher HJC, Russell RCG, Lees WR. Cutaneous seeding of pancreatic carcinoma
by fine needle aspiration biopsy. Br J Radiol 1986;59:1823.
23. Warshaw AL. Implications of peritoneal cytology for staging of early pancreatic cancer. Am J
Surg 1991;161:2630.
24. Ryan ME. Cytologic brushings of ductal lesions during ERCP. Gastrointest Endosc
1991;37:13942.
25. Kurzawinski T, Deery A, Dooley J, et al. A prospective controlled study comparing brush and
bile exfoliative cytology for diagnosing bile duct strictures. Gut 1992;33:16757.
26. Foutch PG, Kerr DM, Harlan JR, Kummet TD. A prospective controlled analysis of endoscopic
cytotechniques for diagnosis of malignant biliary strictures. Am J Gastroenterol
1991;86:57780.
27. Lee JG, Leung JW, Baillie J, et al. Benign, dysplastic, or malignant-Making sense of
endoscopic bile duct brush cytology: Results in 149 consecutive patients. Am J Gastroenterol
1995;90:7226.
28. Foutch PG, Kerr DM, Harlan JR, et al. Endoscopic retrograde wire-guided brush cytology for
diagnosis of patients with malignant obstruction of the bile duct. Am J Gastroenterol
1990;85:7915.
29. Scudera PL, Koizumi J, Jacobson IM. Brush cytology evaluation of lesions encountered during
30. Mansfield JC, Griffin SM, Wadehra V, Matthewson K. A prospective evaluation of cytology
from biliary strictures. Gut 1997;40:6717.
31. Sugiyama M, Atomi Y, Wada N, et al. Endoscopic transpapillary bile duct biopsy without
sphincterotomy for diagnosing biliary strictures: A prospective comparative study with bile and
brush cytology. Am J Gastroenterol 1996;91:4657.
32. Pugliese V, Conio M, Nicolo G, et al. Endoscopic retrograde forceps biopsy and brush
cytology of biliary strictures: A prospective study. Gastrointest Endosc 1995;42:5206.
33. Schoefl R, Haefner M, Wrba F, et al. Forceps biopsy and brush cytology during endoscopic
retrograde cholangiopancreatography for the diagnosis of biliary stenoses. Scand J
34. Howell DA, Beveridge RP, Bosco J, Jones M. Endoscopic needle aspiration biopsy at ERCP
in the diagnosis of biliary strictures. Gastrointest Endosc 1992;38:5315.
35. Howell DA, Parsons WG, Jones MA, et al. Complete tissue sampling of biliary strictures at
ERCP using a new device. Gastrointest Endosc 1996;43:498502.
36. Lee JG, Leung JW, Cotton PB, et al. Diagnostic utility of K-ras mutational analysis on bile
obtained by endoscopic retrograde cholangiopancreatography. Gastrointest Endosc
1995;42:31720.
37. Buffet C, Fourre C, Altman C, et al. Bile levels of carcino-embryonic antigen in patients with
hepatopancreatobiliary disease. Eur J Gastroenterol Hepatol 1996;8:1314.
38. Cotton PB. Nonsurgical palliation of jaundice in pancreatic cancer. Surg Clin North Am
1989;69:61327.
39. Soehendra N, Grimm H, Berger B, Nam VC. Malignant jaundice: Results of diagnostic and
therapeutic endoscopy. World J Surg 1989;13:1717.
40. Ballinger AB, McHugh M, Catnach SM, et al. Symptom relief and quality of life after stenting
for malignant bile duct obstruction. Gut 1994;35:46770.
41. Lambert R, Ponchon T, Chavaillon A, Berger F. Laser treatment of tumors of the papilla of
Vater. Endoscopy 1988;20(suppl 1):22731.
42. Huibregtse K, Tytgat GN. Carcinoma of the ampulla of Vater: The endoscopic approach.
43.
Hawes RH, Jamidar PA. Endoscopic diagnosis and treatment of ampullary tumors.
44. Montemaggi P, Morganti AG, Dobelbower RR Jr, et al. Role of intraluminal brachytherapy in
extrahepatic bile duct and pancreatic cancers: Is it just for palliation?. Radiology
1996;199:8616.
45. Urban MS, Siegel JH, Pavlou W, et al. Treatment of malignant biliary obstruction with a
high-dose rate remote afterloading device using a 10 F nasobiliary tube [see comments].
46.
Cotton PB, Williams CB. Practical Gastrointestinal Endoscopy. 4th ed. Cambridge, MA:
Blackwell Scientific. 1996.
47. Tarnasky PR, Cunningham JT, Hawes RH, et al. Transpapillary stenting of proximal biliary
strictures: Does biliary sphincterotomy reduce the risk of postprocedure pancreatitis?.
48. Scheeres D, O'Brien W, Ponsky L, Ponsky J. Endoscopic stent configuration and bile flow
rates in a variable diameter bile duct model. Surg Endosc 1990;4:913.
49. Kendall BJ, Jutabha R, So L, Jamidar PA. Determining required stent length in endoscopic
retrograde biliary stenting. Gastrointest Endosc 1995;41:2423.
50. Lawrie BW, Pugh S, Watura R. Bile duct stenting: A comparison of the One-Action Stent
Introduction System with the conventional delivery system. Endoscopy 1996;28:299301.
51.
Deviere J, Baize M, Vandermeeren A, et al. Endoscopic stenting for biliary strictures. Acta
Gastroenterol Belg 1991;55:295305.
52. Polydorou AA, Cairns SR, Dowsett JF, et al. Palliation of proximal malignant biliary obstruction
by endoscopic endoprosthesis insertion. Gut 1991;32:6859.
53. Dowsett JF, Vaira D, Hatfield AR, et al. Endoscopic biliary therapy using the combined
percutaneous and endoscopic technique. Gastroenterology 1989;96:11806.
54. Robertson DAF, Hacking CN, Birch S, et al. Experience with a combined percutaneous and
endoscopic approach to stent insertion in malignant obstructive jaundice. Lancet
1987;2:144952.
55. Martin DF. Combined percutaneous and endoscopic procedures for bile duct obstruction. Gut
1994;35:10112.
56. Tam PC, Lai EC, Hui WM, Chan SC. Pitfalls of percutaneous-endoscopic biliary stent
placement. Am J Gastroenterol 1990;85:2079.
57. Ducreux M, Liguory C, Lefebvre JF, et al. Management of malignant hilar biliary obstruction by
endoscopy. Results and prognostic factors. Dig Dis Sci 1992;37:77883.
58. Cheung KL, Lai EC. Endoscopic stenting for malignant biliary obstruction. Arch Surg
1995;130:2047.
59. Motte S, Deviere J, Dumonceau JM, et al. Risk factors for septicemia following endoscopic
60. Huibregtse K, Katon RM, Coene PP, Tytgat GN. Endoscopic palliative treatment in pancreatic
61. Leung JW, Chung SC, Sung JY, Li MK. Acute cholecystitis after stenting of the common bile
duct for obstruction secondary to pancreatic cancer. Gastrointest Endosc 1989;35:10910.
62. Dolan R, Pinkas H, Brady PG. Acute cholecystitis after palliative stenting for malignant
obstruction of the biliary tree. Gastrointest Endosc 1993;39:4479.
63. Ainley CC, Williams SJ, Smith AC, et al. Gallbladder sepsis after stent insertion for bile duct
obstruction: Management by percutaneous cholecystostomy. Br J Surg 1991;78:9613.
64. Johanson JF, Schmalz MJ, Geenen JE. Incidence and risk factors for biliary and pancreatic
stent migration. Gastrointest Endosc 1992;38:3416.

65. Ruffolo TA, Lehman GA, Sherman S, et al. Biliary stent migration with colonic diverticular
impaction. Gastrointest Endosc 1992;38:813.
66. Tarnasky PR, Cotton PB, Baillie J, et al. Proximal migration of biliary stents: Attempted
endoscopic retrieval in forty-one patients. Gastrointest Endosc 1995;42:51320.
67. Brown KA, Carpenter S, Barnett JL, Williams DM. Proximal migration of a biliary stent:
Treatment by combined percutaneous/endoscopic approach. Gastrointest Endosc
1995;41:6112.
68. Novis B, Pomeranz I, Zissin R, Rathaus V. Breakage of an endoscopically placed bile duct
stent. Gastrointest Endosc 1991;37:5823.
69. Donahue DG, Saltzman JR, Krims P. Stent fracture in malignant biliary obstruction [Letter].
70. Mallat A, Saint-Marc Girardin MF, Meduri B, et al. Fracture of biliary endoprosthesis after
endoscopic drainage for malignant biliary obstruction. Report of two cases. Endoscopy
1986;18:2434.
71. Csendes A, Fernandez M, Uribe P. Bacteriology of the gallbladder bile in normal subjects. Am
J Surg 1975;129:62931.
72. Dye M, MacDonald A, Smith G. The bacterial flora of the biliary tract and liver in man. Br J
Surg 1978;65:2857.
73. Sung JY, Costerton JW, Shaffer EA. Defense system in the biliary tract against bacterial
infection. Dig Dis Sci 1992;37:68996.
74. Sung JY, Shaffer EA, Olson ME, et al. Bacterial invasion of the biliary system by way of the
portal-venous system. Hepatology 1991;14:3137.
75. Feretis CB, Contou CT, Manouras AJ, et al. Long term consequences of bacterial colonization
of the biliary tract after choledochostomy. Surg Gynecol Obstet 1984;159:3636.
76. Gregg JA, De Girolami P, Carr-Locke DL. Effects of sphincteroplasty and endoscopic
sphincterotomy on the bacteriologic characteristics of the common bile duct. Am J Surg
1985;149:66871.
77. Sung JY, Leung JW, Shaffer EA, et al. Ascending infection of the biliary tract after surgical
sphincterotomy and biliary stenting. J Gastroenterol Hepatol 1992;7:2405.
78. Speer AG, Cotton PB, Rode J, et al. Biliary stent blockage with bacterial biofilm. A light and
electron microscopy study. Ann Intern Med 1988;108:54653.
79. Leung JW, Ling TK, Kung JL, Vallance-Owen J. The role of bacteria in the blockage of biliary
stents. Gastrointest Endosc 1988;34:1922.
80. Moesch C, Sautereau D, Cessot F, et al. Physicochemical and bacteriological analysis of the
contents of occluded biliary endoprostheses. Hepatology 1991;14:11426.
81. Groen AK, Out T, Huibregtse K, et al. Characterization of the content of occluded biliary
endoprostheses. Endoscopy 1987;19:579.
82. Costerton JW, Lewandowski Z, DeBeer D, et al. Biofilms, the customized microniche. J
Bacteriol 1994;176:213742.
83. Coene PP, Groen AK, Cheng J, et al. Clogging of biliary endoprostheses: A new perspective.
Gut 1990;31:9137.
84. Dowidar N, Kolmos HJ, Lyon H, Matzen P. Clogging of biliary endoprostheses: A morphologic
and bacteriologic study. Scand J Gastroenterol 1991;26:113744.
85. Smit JM, Out MM, Groen AK, et al. A placebo-controlled study on the efficacy of aspirin and
doxycycline in preventing clogging of biliary endoprostheses. Gastrointest Endosc
1989;35:4859.
86. Ghosh S, Palmer KR. Prevention of biliary stent occlusion using cyclical antibiotics and
ursodeoxycholic acid. Gut 1994;35:17579.
87. Barrioz T, Ingrand P, Besson I, et al. Randomised trial of prevention of biliary stent occlusion
by ursodeoxycholic acid plus norfloxacin. Lancet 1994;344:5812.
88. Sung JY, Shaffer EA, Costerton JW. Antibacterial activity of bile salts against common biliary
pathogens. Effects of hydrophobicity of the molecule and in the presence of phospholipids.
Dig Dis Sci 1993;38:210412.

89. Sung JY, Shaffer EA, Lam K, et al. Hydrophobic bile salt inhibits bacterial adhesion on biliary
stent material. Dig Dis Sci 1994;39:9991006.
90.
Geoghegan JG, Branch MS, Costerton JW, et al. Biliary stents occlude earlier if the distal tip
is in the duodenum in dogs. Gastrointest Endosc 1991;37:257.
91. Seitz U, Vadeyar H, Soehendra N. Prolonged patency with a new-design Teflon biliary
prosthesis. Endoscopy 1994;26:47882.
92. McAllister EW, Carey LC, Brady PG, et al. The role of polymeric surface smoothness of biliary
stents in bacterial adherence, biofilm deposition, and stent occlusion. Gastrointest Endosc
1993;39:4225.
93. Jansen B, Goodman LP, Ruiten D. Bacterial adherence to hydrophilic polymer-coated
polyurethane stents. Gastrointest Endosc 1993;39:6703.
94. Hoffman BJ, Cunningham JT, Marsh WH, et al. An in vitro comparison of biofilm formation on
various biliary stent materials. Gastrointest Endosc 1994;40:5813.
95. Leung JWC, Lau GTC, Sung JJY, Costerton JW. Decreased bacterial adherence to
silver-coated stent material. An in vitro study. Gastrointest Endosc 1992;38:33840.
96. Sung JJ, Chung SC, Tsui CP, et al. Omitting side-holes in biliary stents does not improve
drainage of the obstructed biliary system: A prospective randomized trial. Gastrointest Endosc
1994;40:3215.
97. Matsuda Y, Shimakura K, Akamatsu T. Factors affecting the patency of stents in malignant
biliary obstructive disease: Univariate and multivariate analysis. Am J Gastroenterol
1991;86:8439.
98. Libby ED, Leung JW. Prevention of biliary stent clogging: A clinical review. Am J
99. Nishioka NS, Kelsey PB, Kibbi AG, et al. Recanalization of occluded biliary endoprostheses
with pulsed laser radiation. Lasers Surg Med 1987;7:3913.
100. Speer AG, Cotton PB, MacRae KD. Endoscopic management of malignant biliary obstruction.
Stents of 10 FG are preferable to stents of 8 FG. Gastrointest Endosc 1988;34:4127.
101. Dowidar N, Moesgaard F, Matzen P. Clogging and other complications of endoscopic biliary
endoprostheses. Scand J Gastroenterol 1991;26:11326.
102. Pedersen FM. Endoscopic management of malignant biliary obstruction. Is stent size of 10
French gauge better than 7 French gauge?. Scand J Gastroenterol 1993;28:1859.
103. Kadakia SC, Starnes E. Comparison of 10 French gauge stent with 11.5 French gauge stent
in patients with biliary tract diseases. Gastrointest Endosc 1992;38:4549.
104. Frakes JT, Johanson JF, Stake JJ. Optimal timing for stent replacement in malignant biliary
tract obstruction. Gastrointest Endosc 1993;39:1647.
105. Soehendra N, Maydeo A, Eckmann B, et al. A new technique for replacing an obstructed
biliary endoprosthesis. Endoscopy 1990;22:2712.
106. Sherman S, Hawes RH, Uzer MF, et al. Endoscopic stent exchange using a guide wire and
mini-snare. Gastrointest Endosc 1993;39:7949.
107. Tarnasky PR, Morris J, Hawes RH, et al. Snare beside-a-wire biliary stent exchange: A
method that maintains access across biliary strictures. Gastrointest Endosc 1996;44:1857.
108. Huibregtse K, Carr-Locke DL, Cremer M, et al. Biliary stent occlusion: A problem solved with
self-expanding metal stents?. Endoscopy 1992;24:3914.
109. Davids PH, Groen AK, Rauws EA, et al. Randomised trial of self-expanding metal stents
versus polyethylene stents for distal malignant biliary obstruction. Lancet 1992;340:148892.
110.
Carr-Locke DL, Connors PJ, Cotton PB, et al. Interim results of a multicenter randomized trial:
Wallstent biliary endoprosthesis versus plastic stents. Gastrointest Endosc 1992;38:A248.
111. Neuhaus H, Hagenmuller F, Griebel M, Classen M. Percutaneous cholangioscopic or
transpapillary insertion of self-expanding biliary metal stents. Gastrointest Endosc
1991;37:317.
112. Vitale GC, Larson GM, George M, Tatum C. Management of malignant biliary stricture with
self-expanding metallic stent. Surg Endosc 1996;10:9703.
113. Ee H, Laurence BH. Haemorrhage due to erosion of a metal biliary stent through the duodenal
wall. Endoscopy 1992;24:4312.
114. Marano BJ Jr, Bonanno CA. Metallic biliary endoprosthesis causing duodenal perforation and
acute upper gastrointestinal bleeding. Gastrointest Endosc 1994;40:2578.
115. Fockens P, Waxman I, Davids PH, et al. Early recurrence of obstructive jaundice after
placement of a self-expanding metal endoprosthesis. Endoscopy 1992;24:42830.
116. Hausegger KA, Kleinert R, Lammer J, et al. Malignant biliary obstruction: Histologic findings
after treatment with self-expandable stents. Radiology 1992;185:4614.
117. Cremer M, Deviere J, Sugai B, Baize M. Expandable biliary metal stents for malignancies:
Endoscopic insertion and diathermic cleaning for tumor ingrowth. Gastrointest Endosc
1990;36:4517.
118. O'Brien S, Hatfield AR, Craig PI, Williams SP. A three year follow up of self expanding metal
stents in the endoscopic palliation of longterm survivors with malignant biliary obstruction. Gut
1995;36:61821.
119. Van Berkel AM, Bergman JJ, Waxman I, et al. Wallstents for metastatic biliary obstruction.
Endoscopy 1996;28:41821.
120. Glaser M, Laurence BH, Cameron FG. Relief of tumorous obstruction of a metal biliary stent
with palliative intraluminal iridium-192 therapy. Gastrointest Endosc 1992;38:4968.
121. Mixon T, Goldschmid S, Brady PG, Boulay J. Endoscopic management of expandable
metallic biliary stent occlusion. Gastrointest Endosc 1993;39:824.
122. Silvis SE, Sievert CE Jr, Vennes JA, et al. Comparison of covered versus uncovered wire
mesh stents in the canine biliary tract. Gastrointest Endosc 1994;40:1721.
123. Cotton PB. Metallic mesh stentsIs the expanse worth the expense?. Endoscopy
1992;24:4213.
124. Knyrim K, Wagner HJ, Pausch J, et al. A prospective randomized, controlled trial of metal
stents for malignant obstruction of the common bile duct. Endoscopy 1993;25:20712.
125. McLean GK, Burke DR. Role of endoprostheses in the management of malignant biliary
obstruction. Radiology 1989;170:9617.
126. Dick R, Platts A, Gilford J, et al. The Carey-Coons percutaneous biliary endoprosthesis: A
three-center experience in 87 patients. Clin Radiol 1987;38:1758.
127. Speer AG, Cotton PB, Russell RCG, et al. Randomised trial of endoscopic versus
percutaneous stent insertion in malignant obstructive jaundice. Lancet 1987;2:5762.
128. Roddie ME, Adam A. Metallic stents in biliary disease. Baillieres Clin Gastroenterol
1992;6:34154.
129. Salomonowitz EK, Adam A, Antonucci F, et al. Malignant biliary obstruction: Treatment with
self-expandable stainless steel endoprosthesis. Cardiovasc Intervent Radiol 1992;15:3515.
130. Gordon RL, Ring EJ, Laberge JM, Doherty MM. Malignant biliary obstruction: Treatment with
expandable metallic stentsFollow-up of 50 consecutive patients. Radiology
1992;182:697701.
131. Gilliams A, Dick R, Dooley JS, et al. Self-expandable stainless steel braided endoprosthesis
for biliary strictures. Radiology 1990;174:13740.
132.
Adam A, Chetty N, Roddie M, et al. Self-expandable stainless steel endoprostheses for
treatment of malignant bile duct obstruction. AJR 1991;156:3215.
133. Yoshioka T, Sakaguchi H, Yoshimura H, et al. Expandable metallic biliary endoprostheses:
Preliminary clinical evaluation. Radiology 1990;177:2537.
134. Lammer J. Biliary endoprosthesis, plastic versus metal stents. Radiol Clin North Am
1990;28:121122.
135.
Knyrim K, Wagner HJ, Starck E, et al. A prospective controlled randomized trial of expandable
metal stents in malignant biliary obstruction. Gastroenterology 1992;102:A318.
136.
Jackson JE, Roddie ME, Chetty N, et al. The management of occluded metallic
self-expandable biliary endoprostheses. AJR 1991;157:2912.
137. Kubota Y, Seki T, Yamaguchi T, et al. Bilateral internal drainage of biliary hilar malignancy via
a single percutaneous track. Role of percutaneous transhepatic cholangioscopy. Endoscopy
1992;24:1948.
138. Sarr MG, Cameron JL. Surgical palliation of unresectable carcinoma of the pancreas. World J
Surg 1984;8:90618.
139. Smith AC, Dowsett JF, Russell RC, et al. Randomised trial of endoscopic stenting versus
surgical bypass in malignant low bile duct obstruction. Lancet 1994;344:165560.
140. Shimi S, Banting S, Cushieri A. Laparoscopy in the management of pancreatic cancer:
Endoscopic cholecystojejunostomy for advanced disease. Br J Surg 1992;79:3179.
141. Rhodes M, Nathanson L, Fielding G. Laparoscopic biliary and gastric bypass: A useful adjunct
in the treatment of carcinoma of the pancreas. Gut 1995;36:77880.
142. Nagy A, Brosseuk D, Hemming A, et al. Laparoscopic gastroenterostomy for duodenal
obstruction. Am J Surg 1995;169:53942.
143. Bismuth H, Castaing D, Traynor O. Resection or palliation: Priority of surgery in the treatment
of hilar cancer. World J Surg 1988;12:3947.
144. Andersen JR, Sorensen SM, Kruse A, et al. Randomized trial of endoscopic endoprosthesis
versus operative bypass in malignant obstructive jaundice. Gut 1989;30:11325.
145. Shepherd HA, Royle G, Ross AP, et al. Endoscopic biliary endoprosthesis in the palliation of
malignant obstruction of the distal common bile duct: A randomised trial. Br J Surg
1988;75:11668.
146. Bornman PC, Tobias R, Harries-Jones EP, et al. Prospective controlled trial of transpapillary
biliary endoprostheses versus bypass surgery for unresectable carcinoma of the head of the
pancreas. Lancet 1986;1:6971.
147. Little JM. Hilar biliary cancerAre we getting it right?. HPB Surg 1989;1:936.
148. Nordlinger B, Jeppsson B, el-Khoury W, et al. Tumours of Oddi: Diagnosis and surgical
treatment. HPB Surg 1992;5:12331; discussion 1313.
149. Seifert E, Schulte F, Stolte M. Adenoma and carcinoma of the duodenum and papilla of Vater:
A clinicopathologic study. Am J Gastroenterol 1992;87:3742.
150. Komorowski RA, Beggs BK, Geenen JE, et al. Assessment of ampulla of Vater pathology: An
endoscopic approach. Am J Surg Pathol 1991;15:118896.
151. Bismuth H, Nakache R, Diamond T. Management strategies in resection for hilar
cholangiocarcinoma. Ann Surg 1992;215:318.
152. Wilker DK, Izbicki JR, Rohloff R. Is aggressive surgical palliation of proximal bile duct cancer
with involvement of both main hepatic ducts worthwhile?. HPB Surg 1992;5:23549.
153. Hadjis NS, Blenkharn JI, Alexander N, et al. Outcome of radical surgery in hilar
cholangiocarcinoma. Surgery 1990;107:597604.
154. Boerma EJ. Research into the results of resection of hilar bile duct cancer. Surgery
1990;108:57280.
155. Reding R, Buard JL, Lebeau G, Launois B. Surgical management of 552 carcinomas of the
extrahepatic bile ducts (gallbladder and periampullary tumors excluded). Results of the
French Surgical Association survey. Ann Surg 1991;213:23641.
156. Van der Heyde MN, Verbeek PCM, Lygidakis NJ. Decision making in the treatment of patients
with malignant proximal bile duct obstruction. Semin Liver Dis 1990;10:1458.
157. Altaee MY, Johnson PJ, Farrant JM, Williams R. Etiology and clinical characteristics of
peripheral and hilar cholangiocarcinoma. Cancer 1991;68:20515.
158. Pitt AP, Gomes AS, Lois JF, et al. Does pre-operative percutaneous biliary drainage reduce
operative risk or increase hospital cost?. Ann Surg 1985;201:54553.
159. Karsten TM, Allema JH, Reinders M, et al. Preoperative biliary drainage, colonisation of bile
and postoperative complications in patients with tumours of the pancreatic head: A
retrospective analysis of 241 consecutive patients. Eur J Surg 1996;162:8818.
160. . The Gastrointestinal Tumor Study Group: Further evidence of effective adjuvant combined
radiation and chemotherapy following curative resection of pancreatic cancer. Cancer
1987;59:200610.
161. Ede RJ, Williams SJ, Hatfield ARW, et al. Endoscopic management of inoperable
cholangiocarcinoma using iridium-192. Br J Surg 1989;76:8679.
162. Ker CG, Sheen PC, Arcillia CE. Choledochoscopic heat-probe therapy: An adjunctive
palliative treatment for intrahepatic cholangiocarcinoma with hepatolithiasis. Kao Hsiung I
Hsueh Ko Hsueh Tsa Chih 1991;7:34550.
163. Liguory C, Vitale GC. Biliary perestroika. Am J Surg 1990;160:2378.
164. Wagner HJ, Knyrim K, Vakil N, Klose KJ. Plastic endoprostheses versus metal stents in the
palliative treatment of malignant hilar biliary obstruction. A prospective and randomized trial.
Endoscopy 1993;25:2138.
165. Goldin E, Beyar M, Safra T, et al. A new self-expandable and removable metal stent for biliary
obstructionA preliminary report. Endoscopy 1993;25:5979.
166. Hoepffner N, Foerster EC, Hogemann B, Domschke W. Long-term experience in Wallstent
therapy for malignant choledochal stenosis. Endoscopy 1994;26:597602.
167. Schofl R, Brownstone E, Reichel W, et al. Malignant bile-duct obstruction: Experience with
self-expanding metal endoprostheses (Wallstents) in Austria. Endoscopy 1994;26:5926.
Chapter 65 Diagnosis and Management of Nonneoplastic Biliary

Obstruction, Biliary Leakage, and Disorders of the Liver Affecting the
Bile Ducts
(6758)
(6759)
KEES HUIBREGTSE, M.D., PH.D.
JOHN MEENAN, M.D., PH.D., M.R.C.P.
E. A. J. RAUWS, M.D.
VINOOL K. PARASHER, M.D.
Benign biliary strictures occur in a wide spectrum of conditions, arising from a disease process or as
the result of attempts at treatment. The endoscopic management of such strictures borrows heavily
from work in the field of interventional radiology and seeks to provide a flexibility in approach that is not
possible with either percutaneous or surgical therapy.
This chapter deals only with the endoscopic management of the major forms of benign biliary stricture.
There are many case reports of unusual causes of biliary obstructionfor example, venous
varicosities,1,2(6760) Crohn's disease,3(6761) peptic ulcer disease,4,5(6762) hemobilia,6(6763)
tuberculosis,7(6764) sarcoidosis,8(6765) and eosinophilic cholangiopathy9,10(6766)that are not
considered in detail. However, the principles of management are the same for all types of strictures.
Also, technical problems and complications relating to biliary stents, including clogging and migration,
are not discussed in this chapter (see Chapter 64: Diagnosis and Management of Malignant Biliary
Obstruction).
Perhaps the single greatest cause of benign biliary strictures is laparoscopic cholecystectomy (LC).
Although in experienced hands this technique is both safe and effective, the novice can wreak havoc.
With the potential widening of indications for the use of this form of surgery, an ability to treat such
complications is every bit as important for the endoscopist as the ability to treat choledocholithiasis.
Primary sclerosing cholangitis (PSC) remains a challenge for the endoscopist. Although the use of
biliary stenting and irrigation does hold some promise, this is at the cost of a significant risk of
precipitating cholangitis. Although new strategies will undoubtedly unfold, the potential for significant
progress in therapy of PSC probably lies in the hands of the immunologist.
Iatrogenic Injury
After Cholecystectomy
Laparoscopic methods for removal of the gallbladder have largely supplanted the traditional open
cholecystectomy in both the United States and Europe.11,12(6767) In comparison with open
cholecystectomy, the laparoscopic approach is associated with a shorter hospital stay and an earlier
return to normal activities.13(6768) However, the widespread adoption of laparoscopic biliary surgery
carries in its wake a greater incidence of bile duct injury.
Open cholecystectomy is associated with injury to the common bile duct (CBD) in 0.5% of
cases.14,15(6769) In a review of both published series and audit reports, Macintyre and
Wilson16(6770) found that bile duct injury occurred in from 0 to 2% and 0.5 to 0.9%, respectively, of
LCs. The importance of these differences in incidence between the laparoscopic and the open
procedures is reflected in the potentially detrimental impact of this complication on the life of the
affected patient.17(6771)
A prospective study based on a registry of all LCs performed in Denmark found that a bile duct injury
occurred in 57 of 7654 cases (0.74%).18(6772) Year to year, there was no change in the incidence of
complications. The incidence of complications was higher (1.3%) when acute cholecystitis was the
indication for the procedure. Biliary leaks, the majority from the cystic duct stump, were encountered in
2.1% of cases. Cholangiography, whether preoperative or intraoperative, did not appear to reduce the
frequency of complications.
For the most part, ductal damage, whether leakage, stricture formation, or complete occlusion, is not
recognized at the time of surgery but presents some days to years later with abdominal pain (59 to
71%), icterus (59 to 71%), cholangitis (29 to 45%), anorexia (43%), vomiting or nausea (36%), and
ileus (29%).17,1922(6773) Elevated alkaline phosphatase and bilirubin have been reported to
accompany all such injuries.19(6774) However, the presence of a biliary leak proximal to a complete
ductal obstruction or the development of a fistula may distort this picture.
Although the range of postsurgical biliary complications is wide, we have classified our patients
referred for endoscopic retrograde cholangiopancreatography (ERCP) into four major types: Type A,
leakage from cystic duct or peripheral hepatic radicles; Type B, leakage from a major bile duct; Type
C, isolated ductal stricture; and Type D, complete transection of the bile duct.23(6775) ERCP has been
shown to be an effective method of diagnosis and treatment in many patients who sustain bile duct
injuries during surgery. Types A and B in our experience are the ones most amenable to endoscopic
therapy.
Bile Duct Leak
The consensus is that biliary leaks are more common after LC than after the traditional open
procedure. Exact data on the frequency of this complication are difficult to obtain and are likely to vary
in relation to the experience of the operating surgeon. Complications encountered at laparoscopy, such
as excessive bleeding, were found in one study to be associated with an increased risk for
postoperative leakage.24(6776) However, leakage is also known to arise after seemingly
uncomplicated operations.25(6777) The calculated incidence in the study of Barkun et al.24(6778) was
1.1%, but data from open procedures were included in this analysis. Raijman et al.26(6779) identified a
leak in 17 of 465 patients (3.7%) undergoing LC, Sefr et al.27(6780) in 8 of 1223 cases (0.65%), and
Albasini et al.28(6781) in 10 of 500 consecutive patients (5%) who underwent the laparoscopic
procedure.
The clinical presentation of patients with biliary leaks is variable, but typically symptoms do not appear
until several days after the operation. In the study of Barkun et al.,24(6782) patients presented on
average at 5 days after surgery; symptoms included abdominal pain (89%), tenderness (81%), and
fever (74%). Several imaging studies can be utilized to evaluate patients with suspected biliary leaks.
Ultrasound may reveal presumptive evidence in the form of a fluid collection in the right upper
quadrant. Nuclear cholescintigraphy has been found to be highly sensitive, specific, and accurate in the
diagnosis of biliary leakage.29(6783) Although these noninvasive imaging methods offer valuable
information, retrograde cholangiography provides precise details on the site of the leak and the
presence or absence of additional problems such as distal obstruction, and it affords an opportunity for
immediate treatment.
The majority of bile duct leaks after LC arise from the cystic duct stump (Figure
651).24,26,3032(6784) However, leakage may occur from virtually any segment of the extrahepatic
biliary tract as well as the ducts of Luschka; in a few cases, the site cannot be identified. The approach
to treatment has been highly variable and dependent on the clinical status of the patient as well as the
nature and location of the leak. Spontaneous resolution has been described in patients with external
drains.27,28(6785) In clinically stable patients without evidence of sepsis, percutaneous drainage of
fluid collections has resulted in resolution of symptoms and presumably closure of the leak. However,
more interventional approaches to treatment, including reoperation and endoscopic therapy,
sometimes in combination, are required in a large number of patients.
(6786)Figure 651. A, Cystic stump leakage after laparoscopic cholecystectomy. B, A short

10-French endoprosthesis has been placed in the distal common bile duct (CBD). C, After
removal of the endoprosthesis at 6 weeks, no further leakage could be seen.
Bile duct obstruction is thought to promote the development of a bile leak after cholecystectomy.
Barkun et al.24(6787) found an obstructing process, most commonly stones, in 31% of cases. This has
obvious relevance with respect to the choice of therapy.
A wide variety of therapeutic endoscopic maneuvers have been used to treat bile
leaks,24,25,3033(6788) including sphincterotomy alone, sphincterotomy with stent placement
(crossing the stricture or just the main duodenal papilla), and placement of a nasobiliary drain. Despite
this variability, endoscopic treatment has in general been remarkably successful, with closure of the
leak achieved in better than 90% of properly selected patients. Such favorable results would suggest
that ERCP is the procedure of choice for diagnosis and management when a biliary leak is suspected.
However, endoscopic therapy is probably most successful in the treatment of cystic duct stump leaks
(the majority of cases). More complex and serious injuries to the extrahepatic ducts themselves are
more difficult to manage endoscopically, and in such patients, surgical treatment is often more
appropriate.
The choice of endoscopic therapy is problematic, as there are no data to indicate the most efficacious
approach. Because it is thought that bile duct obstruction, whether actual or relative, is a factor in the
development of a biliary leak after cholecystectomy, it is logical to think that establishment of adequate
drainage will promote healing of the leak. In fact, it is well documented that many patients respond to
endoscopic sphincterotomy alone. However, the actual mechanisms that promote leakage are
probably more complex. For example, cystic duct stump leaks have been described in patients who
had undergone endoscopic sphincterotomy before cholecystectomy.34(6789) In cases where an actual
obstruction, such as a stone or stricture, is not identified, the least invasive method of treatment would
be placement of a stent without sphincterotomy. Although there are no data on a possible role for
sphincter of Oddi dysfunction in the development of bile leakage, it is possible that simple placement of
a small-diameter stent across the main duodenal papilla will be adequate treatment. However, the
majority of investigators appear to consider it advisable to bridge the site of the leak with a stent. There
are no data on the ideal duration of stenting; in most series, stents have been left in situ for periods of
1 to 2 months.
There are various other causes of biliary leaks in addition to surgical procedures. There are reports of
successful endoscopic therapy in patients with biliary fistulas due to trauma. The appropriateness of
endoscopic therapy in such cases depends on the clinical status of the patient as well as the nature
and location of the biliary injury.35,36(6790) Endoscopic therapy including stent placement was used to
close a fistula between a transjugular intrahepatic portosystemic shunt (TIPS) and the biliary ducts in a
patient who developed polymicrobial gram-negative sepsis after the TIPS procedure.37(6791)
Stricture
The majority of benign bile duct strictures occur as a complication of cholecystectomy.38(6792) Most
are short, that is, less than 10 mm in length, and distal to the confluence of right and left hepatic ducts,
as described by the Bismuth classification (Figures 652, 653 and 654).39,40(6793)
(6794)Figure 652. Complete occlusion of the CBD after erroneous clip placement at
laparoscopic cholecystectomy.
(6795)Figure 653. Stenosis of the CBD consequent to damage at laparoscopic

cholecystectomy.
(6796)Figure 654. Serial radiographs show leakage of contrast from the CBD after a Billroth
type II gastrectomy (A), with subsequent stricture (B) and stone formation (C).
Most bile duct injuries arising during LC result from poorly defined anatomy.17,41,42(6797) Bleeding
within the liver hilum represents a particular danger, as attempts are made to staunch the flow with
clips, laser, or diathermy.43(6798) Inadvertent injury may also result from the use of diathermy (stray
current, coupling, and sparking) and laser (backstop effect).4446(6799) Delayed injury arising from
ischemia of the bile ducts, in particular of the right hepatic duct, is well recognized.47(6800)
The approximate distribution of bile duct strictures after LC is as follows: mid CBD (42 to 50%),
confluence area (22 to 41%), common hepatic duct (28%), and distal CBD (15%). Lesions at the level
of the cystic duct account for about 20% of cases.19,48(6801) Benign biliary strictures have been
classified by Bismuth39(6802) in relation to distance from the confluence of the right and left hepatic
ducts. Options for therapy include surgery and percutaneous or endoscopic hydrostatic balloon dilation
with placement of an endoprosthesis.
The outcome of surgery for benign biliary strictures is good in 75 to 93% of patients.4850(6803)
Strictures recur in 18% of patients,48(6804) and when a subsequent repair is undertaken, a further
recurrence develops in 26%.50(6805) The majority of stricture recurrences develop within 7 years of
surgery.51(6806) Reported mortality rates are in the range of 3.2 to 27%,4850(6807) the higher rate
being related to patients with coexisting pathology such as portal hypertension. Factors that are
associated with a favorable outcome include greater distance from the confluence of the right and left
hepatic ducts, early referral, no previous repair, and the quality of the proximal duct.52,53(6808) The
use of transhepatic Silastic tubes to facilitate surgical reconstruction has improved the results of
surgery.54,55(6809)
Percutaneous biliary dilation via transhepatic puncture, T-tube, or jejunal loop56(6810) has an
associated morbidity of less than 7%57,58(6811) and a reported success rate of 33 to 100%,59(6812)
with patency being maintained in one series in 76% of patients at 36 months follow-up.57(6813) In
general, such therapy requires several sessions in order to attain a satisfactory outcome, although Lee
et al.60(6814) advocate single-session therapy. The major concerns with the transhepatic approach
are the attendant risks of hemorrhage and bile leakage associated with liver puncture. Additionally,
two-thirds of patients may have a nondilated biliary tract, making ductal puncture technically
difficult.61(6815) A further disadvantage rests in the requirement for long-term transhepatic intubation.
The role of ERCP in the care of patients with LC complications of stone retention, ductal leakage, or
ductal stricture is well established.6265(6816) Endoscopic hydrostatic balloon dilation, with and
without endoprosthesis placement, for post-LC biliary strictures has met with some success. Some
preliminary data for endoscopic balloon dilation alone appeared favorable,66(6817) but this was not
confirmed in other studies.58(6818) An early study by our group found that a combination of balloon
dilation and insertion of a 10-French polyethylene endoprosthesis, yielded satisfactory results in 21 of
27 patients during a follow-up period of 18 months.67(6819) Other investigators obtained similar
results with this combination therapy.19,68(6820) From our group, Davids et al.40(6821) evaluated the
efficacy of endoscopic stenting in 70 patients with incomplete strictures after cholecystectomy and
showed that stent placement could be achieved in 94% of patients, with good or excellent clinical
responses noted in 83% (Table 651). Early complications (minor bleeding from sphincterotomy site,
cholangitis, and pancreatitis) occurred in 9% of patients. A retrospective study comparing surgical with
endoscopic therapy concluded that surgery and endoscopy have similar long-term success rates, with
recurrences being seen in 17% of patients (Table 652).69(6822) These data suggest that surgery
should be reserved for those patients with complete duct transection, failed previous repairs, and failed
endoscopic therapy.
Endoscopic Management of Benign

Postoperative Biliary Strictures*
TABLE 651
Months follow-up (mean, range)

Recurrent stricture
38
8
Endoscopic management
Surgery
(496)
(17%)
3
5
Calculus formation
(4%)
6
Mortality (nonbiliary)
Data from Davids PHP, Rauws EAJ, Coene PPLO, et al. Endoscopic stenting for
postoperative biliary strictures. Gastrointest Endosc 1992; 38:128.
* Data refer to long-term results in 46 patients after removal of endoprotheses.
No. of patients.
Comparison Between Endoscopic and Surgical

Treatment of Benign Postoperative Biliary Strictures
TABLE 652
Number of patients
Early complications
Mortality*
Late complications
ENDOSCOPY
SURGERY
66
8
1
27
35
26
0
0
46
35
Outcome (number of patients)
Excellent
33 (71%)
25 (71%)
Comparison Between Endoscopic and Surgical

Treatment of Benign Postoperative Biliary Strictures
TABLE 652
ENDOSCOPY
SURGERY
Good
5 (10%)
4 (11%)
Poor
8 (17%)
6 (17%)
Data from Davids PHP, Tanka AK, Rauws EAJ, et al. Benign biliary strictures:
Surgery or endoscopy. Ann Surg 1993; 217:23743.
* Mortality figures are for the 30-day period after the procedure.
Late complications refer to those occurring within 1 year of the procedure.
The outcome results in the endoscopy therapy group pertain to those patients
from whom endoprostheses have been finally removed.
Consequent to these studies, our routine practice is to introduce an endoprosthesis as the initial
therapy for postoperative partial stricturing of the biliary tree. At the first endoscopic session, a
sphincterotomy is made and a single 10-French plastic stent is seated over a placement catheter
containing an atraumatic guidewire. Should the stricture prove to be tight, dilation is performed with a
balloon catheter (4 to 8 mm) or a Soehendra dilating catheter. Subsequently, at 6 to 8 weeks, the first
stent is removed and replaced by two 10-French stents. Thereafter, these are exchanged for new
stents at 3-month intervals for a period of 12 months. We have found little benefit from prolonging this
course of therapy beyond 1 year. Antibiotics are given only when cholangitis is present or drainage is
not achieved. Should a complete ductal obstruction be noted at the first endoscopy (see Figure 652),
stenting is not attempted and the patient is referred for surgery.
The use of the metal expandable type of endoprosthesis in benign biliary disease remains
controversial. However, several groups report favorable results, including good outcomes for use of
this type of device in the management of postoperative strictures. Gianturco-Rsch Z stents were
placed percutaneously in 43 such patients by Coons.70(6823) All patients had previously undergone
an unsuccessful balloon cholangioplasty. The 1-year reocclusion rate was 13%. Maccioni et
al.71,72(6824) report long-term patency in 100% in patients with a CBD stricture. Foerster et
al.,73(6825) reporting on endoscopic Wallstent placement in four patients, identified no stent-related
complications and no cases of occlusion during a follow-up period of 53 weeks. Despite such results,
however, the benefits of the metal expandable endoprosthesis seem limited in light of the successful
results obtained with the use of plastic stents. Nevertheless, the advent of removable metal stents may
alter this situation in future years.74(6826)
Anastomotic Strictures
A variety of complications involving the biliary tract may arise in patients who undergo orthotopic liver
transplantation. Many of these are amenable to endoscopic therapy,7578(6827) including
anastomotic stricture, fistula, and stone formation. Additional causes of cholestasis include biliary
sludge and casts. Endoscopic therapy is usually not suitable for certain complications such as bile duct
necrosis, hemobilia, and extensive intrahepatic strictures. Biliary complications occurred in 19% of liver
transplant patients (n = 105) in the series of Theilmann et al.79(6828) Hintze et al.77(6829)
encountered biliary complications in 44 of 500 transplant patients (8.8%). In a review of published data,
Sossenheimer et al.80(6830) found that biliary tract complications have been reported in 13 to 34% of
patients who undergo liver transplantation. Biliary tract complications were responsible for major
morbidity and mortality in 10% of transplanted patients.
Stricture formation after liver transplantation and hepatic or pancreatic resection with reimplantation of
the bile ducts presents additional difficulties for the endoscopist.81,82(6831) Although short and
amenable to both dilation and stenting (Figure 655), such strictures may lie beyond the reach of the
endoscope. However, the fashioning of a jejunocutaneous fistula permits an endoscopic approach in
some cases (see also Chapter 71: Choledochofiberoscopy and Endoscopic Lithotripsy). Further
constraints on manipulations by surgeon, radiologist, and endoscopist may arise from anatomic
distortion and displacement of the hilar structures due to liver lobe atrophy and hypertrophy after
hepatic resection.83(6832) Additionally, stricture patency may be less easily achieved in patients with
biliary-enteric anastomoses.84(6833)
(6834)Figure 655. A, Stenotic biliary-enteric anastomosis fashioned using the cystic duct at
liver transplantation. B, Balloon dilation of the stricture. Indentation of the balloon by the
stricture is evident.
The possibility that some pathologic process other than a postoperative stricture accounts for the
clinical picture of biliary obstruction should always be entertained. Matthews et al.,85(6835) on review
of 24 cases of recurrent cholangitis after biliary-enteric anastomosis for benign disease, found that up
to one-third of patients had no evidence of stricturing involving the extrahepatic bile ducts. Intrahepatic
strictures, intrahepatic calculi, improperly constructed conduits, and potential bacterial overgrowth were
found in 42%, 25%, 13%, and 17% of patients, respectively. Multiple causes were identified in 71% of
patients.
Ward et al.86(6836) described nonanastomotic hilar biliary strictures. In some patients, the entire
donor extrahepatic duct may become stenotic.79(6837) The cause of these lesions is unknown,
although vascular insufficiency and cytomegalovirus infection have been implicated. Such lesions have
been managed by both percutaneous and endoscopic approaches.
Sclerosing Cholangitis
Primary Conditions
Primary Sclerosing Cholangitis
PSC is a chronic cholestatic liver disease that frequently results in bile duct obliteration, biliary
cirrhosis, hepatic failure, and in some cases, cholangiocarcinoma. It is often associated with ulcerative
colitis, although the course of the biliary disease is unrelated to the clinical course of the
colitis.87,88(6838) In a comparative study of patients with PSC, Kaw et al.89(6839) found that those
with biliary tract calculi were more likely to have biliary symptoms including abdominal pain, pruritus,
jaundice, and cholangitis.
Cholangiographically, PSC is characterized by multiple fibrotic strictures of both the intrahepatic and
the extrahepatic tree with sparing of the cystic duct and gallbladder (Figures 656 and
657).90,91(6840) The strictures are short, giving rise to a beaded appearance. Such
pseudodiverticula may be seen in 27% of patients but should not be interpreted as being
pathognomonic of PSC.92(6841) Craig et al.,93(6842) in correlating radiographic appearance with the
clinical course of the disease, noted that intrahepatic disease was of particular prognostic significance.
There was a significant reduction in 3-year survival rate if diffuse or high-grade (narrowing greater than
75%) intrahepatic duct strictures were present.
(6843)Figure 656. Primary sclerosing cholangitis (PSC). There is diffuse involvement of the
intrahepatic ducts where short segmental stenoses alternate with areas of normal caliber or
dilation to produce a beaded appearance. Involvement of the extrahepatic ducts has resulted in
mural irregularities including sacculations. (From the collection of Dr. M. V. Sivak, Jr.)
(6844)Figure 657. PSC with extrahepatic and intrahepatic ductal involvement. The
intrahepatic ducts were difficult to fully opacify, in part owing to the narrowing of the
extrahepatic ducts. Only the ventral pancreatic duct was opacified by injection of contrast
material at the papilla, indicating an incidental finding of pancreas divisum. Retrograde
injection at the minor papilla revealed a normal duct. (From the collection of Dr. M. V. Sivak,
Jr.)
Endoscopic therapy in PSC remains palliative, its aim being to improve bile flow. To this end, several
strategies including hydrostatic balloon dilation of strictures, removal of ductal debris, and
endoprosthesis or nasobiliary drain placement have been advocated. The interpretation of the results
of most studies of endoscopic therapy in PSC is difficult because of the small numbers of cases in
most trials and the unpredictable nature of the disease itself. Furthermore, therapy in most series has
been by means of various combinations of endoscopic maneuvers, thereby making it difficult to
evaluate any one form of treatment.
Endoscopic dilation and stenting with sphincterotomy has met with some success in patients with PSC.
Our group as well as others has obtained clinical response rates of 69 to 94%, with clinical
improvement after stent removal being maintained long-term in 50 to 69% of patients.9395(6845) In
all of these cases, the dominant stricture was extrahepatic in location.
Johnson et al.96(6846) suggested that the use of stents should be avoided in patients with PSC
because of a high incidence of associated cholangitis; these investigators favor hydrostatic balloon
dilation alone. Lombard et al.97(6847) eschew sphincterotomy because of the potential for
contamination of the bile ducts by enteric bacteria.
Placement of a nasobiliary drain has been used by several groups to flush the biliary tree with saline
solution or saline solution together with corticosteroids. The results of such therapy in a small number
of patients have been encouraging.98,99(6848) Wagner et al.100(6849) treated 12 symptomatic
patients with high-grade biliary strictures by hydrostatic balloon dilation followed by ductal perfusion
with normal saline solution by means of a nasobiliary catheter. The number of major strictures in
individual patients varied from 1 to 3 (mean 2.1). Strictures were graded before treatment was initiated.
Patients with high-grade strictures proximal to the left or right hepatic duct were excluded. Saline
solution perfusion was maintained for 1 week, at which time ERCP was repeated with further dilations.
Endoscopic procedures were repeated until all strictures were judged to be dilated satisfactorily.
Between two and nine procedures were required to reach this end point. Thereafter, follow-up ERCPs
(with dilation as needed) were performed at 3, 6, 12, 18, and 24 months. There were no
procedure-related complications. The average length of follow-up was 23 months, with a range of 12 to
50 months. There were significant and sustained improvements in clinical status, liver function tests,
and the scoring of strictures in 8 patients (67%). The disease was progressive in 3 patients; all
eventually underwent liver transplantation. One patient did not respond to treatment, but his clinical
condition remained stable.
Our current approach to the endoscopic treatment of patients with PSC is as follows: When a
significant extrahepatic stricture is identified, dilation is first performed using either a Gruntzig-type
balloon or a Soehendra dilator set (graded coaxial dilators). The biliary tree is then flushed with saline
solution to remove sludge or stones, and a 10-French plastic endoprosthesis is placed across the
stricture (Figure 658). Should it prove to be impossible to insert the larger-caliber stent, we insert
either a nasobiliary drain for additional flushing with saline solution or a 7-French endoprosthesis. In
such cases, ERCP is repeated some days later with insertion of a 10-French stent. In all cases, the
patient is admitted to hospital overnight for observation and broad-spectrum antibiotics are given
intravenously.
(6850)Figure 658. A, PSC. B, An endoprosthesis has been placed to dilate a dominant

stricture at the biliary bifurcation. After 1 week, this stent was removed.
We have retrospectively reviewed the results of our treatment program of prolonged stenting in 25
symptomatic patients with worsening biochemical tests of liver function and major extrahepatic bile
duct strictures.101(6851) Endoscopic therapy was technically successful in 21 patients (84%), and in
all of these, a significant improvement in liver function tests was noted within 6 months of insertion of
the prosthesis. Stents were thereafter exchanged electively at intervals of 2 to 3 months or
nonelectively in patients who developed cholangitis. Four patients underwent more than one period of
prolonged stenting. Complications occurred in relation to 14% of the 105 procedures. Median follow-up
after final stent removal was 29 months, with a range of 2 to 120 months. During follow-up, 12 patients
(57%) remained asymptomatic with stable liver function tests. Clinical and biochemical parameters
deteriorated in 4 patients (19%), but all responded to additional periods of endoscopic treatment.
Our results are comparable with those of Lee et al.102(6852) In this retrospective study of 53 patients,
endoscopic therapywhich included various combinations of dilation, stent placement, stone
extraction, and nasobiliary tube perfusionwas technically successful in 88%. Procedure-related
complications were encountered in 15 patients. Follow-up data, available in 50 of the 53 cases (median
follow-up 31 months), indicated that symptoms were lessened in 28 patients (56%). At 3 months
follow-up, there was a significant improvement in biochemical tests of liver function. No change was
evident in the cholangiographic appearance of the bile ducts in 51% of patients, although some
improvement was thought to have been achieved in 36%; in 13%, the cholangiographic appearance
could not be assessed. By combined assessment of clinical symptoms, liver function tests, and
cholangiograms, Lee et al.102(6853) concluded that some improvement had been achieved in 41 of 53
patients (77%).
In addition to their being at greater risk for the development of procedure-related cholangitis,96(6854)
patients with PSC may present other difficulties to the endoscopist. The initial cannulation of the CBD
may be made difficult because of retraction of the papilla owing to ductal fibrosis and shortening.
Additionally, the presence of sacculations in the bile duct wall increases the risk of false passage of
guidewires and perforation. In advanced cases, it may be difficult to achieve opacification of the
intrahepatic ducts, especially if the gallbladder acts as a low-pressure reservoir. Selective cannulation
of the left and right hepatic ducts may surmount this problem; in other cases, the use of an extraction
balloon to occlude the distal duct during injection ("occlusion cholangiogram") of contrast material may
provide adequate visualization of the intrahepatic ducts.
The association between PSC and cholangiocarcinoma adds yet another layer of complexity to
endoscopy in this patient population. Cholangiocarcinoma coexists with PSC in 7 to 36% of patients
diagnosed at autopsy or at transplantation.103106(6855)
Clinically, the onset of cholangiocarcinoma is suggested by rapid development of persistent jaundice,
pruritus, weight loss, and a rise in bilirubin.106(6856) The presence of carcinoma cannot be
distinguished by cholangiography in the absence of gross findings such as a polypoid mass greater
than 1 cm in diameter, markedly dilated ducts proximal to a stricture, or rapidly progressive stricture
formation.107,108(6857) However, brush cytology may identify up to 85% of cases (see also Chapter
64: Diagnosis and Management of Malignant Biliary Obstruction).109(6858) Cytologic samples should
therefore be obtained from all strictures amenable to brushing. The advent of small-diameter video
endoscopes suitable for cholangioscopy may make it possible to obtain guided forceps biopsies of
ductal strictures as an adjunct to brush cytology (see Chapter 70: Peroral Cholangioscopy and
Pancreatoscopy).110(6859) An index formula based on combined elevations of the serum tumor
markers carcinoembryonic antigen and carbohydrate antigen 199 was found to be 86% accurate for
diagnosis of cholangiocarcinoma in patients with PSC, although the accuracy was less if the tumor was
regarded as occult.111(6860)
Idiopathic Biliary Strictures
The occurrence of idiopathic, nontraumatic inflammatory extrahepatic bile duct strictures has been
reported by Caldwell et al.112(6861) and Standfield et al.113(6862) The latter group described 12 such
cases of benign strictures in patients who ranged in age from 1.5 to 60 years.113(6863) Histologic
examination showed chronic inflammation with ulceration. All patients responded well to surgery.
Secondary Sclerosing Cholangitis

Acquired Immunodeficiency Disease-Related
The clinical presentation of acquired immunodeficiency syndrome (AIDS)-related sclerosing cholangitis
is that of upper abdominal pain and elevated liver function tests.114116(6864) However, marked
elevation of serum alkaline phosphatase per se does not imply such a diagnosis.117(6865) There is a
close association between AIDS-related cholangitis and infection by cryptosporidia, cytomegalovirus,
and microsporidia.118120(6866) A case of necrotizing cholangitis due to infection by Septata
intestinalis, a newly recognized invasive microsporidial species, has been reported.121(6867) The
cholangiographic appearance of AIDS-related cholangitis is similar to that of PSC. Biliary fistula has
also been described.122(6868) Endoscopy remains the most appropriate method of investigation,
although radionucleotide scanning may have a role.123(6869) Ultrasound, which may identify various
types of abnormalities, has been found to be a useful adjunctive test in several reports.124126(6870)
Forbes et al.127(6871) reviewed the natural history of 20 cases of AIDS-related cholangiopathy. All
patients had pain, 13 had elevated serum levels of alkaline phosphatase, but only 3 had an elevated
bilirubin. The median CD4 count at diagnosis was 0.024 106/ml. Infection with cryptosporidiosis and
cytomegalovirus was found in 13 and 6 patients, respectively. Ultrasound was abnormal in 10 cases,
dilation of the bile duct and duct wall thickening being the most common findings. Cholangiography
was abnormal in all patients, with widespread abnormalities present in the majority; cystic lesions of the
CBD were noted in 2 cases. When compared with matched controls, the presence of AIDS-related
sclerosing cholangitis did not appear to have any influence on the progression of the underlying
disease. All patients were treated with analgesics, no endoscopic intervention being attempted, other
than a sphincterotomy in 1 patient.
Benhamou et al.128(6872) described two patterns of biliary lesions: (1) regular dilation of the
intrahepatic ducts associated with a distal, smooth CBD stricture and (2) diffuse, irregular stenoses of
the intrahepatic biliary tree with a distal extrahepatic stenosis. The latter pattern was noted particularly
in patients with concomitant cryptosporidial and cytomegaloviral infections. Kline et al.129(6873)
described protuberant, intrabiliary filling defects in association with cryptosporidial infection;
histopathologic evaluation of these defects revealed squamous metaplasia.
Stenosis of the main duodenal papilla has also been described in patients with AIDS (see Chapter 68:
Sphincter of Oddi Stenosis and Dysfunction).126,130(6874)
Strictures Related to Infections and Parasitic Infestation
Parasites are yet another cause of benign bile duct strictures, particularly in Southeast Asia (Figure
659).131135(6875) Changing demographic patterns owing to immigration have, however, introduced
this disease into the Western hemisphere. Oriental cholangiohepatitis is a syndrome characterized by
recurrent attacks of abdominal pain (often lasting for years), episodic cholangitis and jaundice, and
dilation of the intrahepatic and extrahepatic bile ducts with formation of pigmented stones. Clonorchis
sinensis and Ascaris lumbricoides are the most frequently found parasites; whether they have a direct
role in the causation of the disease remains controversial.132,133(6876) Bile duct strictures may be
seen in 10 to 35% of patients.131,134(6877) There is a preponderance of hilar and left hepatic duct
strictures, with common hepatic duct involvement being less frequent.131135(6878) Dilation of
intrahepatic ducts in the absence of extrahepatic duct obstruction is a pathognomonic finding.
Additionally, ERCP may reveal a filamentous filling defect, blunting of branches of the intrahepatic bile
ducts, and intrahepatic stone formation.131(6879) Treatment consists of eradication of the parasite
and surgical decompression with stone removal or excision of liver tissue should drainage be
inadequate. Percutaneous transhepatic approaches to treatment including endoscopic dilation of
strictures and stone extraction have been attempted in some patients such as those in whom surgery
was unsuccessful (see also Chapter 71: Choledochofiberoscopy and Endoscopic
Lithotripsy).134(6880) Stones and parasites have been removed at ERCP after sphincterotomy;
however, the role of endoscopy in the treatment of such biliary strictures remains undefined.135(6881)
(6882)Figure 659. Helminth infestation of the CBD and right biliary system.
Infestation with the liver fluke, Fasciola hepatica, causes a variety of symptoms including fever,
hepatomegaly, eosinophilia, and abdominal pain; patients may also be asymptomatic. The adult flukes
reside in the bile duct, where they can cause hyperplasia of the ductal epithelium. A variety of features,
including dilation, filling defects, and strictures, may be disclosed by imaging studies. Although this
infestation is at least partially responsive to several pharmacologic agents, several case reports outline
various roles for ERCP in its treatment.136139(6883)
Ramirez et al.140(6884) described cholangiographic findings that were similar to PSC that were due to
Coccidioides immitis in a 43-year-old man who had symptoms of abdominal pain, jaundice, and
pruritus. The patient was not immunocompromised, and his biliary stricture resolved with treatment that
included endoscopic placement of a stent and antifungal agents.
Iatrogenic
Ischemic bile duct strictures have been ascribed to hepatic arterial infusion of
5-fluorodeoxyuridine,141143(6885) the cholangiographic appearance mimicking that of PSC closely.
Therapy includes withdrawal of the drug and drainage of the dilated biliary tree. This may be achieved
percutaneously or surgically, although an endoscopic approach has been successful.143(6886)
Direct toxic injury to the biliary tree after the use of formaldehyde or alcohol solutions in the therapy of
hydatid disease has been described.144(6887)
Transient partial obstruction of the distal CBD is a well-recognized complication of acute pancreatitis. It
is well tolerated, and clinical signs of obstruction disappear as the pancreatic inflammation subsides
and, in some cases, with resolution of a pseudocyst. However, chronic pancreatitis is associated with
the development of persisting strictures that may result in episodes of cholangitis and secondary
hepatic cirrhosis (see also Chapter 73: Endoscopic Retrograde Cholangiopancreatography in Acute
and Chronic Pancreatitis).

Persisting biliary stricture accompanies chronic acalculous pancreatitis in up to 29% of cases in
adults.145,146(6888) In children, chronic pancreatitis being relatively uncommon, only 13 cases have
been reported.147(6889) The strictures tend to be relatively long; the mean length in one retrospective
study of 38 patients was 3.9 cm.148(6890) In a series of 40 patients with well-developed chronic
pancreatitis and bile duct stricture collected by Stabile et al.,149(6891) clinical manifestations included
abdominal pain (65%), icterus (43%), secondary biliary cirrhosis (15%), and cholangitis (15%). A
persistently elevated alkaline phosphatase was the most sensitive laboratory indicator of occult
obstruction. However, biochemical indices correlate poorly with cholangiographic
appearance;148,150(6892) in some patients, long strictures with proximal bile duct dilation were found
in the presence of normal liver function tests.
The management of patients with chronic pancreatitis and related bile duct strictures remains
controversial. It has been suggested that surgical intervention may be reserved for those with acute
cholangitis, protracted icterus, and secondary cirrhosis.148(6893) However, endoscopic stent
placement is relatively safe and raises the question of the most appropriate therapy in patients with
abnormal biochemical tests or a dilated bile duct.
The high morbidity (25%) associated with surgical correction of distal CBD strictures in alcoholic
patients has prompted attempts at endoscopic management of this problem.148,151(6894) Deviere et
al.152(6895) treated 25 symptomatic patients by insertion of a plastic endoprosthesis. Two patients
died 2 months after treatment. Short-term follow-up (7 to 42 months) in 19 patients revealed stent
migration in 10 and stent occlusion in 8 along with relapsing cholestasis. Longer follow-up (12 to 72
months) disclosed only 3 asymptomatic patients with a stent in place. All other patients required
surgical decompression.
We have treated 58 patients with biliary strictures due to chronic pancreatitis by insertion of a plastic
stent (Figure 6510).153(6896) Five patients (9%) developed complications related to the procedure.
In 16 patients (28%), there was regression of the stricture to the point that stents were permanently
removed. Of the 42 patients with persistent strictures, 16 underwent surgery and 26 were managed by
continued use of endoscopically placed stents. Stent-related complications, such as clogging, occurred
over the long-term in 64% of patients and proved to be the major limitation of this form of therapy.
(6897)Figure 6510. Chronic pancreatitis with calculus formation in the pancreatic duct. An
endoprosthesis has been placed in the CBD because of distal stricturing.
Because of the mixed results obtained with insertion of plastic endoprostheses, Deviere et
al.154(6898) used short (3.4 cm) metal self-expandable stents (Wallstent) in the management of 20
patients with persistent cholestasis. Eighteen patients remained well after a mean follow-up of nearly 3
years. Two patients required plastic stent placement, through the Wallstent, because of biliary
epithelial hyperplasia. Deviere et al.154(6899) suggest that the use of such stents should not
compromise any future surgery in these patients.
REFERENCES
1.
2.
3.
Takehara H, Komi N, Okada A, et al. Unusual cases of benign stricture of the biliary tract.
Tokushima J Exp Med 1992;39:13543.
Thervet L, Faulques B, Pissas A, et al. Endoscopic management of obstructive jaundice due
to portal cavernoma. Endoscopy 1993;25:4235.
Spiess SE, Braun M, Vogelzang RL, Craig RM. Crohn's disease of the duodenum
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
complicated by pancreatitis and common bile duct obstruction. Am J Gastroenterol

1992;87:10336.
Fowler CL, Sternquist JC. Choledochoduodenal fistula: A rare complication of peptic ulcer
disease. Am J Gastroenterol 1987;82:26971.
Misra MC, Grewal H, Kapur BM. Spontaneous choledochoduodenal fistula complicating peptic
ulcer diseaseA case report. Jpn J Surg 1989;19:3679.
Kroser J, Rothstein RD, Kochman ML. Endoscopic management of obstructive jaundice
caused by hemobilia. Gastrointest Endosc 1996;44:6189.
Fan ST, Ng IO, Choi TK, Lai EC. Tuberculosis of the bile duct: A rare cause of biliary stricture.
Rezeig MA, Fashir BM. Biliary tract obstruction due to sarcoidosis: A case report. Am J
Tenner S, Roston A, Lichtenstein D, et al. Eosinophilic cholangiopathy. Gastrointest Endosc
1997;45:3079.
Shimomura I, Takase Y, Matsumoto S, et al. Primary sclerosing cholangitis associated with
increased peripheral eosinophils and serum IgE. J Gastroenterol 1996;31:73741.
The Southern Surgeons Club. A prospective analysis of 1518 laparoscopic
cholecystectomies. N Engl J Med 1991;324:10738.
McEntee G, Grace PA, Bouchier-Hayes D. Laparoscopic cholecystectomy and the common
bile duct. Br J Surg 1991;78:3856.
Grace PA, Quereshi A, Coleman J, et al. Reduced postoperative hospitalization after
laparoscopic cholecystectomy. Br J Surg 1991;78:1602.
Hermann RE. A plea for a safer technique of cholecystectomy. Surgery 1976;79:60911.
Gilliland TM, Traverso LW. Modern standards for comparison of cholecystectomy with
alternative treatments for symptomatic cholelithiasis with emphasis on long-term relief of
symptoms. Surg Gynecol Obstet 1990;170:3944.
Macintyre IMC, Wilson RG. Laparoscopic cholecystectomy. Br J Surg 1993;80:5529.
Asbun HJ, Rossi RL, Lowell JA, Munson JL. Bile duct injury during laparoscopic
cholecystectomy: Mechanism of injury, prevention and management. World J Surg
1993;80:54752.
Adamsen S, Hansen OH, Funch-Jensen P, et al. Bile duct injury during laparoscopic
cholecystectomy: A prospective nationwide series. J Am Coll Surg 1997;184:5718.
Berkelhammer C, Kortan P, Haber GB. Endoscopic biliary prostheses as treatment for benign
postoperative bile duct strictures. Gastrointest Endosc 1989;35:95101.
Wootton FT, Hoffman BJ, Marsh WH, Cunningham JT. Biliary complications following
laparoscopic cholecystectomy. Gastrointest Endosc 1992;38:1835.
Davidoff AM, Pappas TN, Murray EA, et al. Mechanisms of major biliary injury during
laparoscopic cholecystectomy. Ann Surg 1992;215:196202.
Woods MS, Traverso LW, Kozarek RA, et al. Characteristics of biliary tract complications
during laparoscopic cholecystectomy: A multi-institutional study. Am J Surg 1994;167:2734.
Bergman JJ, van den Brink GR, Rauws EA, et al. Treatment of bile duct lesions after
laparoscopic cholecystectomy. Gut 1996;38:1417.
Barkun AN, Rezieg M, Mehta SN, et al. Postcholecystectomy biliary leaks in the laparoscopic
era: Risk factors, presentation, and management. McGill Gallstone Treatment Group.
Himal HS. The role of ERCP in laparoscopic cholecystectomy-related cystic duct stump leaks.
Raijman I, Catalano MF, Hirsch GS, et al. Endoscopic treatment of biliary leakage after
laparoscopic cholecystectomy. Endoscopy 1994;26:7414.
Sefr R, Ochmann J, Kozumplik L, et al. The role of relaparoscopy in the management of bile
leaks after laparoscopic cholecystectomy. Int Surg 1995;80:3567.
Albasini JL, Aledo VS, Dexter SP, et al. Bile leakage following laparoscopic cholecystectomy.
29. Sandoval BA, Goettler CE, Robinson AV, et al. Cholescintigraphy in the diagnosis of bile leak
after laparoscopic cholecystectomy. Am Surg 1997;63:6116.
30. Neidich R, Soper N, Edmundowicz S, et al. Endoscopic management of bile duct leaks after
attempted laparoscopic cholecystectomy. Surg Laparosc Endosc 1996;6:34854.
31. Saraswat VA, Choudhuri G, Sharma BC, et al. Endoscopic management of postoperative bile
leak. J Gastroenterol Hepatol 1996;11:14851.
32. Barton JR, Russell RC, Hatfield AR. Management of bile leaks after laparoscopic
33. Born P, Bruhl K, Rsch T, et al. Long-term follow-up of endoscopic therapy in patients with
post-surgical biliary leakage. Hepatogastroenterology 1996;43:47782.
34. Libby ED, Branch MS, Cotton PB. Cystic duct leak after laparoscopic cholecystectomy despite
preoperative sphincterotomy. Gastrointest Endosc 1995;41:5114.
35. Eid A, Almogy G, Pikarsky AJ, et al. Conservative treatment of a traumatic tear of the left
hepatic duct: Case report. J Trauma 1996;41:9123.
36. Poli ML, Lefebvre F, Ludot H, et al. Nonoperative management of biliary tract fistulas after
blunt abdominal trauma in a child. J Pediatr Surg 1995;30:171921.
37. Mallery S, Freeman ML, Peine CJ, et al. Biliary-shunt fistula following transjugular intrahepatic
portosystemic shunt placement. Gastroenterology 1996;111:13537.
38. Lillemore KD, Pitt HA, Cameron JL. Current management of benign bile duct strictures. Adv
Surg 1992;25:11974.
39.
Bismuth H. Postoperative strictures of the bile duct. In Blumgart LH (ed). The Biliary Tract.
Clinical Surgery International. Edinburgh: Churchill Livingstone, 1982;20918.
40. Davids PHP, Rauws EAJ, Coene PPLO, et al. Endoscopic stenting for postoperative biliary
strictures. Gastrointest Endosc 1992;38:128.
41. Rossi RL, Schirmer WJ, Braasch JW, et al. Laparoscopic bile duct injuries: Risk factors,
recognition and repair. Arch Surg 1992;127:596602.
42. Kozarek RA. Laparoscopic cholecystectomy: What to do with the common duct. Gastrointest
Endosc 1993;39:99101.
43. Way L. Bile duct injury during laparoscopic cholecystectomy. Ann Surg 1992;215:196202.
44. McAnena OJ, Wilson PD. Diathermy in laparoscopic surgery. Br J Surg 1993;80:10946.
45. Keane FB, Tanner WA, Gillen P. Operative cholangiography and laparoscopic bile duct
exploration. Br J Surg 1993;80:9578.
46. Moossa AR, Easter DW, van Sonnenberg E, et al. Laparoscopic injuries to the bile duct: A
cause for concern. Ann Surg 1992;215:2038.
47. Terblanche J, Allison HF, Northover JMA. An ischaemic basis for biliary strictures. Surgery
1983;94:527.
48. Raute M, Podlech P, Jaschke W, et al. Management of bile duct injuries and strictures
following cholecystectomy. World J Surg 1993;17:55362.
49. Blumgart LH, Kelley CJ, Benjamin IS. Benign bile duct stricture following cholecystectomy:
Critical factors in management. Br J Surg 1984;71:83643.
50.
Genest JF, Nanos E, Grundfest-Broniatowski S, et al. Benign biliary strictures: An analytic
review (19701984). Surgery 1985;99:40913.
51. Pellegrini CA, Thomas MJ, Way LW. Recurrent biliary stricture: Pattern of recurrence and
outcome of surgical therapy. Am J Surg 1984;147:17580.
52. Csendes A, Diaz C, Burdiles P, et al. Indications and results of hepaticojejunostomy in benign
strictures of the biliary tract. Hepatogastroenterology 1992;39:3336.
53. Pitt HA, Miyamoto T, Parapatis KS. Factors influencing outcome in patients with
post-operative biliary stricture. Am J Surg 1982;144:1421.
54. Pereira-Lima L. Biliary reconstruction in benign postoperative stricture with transhepatic tubes.
Am J Surg 1992;164:1248.
55. Millikan KW, Gleason TG, Deziel DJ, Doolas A. The current role of U tubes for benign and
malignant biliary obstruction. Ann Surg 1993;218:6219.
56. Morrison MC, Lee MJ, Saini S, et al. Percutaneous balloon dilatation of benign biliary
strictures. Radiol Clin North Am 1990;28:1191201.

57. Mueller PR, van Sonnenberg E, Ferrucci JT, et al. Biliary stricture dilatation: Multicenter
review of clinical management in 73 patients. Radiology 1986;160:1722.
58. Gallacher DJ, Kadir S, Kaufman SL, et al. Nonoperative management of benign biliary
strictures. Radiology 1985;156:6259.
59. Citron SJ, Martin LG. Benign biliary strictures: Treatment with percutaneous cholangioplasty.
Radiology 1991;178:33941.
60.
Lee MJ, Mueller PR, Saini S, et al. Percutaneous dilatation of benign biliary strictures:
Single-session therapy with general anaesthesia. AJR 1991;157:12636.
61. Speer AG, Cotton PB, Russell RCG, et al. Randomised trial of endoscopic versus
percutaneous stent insertion in malignant obstructive jaundice. Lancet 1987;2:5762.
62. Traverso LW, Kozarek RA, Ball TJ, et al. Endoscopic retrograde cholangiopancreatography
after laparoscopic cholecystectomy. Am J Surg 1993;165:5816.
63. Kent AL, Cox MR, Wilson TG, et al. Endoscopic retrograde cholangiopancreatography
following laparoscopic cholecystectomy. Aust N Z J Surg 1994;64:40712.
64. Siegel JH, Cohen SA. Endoscopic treatment of laparoscopic bile duct injuries.
Gastroenterologist 1994;2:513.
65. Pencev D, Brady PG, Pinkas H, Boulay J. The role of ERCP in patients after laparoscopic
cholecystectomy. Am J Gastroenterol 1994;89:15237.
66. Siegel JH, Guelrud M. Endoscopic cholangiopancreatoplasty: Hydrostatic balloon dilation in
the bile duct and pancreas. Gastrointest Endosc 1983;29:99103.
67. Huibregtse K, Katon RM, Tytgat GN. Endoscopic treatment of postoperative biliary strictures.
Endoscopy 1986;18:1337.
68. Manoukian AV, Schmalz MJ, Geenen JE, et al. Endoscopic treatment of problems
encountered after laparoscopic cholecystectomy. Gastrointest Endosc 1993;39:914.
69. Davids PHP, Tanka AK, Rauws EAJ, et al. Benign biliary strictures: Surgery or endoscopy.
Ann Surg 1993;217:23743.
70. Coons H. Metallic stents for the treatment of biliary obstruction: A report of 100 cases.
Cardiovasc Intervent Radiol 1992;15:36774.
71. Maccioni F, Bezzi M, Gandini R, et al. Metallic stents in benign biliary stenosis. A four year
follow up. Radiol Med (Torino) 1993;86:294301.
72. Maccioni F, Rossi M, Salvatori FM, et al. Metallic stents in benign biliary strictures: Three year
follow-up. Cardiovasc Intervent Radiol 1992;15:3606.
73. Foerster EC, Hoepffner N, Domschke W. Bridging of benign choledochal stenoses by
endoscopic retrograde implantation of mesh stents. Endoscopy 1991;23:1335.
74. Goldin E, Beyar M, Safra T, et al. A new self-expandable and removable metal stent for biliary
obstructionA preliminary report. Endoscopy 1993;25:5979.
75. Sherman S, Jamidar P, Shaked A, et al. Biliary tract complications after orthotopic liver
transplantation. Endoscopic approach to diagnosis and therapy. Transplantation
1995;60:46770.
76. O'Connor HJ, Vickers CR, Buckels JA, et al. Role of endoscopic retrograde
cholangiopancreatography after orthotopic liver transplantation. Gut 1991;32:41923.
77. Hintze RE, Adler A, Veltzke W, et al. Endoscopic management of biliary complications after
orthotopic liver transplantation. Hepatogastroenterology 1997;44:25862.
78. Macfarlane B, Davidson B, Dooley JS, et al. Endoscopic retrograde cholangiography in the
diagnosis and endoscopic management of biliary complications after liver transplantation. Eur
J Gastroenterol Hepatol 1996;8:10036.
79. Theilmann L, Kuppers B, Kadmon M, et al. Biliary tract strictures after orthotopic liver
transplantation: Diagnosis and management. Endoscopy 1994;26:51722.
80. Sossenheimer M, Slivka A, Carr-Locke D. Management of extrahepatic biliary disease after
orthotopic liver transplantation: Review of the literature and results of a multicenter survey.
Endoscopy 1996;28:56571.
81. Kasai T, Nakatani T, Hirosawa K, Kobayashi K. Benign stricture of the extra-hepatic bile duct
following hepatectomy for traumatic hepatic rupture. Intensive Care Med 1993;19:4724.
82. Stratta RJ, Wood RP, Langnas AN, et al. Diagnosis and treatment of biliary tract
complications after orthotopic liver transplantation. Surgery 1989;106:67583; discussion
6834.
83. Matthews JB, Gertsch P, Baer HU, et al. Biliary stricture following hepatic resection. HPB
Surg 1991;3:18190.
84. Williams HJ, Bender CE, May GR. Benign postoperative biliary strictures: Dilation with
fluoroscopic guidance. Radiology 1987;163:62934.
85. Matthews JB, Baer HU, Schweizer WP, et al. Recurrent cholangitis with and without
anastomotic stricture after biliary-enteric bypass. Arch Surg 1993;128:26972.
86. Ward EM, Kiely MJ, Maus TP, et al. Hilar biliary strictures after liver transplantation:
Cholangiography and percutaneous treatment. Radiology 1990;177:25963.
87. Chapman RWG, Marborgh BA, Rhodes JM, et al. Primary sclerosing cholangitis: A review of
its clinical features, cholangiography and hepatic histology. Gut 1980;21:8707.
88. Chapman RW. Aetiology and natural history of primary sclerosing cholangitisA decade of
progress?. Gut 1991;32:14335.
89. Kaw M, Silverman WB, Rabinovitz M, Schade RR. Biliary tract calculi in primary sclerosing
cholangitis. Am J Gastroenterol 1995;90:725.
90.
Majoie CBLM, Reeders WAJJ, Sanders BJ, et al. Primary sclerosing cholangitis: A modified
classification of cholangiographic findings. AJR 1991;157:4957.
91. Chen LY, Goldberg HI. Sclerosing cholangitis: Broad spectrum of radiographic features.
92.
Gulliver DJ, Baker ME, Putnam W, et al. Bile duct diverticula and webs: Nonspecific
cholangiographic features of primary sclerosing cholangitis. AJR 1991;157:2815.
93.
Craig DA, MacCarty RL, Wiesner RH, et al. Primary sclerosing cholangitis: Value of
cholangiography in determining the prognosis. AJR 1991;157:95964.
94.
Huibregtse K. Therapeutic investigations in primary sclerosing cholangitis. In Endoscopic
Biliary and Pancreatic Drainage. Stuttgart: George Thieme, 1988;7885.
95. Gaing AA, Geders JM, Cohen SA, Siegel JH. Endoscopic management of primary sclerosing
cholangitis: Review and report of an open series. Am J Gastroenterol 1993;88:20008.
96. Johnson GK, Geenen JE, Venu RP, Hogan WJ. Endoscopic treatment of biliary duct
strictures in sclerosing cholangitis: Follow-up assessment of a new therapeutic approach.
97. Lombard M, Farrant M, Karani J, et al. Improving biliary-enteric drainage in primary sclerosing
cholangitis: Experience with endoscopic methods. Gut 1991;32:13648.
98. Grijm R, Huibregtse K, Bartelsman J, et al. Therapeutic investigations in primary sclerosing
cholangitis. Dig Dis Sci 1986;31:7928.
99. Allison MC, Burroughs AK, Noone P, Summerfield JA. Biliary lavage with corticosteroids in
primary sclerosing cholangitis. A clinical, cholangiographic and bacteriological study. J
Hepatol 1986;3:11822.
100. Wagner S, Gebel M, Meier P, et al. Endoscopic management of biliary tract strictures in
primary sclerosing cholangitis. Endoscopy 1996;28:54651.
101. van Milligen de Wit AW, van Bracht J, Rauws EA, et al. Endoscopic stent therapy for
dominant extrahepatic bile duct strictures in primary sclerosing cholangitis. Gastrointest
Endosc 1996;44:2939.
102. Lee JG, Schutz SM, England RE, et al. Endoscopic therapy of sclerosing cholangitis.
103.
Coffey RJ, Wiesner RH, Beaver SJ, et al. Bile duct carcinoma: A late complication of end
stage primary sclerosing cholangitis. Hepatology 1984;4:1056.
104. Aadland E, Schrumpf E, Fausa O, et al. Primary sclerosing cholangitis: A long-term follow up
study. Scand J Gastroenterol 1987;22:65564.
105. Marsh JW, Iwatsuki S, Makowka L, et al. Orthotopic liver transplantation for primary
sclerosing cholangitis. Ann Surg 1988;207:215.
106. Miros M, Kerlin P, Walker N, et al. Predicting cholangiocarcinoma in patients with primary
sclerosing cholangitis before transplantation. Gut 1991;32:136973.
107. MacCarty RL, LaRusso NF, May GR, et al. Cholangiocarcinoma complicating primary
sclerosing cholangitis: Cholangiographic appearances. Radiology 1985;156:436.
108. Van Laethem JL, Deviere J, Bourgeois N, et al. Cholangiographic findings in deteriorating
primary sclerosing cholangitis. Endoscopy 1995;27:2238.
109. Cohan RH, Illescas FF, Newnam GE, et al. Biliary cytodiagnosis: Bile sampling for cytology.
Invest Radiol 1985;20:1779.
110. Meenan J, Schoeman M, Rauws E, et al. A video baby cholangioscope. Gastrointest Endosc
1995;42:5845.
111. Ramage JK, Donaghy A, Farrant JM, et al. Serum tumor markers for the diagnosis of
cholangiocarcinoma in primary sclerosing cholangitis. Gastroenterology 1995;108:8659.
112. Caldwell RJ, Phillips E, Thompford NR. Idiopathic benign stricture of the common bile duct.
113. Standfield NJ, Salisbury JR, Howard ER. Benign non-traumatic inflammatory strictures of the
extrahepatic biliary system. Br J Surg 1989;76:84952.
114. Cello JP. Acquired immunodeficiency syndrome cholangiopathy: Spectrum of disease. Am J
Med 1989;86:53946.
115. Dowsett JF, Miller R, Davidson R, et al. Sclerosing cholangitis in acquired immunodeficiency
syndrome. Scand J Gastroenterol 1988;23:126774.
116. Cappell MS. Hepatobiliary manifestations of the acquired immunodeficiency syndrome. Am J
117. Payne TH, Cohn DL, Davidson AJ, et al. Marked elevations of serum alkaline phosphatase in
patients with AIDS. J Acquir Immune Defic Syndr 1991;4:23843.
118. Angus KW. Cryptosporidiosis and AIDS. Baillieres Clin Gastroenterol 1990;4:42541.
119. Teixidor HS, Godwin TA, Ramirez EA. Cryptosporidiosis of the biliary tract in AIDS. Radiology
1991;180:516.
120. Pol S, Romana CA, Richard S, et al. Microsporidiosis infection in patients with the human
immunodeficiency virus and unexplained cholangitis. N Engl J Med 1993;328:959.
121. Willson R, Harrington R, Stewart B, Fritsche T. Human immunodeficiency virus 1-associated
necrotizing cholangitis caused by infection with Septata intestinalis. Gastroenterology
1995;108:24751.
122. Cacciarelli AG, Freiman HJ, Anderson JD, Robilotti JG. Biliary fistula in a patient with HIV
cholangiopathy. Gastrointest Endosc 1996;44:3458.
123. Quinn D, Pocock N, Freund J, et al. Radionuclide hepatobiliary scanning in patients with
AIDS-related sclerosing cholangitis. Clin Nucl Med 1993;18:41722.
124. Bird GL, Kennedy DH, Forrest JA. AIDS-related cholangitis: Diagnostic features and course in
four patients. Scott Med J 1995;40:534.
125. Daly CA, Padley SP. Sonographic prediction of a normal or abnormal ERCP in suspected
AIDS-related sclerosing cholangitis. Clin Radiol 1996;51:61821.
126. Ducreux M, Buffet C, Lamy P, et al. Diagnosis and prognosis of AIDS-related cholangitis.
AIDS 1995;9:87580.
127. Forbes A, Blanshard C, Gazzard B. Natural history of AIDS-related sclerosing cholangitis: A
study of 20 cases. Gut 1993;34:11621.
128. Benhamou Y, Caumes E, Gerosa Y, et al. AIDS-related cholangiopathy. Critical analysis of a
prospective series of 26 patients. Dig Dis Sci 1993;38:11138.
129. Kline TJ, De las Morenas T, O'Brien M, et al. Squamous metaplasia of extrahepatic biliary
system in an AIDS patient with cryptosporidia and cholangitis. Dig Dis Sci 1993;38:9602.
130. Cello JP, Chan MF. Long-term follow-up of endoscopic retrograde cholangiopancreatography
J Med 1995;99:6003.
131. Leung JW, Sung JY, Chung SC, Metreweli C. Hepatic clonorchiasisA study by endoscopic
retrograde cholangiopancreatography. Gastrointest Endosc 1989;35:22631.
132.
Yellin AE, Donovan AJ. Biliary lithiasis and helminthiasis. Ann Surg 1982;142:12836.
133. Carmond RH, Crass RA, Lim RC Jr, et al. Oriental cholangitis. Am J Surg 1984;148:11724.
134.
Van Sonnenberg E, Casola JR, Cubberley DA, et al. Oriental cholangiohepatitis? Diagnostic
imaging and interventional management. AJR 1986;146:32731.
135. Leung JW, Chung SC. Endoscopic management of biliary ascariasis. Gastrointest Endosc
1988;34:31820.
136. Schwacha H, Keuchel M, Gagesch G, Hagenmuller F. Fasciola gigantica in the common bile
duct: Diagnosis by ERCP. Endoscopy 1996;28:323.
137. Danilewitz M, Kotfila R, Jensen P. Endoscopic diagnosis and management of Fasciola
hepatica causing biliary obstruction. Am J Gastroenterol 1996;91:26201.
138. Dias LM, Silva R, Viana HL, et al. Biliary fascioliasis: Diagnosis, treatment and follow-up by
139. el-Newihi HM, Waked IA, Mihas AA. Biliary complications of Fasciola hepatica: The role of
endoscopic retrograde cholangiography in management. J Clin Gastroenterol
1995;21:30911.
140. Ramirez FC, Walker GJ, Sanowski RA, et al. Obstructive jaundice due to Coccidioides
immitis. Gastrointest Endosc 1996;43:5057.
141. Kemeny MM, Battiford H, Blaney DW, et al. Sclerosing cholangitis after continuous hepatic
artery infusion of FUDR. Ann Surg 1985;202:1768.
142. Ludwig J, Kim CH, Wiesner RH, et al. Floxuridine induced sclerosing cholangitis: An ischemic
cholangiopathy?. Hepatology 1989;9:2158.
143. Siegel JH, Ramsey WJ. Endoscopic biliary stent placement for bile stricture after hepatic
artery infusion of 5-FUDR. J Clin Gastroenterol 1986;8:6736.
144. Belghiti J, Benhamou JP, Houry S, et al. Caustic sclerosing cholangitis: A complication of the
surgical treatment of hydrated disease of the liver. Arch Surg 1986;121:11625.
145. Scott J, Summerfield JA, Elias E, et al. Chronic pancreatitis: A cause of cholestasis. Gut
1977;18:196201.
146. Yadegar J, Williams RA, Passaro E, Wilson SE. Common duct strictures from chronic
pancreatitis. Arch Surg 1980;115:5826.
147. Wheatley MJ, Coran AG. Obstructive jaundice secondary to chronic pancreatitis in children:
Report of two cases and review of the literature. Surgery 1988;104:8639.
148. Stahl TJ, Allen MO, Ansel HJ, Vennes JA. Partial biliary obstruction caused by chronic
pancreatitis. An appraisal of indications for surgical biliary drainage. Ann Surg
1988;207:2632.
149. Stabile BE, Calabria R, Wilson SE, Passaro E Jr. Stricture of the common bile duct from
chronic pancreatitis. Surg Gynecol Obstet 1987;165:1216.
150. Gregg JA, Carr-Locke DL, Gallagher MM. Importance of common bile duct stricture
associated with chronic pancreatitis. Diagnosis by endoscopic retrograde
cholangiopancreatography. Am J Surg 1981;141:199203.
151. Littenberg G, Afroudakis A, Kaplowitz N. Common bile duct stenosis from chronic
pancreatitis: A clinical and pathological spectrum. Medicine 1979;58:385412.
152. Deviere J, Devaere S, Baize M, Cremer M. Endoscopic biliary drainage in chronic
pancreatitis. Gastrointest Endosc 1990;36:96100.
153. Smits ME, Rauws EA, van Gulik TM, et al. Long-term results of endoscopic stenting and
surgical drainage for biliary stricture due to chronic pancreatitis. Br J Surg 1996;83:7648.
154. Deviere J, Cremer M, Baize M, et al. Management of common bile duct stricture caused by
chronic pancreatitis with metal mesh self expandable stents. Gut 1994;35:1226.
Chapter 66 The Gallbladder and Intrahepatic Bile Ducts

(6900)
(6901)
LARS AABAKKEN, M.D., PH.D., B.C.
JARL . JAKOBSEN, M.D., PH.D.

MAGNE OSNES, M.D., PH.D.
Endoscopic retrograde cholangiography (ERC) has brought about major improvements in the
diagnosis and therapy of diseases of the biliary tract. The most important advances are exactness in
diagnosis of extrahepatic biliary diseases, especially those involving the common bile duct, and
differentiation between benign and malignant conditions. There has also been remarkable progress in
endoscopic therapeutic intervention in diseases involving the common bile duct. Despite this emphasis
on the extrahepatic bile ducts, endoscopic cannulation has also enhanced the diagnosis of gallbladder
disease and diseases affecting the intrahepatic duct system.
A major restriction in the recognition of abnormalities in the gallbladder and intrahepatic ducts is that
the endoscopic and radiologic techniques are difficult to master; accurate diagnosis requires
considerable experience. This chapter focuses on the main technical principles, radiologic and
endoscopic, for examinations that permit exact diagnosis and successful treatment. The typical
cholangiographic findings of diseases of gallbladder and intrahepatic ducts where ERC plays or may
play an important role are described.
Technique
Endoscopic Equipment
Some of the endoscopic equipment required for the procedure is listed in Table 661. A side-viewing
duodenoscope with a large instrumental channel (3.2 mm) is preferable. This allows the use of a
relatively large-diameter forceps to obtain intraductal biopsies. Although any flexible biopsy forceps can
be used in either the common bile duct or the intrahepatic ducts, one with a soft tip is preferable. This
will pass the main duodenal papilla and reach the target in the ductal system with greater ease.
Sheathed brushes may be used to obtain cytologic specimens in some pathologic conditions that
involve the intrahepatic ducts.
TABLE 661
Equipment for Endoscopic Retrograde
Cholagiography
DIAGNOSIS
Duodenoscope with large accessory channel (3.2 mm)
Biopsy forceeps with soft tip
Sheathed cytology brush
Balloon catheter for occlusion and selective apacification of
intrahepatic ducts
Guidewires for selective cannulation of intrahepatic ducts
Mother-baby cholangioscopes for direct visualization and access to
intraductal lesions
TREATMENT
Balloon catheter for removal of intrahepatic calculi
Baskets for removal of intrahepatic calculi
Grundzig-type balloon-dilation catheters for dilation of strictures
Catheters with distal tips of varying shapes and diameters may be useful when cannulation of the main
duodenal papilla is difficult (Figure 661). Certain standard catheters can be shaped or drawn into
different configurations if heated carefully; this depends, however, on the material the catheter is made
of.
(6902)Figure 661. Catheters with differently shaped tips, which may be useful when
cannulation of the main duodenal papilla is difficult. Some of these are available
commercially, others are "homemade." (Drawn by Per Engeseth. Reproduced with the
permission of the Scandinavian Association for Digestive Endoscopy.)
For selective radiologic demonstration of an intrahepatic ductal segment, a balloon catheter may be of
value. Contrast medium can be injected with gentle pressure, and fluoroscopic monitoring used when
the balloon has been inflated to occlude the biliary system distal to the segment of interest. In some
instances, a guidewire may also help in cannulating the different intrahepatic ducts. Balloon and
basket-type catheters may also be used for extraction of calculi from within the intrahepatic ducts.
Grntzig-type balloon catheters can be used for dilation of strictures within the biliary system, in some
cases with subsequent extraction of calculi lodged proximal to the stricture.
Peroral cholangioscopes have been introduced for use in conjunction with ERC for direct visualization
of the ductal system and gallbladder. Cholangioscopes make it possible to obtain biopsies under direct
vision from lesions within the ducts and provide direct access to calculi for dissolution or fragmentation
(see Chapter 70: Peroral Cholangioscopy and Pancreatoscopy and Chapter 71:
Choledochofiberoscopy and Endoscopic Lithotripsy).
Radiologic Equipment
The radiologic equipment for the procedure is listed in Table 662. The examination is performed with
an x-ray machine equipped with a tiltable table that allows x-ray exposures in different positions. A
high-resolution fluoroscopic unit is necessary for detection of small abnormalities. At any point in the
examination, it may be necessary to immediately document fine alterations from the normal anatomy.
X-ray films may be made using a standard film size and conventional spot filming system, although a
fast camera system with smaller-size x-ray film and a short exposure time provides a rapid method of
documentation during filling of the duct system. In addition to the standard compression apparatus built
into the x-ray machine, a compression device that operates on the principle of an inflatable rubber ball
is useful. The x-ray contrast agent should be a standard water-soluble iodine medium with a low
concentration of iodine (200 mg/ml).
Radiologic Equipment for

Gallbladder and Intrahepatic Duct
Examination During Endoscopic Retrograde
TABLE 662
X-RAY APPARATUS
Tiltable, movable table
Capability for horizontal beam x-rays (patient either upright
or flat)
STANDARD FLUOROSCOPY CAPABILITY
Additional monitor for endoscopist
On/off foot pedal for endoscopist
IMAGING SYSTEM
Spot-film system, standard film sizes (24 x 30 cm and 18 x
24 cm)
Smaller-size-film camera system (100 mm/70 mm)
COMPRESSION DEVICE
Radiologic Equipment for

Gallbladder and Intrahepatic Duct
Examination During Endoscopic Retrograde
TABLE 662
Built into the x-ray apparatus

Additional inflatable compression device
STANDARD CONTRAST MATERIAL (200 mg iodine/ml)
Gallbladder Examination During Endoscopy

Some important technical principles must be considered in order to obtain a satisfactory examination of
the gallbladder and achieve a high level of diagnostic accuracy (Table 663). After cannulation of the
common bile duct, contrast material is injected under continuous fluoroscopic control. The main
technical problem with demonstration of gallbladder disease is overfilling (excessive filling) of the
gallbladder with contrast medium, defined here in comparison to optimum opacification for satisfactory
x-ray exposures. Dense opacification may easily obscure the presence of small calculi.
Radiologic Procedures for

Gallbladder Examination During Endoscopy
TABLE 663
1. Plain film
2. X-ray exposure with filled gallbladder
3. X-ray exposures with filled cystic duct
The phenomenon of overopacification cannot be avoided by the use of contrast material with a low
iodine concentration. Retrograde injection can be stopped when the cystic duct is filled or when the
medium reaches the infundibulum of the gallbladder (Figure 662), although sometimes excessive
filling during the procedure cannot be avoided. Therefore, it may be necessary to interrupt the injection
when a sufficient quantity of contrast medium has entered the gallbladder to obtain a series of x-rays
before proceeding to complete opacification of the biliary system. Ordinarily, the normal cystic duct and
gallbladder will fill with contrast medium automatically during the injection. However, lack of
opacification does not necessarily indicate blockage of the cystic duct. Therefore, cannulation of the
cystic duct may be required to verify stenosis or occlusion. During the endoscopic procedure, it may
also be necessary to change the patient's position from semiprone (right shoulder raised) to prone to
improve the radiologic view of the gallbladder (Figure 663).
(6903)Figure 662. Contrast injection is stopped before it reaches the infundibulum of the
gallbladder to avoid overfilling. (Drawn by Per Engeseth. Reproduced with the permission of
the Scandinavian Association for Digestive Endoscopy.)
(6904)Figure 663. Selective cannulation of the gallbladder (arrow), with x-ray film being
taken with the patient in a nearly horizontal and prone position.
Gallbladder Examination After Endoscopy

The radiologic examination of the gallbladder after completion of the endoscopic phase of the
procedure is summarized in Table 664. After removal of the endoscope, the patient is turned supine,
and a series of x-rays are made while gradually changing the subject's position from recumbent to
half-upright or to standing (Figure 664). During elevation toward the standing position, the gallbladder
will fill with contrast medium, if not already opacified, unless an occlusion or stenosis of the cystic duct
is present. Ideally, the first x-ray film should be taken when the gallbladder is coated with a thin layer of
contrast medium. By elevating the patient toward the upright position, it is not difficult to demonstrate
movement of a filling defect(s). This confirms the diagnosis of a calculus. Compression over the area
of the gallbladder will remove contrast medium or air-filled loops of bowel from the x-ray field.

Gallbladder Examination After Endoscopy
TABLE 664
1. X-ray films with a thin coating of contrast medium and

the patient supine
2. Spot films in different positions while gradually tilting
the patient upright.
3. Spot film with the patient standing
4. Eventual spot films in the right lateral decubitus
position with the patient horizontal or tilted
5. Delayed spot films
(6905)Figure 664. A-C, Schematic diagrams illustrate the process for obtaining x-ray films
of gallbladder while gradually raising the patient from a horizontal to a standing position to
detect calculi and their motion. (A-C, drawn by Per Engeseth. Reproduced with the permission
of the Scandinavian Association for Digestive Endoscopy.)
If, after these procedures, the status of the gallbladder is still unclear or the diagnosis uncertain, x-ray
films should be obtained with a horizontal x-ray beam with the patient in a right lateral decubitus
position. The patient may be tilted upward and downward in this position (Figure 665). Sometimes the
best x-ray views are obtained during a sustained expiration. To demonstrate floating calculi, it may be
necessary to obtain delayed x-rays with a horizontal beam (Figure 666).
(6906)Figure 665. A-C, X-ray films of the gallbladder taken in the right lateral decubitus
position with the patient lying flat or with the head tilted upward or downward. (A-C, drawn
by Per Engeseth. Reproduced with the permission of the Scandinavian Association for
Digestive Endoscopy.)
(6907)Figure 666. X-ray film taken on a delayed basis with a horizontal beam to
demonstrate multiple small, floating calculi.
After x-ray films are obtained utilizing the right lateral decubitus position, the patient may be turned to
the supine position and further exposures made. As the contrast medium shifts within the gallbladder,
certain parts of the gallbladder will remain coated with a thin layer of contrast medium. This is ideal for
x-ray study of the mucosal surface. By taking advantage of this phenomenon during the various
changes in the patient's position, the entire mucosal surface of the gallbladder may be examined. With
the use of the techniques described, x-ray films made with the aid of drugs or fatty meals that contract
the gallbladder are unnecessary.
Examination of Intrahepatic Ducts During Endoscopy

The main technical points with reference to radiologic examination of the intrahepatic ducts during
endoscopy are listed in Table 665. After contrast medium injection by standard endoscopic retrograde
cholangiopancreatography (ERCP) technique, the intrahepatic ducts will be visualized in most cases.
With the patient in the semiprone (right shoulder slightly raised) position, the left hepatic duct will fill
with contrast medium first. Sometimes significant filling of the intrahepatic ducts in the right lobe can be
obtained with this technique. When this is not accomplished, the right ductal system may be filled by
turning the patient to a prone position during the endoscopic procedure. Although the right hepatic duct
is generally oriented along the long axis of the body, it also has a somewhat posterior course. With the
patient in the prone position, this may mean that the course of the duct is upward and, consequently,
the flow of contrast medium into the duct is opposed by gravity. Thus, in certain positions, the left
system may fill preferentially. Tilting the patient's head downward may also aid in filling the right hepatic
system. Lack of filling of one of the intrahepatic systems based on these normal anatomic features can
be misinterpreted as evidence of obstruction.

Intrahepatic Duct Examination During
Endoscopy
TABLE 665
1. Spot film when contrast medium injection nearly

complete
2. Spot films when intrahepatic ducts completely filled
3. Spot films after injection with occluding balloon
catheter, after selective duct catheterization, or after
positioning the catheter in the mid-extrahepatic ductal
system
Generally, the intrahepatic ducts fill more readily after cholecystectomy. When the gallbladder is
present, it often acts as a low-pressure reservoir. With standard injection technique with the catheter
tip at the papilla, the gallbladder will fill in preference to the intrahepatic system; dense gallbladder
opacification must then occur before the intrahepatic system begins to fill to any significant degree.
Overdistention of the gallbladder may result in nausea and vomiting. When the intrahepatic ducts are
the main object of the examination, better filling may be achieved by deep cannulation to the bifurcation
or selective cannulation of either of the intrahepatic ducts with the standard catheter (Figure 667). If a
higher injection pressure is required for satisfactory opacification, a balloon catheter may be used to
occlude the distal biliary system during the injection (Figure 668). A standard stone retrieval balloon
catheter is suitable for this purpose. Depending on where the balloon is positioned, this catheter can
also be used to prevent runoff of contrast medium into the gallbladder. A guidewire can be useful if
difficulty is encountered in correctly positioning either the standard catheter for selective opacification
or the balloon catheter for occlusion and opacification. High-quality x-ray films should be obtained
during the endoscopic procedure because of the relatively rapid emptying of the intrahepatic ducts
when the injection is ended.
(6908)Figure 667. Selective catheterization of the right hepatic duct (arrow indicates
catheter) in a patient with metastasis from papillary carcinoma. Air in the biliary tree is due to
endoscopic choledochoduodenostomy.
(6909)Figure 668. Catheterization and injection of contrast material using a balloon catheter
to occlude the common hepatic duct in a patient with cirrhosis.
In special cases, deep cannulation of the intrahepatic ducts may be technically difficult. In this situation,
a Dormia basket may be of assistance. The tip of the basket is pushed out and ahead of the sheath
and opened, and then the catheter is advanced as the basket is closed. The net result of this
maneuver is that the sheath will have advanced several millimeters in the direction followed by the
basket. The maneuver is then repeated, allowing a stepwise introduction of the sheath into the desired
position (Figure 669). Contrast material can be injected during the cannulation procedure, an
additional advantage of this technique.
(6910)Figure 669. Use of a Dormia basket to achieve intrahepatic cannulation in a

technically difficult case. From a withdrawn, closed position (A), the catheter is introduced
and opened (B), and the sheath (arrowhead in B), is pushed proximally to the fixed upper end
of the basket (double arrowhead in B) as the basket is gradually closed by an assistant.
Examination of Intrahepatic Ducts After Endoscopy

Table 666 lists the main steps of the radiologic technique after the endoscopic phase of the
procedure has been completed. Once the intrahepatic ducts are filled with contrast material, the
endoscope is removed. The patient assumes a prone position and is then tilted head-downward.
Because contrast medium is heavier than bile, the effect of gravity will be advantageous in outlining
various parts of the intrahepatic ductal system. With the patient in the prone, head-downward position,
the ventral and cephalad parts of the bile ducts will be visualized (Figure 6610). When this maneuver
does not produce complete filling of the upper aspect of the bile ducts, a pillow should be placed under
the abdomen. This position is also best for filling of the confluence of the right and left hepatic duct.
The patient is then turned to the supine position, which results in opacification of the common bile duct
and the dorsal parts of the intrahepatic ducts, particularly the right system. X-ray films should also be
taken under fluoroscopic guidance as the patient is raised toward an upright position. This ensures
complete filling of the right posterior aspect of the intrahepatic duct system as well as the common bile
duct (Figure 6610).

Intrahepatic Duct Examination After
Endoscopy
TABLE 666
1. Spot film with the patient's head tilted downward

2. Spot films after turning the patient to the supine
position
3. Spot films with the patient horizontal
4. Spot films while gradually tilting the patient upright
At each step, well-collimated exposures of regions of
special interest must be obtained. The entire duct system
should be imaged in two projections to avoid
false-negative findings.
(6911)Figure 6610. Diagrams of the various essential positions for visualization of different
parts of the intrahepatic duct system. A, Patient prone, head downward. B, Patient supine, head
downward. C, Patient supine and horizontal. D, Patient supine, head upward. Note the pooling
of contrast material, indicated by the shaded area within the ducts, under the influence of
gravity. (A-D, Drawn by Per Engeseth. Reproduced with the permission of the Scandinavian
Association for Digestive Endoscopy.)
Sometimes it is difficult to demonstrate the biliary ductal system proximal to an occlusion or stenosis. If
the patient can be kept in a prone, head-downward position for a period of time, some contrast material
may gradually flow into the proximal bile ducts. When this maneuver is successful, x-rays often reveal
important information about the nature of the stenosis or obstruction (Figure 6611; see also Figure
6610A). In cases of severely delayed filling of the proximal parts of the ducts, opacification can
sometimes be achieved with the "shaking procedure": With the patient in head-down, prone position,
the endoscopist grasps the abdomen of the patient firmly and shakes it (Figure 6612).
(6912)Figure 6611. Malignant stenosis of the common hepatic duct is demonstrated while
the patient is maintained in the prone, head downward position. Sequence of x-ray films, taken
over the course of several minutes, is arranged from left to right, top to bottom, with the last
film (lower right) demonstrating the tumor (arrowhead). Calculi are also demonstrated in the
gallbladder.
(6913)Figure 6612. "Shaking procedure." A, Standard maneuvers fail to produce sufficient

filling of the left intrahepatic ducts. B, Even after delayed exposure, proximal filling is
incomplete. C, After vigorous shaking of the patient's abdomen, the complete duct system is
eventually opacified.
Some of the common sources of error in the radiologic examination of the gallbladder and intrahepatic
ducts are listed in Table 667.
Technical Problems with Radiology of

Gallbladder and Intrahepatic Ducts
TABLE 667
GALLBLADDER
Overopacification with contrast medium
Contrast in the gut obscuring view
Feces in right colon mimicking small calculi
Failure of mixing of contrast and bile
INTRAHEPATIC DUCTS
Insufficient filling of right hepatic duct
Other Diagnostic Procedures

There are many disorders of the bile ducts in which material for histologic examination is crucially
important. Previously, such material could be readily obtained only at surgery. However, many
nonsurgical methods of management have been introduced for a variety of pathologic conditions of the
biliary tract. A number of techniques are available to obtain histologic confirmation; our preference is
the transpapillary biopsy. The biopsy forceps is guided endoscopically through the papilla and then
under fluoroscopic control to the lesion (Figure 6613). The main indication for this procedure is the
finding of a lesion in the common bile duct (Figure 6614), but lesions in the intrahepatic ducts are also
reachable. In rare cases, it may be necessary to perform an endoscopic papillotomy so that the
forceps may be introduced more easily. This is permissible for diagnostic purposes alone when an
exact diagnosis is essential to the selection of the most appropriate option for treatment. In our
experience with this technique for obtaining endoscopic biopsies from within the bile ducts, malignant
biliary disease was correctly diagnosed in 80% of patients with no false-positive findings (see also
Chapter 64: Diagnosis and Management of Malignant Biliary Obstruction).1(6914)
(6915)Figure 6613. A, Radiograph of stenosis of the bile duct. B, Transpapillary biopsy with
forceps open at the stenosis. C, Biopsy specimen shows adenocarcinoma (175). D, Biopsy
specimen (280).
(6916)Figure 6614. A, Adenoma (arrowheads) has an irregular contour within the common
hepatic duct. B, Transpapillary biopsy with forceps open at the lesion.
The use of a sheathed cytology brush may provide additional information. This accessory is especially
helpful within stenotic lesions (see also Chapter 64: Diagnosis and Management of Malignant Biliary
Obstruction). After introduction through the papilla, the sheath is passed to within or through the
stenotic segment, and then the brush is advanced beyond the sheath and pushed and pulled several
times through the pathologic segment before it is retracted into the sheath. The results of brushing
cytology of the bile ducts are, however, not entirely satisfactory.2(6917) It is possible that new brush
devices with a guidewire at the distal end may make the passage of stenoses easier and improve the
results of brush cytology.
Intraductal cholangioscopy provides direct visualization of pathologic processes in the bile ducts (see
also Chapter 70: Peroral Cholangioscopy and Pancreatoscopy). Newer cholangioscopes have an
instrument channel that makes it possible to obtain biopsies under visual control, although these
specimens are often tiny. Shockwave lithotryptor devices and laser fibers can also be introduced
through cholangioscopes, but at present, these techniques are largely investigational. Transpapillary
cholecystoscopy is a recent addition to the diagnostic repertoire; it even allows for direct contact methyl
tert-butyl ether dissolution of calculi under visual control.36(6918) However, the indications, efficacy,
and safety of this procedure remain to be established.7(6919)
Diseases of the Gallbladder

Calculi
Cholecystolithiasis is usually diagnosed by methods other than ERC, most often ultrasonography. The
use of cholecystography and intravenous cholangiography is steadily decreasing. False-negative
findings may, however, be problematic with these examinations.8(6920) However, results are
improving as a result of the availability of better equipment and the increasing skill of
radiologists.9(6921) The ERC examination of the gallbladder is a very sensitive method for detection of
stones, even when these are very small.8,10(6922) This, however, depends on adherence to the
technical principles outlined previously. The best opportunity for detection of small stones occurs with
the early filling images of the gallbladder when a thin layer of contrast medium is present (Figure
6615). When defects are small, movement differentiates calculi from polyps (Figure 6616). Large
stones are easily detectable, during both early and later filling of the gallbladder (Figures 6617 and
6618). If calculi are adherent to the mucosa, their differentiation from polyps is impossible; in such
patients, acoustic shadowing of the calculi at ultrasonography may resolve the dilemma.
(6923)Figure 6615. Cholelithiasis. Thin contrast material in the gallbladder outlines multiple
small filling defects.
(6924)Figure 6616. Small calculi in the gallbladder. Position of the calculi (arrowheads)
within the gallbladder changes as the patient is raised from flat (A) to standing (B) position.
(6925)Figure 6617. Cholelithiasis. Gallbladder contains a large stone.
(6926)Figure 6618. Thin layer of contrast material demonstrates calculus (arrow) in the
gallbladder. Cystic duct drains into an intrahepatic duct (arrowhead). Recognition of this
anomaly is important to the operating surgeon.
Calculi may occlude the cystic duct. This sometimes appears radiographically only as a stenosis, or the
flowing contrast medium may halt abruptly in a concave contour. If the calculi are small, the contrast
material may outline the entire stone or stones (Figure 6619). Occlusion of the cystic duct is clinically
significant if this finding is verified by adequate filling of the duct (Figure 6620), either by injection
under somewhat higher pressure or by selective catheterization and injection of the cystic duct. Calculi
within a cystic duct remnant after cholecystectomy are also detectable (Figure 6621). These may be
difficult to differentiate from a suture granuloma.
(6927)Figure 6619. Calculi (arrows) in the cystic duct.
(6928)Figure 6620. Occlusion (between arrowheads) of the cystic duct. Note the relative
stenosis of the common duct due to inflammation of the gallbladder.
(6929)Figure 6621. Calculus (arrow) in a cystic duct remnant in a patient who had a prior
cholecystectomy. Calculus was removed after endoscopic papillotomy.
The triad of cystic duct stone, cholecystitis, and benign stenotic occlusion of the bile duct with icterus is
termed the Mirizzi syndrome (see also Chapter 62: Calculus Disease of the Bile Ducts). The main ERC
findings are smooth stenosis of the common hepatic duct or common bile duct, or both, or a short,
rounded indentation.11(6930) Depending on the degree of stenosis, there may be dilation of the
intrahepatic ducts. In some instances, contrast medium may enter the gallbladder, and thereby a stone
and other abnormalities may be visualized. The main differential diagnosis is malignancy involving the
gallbladder or common bile duct.
Stents have been placed endoscopically into the cystic duct in patients with symptomatic
cholecystolithiasis who are poor candidates for surgery.1215(6931) Tamada et al.14(6932) treated 14
patients using this technique. Abdominal pain reportedly improved in 9 patients (64%) and fever
resolved in 5 patients with this finding. There were no procedure-related complications. Shrestha et
al.13(6933) placed stents from the duodenum across the cystic duct and into the gallbladder in 3
patients with advanced cirrhosis who were awaiting orthotopic liver transplantation. Symptoms resolved
in all cases; 2 patients underwent successful transplantation.
Chronic Cholecystitis
Cholelithiasis is the most common finding in chronic cholecystitis. Other abnormalities demonstrable by
ERC are a shrunken gallbladder and an irregular, rigid wall (Figure 6622). Pseudopolyps that produce
immobile filling defects may also be found. In some patients, there is occlusion, by either inflammation
or impacted calculi in the cystic duct, that makes opacification of the gallbladder difficult or impossible
(Figure 6623).
(6934)Figure 6622. Chronic, acalculus cholecystitis. Gallbladder is shrunken, and its wall
(arrowheads) is irregular.
(6935)Figure 6623. "Porcelain" gallbladder. Calcifications in the wall are seen en face and
in profile (arrowheads). There is occlusion of the cystic duct. Histologic study of the resected
specimen revealed chronic cholecystitis.
Tumors
Conventional radiologic examination in carcinoma of the gallbladder is difficult and neither sensitive nor
specific.16(6936) A direct cholangiogram as with ERC will outline the entire biliary duct system if there
is no occlusion. A tumor still localized to the gallbladder may be seen as a polyp or filling defect in a
stiff, rigid, and immobile gallbladder (Figures 6624 and 6625). However, the tumor may infiltrate and
occlude the cystic duct (Figure 6626). Often the common bile duct or the common hepatic duct, or
both, may be involved, with infiltration or stenoses of either duct being the only finding. Infiltration of the
liver is also frequent, and this may be detected with filling of the intrahepatic ducts (Figure 6626). The
presence of calculi, in either the gallbladder or the bile ducts, is a common finding in association with
carcinoma of the gallbladder (Figure 6626). A case of hemobilia due to gallbladder carcinoma has
been reported in which ERC disclosed a castlike filling defect in the common bile duct together with
several stringlike filling defects in the gallbladder.17(6937) An association exists between carcinoma of
the gallbladder and anomalous pancreaticobiliary ductal junction (see Chapter 63: Cystic Disorders of
the Bile Ducts).18(6938)
(6939)Figure 6624. Carcinoma of the gallbladder infiltrates the gallbladder and the cystic
duct (arrowheads) as well as the liver hilus (arrow).
(6940)Figure 6625. Carcinoma of the gallbladder indicated by an immobile, irregular filling

defect (arrowheads).
(6941)Figure 6626. Carcinoma of the gallbladder. Occlusions of the cystic duct by tumor
(large arrow) and calculus (arrowhead) in the common duct are noted. Displacement of the
intrahepatic bile ducts is also present (small arrow).
If carcinoma of the gallbladder is suspected, transpapillary biopsies or material for cytologic
examination should be obtained when possible. The diagnosis may also be established by guided
transcutaneous biopsy or cytology, and further definition of the mass lesion may be obtained by
computed tomography or ultrasonography. At ERC, direct inspection of the upper gastrointestinal tract
may occasionally reveal involvement by the cancer, particularly in the duodenum. In addition,
endoscopic ultrasonography appears to be valuable for the diagnosis as well as the staging of
gallbladder carcinoma (see Chapter 78: Endoscopic Ultrasonography of the Retroperitoneal
Organs).19(6942)
A benign tumor of the gallbladder is a rare finding during ERC. It is probably impossible to differentiate
such a lesion from a carcinoma by any method other than operation. There are no specific ERC
findings for benign gallbladder tumors. Another diagnostic challenge is displacement of the gallbladder,
which may represent developmental anomalies or benign or malignant processes in the vicinity of the
gallbladder. Laparotomy or laparoscopy may be warranted in these cases to establish a definite
diagnosis (Figure 6627).
(6943)Figure 6627. In a patient with chronic cholecystitis, the gallbladder is displaced and
fixed to the liver. Computed tomography demonstrated lesions thought to be indicative of a
liver tumor in close proximity to the gallbladder. At laparotomy, a benign, blood-filled liver
cyst was found.
Cholesterolosis and Adenomyomatosis of the Gallbladder

Morphologically, the polypoid form of cholesterolosis and the annular (segmental) and localized form of
adenomyomatosis are of interest with respect to ERC. The finding, which consists of one or more
persistent filling defects, may be the same for both diseases (Figure 6628). Radiologic views of such
lesions should be obtained en face as well as in profile. Compression must be utilized to determine
whether the gallbladder wall is stiff or pliable. Both cholesterolosis and adenomyomatosis are most
often found in association with cholelithiasis.
(6944)Figure 6628. These defects (arrowheads), which did not move with changes in patient
position, proved to be due to cholesterolosis of the gallbladder.
Trauma
ERC may occasionally be indicated for the patient who has sustained acute trauma. The gallbladder is
injured in less than 2% of cases of major blunt injury to the abdomen.20(6945) However, delayed
rupture has been reported.21(6946) This may occur directly into the peritoneal cavity and is diagnosed
by leakage of contrast medium from the gallbladder at ERC. The rupture may also be into the liver, and
this also may be recognized by ERC.21(6947) Perforation of the gallbladder may occur as a result of
percutaneous liver biopsy.
Fistula
Usually a fistula between the gallbladder and the neighboring organs occurs as the result of infection or
after operation. Occasionally, it may arise in conjunction with a carcinoma in the region. In one review,
diarrhea was the most common symptom, and clinical findings of cholangitis were unusual.22(6948)
The diagnosis is often confirmed by barium contrast x-rays of the colon, but when the location or
presence of fistula is in doubt, it may be necessary to obtain contrast medium studies of the biliary
system. After first achieving satisfactory opacification of the gallbladder (Figure 6629) or cystic duct,
the course, direction, and termination of the escaping contrast may be studied. Endoscopy may offer
the possibility of direct cannulation and inspection of the fistula itself (Figure 6630) if it terminates in
the duodenum or stomach. When carcinoma is suspected as the underlying cause, biopsy specimens
from the intestinal terminus of the fistula may be positive.
(6949)Figure 6629. Cholecystocolonic fistula. Wide fistula (between arrows) is present

between the shrunken gallbladder and the hepatic flexure of the colon. Numerous air bubbles
are present in the biliary ducts.
(6950)Figure 6630. Cholecystoduodenal fistula (arrowheads) demonstrated by direct

injection of contrast material through the fistula into the gallbladder.
Congenital Anomalies
Congenital abnormalities of the gallbladder occur rarely in humans (see also Chapter 58: Anatomy and
Embryology of the Biliary Tract and Pancreas). These include folding of the fundus of the gallbladder
(phrygian cap), diverticulum, double gallbladder (various forms), agenesis, bilobed, and multiseptate
gallbladder. It may be ectopic in position and can lie within the liver, under the left lobe (where the
cystic duct may drain into the left hepatic duct).23(6951) The case of a patient with a stricture of the
gallbladder, believed to be congenital, demonstrated by ERC has been reported.24(6952)
Agenesis of the gallbladder may be associated with other anomalies.25(6953) About half of patients
have symptoms suggestive of biliary colic, but in many cases, this is probably due to the presence of
stones in the bile duct. Approximately one third have only nonspecific dyspeptic symptoms.26(6954)
Absence of the gallbladder can be confirmed by ERC.27(6955)
The common hepatic duct and one or both of the hepatic ducts may drain into the
gallbladder.2830(6956) It may be difficult to differentiate this uncommon type of anomaly from a wide
variety of other disorders of the bile ducts, in which case ERC may be especially helpful.
Diseases of the Intrahepatic Bile Ducts

Calculi
Careful radiographic study of the intrahepatic ducts for calculi should be performed in patients with
cholelithiasis, choledocholithiasis, cholangitis, stenoses and anomalies of the bile ducts, chronic
hepatitis, and the different forms of cirrhosis. Diagnosis is based on the demonstration of movable
filling defects (Figures 6631 and 6632). An apparent filling defect in the column of contrast material
in the region of the liver hilum can be caused by the right hepatic artery as it crosses the bile duct. This
is sometimes termed a pseudostone, and it may be erroneously diagnosed as a tumor or
stone.31(6957) Large stones in the region of the hilus may be difficult or impossible to differentiate
from a tumor (Figure 6633). Biopsy specimens can be of critical importance in such cases (Figure
6634).
(6958)Figure 6631. Calculi (arrows) within the intrahepatic ducts.

(6959)Figure 6632. A, At least four calculi (arrowheads) are present proximal to a

postoperative stricture of the bile duct (arrow). B, Dilation of the stricture (arrow) with a
Grntzig-type hydrostatic balloon catheter. C,Cholangiogram after basket extraction of the
calculi.
(6960)Figure 6633. Large calculi (arrowheads) in the region of the hilus with dilated
intrahepatic ducts. Differentiation between a stone and a tumor is difficult in such a case.
(6961)Figure 6634. Biopsy specimen of a calculus that mimicked a ductal malignancy at

endoscopic retrograde cholangiography.
Primary Sclerosing Cholangitis

Primary sclerosing cholangitis (PSC) is often associated with inflammatory bowel diseases, usually
ulcerative colitis but also Crohn's disease. Diffuse, multifocal, short strictures with a patchy distribution
throughout the liver with involvement of both intrahepatic and extrahepatic ducts are the most common
radiographic findings (Figure 6635).32(6962)
(6963)Figure 6635. A, Typical cholangiographic features of sclerosing cholangitis include

irregular intrahepatic bile ducts and stenosis (arrowhead) of the common bile duct. B,After
dilation, the stenotic area (arrowhead) has been substantially widened. C, Biopsy from the
common bile duct reveals massive infiltration of mainly mononuclear cells and early-stage
lymphoid follicle formation. These findings indicate a high probability of sclerosing
cholangitis (H & E 200).
Some degree of focal dilation is often found, which results in the characteristic "beaded" appearance.
Slight irregularities in the wall of the intrahepatic ducts that produce a "shaggy" appearance may also
be seen. This is a relatively specific finding, although malignancy must be considered if the
irregularities appear more coarse. Dilation peripheral to stenotic segments has a focal character, and
diffuse dilation of the entire biliary tree is extremely uncommon. Some degree of pruning may occur.
Primary biliary cirrhosis (PBC) may cause cholangiographic changes similar to those of PSC, but in
general, stricturing is less pronounced and more peripheral in location, and the extrahepatic ducts are
usually not affected. In PBC, there may also be ductal changes of concomitant cirrhosis. Chen et
al.33(6964) compared retrograde cholangiograms obtained in 26 patients with PBC and 16 patients
with normal bile ducts. Radiographic abnormalities were found in half of the 24 technically satisfactory
studies in patients with PBC. The most commonly encountered findings were diminished arborization
and focal stenosis.
The case of a patient with hepatic sarcoidosis and severe cholestasis has been described in which
ERC disclosed findings that were typical for PSC.34(6965) Treatment with a corticosteroid drug
resulted in substantial resolution of the radiographic findings, including resolution of a dominant
stricture involving the right hepatic duct. There is an intriguing report of a case in which ingestion of
chaparral tablets resulted in painless jaundice, fatigue, and pruritus.35(6966) ERC demonstrated
smooth but severely narrowed bile ducts.
Secondary Cholangitis
Secondary cholangitis may be associated with most diseases of the biliary tree or livercholelithiasis,
cholangiocarcinoma, stricture, previous surgery in the region, and congenital abnormalities. The
findings of the underlying disease usually dominate the radiographic picture. The primary disease
process is usually located in the extrahepatic bile ducts and often causes dilation of these structures,
and consequently, dilation of the intrahepatic biliary tree is the most common finding in association with
secondary cholangitis.36,37(6967) In some patients, short stenoses and some degree of pruning may
be seen (Figure 6636).
(6968)Figure 6636. Secondary cholangitis due to gallstone disease. Stones were removed
endoscopically. Note stenosis and some degree of pruning of the intrahepatic ducts.
Bile Duct Carcinoma

About 4 to 8% of bile duct carcinomas arise from the intrahepatic ducts.38(6969) The usual
radiographic findings are those of a mass lesion. The size of the mass varies from a small discrete
nodule to larger lesions several centimeters in diameter. Thus, the appearance may be that of a small
filling defect, although the most common findings are stricturing, displacement, and occlusion of the
ducts (see Figure 6613). The stricture is often smooth and regular in contour (Figure 6637). A very
unusual and uncommon diffuse infiltrating variant of bile duct carcinoma is difficult to distinguish from
PSC. Cholelithiasis is often found in association with ductal carcinoma.
(6970)Figure 6637. Infiltration by cholangiocarcinoma (arrowhead) in the liver hilus. Right

lobe of the liver was previously resected.
Hepatocellular and Metastatic Carcinoma

It is frequently difficult to differentiate hepatic replacement and insufficiency from biliary obstruction in
patients with primary hepatocellular carcinoma or secondary metastasis to the liver who become
icteric. In the majority of cases, jaundice can be accounted for by replacement of the liver by malignant
tumor. In some cases, however, there may be a component of biliary obstruction, in which case ERC
and endoscopic stent placement may sometimes be beneficial as a method of palliation.39,40(6971)
Parasitic Diseases
One of the most common parasitic diseases encountered in ERC is the unilocular form of
Echinococcus granulosus. The cholangiographic findings are those of a mass lesion, although rupture
of the hydatid cyst into bile ducts has been reported.4143(6972) In the study of Spiliadis et
al.,44(6973) the most common ERC finding in 15 patients who had not previously undergone surgery
was bile duct compression (11 patients), followed by biliary fistula in 3 and echinococcal remnants
within the bile duct in 1 patient. In the study of Ulualp et al.43(6974) of 36 patients in whom the cyst
had ruptured into the biliary ducts, the most common symptoms were abdominal pain, jaundice, fever,
and chills. All patients underwent surgery that included choledochotomy and T-tube drainage in all
cases as well as a variety of surgical maneuvers directed at the cyst itself. Complications were
encountered in 7 cases. Five patients with cyst rupture into the bile duct were treated by Akkiz et
al.42(6975) by endoscopic sphincterotomy with extraction of daughter cysts; there were no
procedure-related complications and all patients recovered completely. Endoscopic sphincterotomy for
a variety of indications including cyst rupture into the bile duct was also found to be efficacious and
safe by Spiliadis et al.44(6976) External biliary fistulas related to surgical treatment of hydatid liver
disease were successfully treated by endoscopic sphincterotomy in 9 of 10 patients in the series of

Tekant et al.45(6977)
Clonorchis sinensis infestation is common in Asia. ERC findings are not fully established nor correlated
with the clinical stages of this disease. Cholangiography will typically demonstrate wavy or elliptical
filling defects of varying size as well as varying dilation of bile ducts (Figure 6638).46(6978)
(6979)Figure 6638. Clonorchis sinensis. Arrows mark filling defects produced by worms in
the bile ducts.
The presenting symptoms of ascariasis may be jaundice, biliary colic, or biliary infection. In children,
shock and acute abdominal pain may occur.47(6980) ERC may play a role in the evaluation of patients
with these signs and symptoms. The radiographic picture is a characteristic filling defect caused by the
adult worm (Figure 6639). Complications such as strictures, calculi, or abscesses may also be found.
Sometimes the worm may be seen in the papillary orifice during the endoscopic procedure.
(6981)Figure 6639. Ascariasis. Moving filling defects produced by worms are seen in both
the intrahepatic (small arrows) and the extrahepatic (large arrows) bile ducts. The worms
were later removed endoscopically.
Caroli's Disease
Most investigators now regard Caroli's disease as a variant of periportal fibrosis. This fibrosis results in
dilation of the large subsegmental intrahepatic ducts and produces a characteristic appearance
sometimes likened to a "lollipop tree" (Figure 6640). Smooth-walled, long stenotic ductal segments
may be present between dilated sections. Stone formation is common. Complications such as
cholangitis and abscess formation may result in superimposition of other changes on the primary
pattern and may make interpretation of the cholangiogram difficult. Endoscopic treatment has been
reported as beneficial in a few patients with Caroli's disease.48,49(6982)
(6983)Figure 6640. Caroli's syndrome. Subsegmental dilation of the intrahepatic ducts

(arrows) and stenoses (large arrowheads). A calculus is also noted (small arrowhead) within
one dilated segment.
Human Immunodeficiency Virus Infection

The liver and biliary tract are frequently involved with opportunistic infections (mycobacteria,
cytomegalovirus, and cryptosporidium) and neoplasms, most often Kaposi's sarcoma in patients with
human immunodeficiency virus (HIV) infection.50(6984) However, apart from moderate, nonspecific
changes in biochemical parameters of liver function, the hepatobiliary involvement often runs a quiet
clinical course, and the diagnosis is often coincidental. Acalculus cholecystitis and sclerosing
cholangitis are found in HIV patients (Figure 6641), and cryptosporidium and cytomegalovirus
infection appear to be associated with these conditions (Figure 6642).
(6985)Figure 6641. Sclerosing cholangitis with the typical, "ragged" appearance of the bile
ducts (arrowhead) in a patient infected with human immunodeficiency virus.
(6986)Figure 6642. Cryptosporidium. A, Biopsy from the common bile duct of the patient
shown in Figure 6641, obtained through the duodenoscope, shows severe chronic
inflammation with dilated crypts (H & E 80). B, Greater magnification from the same section
shows small basophilic bodies (arrowheads) related to the luminal side of the crypt epithelium
(H & E 300). Characteristic Cryptosporidia sporozoites were also found in colonic biopsies,
and the encysted form was found in the feces.
If the clinical course suggests extrahepatic cholestasis, ERC may be of value for diagnostic, as well as
for therapeutic purposes. Otherwise, extensive diagnostic procedures in the hepatobiliary system are
rarely warranted in these patients because manifestations of disease in other organs generally
dominate the clinical picture.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
Aabakken L, Karesen R, Serck-Hanssen A, Osnes M. Transpapillary biopsies and brush

cytology from the common bile duct. Endoscopy 1986;18:4951.
Harada H, Sasaki T, Yamamoto N, et al. Assessment of endoscopic aspiration cytology and
endoscopic retrograde cholangiopancreatography in patients with cancer of the hepatobiliary
tract. Part II. Gastroenterol Jpn 1977;12:5964.
Foerster EC, Schneider MU, Matek W, Domschke W. Transpapillary cholecystoscopy.
Foerster EC, Matek W, Domschke W. Endoscopic retrograde cannulation of the gallbladder:
Direct dissolution of gallstones. Gastrointest Endosc 1990;36:44450.
Edison SA, Maier M, Kohler B, et al. Direct dissolution of gallstones with methyl tert-butyl
ether by endoscopic cannulation of the gallbladder. Am J Gastroenterol 1993;88:12428.
Ponchon T, Baroud J, Mestas JL, Chayvialle JA. Gallbladder lithotripsy: Retrograde
dissolution of fragments. Gastrointest Endosc 1988;34:4689.
Soehendra N. Access to the cystic duct: A new endoscopic therapy for gallbladder diseases?.
Osnes M, Grondseth K, Larsen S, et al. Comparison of endoscopic retrograde and
intravenous cholangiography in diagnosis of biliary calculi. Lancet 1978;2:230.
Qang HL, Li ZZ, Sun YG. Reliability of ultrasonography in diagnosis of biliary lithiasis. Chin
Med J (Engl) 1990;103:63841.
Venu RP, Geenen JE, Toouli J, et al. Endoscopic retrograde cholangiopancreatography.
Diagnosis of cholelithiasis in patients with normal gallbladder x-ray and ultrasound studies.
JAMA 1983;249:75861.
Kauffmann GW, Hoppe-Seyler P, Noldge G, Waninger J. Mirizzi-syndrome. MMW Munch
Gaglio PJ, Buniak B, Leevy CB. Primary endoscopic retrograde cholecystoendoprosthesis: A
nonsurgical modality for symptomatic cholelithiasis in cirrhotic patients. Gastrointest Endosc
1996;44:33942.
Shrestha R, Bilir BM, Everson GT, Steinberg SE. Endoscopic stenting of gallbladder for
symptomatic cholelithiasis in patients with end-stage liver disease awaiting orthotopic liver
transplantation. Am J Gastroenterol 1996;91:5958.
Tamada K, Seki H, Sato K, et al. Efficacy of endoscopic retrograde cholecystoendoprosthesis
(ERCCE) for cholecystitis. Endoscopy 1991;23:23.
Kalloo AN, Thuluvath PJ, Pasricha PJ. Treatment of high-risk patients with symptomatic
cholelithiasis by endoscopic gallbladder stenting. Gastrointest Endosc 1994;40:60810.
Petterson H. Carcinoma of the gallbladder. Acta Radiol Diagn (Stockh) 1974;15:22536.
Osawa H, Mori Y, Inoue F. Case report: Malignant haemobilia detected in the
gallbladderRetrograde cholangiographic findings. Br J Radiol 1996;69:7981.
18. Chao TC, Jan YY, Chen MF. Primary carcinoma of the gallbladder associated with anomalous
pancreaticobiliary ductal junction. J Clin Gastroenterol 1995;21:3068.
19. Mitake M, Nakazawa S, Naitoh Y, et al. Endoscopic ultrasonography in diagnosis of the extent
of gallbladder carcinoma. Gastrointest Endosc 1990;36:5626.
20.
Mitre JR, Brodmerkel GJ Jr. Traumatic rupture of the gallbladder. Preoperative diagnosis by
endoscopic retrograde cholangiopancreatogram (ERCP). Gastrointest Endosc 1979;27:745.
21. Safrany L, van Husen N, Kautz G, et al. Endoscopic retrograde cholangiography (ERC) in
surgical emergencies. Ann Surg 1978;187:203.
22. Hession PR, Rawlinson J, Hall JR, et al. The clinical and radiological features of
cholecystocolic fistulae. Br J Radiol 1996;69:8049.
23. Szanto I, Voros A, Altorjay A, Kiss J. Gallbladder in the left side of the lower abdomen.
24. Jacob H, Fenster JS, Zinberg J, Kadish S. Stricture of the gallbladder. Am J Gastroenterol
1996;91:18623.
25. Goel AK, Seenu V, Khosla NK, et al. Agenesis of gallbladder with choledochal cystAn
unusual combination. Trop Gastroenterol 1994;15:336.
26. Richards RJ, Taubin H, Wasson D. Agenesis of the gallbladder in symptomatic adults. A case
and review of the literature. J Clin Gastroenterol 1993;16:2313.
27. Singh B, Moodley J, Haffejee AA, Rajaruthnam P. Laparoscopic diagnosis of gallbladder
agenesis. Surg Laparosc Endosc 1997;7:12932.
28. Hashmonai M, Kopelman D. An anomaly of the extrahepatic biliary system. Arch Surg
1995;130:6735.
29. Stringer DA, Dobranowski J, Ein SH, et al. Interposition of the gallbladderOr the absent
common hepatic duct and cystic duct. Pediatr Radiol 1987;17:1513.
30. Walia HS, Abraham TK, Baraka A. Gall-bladder interposition: A rare anomaly of the
extrahepatic ducts. Int Surg 1986;71:11721.
31. Baer JW, Abiri M. Right hepatic artery as a cause of pseudocalculus in the biliary tree.
32. MacCarty RL, LaRusso NF, Wiesner RH, Ludwig J. Primary sclerosing cholangitis: Findings
on cholangiography and pancreatography. Radiology 1983;149:3944.
33. Chen CC, Meng HC, Chang FY, et al. Characteristics of endoscopic retrograde
cholangiopancreatogram in patients with primary biliary cirrhosis. Chung Hua I Hsueh Tsa
Chih (Taipei) 1996;57:2549.
34. Alam I, Levenson SD, Ferrell LD, Bass NM. Diffuse intrahepatic biliary strictures in
sarcoidosis resembling sclerosing cholangitis. Case report and review of the literature. Dig Dis
Sci 1997;42:1295301.
35. Alderman S, Kailas S, Goldfarb S, et al. Cholestatic hepatitis after ingestion of chaparral leaf:
Confirmation by endoscopic retrograde cholangiopancreatography and liver biopsy. J Clin
36.
Rohrmann CA Jr, Ansel HJ, Ayoola EA, et al. Endoscopic retrograde intrahepatic
cholangiogram: Radiographic findings in intrahepatic disease. AJR 1977;128:4552.
37. Ayoola EA, Vennes JA, Silvis SE, et al. Endoscopic retrograde intrahepatic cholangiography in
liver diseases. Gastrointest Endosc 1976;22:1569.
38.
Berk RN, Ferrucci JT Jr, Leopold GRK. Radiology of the Gallbladder and Bile Ducts:
Diagnosis and Intervention. Philadelphia: WB Saunders. 1983.
39. Lau W, Leung K, Leung TW, et al. A logical approach to hepatocellular carcinoma presenting
with jaundice. Ann Surg 1997;225:2815.
40. Pereira-Lima JC, Jakobs R, Maier M, et al. Endoscopic stenting in obstructive jaundice due to
liver metastases: Does it have a benefit for the patient?. Hepatogastroenterology
1996;43:9448.
41. van Steenbergen W, Fevery J, Broeckaert L, et al. Hepatic echinococcosis ruptured into the
biliary tract. J Hepatol 1987;4:1339.
42. Akkiz H, Akinoglu A, Colakoglu S, et al. Endoscopic management of biliary hydatid disease.
Can J Surg 1996;39:28792.

43. Ulualp KM, Aydemir I, Senturk H, et al. Management of intrabiliary rupture of hydatid cyst of
the liver. World J Surg 1995;19:7204.
44. Spiliadis C, Georgopoulos S, Dailianas A, et al. The use of ERCP in the study of patients with
hepatic echinococcosis before and after surgical intervention. Gastrointest Endosc
1996;43:5759.
45. Tekant Y, Bilge O, Acarli K, et al. Endoscopic sphincterotomy in the treatment of
postoperative biliary fistulas of hepatic hydatid disease. Surg Endosc 1996;10:90911.
46. Leung JW, Sung JY, Chung SC, Metreweli C. Hepatic clonorchiasisA study by endoscopic
retrograde cholangiopancreatography. Gastrointest Endosc 1989;35:22631.
47. van Severen M, Lengele B, Dureuil J, et al. Hepatic ascaridiasis. Endoscopy 1987;19:1402.
48. Ciambotti GF, Ravi J, Abrol RP, Arya V. Right-sided monolobar Caroli's disease with
intrahepatic stones: Nonsurgical management with ERCP. Gastrointest Endosc
1994;40:7614.
49. Gold DM, Stark B, Pettei MJ, Levine JJ. Successful use of an internal biliary stent in Caroli's
50. Bonacini M. Hepatobiliary complications in patients with human immunodeficiency virus
infection. Am J Med 1992;92:40411.
Chapter 67 Physiology and Pharmacology of the Sphincter of Oddi

(6987)
(6988)
JAMES TOOULI, M.B.B.S., PH.D., F.R.A.C.S.
The sphincter of Oddi (SO) is a small, smooth muscle structure that is positioned at one of the busiest
junctions in the human gastrointestinal tract. Despite its small size (2 to 3 cm in length), the strategic
position of the SO at the junction of the bile duct, pancreatic duct, and duodenum makes knowledge of
its function important in the proper management of a variety of biliary and pancreatic disorders.
Although the presence of a thickening at the distal end of the bile duct had been described by early
anatomists, it was not until 1887 that Rugero Oddi1(6989) described in detail the anatomy of the
structure that bears his name and suggested that it may have a sphincteric function. In a subsequent
publication, Oddi2(6990) demonstrated that this sphincteric action controlled the flow of bile and further
postulated that malfunction might produce clinical disorders.
Study Techniques
Anatomy
Gross dissection of the extrahepatic biliary tract has provided information on the overall structure and
relationships of its various components.3(6991) Histologic studies have resulted in an appreciation of
the complex relationships between the biliary ducts and the related extrinsic nerves and blood vessels.
Identification of extrinsic neural projections has allowed the performance of experiments to define the
distribution of extrinsic as opposed to intrinsic neural elements.4(6992)
Standard histologic techniques have been used to define the relationships among the smooth muscle
layers, mucosal lining, serosa, neural plexuses, and the vessels. The whole-mount preparation has
also been used to examine nerves in the different layers of the viscus.5(6993) Using these
preparations, nerves may be identified by selective staining with silver and other stains. The mucosa of
the gallbladder and duodenum can be readily stripped from the underlying muscle layer, but in the bile
duct and SO, the mucosa is tightly adherent to the underlying layer, thereby preventing its separation
without damage.
Fluorescence immunohistochemistry techniques have been used on the biliary tracts of experimental
animals and some human specimens to define the presence and distribution of a variety of
peptide-containing neurons.6(6994) Detailed studies have been conducted in the Australian
brush-tailed possum*(6995) (Trichosurus vulpecula), systematically examining the gallbladder, bile
duct, and SO. Tissue specimens are fixed in Zamboni's solution (2% formaldehyde in 0.1 M phosphate
buffer plus 15% picric acid) and sectioned.4(6996) Antisera to a variety of available neuropeptides are
applied and labeled with fluorescein coupled to specific antibodies, thus allowing observation of peptide
localization with a fluorescence microscope. Further advances in these techniques have allowed
double labeling of nerves, and colocalization of neuropeptides in the nerve supply of the biliary tract
has been demonstrated.7(6997)
The distribution and relationships of the blood supply to the biliary tract have been determined by
vessel-casting techniques.8(6998) Arteries and veins are infused separately with a colored resin that
hardens and is resistant to corrosion. The surrounding tissue can then be removed by tissue-corrosive
agents, leaving the architecture of the vessels intact. The distribution of the layers of small vessels in
the biliary tract is defined by examination of the casts. Other methods of defining the vasculature of the
biliary tract include injection of India ink9(6999) and the use of radiopaque solutions that allow
examination by radiographic methods.10(7000)
Motility Studies
In Vitro Techniques
The biliary tract, as with other parts of the gut, readily lends itself to detailed electrical and
pharmacologic study using in vitro techniques. Tissue specimens of standard size from gallbladders,
cystic duct, bile duct, SO, and adjacent duodenum have been studied by immersion in a bath of
oxygenated Tyrode's or Kreb's solution.11(7001) Under these conditions, the tissue remains viable for
up to 12 hours and exhibits spontaneous contractile activity. The tissue is suspended at a fixed length
that is established by a predetermined weight, and motility is recorded by force transducers (Figure
671).
(7002)Figure 671. Diagram of a full-thickness strip of the sphincter of Oddi (SO) suspended
in a bath of Kreb's solution. The isometric force transducer records changes in tone.
Field stimulation has been applied to such specimens by the immersion of electrodes in the bath.
Electrical stimulation may selectively activate either the nerves or the smooth muscle. Using a variety
of pharmacologic agonists or antagonists, the function of various nerves associated with the specimen
can be determined. Pharmacologic experiments have been conducted using such in vitro techniques to
determine the dose-response relationships for a number of potential agonists in the biliary
tract.12(7003)
The simultaneous recording of changes in tone and electrical activity from the smooth muscle has
revealed the relationship between contraction and electrical events. Intracellular recordings from
smooth muscle cells have also been made, and these demonstrate that the muscle of the biliary tract
behaves very much like that in other areas of the gastrointestinal tract.13(7004)
Manometry of the Gallbladder and Sphincter of Oddi
The recording of pressure changes within the lumen of the extrahepatic biliary tract has been the most
important technique in advancing our understanding of biliary motility. Such manometric techniques
have been used to record the motility of the gallbladder, bile duct, and SO, in both humans and a
variety of experimental animals.
Human Manometry
The measurement of intraluminal pressure changes within the human biliary tract was pioneered by
surgeons recording pressures during operations on the biliary tract.14(7005) Pressure measurements
from the bile ducts have been obtained by inserting a tube into the bile duct by way of the cystic duct.
The pressure changes have been recorded either directly from fluid-filled manometers or by means of
a transducer linked to a polygraph. The fluid manometric technique allows for recording of the opening,
passage, and closing pressures of the SO by either increasing or decreasing the height of a reservoir
connected in series with the tube inside the bile duct.15(7006)
Intraluminal pressure measurements of SO contractions became possible only when manometry
catheters were miniaturized so that their introduction into the SO lumen produced minimum distortion
of SO activity. In addition, the development of manometric systems of low compliance made it possible
to accurately record pressures.16(7007) Polyethylene or Teflon catheters with three lumens of 0.5 mm
internal diameter are used. The outer diameter of these catheters varies from 1.5 to 1.7 mm, calibers
that have been shown to produce little or no distortion of the SO.17(7008) Three side-holes are placed
2 mm apart starting at 10 mm from the distal tip. Therefore, pressures are recorded via the three
lumens across a length of 5 mm from within the SO. The catheter is 200 cm long and is perfused with
deionized, bubble-free water by means of a pneumohydraulic capillary system at a pressure of 750 mm
Hg. Water is perfused through the catheter at 0.125 ml/min; the entire system is capable of accurately
recording pressure changes of up to 300 mm Hg/sec (Figure 672). In some centers, normal saline
solution is used as the perfusing fluid; this does not alter the recording fidelity of the system. However,
the catheter must be thoroughly washed between recordings to ensure that the small-diameter lumens
are not blocked by encrustations of salt. The recording from three lumens allows the evaluation of the
progressive nature of SO contractions as well as the determination of basal pressure, amplitude, and
frequency of contractions.18(7009)
(7010)Figure 672. Equipment used for biliary manometry. A triple-lumen catheter is

constantly perfused by means of a pneumohydraulic pump set at a pressure of 750 mm Hg. For
endoscopic manometry, the tip of the catheter is introduced into the bile duct through the
accessory channel of a side-viewing endoscope. SOsphincter of Oddi; DUOduodenum.
Recordings of human SO motility may be made under three different clinical circumstances.
The most common technique, used for diagnosis of SO motility disorders, involves the introduction of
the catheter during endoscopic retrograde cholangiopancreatography.19(7011) The patient is lightly
sedated and the oropharynx anesthetized with a topical agent to assist passage of the duodenoscope.
The manometry catheter is introduced via the accessory channel, through the papilla, and into either
the bile duct or the pancreatic duct (see Figure 672). The catheter is then withdrawn, so that all three
recording ports are situated within the SO. To assist in the positioning of the catheter across the SO,
catheters are marked with circular bands at 2-mm intervals from the most proximal hole over a total
length of 12 mm. Recordings are made for approximately 3 to 10 minutes. It is difficult to record for
much longer by the endoscopic technique because patients often become intolerant of the procedure.
A second manometry catheter may be attached to the endoscope for simultaneous recording of
duodenal pressure charges. The manometric recordings are scored and pressures reported relative to
duodenal pressure, which is taken as the zero baseline. The SO basal pressure is the pressure at the
bottom of the pressure waves above duodenal pressure. The amplitude of the phasic contractions is
the pressure at the peaks above the basal pressure. Propagation of the phasic contractions is
determined by drawing a line through the onset of the wave recorded at each of the three lumens.
Because the locations of the three lumens on the catheter are known, it can be determined whether
the waves propagate antegrade or retrograde or whether the contractions are simultaneous.
During elective operation for gallstones, recordings from the SO may be made by introducing the
triple-lumen catheter through the cystic duct, into the bile duct, and through the SO into the
duodenum.17(7012) The catheter is then withdrawn, so that the three lumens are recording from within
the SO (Figure 673). The duration of recording may be greater than that obtained at endoscopy, but
the fact that the patient is anesthetized is a major disadvantage. The effects of anesthetic agents on
biliary motility are unknown. However, in studies where manometry of the SO was obtained in both
anesthetized and awake patients, SO motility did not differ significantly when nonopiate anesthetics
were used.20(7013)
(7014)Figure 673. Intraoperative SO manometry. The triple-lumen manometry catheter is

introduced through the cystic duct. The three recording lumens of the catheter are positioned
within the SO. CBDcommon bile duct; PDpancreatic duct. (From Toouli J, Bushell M,
Iannos J, et al. Peroperative sphincter of Oddi manometry: Motility disorder in patients with
cholelithiasis. Aust N Z J Surg 1986; 56:6259.)
Prolonged studies of human SO motility can be made by introducing the manometry catheter through
the T-tube in patients who have undergone exploration of the bile duct for stones or via a percutaneous
tract.21,22(7015) The manometry catheter is passed into the bile duct, through the SO, and into the
duodenum. It is then withdrawn to position the side-holes within the SO. This is confirmed
fluoroscopically and by the characteristic changes on the pressure tracings as the catheter is
withdrawn from the duodenum into the SO. Recordings can be made for as long as 8 hours. This
technique is useful for physiologic studies on the human SO, but it has little practical value as a
diagnostic test. It has the disadvantage that, by necessity, only patients who have had stones in the bile
duct or other biliary disorders can be studied.
Animal Manometry
Pressure changes from within the biliary tract have been studied in a variety of animal species
including dogs, cats, possums, opossums, rabbits, pigs, prairie dogs, and guinea pigs. The methods
used are similar to those described previously for humans. Two techniques are used to record the
resistance to flow across the SO of animals in acute experiments performed under anesthesia. In one
technique, the catheter is inserted by way of the bile duct, so that it is jammed above the SO. The
catheter is ligated in this position, and water is perfused. The resistance to flow is measured by a
pressure transducer connected to the catheter.23(7016) In the other technique, flow across the SO is
measured by recording the volume change in a reservoir that connects to a catheter ligated in the bile
duct proximal to the SO.24(7017) The rate of change in volume in the reservoir maintained at a
physiologic pressure head determines flow through the SO. Outflow from the bile duct has been
measured using a similar concept, where the fluid is collected via a funnel sutured over the opening.
By means of these methods, a variety of neuropeptides and their action on the SO have been studied.
Prolonged recordings of pressure have been made from conscious pigs after insertion of a
nonperfused manometry catheter in the gallbladder and a triple-lumen perfused catheter in the SO.
Studies carried out in these animals can elucidate changes in motility after ingestion of food and the
infusion of a variety of hormones.25(7018)
An ingenious method for studying gallbladder tone has been developed by adaptation of the gastric
barostat technique. The gallbladder barostat consists of a polyurethane bag that is connected by a
double-lumen polyvinyl tube to a strain gauge and a syringe air injection system. The bag is inserted
into the gallbladder via a cannula that has been previously implanted in the animal gallbladder and
exteriorized. The bag is inflated, so that it takes up the gallbladder volume and is maintained at a
constant low pressure (2 mm Hg). At this pressure, it is possible to register gallbladder wall contraction
and relaxation by recording the changes in the volume of the bag. The studies are performed by
occluding the cystic duct by means of a previously implanted inflatable hydraulic occluding device, thus
allowing study of gallbladder motility changes without the influence of changes in SO motility.26(7019)
Electromyography
Extracellular bipolar or monopolar electrodes sutured to the outside surface of the gallbladder and to
the SO in animals have been used to record contractions of the biliary tract.27(7020) The American
opossum is a particularly useful animal model for this technique because the SO has a long
extraduodenal segment that is readily accessible for the recording of electrical activity. Bipolar
electrodes sutured to the SO and gallbladder of the American opossum have shown that, as in other
parts of the gastrointestinal tract, the SO exhibits two forms of myoelectrical activity: slow regular
changes in membrane potential, known as electrical slow waves, and rapid depolarization spikes,
called spike bursts. The latter are virtually always associated with phasic smooth muscle contractions
and therefore produce transient intraluminal pressure changes.28(7021)
Recordings of electrical signals from the human SO have been made by introducing electrodes into the
lumen of the SO via the accessory channel of a duodenoscope. These recordings resemble those
obtained in experimental animals and are believed to reflect contractions of the human SO.29(7022)
Ultrasonography
Techniques that make use of ultrasonography have been developed for measuring the volume of the
gallbladder and the diameters of the biliary and pancreatic ducts. After stimulation, either by eating or
by hormone infusion, changes in these measurements can be followed up, thus providing a
noninvasive means of evaluating the motility of the biliary and pancreatic ductal systems in clinical
situations.
Human gallbladder volume is determined by measuring the maximum length and maximum transverse
diameter of the gallbladder. The volume is then calculated from the ultrasonographic images as the
sum of a series of cylinders.30(7023) Serial measurements of volume changes can reveal the effects
of various stimuli on gallbladder volume.
The maximum diameter of the bile duct can also be measured by ultrasonography. A fixed point
adjacent to the point where the right hepatic artery crosses the common hepatic duct has been taken
as the point of maximum diameter. After the patient ingests a fatty meal, it is expected that normal
changes in SO motility will enhance bile flow into the duodenum, and so, the diameter of the common
hepatic duct should decrease. An increase in ductal diameter is regarded as reflecting a resistance to
outflow that may be indicative of SO dysfunction.31(7024)
The diameter of the pancreatic duct as determined by ultrasonography in normal human subjects is
less than 1 mm. After infusion of secretin, there is an increase in ductal diameter followed by a rapid
return to normal diameter within 30 minutes. If the ductal diameter remains in excess of 1.5 mm after
30 minutes, this suggests an increase in resistance to the outflow of pancreatic juice that may
represent dysfunction of the SO.32(7025)
These ultrasonographic techniques for evaluating biliary motility have been developed as noninvasive
methods for diagnosis of SO dysfunction. Their specificity, sensitivity, and usefulness as diagnostic
tools are under evaluation at present.
Isotope Studies
Technetium-99m-iminodiacetic acid (99mTc-IDA) compounds administered intravenously are cleared
through the liver into bile. 99mTc gives off 140-keV photons that are ideally suited for imaging and
counting by a camera. These properties of the 99mTc-IDA compounds have been used to image the
gallbladder33(7026) and the biliary tract; by producing temporal profiles, monitoring of gallbladder
emptying and flow across the bile duct in subjects after cholecystectomy can be accomplished. Flow
through the bile duct has been determined in patients after cholecystectomy to define the
characteristics of flow across the SO.34(7027) After the injection of 99mTc-IDA, the subject is
positioned under a camera and continuous recordings are made from the liver, biliary tract, and
duodenum. To analyze flow, areas of interest are outlined and the counts per second within those
areas are determined and graphically displayed. The shape and slope of the curve of counts from the
bile duct reflect flow through the biliary tree. Delay in outflow may represent increased resistance to
flow across the SO owing to SO dysfunction.
Anatomy of the Biliary System

Gross Anatomy
Embryology
The biliary ducts and gallbladder arise from the caudal portion of a diverticular anlage that originates
from the ventral floor of the foregut; the pancreas arises from two foregut buds in the region of the
future duodenum (see Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas).
Anatomic dissections have shown that the distal muscularis propria of the bile duct and pancreatic duct
are independent from the duodenal musculature.35(7028) In studies of the human fetus,
Boyden35(7029) demonstrated that the SO musculature arises de novo from mesenchyme, appearing
approximately 5 weeks after the intestinal musculature.
Morphology
The intrahepatic canaliculi communicate with numerous interlobular ducts; these in turn drain into two
main hepatic ducts. The main right and left hepatic ducts fuse at the porta hepatis into the common
hepatic duct; the cystic duct joins the common hepatic duct at a variable distance caudal to the porta
hepatis to form the common bile duct.36(7030)
The human gallbladder is situated in a fossa on the right inferior surface of the liver and is anatomically
divided into the blunt fundus, the body, and the neck, which leads to the cystic duct. A sacculation at
the neck of the gallbladder is known as Hartman's pouch. The cystic duct is of variable length, usually
joining the common hepatic duct at an acute angle to form the common bile duct.
The common bile duct passes dorsal to the first part of the duodenum, lying in a groove either within or
posterior to the head of the pancreas.37(7031) It enters the second part of the duodenum through the
major duodenal papilla in association with the main pancreatic duct. The junction of the terminal
common bile duct, pancreatic duct, and duodenum at the papilla may be in one of three configurations
that can be likened to a Y, V, or U. In approximately 70% of subjects, the ducts open into a common
channel and thus have a Y-shaped configuration. This common channel drains into the duodenum
through a single orifice on the main duodenal papilla. In approximately 20% of subjects, the common
channel is almost nonexistent and the two ducts have a V-shaped configuration as they approach the
opening on the papilla. In 10% of subjects, the common bile duct and pancreatic duct have separate
openings on the tip of the papilla; the adjacent openings therefore have a U-shaped configuration. The
terminal parts of the common bile duct and pancreatic duct, the common channel, and the main
duodenal papilla are invested by smooth muscle of varying thickness, and together these muscle coats
form the SO.
The human SO lies mainly within the duodenal wall, but it has been shown to be distinctly separate
from the wall. Boyden,38(7032) in a series of publications on the anatomy of the SO, described distinct
sphincters at the terminal end of the common bile duct (sphincter choledochus), the terminal end of the
pancreatic duct (sphincter pancreaticus), and the common channel (sphincter ampullae). However,
Hand,39(7033) using a combination of radiographic methods, duct-casting techniques, and histologic
sectioning methods, could not distinguish separate sphincters. His studies showed that the common
bile duct and the pancreatic duct become fused in a common connective tissue sheath outside the
duodenal wall and then pass together through a slit in a duodenal muscle called the choledochal
window. The lumina, however, do not join at this level but are separated by a thick muscular septum. In
most subjects, fusion of the two lumina occurs in the submucosal layer of the duodenum to form a
common channel varying in length from 2 to 17 mm. Before entering the duodenum, each duct
becomes completely surrounded by circular muscle, some of which forms a figure-of-eight pattern
around the two ducts. The junction point between the smooth muscle and each duct is readily identified
radiographically as a notch. Distal to the notch, each lumen becomes narrow as it traverses the
duodenal wall; this narrowing is associated with a thickening of the duct wall that consists of smooth
muscle, connective tissue, and mucous glands. As the ducts pass through the duodenal wall,
longitudinal muscle fibers interdigitate between the circular muscle fibers of the ducts and the duodenal
muscle. The ducts emerge from the duodenal muscle layers to follow a course of variable length
through the duodenal submucosa before opening on the main duodenal papilla. Throughout this
submucosal course, the ducts are ensheathed by circularly oriented smooth muscle bundles.
Manometric studies in humans support Hand's description of the SO in that separate sphincteric zones
were not identified.18(7034)
The mucosa of the human SO segment is lined by a columnar epithelium and contains numerous
mucus-secreting glands. It is thrown into longitudinal folds that have been called mucosal
valvules.40(7035) These folds are least prominent proximally and become more conspicuous distally,
being maximum in the common channel. The mucosal folds may occasionally be seen projecting
through the orifice of the duodenal papilla when it is viewed at endoscopy.
Investigators have used a variety of experimental animals to study the motility of the biliary tract and its
controlling mechanisms. It should be noted, however, that the anatomy of the biliary tract differs
between species, and these differences probably reflect differences in function.
Some species, such as the horse, rat, pigeon, and pocket gopher, do not have gallbladders. The SO is
also rudimentary in these species.41(7036) In dogs, pigs, and rabbits, the bile duct and pancreatic duct
enter the duodenum separately. The SO is present in these species and, as in humans, is mainly
contained within the wall of the duodenum (Figure 674).38(7037) The guinea pig bile duct and
pancreatic duct enter a common ampulla situated within the duodenal wall. This animal has the most
prominent ampulla of any species, and a distinct SO is present at its junction with the duodenal papilla.
The American opossum and the Australian possum have an extraduodenal SO (Figure 675). In both
of these marsupials, the bile and pancreatic ducts join to form a common channel before reaching the
duodenal wall. In the American opossum, however, circular muscle surrounds both ducts, whereas in
the Australian species, the arrangement is that of a figure-of-eight, thus creating a muscular septum
between the bile duct and the pancreatic duct.
(7038)Figure 674. Schematic representation of the human SO. Most of the SO is situated
within the wall of the duodenum. The part of the SO surrounding the terminal bile duct is
longer than that surrounding the terminal pancreatic duct. (From Toouli J. Sphincter of Oddi
motility. Br J Surg 1984; 71:2516, by permission of Blackwell Science Ltd.)
(7039)Figure 675. Schematic representation of the opossum SO. The SO is extraduodenal in

position. The bile duct and pancreatic duct join external to the wall from a common opening
into the duodenum. (From Toouli J. Sphincter of Oddi motility. Br J Surg 1984; 71:2516, by
permission of Blackwell Science Ltd.)
The SO in cats, chimpanzees, and monkeys appears to most closely approximate the human anatomy.
In these species, the bile duct and the pancreatic duct join to form a short common channel that is
surrounded by a sphincter that lies largely within the wall of the duodenum.
Innervation
The extrahepatic bile ducts are richly supplied by both extrinsic and intrinsic nerves. Histologic and
immunohistochemical studies have shown the extensive distribution of ganglia and nerve fibers
throughout the different layers of the biliary system.
Extrinsic Nerves
The biliary tract is supplied by extrinsic nerves of the parasympathetic and sympathetic systems.
Parasympathetic nerves reach the upper abdomen by way of the vagus nerves. In excess of 90% of
fibers in the vagi are afferent and 10% efferent, whereas 10% of all of these fibers supply the biliary
tract.42(7040) The cell bodies of the vagal efferent fibers are situated in the medulla oblongata, and
the afferent cell bodies are found in the jugular and nodose ganglia.43(7041) In addition to the
cholinergic nerves in the vagi, a variety of peptide-containing fibers are present. Identified peptides
include substance P (SP), somatostatin, vasoactive intestinal polypeptide (VIP), galanin (GAL), and
enkephalin (ENK).44(7042)
The sympathetic fibers that supply the gastrointestinal tract and biliary tract project from the thoracic
and lumbar spinal cord to the prevertebral sympathetic ganglia. In these ganglia are found the cell
bodies of the majority of noradrenergic fibers that supply the gut. In general, the distribution of
noradrenergic fibers to the biliary tract follows the pattern of the blood supply.45(7043)
Parasympathetic and sympathetic extrinsic nerves innervating the biliary tract form an anterior and
posterior hepatic plexus that is situated in the porta hepatis, anterior and posterior to the bile duct,
hepatic artery, and portal vein. Contributions to these plexuses are made from the anterior and
posterior vagi and from the celiac ganglion. The relative proportions in the mix of fibers from each of
these sources is uncertain. A further sympathetic contribution to the biliary tract arises from the
superior mesenteric ganglion and supplies the SO by way of the inferior pancreaticoduodenal artery.
Intrinsic Nerves
Studies in a number of different species have demonstrated variations in the architecture and
distribution of intrinsic nerves of the biliary tract. The gallbladder is innervated via three ganglionated
plexuses: an outer subserosal, an intramuscular, and an innermost plexus in the lamina propria. This
latter plexus has not been found in all species studied.46(7044) Nerve fibers in the gallbladder wall are
densely distributed along a subepithelial plexus; a lesser concentration of fibers is found in the muscle
coat with fibers oriented in the direction of the muscle bundles. Fluorescence histochemistry has
demonstrated nonadrenergic fibers and nerves with cholinesterase reactivity among these
fibers.47(7045) Furthermore, a relatively rich distribution of VIP- and SP-containing nerves has been
described.48(7046) Other peptides, neuropeptide Y (NPY), somatostatin (SOM), gastrin-releasing
peptide (GRP), GAL, and ENK have also been described in a less concentrated
distribution.49,50(7047)
The common bile duct also is extensively supplied with nerves, but their distribution is not as well
organized as in the gallbladder.51,52(7048) A prominent feature of the bile duct nerve supply is the
presence of two or more large nerve trunks that run along its length in the submucosal region. These
trunks may be important in the coordination of motor activities of the gallbladder and the SO.
Histochemical studies of nerves to the bile duct have demonstrated noradrenergic fibers and some
small cholinesterase-containing ganglia.53(7049) In addition, sparsely situated VIP, ENK, SOM, NPY,
and GRP fibers have been identified.54(7050)
The SO is extensively innervated by a dense collection of ganglia and nerve fibers. The SO contains
two ganglionated plexuses: an outer plexus between the muscle layers and a submucosal
plexus.47,55(7051) Most studies have described connections between these plexuses and those of the
duodenum and the gallbladder. These connections may have important functions in the control of SO
activity. In addition, a dense distribution of nerve fibers supply the SO muscle as well as blood vessels
to the SO. Noradrenergic and cholinesterase-containing fibers have been described.56(7052)
Furthermore, dense innervation of the muscle by nerves containing VIP, SP, GRP, NPY, SOM, and
ENK has been demonstrated.48,49,57(7053)
Blood Supply
The arterial blood supply of the biliary tract is via the hepatic artery and the gastroduodenal artery.
Anatomically, these vessels exhibit marked variation in origin and relations to the biliary tract. However,
they feed an extensively anastomosing arterial plexus that ensheaths the biliary tract and from which
the intramural blood supply is derived.
The blood supply of the extrahepatic biliary tract is divided relative to the bile duct into anterior and
posterior components that anastomose freely via numerous connecting plexuses.58(7054) Anterior to
the bile duct, two to five ascending vessels arise from the retroduodenal branch of the gastroduodenal
artery. Three or four vessels originate from the right hepatic artery and the cystic artery and
anastomose with the ascending arteries. In particular, the ascending and descending vessels form two
prominent arterial channels at the left and right of the bile duct; these arteries are the major source of
the axial blood supply to the biliary tract. Thus, the major (60%) axial blood supply of the biliary tract is
derived from the gastroduodenal artery; 38% comes from the right hepatic artery and 2% from other
sources.59(7055)
A retroportal artery posterior to the bile duct has been described that arises from either the celiac axis
or the superior mesenteric artery and runs cephalad on the posterior aspect of the portal vein. It has a
variable course, most commonly ending by joining the retroduodenal artery close to the distal end of
the extraduodenal bile duct. Less frequently, it joins the right hepatic artery.
The position of these major arteries in relation to the bile duct is important in that injury during surgery
may cause portions of the bile duct to become ischemic, thereby leading to stricture formation.
Because most of the blood supply is derived from the right hepatic artery, enteric anastomoses to the
bile duct should be constructed as far proximally along the course of the duct as possible in order to
avoid stricture formation.
The intrinsic blood supply of the bile duct comprises arterioles that penetrate the wall to form two rich
plexuses of capillaries.60(7056) The most superficial plexus lies between the outer fibrous sheath and
the lamina propria. The other plexus is situated directly under the mucosa. Both of these plexuses form
an extensive capillary network of fine vessels.61(7057) The site and configuration of this vascular bed
suggest that it is probably involved in physiologic absorption of substances from the lumen of the bile
duct. Of practical importance to interventional endoscopists is a small branch of the gastroduodenal
artery that supplies the SO. This arteriole may be divided during endoscopic sphincterotomy.
Motility of the Biliary Tract

The gallbladder, bile ducts, and SO and their respective functions are integrated in a motor action that
normally promotes the orderly flow of bile from the liver to the duodenum. This activity is finely
controlled during both the interdigestive period and the response to a meal stimulus.
Gallbladder and Cystic Duct

The human gallbladder has a fluctuating volume during the interdigestive period.62(7058) Its response
to a meal stimulus is a steady contraction that slowly delivers bile through the bile duct and SO into the
duodenum.
Common Bile Duct

Data from most studies lead to the conclusion that the human bile duct does not exhibit primary
propulsile activity. This is consistent with the histology of the bile duct; it comprises mainly elastic fibers
and longitudinally oriented smooth muscle cells that are suited to providing a tonic pressure but not
lumen-occluding contractions.11(7059) This tone may help to overcome the tonic resistance to flow
offered by the SO.
The diameter of the bile duct may change as a result of a change in resistance to outflow of bile
provided by the SO. After cholecystectomy, bile duct diameter does not change
significantly.63,64(7060) Therefore, the finding of an abnormally dilated duct in the absence of a clear
pathologic cause reflects increased resistance across the SO owing to a motor disturbance of the SO.
Sphincter of Oddi
Manometric recordings from the SO demonstrate prominent pressure peaks representing phasic
contractions (Figure 676) that are superimposed on a low and stable basal pressure (tone). The
function of these phasic contractions and their role in promoting flow from the bile duct and pancreas
into the duodenum vary between different species.
(7061)Figure 676. A manometric recording from the human SO with a triple-lumen catheter.
The SO is characterized manometrically by prominent phasic contractions superimposed on a
low and stable basal pressure. (From Toouli J. Biliary tract motor dysfunction. In Dent J [ed].
Practical Issues in Gastrointestinal Disorders. Vol. 5, No. 2 of Baillire's Clinical
Gastrointestology. London: Baillre Tindall, 1991; 40930.)
The opossum has been a useful animal model in the effort to understand the relationship between bile
flow and contractions of the SO.65(7062) Simultaneous cineradiography and electromyography to
register contractions at the same time as flow have demonstrated that flow occurs both passively
between phasic contractions and by active peristaltic propulsion (Figure 677). In the opossum, the SO
relaxes and bile flows from the bile duct into the SO segment. Some of this bile then flows passively
from the SO segment into the duodenum, but the major flow is accomplished by an active propulsive
contraction that sweeps across the length of the SO segment, thereby emptying its contents into the
duodenum. This cycle then repeats. Most of the net flow occurs through active propulsion, although a
significant amount of bile still flows passively.
(7063)Figure 677. Schematic representation of a cineradiographic sequence illustrates the

flow of contrast medium from the common bile duct (CBD) through the sphincter of Oddi
(SO) as drops of contrast medium from the reservoir into the CBD. (From Toouli J, Dodds
WJ, Honda R, et al. Motor function of the opossum sphincter of Oddi. J Clin Invest 1983;
71:20820.)
In humans and dogs, most of the flow from the bile duct into the duodenum occurs between the
contractions, hence passively.21(7064) Phasic contractions expel a relatively small proportion of the
bile that flows across the SO segment.14,66(7065) Inevitably, the phasic contractions produce a
resistance to flow between the bile duct and the duodenumeach time the SO contracts, the outflow
from the bile duct is occluded. However, this effect is no more obstructive than that of the ventricle of
the heart in obstructing blood flow from the atria every time it contracts. The overall effect of the phasic
contractions of the SO is that of promoting flow from the ducts into the duodenum.
The frequency of SO contractions demonstrates a periodicity during the interdigestive cycle or
migrating motor complex (MMC). Unlike the small intestine, SO contractions do not pass through a
Phase I of absolute quiescence; hence, they are omnipresent. In the opossum, the frequency of
contractions gradually increases during Phase II and culminates in a short burst of high-frequency
contractions just before duodenal Phase III activity.28(7066) In humans, SO contractions are
constantly present during all phases of the interdigestive cycle.21(7067) The SO frequency remains
constant until just before Phase III of the duodenal MMC, when it increases to approximate the
frequency of duodenal contractions. This burst of maximum activity lasts throughout the length of
Phase III of the duodenal MMC and then returns to its normal contractile frequency after the passage
of Phase III of the duodenal MMC (Figure 678).
(7068)Figure 678. Histogram demonstrates the frequency of human SO contractions in

relation to the duodenal interdigestive activity (phases I to IV). SO contractions are present
throughout the interdigestive period. The SO contraction frequency increases just before
duodenal phase III. (From Worthley CS, Baker RA, Iannos J, et al. Human fasting and
post-prandial sphincter of Oddi motility. Br J Surg 1989; 76:70914, by permission of
Blackwell Science Ltd.)
The responses of the opossum SO and the human SO after a meal differ in a way that may reflect the
different functions of the contractions in these two species. In the opossum, an increase in contractile
frequency enhances flow across the SO. In humans, there is no alteration in the frequency but the
amplitude of the contractions decreases, so that the diastolic interval between contractions is
increased. In addition, the SO basal pressure decreases. The overall effect is the promotion of passive
flow of fluid from the ducts to the duodenum. It is significant, however, that the SO contractions are not
abolished, so that even in a weakened form they tend to keep the SO segment free of debris and to
promote flow toward the duodenum.
It is becoming evident that the enteric nervous system has an important role in the control of SO
activity. The functional significance of the extrinsic nerves is not as yet clearly defined. Vagotomy has
been shown in some studies to increase resistance to flow across the SO;67(7069) in other studies,
little effect was seen.68(7070) Studies that involve stimulation of the vagus or sympathetic nerves have
not demonstrated reproducibly consistent results. This may suggest that the extrinsic nerves set only a
background tone for the superimposed contractile events that are modulated at a peripheral site by the
actions of intrinsic nerves and hormones.
Studies in cats and guinea pigs have provided evidence for neural communication among the
gallbladder, cystic duct, and SO, as well as between the gastric antrum and the SO. Distention of the
gallbladder has been shown to promote relaxation of the SO by way of a reflex mediated by these
connections.24(7071) A similar reflex action has also been demonstrated in humans. In patients
undergoing cholecystectomy, SO manometry was performed using a triple-lumen catheter introduced
via the cystic duct. Distention of the gallbladder produced an inhibition in the amplitude of the SO
phasic contractions and a fall in basal pressure.69(7072) This may represent a normal physiologic
response of the SO to increased intraluminal pressure within the gallbladder. Relaxation of the SO
would promote flow and reduce bile duct intraluminal pressure.
A variety of gastrointestinal hormones influence the motility of the SO. Prominent among these is
cholecystokinin (CCK), which has been shown to relax the SO and promote flow. In cats, CCK inhibits
SO phasic contractions23(7073) and reduces basal pressure by an effect on nonadrenergic
noncholinergic inhibitory nerves. Its action in humans is believed to occur through a similar mechanism
that produces inhibition of phasic contractions and a fall in SO basal pressure.70(7074) The actions of
other gastrointestinal hormones on the SO are not so well established as that of CCK. However,
secretin, glucagon, histamine, gastrin, VIP, and SP have all been shown to alter SO contractions in
doses that may apply in the physiologic state.7173(7075) Kalloo and Pasricha74(7076) demonstrated
by endoscopic manometry that intraduodenal infusion of fat in normal subjects resulted in a significant
decrease in mean SO basal pressure; distention of the stomach with air had no appreciable effect.
A variety of smooth muscle-active agents have an effect on the SO. Morphine increases the resistance
to flow across the SO and has been shown manometrically to produce an increase in phasic
contraction frequency and a rise in basal pressure. At nonanalgesic doses, comparable concentrations
of meperidine do not have an excitatory action, but instead they inhibit the amplitude of the SO phasic
contractions.75(7077) These differences in action of the two opiate analgesics support the
recommendation that meperidine be used in preference to morphine for alleviation of biliary colic. The
cholinergic antagonist atropine also has an inhibitory effect on the SO, producing a fall in basal
pressure and a reduction in phasic contraction amplitude. A similar effect is produced by the calcium
channel blocker nifedipine and the smooth muscle relaxant amyl nitrate. Topical application of glyceryl
trinitrate to the main duodenal papilla reduced both tonic and phasic contractions of the SO in the study
of Luman et al.76(7078) In another study by Staritz et al.,77(7079) sublingual glyceryl trinitrate resulted
in a decrease in amplitude of SO contractions as well as baseline pressure. In one study, midazolam
was found to have no significant effect on the SO.78(7080) In another investigation, intravenous
administration of midazolam had no effect in patients with normal manometric pressures, but SO
pressure was reduced in patients with initially high readings.79(7081) The somatostatin analog
octreotide was shown to increase basal SO pressure and frequency of contractions and to decrease
wave amplitude in a study involving a small number of patients with abdominal pain.80(7082)
Endoscopic Sphincter of Oddi Manometry

The recording of pressures from the bile duct and SO is the most objective method currently available
for evaluating the motility of the SO. A number of manometric studies have been conducted in normal
individuals, and in general, there is widespread agreement as to the normal values for SO pressures
(Table 671).19,8183(7083) In a study conducted in our unit, the reproducibility of endoscopic SO
manometry was assessed in 12 patients who underwent repeat studies over a 3-month
period.84(7084) This study showed that endoscopic SO manometry is reproducible and therefore
suitable for making the diagnosis of SO stenosis or establishing the presence of normal motility.
However, reproducibility was poor in the diagnosis of SO dyskinesia. This is not surprising because the
dyskinetic disorders may be episodic in nature. In addition, SO manometry, conducted via endoscopy,
can sample the motility of the SO over a brief period only, hence episodic abnormalities of the SO may
be missed.
Normal Values and Ranges for Sphincter

of Oddi Manometry in Humans
TABLE 671
NORMAL
Basal pressure (mm Hg)
15 (535)
Amplitude (mm Hg)
135 (95195)
Frequency (n/min)
4 (26)
Sequences
Antegrade (%)
80 (12100)
Simultaneous (%)
13 (050)
Retrograde (%)
9 (050)
CCK-OP 20 ng/kg
Inhibits
CCK-OPcholecystokinin octapeptide.
ABNORMAL
>40
>300
>7
>50
Contracts
Endoscopic SO manometry is an invasive technique and is thus associated with a significant morbidity.
In one series, the frequency of pancreatitis after SO manometry was almost 20%.85(7085) In our
series, mild pancreatitis occurred in 8% of patients, the majority of whom were being assessed for
idiopathic recurrent pancreatitis. A similar experience has been reported by Rolny et al.86(7086) in a
study from Sweden. In many centers, including ours, it is common practice to perform endoscopic SO
manometry and diagnostic endoscopic retrograde cholangiopancreatography at separate sessions on
different days. This avoids overinjection or overperfusion of the pancreatic duct and might account for
the lower incidence of pancreatitis in our experience.
Summary
During fasting, approximately 50% of secreted hepatic bile passes into the gallbladder where it is
concentrated by absorption of water. Resistance to flow through the SO is produced by the basal
pressure and by phasic contractions that modulate the flow of bile into either the gallbladder or the
duodenum. During the interdigestive period, the expulsion of small volumes of concentrated bile from
the gallbladder occurs just before Phase III of the duodenal MMC. This action of the gallbladder is
associated with an increased frequency of SO contractions, which may promote the flow of small
volumes of bile into the duodenum. After a meal, the gallbladder undergoes a slow sustained
contraction, so that most of its contents are expelled into the duodenum. The SO contractions diminish
in amplitude, allowing an enhancement of the passive flow of bile.
The resting gallbladder tone depends on an intact vagal innervation, whereas SO contractions are
influenced by intrinsic cholinergic mechanisms. The changes in gallbladder volume and SO activity
during the interdigestive cycle are associated with changes in serum motilin concentrations, but a
causal relationship has not been demonstrated. The alterations in biliary tract motility after a meal are
mainly produced through the action of CCK released from the duodenum. However, cephalic
stimulation by way of the vagus has an initial effect, and circumstantial evidence suggests that gastric
and duodenal distention may also alter both gallbladder and SO motility through local reflexes. The
motility of the SO is best studied via endoscopic manometry, which allows the recording of SO
pressures. Normal values have been determined and appear to be highly reproducible at different
centers and at different times of study.
Abnormalities in biliary tract motility have been described in both the gallbladder and the SO. These
disorders of biliary tract motility are discussed further in Chapter 68: Sphincter of Oddi Stenosis and
Dysfunction.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
Oddi R. D'une disposition a sphincter speciale de l'ouverture du canal cholidoque. Arch Ital
Biol 1887;8:31722.
Oddi R. Sulla tonicita dello sfintere del coledoco. Arch Sci Med 1888;12:3339.
Smadja C, Blumgart LH. The biliary tract and the anatomy of biliary exposure. In Blumgart LH
(ed). Surgery of the Liver and Biliary Tract. Vol. 1. New York: Churchill Livingstone,
1988;1122.
Costa M, Furness JB, Llewellyn-Smith IJ. Histochemistry of the enteric nervous system. In
Johnson LR (ed). Physiology of the Gastrointestinal Tract. 2nd ed. New York: Raven,
1987;140.
Costa M, Furness JB. Immunohistochemistry on wholemount preparations. In Cuello AC (ed).
Methods of Neuroscience. Vol 3: Immunohistochemistry. New York: John Wiley & Sons,
1983;37398.
Padbury RTA, Furness JB, Toouli J. Intrinsic catecholamine containing neurones in the
sphincter of Oddi and duodenum of the Australian possum. Gastroenterology 1989;96:381.
Costa M, Furness JB, Gibbins IL. Chemical coding of enteric neurones. In Hokfeldt T, Fuxe K,
Pernow B (eds). Progress in Brain Research. New York, Amsterdam, London: Elsevier
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
Science (Biomedical Division), 1986;68.

Northover JMA, Williams EDF, Terblanche J. The investigation of small vessel anatomy by
scanning electron microscopy of resin casts. J Anat 1980;130:4354.
Shapiro AL, Robillard GL. The arterial blood supply of the common and hepatic bile ducts with
reference to the problems of common duct injury and repair. Surgery 1948;23:111.
Munkacsi I, Siklos I. Blood supply of the common bile duct. Acta Morphol Hung
1962;11:17988.
Toouli J, Watts JM. In-vitro motility studies on the canine and human extrahepatic biliary
tracts. Aust N Z J Surg 1971;40:3807.
Helm JF, Dodds WJ, Christensen J, Sarna SK. Control mechanism of spontaneous in vitro
contractions of the opossum sphincter of Oddi. Am J Physiol 1985;249:G5729.
Bauer AJ, Go VLW, Koch TR, Szurszewski J. Non-adrenergic, non-cholinergic (NANC)
inhibitory innervation of the canine and opossum sphincter of Oddi. Gastroenterology
1987;92:1311.
Hess W. Physiology of the sphincter of Oddi. In Classen M, Geenen J, Kawai K (eds). The
papilla Vateri and its diseases. Proceedings of the International Workshop of the World
Congress of Gastroenterology held in Madrid. Kohn, Gerhard Witzshock, 1978;1421.
Cuschieri A, Hughes JH, Cohen M. Biliary pressure studies during cholecystectomy. Br J Surg
1972;59:26773.
Arndorfer RC, Steff JJ, Dodds WJ, et al. Improved infusion system for intraluminal
esophageal manometry. Gastroenterology 1977;73:237.
Saccone GTP, Toouli J, Iannos J, et al. The effect of catheter diameter on sphincter of Oddi
motility. Surg Res Commun 1988;2:31521.
Toouli J, Geenen JE, Hogan WJ, et al. Sphincter of Oddi motor activity: A comparison
between patients with common bile duct stones and controls. Gastroenterology
1982;82:1117.
Geenen JE, Hogan WJ, Dodds WJ, et al. Intraluminal pressure recording from the human
sphincter of Oddi. Gastroenterology 1980;78:31724.
Toouli J, Bushell M, Iannos J, et al. Peroperative sphincter of Oddi manometry: Motility
disorder in patients with cholelithiasis. Aust N Z J Surg 1986;56:6259.
Worthley CS, Baker RA, Iannos J, et al. Human fasting and post-prandial sphincter of Oddi
motility. Br J Surg 1989;76:70914.
Akita Y, Nimura Y, Yasui A. Percutaneous transhepatic manometry of sphincter of Oddi. Dig
Dis Sci 1991;36:14107.
Behar J, Biancani P. Effect of cholecystokinin and the octapeptide of cholecystokinin on the
feline sphincter of Oddi and gallbladder. Mechanisms of action. J Clin Invest 1980;66:12319.
Thune A, Thornell E, Svanvik J. Reflex regulation of flow resistance in the feline sphincter of
Oddi by hydrostatic pressure in the biliary tract. Gastroenterology 1986;91:13649.
McIntosh GH, Baker RA, Worthley CS, et al. Motility of the biliary tract in a chronic pig model.
Surg Res Commun 1988;4:93102.
Tanaka M, Sarr MG, Miller LJ, Malagelada JR. Changes in tone of canine gallbladder
measured by an electronic pneumatic barostat. Am J Physiol 1990;259:G31220.
Sarles JC, Midejean A, Devaux MA. Electromyography of the sphincter of Oddi. Am J
Honda R, Toouli J, Dodds WJ, et al. Relationship of sphincter of Oddi spike bursts to
gastrointestinal myoelectric activity in conscious opossums. J Clin Invest 1982;69:7708.
Bortolotti M, Caletti GC, Brocchi E, et al. Endoscopic electromyography and manometry of the
human sphincter of Oddi. Hepatogastroenterology 1985;32:2502.
Everson GT, Braverman DZ, Johnson ML, Kern F Jr. A critical evaluation of real-time
ultrasonography for the study of gallbladder volume and contraction. Gastroenterology
1980;79:406.
Darweesh RMA, Dodds WJ, Hogan WJ, et al. Efficacy of quantitative hepatobiliary
scintigraphy and fatty meal sonography for evaluating patients with suspected partial common
bile duct obstruction. Gastroenterology 1988;94:77986.

32. Bolondi L, Gaiani S, Gullo L, Labo G. Secretin administration induces a dilatation of main
pancreatic duct. Dig Dis Sci 1984;29:8028.
33. Shaffer EA, McOrmond P, Duggan H. Quantitative cholescintigraphy: Assessment of
gallbladder filling and emptying and duodenogastric reflux. Gastroenterology
1980;79:899906.
34. Roberts-Thomson IC, Toouli J, Blanchett W, et al. Assessment of bile flow by
radioscintigraphy in patients with biliary-type pain after cholecystectomy. Aust N Z J Med
1986;16:78893.
35.
1957;104:64152.
36.
Dowdy GS, Waldron GW, Brown WG. Surgical anatomy of the pancreatico-biliary system:
Observations. Arch Surg 1962;84:22946.
37.
Lytle WJ. The common bile duct groove in the pancreas. Br J Surg 1959;47:20912.
38.
Boyden EA. The sphincter of Oddi in man and certain representative mammals. Surgery
1937;1:2537.
39.
1963;50:48694.
40. Tansy MF, Salkin L, Innes DL, et al. The mucosal lining of the intramural common bile duct as
a determinant of ductal opening pressure. Am J Dig Dis 1975;20:61325.
41.
Hallenbeck GA. Biliary and pancreatic intraductal pressures. In Code CF (ed). The Handbook
of Physiology. Sect 6: The Alimentary Canal. Vol 2: Secretion. Washington, DC: American
Physiological Society, 1967;100725.
42.
Harkins HN, Stavney LS, Griffith CA, et al. Selective gastric vagotomy. Ann Surg
1963;158:44860.
43.
Davison JS. Innervation of the gastrointestinal tract. In Christensen J, Wingate PL (eds). A
Guide to Gastrointestinal Motility. Vol 1. New York: Wright, 1983;47.
44. Lundberg JM, Hokfelt T, Nilsson G, et al. Peptide neurones in the vagus, splanchnic and
sciabi nerves. Acta Physiol Scand 1978;104:499501.
45.
Bingham JR, Ingelfinger FJ, Smithwick RH. The effects of sympathectomy on the motility of
the human gastrointestinal and biliary tracts. Gastroenterology 1950;15:617.
46.
Keast JR, Furness JB, Costa M. Distribution of certain peptides containing nerve fibres and
endocrine cells in the gastrointestinal mucosa in five mammalian species. J Comp Neurol
1985;236:41322.
47. Cai WQ, Gabella G. The musculature of the gallbladder and biliary pathways in the guinea
pig. J Anat 1983;136:23750.
48. Cai WQ, Gu J, Huang W, et al. Peptide immunoreactive nerves and cells of the guinea-pig
gallbladder and biliary pathways. Gut 1983;24:118693.
49. Allen JM, Gu J, Adrian TE, et al. Neuropeptide Y in the guinea-pig biliary tract. Experientia
1984;40:7657.
50. Cai WQ, Gabella G. Innervation of the gallbladder and biliary pathways in the guinea-pig. J
Anat 1983;136:97109.
51.
Alexander WF. The innervation of the biliary system. J Comp Neurol 1940;72:35770.
52.
Burnett W, Cairns FW, Bacsich P. Some observations on the innervation of the extrahepatic
biliary system in man. Ann Surg 1964;159:826.
53. Kyosola K. Adrenergic and cholinergic innervation of the supraduodenal common bile duct.
54. Dancygier H, Klein U, Leaschner U, et al. Somatostatin containing cells in the extrahepatic
biliary tract of humans. Gastroenterology 1984;86:8926.
55.
Van Buskirk C, Boyden EA. The distribution of nerve cells and fibres to the
choledochalduodenal junction of the cat. Anat Rec 1944;88:423.
56.
Mori J, Azuma H, Fujiwara M. Adrenergic innervation and receptors in the sphincter of Oddi.
Eur J Pharmacol 1979;14:36573.
57. Alumets J, Fahrenkrug J, Hakanson R, et al. A rich VIP nerve supply is characteristic of
sphincters. Nature 1979;280:15556.
58. Northover J, Terblanche J. Bile duct blood supply: Its importance in human liver
transplantation. Transplantation 1978;26:679.
59. Northover J, Terblanche J. A new look at the arterial supply of the bile duct in man and its
surgical implications. Br J Surg 1979;66:37984.
60.
Parke WW, Michels NA. Blood supply of the common bile duct. Surg Gynecol Obstet
1963;177:4755.
61. Cho KJ, Lunderquist A. The peribiliary vascular plexus: The microvascular architecture of the
bile duct in the rabbit and in clinical cases. Radiology 1983;147:35764.
62. Toouli J, Bushell M, Stevenson G, et al. Gallbladder emptying in man related to fasting
duodenal migrating motor contractions. Aust N Z J Surg 1986;56:14751.
63.
Lequesne LP, Whiteside CG, Hand BH. The common bile duct after cholecystectomy. BMJ
1959;1:32932.
64. Hunt DR, Scott AJ. Changes in bile duct diameter after cholecystectomy. A 5 year prospective
study. Gastroenterology 1989;97:14858.
65. Toouli J, Dodds WJ, Honda R, et al. Motor function of the opossum sphincter of Oddi. J Clin
Invest 1983;71:20820.
66. Watts JM, Dunphy JE. The role of the common bile duct in biliary dynamics. Surg Gynecol
Obstet 1966;122:120781.
67. Pitt HA, Doty JE, Roslyn JJ, DenBesten L. The role of altered extrahepatic biliary function in
the pathogenesis of gallstones after vagotomy. Surgery 1981;90:41825.
68.
Tansy MF, Innes DL, Martin JS, Kendall FM. An evaluation of neural influences on the SO in
the dog. Dig Dis Sci 1974;19:42337.
69. Thune A, Saccone GT, Scicchitano JP, Toouli J. Distension of the gall bladder inhibits
sphincter of Oddi motility in humans. Gut 1991;32:6903.
70. Toouli J, Hogan WJ, Geenen JE, et al. Action of cholecystokinin-octapeptide on sphincter of
Oddi basal pressure and phasic wave activity in humans. Surgery 1982;92:497503.
71. Carr-Locke DL, Gregg JA, Aoki TT. Effects of exogenous glucagon on pancreatic and biliary
ductal and sphincteric pressures in man demonstrated by endoscopic manometry and
correlation with plasma glucagon. Dig Dis Sci 1983;28:31220.
72. Dahlstrand C, Bjorck S, Edin R, et al. Substance P in the control of extrahepatic biliary motility
in the cat. Regul Pept 1988;20:1124.
73. Dahlstrand C, Dahlstrom A, Ahlman H. Adrenergic and VIPergic relaxatory mechanisms of
the feline extrahepatic biliary tree. J Auton Nerv Syst 1989;26:97106.
74. Kalloo AN, Pasricha PJ. Effect of gastric distension and duodenal fat infusion on biliary
sphincter of Oddi motility in healthy volunteers. Dig Dis Sci 1995;40:7458.
75. Thune A, Baker RA, Saccone GT, et al. Differing effects of pethidine and morphine on human
sphincter of Oddi motility. Br J Surg 1990;77:9925.
76. Luman W, Pryde A, Heading RC, Palmer KR. Topical glyceryl trinitrate relaxes the sphincter
of Oddi. Gut 1997;40:5413.
77. Staritz M, Poralla T, Ewe K, Meyer zum Buschenfelde KH. Effect of glyceryl trinitrate on the
sphincter of Oddi motility and baseline pressure. Gut 1985;26:1947.
78. Cuer JC, Dapoigny M, Bommelaer G. The effect of midazolam on motility of the sphincter of
Oddi in human subjects. Endoscopy 1993;25:3846.
79. Rolny P, Arleback A. Effect of midazolam on sphincter of Oddi motility [see comments].
Endoscopy 1993;25:3813.
80. Binmoeller KF, Dumas R, Harris AG, Delmont JP. Effect of somatostatin analog octreotide on
human sphincter of Oddi. Dig Dis Sci 1992;37:7737.
81. Guelrud M, Mendoza S, Rossiter G, Villegas MI. Sphincter of Oddi manometry in healthy
volunteers. Dig Dis Sci 1990;35:3846.
82. Funch-Jensen P, Kruse A, Ravnsbaek J. Endoscopic sphincter of Oddi manometry in healthy
volunteers. Scand J Gastroenterol 1987;22:2439.
83. Carr-Locke DL, Gregg JA. Endoscopic manometry of pancreatic and biliary sphincter zones in
man. Basal results in healthy volunteers. Dig Dis Sci 1981;26:715.
84. Thune A, Scicchitano J, Roberts-Thomson I, Toouli J. Reproducibility of endoscopic sphincter
of Oddi manometry. Dig Dis Sci 1991;36:14015.
85.
King CE, Kalvaria I, Sninsky CA. Pancreatitis due to endoscopic biliary manometry: Proceed
with caution. Gastroenterology 1988;94:A227.
86. Rolny P, Anderberg B, Ihse I, et al. Pancreatitis after sphincter of Oddi manometry. Gut
1990;31:8214.
Chapter 68 Sphincter of Oddi Stenosis and Dysfunction

(7087)
(7088)
RAMA P. VENU, M.D., F.A.C.P., F.A.C.G.
JOSEPH E. GEENEN, M.D.
W. J. HOGAN, M.D.
The sphincter of Oddi (SO), located strategically at the choledochoduodenal junction, modulates the
flow of bile and pancreatic secretion into the duodenum. Previously, because of its inaccessibility, the
structure and function of the SO in humans were of marginal interest to clinicians and clinical
investigators. With the introduction of fiberoptic instruments and the development of advanced
diagnostic and therapeutic capabilities, this once-remote region of the gastrointestinal tract has been
"rediscovered." This chapter is not a comprehensive presentation of the SO's complex role as
integrator of pancreaticobiliary flow dynamics; rather it highlights anatomic, physiologic, and
pathophysiologic features that are relevant to clinical problems encountered by the gastrointestinal
endoscopist who performs endoscopic retrograde cholangiopancreatography (ERCP). Much of the
information included in this section is based on the results of clinical and laboratory studies of the SO.
This chapter endeavors to offer the clinician a practical understanding of the function and dysfunction
of the SO.
Sphincter of Oddi and Related Structures

Anatomy and Structure
Pancreaticobiliary Tract
The right and left hepatic ducts are the culmination of an extensive network of progressively larger
tributaries originating within the liver that transport bile from the hepatocytes. These two mains ducts
fuse at the portal fissure to form the common hepatic duct. Four centimeters distal to the origin of the
hepatic duct, the cystic duct branches off as a continuation of the gallbladder. From this juncture, the
common bile duct becomes the continuation of the hepatic duct. The common bile duct coursing
downward encounters the posterior aspect of the pancreas, angles to the right, and enters the second
portion of the duodenum. The anatomy of the biliary ducts and related structures is described in
Chapter 58: Anatomy and Embryology of the Biliary Tract and Pancreas.
The intramural portion of the common bile duct ranges from 10 to 30 mm in length. The confluence of
the distal common bile duct and the pancreatic duct usually occurs just before the ductal structures
pierce the duodenal wall. The common bile duct and main pancreatic duct merge into a common
channel in over 80% of patients. Two or more orifices in the same papilla or separate papillae for bile
and pancreatic ducts occur rarely. The dimension of the ampulla varieslength ranges from 2 to 17
mm, diameter from 2 to 7 mm.
Sphincter of Oddi
There is an interchange of fibers between the duodenal muscle layers and the intrinsic musculature of
the distal bile duct as it traverses the wall of the duodenum. These connecting and reinforcing fibers
are highly variable. Boyden's1(7089) meticulous dissection of the human SO demonstrated the
existence of an intrinsic choledochal muscle that is anatomically distinct from the surrounding duodenal
musculature. There is now ample physiologic evidence to indicate that a zone of intrinsic electric and
contractile activity exists at the distal choledochus in humans and most other animals.
At anatomic dissection, this intrinsic choledochal muscle is approximately 14 mm in length.2(7090) It is
divided into two sphincteric zones: a superior choledochal sphincter encircling the bile duct just
proximal to the duodenum and an inferior choledochal sphincter muscle surrounding the submucous
portion of the duct.3(7091) The terminal pancreatic duct is also surrounded by a portion of the
sphincter that is continuous with that of the common bile duct (Figure 681). Despite this anatomic
arrangement that suggests separate sphincteric zones, manometric studies of the human SO have
failed to identify discrete small sphincters or "minisphincters" corresponding to those found by
dissection. A contiguous zone of contractile activity and baseline pressure elevation is recorded over a
ductal length of 6 to 8 mm extending proximally into the common duct or pancreatic duct from the
aperture of the main duodenal papilla. This segment is referred to as the sphincter of Oddi
zone.4(7092)
(7093)Figure 681. Schematic diagram depicts the relative anatomic arrangement of the
intrinsic musculature of the sphincter of Oddi (SO), in which the presence of sphincter zones
is suggested. (Adapted from Hess W. Manometry and radiography in the biliary system during
surgery. In Demling L, Classen M [eds]. Endoscopic Sphincterotomy of the Papilla of Vater.
Proceedings of the International Workshop of the World Congress of Gastroenterology,
Munich, 1976. Stuttgart: Georg Thieme, 1978.)
Embryology
The muscularis proprius of the distal choledochus first appears in the human fetus 5 weeks after the
intestinal musculature. It arises as a linear wall from the mesenchyme. In his definitive study of the
embryology of the SO, Boyden1(7094) remarked that the distal choledochus is a different order of
musculature than the other structures. It originates at the anatomic point where it subsequently
penetrates the duodenal wall. Although the SO segment appears to be embryologically and
anatomically distinct from the duodenal musculature, there is confirmation of an intricate networking
between the distal choledochus and the duodenum.5(7095)
Innervation
The extrahepatic bile ducts are innervated by sympathetic nerves originating from spinal segments 7 to
10 that reach the celiac ganglia via the splanchnic nerves.6(7096) The splanchnic nerves supply both
motor and inhibitory fibers. Parasympathetic innervation is derived from the vagal nerves, which supply
both motor and sensory fibers. At the hiatus of the liver, parasympathetic fibers from both the right and
the left vagal nerves form the hepatic plexus, which subsequently supplies parasympathetic innervation
to the extrahepatic bile ducts. The bulk of afferent nerve supply from the extrahepatic biliary system is
believed to pass via the sympathetic afferents through both splanchnic nerves and also through the
right phrenic nerve in many cases.
The choledochoduodenal junction is innervated by both components of the autonomic nervous system
via the gastroduodenal nerve, which ends in a plexus containing numerous ganglion cells within the SO
zone. Adrenergic nerves appear to be more sparse, but their fibers may directly innervate smooth
muscle cells in the same pattern as that described for the gut wall. The intrinsic cholinergic nerve
apparatus of the SO appears richly developed in comparison with that of the gut. The ganglion cells of
the myenteric plexus and submucosal plexus appear to have a predominance of cholinergic nerves.
Nitric oxide synthase has been localized to neural elements in the human SO,7(7097) and
cholecystokinin (CCK) receptors that mediate relaxation have been demonstrated on the human
SO.8(7098)
Blood Supply
The major blood supply to the papillary region is derived from the retroduodenal artery, a branch of the
inferior pancreaticoduodenal artery. Stolte9(7099) described variations of the arterial supply to the
papillary region using combined angiographic and histologic studies of 50 postmortem specimens
(Figure 682). A ventral and a dorsal branch of the retroduodenal artery form the arterial plexus of the
papilla in over 50% of cases. The dorsal branch predominated in 25% and the ventral branch was
prevalent in 8%. In another 8% of cases, the arterial plexus was made up of small lateral vessels
arising in the wall of the duodenum. In most specimens, the retroduodenal artery was located more
than 35 mm from the papillary orifice. In 5%, the artery resided within the range of endoscopic
papillotomy incision length. This latter variation, therefore, poses a threat of severe hemorrhage during
endoscopic sphincterotomy (ES). Doppler probes have been shown to be helpful in localizing the
retroduodenal artery.10(7100) This technique has not gained wide clinical utilization.
(7101)Figure 682. Diagram of blood supply in the region of the papilla of Vater. (Adapted
from Stolte M. Some aspects of the anatomy and pathology of the papilla of Vater. In Classen
M. Greenen J, Kawai K [eds]. The Papilla of Vateri and Its Diseases. Baden-Baden, Koln,
New York: Witzstrock, 1979.)
Study Techniques
There have been investigators of biliary tract function for more than 4 centuries. In 1543, Vesalius
noted a membrane at the orifice of the common bile duct that he postulated impeded bile flow and
prevented regurgitation of duodenal contents. A possible sphincter mechanism at the distal
choledochus was first described in 1879. Eight years later, Rugero Oddi11(7102) performed such a
thorough analysis of this tiny bundle of circular muscle fibers at the confluence of the choledochus and
pancreatic ducts in animals that his name is given to this structure.
At the turn of this century, attention was focused on research into the mechanism of bile delivery into
the duodenum. Many renowned physicians including Krukenberg, Borghi, Aschoff, Meltzer, Judd, and
Mann studied gallbladder motor function and its interaction with the SO and bile flow. This set the
stage for considerable speculation about the possibility that some clinical symptoms and disease
states were related to disordered motor function of the biliary tract.
Experimental design, instrumentation, and methodology were relatively crude by present standards but
improved during the decade from 1920 to 1930. During this period, humoral control mechanisms were
suggested as primary modulating factors in biliary tract dynamics, particularly gallbladder evacuation of
bile. For instance, it was found that specific food substances introduced into the duodenum prompted
evacuation of bile and increased the contractile force of the gallbladder. Subsequently, Ivy and
Oldberg12(7103) extracted a substance from hog mucosa that stimulated the evacuation of bile from
the canine gallbladder. Ivy called this substance cholecystokinin. Other investigators confirmed its
status as a true hormone shortly thereafter.
In the next half century, important advances in experimental design and technology enabled more
precise investigation and measurement of parameters of biliary tract dynamics in humans. Intravenous
cholangiography, cineradiography, operative and postoperative cholangiography, and various
manometric recording techniques have been used to study pressure and flow dynamics of the
extrahepatic biliary system. Some of these investigative techniques are described in the following
section.
Cholangiographic Techniques
Bile flow kinetics have been studied by cholangiographic techniques that involve opacification of the
bile duct by intravenously administered contrast medium or by injection of contrast medium through a
cannula at laparotomy or postoperatively. Radiomanometry consists of a sequence of bile duct
radiographs obtained at low hydrostatic pressure, sufficient to initiate common bile duct outflow (usually
<15 cm H2O), and at high hydrostatic pressure, approximately 40 to 50 cm H2O.
Bile duct flow is measured by a drop counter during infusion of the common bile duct through a
cannula. At a given pressure, flow is determined by the resistance of the SO. It is assumed that the
inflow rate of the infusate reflects the rate of outflow into the duodenum. Using a gravity fluid reservoir,
the resting choledochal pressure and passage pressure (i.e., the pressure at which the sphincter
yields) can be determined. Choledochal basal pressure of 5 to 10 cm H2O and yield pressure of 12 to
15 cm H2O have been measured. A contrast infusion pump has been added to the system, and
variable infusion rates have been used to study sphincter activity. An infusion rate of 10 ml/min
associated with the flow resistance in the tubing of 25 to 30 mm Hg is commonly used. By substituting
contrast material, continuous radiographic monitoring of the biliary tree can be performed during a
period of dye injection. Average choledochal resting pressure of 6 + 4 mm Hg has been measured in
the human bile duct. Cineradiographic imaging has shown phasic wave pressure activity, independent
of respiration or heartbeat, in the SO segment. Phasic pressure waves with an average amplitude of 6
mm Hg have been recorded. These were not accompanied by significant changes in intraduodenal
pressure or contractile activity.13(7104)
A series of pressure filling curves have been determined from these constant infusion studies of the
biliary tract. The data obtained have been interpreted as reflecting SO dynamics, specifically that high
common bile duct pressure indicates excessive SO contraction, whereas low common bile duct
pressures is indicative of little or no resistance by the sphincter to the flow of material into the
duodenum. Unfortunately, direct measurement of SO motor activity and pressure has not been
correlated with this type of investigation.
ERCP Manometry
The direct measurement of the human SO motor function with a pressure recording catheter is a
logical evolution of diagnostic ERCP. Pressures from the distal choledochus were first measured by
this method in 1974,14(7105) and in 1977, distinctive phasic wave contractions were described in the
SO zone using a minimally compliant catheter perfusion system.15(7106)
Pressure in the common bile duct, pancreatic duct, and the SO segment are recorded using a
triple-lumen polyethylene catheter (outside diameter 1.7 mm), with a luminal diameter of 0.5 mm, and a
length of 200 cm. The entire catheter is extruded as a single tube (Medi-Tech, Inc., Watertown, MA).
Each of the three recording lumens has a lateral recording orifice 0.5 mm in diameter. The distal orifice
is located 5 mm from the end of the catheter. The three orifices are 2 mm apart and, therefore, span a
length of 4 mm on the catheter. Six black rings are etched on the catheter at 2-mm intervals to permit
endoscopic determination of the depth of catheter insertion.
A new manometry catheter has also been designed that has an added aspiration channel.16(7107)
Continuous aspiration prevents a rise in intraductal pressure and has been reported to decrease the
incidence of procedure-induced pancreatitis.
To obtain recordings from the biliary tree, the catheter is passed through the accessory channel of the
duodenoscope. Continuous recording of intraductal pressure is obtained by a single-lumen catheter
taped to the endoscope so that its recording lateral orifice is 4 mm proximal to the accessory channel
opening (Figure 683). During pressure recording, each catheter lumen is infused with bubble-free
water at the rate of 0.25 ml/min with a minimally compliant hydraulic capillary infusion system driven by
a constant reservoir pressure of 375 to 400 mm Hg.
(7108)Figure 683. Overview of endoscopic retrograde cholangiopancreatography (ERCP)

manometry. A, Recording system shows internal/external catheters and noncompliant pump
system. B, ERCP manometric technique shows a catheter oriented in the common bile duct
(CBD) with three recording tips within the sphincter of Oddi (SO) zone. PDpancreatic duct.
C, SO pressure profile obtained during catheter pull-through from the CBD to the duodenum.
Appropriate reference points are indicated. Pphasic wave.
The performance of the recording system at the reservoir pressure and infusion rate is tested as
follows. The inherent postocclusion rise rate of the infused catheter system is determined by abruptly
obstructing the recording orifice of each recording lumen. At these conditions, the system will yield a
postocclusion pressure rise rate of 400 mm Hg/sec or greater as measured during the initial 200-mm
Hg pressure rise.
Preferably, patients who undergo ERCP manometry receive either no sedation or only diazepam in
small amounts (5 to 15 mg intravenously). Diazepam and water-soluble contrast media do not appear
to alter the motor activity of the SO. It was demonstrated in one study that meperidine increases phasic
SO contractions but that it had no effect on basal SO pressure during ERCP manometry, and the use
of this drug was therefore recommended for additional analgesia during SO pressure
recording.17(7109) In another report, however, this agent was shown to reduce phasic SO
activity.18(7110) A third study found that meperidine when used for conscious sedation and analgesia
did not affect basal SO pressure.19(7111)
After diagnostic ERCP, the manometry catheter is inserted through the endoscope into the papillary
orifice and directed into the common bile duct or the pancreatic duct (at the completion of the
manometry procedure, contrast material is instilled through the manometry catheter to verify its
location). The appropriate ductal pressure is recorded initially. The catheter is then slowly withdrawn by
2-mm increments into the duodenum using the black marks on the catheter to determine the length of
the sphincter. At each station, pressure recordings are obtained for 2 to 3 min. When the catheter is
positioned in the SO segment so that all three recording orifices record phasic SO contractions or
basal SO pressure, a 3- to 5-min recording is obtained. A polygraph paper speed at 1 cm/sec is best
for demonstrating the onset of phasic wave contractions and their sequential relationship. Zero or
baseline pressure is atmosphere.
Physiology of the Sphincter of Oddi

The biliary tract is a low-pressure, low-flow system. Bile flow into the duodenum is determined mainly
by the rate of bile production, contraction of the gallbladder, and motor activity of the SO. Other factors
such as cystic duct resistance, bile viscosity, duodenal contractions, and gravity are of lesser
importance. There is ample evidence that SO function can be independent of duodenal contractile and
electric activity.20(7112) Furthermore, certain pharmacologic agents produce opposite effects in the
sphincter and duodenum. For example, CCK in humans generally causes duodenal contractions and
SO relaxation.
The physiologic function of the SO is threefold: (1) regulation of flow of bile and pancreatic juice into
the duodenum, (2) diversion of hepatic bile into the gallbladder, and (3) prevention of reflux of
duodenal contents into the pancreatobiliary system. The function of the SO has been regarded solely
as a resistor to bile flow; it has an active and passive component basal SO pressure that gradually
changes during intervals of minutes to hours. These changes in tone regulate bile flow during fasting
and postprandial periods.
Sphincter of Oddi Contractile Activity

Cineradiographic studies provide rapid imaging of the SO segment and demonstrate spontaneous,
rhythmic contractions that have been compared with intervals of "systolic contraction" separated by
intervals of "diastolic relaxation" (Figure 684). Subsequently, ERCP manometric pressure recording of
sphincter motor activity using the noncompliant infusion system has confirmed these radiographic
observations and recorded phasic SO pressure within the profile of the SO zone. These phasic SO
contractions are believed to be more important in "housekeeping" than in metering fluid.
(7113)Figure 684. Radiographic sequence of sphincter of Oddi contractile activity (over a

15-sec interval) demonstrates opening (emptying) and closing (filling) of sphincter segment
after retrograde contrast material injection.
In patients considered to be normal, the SO pressure profile obtained during pull-through across the
sphincter has the following features (see Figure 683). The SO segment measures 4 to 6 mm in
length. Within this zone, there is a modest basal pressure with superimposed phasic contraction
waves. The SO basal pressure is intermittently isobaric with the common bile or pancreatic duct
pressure. The amplitude of the phasic contractions is calculated by subtracting the basal sphincteric
pressure from the peak pressure. Duodenal pressure averages 6 to 10 mm Hg above atmospheric
pressure, and common bile duct pressure averages 10 mm Hg above duodenal pressure. Basal
sphincter pressure is 8 to 10 mm Hg higher than common bile duct pressure, approximately 20 to 25
mm Hg higher than duodenal pressure. The characteristics of SO manometric pressures have been
extensively detailed in a group of 50 healthy volunteers.21(7114) A basal SO pressure of 50 mm Hg or
greater is considered abnormal.
The amplitude of the phasic SO waves averaged from the three recording tips is 128 mm Hg above
SO basal pressure. Wave amplitudes do not differ statistically between multiple recording sites within
the SO segment (Figure 685). The frequency of the phasic contraction waves averages 4/min, and
the wave duration is about 6 sec. The temporal relationship of the phasic contraction waves within the
SO zone has been analyzed.22(7115) In individuals with normal biliary and pancreatic ducts, 60% of
SO phasic contractions were oriented in an antegrade direction, that is, toward the duodenum. About
14% of wave contractions were recorded in a retrograde direction and 26% occurred simultaneously.
Comparison of phasic contraction sequence propagation between these individuals and patients with
stones in the common bile duct is shown in Table 681. The antegrade orientation of most SO
contractions would suggest a propulsive role, but another study demonstrated that the majority of
phasic SO waves are simultaneous in propagation sequence.21(7116)
TABLE 681
Sphincter of Oddi: Phasic Contraction Sequence Propagation

PROPAGATION (%)*
PATIENT GROUP
AGE (yr) *
PATIENTS (n)
Controls
42 3
20
CBD stone
67 4
15
Antegrade
60 4
18 5
Simultaneous
26 3
29 6
CBDcommon bile duct.

* x SEM.
p < .001.
p < .005.
Retrograde
14 4
53 9
(7117)Figure 685. Sphincter of Oddi (SO) basal (B) and peak phasic (P) wave amplitudes
obtained from multiple points within the SO zone show no significant differences between
recording sites whether the catheter is directed into the common bile duct (CBD) or the
pancreatic duct (PD). Pressures are listed on the vertical axis.
Sphincter of Oddi Electric Activity

Myoelectric recording from the human SO has been obtained by positioning electrodes during
laparotomy. Discrete electric spike bursts were attributed to sphincter activity and clearly were not
duodenal in origin (Figure 686). Fluid flow through an indwelling T-tube was interrupted concurrently
with SO spike bursts. SO electric activity has been recorded in a few subjects by positioning catheter
ring electrodes within the sphincter during ERCP.23(7118) However, detailed information obtained
from electric recordings from the human SO is limited by the number of clinical variables. However,
comprehensive evaluation of the myoelectric activity of the SO has been obtained in an opossum
animal model.
(7119)Figure 686. Myoelectric recording from human sphincter of Oddi (SO). Top,
Electrodes are positioned in the SO (Oddi's muscle) and duodenum at the time of laparotomy.
Below, Phasic distal choledochal pressure was recorded concurrent with discrete electric spike
bursts form the SO zone. The latter clearly are not duodenal in origin. (From Ono K,
Watanabe N, Suzuki K. Bile flow mechanisms in man. Arch Surg 1968; 96:86974. Copyright
1968, American Medical Association.)
Miniaturized bipolar electrodes implanted within the SO segment of the opossum show omnipresent
rhythmic spike bursts occurring at the rate of 1 to 8/min in the conscious, fasted animal. Concurrent
peristaltic contractions are recorded manometrically in the SO zone as a propagating monophasic
pressure wave. The spike bursts correspond to a peristaltic contraction that originates proximally in the
sphincter and generally propagates to the sphincter-duodenal junction.24(7120)
Electrodes implanted within both the musculature of the opossum SO and upper gastrointestinal tract
provide a method for studying the relationship of biliary tract and gastrointestinal function in awake
animals over the course of time. During Phase II of the duodenal migrating myoelectric complex
(MMC) activity, the rate of SO spike bursts increases gradually and culminates in a 5-min interval of
maximum activity concurrent with the passage of the Phase III MMC activity through the duodenum.
After feeding, the rate of SO spike bursts increases during a 45-min interval to reach a plateau of about
6 spike bursts/min that last a minimum of 3 hr. Thus, contractile activity in the opossum SO is
coordinated with patterns of gastrointestinal contractility occurring during the fasted and the fed states.
A similar coordination has been demonstrated in long-term recordings from the human SO.25(7121)
Sphincter of Oddi Control Mechanism

The control mechanisms that regulate contractile activity of the SO are incompletely defined. The
prevalent concept is that the main controlling mechanism is hormonal, for example, the reciprocal
relationship between gallbladder contraction and SO relaxation appears to be controlled by CCK
released in response to a meal.26(7122) Neural mechanisms, however, may provide an important
auxiliary function for controlling SO contractile activity. Nonadrenergic inhibitory nerves, for example,
are present in several mammalian animal species and may also be present in humans. Finally,
vascular engorgement in the region of the sphincter can cause significant alteration in the spongelike
network of the mucosal folds lining the segment, thereby altering luminal resistance and influencing
flow across the sphincter.27(7123)
Physiologic Role of the Sphincter of Oddi

The physiologic role of the SO is that of controlling delivery of bile or pancreatic juice into the
duodenum. The SO exhibits contractile activity as well as variable tone. Antegrade phasic contractions
may serve to actively transport sphincter contents into the duodenum, but the phasic contractions
probably have a major role in preventing reflux of duodenal contents into the common bile duct and
may play an important role in keeping the distal choledochus free of sludge and particulate matter.
The SO basal pressure appears to provide a variable tonicity to the sphincteric mechanism and is the
most important feature of the SO function. Pharmacologic studies in the human show that sphincteric
basal pressure and phasic contractions differ in their response to intravenous doses of
morphine.28(7124) Low-dose morphine increases the rate of phasic contractions, whereas higher
doses of morphine increase basal pressure as well. These findings suggest a separation of function for
these two components.29(7125) In order for bile to flow into the duodenum, the basal pressure must
be reduced to common bile duct pressure. The mechanisms regulating this feature of the SO segment
have yet to be defined.
Sphincter of Oddi Pharmacology

Enteric Hormones
A number of enteric hormones have been administered to humans during direct manometric
measurement of SO motor activity. Cholecystokinin octapeptide (CCK-OP) promptly inhibits both tonic
and phasic contractions in humans (Figure 687) at an intravenous bolus dose of 20 to 40
ng/kg.30(7126) CCK receptors have been identified on human SO tissue that are sulfate-dependent
and mediate sphincter relaxation.8(7127) Studies in the cat indicate CCK-OP releases a nonadrenergic
neural transmitter that inhibits sphincter motor activity. This inhibitory action masks a direct excitatory
effect of CCK in the feline sphincter muscle that is observed after pharmacologic denervation of this
segment.31(7128) This phenomenon is of particular interest because a possible counterpart in
humans is the paradoxical contraction of the human SO after CCK-OP administration. Such
paradoxical responses have been observed in eight patients with suspected SO dyskinesia,32(7129)
and in four post-liver transplant patients with elevated basal SO pressure (Figure 687).33(7130)
Continuous intravenous infusion of CCK-OP at a rate sufficient to reproduce physiologic blood levels
has also been shown to inhibit human SO motor activity.34(7131) Pentagastrin contracts the SO in
humans after intravenous bolus administration, but this effect is impossible to quantitate because of
concurrent small intestinal contractions.
(7132)Figure 687. Effect of cholecystokinin octapeptide (CCK-OP) on human SO motor

activity. A, Dramatic reduction in basal and phasic wave pressure was recorded from three
sites within the SO segment (SO1, SO2, SO3), immediately after an intravenous bolus of
CCK-OP (20 ng/kg). This is considered a normal response. B, Sustained contraction response
was recorded from three sites within the SO segment (SO1, SO2, SO3) after intravenous
administration of CCK-OP. This is considered a paradoxical response and suggests possible
SO denervation.
Secretin relaxes the human SO presumably by a direct effect on the muscle. After intravenous bolus
administration (GIH secretin, 1 U/kg; Gastrointestinal Hormone Research, Karolinska Institute,
Stockholm, Sweden), there is a transient increase in phasic wave contractions of the sphincter followed
rapidly by inhibition of all motor activity for several minutes. In one report, secretin administered
intravenously to patients with chronic pancreatitis caused elevation of the main pancreatic duct
pressure and produced a pattern of SO dyskinesia when compared with a control group.35(7133) In
earlier studies, glucagon was shown to inhibit SO motor activity but a more recent study using
glucagon 129 demonstrated no effect on SO motility.36(7134) Octreotide, a long-acting somatostatin,
increases both basal SO pressure and frequency of phasic SO waves when administered
intravenously.37(7135)
Drugs
In humans, intravenous administration of small doses of morphine increases the rate of phasic
contractions of the SO but has little effect on basal pressure. One study, using intravenous morphine
(10 g/kg) demonstrated an increase in phasic SO waves from a mean of 2.4 to 7.9 per min.18(7136)
Larger doses of morphine cause "spasm" of the SO segment. Naloxone abolishes the effect of
morphine on the human SO. Meperidine, in clinical analgesia-producing doses (100 g/kg), appears to
cause an increase in sphincter phasic activity.18(7137) Atropine decreases motor activity and also
partially antagonizes the effect of morphine and meperidine on the human SO. Amyl nitrite relaxes the
SO in most mammalian species. Administered as an inhalant (four "sniffs" from a ruptured pulvule), the
rapid inhibitory action of amyl nitrite has been used clinically to aid in distinguishing SO "spasm" from
"stenosis" (Figure 688). Calcium channel blockers (intravenous nicardipine) effectively decrease SO
pressure,38(7138) whereas oral nifedipine has been demonstrated to relieve postcholecystectomy pain
in patients with elevated basal SO pressure.39(7139) The topical application of glycerol trinitrate, a
nitrous oxide donor, inhibits SO motor activity,40(7140) whereas the instillation of alcohol increases
basal SO pressure.41(7141) The effects, if any, of midazolam on the motor activity remain
controversial to date.42,43(7142)
(7143)Figure 688. SO basal and phasic wave pressure amplitudes recorded from three sites
within the SO segment are suddenly and profoundly reduced after inhalation of amyl nitrite by
the patient.
Sphincter of Oddi Dysfunction

The term sphincter of Oddi dysfunction is applied to a collection of symptoms and signs associated
with either actual or presumed abnormalities of the SO. In different areas of the world, this syndrome is
diagnosed in 1 to 40% of patients complaining of postcholecystectomy pain.
SO dysfunction may be caused by structural narrowing of the sphincteric segment as a result of direct
trauma during passage of a common bile duct stone, injury at the time of an operative procedure, or
inflammation secondary to pancreatitis. Acquired immunodeficiency syndrome (AIDS)-associated
papillary stenosis has been described in a report of 25 patients with AIDS and right upper quadrant
pain and markedly elevated alkaline phosphatase whose symptoms were relieved or reduced by
ES.44(7144) These observations remain to be confirmed in further studies. SO dysfunction may also
result from a functional motor abnormality with associated spasm that produces intermittent bile or
pancreatic duct obstruction.
Clinical Features
The clinical features of SO dysfunction are as follows:
1. Patients experience recurring bouts of epigastric or right upper quadrant pain that often radiates to
the back and may be precipitated by meals.
2. Some patients may have intermittent or transient abnormalities in liver enzymes, particularly
serum alkaline phosphatase and transaminases, or intermittent elevation of serum amylase

levels.
3. The common bile duct or pancreatic duct may be dilated as demonstrated by contrast material
injection, for example, at ERCP. A common bile duct diameter greater than 12 mm or a pancreatic
duct diameter greater than 7 mm is considered abnormal.
4. Delayed emptying of contrast material from the common bile duct (>45 min) or pancreatic duct
(>10 min) with the patient in the supine position may occur. If the emptying time is to be
evaluated, patients should not be given narcotics or anticholinergics before the study because
these drugs impair SO motor function.
Diagnosis
There is no definitive standard by which the diagnosis of SO dysfunction can be established with
certainty. The current methods used to diagnose and treat patients with SO dysfunction are for the
most part without consensus approval, and most reports base their results on uncontrolled studies or
incomplete follow-up periods. Nevertheless, it is logical to assume that this complex sphincteric
mechanism may be the object of motor dysfunction or structural narrowing caused by certain
disorders. In the past, several relatively crude methods have been used to confirm the diagnosis of SO
dysfunction at operation. The decision for or against surgical sphincterotomy has been based on these
findings: (1) Resistance to flow of a column of fluid introduced through a T-tube into the distal common
bile duct has been considered a reflection of SO resistance. A presumptive diagnosis of sphincter
stenosis is suggested if measured resistance is in excess of 30 cm of H2O and calculi are not present
in the bile duct. (2) The inability of the surgeon to pass a No. 3 Bakes (3 mm) dilator via the common
bile duct through the SO and into the duodenum or the inability to pass a lacrimal duct probe (2 mm
diameter) into the pancreatic duct has defined the diagnosis of papillary "stenosis."
The introduction of ERCP provided a nonsurgical method for evaluating possible SO dysfunction.
Subsequently, the development of ES offered a nonoperative treatment of this disorder. The methods
for assessing possible SO dysfunction at the time of ERCP study have included the following:
1. Evaluation of the "ease" of cannulation of the papilla or, more precisely, the endoscopist's
estimate of "tightness" encountered during insertion of the catheter.
2. Re-creation of a patient's typical abdominal pain after instillation of a minimum amount of contrast
medium into the bile duct during ERCP.
3. Evaluation of ductal emptying rate after instillation of contrast medium at ERCP.
4. Direct measurement of the pressure and motor patterns of the SO zone by ERCP manometry
techniques.
In our experience, pain during contrast instillation45(7145) and the endoscopist's ability to define a tight
sphincter segment or difficulty in cannulating the ducts do not correlate with the presence or absence
of recognizable sphincteric disease.46(7146) Also, radiologic evaluation of a narrowed distal segment
and the opening and closing of the sphincteric segment has not yielded definitive information
correlating with SO function or dysfunction.
Classification
With the use of a high-fidelity perfusion system, we have attempted to correlate SO pressures obtained
at ERCP manometry with a clinical syndrome of SO dysfunction. It is not possible, however, to
separate all those patients with a functional motility disorder of the sphincter (dyskinesia) from those
with an anatomic stricture (papillary stenosis). Nonetheless, to enhance identification and future study,
we have subdivided patients into biliary and pancreatic categories of SO dysfunction. These
classifications, although imperfect, have been useful in better defining SO dysfunction.
Biliary Group
The biliary type of SO dysfunction can be further subdivided into three groups.
Biliary Group IProbable SO dysfunction. Patients in this group must meet all the following criteria: (1)
Pain that is typical for a biliary tract disorder. (2) Elevated serum alkaline phosphatase or
transaminase, or both (twice-normal values), documented on at least two occasions. (3) Dilated
common bile duct with a diameter greater than 12 mm on retrograde cholangiography. (4) Delayed
drainage of contrast material from the common bile duct (>45 min after ERCP injection with the patient
in the supine position).
Group IIPresumptive SO dysfunction. Patients in this category must meet the following criteria: (1)
Pain that is typical for a biliary tract disorder. (2) One or two of criteria 2 through 4 listed for Group I.
Group IIIPossible SO dysfunction. Patients presenting with only biliary type pain and none of the
other criteria for Group I.
Pancreatic Group
Approximately 20% of patients in our practice who present with idiopathic recurrent pancreatitis have
SO dysfunction. Patients in this group must meet the following criteria: (1) Two or more documented
episodes of unexplained clinical pancreatitis and (2) no structural abnormalities of the biliary tree or
pancreas. A dilated pancreatic duct or delayed pancreatic drainage is obviously helpful in focusing on a
potential obstruction to flow of pancreatic juice.
Study Results
Biliary Group I SO dysfunction is the most easily recognized clinical entity. The diagnosis of SO
dysfunction in Group I patients is readily appreciated by the clinician. These patients are often older in
age than Group II and Group III patients. SO manometry reveals a basal pressure greater than 40 mm
Hg in about 65 to 70% of Group I patients.47(7147) Interestingly, not all Group I patients have SO
abnormalities. However, manometry is not essential to establish the diagnosis of SO
dysfunction.47(7148) Group I patients respond to ES with a successful therapeutic outcome.47(7149)
On the other hand, Group II biliary patients with suspected SO dysfunction are somewhat more difficult
to diagnose. Although these patients also have typical biliary pain, laboratory and bile duct
abnormalities are less frequent or transient. SO manometry is essential for the diagnosis of SO
dysfunction in this group. In a randomized study of Group II patients, the efficacy of ES in patients with
abnormal basal SO pressures was evaluated.48(7150) Forty-seven patients in this study were
randomly assigned to either a sphincterotomy group (23) or a nonsphincterotomy sham group (24). ES
resulted in significant improvement in pain scores and liver function tests at 1 year follow-up in 10 of 11
patients with elevated SO pressure. In contrast, there was improvement in only 3 of 12 patients with
elevated SO pressure who underwent sham sphincterotomy. Patients with normal basal SO pressure
and abdominal pain remain unchanged regardless of the type of treatment, sham or sphincterotomy. At
1 year, ES was performed in 12 symptomatic patients who had undergone a sham procedure. Seven
of these patients had elevated basal SO pressure and 5 had normal basal SO pressure. Forty patients
were followed up for 4 years. Of the 19 patients with elevated basal SO pressures who had ES, 17
remained symptom-free. In Group II biliary patients, an abnormal basal SO pressure is a positive
predictor to a successful clinical outcome after ES.
Biliary Group III patients with suspected SO dysfunction remain the most difficult and perplexing clinical
problem. Although ERCP manometry is the most useful diagnostic modality to establish basal SO
pressure,49(7151) there is no uniform agreement as to its significance.50(7152) Many patients
suspected to have SO dysfunction may be suffering from irritable bowel syndrome or a somatization
disorder.51(7153) These patients seldom respond to medical therapy. Endoscopic therapy in Group III
patients is currently recommended within study protocols or where SO manometry is routinely
performed.
Sphincter of Oddi DysfunctionPancreatic Type

The SO encircles the distal common bile duct and the pancreatic duct as they course through the
duodenal wall. It has been reported that approximately 20 to 25% of patients with idiopathic recurrent
pancreatitis may have SO dysfunction. Whereas the motor abnormality may be limited to the
pancreatic sphincter in the majority of these patients, abnormal pressure characteristics may also be
demonstrated for the sphincter choledochus.52(7154) Therefore it is important to perform manometric
studies from both the common bile duct segment of the SO and the pancreatic segment of the SO in
patients with recurrent pancreatitis. This is a matter of great practical importance, as ES can be
routinely undertaken if the pressure elevation is limited to the sphincter choledochus. However, if the
basal SO pressure of the pancreatic sphincter is elevated, ES of this segment must be undertaken with
a great deal of caution and requires an experienced operator. In a report of 24 patients with acute
recurrent pancreatitis, 9 patients were noted to have elevated SO basal pressure greater than 40 mm
Hg.53(7155) Among these patients, 2 had elevated basal pressure involving the common bile duct and
7 patients had elevated basal SO pressure confined to the pancreatic segment of the SO. Many
patients who have idiopathic recurrent pancreatitis and elevated basal SO pressure respond
successfully to ES or surgical sphincteroplasty.54(7156)
Complications of Endoscopic Sphincterotomy

The complication rate associated with ES for SO dysfunction has been reported to be higher than that
for treatment of common bile duct stones.55(7157) A 9% complication rate and a 2% mortality rate for
patients with papillary stenosis after ES have been reported from Germany.56(7158) In our series,
complications developed in 5 of 80 patients (6.3%) who underwent ES for SO dysfunction. There were
no deaths. This is similar to our complication rate for patients who have undergone ES for common bile
duct stones.
Speculation on the Pathophysiology of Sphincter of Oddi Dyskinesia

The following list of abnormalities in SO motor dysfunction derived from our clinical observations could
represent examples of SO dysfunction.
Hypertonicity or Spasm
Elevation of basal and phasic wave contraction pressures has been recorded at ERCP manometry in
patients with suspected SO motor dysfunction.57(7159) This does not of itself indicate a smooth
muscle abnormality, although a transient decrease in SO pressure has been demonstrated in several
of these patients after intravenous administration of glucagon or CCK-OP after inhalation of amyl
nitrite.
Denervation
A paradoxical elevation of SO pressure in response to intravenous administration of CCK-OP has been
demonstrated in a subset of patients with biliary-type pain (see Figure 687). Eight patients with this
type of pain demonstrated a paradoxical rise in SO basal pressure after the intravenous bolus
administration of CCK-OP during ERCP manometry. Whether this atypical response indicates a
neuromuscular abnormality of the sphincter is not known. A similar paradoxical response to
intravenous CCK has been shown in the denervated feline SO and lower esophageal sphincter. The
lower esophageal sphincter in the great majority of patients with achalasia demonstrates paradoxical
contraction after administration of CCK-OP.58,59(7160) This type of response suggests a defect in
inhibitory innervation of the sphincter that unmasks the direct stimulatory effect of CCK on the smooth
muscle of the SO.
Phasic Wave Sequence Abnormality
The majority of phasic wave sequences in the SO zone appear to be simultaneous.21(7161) The
phasic wave sequence is altered in patients with common bile duct stones (Figure 689). There is a
significant increase in the frequency of retrograde SO phasic waves. Conceivably, alteration in the
direction of phasic wave propagation could impede bile flow and subsequently produce pain.
(7162)Figure 689. Phasic wave sequences recorded from three sites (4-mm segment) within
the sphincter of Oddi (SO) zone in a patient with a retained common bile duct stone. The
majority of sequences shown here are retrograde (R) in direction. Ssimultaneous.
"Tachyoddia"
The frequency of SO phasic contractions appears to regulate the rate of common bile duct emptying in
the opossum. Increasing this contraction rate appears to shorten the filling phase of the ampullary
region. Theoretically, high-frequency SO contraction waves could obstruct emptying of the common
bile duct. Phasic contraction waves have been recorded at a frequency up to 12/min in the human
during ERCP manometry (Figure 6810). Low-dose morphine causes a profound increase in the rate
of phasic contraction waves. Theoretically, a prolonged high rate of SO contractions could disrupt the
flow of bile into the duodenum and cause symptoms.
(7163)Figure 6810. Rapid sphincter of Oddi (SO) phasic wave contractions of 12 per minute
(tachyoddia) were recorded during catheter pull-through from the pancreatic duct into the
duodenum. The SO basal pressure is elevated in the proximal segment of the SO zone.
Clinical Relevance of Sphincter of Oddi Manometry

For patients with overt symptoms and signs suggestive of SO dysfunction, ERCP manometry pressure
study may not be necessary for the diagnosis. For example, ERCP manometry may be unnecessary
for patients with all the criteria for inclusion in Group I; however, the effect of hormones or drugs on the
SO basal pressure may be useful in determining whether sphincter dysfunction is the result of organic
stricture or a primary motor disorder. Amyl nitrite and CCK lower the basal sphincteric pressure in
patients with motor dysfunction or spasm of the sphincter segment, but these do not affect the elevated
SO basal pressure caused by a structural alteration.
SO manometry is also useful in evaluating the effectiveness of ES. If the common bile duct to
duodenal pressure gradient does not approximate zero or if phasic waves are not completely
abolished, the sphincterotomy may not be adequate to alter sphincter dynamics and affect symptoms.
In biliary Groups II and III patients and in patients with idiopathic recurrent pancreatitis, SO motility
studies may be the only direct method for selecting those patients who may benefit from ES or surgical
sphincterotomy.60(7164)
At the present time, without manometric studies, it is difficult to determine which patient with suspected
SO dysfunction should undergo sphincterotomy. However, it is conceivable that sphincter dysfunction
may be intermittent and SO function may be normal during a manometric study period. SO manometry
remains a difficult technique to acquire and is not available in most centers. In this context, a number
of noninvasive tests have been used to help diagnose SO dysfunction.
Secretin Ultrasonography
Secretin, an enteric peptide, is a potent secreatagogue that causes exocrine pancreatic secretion.
Intravenous administration of secretin also has been shown to cause increased basal tone in the
human SO. The voluminous output of pancreatic secretion coupled with increased basal tone in the
sphincter segment results in distention of the pancreatic duct in normal subjects. In patients with SO
dysfunction, this effect is exaggerated. An increase in the diameter of the main pancreatic duct of 1 to
2 mm or greater demonstrated by ultrasonography has been reported to be diagnostic for papillary
stenosis.61(7165) Although the principle behind this noninvasive test is quite attractive, verification of
these results based on larger clinical studies has not been reported. The secretin-ultrasound test is
discussed in Chapter 74: Adjunct Diagnosis of Pancreatic Disease and Pancreatic Physiology.
Fatty Meal Ultrasonography
Fatty meal administration can lead to endogenous release of CCK, which enhances bile flow. In SO
dyskinesia, the impedance of biliary flow at the sphincteric level can lead to dilation of the bile duct.
The fatty meal ultrasound test consists of measuring the diameter of the bile duct at baseline followed
by oral administration of corn oil (Lipomul) 1.5 ml/kg. The diameter of the bile duct is measured by
ultrasonography every 15 min for the next 60 min. After administration of corn oil, an increase in the
diameter of the bile duct of 2 mm or greater from the baseline (at 45 min) is considered to be
diagnostic for SO dysfunction.62(7166) Despite positive reports, the fatty meal ultrasound test has not
been used routinely in clinical practice.
Quantitative Hepatobiliary Scintigraphy
Quantitative hepatobiliary scintigraphy is another noninvasive test found to be useful in identifying
patients with SO dysfunction. Hepatobiliary scintigraphy is performed after a minimum fast of 4 hr.
Following intravenous administration of 5 mCi of technetium-99m diisopropyl phenol carbamyl methyl
aminoacetic acid, counts over the abdomen are recorded for 90 min. Isotope distribution curves are
generated over the liver, hepatic hilum, and common bile duct. In one report, the most sensitive
indicator for a positive test was 45-min isotope clearance less than 63%.63(7167) In a more recent
study, the hepatic hilum-duodenal transit time showed a direct correlation with maximum manometric
basal SO pressure but not with phasic SO activity.64(7168)
Management of Sphincter of Oddi Dysfunction

Medical Therapy
Pharmacologic agents known to cause relaxation of the SO are often the first line of treatment for
possible SO dysfunction. Anticholinergic agents such as dicyclomine, nitrates, and calcium channel
blockers (e.g., nifedipine) are among the commonly employed drugs. In addition, progesterone
derivatives and bile acids also have been reported to be useful in patients with SO dysfunction.
However, pharmacologic therapy is generally disappointing, especially in patients with suspected SO
dyskinesia.
Surgical Therapy
Operative therapy is directed toward severing the sphincter muscle by a number of techniques, for
example, sphincteroplasty, sphincterotomy, or septectomy. Surgical therapy is successful in the
majority of patients and provides long-term relief.
Endoscopic Therapy
ES, similar to surgical therapy, is a viable alternative for patients with SO dysfunction.65(7169) It is
associated with a successful outcome in 75 to 80% of patients. Although the complication rate is
similar to that of operative intervention, endoscopic therapy is more cost-effective. However, ES for
suspected SO dysfunction in patients with only pain may be associated with an increased complication
rate compared with patients undergoing ES for common bile duct stone.
Summary
The human SO is a uniquely adaptive structure that possesses ejecting as well as occluding
mechanisms. Its primary physiologic role is regulation of flow pressure within the biliary and pancreatic
system. Irrespective of the flow rate, a narrow range of low pressure is maintained within the bile duct
or pancreatic duct that allows hepatic or pancreatic secretion to proceed against negligible hydrostatic
pressure. Structural abnormalities of the SO can occur as a result of passage of biliary calculi or injury
at the time of operation. These abnormalities can be more easily identified and can be managed by
appropriate relief of the obstruction.

The true incidence of primary SO dysfunction is unknown. It is not a common disorder. Possible
causes of SO dysfunction include muscular SO spasm, denervation, or an abnormality of phasic
contraction wave sequencing or rate. Patients with suspected SO dysfunction and elevated SO basal
pressure may be aided by sphincterotomy. Continued refinement of current techniques and new
methodologies evaluated in prospective studies with long-term outcomes may eventually define the
clinical entity of SO dysfunction with certainty.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
1957;104:64152.
Teilum D. In vitro measurement of the length of the sphincter of Oddi. Endoscopy
1991;23:1146.
Boyden EA. The sphincter of Oddi in man and certain representative mammals. Surgery
1937;1:2537.
Geenen JE, Hogan WJ, Dodds WJ, et al. Intraluminal pressure recording from the human
sphincter of Oddi. Gastroenterology 1980;78:31724.
Toouli J, Baker RA. Innervation of the sphincter of Oddi: Physiology and consideration of
pharmacological interventions in biliary dyskinesia. Pharmacol Ther 1991;49:26981.
Burnett W, Cairns FW, Bacsich P. Some observations on the innervation of the extrahepatic
biliary system in man. Ann Surg 1964;159:826.
Slivka A, Chuttani R, Carr-Locke DL, et al. Inhibition of sphincter of Oddi function by the nitric
oxide carrier S-nitroso-N-acetylcysteine in rabbits and humans. J Clin Invest 1994;94:17928.
Tokunaga Y, Cox KL, Itasaka H, et al. Characterization of cholecystokinin receptors on the
human sphincter of Oddi. Surgery 1993;114:94250.
Stolte M. Some aspects of the anatomy and pathology of the papilla of Vater. In Classen M,
Geenen J, Kawai K (eds). The Papilla Vateri and Its Diseases. Baden-Baden, Koln, New York:
Witzstrock, 1979;313.
Martin RW, Gilbert DA, Silverstein FE, et al. An endoscopic Doppler probe for assessing
intestinal vasculature. Ultrasound Med Biol 1985;11:619.
Oddi R. D'une disposition a sphincter speciale de l'ouverture du canal cholidoque. Arch Ital
Biol 1887;8:31722.
Ivy AC, Oldberg E. A hormone mechanism for gallbladder contraction and evacuation. Am J
Physiol 1928;86:599613.
Cuschieri A, Hughes JH, Cohen M. Biliary pressure studies during cholecystectomy. Br J Surg
1972;59:26773.
Vondrasek P, Eberhardt G, Classen M. Endoscopic semiconductor manometry. Int J Med
1974;3:18892.
Geenen JE, Hogan WJ, Shaffer RD, et al. Endoscopic electrosurgical papillotomy and
manometry in biliary tract disease. JAMA 1977;237:20758.
Sherman S, Troiano FP, Hawes RH, Lehman GA. Does continuous aspiration from an end
and side port in a sphincter of Oddi manometry catheter alter recorded pressures?.
Elta GH, Barnett JL. Meperidine need not be proscribed during sphincter of Oddi manometry.
Thune A, Baker RA, Saccone GT, et al. Differing effects of pethidine and morphine on human
sphincter of Oddi motility. Br J Surg 1990;77:9925.
Sherman S, Gottlieb K, Uzer M, et al. Effects of meperidine on the pancreatic and biliary
sphincter. Gastrointest Endosc 1996;44:23942.
Ono K, Watanabe N, Suzuki K, et al. Bile flow mechanisms in man. Arch Surg
1968;96:86974.
Guelrud M, Mendoza S, Rossiter G, Villegas MI. Sphincter of Oddi manometry in healthy
volunteers. Dig Dis Sci 1990;35:3846.

22. Toouli J, Geenen JE, Hogan WJ, et al. Sphincter of Oddi motor activity: A comparison
between patients with common bile duct stones and controls. Gastroenterology
1982;82:1117.
23.
Salducci J, Naudi B, Pin G, et al. Papilla electromyography: Endoluminal recording performed
in man by perduodenoscopic cannulation. In Delmont J (ed). The Sphincter of Oddi.
Proceedings of the Third Gastroenterological Symposium, Nice, June 89, 1976. New York:
S. Karger, 1976;7780.
24. Honda R, Toouli J, Dodds WJ, et al. Relationship of sphincter of Oddi spike bursts to
gastrointestinal myoelectric activity in conscious opossums. J Clin Invest 1982;69:7708.
25. Torsoli A, Corazziari E, Habib FI, et al. Frequencies and cyclic pattern of the human sphincter
of Oddi phasic activity. Gut 1986;27:3639.
26. Lin TM. Actions of gastrointestinal hormones and related peptides on the motor function of the
biliary tract. Gastroenterology 1975;69:100622.
27. Tansy MF, Salkin L, Innes DL, et al. The mucosal lining of the intramural common bile duct as
a determinant of ductal opening pressure. Am J Dig Dis 1975;20:61325.
28.
Venu R, Toouli J, Geenen JE, et al. Effect of morphine on motor activity of the human
sphincter of Oddi [Abstract]. Gastroenterology 1983;84:1342.
29. Helm JF, Venu RP, Geenen JE, et al. Effects of morphine on the human sphincter of Oddi.
Gut 1988;29:14027.
30. Toouli J, Hogan WJ, Geenen JE, et al. Action of cholecystokininoctapeptide on sphincter of
Oddi basal pressure and phasic wave activity in humans. Surgery 1982;92:497503.
31. Behar J, Biancani P. Effect of cholecystokinin and the octapeptide of cholecystokinin on the
feline sphincter of Oddi and gallbladder. Mechanisms of action. J Clin Invest 1980;66:12319.
32.
Hogan WJ, Geenen JE, Dodds WJ, et al. Paradoxical motor response to cholecystokinin
(CCK-OP) in patients with suspected sphincter of Oddi dysfunction [Abstract].
33. Douzdjian V, Abecassis MM, Johlin FC. Sphincter of Oddi dysfunction following liver
transplantation. Screening by bedside manometry and definitive manometric evaluation [see
comments]. Dig Dis Sci 1994;39:2536.
34.
Jensen DM, Weiss S, Tapia JI, Beilin DB. Control of motility and pressures of the human
sphincter of Oddi. Proceedings of the 4th Annual Meeting of the International Biliary
Association, Paris, June 1012. 1982.
35. Laugier R. Dynamic endoscopic manometry of the response to secretin in patients with
chronic pancreatitis. Endoscopy 1994;26:2227.
36. Biliotti D, Habib FI, De Masi E, et al. Effect of glucagon on sphincter of Oddi motor activity.
Digestion 1989;43:1859.
37. Binmoeller KF, Dumas R, Harris AG, Delmont JP. Effect of somatostatin analog octreotide on
human sphincter of Oddi. Dig Dis Sci 1992;37:7737.
38. Fullarton GM, Falconer S, Campbell A, Murray WR. Controlled study of the effect of
nicardipine and ceruletide on the sphincter of Oddi. Gut 1992;33:5503.
39. Khuroo MS, Zargar SA, Yattoo GN. Efficacy of nifedipine therapy in patients with sphincter of
Oddi dysfunction: A prospective, double-blind, randomized, placebo-controlled, cross over
trial. Br J Clin Pharmacol 1992;33:47785.
40. Luman W, Pryde A, Heading RC, Palmer KR. Topical glyceryl trinitrate relaxes the sphincter
of Oddi. Gut 1997;40:5413.
41. Guelrud M, Mendoza S, Rossiter G, et al. Effect of local instillation of alcohol on sphincter of
Oddi motor activity: Combined ERCP and manometry study. Gastrointest Endosc
1991;37:42832.
42. Cuer JC, Dapoigny M, Bommelaer G. The effect of midazolam on motility of the sphincter of
Oddi in human subjects. Endoscopy 1993;25:3846.
43. Rolny P, Arleback A. Effect of midazolam on sphincter of Oddi motility. Endoscopy
1993;25:3813.
44. Cello JP, Chan MF. Long-term follow-up of endoscopic retrograde cholangiopancreatography
J Med 1995;99:6003.
45. Schmalz MJ, Geenen JE, Hogan WJ, et al. Pain on common bile duct injection during ERCP:
Does it indicate sphincter of Oddi dysfunction?. Gastrointest Endosc 1990;36:45861.
46. Venu RP, Geenen JE, Hogan W, et al. Idiopathic recurrent pancreatitis. An approach to
diagnosis and treatment. Dig Dis Sci 1989;34:5660.
47. Rolny P, Geenen JE, Hogan WJ. Post-cholecystectomy patients with objective signs of partial
bile outflow obstruction: Clinical characteristics, sphincter of Oddi manometry findings, and
results of therapy. Gastrointest Endosc 1993;39:77881.
48. Geenen JE, Hogan WJ, Dodds WJ, et al. The efficacy of endoscopic sphincterotomy after
cholecystectomy in patients with sphincter-of-Oddi dysfunction. N Engl J Med 1989;320:827.
49. Sherman S, Troiano FP, Hawes RH, et al. Frequency of abnormal sphincter of Oddi
manometry compared with the clinical suspicion of sphincter of Oddi dysfunction. Am J
50. Thune A, Scicchitano J, Roberts-Thomson I, Toouli J. Reproducibility of endoscopic sphincter
of Oddi manometry. Dig Dis Sci 1991;36:14015.
51. Abraham HD, Anderson BA, Lee D. Somatization disorder in sphincter of Oddi dysfunction.
Psychosom Med 1997;59:5537.
52. Silverman WB, Ruffolo TA, Sherman S, et al. Correlation of basal sphincter pressures
measured from the bile duct and the pancreatic duct in patients with suspected sphincter of
Oddi dysfunction. Gastrointest Endosc 1992;38:4403.
53. Raddawi HM, Geenen JE, Hogan WJ, et al. Pressure measurements from biliary and
pancreatic segments of sphincter of Oddi. Comparison between patients with functional
abdominal pain, biliary, or pancreatic disease [see comments]. Dig Dis Sci 1991;36:714.
54. Toouli J, Di Francesco V, Saccone G, et al. Division of the sphincter of Oddi for treatment of
dysfunction associated with recurrent pancreatitis. Br J Surg 1996;83:120510.
56. Siefert E, Gail K, Weismueller J. Langzeitresultate nach endoskopischer Sphinkterotomie.
57. Fullarton GM, Hilditch T, Campbell A, Murray WR. Clinical and scintigraphic assessment of
the role of endoscopic sphincterotomy in the treatment of sphincter of Oddi dysfunction. Gut
1990;31:2315.
58. Behar J, Biancani P. Effect of cholecystokinin-octapeptide on the lower esophageal sphincter.
59. Dodds WJ, Dent J, Hogan WJ, et al. Paradoxical lower esophageal sphincter contraction
induced by cholecystokinin-octapeptide in patients with achalasia. Gastroenterology
1981;80:32733.
60. Nardi GL, Acosta JM. Papillitis as a cause of pancreatitis and abdominal plain. Role of
evocative test, operative pancreatography and histological evaluation. Ann Surg
1966;164:611.
61. Warshaw AL, Simeone J, Schapiro RH, et al. Objective evaluation of ampullary stenosis with
ultrasonography and pancreatic stimulation. Am J Surg 1985;149:6572.
62.
Darweesh R, Dodds WJ, Hogan WJ, et al. Fatty meal sonography for evaluating patients with
suspected partial common bile duct obstruction. AJR 1988;151:63.
63.
Darweesh R, Dodds WJ, Hogan WJ, et al. Efficacy of quantitative hepatobiliary scintigraphy
and fatty meal sonography for detecting partial common bile duct obstruction.
64. Corazziari E, Cicala M, Habib FI, et al. Hepatoduodenal bile transit in cholecystectomized
subjects. Relationship with sphincter of Oddi function and diagnostic value. Dig Dis Sci
1994;39:198593.
65.
Venu RP, Geenen JE. Postcholecystectomy syndrome. In Yamada T, Alpers DH, Owyang C,
et al. (eds). Textbook of Gastroenterology. Philadelphia: JB Lippincott, 1991;203344.
Chapter 69 Tumors of the Main Duodenal Papilla

(7170)
(7171)
W. G. PARSONS, M.D.
P. J. CONNORS, M.D.
DAVID L. CARR-LOCKE, M.B., B.CHIR., D.R.C.O.G., F.R.C.P.

Benign and malignant tumors arise from the main duodenal papilla. In this chapter, the phrase
ampullary tumor is used interchangeably with tumor of the main duodenal papilla. Periampullary tumor
is a term that encompasses not only tumors of the main duodenal papilla but also tumors of the head
of the pancreas, distal common bile duct, and duodenum. Malignant ampullary tumors are of particular
interest because they represent a potentially curable lesion if the correct diagnosis is made at an early
stage. Preoperative diagnosis of these tumors, previously a rare occurrence, has become more
common with the widespread use of endoscopy of the descending duodenum, application of
side-viewing duodenoscopes, and the dramatic growth of endoscopic retrograde
cholangiopancreatography (ERCP) and related techniques.
Carcinoma
The age range of patients at the time of diagnosis of carcinoma of the ampulla is 20 to 91 years with a
peak in the seventh decade of life.18(7172) A review of eight series reveals that of 424 cases, 234
(55%) occurred in men.18(7173)
Jaundice is the most common presenting symptom, occurring in 84% of patients.114(7174) The triad
of fluctuating painless jaundice, anemia with or without symptoms of gastrointestinal bleeding, and a
palpably enlarged gallbladder is considered relatively specific for an ampullary or periampullary
tumor.15(7175) Unfortunately, this classic triad is seen infrequently, being found in only 2 of 35 patients
in one report (5.7%).7(7176) Fluctuating jaundice was noted in 5 of 32 patients (16%) by Makipour et
al.,16(7177) in 2 of 18 patients (11.1%) of Russell et al.,3(7178) and in only 9 of 109 patients (8.2%) in
the series of Yamaguchi and Enjoji.5(7179) Those patients with ampullary adenocarcinoma who
present without jaundice have a more favorable outcome than those who present with jaundice. A
careful retrospective analysis by Yamaguchi et al.17(7180) of 31 patients with nonicteric ampullary
carcinoma and 111 patients with icteric ampullary carcinoma revealed cumulative 10-year survival
rates of 57% and 23%, respectively.17,18(7181) Multivariate regression analysis disclosed that
jaundice had no independent prognostic significance but that a greater number of early ampullary
carcinomas and papillary-type adenocarcinomas were present in the nonicteric group, thereby partially
explaining their observations.17(7182)
Other symptoms of ampullary adenocarcinoma include weight loss (83% of patients), abdominal pain
(53%), anorexia (46%), pruritus (43%), vomiting (43%), nausea (41%), fever and chills (27%), diarrhea
(18%), constipation (15%), dyspepsia (13%), and rectal bleeding (9%).114(7183) When questioned
carefully, many patients will admit to having constitutional symptoms for weeks or months before the
onset of jaundice. Walsh et al.2(7184) found that malaise and dyspepsia began, on average, 150 and
365 days, respectively, before the diagnosis of ampullary carcinoma. They also noted a median delay
of 30 days from the onset of jaundice to presentation to a physician. Some patients with carcinoma of
the ampulla present with a clinical picture of cholangitis or pancreatitis, or both. Schlippert et al.1(7185)
found that the serum amylase was elevated in 5 of 19 patients (26%) with ampullary carcinoma, but
these authors did not comment on whether this laboratory abnormality was associated with abdominal
pain consistent with pancreatitis. Acute pancreatitis was the presenting picture in 6 of 41 patients
(14.6%) reported by Robertson et al.,6(7186) 7 of 40 patients (18%) in the study of Jones et
al.,19(7187) in 3 of 52 patients (6%) of Neoptolemos et al.,7(7188) and in 2 of 35 patients (5%) in the
report of Hayes et al.8(7189) Cholangitis is an uncommon presentation of ampullary carcinoma; it
occurred in only 6 of 102 cases (5.8%) reported by Yamaguchi and Enjoji.5(7190) A large number of
patients with ampullary carcinoma have coexisting cholelithiasis or choledocholithiasis. However, no
data support the hypothesis that biliary calculi are a factor in the pathogenesis of ampullary
neoplasms.15(7191) Stones and ampullary carcinoma coexisted in 12 of 37 patients (32%) in the study
by Hayes et al.,8(7192) and in 22 of 58 patients (38%) reported by Baczako et al.20(7193) As might be
expected, this has led to diagnostic errors and inappropriate operations. In the 12 patients of Hayes et
al.8(7194) with both gallstones and ampullary carcinoma, 9 had undergone cholecystectomy within the
preceding 10 months in the mistaken belief that stones had caused the patients' clinical syndromes.
Preoperative findings of anemia, occult blood in stool specimens, or weight loss in a patient with biliary
stone disease warrant both endoscopic visualization of the main duodenal papilla and additional
intraoperative maneuvers including palpation of the ampulla, choledochoscopy, or duodenotomy to
exclude the possibility of malignancy.7(7195)
Precursor Lesions
Benign adenomas of the main duodenal papilla (Figure 691) are considered to be premalignant
lesions.5,2028(7196) This conclusion is based primarily on the fact that elements of adenomatous
tissue have been found within malignant tumors and on the presence of carcinoma in situ within
resection margins of adenomas (Figure 692). Among 109 surgical cases of carcinoma and 5 cases of
adenoma of the main duodenal papilla, Yamaguchi and Enjoji5(7197) found that 20 (18%) of the
carcinomas contained areas of unequivocal adenoma at the margins, and that 2 of the 5 adenomas
had foci of carcinoma in situ. A benign adenomatous component was found by Hayes et al.8(7198) in
11 of 31 (35%) surgically resected ampullary carcinomas. In this study, 11 of 23 patients (48%) had
tumors less than 3 cm in diameter that contained adenomatous tissue but only 1 of 8 (12%) tumors
greater than 3 cm in diameter had a benign component. Hayes et al.8(7199) concluded that this finding
raises the possibility that an advancing invasive carcinoma might overgrow a preexisting benign
adenoma. Baczako et al.20(7200) studied 58 invasive carcinomas and found residual villous and
tubular adenomas as well as microadenomas in 53 (91%) of the cases. Kozuka et al.24(7201) reported
similar findings in 18 of 22 cases (82%) of ampullary carcinoma; Perzin and Bridge27(7202) in 22 of 73
cases (30%). In addition, Perzin and Bridge27(7203) found that 22 of 28 (79%) adenomatous lesions
of the ampulla had histologic evidence of carcinoma, whereas only 5 of 11 (45%) nonampullary
duodenal adenomas had foci of carcinoma. They concluded that ampullary adenomas may be more
likely to undergo malignant transformation than adenomas arising elsewhere in the
duodenum.27(7204) Seifert et al.28(7205) reached the same conclusion in a retrospective study of 68
patients with tumors of the main duodenal papilla.
(7206)Figure 691. Endoscopic appearances of the main duodenal papilla. A, Normal. B,

Pseudotumor. C, Adenoma. D, Adenocarcinoma.
(7207)Figure 692. Photomicrograph shows the histopathology of an adenoma of the main

duodenal papilla with foci of severe dysplasia. (Courtesy of Dr. A. Nusrat.)
Patients with familial adenomatous polyposis (FAP) and Gardner's variant (Figure 693) have an
increased incidence of ampullary neoplasms (see also Chapter 49: Diseases of the
Duodenum).2931(7208) Iida et al.32(7209) obtained duodenoscopic biopsies in 24 patients with FAP.
The mean age of their patients was 30.8 years, and in half of the patients the biopsies from the
ampulla revealed adenoma. In 2 of the 12 biopsy-proven cases of adenoma, the endoscopic
appearance of the papilla was normal. Hypotonic duodenography demonstrated a duodenal
abnormality in only 1 of the 12 cases of adenoma. These results raise the possibility that screening
endoscopy, with visualization and procurement of biopsies from the main duodenal papilla, may detect
a significant number of premalignant lesions and therefore should be considered in all patients with
FAP.
(7210)Figure 693. Endoscopic appearance of the main duodenal papilla and duodenum in a
patient with Gardner's syndrome.
Evidence indicates that an association may exist between the presence of duodenal adenomas and
that of colonic adenomas. In 11 of the 21 patients with duodenal adenomas of Seifert et al.,28(7211)
colonoscopy revealed adenomas in 8 (72.7%). Four of these 8 patients were known to have Gardner's
syndrome. These investigators concluded that when an adenoma of either the ampulla or the
duodenum is diagnosed, colonoscopy is mandatory to exclude colonic adenomas. Further data,
especially data from prospective studies, are needed to corroborate these findings.
Diagnosis
Once the clinical suspicion of an ampullary tumor exists, duodenoscopy and ERCP should be
performed. These two techniques frequently permit identification and localization of the tumor,
confirmation of its malignant nature by obtaining biopsies, and differentiation from other lesions such
as pancreatic carcinoma, bile duct carcinoma, and choledocholithiasis.
The yields of other diagnostic modalities, such as barium contrast studies of the upper gastrointestinal
tract (Figure 694), abdominal ultrasonography, and computed tomography (CT) (Figure 695), are
variable. Walsh et al.2(7212) found that an upper gastrointestinal series was abnormal in 60.6% and
diagnostic in 27.3% of their patients with ampullary adenocarcinoma. Barton and Copeland33(7213)
noted periampullary abnormalities on upper gastrointestinal x-rays in 45 of 56 patients (80.3%) with
ampullary adenocarcinoma. Makipour et al.16(7214) obtained upper gastrointestinal series in 24 of
their 38 patients of ampullary carcinoma and found nonspecific abnormalities in only 11 (46%). In a
prospective, multicenter trial that included 30 patients with ampullary carcinoma, Bakkevold et
al.34(7215) found that the sensitivities of several diagnostic studies in making a conclusive diagnosis
were as follows: abdominal ultrasonography 23%, CT 58%, ERCP 78%, ERCP and duct cytology 91%,
and percutaneous transhepatic cholangiography 100%.
(7216)Figure 694. Upper gastrointestinal barium x-ray series demonstrates an ampullary

tumor (between arrows).
(7217)Figure 695. Computed tomography (CT) appearance of an ampullary tumor (arrow).

The appearance of the papilla in patients with ampullary adenocarcinoma varies from normal to an
obvious, ulcerated mass (see Figure 691). Tasaka35(7218) classified these tumors by their
endoscopic appearance as "intramural protruding" if the only abnormality was an enlarged papilla,
"exposed protruding" if the papilla was enlarged and tumor was seen protruding from it, or "ulcerative"
if an obvious, ulcerated, exophytic mass was present. Ponchon et al.36(7219) noted a normal
endoscopic appearance of the papilla in 37% of patients in their series of 52 patients with
biopsy-proven ampullary adenomas or carcinomas. In addition, they found that 42% of intraampullary
tumors became apparent only after endoscopic sphincterotomy. In those cases, biopsies were
indicated because of prominence of the infundibulum, unexplained common bile duct dilation, or an
exophytic appearance of the papilla after sphincterotomy.
Endoscopic Biopsy
Although forceps biopsy during ERCP is recommended for suspected ampullary tumors, accurate
histologic diagnosis is not always possible. Yamaguchi et al.37(7220) retrospectively reviewed their
series of 73 patients with ampullary carcinoma and found an overall biopsy sensitivity of 70%. The vast
majority of the errors were accounted for by the classification of malignancies as "moderate atypia or
adenoma," which suggests that a finding of this degree of dysplasia should dictate, at a minimum, that
additional biopsies be obtained. In the study of Yamaguchi et al.,37(7221) biopsy accuracy was 50%
for tumors of the intramural protruding type, 64% for the exposed protruding type, and 88% for the
ulcerative type. In other studies, the yield from endoscopic forceps biopsies has ranged from 50 to
90%.3842(7222)
The use of a polypectomy snare to obtain ampullary biopsies increases the size of the specimen but
also the risk for bleeding as compared with the use of a standard forceps. According to
Safrany,38(7223) the diagnostic yield increases from 60 to 83% when a snare biopsy is obtained rather
than standard endoscopic forceps biopsies.
The location and timing of endoscopic biopsies are important parameters in the diagnosis of ampullary
carcinoma. The exact biopsy site should be communicated to the pathologist because the morphology
of the epithelium changes proximal to the papilla. Dark columnar cells that can be present normally in
the region of the papilla could be interpreted as malignant in a specimen taken higher in the bile
duct.43(7224) In ulcerated tumors, biopsies should be taken at the nodular edge and not from the
base, as the latter often consists of necrotic exudate.17(7225) Should biopsies be needed in a situation
where papillotomy has been performed, they should be taken during the index procedure. Bourgeois et
al.41(7226) performed biopsies in 22 patients with benign biliary tract disease at the time of, or 2 days
after, an endoscopic papillotomy. In the 10 patients with a prominent papilla in this study, biopsies
obtained 2 days after papillotomy revealed moderate or severe atypia that had not been appreciated at
the time of papillotomy. In the remaining 12 patients with a normal-appearing papilla, 3 manifested
cellular atypia 48 hours after, but not at the time of papillotomy. One of these patients developed
cellular atypia resembling a neoplasm that persisted, albeit to a lesser extent, for at least 6 days. The
results of this study suggest that papillotomy may create histologic changes, lasting at least 6 days,
that mimic dysplasia.
Pseudotumors
There is a wide variation in the endoscopic appearance of the normal major duodenal papilla (see
Figure 691).44(7227) In addition, both malignant and benign lesions can produce a tumor-like
appearance.45(7228) A retrospective review of 1500 consecutive ERCP examinations revealed that 63
patients had a tumor-like appearance of the main duodenal papilla.46(7229) Forty-seven of these
patients underwent careful photographic, radiologic, and histologic examination as well as follow-up
that ranged from 18 to 45 months; 14 patients had benign lesions. Five of these patients with benign
lesions also had gallstone disease. The only characteristic distinguishing between benign and
malignant disease was the presence of an ulcerative tumor mass, this being seen in 21% of the
patients with malignancy but in none of those with benign disease.
Leese et al.45(7230) reported that 49 of 2000 patients who underwent ERCP had the appearance of a
periampullary carcinoma on duodenoscopy. Biopsies were taken both before and after sphincterotomy,
and ERCP was attempted by insertion of a catheter through the tumor mass. Thirty-eight patients had
neoplasms, whereas 11 had inflammatory, nonneoplastic lesions that were termed pseudotumors (see
Figure 691B). There were no statistically significant differences between these two groups with regard
to any of the clinical features recorded except that a higher proportion of patients with pseudotumors
had epigastric pain. None of the patients with pseudotumors had a palpable gallbladder, whereas this
finding was present in 16 patients with ampullary tumors. Comparison of hematologic and biochemical
tests revealed significantly lower values for hemoglobin, total protein, albumin, and alanine
transaminase in the tumor group. Comparison of ERCP results disclosed that, although the diameter of
the pancreatic duct tended to be higher in the tumor group, there was no difference with respect to
common bile duct diameter, and the cholangiograms appeared similar in many cases (Figure 696).
Gallstones were present significantly more often in the pseudotumor group. Two of the 11 patients with
pseudotumors were subjected to surgical excision because of histologic features on endoscopic
biopsies that raised a suspicion of carcinoma. Thus, although certain clinical, laboratory, and
endoscopic findings tend to be associated with ampullary carcinoma, there is much overlap with the
findings in benign disease. When endoscopic biopsies are equivocal for carcinoma and the suspicion
of a pseudotumor exists, repeating duodenoscopy with biopsies in 2 to 4 weeks is indicated. A
significant proportion of pseudotumors, especially those relating to gallstone disease, will have
resolved in this time period, thus preventing unnecessary surgery.
(7231)Figure 696. Endoscopic retrograde cholangiopancreatography (ERCP) findings in

adenocarcinoma of the main duodenal papilla. A, Tumor projects into the dilated common bile
duct with dimensions defined by contrast material in the duodenum. B, Endoscopic biopsy of
the same lesion. C, Pseudotumor.
Endoscopic Sphincterotomy
Endoscopic sphincterotomy plays an important role in the diagnosis of ampullary tumors (Figure 697),
especially those of the intramural protruding type. Ponchon et al.36(7232) noted that in 6 of 13 patients
with macroscopically normal ampullas, a polypoid mass evaginated from the ampulla after
sphincterotomy. In the study of Nakao et al.,43(7233) two patients with ampullary carcinoma had a
normal-appearing papilla at endoscopy, but an irregular filling defect was seen during fluoroscopy at
ERCP. In 1 patient, the defect was close to the ampulla, prompting the investigators to advance a
biopsy forceps under fluoroscopic guidance into the area to obtain biopsies that subsequently revealed
malignant tissue. A defect was demonstrated in the distal portion of the common bile duct of the
second patient and papillotomy resulted in prolapse of a large polypoid mass. Nakao et al.43(7234)
stated that papillotomy should be considered only when a definitive abnormality is visualized
radiographically during ERCP and a blind biopsy is not technically possible. Huibregtse and
Tytgat39(7235) state that, as a rule, an endoscopic papillotomy should be performed only when it is
considered essential for diagnostic purposes or to eliminate the possibility of a tumor simulating stone
impaction. Other authorities routinely employ papillotomy, noting that it doubles the sensitivity of
subsequent forceps biopsies in detecting ampullary carcinoma.41(7236)
(7237)Figure 697. Endoscopic appearance during (A) and after (B) sphincterotomy.
Treatment
Surgical Treatment
Pancreaticoduodenectomy
Surgical excision remains the mainstay of treatment for adenocarcinoma of the main duodenal papilla
in good-risk patients without metastatic disease (Figure 698). The four main choices for operative
intervention with curative intent include pancreaticoduodenectomy, total pancreatectomy, regional
pancreatectomy, and local excision. Although the ability to detect regional and distant metastases by
noninvasive means is improving, the surgeon must carefully examine the liver and peritoneum; search
for involved lymph nodes in the gastrohepatic ligament, the celiac axis, and the origin of the superior
mesenteric artery; and finally, mobilize the duodenum by a Kocher maneuver to search for
peripancreatic metastatic nodes and posterior invasion.47(7238) If the tumor is deemed resectable, the
procedure of choice is pancreaticoduodenectomy, the Whipple version being the most popular until the
more recent development of pylorus-preserving variations.
(7239)Figure 698. Surgical resection specimen of adenocarcinoma of the main duodenal

papilla. A, Appearance with a grossly dilated bile duct. B, Photomicrograph of a typical poorly
differentiated adenocarcinoma. (A and B, Courtesy of Dr. I. C. Talbot.)
Pylorus-preserving pancreaticoduodenectomy is reported to be an easier and less time-consuming
operation with less blood loss, a shorter hospital stay, and better weight gain during follow-up48(7240)
with no differences in recurrence rate and patient survival compared with the standard Whipple
procedure. Operative mortality for the Whipple procedure has declined from 26 to less than 5%,
although the associated morbidity remains in the range of 30 to 60%.7,4956(7241)
Typically, total pancreatectomy is performed only in those patients in whom a safe pancreatojejunal
anastomosis cannot be constructed.5760(7242) Regional pancreatectomy, an operation developed
by Fortner,61(7243) involves the en bloc removal of a tumor in or immediately adjacent to the
pancreas, with an adequate soft tissue margin, and with its regional lymphatic drainage. This operation
may allow for a significantly higher resectability rate, compared with the Whipple procedure, of
ampullary carcinomas with vascular involvement because it involves resection of the pancreatic
segment of the portal vein, resection of involved arterial structures if necessary, and a total or subtotal
pancreatectomy.62(7244)
Local Excision
An argument in favor of local surgical excision for small ampullary adenocarcinomas can be made only
for patients who are not fit to undergo pancreaticoduodenectomy.7,5056,63(7245) Local surgical
excisions of ampullary carcinoma in good-risk patients is controversial. Halsted62(7246) described the
first such local excision in 1899. The patient succumbed 6 months later from recurrent carcinoma in
the duodenum and pancreas. The current surgical technique as outlined by Sharp and
Brandes64(7247) involves a generous Kocher maneuver, palpation of the tumor transduodenally, and
passage of a biliary Fogarty catheter through the cystic duct and across the tumor, or through a
choledochotomy, to aid in reconstruction of the common duct after the excision. Next, a vertical
duodenotomy is made opposite the papilla, and the mobility and fixation of the tumor are assessed.
The tumor is excised with a rim of normal duodenal tissue, and biopsies are obtained from surrounding
tissues for evaluation as frozen sections. Finally, the bile and pancreatic ducts are opened
longitudinally and anastomosed to the duodenal edges. The problem of missing foci of carcinoma
within endoscopic biopsies or frozen section biopsies of polypoid lesions of the ampulla mitigates in
favor of more extensive excision. Farouk et al.65(7248) performed local surgical resection of tumors of
the main duodenal papilla in six patients, three of whom had adenocarcinoma. Although none of the
patients had postoperative complications, all three with malignant disease developed recurrences at 9,
24, and 32 months after surgery. One patient died 35 months after surgery, and the other two patients
were alive at 6 and 8 months after recurrence was diagnosed. For patients with malignant ampullary
tumors, Farouk et al.65(7249) concluded that local surgical resection should be considered only a
palliative procedure and reserved for elderly patients with severe concomitant disease or patients in
whom a more radical procedure would incur undue surgical risk.
Retrospective reviews demonstrate equivalent survival data in patients with ampullary carcinoma who
undergo local surgical excision compared with those who undergo
pancreaticoduodenectomy.6,66(7250) Although formal guidelines for following up patients who have
undergone local excision of ampullary tumors do not exist, duodenoscopic surveillance is logical
because the recurrence rate is significant. Semiannual examination for 1 or 2 years followed by annual
examination is a suggested surveillance schedule.
Biliary Bypass
A palliative proximal biliary-enteric and gastrojejunal bypass procedure can often be performed in
patients unfit for radical procedures who have large bulky tumors that obstruct a significant portion of
the duodenal lumen and the bile duct. The operative mortality rate of patients with malignant ampullary
tumors undergoing palliative surgical bypass ranges from 17 to 60% with a mean of
32%.1,2,4,7,67,68(7251) For patients with inoperable disease but without significant duodenal
obstruction, endoscopic palliation should be considered in lieu of surgical bypass because the
morbidity and mortality are lower (see later).
In carefully selected patients, there is a place for reoperation for recurrent periampullary
adenocarcinoma. Of 55 patients in the series of McGuire et al.69(7252) who underwent reoperation for
periampullary adenocarcinoma between 1979 and 1990, 11 had ampullary adenocarcinoma. The
median interval between initial surgery and second exploration for patients with nonpancreatic
periampullary adenocarcinoma was 2 to 4 months; the most common symptoms on presentation were
weight loss (31%), pain (25%), jaundice (25%), fever and chills (12%), and nausea and vomiting (6%).
McGuire et al.69(7253) recommended that patients being considered for reoperation undergo CT and
mesenteric arteriography to exclude those with vascular encasement. All 11 patients with ampullary
adenocarcinoma were deemed resectable and were reoperated with intent to cure. All but one
subsequently underwent pancreaticoduodenectomy. There were no postoperative deaths, and the
5-year actuarial survival rate was 31%. Median survival calculated from the time of the initial
exploratory surgery was 36 months. This study suggests that excellent results can be obtained with
reoperation for ampullary adenocarcinoma in carefully selected patients at a referral center for
pancreatobiliary surgery.
The role of endoscopy in the management of tumors of the main duodenal papilla continues to evolve.
In 1980, Safrany70(7254) first reported endoscopic tumor sphincterotomy to be efficacious (see Figure
697) in a series of 21 patients with ampullary carcinoma who were deemed inoperable. The average
symptom-free interval after sphincterotomy was 5.5 months; 5 patients underwent a second successful
sphincterotomy for recurrent jaundice.
The role of preoperative biliary drainage in a patient with known ampullary carcinoma is controversial.
Most studies have employed percutaneous drainage procedures that are associated with a significantly
higher complication rate than that of endoscopic drainage techniques.71(7255) Although there have
been no controlled trials to determine whether preoperative endoscopic drainage lowers morbidity and
mortality, this approach improves liver function and nutrition72(7256) and provides an opportunity to
obtain additional biopsies if the initial histologic diagnosis is equivocal. Preoperative endoscopic biliary
drainage is therefore employed routinely by some groups in patients with evidence of an obstructed
biliary system.7(7257)
Neoptolemos et al.7(7258) assessed treatment and outcome in 52 consecutive cases of ampullary
adenocarcinoma. Twenty-four patients had a Whipple resection (12.5% mortality), 4 patients had a
local resection (no deaths), 10 patients had surgical bypass (60% mortality), and 13 patients had
endoscopic drainage alone (23% mortality). The group of patients selected for endoscopic drainage
alone had more risk factors (greater age, lower hemoglobin, lower serum albumin) compared with
those undergoing surgery. Despite this, successful endoscopic drainage was achieved in 10 of the 13
patients. Only 1 patient received an endoprosthesis because the sphincterotomy was inadequate. Two
of the 10 patients required needle-knife papillotomy because the common bile duct could not be
otherwise cannulated. Of the 3 patients in whom endoscopic treatment was unsuccessful, 1 had a
previous partial gastrectomy and Billroth type II anastomosis, the tumor was very large in another, and
in 1 patient sphincterotomy was attempted during the early stage of experience with this technique. No
immediate complications occurred owing to the endoscopic procedures. This experience clearly
illustrates that not only is endoscopic drainage feasible but also that it is associated with a much lower
mortality rate compared with surgical bypass procedures.
Endoscopic Sphincterotomy
Bickerstaff et al.73(7259) found endoscopic sphincterotomy to be a safe and effective palliative
technique in patients with ampullary adenocarcinoma. Seventeen patients deemed unfit for surgical
resection because of comorbid disease (10 patients), hepatic metastases (1 patient), or dementia or
general frailty (6 patients) underwent sphincterotomy with (4 patients) or without (13 patients) insertion
of a single 10-French endoprosthesis (straight type). Three patients had large tumors that prevented
cannulation of the papilla, prompting surgical bypass procedures. Early complications after
sphincterotomy included fatal hemorrhage in 1 (6%) and cholangitis in 2 patients (12%), both of whom
had received an endoprosthesis. Cholangitis resolved in 1 of these 2 patients after endoprosthesis
removal. Jaundice resolved in a median of 3 weeks in the 15 patients successfully treated by
endoscopy, and the median stay in the hospital after sphincterotomy was 3 days. Recurrent jaundice
developed between 4 and 11 months after sphincterotomy in 8 of 16 (50%) patients. Cholangitis also
occurred in 2 of these 8 patients; both had an endoprosthesis in place. Repeat sphincterotomy or stent
replacement was successful in 5 of the 8 patients who developed recurrent jaundice. Eleven patients
died between 4 days and 23 months after the sphincterotomy; median survival was 12 months.
Endoscopic sphincterotomy is thus an effective means of palliating patients with ampullary
adenocarcinoma and is associated with a shorter hospital stay, fewer complications, and a lower
30-day mortality rate compared with surgical bypass.
Stent Placement
Although several authors recommend endoscopic sphincterotomy without stent placement as the best
preoperative or palliative method, this is not universally accepted. Huibregtse and Tytgat39(7260)
reported a 96% success rate of endoprosthesis placement without sphincterotomy (Figure 699) in a
series of 71 patients with ampullary carcinoma. Jaundice resolved in 95% of patients who received
stents, and there were no procedure-related deaths. In those patients who did not subsequently
undergo surgical resection, the average survival was 466 days, stent clogging occurred in 36% after a
mean of 312 days, and duodenal obstruction occurred in 23% of patients after a mean interval of 256
days. For palliative endoscopic management of ampullary adenocarcinoma, Ponchon et al.36(7261)
concluded that sphincterotomy should be considered as a temporary drainage technique that should
be performed in conjunction with the placement of an endoprosthesis. Although the issue of stent
placement for palliative management of ampullary adenocarcinoma remains unsettled, it must be
emphasized that patients who are discharged from the hospital with an endoprosthesis should be
instructed to seek medical attention immediately if they develop any signs or symptoms of biliary
obstruction or cholangitis.
(7262)Figure 699. Sequence of endoscopic views shows palliation by endoprosthesis

insertion. A, An obstructing tumor of the main duodenal papilla. B, Cannulation and
cholangiography. C, Guidewire insertion. D, Endoprosthesis insertion with drainage of pus.
Permanent placement of an expandable metal stent should be considered in patients with inoperable
ampullary adenocarcinoma or those who are not fit for surgery. Metal expandable stents, such as the
Wallstent (Schneider U.S. Stent, Plymouth, MN), the Gianturco-Rosch Z stent (Wilson-Cook Medical,
Inc., Winston-Salem, NC), and the Strecker stent (Boston Scientific, Watertown, MA), offer two major
advantages: an introduction system with a diameter that is smaller than that of the standard plastic
endoprosthesis, and the potential to achieve lasting palliation with a single endoscopic
procedure.7477(7263) In addition, employing a 1-cm-diameter Wallstent (Figure 6910) significantly
reduces the rate of occlusion by biliary sludge compared with conventional polyethylene stents,
although the problem of tumor ingrowth remains.76(7264) For patients who develop stent occlusion
and recurrent jaundice as a result of tumor ingrowth, insertion of a plastic prosthesis inside the
occluded metal stent78(7265) or thermal ablation of the tumor within the stent79(7266) have been
employed. Metal stents have two major disadvantages: they cannot be removed, and they are relatively
expensive, although a cost-benefit analysis has demonstrated that they are cost-effective.77(7267)
(7268)Figure 6910. Palliation of adenocarcinoma of the main duodenal papilla by insertion

of an expandable metal stent. A, Adenocarcinoma of the main duodenal papilla. B,
Cannulation of the tumor. C, Insertion of an expandable metal prosthesis into the common bile
duct. D, Stent deployment with copious bile drainage. E, Closer view shows a wide lumen
through the tumor into the common bile duct. Radiographic views of the expandable metal
stent immediately after deployment (F) and after removal of the endoscope (G) with rapid
drainage of contrast material. Note initial constriction of the metal stent at the level of the
tumor.
Laser Therapy
Information on endoscopic laser treatment of ampullary tumors is very limited. One half of the 52
patients treated by Lambert et al.80(7269) had adenocarcinoma. Photoablation using a
neodymium:yttrium-aluminum-garnet (Nd:YAG) laser was attempted in 8 patients with papillary
adenocarcinoma. In 7 of the 8 patients who underwent laser photoablation of the exophytic
intraduodenal portion of an ampullary adenocarcinoma, the procedure was done as an adjunct to
drainage by endoscopic sphincterotomy, and eventually stent placement, in nonoperable patients in
poor general condition. Laser treatment did not extend the duration of survival and did not appear to
completely destroy these tumors. In 1 patient in whom an attempt was made to completely destroy the
tumor before resection, histology of the surgical specimen revealed residual tumor. Complications
seen in the 16 patients undergoing laser ablation included 1 case of ulcerative duodenitis and 2 cases
of mild, self-limited pancreatitis.
Brachytherapy
In an uncontrolled clinical trial, Siegel et al.81(7270) treated malignant biliary obstruction by endoscopic
implantation of iridium-192 using a double-lumen endoprosthesis. Four of the 14 patients had
unresectable ampullary adenocarcinoma. Most underwent sphincterotomy, followed by 5000 rads (48
hours) via the double-lumen endoprosthesis, with subsequent removal of the iridium-192
endoprosthesis and placement of a large-caliber prosthesis for long-term drainage. No complications
were encountered and reduction of tumor mass was observed in 2 of the patients.
Pathology and Staging

Staging Schemes
Initially, adenocarcinoma of the main duodenal papilla was classified into two general types based on
the gross appearance: polypoid and ulcerative.10(7271) Tasaka35(7272) later divided the protruding
type into intramural and exposed forms. The study by Tasaka35(7273) and that of Yamaguchi and
Enjoji5(7274) both demonstrated that the intramural protruding type of tumor represents an earlier
phase that can progress to either the exposed protruding or the ulcerative type. Using this macroscopic
classification scheme, Yamaguchi and Enjoji5(7275) found lymph node metastases in 29% of cases of
the intramural protruding type, 53% of those classified as exposed protruding, and 69% of
ulcerative-type cases.
Histopathologic studies of malignant tumors of the main duodenal papilla have demonstrated that
these lesions are almost exclusively adenocarcinoma. Rarely, squamous and mucinous elements may
be found within these malignant tumors. In addition, these tumors may be characterized as having a
papillary (villous) or a tubular architecture as well as a wide range of differentiation from highly
undifferentiated to well differentiated. Edmondson82(7276) and Blumgart and Kennedy83(7277)
divided ampullary carcinoma histologically into three types: papillary, nonpapillary, and mixed.
Carcinoma of the ampulla spreads by extension to contiguous organs, especially the pancreas,
duodenal wall, or bile ducts, and by permeation of lymphatic or venous channels. At least five different
systems have been proposed for staging tumor spread.5,33,8486(7278) Variations in these staging
schemes make comparison of results from different series difficult. Barton and Copeland33(7279)
proposed seven stages based on histologic findings. Of note, 27 of 56 patients (48%) in this study had
either Stage I, II, or III disease that was potentially curable by operation.33(7280)
Fortner84(7281) devised a TNM (Tumor size, lymph Node involvement, Metastases) postsurgical
staging system for cancer of the pancreatic region. Only 4 of 61 patients in this study had carcinoma of
the main duodenal papilla, and the stage of these patients was not described. In 1987, the staging of
ampullary carcinomas was added to the TNM classification of malignant tumors by the International
Union Against Cancer (UICC),83(7282) and since then this classification has been utilized widely
(Table 691). Of the 24 patients with carcinoma of the ampulla described by Mori et al.,87(7283) 5
(20.8%) had Stage I disease, 11 (45.8%) had Stage II disease, 7 (29.2%) Stage III disease, and 1
(4.2%) had Stage IV disease at diagnosis.
Ampullary Carcinoma
TABLE 691
Tx:
Primary tumor
No tumor assessable
Tumor limited to ampulla of Vater
Tumor invading duodenal wallin particular,
muscularis of duodenum
Tumor invading 2 cm or less into pancreas
Tumor invading more than 2 cm into pancreas or
other adjacent organsin particular, major blood
vessels
N:
N0:
Regional lymph nodes

T:
T0:
T1:
T2:
T3:
T4:
Ampullary Carcinoma
TABLE 691
N1:
Nx:
M:
M0:
M1:
Mx:
Regional lymph node metastasis

Distant metastasis
Distant metastasis: hepatic metastasis, peritoneal
dissemination, or lymph node metastasis along
splenic vein or at splenic hilum
STAGE GROUPING
Stage I:
T1 N0 M0
Stage II:
T2 N0 M0, T3 N0 M0
Stage III: T1 N1 M0, T2 N1 M0, T3 N1 M0
Stage IV: T4 any N M0, any T any N M1
TNMtumor-node-metastasis.
Yamaguchi and Enjoji5(7284) used the classification of Tasaka35(7285) in a detailed histopathologic

study of 109 cases of carcinoma of the ampulla; 23% of the tumors were of the intramural protruding
form, 38.5% were classified as exposed protruding, and the remaining 38.5% were of the ulcerating
form. Tumor size was largest in the ulcerating form, with an average of 3.3 cm in greatest diameter,
and smallest in the intramural protruding form, with an average of 1.8 cm. Microscopic examination
revealed that 56 (51.4%) of the tumors were well-differentiated papillary adenocarcinomas, 49 (44.9%)
were tubular adenocarcinomas including 21 well differentiated, 20 moderately differentiated, and 8
poorly differentiated; 3 (2.8%) of the tumors were adenosquamous carcinomas, and 1 (0.9%) was a
mucinous carcinoma.
Yamaguchi and Enjoji5(7286) described four tumor stages: Stage I (tumor restricted to inside the line
of the sphincter muscle), Stage II (tumor invading the submucosa of the periampullary duodenum
without further extension), Stage III (tumor involving the muscularis propria of the duodenum), Stage IV
(tumor extending into the pancreas). All 12 tumors in Stage I were protruding types (9 intramural, 3
exposed); 19 of the 22 tumors in Stage II were of the protruding type (10 intramural, 9 exposed), and
the remaining 3 were of the ulcerating type. Of the 33 tumors in Stage III and 40 tumors in Stage IV, 33
(47%) were protruding and 37 (53%) were ulcerating tumors. These data support the contention that
intramural protruding tumors of the ampulla progress to exposed protruding or ulcerating tumors in the
course of their development.5(7287) The Stage I tumors were the smallest, measuring on average 2.1
cm in diameter, whereas the Stage IV tumors were the largest, having an average diameter of 3.1 cm.
Lymphatic invasion was found in 25% of the patients with Stage I tumors, 45% of those with Stage II
tumors, 88% of those with Stage III tumors, and in 90% of those with Stage IV tumors. In the
respective stages, venous infiltration by tumor cells was seen in 17%, 18%, 39%, and 50% of the
patients with a periampullary tumor. Lymph node metastases were present in 0%, 27%, 55%, and 78%
of patients with tumors of Stages I through IV, respectively.
Talbot et al.86(7288) reviewed the histologic characteristics of 26 surgically resected adenocarcinomas
of the ampulla including size of tumor, histologic type, degree of differentiation, and extent of local
invasion. These investigators defined Stage I disease as local invasion confined to the wall of the
common bile duct; Stage II as invasion outside the bile duct wall into duodenal submucosa,
muscularis, or adventitial fat; Stage III disease as invasion of pancreatic tissue; and Stage IV disease
as the presence of lymph node metastases, regardless of local invasion. This staging scheme differs
from that of the TNM classification system in that the former divides patients in the TNM Stage II group
into two different stages according to the presence or absence of pancreatic invasion. This subdivision
makes it possible to evaluate the effect of pancreatic invasion on survival. Of note, 25 of the 26 (96%)
ampullary carcinomas were of intestinal morphology, and there was only 1 adenocarcinoma that was
described as truly papillary in nature. This finding is consistent with the conclusion of Morson and
Sobin88(7289) that ampullary carcinomas are essentially of intestinal type, but it is contrary to the
findings of others, specifically that these tumors fall into two categories, either papillary or
nonpapillary.83(7290)
The most significant recent advance in the staging of periampullary tumors is endoscopic
ultrasonography (EUS).89,90(7291) This staging modality is distinguished from all others by its ability
to provide a nonsurgical assessment of the histology of the gastrointestinal tract wall and the extent of
tumor invasion in relation to the gut wall. Although it is well known that transabdominal
ultrasonography, CT, and magnetic resonance imaging are helpful in assessing for the presence of
distant metastases and for ductal dilation, their ability to identify the presence of tumor or to define the
extent of tumor invasion in this region is less clear. EUS of the main duodenal papilla and
periampullary area is discussed in Chapter 78: Endoscopic Ultrasonography of the Retroperitoneal
Organs.
Prognosis
The prognosis for adenocarcinoma of the papilla is the most favorable of all the periampullary tumors.
As a result, mortality rates for various studies can be compared only if pancreatic adenocarcinoma,
distal bile duct carcinoma, and duodenal adenocarcinoma are excluded from analysis. Prognosis for
patients with malignant tumors of the papilla depends on the extent of local tumor invasion and the
presence or absence of vascular invasion, regional lymph node involvement, and distant metastases.
As the ability to accurately stage ampullary adenocarcinomas improves, the percentage of lesions
deemed resectable will decrease. In one study, patients with ampullary tumors producing
predominantly sialomucins had a better prognosis than those with tumors secreting sulfated
mucins.91(7292) In addition, expression of carbohydrate antigen 199 or carcinoembryonic antigen by
ampullary carcinomas is associated with a poor prognosis compared with that of ampullary carcinomas
that do not express these antigens.92(7293)
Currently, approximately 50% of ampullary lesions are deemed resectable. With the seven-stage
system of Barton and Copeland,33(7294) differences in survival by stages were not significant. Using
the four-stage system of Yamaguchi and Enjoji,5(7295) the overall 5-year survival rate was 28%, with
85%, 11%, 25%, and 24% 5-year survival rates for Stages I, II, III, and IV, respectively. The survival
rates for patients with tumors staged as II, III, and IV do not decrease with advancing stage, thereby
raising doubt about the practical value of this staging system. In the novel staging system of Talbot et
al.,86,93(7296) a score is determined by adding the tumor grade to the stage. Grade I is well
differentiated, Grade II is moderately differentiated, and Grade III is poorly differentiated. A score of two
to four is associated with a 79% 5-year survival rate, whereas a score of five to seven is associated
with a 0% 5-year survival.
The overall mean 5-year survival rate in 12 series reported since 1982 for patients with ampullary
carcinoma who underwent radical resection is 40% with a range of 16 to 61%.2,3,68,5055,86(7297)
As operative mortality continues to decrease and selection of patients improves, the 5-year survival
rate should improve accordingly.
Benign Tumors
Although malignant tumors of the main papilla are more common, a number of benign neoplasms may
also arise in this structure including benign adenomas (tubular and villous), lipoma, hamartoma,
fibroma, neurogenic tumors including neurofibroma, granular cell tumor, leiomyoma, lymphangioma,
and hemangioma.
The most common benign tumor of the ampulla is the villous adenoma (Figure 6911). The incidence
of this tumor is lower than that of ampullary carcinoma, being between 0.04 and 0.12% in autopsy
series.94(7298) There is no sex predilection, and the average age at diagnosis is 62 years (range 4
months to 79 years).95104(7299) Villous adenomas of the ampulla range in size from 4 mm to 7 cm
in diameter.23,105(7300) In a review of the literature, Schulten et al.98(7301) found that 31 of 42
(74%) villous tumors of the duodenum were located in areas other than the second part. However, of
the 15 cases containing malignancy, approximately half were located in the second portion. In the
more recent series of Ryan et al.,97(7302) 16 of 19 patients with villous tumors of the duodenum had
involvement of the main duodenal papilla.
(7303)Figure 6911. Surgical resection specimen of villous adenoma of the main duodenal
papilla. (Courtesy of Dr. I. C. Talbot.)
Small tumors of the main papilla can be asymptomatic, but symptoms were present in more than 75%
of reported cases.23,95,98(7304) Symptoms often begin when the tumor obstructs the biliary system
or duodenum.97,106,107(7305) The most frequent symptoms are nonspecific upper abdominal
discomfort, jaundice, and upper gastrointestinal bleeding, either occult or overt. Other patients present
with weight loss, fever, lethargy, or pancreatitis (see Figure 691C).103,108(7306) Although villous
tumors have a soft, friable consistency, and are therefore predisposed to bleeding, only half of the
patients with this tumor in the review of Schulten et al.98(7307) presented with gastrointestinal
hemorrhage. Probably the most frequently described sign of ampullary villous adenoma is jaundice,
occurring in 70% of cases.101(7308) According to Sobol and Cooperman,95(7309) coexisting stones
in the gallbladder or common bile duct occur in 13 to 20% of patients. Mucorrhea and electrolyte loss
have not been reported in association with villous tumors of the small bowel, possibly because
absorption of fluid in the distal bowel compensates for any fluid and electrolyte losses from these
tumors.98,109,110(7310) Goldberg et al.99(7311) described an unusual case of a 67-year-old man
with intermittent biliary obstruction due to a 3-cm mobile villous adenoma of the ampulla that prolapsed
in and out of the distal common bile duct.
Diagnosis
Laboratory studies in patients with villous adenoma of the main papilla are likely to show evidence of
only cholestatic jaundice. Imaging studies, as with malignant tumors, occasionally reveal a filling defect
in the duodenum, but this occurs only with large tumors. Indeed, a preoperative diagnosis of villous
adenoma of the main papilla was rarely made before the widespread application of duodenoscopy.
Villous adenomas of the ampulla have a high incidence of coexisting malignancy. In surgical series, the
incidence of malignant change (both carcinoma in situ and invasive carcinoma) ranges from
2695(7312) to 63%.97(7313) Any assessment of prevalence of malignancy within ampullary villous
adenomas must be based on histopathologic analysis of the entire specimen because endoscopic
biopsies have an unacceptably high false-negative rate. In one series, endoscopic examination found
the tumor in 9 of 10 patients, but endoscopic biopsies missed the areas of malignant change in 5 of 9
patients (56%).97(7314)
Treatment
Surgical Treatment
Given the high false-negative rate for endoscopic biopsy in detecting foci of malignancy, local surgical
excision and pancreaticoduodenectomy have been the mainstay of treatment for villous adenomas of
the main duodenal papilla (Figure 6912). The selection of a surgical procedure depends primarily on
the results of endoscopy and intraoperative frozen section biopsies. For tumors without evidence of
malignancy, submucosal excision is the procedure of choice, combined with sphincterotomy or
sphincteroplasty if the tumor is growing within the ampullary orifice.23,98,105107,111(7315) Greater
than 85% of patients with benign villous adenoma treated with local excision do well clinically and
remain recurrence-free.106,107,111,112(7316) If intraoperative frozen sections or examination of the
local excision specimen reveals carcinoma in situ, most authorities recommend radical
pancreaticoduodenectomy,21,97,101,107(7317) although others, especially in older reports, have
recommended a conservative approach with local excision alone.27,110,113116(7318) Celik et
al.101(7319) reported two cases of recurrent tumor with liver metastases when local excision and
choledochoduodenostomy were performed for carcinoma in situ and severe atypia, respectively. For
lesions that contain invasive carcinoma, the operation of choice is
pancreaticoduodenectomy.21,98,105,107,111,117(7320)
(7321)Figure 6912. Ampullary adenoma. A, Surgical resection specimen. B,

Photomicrograph of the resection specimen shows predominant villous architecture. (A and B,
Courtesy of Dr. A. Nusrat.)
Laser Therapy
Endoscopic treatment of benign adenomas of the papilla is assuming a larger role. Lambert et
al.80(7322) attempted Nd:YAG laser photoablation in seven patients with papillary adenomas and one
patient with an adenoma that contained a malignant focus. In three of these eight patients, a previous
snare resection (see later) had been performed with incomplete removal of the tumor. Complete tumor
destruction was achieved in seven of eight cases; follow-up ranged from 14 to 53 months. Recurrence
of the villous adenoma was observed after 24 months in only one patient. No complications were
noted. In this study, argon laser photoablation seemed to be much less effective in destroying
adenomatous tissue residue compared with the Nd:YAG laser. In the series of 52 patients with
ampullary tumors of Ponchon et al.,36(7323) 11 patients underwent attempted adenoma ablation by
snare resection or laser photocoagulation, or both, after sphincterotomy. Adenomatous tissue was
completely ablated in 10 patients with a mean duration of follow-up of 39 months. Five other patients
with severe comorbid illness or metastatic malignancy underwent successful palliative treatment with
sphincterotomy alone (4 patients) or with a transhepatically placed stent (1 patient). Four of these
patients died of cardiac disease at a mean of 45 months after procedure, and 1 was alive at 15 months
after sphincterotomy.
Diathermic Fulguration
Shemesh et al.118(7324) reported five patients with adenomas of the papilla who presented with
obstructive jaundice and underwent successful drainage via endoscopic sphincterotomy. Four of these
patients underwent local surgical excision with subsequent adenoma recurrence in all four within 6 to
18 months after the operation. The recurrent adenomas were eradicated by endoscopic diathermic
fulguration using the tip of a polypectomy snare or a "hot biopsy" forceps; no recurrences were noted
during follow-up that ranged from 12 to 24 months. One patient with a sessile adenoma of the papilla
refused surgery and therefore underwent incomplete piecemeal resection of the adenoma by snare
polypectomy; adenocarcinoma of the head of the pancreas developed 40 months later.
Snare Papillectomy
Binmoeller et al.119(7325) described 25 patients with adenomatous tumors of the main papilla who
underwent radical endoscopic resection by snare papillectomy. The three criteria for inclusion in this
study were: size less than 4 cm, benign endoscopic appearance, and benign histologic results on at
least six forceps biopsies. Twenty-three patients had de novo tumors and 2 had recurrent adenomas
after local surgical excision. Endoscopic excision was performed with a pure cutting current to the level
of the muscularis propria. Residual tissue was fulgurated using monopolar current. ERCP was
performed after papillectomy, then biliary or pancreatic duct papillotomies, or both, were performed as
well as endoprosthesis placement (Figure 6913) in some patients as needed to achieve satisfactory
drainage.
(7326)Figure 6913. Endoscopic snare excision of an adenoma of the main duodenal papilla.
A and B, Before resection. C, During cannulation. D, With the snare in position. E,
Immediately after removal. F, During pancreatic stent insertion. G, During biliary stent
insertion. H, Final appearance after ductal stenting at the completion of the procedure.
Surveillance duodenoscopy was performed at 1, 6, and 12 months after papillectomy and yearly
thereafter. Immediate complications included postpapillectomy bleeding in 2 patients necessitating
local injection of epinephrine (1:20,000 solution), and self-limited pancreatitis in 3 patients. There were
no procedure-related deaths. Two patients had evidence of intraductal tumor extension by ERCP and
were referred for surgery. The remaining 23 patients were observed for a median of 37 months; 6
patients (26%) had benign recurrences, usually within the first year of follow-up. Of the 6 patients with
recurrences, 1 had tumor extension into the distal common bile duct and later succumbed to
complications of a Whipple procedure. The remaining 5 patients with recurrence of adenoma were
treated endoscopically with snare resection (1 patient), diathermic fulguration (2 patients), or combined
treatment (2 patients). Three of these 5 patients were free of disease at 13, 52, and 53 months. One
patient was lost to follow-up, and 1 underwent pancreatoduodenectomy after repeated fulgurations
failed to completely ablate adenomatous tissue.
Silvis120(7327) correctly urged caution and careful consideration of several points before embarking
on endoscopic papillectomy of benign ampullary tumors. First, the group of patients in the study of
Binmoeller et al.119(7328) is unusual in that no patient had cancer within the adenoma, whereas the
incidence in other reports ranges from 26 to 63%.95,97(7329) Second, papillectomy is associated with
a significant recurrence rate and should therefore be considered only for patients who are likely to be
compliant with an extended schedule of follow-up endoscopic surveillance procedures. Third,
papillectomy should be performed only by an endoscopist expert in ERCP. Lastly, this procedure
should be restricted to either patients who refuse surgery or those who are at high risk for extensive
surgical procedures.
Prognosis
Based on the few case reports and reports of small series available, it appears that long-term survival
should be expected for patients who undergo resection of benign ampullary tumors and for those with
villous adenoma containing in situ carcinoma.98,113,121125(7330) Schulten et al.,98(7331) based on
a review of available follow-up data, calculated a 5-year survival rate of 21% for patients with villous
adenoma of the duodenum harboring malignant change. Celik et al.101(7332) concluded that radical
excision of malignant ampullary villous adenomas in good-risk patients provides long-term survival in a
third of these.
Neuroendocrine and Metastatic Tumors

Neuroendocrine Tumors
Neuroendocrine tumors of the main duodenal papilla are rare malignant neoplasms.15(7333) These
tumors have both neurogenic and endocrine elements in a number of cases. Lesions of endocrine
origin have been termed carcinoid tumors. Duodenal carcinoids have both morphologic and functional
similarities to pancreatic islet cell tumors and have been termed carcinoid islet cell tumors.126(7334)
Although the carcinoid tumor is the most common neoplasm of the small intestine, duodenal carcinoids
(Figure 6914) account for only 1 to 5% of the total.127(7335) Fewer than 40 cases of ampullary
carcinoids exist in the world literature; there appears to be no age or sex predilection for these
tumors.128(7336)
(7337)Figure 6914. Endoscopic view of ampullary carcinoid tumor.

Although most carcinoid tumors are asymptomatic and are found incidentally, those arising at the
ampulla tend to cause symptoms related to intermittent biliary obstruction.129(7338) Symptoms of
ampullary carcinoid/islet cell tumors are the same as those for any tumor in this location: abdominal
pain, melena, recurrent acute pancreatitis, and jaundice. Acute pancreatitis as a presenting feature in
the absence of cholestasis has been reported.130(7339) The carcinoid syndrome has not been
reported in association with an ampullary carcinoid tumor.131(7340) This observation probably reflects
the smaller size of these tumors when detected at the ampulla, a key factor in predicting the incidence
of metastases.132(7341)
The association between ampullary carcinoid/islet cell tumors and von Recklinghausen's disease was
first recognized by Lee and Garber133(7342) in 1970; there have been a number of subsequent
reports.134136(7343) Both cutaneous and viscerocutaneous neurofibromatosis have been
associated with ampullary carcinoid/islet cell tumors. A case of ampullary somatostatinoma has been
reported.137(7344)
Ampullary carcinoid/islet cell tumors have no distinctive endoscopic features. The pitfalls associated
with endoscopic biopsies of ampullary tumors apply to these lesions as well. Neuroendocrine tumors of
the main duodenal papilla may spread locally to the duodenal wall, pancreas, or bile duct or by
metastases to local and regional lymph nodes. In approximately one fourth of the cases of duodenal
carcinoids, lymph node metastases are present at diagnosis.138(7345) In the four cases of ampullary
carcinoid/islet cell tumors reported by Stamm et al.,128(7346) two had lymph node metastases.
Management of these tumors is primarily surgical, with local excision and pancreaticoduodenectomy
being the two operations most commonly employed. Although available data are limited,
pancreaticoduodenectomy may be assumed to be associated with a higher perioperative morbidity and
mortality but a lower risk of recurrence and tumor spread compared with local excision.15(7347) The
5-year survival rate for patients with carcinoid tumors without apparent metastasis who undergo
resection was 68 and 95% in two different studies.132,139(7348) Of note, the 5-year survival rate in
patients with regional lymph node metastasis or distant metastasis is 83% and 38%,
respectively.132(7349)
Other Malignant Tumors

Metastases to the main duodenal papilla are rare, but they have been reported with renal cell
carcinoma,140142(7350) melanoma,143(7351) lymphoma,143(7352) lymphangioma,144(7353) and
endometrial adenocarcinoma.145(7354) Symptoms may arise as a result of bleeding, obstruction,

perforation, or malabsorption caused by blockage of the pancreatic duct. Treatment is usually
palliative.
Approach to the Patient with an Ampullary Tumor

A rational approach to management of tumors involving the main duodenal papilla must recognize that
endoscopic biopsy has limitations, that carcinoma often coexists within an adenoma, and that
"pseudotumors" may occur that are often difficult to distinguish from true neoplasms. Whenever an
ampullary "tumor" is discovered, a complete ERCP should be performed, multiple biopsies should be
taken, a sphincterotomy should be performed, and then more biopsies should be taken from the cut
surface. A stent should be placed if bile drainage is suboptimum, cholangitis is present, or the patient
is not a surgical candidate. If biopsies reveal adenocarcinoma or malignant changes within an
adenoma, and EUS and CT reveal only localized disease, then a pancreaticoduodenectomy should be
performed. If the patient is not a surgical candidate or metastases are present, then palliative
endoscopic interventions, such as placement of a metal stent, should be considered. Palliative surgical
bypass procedures should be reserved for patients with incurable disease associated with duodenal
obstruction.
If histologic assessment of endoscopic biopsies is equivocal for malignancy, the area should be
reexamined, and more biopsies should be obtained, using both standard forceps and snare
techniques, no sooner than 2 weeks after the index procedure. Patients in whom biopsies reveal a
benign tumor without foci of malignancy may be treated by local surgical excision or, in highly selected
patients, endoscopic interventions including snare resection, Nd:YAG laser photoablation, diathermic
fulguration, or snare papillectomy. Should analysis of the intraoperative frozen section biopsies, the
local excision specimen, or biopsies obtained during follow-up surveillance or endoscopic treatment
reveal carcinoma in situ, a pancreaticoduodenectomy should be performed if the patient is a surgical
candidate.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Schlippert W, Lucke D, Anuras S, et al. Carcinoma of the papilla of Vater. A review of

fifty-seven cases. Am J Surg 1978;135:76370.
Walsh DB, Eckhauser FE, Cronenwett JL, et al. Adenocarcinoma of the ampulla of Vater:
Diagnosis and treatment. Ann Surg 1982;195:1527.
Russell DM, Roberts-Thomson IC, Macrea FA, et al. Carcinoma of the papilla of Vater. Aust N
Z J Surg 1982;52:447.
Knox RA, Kingston RD. Carcinoma of the ampulla of Vater. Br J Surg 1986;73:723.
Yamaguchi K, Enjoji M. Carcinoma of the ampulla of Vater. A clinicopathologic study and
pathologic staging of 109 cases of carcinoma and 5 cases of adenoma. Cancer
1987;59:50615.
Robertson JFR, Imrie CW, Hole DJ, et al. Management of periampullary carcinoma. Br J Surg
1987;74:8169.
Neoptolemos JP, Talbot IC, Carr-Locke DL, et al. Treatment and outcome in 52 cases of
ampullary carcinoma. Br J Surg 1987;74:95761.
Hayes DH, Bolton JS, Willis GW, et al. Carcinoma of the ampulla of Vater. Ann Surg
1987;206:5727.
Cooper AW. Carcinoma of the ampulla of Vater. Ann Surg 1937;106:1009.
Miller EM, Dockerty MB, Wollaeger EE, et al. Carcinoma in the region of the papilla of Vater:
A study of cases in which resection was performed. Surg Gynecol Obstet 1951;92:17282.
Hayden WDG, Hightower NC. Primary adenocarcinoma of the ampulla of Vater. South Med J
1958;51:7308.
Howard JM, Jordan GL. Surgical Diseases of the Pancreas. Philadelphia: JB Lippincott. 1960.
Moody F, Thorbjornson B. Carcinoma of the ampulla of Vater. Am J Surg 1964;107:5729.
14. Beall MS, Dyer GA, Stephenson HE. Disappointments in the management of patients with
malignancy of pancreas, duodenum and common bile duct. Arch Surg 1970;101:4615.
15. Sivak MV Jr. Clinical and endoscopic aspects of tumors of the ampulla of Vater. Endoscopy
1988;20:2117.
16. Makipour H, Cooperman A, Danzi JT, et al. Carcinoma of the ampulla of Vater: Review of 38
cases with emphasis on treatment and prognostic factors. Ann Surg 1976;183:3414.
17. Yamaguchi K, Enjoji M, Kitamura K. Non-icteric ampullary carcinoma with a favorable
prognosis. Am J Gastroenterol 1990;85:9949.
18. Yamaguchi K, Enjoji M. Carcinoma of the pancreas: A clinicopathologic study of 96 cases and
immunohistochemical observations. Jpn J Clin Oncol 1989;19:1422.
19. Jones BA, Langer B, Taylor BR, et al. Periampullary tumors: Which ones should be
resected?. Am J Surg 1985;149:4651.
20.
Baczako K, Buchler M, Beger HG, et al. Morphogenesis and possible precursor lesions of
invasive carcinoma of the papilla of Vater: Epithelial dysplasia and adenoma. Hum Pathol
1985;16:30710.
21.
Cattell RB, Pyrtek J. Premalignant lesions of the ampulla of Vater. Surg Gynecol Obstet
1950;90:2130.
22. Gouma DJ, Obertop H, Visman J, et al. Progression of a benign epithelial ampullary tumor to
adenocarcinoma. Surgery 1987;101:5014.
23.
Oh C, Jemerin EE. Benign adenomatous polyps of the papilla of Vater. Surgery
1965;57:495503.
24. Kozuka S, Yamaguchi A, Hachisuka K. Adenomatous residue in cancerous papilla of Vater.
Gut 1981;22:10314.
25. Yamaguchi K, Enjoji M. Adenoma of the ampulla of Vater: Putative precancerous lesion. Gut
1991;32:155861.
26. Neumann RD, LiVolsi VA, Rosenthal NS, et al. Adenocarcinoma in biliary papillomatosis.
27. Perzin KH, Bridge MF. Adenomas of the small intestine. Cancer 1981;48:799819.
28. Seifert E, Schulte F, Stolte M. Adenoma and carcinoma of the duodenum and papilla of Vater:
A clinicopathologic study. Am J Gastroenterol 1992;87:3742.
29. Yao T, Iida M, Watanabe H, et al. Duodenal lesions in familial polyposis of the colon.
30. Ohsato K, Yao T, Watanabe H, et al. Small intestine involvement in familial polyposis
diagnosed by operative intestinal fiberoscopy: A report of four cases. Dis Colon Rectum
1977;20:41420.
31.
Jones TR, Nance FC. Periampullary malignancy in Gardner's syndrome. Ann Surg
1977;185:16573.
32. Iida M, Yao T, Itoh H, et al. Endoscopic features of adenoma of the duodenal papilla in familial
polyposis of the colon. Gastrointest Endosc 1981;27:68.
33. Barton RM, Copeland EM. Carcinoma of the ampulla of Vater. Surg Gynecol Obstet
1983;156:297301.
34. Bakkevold KE, Arnesjo B, Kambestad B. Carcinoma of the pancreas and papilla of
VaterAssessment of resectability and factors influencing resectability in stage I carcinomas.
A prospective multicentre trial in 472 patients. Eur J Surg Oncol 1992;18:494507.
35.
Tasaka K. Carcinoma in the region of the duodenal papilla: A histopathological study.
Fukuoka Acta Med [Japanese, with English. 68:204 1977;abstract].
36. Ponchon T, Berger F, Chavaillon A, et al. Contribution of endoscopy to diagnosis and
treatment of tumors of the ampulla of Vater. Cancer 1989;64:1617.
37. Yamaguchi K, Enjoji M, Kitamura K. Endoscopic biopsy has limited accuracy in diagnosis of
ampullary tumors. Gastrointest Endosc 1990;36:58892.
38.
Safrany L. Duodenoscopy and biopsy: International Workshop. In Classen M, Geenen J,
Kawai K (eds). The Papilla of Vater and Its Diseases. Witzstrock: Baden-Baden, 1979;6671.
39. Huibregtse K, Tytgat GN. Carcinoma of the ampulla of Vater: The endoscopic approach.

40. Komorowski RA, Beggs BK, Geenen JE, et al. Assessment of ampulla of Vater pathology: An
endoscopic approach. Am J Surg Pathol 1991;15:118896.
41. Bourgeois N, Dunham F, Verhest A, Cremer M. Endoscopic biopsies of the papilla of Vater at
the time of endoscopic sphincterotomy: Difficulties in interpretation. Gastrointest Endosc
1984;30:1636.
42. Classen M, Phillip J. Endoscopic retrograde cholangiopancreatography (ERCP) and
endoscopic therapy in pancreatic disease. Clin Gastroenterol 1984;13:81942.
43. Nakao NL, Siegel JH, Stenger RJ, Gelb AM. Tumors of the ampulla of Vater: Early diagnosis
by intraampullary biopsy during endoscopic cannulation. Two case presentations and a review
of the literature. Gastroenterology 1982;83:45964.
44. Silvis SE, Vennes JA, Dreyer M. Variations in the normal duodenal papilla. Gastrointest
Endosc 1983;29:1323.
45. Leese T, Neoptolemos JP, West KP, et al. Tumours and pseudotumours of the region of the
ampulla of Vater: An endoscopic, clinical and pathological study. Gut 1986;27:118692.
46. Chang CP, Ng WW, Tsay SH, et al. Differential diagnosis of tumor-like appearance of the
papilla of Vater. Gastrointest Endosc 1990;36:6356.
47. Piorkowski RJ, Blievernicht SW, Lawrence W, et al. Pancreatic and periampullary carcinoma:
Experience with 200 patients over a 12-year period. Am J Surg 1982;143:18993.
48. Klinkenbijl JHG, van der Schelling GP, Hop WCJ, et al. The advantages of pylorus-preserving
pancreatoduodenectomy in malignant disease of the pancreas and periampullary region. Ann
Surg 1992;216:1425.
49. Grace PA, Pitt HA, Tomkins RK, et al. Decreased morbidity and mortality after
pancreatoduodenectomy. Am J Surg 1986;151:1419.
50.
Cohen JR, Kuchta N, Geller N, et al. Pancreaticoduodenectomy. Ann Surg 1982;95:60817.
51. Monson JRT, Donohue JH, McEntee GP, et al. Radical resection for carcinoma of the
ampulla of Vater. Arch Surg 1991;126:3537.
52. Michellassi F, Erroi F, Dawson PJ, et al. Experience with 647 consecutive tumors of the
duodenum, ampulla, head of the pancreas, and distal common bile duct. Ann Surg
1989;210:54456.
53. Tavazi RY, Hermann RE, Vogt DP, et al. Results of surgical treatment of periampullary
tumors: A thirty-five year experience. Surgery 1986;100:71623.
54. Herter FP, Cooperman AM, Ahlborn TN, et al. Surgical experience with pancreatic and
periampullary cancer. Ann Surg 1982;195:27481.
55. Shutze WP, Sack J, Aldrete JS. Long-term follow-up of 24 patients undergoing radical
resection for ampullary carcinoma, 1953 to 1988. Cancer 1990;66:171720.
56. Yamaguchi K, Nishihara K. Long- and short-term survivors after pancreatoduodenectomy for
ampullary carcinoma. J Surg Oncol 1992;50:195200.
57. van Heerden JA, Judd ES, Doukerty MB. Carcinoma of the extrahepatic bile ducts: A
clinicopathologic study. Am J Surg 1967;113:4956.
58.
Akwari OE, van Heerden JA, Adson MA, et al. Radical pancreatoduodenectomy for cancer of
the papilla of Vater. Arch Surg 1977;1125:4516.
59. Joesting DR, Beart RW, van Heerden JA, et al. Improving survival in adenocarcinoma of the
duodenum. Am J Surg 1981;141:22831.
60. van Heerden JA. Pancreatic resection for carcinoma of the pancreas: Whipple versus total
pancreatectomyAn institutional perspective. World J Surg 1984;8:8808.
61. Fortner JG. Regional resection of cancer of the pancreas: A new surgical approach. Surgery
1973;73:30720.
62.
Halsted WS. Contributions to the surgery of the bile passages, especially of the common bile
duct. Boston Med Surg J 1899;141:6524.
63. Branum GD, Pappas TN, Meyers WC. The management of tumors of the ampulla of Vater by
local resection. Ann Surg 1996;224:6217.
64. Sharp KW, Brandes JL. Local resection of tumors of the ampulla of Vater. Am Surg
1990;56:2147.
65. Farouk M, Niotis M, Branum GD, et al. Indications for and the technique of local resection of
tumors of the papilla of Vater. Arch Surg 1991;126:6502.
66.
Isakson G, Ihse I, Andren-Sandberg A, et al. Local excision for ampullary carcinoma: An
alternative treatment in patients unfit for pancreaticoduodenectomy. Acta Chir Scand
1981;189:12939.
67. Anderson JB, Cooper MJ, Williamson RCN. Adenocarcinoma of the extrahepatic biliary tree.
Ann R Coll Surg Engl 1985;67:13943.
68. Kapur BML. Pancreaticogastrostomy in pancreatoduodenal resection for ampullary cancer:
Experience in 31 cases. Surgery 1986;100:48993.
69. McGuire GE, Pitt HA, Lillemoe KD, et al. Reoperative surgery for periampullary
adenocarcinoma. Arch Surg 1991;126:120512.
70.
Safrany L. Palliative endoscopic therapy of ampullary carcinoma. Gastrointest Endosc
1980;26:77.
71. McPherson GAD, Benjamin IS, Hodgson HJF, et al. Pre-operative percutaneous transhepatic
biliary drainage: The results of a controlled trial. Br J Surg 1984;71:3715.
72.
Alderson D, Lavelle MI, Venables CW. Endoscopic sphincterotomy before
pancreaticoduodenectomy for ampullary carcinoma. BMJ 1981;282:110911.
73. Bickerstaff KI, Berry AR, Chapman RW, Britton BJ. Endoscopic sphincterotomy for the
palliation of ampullary carcinoma. Br J Surg 1990;77:1602.
74. Huibregtse K, Cheng J, Coene PPLO, et al. Endoscopic placement of expandable metal
stents for biliary strictures. A preliminary report on the experience in 33 patients. Endoscopy
1989;21:2802.
75. Domschke W, Foerster E. Endoscopic implantation of large-bore self-expanding biliary mesh
stent. Gastrointest Endosc 1990;36:557.
76. Huibregtse K, Carr-Locke DL, Cremer M, et al. Biliary stent occlusion: A problem solved with
self-expanding metal stents?. Endoscopy 1992;24:3914.
77. Knyrim K, Wagner HJ, Pausch J, et al. A prospective randomized, controlled trial of metal
stents for malignant obstruction of the common bile duct. Endoscopy 1993;25:20712.
78. Fockens P, Waxman I, Davids PH, et al. Early recurrence of obstructive jaundice after
placement of a self-expanding metal endoprosthesis. Endoscopy 1992;24:42830.
79. Cremer M, Deviere J, Sugai B, Baize M. Expandable biliary metal stents for malignancies:
Endoscopic insertion and diathermic cleaning for tumor ingrowth. Gastrointest Endosc
1990;36:4517.
80. Lambert R, Ponchon T, Chavaillon A, Berger F. Laser treatment of tumors of the papilla of
Vater. Endoscopy 1988;20(suppl 1):22731.
81. Siegel JH, Lichtenstein JL, Pullano WE, et al. Treatment of malignant biliary obstruction by
endoscopic implantation of iridium 192 using a new double lumen endoprosthesis.
82.
Edmondson HA. Tumors of the gallbladder and extrahepatic bile ducts. In Atlas of Tumor
Pathology. Ser. 7. Fasc. 26. Washington, DC: Armed Forces Institute of Pathology, 1967;122.
83. Blumgart LH, Kennedy A. Carcinoma of the ampulla of Vater and duodenum. Br J Surg
1973;60:3340.
84. Fortner JG. Regional pancreatectomy for cancer of the pancreas, ampulla, and other related
sites. Ann Surg 1984;199:41825.
85.
Hermanek P, Sobin LH. TNM Classification of Malignant Tumors. 4th ed. Berlin:
Springer-Verlag. 1987.
86. Talbot IC, Neoptolemos JP, Shaw DE, et al. The histopathology and staging of carcinoma of
the ampulla of Vater. Histopathology 1988;12:15565.
87. Mori K, Ikei S, Yamane T, et al. Pathological factors influencing survival of carcinoma of the
ampulla of Vater. Eur J Surg Oncol 1990;16:1838.
88.
Morson BC, Sobin LH. Histological typing of intestinal tumors. In International Histological
Classification of Tumors. No. 15. Geneva: World Health Organization, 1976;47.
89.
Hawes RH, Jamidar PA. Endoscopic diagnosis and treatment of ampullary tumors.
90.
Tio TL, Tytgat GNJ. Endoscopic ultrasonography in staging local resectability of pancreatic
and periampullary malignancy. Scand J Gastroenterol 1986;21:13542.
91. Rosch T, Braig C, Gain T, et al. Staging of pancreatic and ampullary carcinoma by
92. Kamisawa T, Fukayama M, Koike M, et al. Carcinoma of the ampulla of Vater: Expression of
cancer-associated antigens inversely correlated with prognosis. Am J Gastroenterol
1988;83:111823.
93. Neoptolemos JP, Talbot IC, Shaw DC, et al. Long-term survival after resection of ampullary
carcinoma is associated independently with tumor grade and a new staging classification that
assesses local invasiveness. Cancer 1988;61:14037.
94.
Shapiro PF, Litvendahl RA. Tumors of the extrahepatic bile ducts. Ann Surg 1931;94:6179.
95. Sobol S, Cooperman AM. Villous adenoma of the ampulla of Vater: An unusual cause of
biliary colic and obstructive jaundice. Gastroenterology 1978;75:1079.
96. Shabot MM, Asch MJ, Moore TC, et al. Benign obstructing papilloma of the ampulla of Vater
in infancy. Surgery 1975;78:5603.
97. Ryan DP, Schapiro RH, Warshaw AL. Villous tumors of the duodenum. Ann Surg
1986;203:3016.
98. Schulten MF, Oyasu R, Beal JM. Villous adenoma of the duodenum. A case report and review
of the literature. Am J Surg 1976;132:906.
99. Goldberg RI, Hyder SA, Barkin JS, Phillips RS. Intermittent biliary obstruction secondary to a
mobile ampullary adenoma. Gastrointest Endosc 1989;35:5746.
100. Kalmer JA, Merritt CRB. Villous adenoma of the duodenum with intussusception. South Med J
1980;73:6513.
101. Celik C, Venditti JA, Satchindanand S. Villous tumors of the duodenum and ampulla of Vater.
J Surg Oncol 1986;33:26872.
102. Taxier M, Sivak MV, Cooperman A. Villous adenoma of the ampulla of Vater. Gastrointest
Endosc 1979;25:1556.
103. Reynolds R, Lloyd DA, Sweeney JP, et al. Villous adenoma of the duodenum presenting as
acute pancreatitis. J Clin Gastroenterol 1982;4:5336.
104.
Greenwald ES, Parker JG, Schultz S, et al. Benign papillary adenoma of the ampullary region
of the duodenum with intussusception. Gastroenterology 1962;43:34450.
105. Denjoy R. Les tumeurs benignes de la papille de Vater. J Chir (Paris) 1968;95:21126.
106.
Anderson MC, Gregor WH. Papilloma of the ampulla of Vater. Am J Surg 1961;102:86571.
107.
Cattell RB, Braasch JW, Kahn F. Polypoid epithelial tumors of the bile ducts. N Engl J Med
1962;266:5761.
108.
White SH, Nazarian NA, Smith AM, et al. Periampullary adenoma causing pancreatitis. BMJ
1981;283:527.
109.
Malmed L, Levin B. Villous adenomas of the duodenum. AJR 1965;94:3625.
110.
Makkar J, Song KT, Coghill CL. Villous adenoma of the duodenum. Report of a case with
massive hemorrhage. Ann Surg 1969;35:6569.
111. Poilleux J. Benign tumors of the Vater's ampulla region [Abstract]. Med Chir Dig 1975;4:25.
112.
Calzavara C. Ueber adenome des uerdaungskanales. Virchows Arch Pathol Anat Physiol Klin
Med 1895;141:221.
113. Meltzer AD, Ostrum BJ, Isard HJ. Villous tumors of the stomach and duodenum. Radiology
1966;87:5113.
114. Mir-Madjlessi S, Farmer RG, Hawk WA. Villous adenomas of the duodenum and jejunum. Am
J Dig Dis 1973;18:46776.
115. Shah SN, Elibol T, Aguilina JT. Villous adenoma of the duodenum and multiple duodenal
polyps. Gastrointest Endosc 1972;18:17980.
116. Ring EJ, Ferrucci JT, Eaton SB, et al. Villous adenomas of the duodenum. Radiology
1972;104:458.
Barber KW, ReMine WH, Harrison EG, et al. Benign neoplasms of extrahepatic bile ducts,
including papilla of Vater. Arch Surg 1960;81:47984.
118. Shemesh E, Nass S, Czerniak A. Endoscopic sphincterotomy and endoscopic fulguration in
the management of adenoma of the papilla of Vater. Surg Gynecol Obstet 1989;169:4458.
119. Binmoeller KF, Boaventura S, Ramsperger K, Soehendra N. Endoscopic snare excision of
benign adenomas of the papilla of Vater. Gastrointest Endosc 1993;39:12731.
120. Silvis SE. Endoscopic snare papillectomy. Gastrointest Endosc 1993;39:2057.
121. Deucher F, Villiger KJ. Zwei maligne degenerierte duodenal papillome: 10- und 14-Jahres
Heilung nach duodenpankreatektomie. Schweiz Med Wochenschr 1968;98:18225.
122. Serrano JF, McPeak CJ. Primary neoplasms of the duodenum. Surgery 1966;59:199202.
123.
Hancock RJ. An 11-year review of primary tumors of the small bowel including the duodenum.
Can Med Assoc J 1970;21:11779.
124.
Steinberg LS, Shieber W. Villous adenoma of the small intestine. Surgery 1971;72:4238.
125. Anderson A, Iverston O, Ostberg J. Primary tumors of the duodenum. Acta Chir Scand Suppl
1965;343:2836.
126. Weichert R, Reed R, Creech O. Carcinoid-islet cell tumors of the duodenum. Ann Surg
1967;165:6609.
127. Godwin JD. Carcinoid tumors: An analysis of 2837 cases. Cancer 1975;36:5609.
128.
Stamm B, Hedinger CE, Saremaslani P. Duodenal and ampullary carcinoid tumors. Virchows
Arch Pathol Anat Physiol Klin Med 1986;408:47580.
129.
Moreaux J, Martin E, Bloch P. Les tumeurs carcinoides de l'ampoule de Vater. Chir Paris
1971;97:4739.
130. Mergener K, Gottfried MR, Feldman JM, Branch MS. Carcinoid of the ampulla of Vater
presenting as acute pancreatitis. Am J Gastroenterol 1996;91:24267.
131. Vitaux J, Salmon RJ, Languille O, et al. Carcinoid tumor of the common bile duct. Am J
132. Zeitels J, Naunheim K, Kaplan EL, et al. Carcinoid tumors: A 37-year experience. Arch Surg
1982;117:7327.
133. Lee HY, Garber PE. Von Recklinghausen's disease associated with pheochromocytomas and
carcinoid tumor. Ohio State Med J 1970;66:5836.
134. Barber PV. Carcinoid tumor of the ampulla of Vater associated with cutaneous
neurofibromatosis. Postgrad Med J 1976;52:5147.
135. Kapur BML, Savin SK, Anand CS, et al. Carcinoid tumour of ampulla of Vater associated with
viscero-cutaneous neurofibromatosis. Postgrad Med J 1983;59:7345.
136. Hough DR, Chan A, Davidson H. Von Recklinghausen's disease associated with
gastrointestinal carcinoid tumors. Cancer 1983;51:22068.
137. Mahajan SK, Mahajan LA, Malangoni MA, Jain S. Somatostatinoma of the ampulla of Vater.
138. Weichert RF, Roth LM, Krementz ET, et al. Carcinoid-islet cell tumors of the duodenum.
Report of twenty-one cases. Am J Surg 1971;121:195205.
139.
Moertel CG, Sauer WG, Dockerty MB, et al. Life history of the carcinoid tumor of the small
intestine. Cancer 1961;14:90112.
140. McKenna JI, Kozarek RA. Metastatic hypernephroma to the ampulla of Vater: An unusual
cause of malabsorption diagnosed at endoscopic sphincterotomy. Am J Gastroenterol
1989;84:813.
141. Venu RP, Rolny P, Geenen JE, et al. Ampullary tumor caused by metastatic renal cell
carcinoma. Dig Dis Sci 1991;36:3768.
142. Robertson GS, Gertler SL. Late presentation of metastatic renal cell carcinoma as a bleeding
ampullary mass. Gastrointest Endosc 1990;36:3046.
143. Venu RP, Geenen JE. Diagnosis and treatment of diseases of the papilla. Clin Gastroenterol
1986;15:43956.
144. Artaza T, Potenciano JM, Legaz M, et al. Lymphangioma of Vater's ampulla: A rare cause of
117.
obstructive jaundice. Endoscopic therapy. Scand J Gastroenterol 1995;30:8046.

145. Silva R, Paiva ME, Santos CC. Obstructive jaundice caused by ampullary metastases of an
endometrioid adenocarcinoma. Gastrointest Endosc 1996;44:1957.
Chapter 70 Peroral Cholangioscopy and Pancreatoscopy

(7355)
(7356)
MASATSUGU NAKAJIMA, M.D., PH.D.
HIDEKAZU MUKAI, M.D., PH.D.
KEIICHI KAWAI, M.D., PH.D.
Advances in fiberoptic duodenoscopy with cannulation of the main duodenal papilla have greatly
improved the management of disorders of the bile and pancreatic ducts. Endoscopic retrograde
cholangiopancreatography (ERCP) and endoscopic sphincterotomy (EST) are major diagnostic and
therapeutic modalities. Direct endoscopic visualization and instrumentation of the biliary and pancreatic
duct systems by peroral cholangiopancreatoscopy (PCPS) are logical extensions of these techniques.
PCPS has become feasible through the development of very small caliber fiberoptic instruments.
Development of Peroral Cholangiopancreatoscopy

PCPS under duodenoscopic guidance was first reported by Takekoshi and Takagi1(7357) and
Nakamura et al.2(7358) in 1975 and 1976, respectively. These investigators were able to examine the
biliary and pancreatic ducts by means of a "mother" duodenoscope and a "baby" endoscope. The very
thin baby endoscope was designed for passage through the mother instrument. However, the quality of
the images provided by the baby endoscope was poor, and damage to this instrument occurred in the
course of only a few procedures. In 1976, we developed new endoscopic instruments for PCPS in
conjunction with Olympus Optical Co., Ltd. (Tokyo).3,4(7359) Based on the mother-baby concept, this
system provided reasonably good visualization of both the bile and the pancreatic ducts and led to
interesting observations not previously described.5(7360) Using the same instruments, Rsch et
al.6(7361) also performed peroral cholangioscopy (PCS) successfully with this system. However, it was
often impossible to completely examine the ductal system because the prototype baby endoscopes
had neither a tip deflection mechanism nor an effective accessory channel for instrumentation and
irrigation. These major limitations were subsequently resolved by the development of baby endoscopes
with two-way tip deflection and an accessory channel, as well as the use of EST to facilitate the PCPS.
Since the mid-1970s, EST has become an acceptable alternative to surgery for treatment of various
biliopancreatic diseases, especially choledocholithiasis.713(7362) The enlarged orifice of the distal
common bile duct after EST allows passage of a larger-caliber endoscope that has a controllable tip as
well as a satisfactory channel for instrumentation and irrigation. Urakami et al.14(7363) reported a
case of peroral direct biliary endoscopy in which a small-caliber upper gastrointestinal endoscope
(Olympus GIF-P) was used, but only the distal portion of the common bile duct was entered. In 1978,
we developed new instruments for use with a sliding tube guidance technique for PCS after EST and
found this to be an effective approach to the diagnosis and management of biliary tract
diseases.15,16(7364) The sliding tube, which resembles a duodenoscope, has most of the mechanical
features of a side-viewing endoscope, except that it has no optical or light guide bundles. Endoscopic
observation is by means of the optical components of the baby instrument, with the distal tip of this
instrument being positioned at the distal tip of the sliding tube. Kimura et al.17(7365) also reported the
results of their technique for PCS, which employed a specially designed cholangioscope and
balloon-tipped catheter guidance. Although PCS using the sliding tube guidance and balloon-tipped
catheter techniques provided good visualization of the bile ducts and made transendoscopic
instrumentation possible, PCS with these systems was technically difficult and they were not used in
the pancreas.
Efforts to develop better instruments and techniques for PCPS were sustained because of the
limitations of the early systems.18(7366) In 1985, an improved mother-baby system was developed by
Olympus Optical Co., Ltd., based on specifications and recommendations provided by the
authors.19(7367) In this refined system, the mother duodenoscope has a large-diameter instrument
channel that accommodates a larger-caliber baby cholangiopancreatoscope that has a controllable tip
and an effective accessory channel. This instrument system has proved to be considerably easier to
manipulate than earlier versions, and it provides better visualization for intubation, irrigation, and
instrumentation of both the bile and the pancreatic ducts.19,20(7368) This second-generation
mother-baby system has gained wider acceptance for direct endoscopic diagnosis and management of
biliary and pancreatic diseases.2130(7369) As the indications for PCPS have expanded, various
types of endoscopic accessories have also been developed for use in conjunction with these
instruments.26,31(7370)
Instruments
PCPS under duodenoscopic guidance is performed with a duodenoscope (mother endoscope) and a
small-caliber cholangiopancreatoscope (baby endoscope) that is passed through the accessory
channel of the mother instrument. Two standard cold light sources are required. The most improved
instruments for PCPS using the mother-baby approach are shown in Figure 701. The mother
endoscope (TJF-M20), a modified side-viewing duodenoscope, resembles the Olympus standard
duodenoscopes used for ERCP except for the larger diameter of the insertion tube (13.0 mm) and the
larger accessory channel (5.0 mm). The baby endoscope (CHF-B20), designed for passage through
the mother duodenoscope, is a forward-viewing instrument with a working length of 1850 mm and an
external diameter of 4.5 mm. It has a mechanism for two-way tip deflection (90 degrees up and down)
and an instrument channel 2.0 mm in diameter. A variety of accessory instruments are available for
passage through the instrument channel of the baby endoscope (biopsy forceps, cytology brush,
stone-retrieval basket, electrohydraulic lithotripsy probe).
(7371)Figure 701. Improved instrument system for peroral cholangiopancreatoscopy

(PCPS). A, Mother duodenoscope (TJF-M20) for retrograde cannulation. B, Baby
cholangiopancreatoscope (CHF-B20) with a mechanism for two-way tip control and an
effective accessory channel. C, Control section of the baby cholangiopancreatoscope shows
the basket catheter in the accessory channel port. D, Distal tip of the baby endoscope with the
basket catheter inserted through the accessory channel.
Various types of mother duodenoscopes and baby cholangiopancreatoscopes, with or without a tip
control or an accessory channel, have been designed by Olympus Optical Co., Ltd., for specific
diagnostic and therapeutic purposes (Table 701). The larger-diameter CHF-B20, CHF-B20QY, and
CHF-B34Y baby endoscopes are commonly used after EST for PCS, or for peroral pancreatoscopy
(PPS) if the main pancreatic duct is dilated. Transendoscopic instrumentation of the ducts is possible
with these instruments because they have relatively large accessory channel diameters. The
smaller-caliber baby endoscopes (CHF-B27, PF-23MN) are employed for PCPS in patients with an
intact papilla. The smallest-diameter baby instrument (PF-8P), which is usually inserted through a
cannula, is used for PPS of the entire main pancreatic duct when the duct is not dilated. A new
deflectable catheter system with a shape memory alloy (SMA) has been specially designed for
manipulation and control of this ultrathin baby endoscope (Figure 702).31(7372)
CHF-B34Y baby endoscopes are commonly used after EST for PCS, or for peroral pancreatoscopy
(PPS) if the main pancreatic duct is dilated. Transendoscopic instrumentation of the ducts is possible
with these instruments because they have relatively large accessory channel diameters. The
smaller-caliber baby endoscopes (CHF-B27, PF-23MN) are employed for PCPS in patients with an
intact papilla. The smallest-diameter baby instrument (PF-8P), which is usually inserted through a
cannula, is used for PPS of the entire main pancreatic duct when the duct is not dilated. A new
deflectable catheter system with a shape memory alloy (SMA) has been specially designed for
manipulation and control of this ultrathin baby endoscope (Figure 702).31(7372)
Specifications of Instruments for PCPS by Means of the Mother-Baby
TABLE 701
Scope System
MOTHER
SCOPES
TJF-M20
OUTER
DIAMETER
(mm)
ACCESSORY
CHANNEL
(mm)
13.0
5.5
BABY
SCOPES
BENDING
SECTION
(n)
OUTER
DIAMETER
(mm)
ACCESSORY
CHANNEL
(mm)
CHF-B20
CHF-B20QY
2(u, d)
4 (u, d, r, l)
4.5
4.5
1.7
1.2
3.4
2.7
2.3
0.8
1.2
1.2
TJF-20
12.5
4.2
CHF-B34Y
2 (u, d)
TJF-200*
12.0
4.2
CHF-B27
2 (u, d)
JT-1T20
11.0
3.2
PF-23MN
JF-200*
11.0
3.2
PF-8P
PCPSperoral cholangiopancreatoscopy; uup; ddown; rright; lleft.

* Electronic video endoscopes.
(7373)Figure 702. Shape memory alloy (SMA) catheter system for PCPS using an ultrathin
baby scope. A, Generator and control box. B, SMA catheter (external diameter 2.2 mm;
internal diameter 1.2 mm). C, Ultrathin PF-8P baby endoscope (external diameter 0.8 mm;
without controllable tip and instrument channel). D, Bent tip of SMA catheter through which
the ultrathin baby endoscope has been passed.
Technique
PCPS with the mother-baby system is performed in a room equipped for fluoroscopy. The procedure is
performed after ERCP with or without EST. Two endoscopists are required: One operator manipulates
the mother duodenoscope for identification of the main duodenal papilla and insertion of the baby
endoscope into the biliary and pancreatic ducts. The second operator controls the baby endoscope for
visualization and instrumentation of the ducts. The endoscopic images of the duodenum and ducts can
be viewed on separate television monitors by means of video attachments.
Most aspects of the technique are similar to those of ERCP. With the patient under topical pharyngeal
anesthesia and intravenous premedication for sedation and duodenal atony, the mother duodenoscope
is introduced into the descending duodenum. After the main duodenal papilla has been identified (and
EST has been performed in some cases), the tip of the baby endoscope is guided from the mother
instrument into either the bile duct or the pancreatic duct. It is then carefully and gently advanced
toward the liver or the tail of the pancreas under endoscopic and fluoroscopic guidance (Figures 703,
704, 705 and 706). The entire ductal system can be visualized with the aid of tip deflection and the
automatic irrigation system of the baby endoscope. If necessary, transendoscopic instrumentation of
the ducts can be performed under direct vision using various accessory instruments. With the use of
the smallest baby endoscope (PF-8P) in patients with an intact papilla and without dilation of the ducts,
especially in the pancreas, the SMA catheter system is employed for angulation and deeper insertion
of the instrument (Figure 707).
(7374)Figure 703. Radiographs depict PCPS with a mother-baby system. A, Baby endoscope
(CHF-B20) in the left hepatic duct (arrow). B, Baby endoscope in the main pancreatic duct in
the body of the pancreas (arrow).
(7375)Figure 704. Duodenoscopic views of PCPS corresponding to Figure 703. A, Main

duodenal papilla before endoscopic sphincterotomy (EST). B, Papilla after EST. C, Baby
endoscope (CHF-B20) inserted into the ductal system through the incised papilla.
(7376)Figure 705. Endoscopic view of biliary tract by peroral cholangioscopy (PCS)

corresponding to Figure 703A. A, Lumen of the common bile duct. B, Intrahepatic ducts at
the bifurcation.
(7377)Figure 706. Endoscopy of main pancreatic duct by peroral pancreatoscopy (PPS)

corresponding to Figure 703B. A, Lumen of the main pancreatic duct. B, Orifice of the
branch duct recognized as a pinhole.
(7378)Figure 707. PPS with SMA catheter system. A, Endoscopic retrograde pancreatogram
demonstrates normal ductal diameter and a radiolucent shadow (arrow) in the tail of the
pancreas. B, Ultrathin baby endoscope (PF-8P) within SMA catheter (arrow) is introduced
into the distal pancreas. The patient has not undergone prior EST. C, Duodenoscopic view of
the normal main duodenal papilla. D, Duodenoscopic view of the SMA catheter inserted into
the papilla without prior EST. E, PPS view demonstrates absence of ductal abnormalities in
the tail of the pancreas.
Results
During the 18 years from 1976 to 1993, we attempted PCPS in 329 patients: 80 patients with the first
mother-baby system, 64 patients using the sliding tube system, and 185 patients with the second
mother-baby endoscope system (Table 702).
TABLE 702
PROCEDURE
Total Results of PCPS

FIRST-GENERATION
MOTHER-BABY
SYSTEM (19761985)
SLIDING TUBE
SYSTEM
(19781985)
SECOND-GENERATION
MOTHER-BABY
SYSTEM (19851993)
Cholangioscopy
57 (48)
64 (47)
107 (96)
Cholelithiasis
38 (35)
49 (37)
41 (35)
Biliary tumors
9 (5)
7 (5)
37 (34)
Others
10 (3)
8 (5)
29 (27)
TABLE 702
PROCEDURE
Total Results of PCPS

FIRST-GENERATION
MOTHER-BABY
SYSTEM (19761985)
SLIDING TUBE
SYSTEM
(19781985)
SECOND-GENERATION
MOTHER-BABY
SYSTEM (19851993)
Pancreatoscopy
23 (18)
0
Pancreatolithiasis
8 (5)
0
Pancreatic tumors
8 (6)
0
Others
7 (7)
0
Total
80 (66)
64 (47)
PCPSperoral cholangiopancreatoscopy; ( )no. of adequate visualizations.
78 (68)
6 (5)
36 (30)
36 (33)
185 (164)
PCS with the first mother-baby system was successful in 48 of 57 patients (84.2%) and PPS was
possible in 18 of 23 patients (78.3%). The procedure was completed in many patients with an intact
papilla, as well as in some patients with a fistula at the choledochoduodenal junction that had been
created either surgically or by EST. The brightness of the images provided by the baby instruments
was adequate for examination, and the lumen of both the bile and the pancreatic ducts was inspected
without the use of an irrigation system. However, the procedure was limited to visualization of the ducts
and lesions because the baby endoscopes available during this earlier developmental period did not
have effective accessory channels.
PCS using the sliding tube guidance system was successful in 47 of 65 patients (73.4%) who had
previously undergone EST. Although this system was technically difficult to use, it provided adequate
visualization of the ducts as well as effective instrumentation within the biliary tract because the baby
instrument offered controllable tip deflection and an accessory channel.
PCS with the newer mother-baby system was successful in 96 of 107 patients (89.7%) and PPS was
completed in 68 of 78 patients (87.2%), including patients with or without previous EST. Adequate
visualization and instrumentation of the duct systems, including the intrahepatic ducts, gallbladder, and
tail of the pancreas along with the common bile duct and the main duct in the head of the pancreas,
were easily accomplished with this system and a variety of baby instruments. In this series of
examinations, the baby endoscopes were found to be durable despite their small diameters, and there
were no clinically significant complications during or after the procedures.
Peroral Cholangioscopy
The normal bile duct mucous membrane appears yellow when seen through bile. The bifurcation of the
common hepatic duct is recognized as a carina that is similar in appearance to the tracheobronchial
bifurcation, and the hepatic ducts divide into two or three segments (see Figure 705). Stones within
the duct are clearly identified and can be removed, after EST, with a basket catheter under direct
vision. Even huge stones impacted within the duct that cannot be removed by mechanical lithotripsy
can be fragmented under direct vision using an electrohydraulic lithotripsy probe and removed in
pieces (Figure 708). After endoscopic removal of stones, usually no additional abnormalities of the
bile duct are noted endoscopically except for some erosions or redness of the mucosa suggestive of
chronic inflammation. Intrahepatic stones located near the hepatic hilus can also be visualized clearly
and removed under direct visual control by employing a grasping forceps or basket catheter with or
without the use of electrohydraulic lithotripsy. Biliary endoscopy is also effective for differential or
histologic diagnosis in patients with primary tumors. The endoscopic appearance of biliary cancer is
often an abrupt stricture or obstruction of the duct with a reddish and uneven mucosa (Figure 709).
However, early-stage tumors sometimes appear as small polypoid lesions or as an irregular reddish
area within the duct wall. Endoscopic inspection and procurement of biopsies under direct vision will
confirm the diagnosis (Figures 7010 and 7011). The gallbladder can also be inspected by peroral
cholecystoscopy and biopsy specimens obtained for histologic diagnosis (Figure 7012).
(7379)Figure 708. Peroral cholangioscopic lithotripsy in a patient with a huge stone

impacted in the biliary tract. A, Endoscopic retrograde cholangiogram shows the stone (arrow)
at the hepatic hilus. B, Electrohydraulic lithotripsy under direct vision. Arrow shows the stone.
C, Retrograde cholangiogram demonstrates complete removal of the stone. D, Endoscopic
view of the huge stone in the duct. E, Endoscopic view during lithotripsy. F, Lumen of the
biliary tract after removal of the stone.
(7380)Figure 709. PCS in a patient with biliary tract cancer. A, Retrograde cholangiogram
shows an irregular stricture (arrow) of the common duct. B, Baby endoscope (CHF-B20) is
inserted to the lesion for inspection and to obtain biopsy specimens under direct vision. C,
Cholangioscopic view of the lesion. Cancer of the common bile duct was histologically
confirmed in cholangioscopic biopsy specimens.
(7381)Figure 7010. PCS in a patient with a small polypoid cancer of the common bile duct.
A, Retrograde cholangiogram demonstrates a round shadow (arrow) within the dilated
common bile duct suggestive of a tumor or a stone. B, Cholangiogram after EST shows the
cholangioscope approaching the lesion. C, Cholangioscopic view of the polypoid tumor. D,
Cholangioscopic view shows the reddish vascular surface of the tumor. Carcinoma of the
common bile duct was histologically confirmed in biopsies taken under direct vision.
(7382)Figure 7011. A, Resection specimen from the patient in Figure 7010. A 6- 13-mm
polypoid tumor (arrowhead) is present within the common bile duct. B, Photomicrograph of
the specimen shows well-differentiated adenocarcinoma limited to the muscular layer of the
common bile duct.
(7383)Figure 7012. Peroral cholecystoscopy (PCCS) in a patient with a polypoid tumor of

the gallbladder. A, Retrograde cholangiogram shows the introduction of a guidewire into the
gallbladder. Arrow indicates the polypoid lesion within the gallbladder. B, Cholangiogram
shows the introduction of the baby endoscope (CHF-B34) into the gallbladder for inspection
of the lesion (arrow). C, Endoscopic view within the gallbladder reveals the polypoid tumor.
Biopsies obtained under direct vision revealed a hyperplastic polyp.
Peroral Pancreatoscopy
The normal appearance of the main pancreatic duct at PPS is as follows: the mucous membrane
appears pale pink when seen through pancreatic juice and has a delicate submucosal vascular
reticulum (see Figure 706A); orifices of the main duct branches can sometimes be recognized as
pinhole openings (see Figure 706B). In patients with pancreatic calcification, pancreatic stones are
white or whitish-yellow. Direct endoscopic removal of stones is possible in some cases after EST of the
main pancreatic duct, with or without prior extracorporeal shockwave lithotripsy (Figure 7013).
Pancreatoscopy in patients with tumors reveals an abrupt stricture or obstruction of the main duct with
an irregularly reddish and uneven granular mass or a small villous or papillary polypoid lesion within the
duct (Figure 7014). Direct endoscopy of the pancreas with optional procurement of biopsy or cytology
specimens is of value in the early detection and differential diagnosis of tumors within the main
pancreatic duct (Figure 7015). In patients with a dilated main pancreatic duct who have undergone
EST, the entire duct is easily inspected at PPS and tissue samples can be obtained without difficulty
when the procedure is performed with a baby endoscope that has tip deflection control and an
accessory channel. However, use of the larger-diameter baby endoscopes is difficult in patients without
ductal dilation. In such cases, use of an ultrathin-caliber baby endoscope within the SMA catheter
system is more effective (Figures 7016 and 7017).
(7384)Figure 7013. PPS in a patient with pancreatolithiasis. A, Endoscopic retrograde

pancreatogram demonstrates several stones (arrowheads) within the main duct in the head of
the pancreas. EST was performed. B, Pancreatogram after removal of the stones with a
pancreatoscope within the main duct. C, PPS view of white-yellow stones within the main
pancreatic duct.
(7385)Figure 7014. PPS in a patient with intraductal papillary carcinoma of the pancreas. A,
Endoscopic retrograde pancreatogram shows a polypoid shadow (arrow) within the first
branch of the main duct in the head of the pancreas. B, Pancreatogram shows the insertion of
the baby endoscope to the lesion. EST was performed to facilitate insertion. C, PPS shows the
polypoid tumor within the duct, which was histologically confirmed as malignant in biopsy
specimens taken under direct vision. D, Resection specimen with the papillary tumor
(arrowhead) within the dilated duct. E, Photomicrograph reveals intraductal papillary
carcinoma in an adenoma.
(7386)Figure 7015. PPS in patient with tiny carcinoma of the pancreas. A, Retrograde
pancreatogram demonstrates two tiny radiolucent shadows (arrowheads) within the main duct
in the head of the pancreas. B, Radiograph shows the insertion of the baby endoscope after
EST. C, PPS shows the small villous tumor within the main duct. D, Cytology specimen from
the lesions shows malignant changes.
(7387)Figure 7016. PPS with SMA catheter system in a patient with an intact papilla and
pancreatic cancer. A, Retrograde pancreatogram demonstrates obstruction (arrow) of the main
pancreatic duct. B, Radiograph shows the insertion of the SMA catheter with the baby
endoscope. C, Pancreatoscopic view of the reddish, irregular mucosa at the point of
obstruction of the main duct.
(7388)Figure 7017. PPS with SMA catheter system in a patient with chronic pancreatitis. A,
Retrograde pancreatogram demonstrates a stricture (arrow) of the main pancreatic duct. B,
Radiograph shows the insertion of the SMA catheter with the baby endoscope. C,
Pancreatoscopic view of the stricture shows the smooth narrowing of the main duct.
Comment
Direct observation of any phenomenon is always preferable to the use of indirect methods. The ability
to visualize the lumen of the bile and pancreatic ducts results in a level of diagnostic precision that is
not attainable by other means. Intraoperative biliary endoscopy through a choledochotomy incision and
postoperative choledochoscopy through a T-tube tract have become valuable diagnostic tools in biliary
tract diseases, especially for management of biliary neoplasms as well as identification and retrieval of
retained stones (see Chapter 71: Choledochofiberoscopy and Endoscopic Lithotripsy).3234(7389)
Intraoperative pancreatic endoscopy has also been found to be valuable in a few cases.35,36(7390)
There are in general two approaches to direct endoscopic evaluation and therapy in the bile duct:
percutaneous and peroral transpapillary.20(7391) The desire for less invasive methods for diagnosis
and management of disorders of the bile and pancreatic ductal systems in conjunction with major
advances in the design of small-caliber fiberoptic instruments has facilitated the development of
PCPS.16(7392) Percutaneous transhepatic cholangioscopy and cholecystoscopy can provide
excellent results in the diagnosis and management of biliary diseases.37,38(7393) However, these
procedures entail a certain amount of patient discomfort and the preparation and maintenance of the
access/drainage tract requires time and effort. By contrast, PCPS via the peroral transpapillary
approach can be performed quickly and easily by an endoscopist experienced in ERCP with relatively
little patient discomfort.18,20,26(7394)
Despite early enthusiasm and reasonably favorable results with the first mother-baby PCPS systems,
this method did not gain wide acceptance because of technical limitations, especially the lack of tip
deflection and an effective accessory channel in the prototype instruments.35(7395) Nevertheless, it
seemed that refinements in instrumentation and technique would increase the usefulness of this
procedure. The sliding tube system provided a cholangioscope with the desired features.14(7396)
However, prior EST was required and PCPS using this system proved to be technically difficult,
although it provided excellent visualization and transendoscopic instrumentation of the biliary
tract.15,18(7397) These limitations have been largely overcome with development of a
second-generation mother-baby instrument system in which the baby instrument is inserted under
duodenoscopic guidance.19,20(7398) In our experience, this is the most satisfactory system for PCS
and PPS because it offers maneuverability, a high rate of success, and a wide range of instrumentation
of the ductal systems. It is now employed for inspection and instrumentation of the gallbladder and
proximal pancreas as well as the common bile duct and distal pancreas.2931(7399)
Before the development of PCPS, preoperative diagnosis of bile and pancreatic diseases depended on
various radiographic procedures. The radiographic features of the duct systems are easily and clearly
defined with a variety of direct approaches, including ERCP and percutaneous transhepatic
cholangiography. These special procedures permit a high degree of confidence in the diagnosis and
management of diseases of the bile and pancreatic ducts. However, there are still problems in
interpretation of cholangiograms and pancreatograms. These usually relate to the uncertain nature of
filling defects, ductal deformity, obstruction to the flow of contrast medium, poor-quality radiographs, or
technically inadequate studies.3740(7400) Therefore, the precise observation possible with PCS and
PPS can be valuable in numerous ways.1831,4145(7401) The nature of doubtful, suspicious, and
uncertain radiographic findings can be resolved. Stones within the bile duct or main pancreatic duct
can be visualized and removed under visual control. Tumors within the ducts can be clearly identified
and biopsies can be obtained under direct vision. Diagnosis of biliary tract and pancreatic tumors at an
early stage is possible, sometimes as a result of investigation of findings that appear to be minimum by
conventional methods.3,18(7402)
Conclusion
PCS and PPS have been developed for noninvasive, direct endoscopic examination of the bile and
pancreatic duct systems. The atraumatic nature of PCPS is demonstrated by the absence of injury to
the ducts and the low incidence of complications. The second-generation mother-baby endoscope
system for PCPS has now gained wider acceptance by endoscopists with special expertise in ERCP.
Although this system still has certain mechanical limitations, these instruments provide satisfactory
visualization and instrumentation of the bile and pancreatic ducts. With expected improvements in
instrument design, PCPS promises to become a reliable method for diagnosis and management of a
variety of biliary and pancreatic diseases.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Takekoshi T, Takagi K. Retrograde pancreatocholangioscopy. Gastroenterol Endosc (Jpn)

1975;17:67883.
Nakamura M, Toki F, Ohi I, et al. The research of transduodenoscopic
pancreatocholangioscopy. Gastroenterol Endosc (Jpn) 1976;18:1528.
Kawai K, Nakajima M, Akasaka Y, et al. Eine neue endoskopische Technik: Die perorale
Choledocho-Pankreatoscopie. Leber Magen Darm 1976;6:1214.
Nakajima M, Akasaka Y, Fukumoto K, et al. Peroral cholangiopancreatoscopy (PCPS) under
duodenoscopic guidance. Am J Gastroenterol 1976;66:2417.
Nakajima M, Akasaka Y, Yamaguchi K, et al. Direct endoscopic visualization of the bile and
pancreatic duct systems by peroral cholangiopancreatoscopy (PCPS). Gastrointest Endosc
1978;24:1415.
Rsch W, Koch H, Demling L. Peroral cholangioscopy. Endoscopy 1976;8:1725.
Kawai K, Akasaka Y, Hashimoto Y, et al. Preliminary report on endoscopical papillotomy. J
Kyoto Pref Univ Med 1973;82:3535.
Kawai K, Akasaka Y, Murakami K, et al. Endoscopic sphincterotomy of the ampulla of Vater.
aus dem Ductus Choledochus. Dtsch med Wochenschr 1974;99:4967.
Nakajima M, Kimoto K, Fukumoto K, et al. Endoscopic sphincterotomy of the ampulla of Vater
and removal of common duct stones. Am J Gastroenterol 1975;64:3443.
Safrany L. Endoscopic treatment of biliary-tract diseases. An international study. Lancet
1978;2:9835.
Demling L, Classen M. Endoscopic Sphincterotomy of the Papilla of VaterInternational
Workshop, Munich. Stuttgart: Georg Thieme. 1978.
Nakajima M, Kizu M, Akasaka Y, Kawai K. Five years' experience of endoscopic
sphincterotomy in Japan: A collective study from 25 centres. Endoscopy 1979;11:13841.
Urakami Y, Seifert E, Butke H. Peroral direct cholangioscopy (PCS) using routine
straight-view endoscope, first report. Endoscopy 1977;9:2730.
Nakajima M, Kizu M, Akasaka Y, et al. Peroral cholangioscopy (PCS) under sliding tube
guidanceA preliminary report of the instruments and techniques. In Classen M, Geenen J,
Kawai K (eds). International WorkshopThe Papilla Vateri and Its Diseases. Baden-Baden:
Gerhard Witzstrock, 1979;99102.
Nakajima M, Fujimoto S. Peroral cholangioscopy (PCS) under sliding tube guidance. In Kawai
K (ed). Frontiers of GI Endoscopy. Tokyo: Olympus Optical Co., 1982;438.
Kimura K, Sakai H, Yoshida Y, et al. Development of a new peroral cholangioscope
(FDS-CP). Am J Gastroenterol 1982;77:5804.
Nakajima M, Fujimoto S, Kawai K. Peroral cholangioscopy and pancreatoscopy. In Sivak MV
Jr (ed). Gastroenterologic Endoscopy. 1st ed. Philadelphia: WB Saunders, 1987;23146.
Yoshida S, Kimoto K, Nakajima M, et al. Progress of peroral cholangioscopy. J Biliary Tract
Pancreas (Jpn) 1989;10:498.
20. Yasuda K, Nakajima M, Cho E, et al. Comparison of peroral and percutaneous

cholangioscopy. Endoscopy 1989;21:34750.
21. Liguory CL, Lefebvie JF, Bonnel D, et al. Indications for cholangioscopy. Endoscopy
1989;21:3413.
22.
Rieman JF, Kohler B, Harloff M, et al. Peroral cholangioscopy: An improved method in the
diagnosis of common bile duct diseases. Endoscopy 1989;21:4357.
23. Leung JWC, Chung SSC. Electrohydraulic lithotripsy with peroral choledochoscopy. BMJ
1989;299:5958.
24. Cotton PB, Kozarek RA, Schapiro RH, et al. Endoscopic laser lithotripsy of large bile duct
stones. Gastroenterology 1990;99:112833.
25. Ponchon T, Gagnon P, Valette PJ, et al. Pulsed dye laser lithotripsy of bile duct stones.
26.
Nakajima M, Yasuda K, Cho E, et al. Early diagnosis of biliary tract cancer by peroral
cholangioscopy (PCS). J Biliary Tract Pancreas (Jpn) 1991;12:38795.
27. Neuhaus H. Cholangioscopy. Endoscopy 1992;24:12532.
28. Neuhaus H, Hoffmann W, Classen M. Laser lithotripsy of pancreatic and biliary stones via
3.4-mm and 3.7-mm miniscopes: First clinical results. Endoscopy 1992;24:20814.
29. Fujita R, Hirata N, Fujita Y. Peroral cholecystoscopy. Endoscopy 1989;21:37880.
30. Foerster EC, Schneider MU, Matek W, Domschke W. Transpapillary cholecystoscopy.
31. Mizuno S, Nakajima M, Yasuda K, et al. Shape memory alloy (SMA) catheter system for
peroral pancreatoscopy using an ultra thin caliber endoscope. Endoscopy 1994;26:67680.
32. Shore JM, Berci G, Morgenstern L. The value of biliary endoscopy. Surg Gynecol Obstet
1975;140:6014.
33. Tompkins RK, Johnson J, Strome FK, et al. Operative endoscopy of biliary tract neoplasms.
Am J Surg 1976;132:17482.
34. Yamakawa T, Mieno K, Noguchi T, et al. An improved choledocho-fiberscope and
non-surgical removal of retained biliary calculi under vision control. Gastrointest Endosc
1976;22:1607.
35.
Nakamura M, Hamano K, Endo H, et al. Endoscopic study of pancreatic diseases.
Gastroenterol Endosc (Jpn) 1971;13:4268.
36.
Classen M, Schwemmule K, Demling L. Endoscopy of the pancreatic duct. Endoscopy
1972;4:2213.
37. Siegel JH, Mayer LF. Percutaneous choledochoscopy and cholecystoscopy: Diagnostic and
therapeutic uses. Endoscopy 1981;13:1247.
38. Nimura Y, Shionoya S, Hayakawa N, et al. Value of percutaneous transhepatic
cholangioscopy (PTCS). Surg Endosc 1988;2:2139.
39. Block MA, Schuman BM, Weckstein ML. Interpretative problems in endoscopic retrograde
cholangiopancreatography. Am J Surg 1975;129:2933.
40.
Kittredge RD, Baer JW. Percutaneous transhepatic cholangiography: Problems in
interpretation. AJR Am J Roentgenol 1975;125:3546.
41. Caldwell SH, Oelsner DH, Bickston SJ, et al. Intrahepatic biliary endoscopy in sclerosing
cholangitis. J Clin Gastroenterol 1996;23:1526.
42. Ozkan H, Saisho H, Yamaguchi T, et al. Clinical usefulness of a new miniscope in the
diagnosis of pancreatic disease. Gastrointest Endosc 1995;42:4805.
43. Sander R, Poesl H. Initial experience with a new babyscope for endoscopic retrograde
cholangiopancreaticoscopy. Gastrointest Endosc 1996;44:1914.
44. Sackmann M, Weinzierl M, Holl J. Minicholangioscopy during routine endoscopic retrograde
cholangiography. Endoscopy 1995;27:280.
45. Soda K, Yamanaka T, Yoshida Y, et al. A newly developed fine-caliber endoscope for peroral
cholangiopancreatoscopy. Endoscopy 1994;26:6715.
Chapter 71 Choledochofiberoscopy and Endoscopic Lithotripsy

(7403)
(7404)
KOJI GOCHO, M.D.
The technique of nonsurgical removal of postoperative residual bile duct stones using fiberoptic
choledochoscopes and retrieval baskets improved the treatment of choledocholithiasis. This was
laborious when surgical removal was the only available technique, especially when stones were lodged
within the intrahepatic ducts. Moreover, surgery has been associated with a discouragingly high
incidence of residual stones. To address the latter problem, intraoperative cholangiography and
choledochoscopy were developed along with a variety of stone-grasping forceps, retrieval baskets, and
balloon catheters for the capture and removal of stones. Although it is usually possible to remove
stones by means of these devices, there are still many cases in which these techniques do not provide
a definitive solution.
The introduction of fiberoptic technology into medical practice promoted the development of
small-caliber fiberscopes. The first fiberoptic choledochoscope was introduced in 1965.1(7405) The
attribute of flexibility was found to be of value during surgery, and new routes into the various bile duct
segments, both before and after surgery, were devised for the choledochoscope.25(7406)
Development of various ancillary devices such as stone retrieval baskets also accelerated the use of
percutaneous transhepatic cholangioscopy and postoperative choledochoscopy for both diagnostic and
therapeutic purposes.613(7407) These made it possible to remove residual bile duct stones in most
cases and significantly improved overall therapeutic efficacy in the treatment of choledocholithiasis.
Nevertheless, large or impacted stones remained problematic, and this in turn prompted the
development of effective methods for disintegration of stones. Initial attempts at disintegration did not
have favorable results because of a lack of suitable technology when the choledochoscope was first
introduced. Technologic advances subsequently led to techniques that are both safe and effective,
including the application of a laser in 198114(7408) and choledochoscopic electrohydraulic shockwave
lithotripsy (ESWL) in 1985.15(7409) These new modalities have been used successfully in the
treatment of stones not amenable to conventional surgical techniques.
A number of methods are used for treatment of choledocholithiasis via a T-tube or T-tube tract
including dissolution therapy of various types and extraction under fluoroscopic control using a
steerable catheter as introduced by Burhenne.16(7410) The latter method is used frequently in many
countries with good success in experienced hands.1719(7411) However, this method also has some
disadvantages. Radiation exposure can be considerable, especially if there are numerous stones and
the procedure is prolonged or must be repeated. Calculi smaller than 5 mm may be difficult to define
radiographically because they may be obscured as increasing amounts of contrast medium in the
ducts produce denser opacification. It may also be difficult to remove impacted intrahepatic stones by
the method of Burhenne, and it is sometimes necessary to allow a period of time for these to pass into
the more distal intrahepatic or extrahepatic ducts. Some technical problems are common to both the
endoscopic and the fluoroscopically guided methods of stone extraction. For example, some stones
may be too large to be withdrawn through the existing T-tube tract. The choice between endoscopic
and fluoroscopically guided stone extraction depends on the clinical and anatomic circumstances and
expertise in one or both procedures. Among different methods currently used to remove residual
stones in the common bile duct, I consider endoscopic extraction under direct visual control to be the
least traumatic to the ductal epithelium.
The development and widespread utilization of lap-aroscopic cholecystectomy has changed and will
continue to alter the fundamental roles of intraoperative cholangiography, endoscopic retrograde
cholangiopancreatography (ERCP), and choledochoscopy in the management of patients with
choledocholithiasis (see also Chapter 56: Indications, Contraindications, and Complications of
Diagnostic Endoscopic Retrograde Cholangiopancreatography).20(7412) In effect, laparoscopic
cholecystectomy eliminates the possibility of an approach route to the bile duct by which
choledochoscopy can be performed should stones be discovered in the bile duct during or after
removal of the gallbladder. There is, therefore, a greater reliance on endoscopic methods of treatment
in conjunction with laparoscopic cholecystectomy. Laparoscopic choledochoscopy via the cystic duct
has been performed using small-diameter fiberoptic instruments.21(7413) However, this approach
currently affords little or no possibility for removing stones.
Approaches to the Bile Ducts

Unlike the rigid choledochoscope, which can be used only during an open surgical
procedure,22,23(7414) the flexible fiberoptic choledochoscope can reach stones in the bile duct via
several different approach routes (Figure 711). These can be either surgical or nonsurgical (minimally
invasive).4,6,2436(7415) The minimally invasive routes include a percutaneous transhepatic tract
after percutaneous transhepatic biliary drainage (PTBD),2,3,8,3741(7416) T-tube
tract,5,12,42,43(7417) jejunostomy tract,44,45(7418) and cholecystostomy tract.4648(7419) Peroral
cholangioscopy, another route, is discussed in Chapter 70: Peroral Cholangioscopy and
Pancreatoscopy.
(7420)Figure 711. Diagram of endoscopic approaches to the biliary system before, during,
and after surgery.
Preoperative and Nonoperative Choledochoscopy

I consider this approach to be less invasive than surgery. Repeated examinations are possible, and the
procedures are particularly useful in aged and high-risk patients who are not suitable candidates for
surgery.
Percutaneous Transhepatic Cholangioscopy
Percutaneous transhepatic cholangiography (PTC) and PTBD under fluoroscopic guidance must be
performed before percutaneous transhepatic cholangioscopy.49(7421) Then the intrahepatic tract is
dilated using a fibrous tract dilator, and a 6- to 7-mm-diameter tube is inserted into the intrahepatic bile
duct and left in place. After approximately 3 weeks, a solid intrahepatic fibrous tract will form, and a
flexible fiberoptic choledochoscope can then be inserted via this tract into the intrahepatic bile duct and
distally to the common bile duct (Figures 712 and 713).
(7422)Figure 712. Percutaneous transhepatic cholangioscopy.
(7423)Figure 713. Endoscopic view of the retained stone demonstrated in Figure 712.
Percutaneous transhepatic cholangioscopy can be performed in patients with dilated intrahepatic bile
ducts and intrahepatic duct calculi or in those in whom obstruction of the extrahepatic bile duct distal to
retained calculi precludes extraction via endoscopic sphincterotomy. Biopsies may also be obtained of
obstructing lesions such as bile duct carcinoma, and selective cholangiography can be performed.
Percutaneous Transhepatic Cholecystoscopy
Percutaneous transhepatic drainage of the gallbladder is first performed under ultrasonography. The
fibrous tract is prepared in the same manner as for percutaneous transhepatic cholangioscopy. This
secures the approach for routing the fiberoptic choledochoscope into the gallbladder (Figure
714).38,40,41,50(7424) With this method of cholecystoscopy, biopsies may be obtained of the

gallbladder under direct vision, and stones may be removed.
(7425)Figure 714. Percutaneous transhepatic cholecystoscopy.
Operative Choledochoscopy
A choledochoscope can be inserted directly into the bile duct after opening the common duct during
laparotomy. Operative choledochoscopy is usually performed as a method of exploration of the bile
ducts in patients with dilated ducts due to choledocholithiasis to remove stones or to confirm the
absence of residual stones. Both flexible fiberoptic and rigid choledochoscopes are available. One
advantage of operative choledochoscopy is that any abnormalities discovered endoscopically can be
dealt with during the operation. However, the technique also adds to the length of the operative
procedure.
Postoperative Choledochoscopy
In this method, an approach route to the bile duct is secured during surgery, and endoscopic
examination of the bile duct is performed at some point after operation.27,29,3436,42,46,51(7426)
One major advantage of postoperative choledochoscopy is that although the time of each examination
may be limited, the procedure may be repeated many times using the fibrous tract.9,28,52(7427)
T-Tube Tract Choledochoscopy
A T-tube left in place in the common bile duct at lap-arotomy provides access to the biliary system.
After a period of time, the tract becomes solid and fibrotic. Various authors differ as to how much time
is required for formation of a solid tract. I introduce a flexible fiberoptic choledochoscope after about 3
weeks (Figure 715).5,12,43(7428) This method makes possible the endoscopic exploration of the
distal common bile duct as well as the intrahepatic ducts. It is the simplest and most commonly used of
the several postoperative choledochoscopy procedures. The technical details are discussed later.
(7429)Figure 715. T-tube tract choledochoscopy.

Jejunostomy Tract Choledochoscopy
This technique for postoperative choledochoscopy requires that a jejunostomy tract be constructed at
the time of laparotomy. A choledochojejunostomy is performed, and part of the jejunum is fixed to the
parietal peritoneum.45,5355(7430) A flexible fiberoptic choledochoscope is introduced into the bile
duct through this jejunostomy tract after the operation (Figure 716). Jejunostomy tract
choledochoscopy is especially useful when residual intrahepatic stones are expected to be a problem
after operation or recurrence of intrahepatic gallstones is anticipated.32,56(7431) Endoscopic
examination of the bile duct may be repeated long after the tract to the jejunostomy is closed because
this may easily be reopened by making a small incision in the skin.54(7432)
(7433)Figure 716. Jejunostomy tract choledochoscopy. The section of the fiberscope within
the jejunal loop is indicated by the span between the arrows.
Cystic Duct Choledochoscopy
After cystic duct drainage or cholecystostomy, the cystic duct can be dilated and a flexible fiberoptic
choledochoscope introduced into the common bile duct through this tract (Figures 717 and
718).47,48(7434) The distal common bile duct can be reached easily by means of this procedure, but
the common hepatic duct is not accessible.
(7435)Figure 717. Cystic duct choledochoscopy after cholecystostomy. The section of the
fiberscope within the cystic duct is indicated by the span between the arrows.
(7436)Figure 718. Cystic duct choledochoscopy after cystic duct drainage.
Instruments and Equipment

Fiberoptic Choledochoscopes
As with any flexible endoscope, adequate flexibility of the insertion tube and range of motion at the
bending section of the fiberoptic choledochoscope are essential prerequisites for exploration of the bile
ducts. Unlike most other hollow organ systems accessible to endoscopy, the biliary system is full of
substances, such as bile and sludge, that can obscure the visual field and therefore require active
evacuation through the endoscope. The multipurpose choledochoscope, suitable for both diagnostic
and therapeutic applications, should be equipped with two independent channels: an
irrigation/evacuation channel for high-pressure infusion of an irrigation solution and aspiration of fluid
and particles and an operating channel that is suitable for insertion of various devices such as
lithotriptor probes, grasping/retrieval forceps, and basket catheters (Figure 719). Excessive irrigation
of the bile ducts with saline solution during choledochoscopy increases the internal pressure within the
ductal system and causes discomfort to patients. This may also result in diarrhea and vomiting after
the examination. Therefore, instruments must have an aspiration mechanism by which excess saline
solution can be removed from the system.
(7437)Figure 719. A choledochoscope with a 4.9-mm outer diameter. Note the dual
irrigation/evacuation and operating channels (inset). A 3-French electrohydraulic lithotriptor
electrode has been passed through the operating channel, but there is still adequate room for
evacuation of fluid and particles (inset). Because the lithotriptor is flexible, it is possible to
deflect the tip of the instrument 180 degrees.
Accessories
A variety of accessory instruments are needed to achieve maximum results during choledochoscopy.
The essential items are a stone-retrieval basket catheter (see later) and biopsy forceps. Other optional
accessories that increase therapeutic capability and aid in surmounting technical problems are a
flushing tube (catheter for washing), alligator forceps, three-pronged grasping forceps, biopsy forceps
with stylet, and guidewires (Figure 7110).
(7438)Figure 7110. Optional accessories. A, Flushing tube. B, Alligator forceps. C,

Three-armed forceps. D, Loop-type stone-crushing forceps. E, Basket-type stone-crushing
forceps. F, Basket-type stone-crushing forceps. G, Biopsy forceps with stylet. H, Cutting knife
for use with high-frequency electrosurgical generator. I, Guidewire.
The epithelium of the biliary tract is harder and more resistant to endoscopic "pinch-type" biopsy than
other mucosal surfaces of the gastrointestinal tract. Therefore, a biopsy forceps with a bayonet or stylet
within the forceps cups is useful for targeting and fixing the forceps in position to obtain a specimen
(see Figure 7110). When thick fluid, sludge, and floating debris are present in the bile ducts, a
flushing tube can be used to wash the material from the endoscopic field. Hydrostatic balloon dilators
may be useful for dilating strictures or tortuous branches of the bile ducts. The fibrous tract through
which the choledochoscope must pass to enter the biliary system may become narrowed, in which
case either rigid or flexible dilators may be used to open the channel.
After stone disintegration, the larger fragments that cannot be evacuated by suctioning must be
retrieved using a stone basket catheter. The maximum diameter of the stone fragments that can be
retrieved in this way is limited by the diameter or the condition of the approach tract. If the size of the
fragment is too large, it is often difficult or impossible to remove, even if it can be grasped or captured
in the basket. It is best, therefore, to use a basket with a diameter when opened that is equal to that of
the approach route. If the fragment is too large to be captured by this basket, further disintegration is
indicated.
Two characteristics are indispensable for the stone-holding basket catheter: it should permit easy
capture of a stone, and it should not allow the stone to slip from within the basket during extraction.
Catheters are available with baskets of various sizes, and the basket should be selected according to
the size, characteristics, position of the stone(s) to be extracted, and diameter of the approach route.
Modalities for Stone Disintegration

Choledochoscopic lithotripsy requires equipment for stone disintegration and instruments for retrieval
of disintegrated stone fragments in addition to a fiberoptic choledochoscope. With suitable equipment,
almost any bile duct stone can be removed. The availability of lightweight, compact instrument systems
permits treatment in virtually any patient care area within a hospital.
Several different types of stone disintegrators are available for various clinical applications. For
endoscopic applications, the device selected should have the following prerequisites: (1) the
throughput portion of the lithotriptor probe should be of a diameter that is compatible with the operating
channel of the endoscope and should have adequate flexibility so that the tip deflection mechanism of
the endoscope is not restricted; (2) a power output that is high enough to disintegrate stones but not so
high as to damage adjacent tissues or the endoscope; and (3) portability.
Various lithotripsy methods have been developed (Table 711). Each modality has specific
characteristics, advantages, and disadvantages.
TABLE 711
Modalities for Stone Disintegration and Their Characteristics
MECHANICAL
Disintegration
mechanism
Mechanical
force
LASER
Thermal energy of
light
ELECTROHYDRAULIC
MICROWAVE
Electrohydraulic shock waves
Thermal energy
(internally
induced)
Ultrasonic vib
TABLE 711
Disintegration
process
MECHANICAL
LASER
ELECTROHYDRAULIC
MICROWAVE
Alligator jaws:
biting off
small bits at a
time
Intensive heat by
laser irradiation
vaporizes stone,
causing divots and
cracks
Disintegration begins at stone

surface; powerful and rapid
disintegration with scattering
of fragments
Electrode
inserted
inside a soft
stone; heat
generated by
microwave;
boiling water
content
vaporizes and
cracks
stones;
destruction
localized to
electrode site
Mechanical vi
Three-prong:
grasping and
crushing with
claws
Noncontact: aiming
beam indicates
spot to be
irradiated; distal tip
kept 5 mm from
stone surface;
power, 5070
watts
Contact:
bullet-shaped
ceramic tip directly
coupled to stone
surface; power, 15
watts
Bilirubinate and
cholesterol stones;
freely moving
stones difficult to
target
Any type of stone; both

floating and impacted
stones can be disintegrated
For soft
bilirubinate
stones only;
not adequate
for hard or
floating
stones
Thermal energy
may be
transmitted to
surrounding
tissue
causing
potential
burns
Relatively sof
Basket:
basketing
and crushing
with wires
Type of
stones and
effects
Effect on the
adjacent
tissues
Soft bilirubinate
stones; good
for freely
moving
stones; not
effective for
hard stones
Mechanical
injury may
occur if not
used under
direct vision
Keep 5-mm distance

from tissue;
misalignment may
cause bleeding or
ductal wall injury
Keep 3-mm distance from

tissue; misalignment may
cause ductal wall injury or
bleeding
May perforate other

side of the stone if
the stone is
completely
penetrated
Less damage
TABLE 711
Size and
nature of
electrodes/
probes
Equipment
Effect on the
deflection of
choledochoscope
Limitation for
approach
routes
Remarks
MECHANICAL
LASER
ELECTROHYDRAULIC
MICROWAVE
Shaft diameter:
2.4 to 3.6
mm; metal
sheath is less
flexible
No equipment
needed
O.D. 1.6 and 2.1

mm; quartz fiber is
flexible but
somewhat hard
O.D. 1 mm (3 French); woven

electric wire is flexible; distal
electrode may wear
O.D. 1.7 to 2.4

mm; reusable
electrode
O.D. 4 mm (1
Nd:YAG laser unit

heavy and not
mobile
Relatively limited
deflection
Small and light-weight

generator; easy to carry
Microtaze unit
easy to carry
Ultrasonic lith
Full deflection
Relatively
limited
deflection
Not compatib
Unsuitable for
markedly curved
routes
Can be used for
other surgical
applications
No limitation
Unsuitable for
markedly
curved routes
Can be used
for other
surgical
applications
Straight route
Limited
deflection
Acceptable for
slightly
curved routes
Basket may
break
Dedicated lithotriptor; must

follow precautions
O.D.outer diameter; Nd:YAGneodymium:yttrium-aluminum-garnet.

Devices for Mechanical Lithotripsy
Modified biopsy forceps, stone-grasping forceps, and stone-crushing basket catheters are simple in
principle and design. This type of device is inserted through the operating channel of the
choledochoscope for grasping and crushing stones under direct vision.7,5761(7439) Devices for
mechanical lithotripsy (see Figure 7110) include the following:
1. "Alligator" forceps. This instrument is adequate for crushing immobile and fragile bilirubinate
stones. The stone is gradually reduced in size by repeated removal of small pieces.
2. Three-pronged forceps. Fragile, floating stones can be grasped and crushed with the prongs or
claws of this device.
3. Stone-crushing basket catheter. Stones are grasped and crushed by constricting the basket wires.
This technique is effective for fragile, floating stones but not useful for impacted or hard stones.
Mechanical lithotripsy devices have a number of advantages. These instruments are (1) simple in
design and relatively easy to use; (2) inexpensive and do not require additional equipment for use; (3)
effective for stones that do not require extensive manipulation for extraction; and (4) safe when used
under direct vision.
These devices also have disadvantages. They are (1) slightly more rigid than other accessories and
therefore impede motion of the bending section of the choledochoscope; (2) relatively larger in
diameter and not compatible with smaller-caliber endoscopes; and (3) difficult to manipulate if the bile
duct is tortuous.
Lasers
Several types of lasers have been used for lithotripsy of bile duct stones, including
neodymium:yttrium-aluminum-garnet (Nd:YAG)14,6265(7440) and tunable, pulsed dye
Dedicated lith
lasers.6672(7441)
The Nd:YAG laser generates electromagnetic energy at a frequency that is not absorbed by water (see
Chapter 10: Laser Physics and Laser-Tissue Interactions). The beam from the laser unit is transmitted
through a flexible quartz optical fiber (waveguide). The beam as it emerges from the end of the fiber
may be aimed at the stone (noncontact method), or the fiber may be placed directly against the stone
(contact method). Disintegration is performed under direct vision. The newer pulsed dye laser, which
emits intermittent pulses, is now used more commonly for laser lithotripsy.
In the noncontact method of Nd:YAG laser lithotripsy, the distal tip of the waveguide fiber is placed at a
distance of 0.5 to 1.0 cm from the surface of the stone. The laser energy at a power output of 70 watts
is repeatedly directed at the stone in pulses of 0.5 to 1.0 sec. Laser irradiation will generate divots or
cracks along the surface of the stone that make the stone fragile and amenable to mechanical
crushing using alligator forceps or similar devices as outlined previously. With the contact method, a
bullet-shaped ceramic head is attached to the distal tip of the quartz fiber waveguide. The ceramic
head is placed in direct contact with the stone surface. The power output of the laser must be reduced
for contact laser lithotripsy. A power output of 15 watts is used.
Laser lithotripsy is both powerful and fast with regard to its ability to disintegrate stones. Its
shortcomings are:
1. Deflection of the tip of the choledochoscope is limited owing to the relative stiffness of the quartz
waveguide.
2. There is a possibility of ductal wall perforation on penetration of a stone.
3. Laser machines are relatively large and heavy and are not very portable for practical purposes.
Electrohydraulic Lithotriptor
A high-voltage electric discharge within a fluid medium generates a hydraulic shockwave that is of
sufficient energy to fragment or disintegrate gallstones.7376(7442) For this method of lithotripsy to be
effective and safe, precise alignment between the electrode and the stone is required. Electrohydraulic
lithotriptors have been in clinical use for many years, mostly for urologic applications. For biliary
applications, this technology was originally adapted for use through a duodenoscope.7781(7443)
Electrode catheters were also designed for percutaneous use.82,83(7444) Disintegration was
performed under fluoroscopic guidance. However, fluoroscopy is not entirely satisfactory for alignment
of the distal end of the electrode against the stone. With improper alignment, there is always the
possibility of perforation of the bile duct wall as a result of either mechanical penetration or the burn
that occurs when an electrode comes in direct contact with tissue. Use of a small-diameter electrode
through a flexible choledochoscope makes it possible to approach various areas of the bile duct and to
disintegrate stones under direct vision.15,8490(7445) These refinements in design and methodology
make this approach both safe and effective.
Evaluation of various stone disintegration methods for intrahepatic duct, common bile duct, and
gallbladder stones indicates that ESWL is the most useful. To perform ESWL effectively and safely,
several precautions must be observed, which are summarized in Table 712.
Precautions for
Electrohydraulic Lithotripsy
TABLE 712
GENERATOR
Ground the generator power supply
Select the adequate power setting
ELECTRODE
Do not overuse the electrode
Do not discharge in air or within the endoscope
Precautions for
Electrohydraulic Lithotripsy
TABLE 712
Electrode tip should protrude from endoscope by 3 mm or

more
DISINTEGRATION PROCESS
Electrode tip can contact stone
Electrode tip must be kept at least 3 mm away from tissue
Do not discharge for long periods of time
Use repeated short-duration discharges
Do not discharge when field of view is obscured
OTHER CAUTIONS
Do not ground the patient
Insulate the operator's hands (surgical gloves are
adequate)
Contraindicated in patients with implanted cardiac
pacemakers
ESWL has a number of distinct advantages:

1.
2.
3.
4.
It is a very powerful method for disintegration of stones.

It is effective for treatment of impacted as well as floating stones.
The electrode shaft is very flexible and does not interfere with tip deflection of choledochoscopes.
Because of these features, it is useful for the treatment of intrahepatic duct stones where fairly
marked deflection of the instrument tip is frequently necessary.
5. The system is relatively small and lightweight and can be used almost anywhere (Figure 7111).
(7446)Figure 7111. Electrohydraulic lithotriptor and accessories.

This method of lithotripsy has some disadvantages:
1. The electrode wears out with repeated use and is therefore manufactured as a disposable device,
which adds to the cost of the procedure.
2. Damage to the ductal wall may occur if the electrode tip is extremely close to it when fired.
3. The system cannot be used for any purpose other than lithotripsy.
Microwave Coagulator
This system, the Microtaze unit, was originally designed for coagulation of the liver during
hepatectomy.91(7447) It utilizes microwave energy generated at the tip of an antenna-like electrode
that has to be inserted into tissue. For biliary lithotripsy, heat generated by the microwave lithotriptor
seems to promote the disintegration process.9294(7448) Because the electrode must be inserted
within the stone, the system can be used only for immobile bilirubinate stones.
Microwave lithotripsy has the following advantages:
1. The risk of tissue damage is minimal.
2. The electrode does not wear readily and is reusable.
3. The system can be used for other surgical purposes.
Its value is reduced by the following disadvantages:
1. The ability to disintegrate stones is less than that for electrohydraulic and laser systems.
2. It cannot be used for cholesterol stones.
3. It is unsuitable for floating stones.

Ultrasonic Lithotriptor
The ultrasonic lithotriptor, designed for urinary tract stone lithotripsy through the ureteroscope,
cystoscope, or nephroscope, has also been used to disintegrate bile duct stones.9597(7449)
Ultrasonic energy, generated by a piezoelectric transducer, is transmitted through a hollow metal tube
that is pressed against the stone. Fragments of the stone are removed by transmission of the
vibrational energy. Continuous irrigation is required to remove the heat generated by ultrasonic
vibration. Disintegrated particles together with irrigation solution must be aspirated through an internal
lumen.
Conduction of the ultrasonic vibrational energy requires a specific acoustic resonance, and therefore,
the probe must be made of metal with a design that is solid and straight. Application of this technology
to flexible endoscopic lithotripsy is thus limited and not practical.
Where a straight (rigid) choledochoscope can be used, ultrasonic lithotripsy may have some
advantages, including (1) simultaneous disintegration and suction, which allows for rapid evacuation of
stone particles after disintegration and (2) less tissue damage. Conversely, this method of lithotripsy
also has some disadvantages: (1) the probes are usually straight and rigid and can be used only with
rigid endoscopes (e.g., nephroscope, ureteroscope); and (2) a straight route of approach is required,
and thus, this method has very limited application in the bile duct.
Technique of Postoperative Choledochofiberoscopy Via T-Tube Tract

A T-tube tract of sufficient diameter is required for passage of the choledochoscope into the bile duct.
Generally, this means that the portion of the T-tube that exits to the skin must have a diameter of 18
French. The tract produced by a tube of this diameter will allow a fiberscope with an insertion tube
diameter of 6 mm to pass without difficulty after the tract has become fibrotic over the course of about
3 weeks.
It is preferable to perform flexible choledochoscopy under fluoroscopic control because it is useful to
know the position of the instrument tip as well as the stone or stones within the bile duct. X-ray and
endoscopic views are complementary and increase the reliability and thoroughness of the procedure,
so that residual stones and other lesions are less likely to be overlooked when both methods of
observation are used.
Premedication and anesthesia are not necessary because percutaneous choledochoscopy is seldom
associated with discomfort. However, sedatives may be administered parenterally if necessary.
Atropine is given intramuscularly before the examination.
The skin around the fibrous tract is sterilized in the same manner as for any operative procedure and
the patient, except for the site of the fibrous tract, is covered with a sterile drape.
Cholangiography is then performed by injecting a 30% solution of an iodinated contrast agent through
the T-tube. The use of more concentrated solution (e.g., 60%) will result in dense opacification that
may obscure small stones. Cholangiography provides a clear representation of the configuration of the
bile ducts, residual stones, and any areas of narrowing or stricture formation. The contrast medium
must be well mixed with the bile, so that all areas of the biliary system are opacified including the more
peripheral intrahepatic ducts.
After satisfactory radiographs of the entire biliary system are obtained, the contrast material is washed
out of the duct with saline solution, and then the T-tube is carefully withdrawn.
The fiberscope is inserted under direct vision, observing the inner wall of the fibrous tract as the
fiberscope is passed toward and into the bile duct. During insertion, saline solution is injected through
the irrigation channel of the fiberscope to maintain patency of the fibrous tract and bile duct and to
provide a good visual field by eliminating bile and floating substances that obscure the view. One of the
major merits of the choledochofiberscope is that it provides an excellent visual field. This allows the
endoscopist to proceed using direct vision at all times. The fiberscope must be inserted carefully. The
tip angulation mechanism should be used to maintain a satisfactory visual field. The endoscope should
never be inserted blindly or forcefully. Abnormal endoscopic findings can be documented by
endoscopic photography.
Once floating stones are discovered and their sizes estimated, a basket forceps of comparable size is
introduced through the accessory channel of the fiberscope. With the basket in the closed position, it is
passed into the bile duct beyond the stone. Then the basket is opened and drawn back toward the
stone. Generally, stones can be easily caught with the basket. It is then closed slowly until the stone is
encountered, and then a firm, steady force is maintained. The basket must not be closed with too
much force because this will crush the stone, and it will then be more difficult and time consuming to
remove the fragments. Once the stone is secure in the basket, the choledochoscope is removed slowly
from the bile duct and through the T-tube tract.
If the T-tube is left in place after the procedure to provide future access, it must be fixed securely to the
skin to prevent dislodgment. If this should happen, the fibrous tract will close in a few days. It is
possible to insert a thin tube through the tract within 24 hr of dislodgment. This can be followed by
dilation using a fibrous tract dilator, so that eventually the T-tube may be reinserted.
Once there is confirmation that the bile duct is free of stones, the T-tube can be removed permanently.
Bile leakage may be avoided by changing the T-tube to one with a smaller diameter before permanent
removal. Thereafter, the fibrous tract will close completely within a few days.
Technique of Stone Disintegration and Retrieval

Small floating stones within the bile duct may be easily grasped and retrieved using one of the various
grasping devices. Stones that are larger than the internal diameter of the approach route or impacted
within the bile duct cannot be removed en bloc and must be disintegrated into smaller fragments.
Floating stones may be crushed using a stone-crushing forceps; an impacted stone may be crushed
with an alligator or three-pronged forceps. Those that are too hard to crush with a mechanical device
are disintegrated, preferably by ESWL (Figure 7112). The field of view is cleared by irrigating and
evacuating the floating particles, which allow for better visualization of the stone and its relation to the
surrounding structures. The lithotriptor electrode is introduced into the field through the operating
channel of the endoscope toward the stone to be disintegrated. Disintegration is performed, observing
the precautions outlined in Table 712. Disintegrated fragments can be removed or evacuated.
(7450)Figure 7112. Diagram of choledochoscopic lithotripsy. EHLelectrohydraulic

lithotripsy. (Courtesy of Northgate Research Corporation, Plattsburgh, New York.)
After lithotripsy of a stone or stones, the flushing tube should be used to wash the duct. Otherwise, fine
particles and debris from the crushed stone will obscure the visual field. Fragments of stones must be
removed with a stone-holding forceps or basket. When there are many stones or stone fragments,
crushing or disintegration and extraction must be repeated as necessary until the bile duct has been
completely cleared. The necessary maneuvers may be performed repeatedly for as long as 1 to 2 hr. If
the duct cannot be cleared during this time, the procedure should be repeated at another time. To
facilitate subsequent procedures, the patency of the fibrotic T-tube tract (or other approach route) must
be maintained by reinsertion of the T-tube. This is relatively easy to accomplish. Damage to the bile
duct wall by the tip of the tube occurs less often with a T-tube than with a straight tube.
After reinsertion of the T-tube, small stones and debris should be washed from the bile duct by
introducing saline solution through the reinserted T-tube. A final cholangiogram should be obtained to
demonstrate either that no stones are present or the number and position of any residual stones.
Indications
Stones anywhere in the biliary system including the common hepatic and bile ducts, intrahepatic ducts,
and gallbladder may be an indication for choledochoscopic extraction or lithotripsy.9,98(7451) In
patients with stones located within the right and left lobes of the liver, a problem that cannot be readily
managed by an open surgical procedure, choledochoscopic lithotripsy can be the technique of choice.
Frequently, it is the only practical method of treatment.
Residual stones in the common ducts and intrahepatic ducts found by postoperative T-tube tract
cholangiography, stones causing obstructive jaundice where PTBD has been performed, and stones in
the gallbladder where percutaneous transhepatic gallbladder drainage or percutaneous gallbladder
drainage has been performed can be readily approached with choledochoscopes through the available
routes. As long as an approach tract of adequate diameter can be established, choledochoscopic
lithotripsy can be attempted in these clinical circumstances.
Stones or fragments of stones larger than the diameter of the tract (or the endoscope) or stones
located beyond a stricture of the bile duct can be disintegrated by lithotripsy to smaller, retrievable
sizes. Any stone that can be observed by the choledochoscope can be removed by a combination of
various modalities.
Choledocholithiasis
When choledocholithiasis has been diagnosed pre-operatively, the combined use of conventional
surgical exploration and operative choledochoscopy reduces the frequency of postoperative retained
bile duct stones.11,23,30,99(7452) However, small stones may stray into the intrahepatic ducts.
Impacted stones may require a considerable amount of time for removal by operative
choledochoscopy. If these problems excessively increase operation time or if there is a possibility of
residual stones, then biliary drainage should be established. This also provides endoscopic access
during the postoperative period. Stones can then be removed safely by postoperative choledochoscopy
(Figure 7113).
(7453)Figure 7113. Cholangiogram demonstrates the choledochofiberscope in relation to a

retained stone (arrow) in the distal common bile duct.
Birkett and Williams36(7454) performed choledochoscopy via a T-tube tract in 28 patients. Fifty-nine
stones were extracted in 22 patients. In 6 patients, choledochoscopy resolved the nature of filling
defects noted by T-tube cholangiography that proved to be due to findings other than calculi. The
endoscopic findings in 3 of these patients were multiple papillary adenocarcinomas, malignant
stricture, and a benign polyp. No abnormalities were found at endoscopy in a 4th patient, and in the last
2 cases the radiographic abnormalities were explained by the presence of normal mucosal folds.
Yamakawa et al.51(7455) performed 808 choledochoscopies via a T-tube tract in 292 patients. One
hundred and four patients had bile duct stones; the bile ducts were cleared of stones in 101 cases. The
average number of sessions for removal of stones from the extrahepatic bile ducts was 1.5. If the
stones were in the intrahepatic ducts, from 3 to 58 sessions were necessary, with a mean of 15.4.
Failure of stone extraction was attributed to improper insertion of the T-tube; a long, narrow, or
tortuous T-tube tract; giant stones; and stones lodged in the peripheral intrahepatic ducts.
Intrahepatic Stones
Intrahepatic bile duct stones are encountered more frequently in some areas of the worldAsia, for
examplethan others. Unilateral intrahepatic stones have been treated by left hepatic lobectomy.
Bilateral intrahepatic stones can be even more troublesome because they frequently cannot be
completely removed by conventional surgical stone extraction procedures. These technical problems
and more drastic methods of treatment are now obviated by operative and nonoperative stone
extraction procedures. The use of a flexible fiberoptic choledochoscope makes it possible to
completely remove intrahepatic stones in the majority of cases, and choledochoscopy is considered by
numerous authors to be a satisfactory method for the treatment of intrahepatic
stones.9,25,34,5456,100,101(7456)
When intrahepatic stones are discovered preoperatively, cholangiography may not reveal the entire
biliary system because the intrahepatic ducts are often packed with stones and sludge. The degree
and extent of stone impaction may not be demonstrated because the contrast medium does not
penetrate beyond the impacted stones and debris. In such cases, preoperative choledochofiberoscopy
can be performed to eliminate some stones, debris, and turbid bile (Figure 7114). Then, endoscopic
selective cholangiography can be performed to outline the entire intrahepatic bile duct system. The
operative approach can then be planned based on accurate radiographic assessment.32,100(7457)
(7458)Figure 7114. Cholangiogram demonstrates retained stones (arrows) in the left hepatic
bile duct.
Hwang et al.34(7459) treated intrahepatic stones in 30 patients with a 7-mm-diameter fiberoptic
choledochoscope via T-tube tracts. Complete extraction of all stones was achieved in 25 cases (83%).
The average number of sessions required was 4.3, with a range of 1 to 15. Chen et al.102(7460)
performed endoscopic extraction of calculi via T-tube tracts in 66 cases. In 48 of these, the stones
were located in the intrahepatic ducts, and the authors successfully extracted these in 38 cases.
Murata et al.32(7461) extracted intrahepatic stones via T-tube tracts in six of nine patients. In three
other patients, these authors extracted intrahepatic calculi with the aid of percutaneous transhepatic
cholangioscopy. The average number of procedures for these three patients was 5.3. Murata et
al.32(7462) noted that the percutaneous approach had advantages in some circumstances and
considered it to be indicated when surgery was precluded by dense adhesions as a result of multiple
prior operations and when other methods of extractionvia a T-tube tract, for examplehad failed.
The latter circumstance may be due to impaction of stones beyond a stricture. The procedure may also
be considered when technical problems such as a prior partial gastrectomy with a Billroth II
anastomosis or Roux-en-Y anastomosis renders endoscopic sphincterotomy and extraction of stones
difficult or impossible.39(7463)
Ponchon et al.103(7464) performed percutaneous choledochoscopy in 123 patients (161 procedures).
The access route was transhepatic in 96 patients and via a T-tube tract in 27 cases. Gallstones were
extracted in 75 cases, intrahepatic stones being present in 35 patients. The bile ducts were cleared in
92% of cases, with lithotripsy being necessary in 85%. Stones recurred in 18% at a median follow-up of
32 months. Biopsies obtained from bile duct stenoses in 52 patients had a sensitivity of 78% for the
diagnosis of malignancy. In 2 of 3 patients with intraductal adenomas laser photoablation was
successful, but in 8 patients with cholangiocarcinoma the duration of duct patency after treatment was
too short for this form of treatment to be considered successful. Yeh et al.104(7465) performed PTC
and lithotripsy and stone removal in 165 patients. Complete clearance of calculi was accomplished in
132 (80%). The rate of stone recurrence was 32.6% at a mean follow-up of 58 months.
Jan and Chen105(7466) achieved complete clearance of intrahepatic stones via percutaneous
transhepatic cholangioscopy in 40 of 48 patients (83.3%). During follow-up, which ranged from 4 to 10
years, 14 patients (35%) had symptomatic recurrent stones, 2 patients (5%) developed recurrent
stones but remained asymptomatic, and 2 patients developed cholangitis but stones were not found.
The presence of a bile duct stricture was the most common reason for failure to clear the ducts, and
the rate of recurrent stone formation was highest in patients with strictures.
When intrahepatic stones are discovered intraoperatively, removal of as many as possible should be
attempted at surgery. However, if the stones are numerous and impacted into many of the peripheral
intrahepatic radicles, removal of all stones may necessitate excessive operating time. If complete
removal is precluded because of the length of surgery, or if there is uncertainty that the biliary system
has been cleared of stones, then an access route to the biliary system should be established so that
postoperative choledochofiberoscopy can be performed.
T-tube choledochofiberoscopy can be performed when intrahepatic stones have been discovered
postoperatively by T-tube cholangiography (Figures 7115 and 7116). As stones are removed, the
status of the intrahepatic ducts and any remaining stones can be defined by repeated T-tube
cholangiography.
(7467)Figure 7115. Choledochoscopic approach via T-tube tract to retained stones (arrows)
in the right hepatic bile ducts.
(7468)Figure 7116. Endoscopic view of the impacted stone demonstrated by

cholangiography in Figure 7115.
If postoperative recurrence of intrahepatic stones is expected, a provision can be made at surgery for
jejunostomy tract choledochoscopy by performing a choledochojejunostomy and then fixing a portion of
the jejunum near the biliary enteric anastomosis to the parietal peritoneum. This provides an access
route to the intrahepatic ducts via the jejunostomy tract that can be used in the event of a recurrence of
intrahepatic stones long after closure of the choledochojejunostomy.54,106,107(7469) In the series of
29 patients of Ker et al.107(7470) who underwent choledochojejunostomy with subcutaneous
jejunostomy, 7 patients developed cholangitis more than 5 years later, including 4 with recurrent
stones. In the latter patients, the fistula into the jejunal loop was easily established under local
anesthesia and a choledochoscope was inserted for removal of stones.
Contraindications and Complications

There are no absolute contraindications to choledochofiberoscopy except for a bile duct that is too
small in diameter to introduce a choledochofiberscope.
Relatively few complications have been repor-ted.36,51,108111(7471) In the series of Birkett and
Williams36(7472) of 28 patients treated by T-tube choledochoscopy, transient fever developed in 5
patients and pancreatitis developed in 2. Of the 218 procedures performed by Chen et al.,102(7473)
there were 8 subsequent episodes of abdominal pain, 6 of fever and chills, and 2 of bleeding from the
fistula tract. A duodenocutaneous fistula developed in 1 patient. Yamakawa et al.51(7474) performed
808 endoscopic procedures via T-tube tracts in 292 patients. Limited extravasation of contrast medium
occurred in 5 patients, 3 of whom had retained stones. Free extravasation of contrast from the T-tube
tract occurred in 4 patients with formation of a subhepatic collection of bile in 2 patients, a subphrenic
bile collection in 1, and free intraabdominal extravasation in 1. Fistula formation between the T-tube
tract and the duodenum occurred in 1 patient. In the report of 30 cases by Hwang et al.,34(7475) fever
and chills developed in 6 patients, and a pyogenic liver abscess occurred in 1. Moss et al.42(7476)
performed T-tube choledochoscopy in 17 patients (19 procedures) and encountered 1 episode of
postprocedure bacteremia and 1 of bacteremia after dilation of the T-tube tract.
Chen and Jan112(7477) demonstrated bacteremia during postoperative choledochofiberoscopy in 15
of 100 patients. Blood cultures were positive at 5 min in 7 patients, at 15 min in 8 patients, and at hr
in 2 patients; in 2 cases, cultures at two different time intervals were positive. There was no correlation
between bacteremia and the presence of residual stones or the duration of the procedure. Six of the
patients with positive blood cultures developed clinical evidence of cholangitis within 24 hr of the
procedure. Cholangitis resolved in all cases with antibiotic therapy. All bacteria cultured were aerobes,
with the most common species being Enterococcus, Escherichia coli, and Klebsiella.
All the lithotripsy modalities were originally designed to destroy target objects by various mechanisms
and thus can potentially damage adjacent tissue or structures when used intracorporeally.
Complications can also occur during establishment of the various approach tracts, manipulation of
endoscopes, and retrieval of stones and stone fragments through the tract. These include injury to the
bile duct wall, perforation, and hemorrhage.
Injury can be avoided if a clear endoscopic field of view is maintained, so that all manipulations of
stones are performed under direct endoscopic observation. A choledochoscope with both irrigation and
suction capability is therefore very useful and effective. Fluoroscopic observation during
choledochoscope insertion and stone retrieval can avoid inadvertent damage to tissues and organs en
route. A careful, discrete, step-by-step approach with stringent monitoring and observation is essential
for safe and effective choledochoscopic lithotripsy.
Conclusions
Choledochoscopic lithotripsy was originally used to remove postoperative residual stones without
repetitive surgical interventions. The application of this method has been expanded to include
treatment of stones not amenable to traditional surgical procedures alone. Development of various
minimally invasive approaching routes has made this method a primary therapeutic modality for stones
in virtually any part of the biliary system. Lithotripsy always has a potential for damage to adjacent
tissues and organs if used improperly. Careful observation and attention to small details during the
procedure are essential if the procedure is to be performed safely and effectively. Selection of an
appropriate approach route, devices, and equipment should be based on a thorough understanding of
the characteristics, advantages, and disadvantages of the various options.
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
Shore JM, Lippman HN. A flexible choledochoscope. Lancet 1965;1:12001.

Takada T, Hanyu F, Kobayashi S, et al. Percutaneous transhepatic cholangiodrainage and
cholangioscopy for the cases of severe obstructive jaundice caused by pancreatic tumor.
Proceedings of the 18th World Congress of the International College of Surgeons.
Amsterdam, Neth: Excerpta Medica, 1972;478.
Gocho K, Hiratsuka H, Hasegawa M, et al. Percutaneous intrahepatic fistula dilation for
nonoperative choledochoscopy in intrahepatic stones. Jpn J Gastroenterol 1974;71:526.
Yamakawa T, Mieno K, Shikata J. Improved choledochofiberscopes and non-surgical removal
of retained biliary calculi under direct visual control [Abstract]. Gastroenterology 1975;68:
10512.
Moss JP, Whelan JG Jr, Powell RW, et al. Postoperative choledochoscopy via the T-tube
tract. JAMA 1976;236:27812.
Gocho K, Ikezawa H, Hiratsuka H. Endoscopic approaches in intrahepatic calculi. Report 1:
Significance of preoperative and postoperative choledochoscopy. Abstracts of the 8th
Congress of the Japanese Society of Gastroenterological Surgery, Aomori, Japan, 1975;24.
Gocho K, Hasegawa M, Masuda H, et al. Endoscopic approaches in intrahepatic calculi.
Report 2: In special reference to developments of forceps of choledochoscopy. Prog Dig
Endosc (Jpn) 1975;7:1703.
Gocho K, Hiratsuka H. Percutaneous intrahepatic fistula dilationCase report of intrahepatic
stones. Diagnosis and treatment of obstructive jaundice cases. Tokyo: Igakushoin Shuppan,
1976;8590.
Gocho K, Hiratsuka H. Postoperative choledochofiberscopic removal of intrahepatic stones.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Jpn J Surg 1977;7:1827.

Shikata J, Yamakawa T, Watanabe J, et al. Clinical evaluation of choledochofiberscope
[English Abstract]. Jpn J Gastroenterol 1976;73:122231.
Rattner DW, Warshaw AL. Impact of choledochoscopy on the management of
choledocholithiasis: Experience with 499 common duct explorations at the Massachusetts
General Hospital. Ann Surg 1981;194:769.
Yamakawa T, Komaki F, Shikata J. Experience with routine choledochoscopy via the T-tube
sinus tract. World J Surg 1978;2:37985.
Gocho K, Kameya S, Iijima N. Choledochofiberscopy. In Sivak MV Jr (ed). Gastroenterologic
Endoscopy. Philadelphia: WB Saunders, 1987;75269.
Orii K, Nakahara A, Takase Y. Choledocholithotomy by YAG laser with a
choledochofiberscope: Case reports of two patients. Surgery 1981;90:1202.
Gocho K, Ikezawa H, Yasuo M, et al. Experiences of choledochofiberscopic electrohydraulic
shock wave lithotripsy. Prog Dig Endosc (Jpn) 1985;26:1837.
Burhenne HJ. Percutaneous extraction of retained biliary tract stones: 661 patients. AJR Am J
Roentgenol 1980;134:88898.
Tsuchiya Y. A new safer method of percutaneous transhepatic cholangiography. Jpn J
Mazzariello RM. A fourteen-year experience with nonoperative instrument extraction of
retained bile duct stones. World J Surg 1978;2:44755.
Ellman BA, Berman HL. Treatment of a common duct stone via transhepatic approach.
Madhavan KK, Macintyre IM, Wilson RG, et al. Role of intraoperative cholangiography in
laparoscopic cholecystectomy. Br J Surg 1995;82:24952.
Kelley WE Jr, Sheridan VC. Laparoscopic choledochoscopy with a small-caliber endoscope.
A safe and effective technique for lap-aroscopic common bile duct exploration. Surg Endosc
1995;9:2936.
Iseli A, Marshall VC. Choledochoscopy: A comparison of a rigid and a flexible fiberoptic
instrument. Med J Aust 1978;1:1312.
Berci G, Shore JM, Morgenstern L, Hamlin JA. Choledochoscopy and operative
fluorocholangiography in the prevention of retained bile duct stones. World J Surg
1978;2:41127.
Yamakawa T, Mieno K, Gocho K, Shikata J. Improved choledochoscopes. Gastroenterol
Endosc (Jpn) 1975;17:45967.
Gocho K, Hiratsuka H, Yamakawa T. Endoscopic management of intrahepatic calculi using
the choledochofiberscope. Proceedings of the 2nd Asian Pacific Congress of Endoscopy,
Singapore, 1976;2148.
Ashby BS. Fiberoptic choledochoscopy in common bile duct surgery. Ann R Coll Surg (Engl)
1978;60:399403.
Chen MF, Chou FF, Wang CS. Removal of retained biliary stone by choledochofiberscope. J
Formos Med Assoc 1979;78:7616.
Okabe N, Kawai K, Kondo O, et al. Operative and postoperative choledochofiberscopy. Am J
Surg 1979;137:81620.
Ho LC. Postoperative choledochofiberscopy: An experience of 39 patients. Southeast Asian J
Surg 1980;3:6571.
Bauer JJ, Salky BA, Gelernt IM, Kreel I. Experience with the flexible fiberoptic
choledochoscope. Ann Surg 1981;194:1616.
Siegel JH, Sherman HI, Davis RC Jr, Webb WA. Modification of a bronchoscope for
percutaneous choledochoscopy. Gastrointest Endosc 1981;27:17981.
Murata N, Beppu T, Bandai Y, et al. Treatment of intrahepatic lithiasis using the
choledochofiberscope. Endoscopy 1981;13:2402.
Gocho K, Ikezawa H, Kameya S, et al. Experience using Machida fiberoptic choledochoscope
FCH-6T. Gastroenterol Endosc (Jpn) 1981;23:1862.
34. Hwang MH, Yang JC, Lee SA. Choledochofiberoscopy in the postoperative management of
intrahepatic stones. Am J Surg 1980;139:8604.
35.
Ker CG, Sheen PC. Postoperative choledochofiberscopic removal of retained intrahepatic
stones. Taiwan I Hseuh Hui Tsa Chi 1981;80:15869.
36. Birkett DH, Williams LF Jr. Postoperative fiberoptic choledochoscopy. Ann Surg
1981;194:6304.
37.
Nimura Y, Hayakawa N, Toyoda S, et al. Percutaneous transhepatic cholangioscopy (PTCS).
Stomach Intestine (Jpn) 1981;16:6819.
38.
Ichikawa K, Hayashi S, Ema Y, et al. Percutaneous transhepatic cholangioscopic laser
lithotomy (PTCS-LL) and percutaneous transhepatic cholecystoscopy (PTCCS).
39. Shimada H, Nakagawara G, Abe T, et al. Extraction of retained gallstones using a
fibercholedochoscope through a PTC-drainage fistula. Jpn J Surg 1983;13:4159.
40.
Inui K, Nakae Y, Nakamura J, et al. Percutaneous transhepatic cholecystoscopy (PTCCS).
41.
Ichikawa K, Nakazawa S, Naito Y, et al. Percutaneous transhepatic cholecystoscopy
(PTCCS)The second report. Gastroenterol Endosc (Jpn) 1983;25:798.
42.
Moss JP, Whelan JG Jr, Dedman TC III, Voyles RG. Postoperative choledochoscopy through
the T-tube tract. Surg Gynecol Obstet 1980;151:8079.
43. Birkett DH, Williams LF. Choledochoscopic removal of retained stones via a T-tube tract. Am
J Surg 1980;139:5314.
44.
Miyazaki I, Nagakawa T, Kurachi M, et al. Choledochojejunos-tomy (end-to-side anastomosis)
for postoperative choledocho-fiberscopy. In Japanese Research Group. Reports on Hepatic
Duct. 303 1979;Disorder.
45.
Hwang MH, Kuo RJ. Modified Roux-en-Y choledochojejunostomy with postoperative
choledochofiberoscopy in the treatment of intrahepatic stones. J Formos Med Assoc
1980;79:6316.
46. Siegel JH, Mayer LF. Percutaneous choledochoscopy and cholecystoscopy: Diagnostic and
therapeutic uses. Endoscopy 1981;13:1247.
47. Amberg JR, Chun G. Transcystic duct treatment of common bile duct stones. Gastrointest
Radiol 1981;6:3612.
48. Mason RR, Shorvon PJ, Cotton PB. Percutaneous descending biliary sphincterotomy with a
choledochoscope passed through the cystic duct after cholecystostomy. Br J Radiol
1982;55:5957.
49. Nakayama T, Ikeda A, Okuda K. Percutaneous transhepatic drainage of the biliary tract.
50.
Miyoshi T, Yamakawa T, Itoh S, et al. Percutaneous transhepatic cholangioscopy for
malignant biliary tract diseasesPTCS or PTCCS. Gastroenterol Endosc (Jpn)
1983;25:53741.
51. Yamakawa T, Komaki F, Shikata J. The importance of postoperative choledochoscopy for
management of retained biliary tract stones. Jpn J Surg 1980;10:3029.
52.
Gocho K, Ikezawa H, Hasegawa M, et al. Postoperative choledochoscopic follow-up of a
common bile duct cancer. Prog Dig Endosc (Jpn) 1977;10:2446.
53.
Nagakawa T, Asano E, Konishi I, et al. Results of choledochojejunostomy (end-to-side
anastomosis) for postoperative choledochofiberscopy. Operation (Jpn) 1983;37:10439.
54.
Sheen PC, Ker CG. Intrahepatic stones. II: Clinical study. J Formos Med Assoc
1981;80:12176.
55.
Ho LC. Surgical treatment of intrahepatic stones: Clinical significance of additional procedure.
J Chun Shan Med Dent Coll 1981;3:959.
56.
Nimura Y, Hayakawa N, Toyoda S, et al. Endoscopic treatment for intrahepatic stone. Surg
Ther (Jpn) 1982;46:57786.
57.
Komaki F, Yamakawa T, Shikata J. Newly devised forceps for non-surgical removal of
intrahepatic stones. Prog Dig Endosc (Jpn) 1976;9:658.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
Gocho K, Iizuka K, Ishimoto K, et al. Development and improvement of basket forceps for
choledochofiberscopy. Prog Dig Endosc (Jpn) 1979;14:658.
Demling L, Seuberth K, Rieman JF. A mechanical lithotriptor. Endoscopy 1982;14:1001.
Rieman JF, Seuberth K, Demling L. Mechanical lithotripsy of common bile duct stones.
Shimada K, Niimoto S, Fukushima Y, et al. Intra- and postoperative mechanical lithotripsy.
Biliary Tract Pancreas (Jpn) 1988;9:9991007.
Hayakawa N, Nimura Y, Kamiya J, et al. Studies of laser cholangioscopyApplication for
lithotripsy of gallstone. In Atsumi K (ed). Proceedings of the IVth Congress of the International
Society for Laser Surgery, 1981, Tokyo, Japan. Session 23, General Surgery (II), pp 1618.
1981.
Orii K, Ozaki A, Takase Y, Iwasaki Y. Lithotomy of intrahepatic and choledochal stones with
YAG laser. Surg Gynecol Obstet 1983;156:4858.
Kohzu T, Satoh H, Yamazaki Y. Application of Nd:YAG laser to crush calculus. Gastroenterol
Surg (Jpn) 1984;7:58793.
Sumiyama Y, Suzuki S. Clinical experiences in choledochoscopic lithotomy by laser. Biliary
Tract Pancreas (Jpn) 1988;9:10439.
Kozarek RA, Low DE, Ball TJ. Tunable laser lithotripsy: In vitro studies and in vivo treatment
of choledocholithiasis. Gastrointest Endosc 1988;34:41821.
Murray A, Basu R, Fairclough PD. Gallstone lithotripsy with the pulsed dye laser: In vitro
studies. Br J Surg 1989;76:45760.
Bogan ML, Robert HH, Kopecky KK. Percutaneous cholecystolithotomy with endoscopic
lithotripsy by using a pulsed dye laser. AJR Am J Roentgenol 1990;155:7814.
Ponchon T, Gagnon P, Valette PJ, et al. Pulsed dye laser lithotripsy of bile duct stones.
Ohshima I, Kozu T, Iaono K. The investigation for the cholangioscopic lithotripsy. Jpn J
Gastroenterol Surg 1991;24:26269.
Takekuma Y, Hirata N, Fuiita R. PTCS and lithotripsy (using a pulsed dye laser). Endoscopia
Digestiva 1992;4:30914.
Prat F, Fritsch J, Choury AD, et al. Laser lithotripsy of difficult biliary stones. Gastrointest
Endosc 1994;40:2905.
Martin EC, Wolff M, Neff RA, et al. Use of the electrohydraulic lithotriptor in the biliary tree in
dogs. Radiology 1981;139:2157.
Harrison J, Morris DL, Hayres J. Electrohydraulic lithotripsy of gallstonesIn vitro and animal
studies. Gut 1987;28:26771.
Frimberger E, Kuhner W, Weingart J, et al. Eine neue Methode der elektrohydraulischen
Cholelithotripsie (Lithoklasie). Dtsch Med Wochenschr 1982;107:2135.
Tanaka M, Yoshimoto H, Ikeda S, et al. Two approaches for electrohydraulic lithotripsy in the
common bile duct. Surgery 1985;98:3138.
Koch H, Stolte M, Walz V. Endoscopic lithotripsy in the common bile duct. Endoscopy
1977;9:958.
Manegold BC, Mennicken C, Jung M. Endoscopic electrohydraulic disintegration of common
bile duct concrements [Abstract]. Scand J Gastroenterol Suppl 1982;78:144.
Liguory CL, Bonnel D, Canard JM. Intracorporeal electrohydraulic shock wave lithotripsy of
CBD stones. Endoscopy 1987;19:23740.
Siegel JH, Ben-Zvi JS, Pullano WE. Endoscopic electrohydraulic lithotripsy. Gastrointest
Endosc 1990;36:1346.
Nakajima M, Fujimoto S, Imaoka W, et al. Endoscopic electrohydraulic lithotripsy (EEL) of
biliary tract stones. Gastroenterol Endosc (Jpn) 1982;24:13919.
Burhenne HJ. Electrohydraulic fragmentation of retained common bile duct stones. Radiology
1975;117:7212.
Lear JL, Ring EA, Macoviac JA, et al. Percutaneous lithotripsy. Radiology 1984;150:58990.
Ker CG, Sheen PC. Choledochoscopic electrohydraulic shock wave for treatment of residual
intrahepatic stones. Kao Hsiung I Hsueh Ko Hsueh Tsa Chih 1985;1:42531.

Gocho K, Miyamoto H, Ohdate K, et al. Choledochofiberscopic electrohydraulic lithotripsy.
86. Matsumoto S, Tanaka M, Yoshimoto H, et al. Electrohydraulic lithotripsy of intrahepatic stones
during choledochoscopy. Surgery 1987;102:8526.
87.
Gocho K, Miyamoto H, Sekizawa H, et al. Progress of gallstone disintegrator (EHL) and other
related equipments. Biliary Tract Pancreas (Jpn) 1988;9:32935.
88.
Gocho K, Miyamoto H, Kameya S, et al. Experiences of choledochofiberscopic
electrohydraulic lithotripsy. Biliary Tract Pancreas (Jpn) 1988;9:102933.
89. Yoshimoto H, Ikeda S, Tanaka M, et al. Choledochoscopic electrohydraulic lithotripsy and
lithotomy for stones in the common bile duct, intrahepatic ducts, and gallbladder. Ann Surg
1989;210:57682.
90.
Gocho K, Ohya K, Miyamoto H, et al. Characteristics of endoscopic direct visual biliary
lithotripsy. Kan Tan Sui (Jpn) 1990;20:5137.
91.
Tabuse K. A new operative procedure of hepatic surgery using a microwave tissue
coagulator. Arch Jpn Chir 1979;48:16072.
92.
Inoue S, Endoh M, Itoh T, et al. Microwave lithotripsy for intrahepatic gallstones. Hirosaki Med
1985;37:102939.
93.
Tabuse K, Tanimura H, Shimoma N. Role of lithotomy for gallstones by microwave tissue
coagulator. Biliary Tract Pancreas (Jpn) 1988;9:10518.
94.
Endoh M, Inoue S, Sasaki M, et al. Cholangioscopic microwave lithotomy for cholelithiasis.
95. Gacetta DJ, Cohen MJ, Crumy AB, et al. Ultrasonic lithotripsy of gallstones after
cholecystostomy. AJR Am J Roentgenol 1984;143:10889.
96.
Hwang MH, Chen GD, et al. Percutaneous transhepatic ultrasonic lithotripsy in the treatment
of biliary stones. Jpn J Med Ultrasonics 1986;13:3418.
97.
Ohtsubo K, Fujio T, Furukawa M. Ultrasonic lithotripsy in the treatment of biliary stones. Biliary
Tract Pancreas (Jpn) 1988;9:107580.
98.
Gocho K, Iwasaki M, Ohya K, et al. Application and complications of percutaneous
transhepatic cholangioscopy. Biliary Tract Pancreas (Jpn) 1985;6:132737.
99.
Samama G, Dupuis F, Descamps L. Cent lithiases du choledoque sans lithiase residuelle.
Interet de la choledocoscopie. Med Chir Dig (Fr) 1981;10:5756.
100.
Nimura Y, Maeda S, Kamiya J, et al. Endoscopic treatments of intrahepatic stones. Operation
(Jpn) 1982;36:1618.
101.
Chen MF, Chou FF, Wang CS, et al. Surgical treatment of intrahepatic gallstoneAn
experience of 194 operations. Scand J Gastroenterol Suppl 1982;17:389.
102. Chen MF, Chou FF, Wang CS, Jang YI. Experience with and complications of postoperative
choledochofiberoscopy for retained biliary stones. Acta Chir Scand 1982;148:5039.
103. Ponchon T, Genin G, Mitchell R, et al. Methods, indications, and results of percutaneous
choledochoscopy. A series of 161 procedures. Ann Surg 1996;223:2636.
104. Yeh YH, Huang MH, Yang JC, et al. Percutaneous transhepatic cholangioscopy and lithotripsy
in the treatment of intrahepatic stones: A study with 5 year follow-up. Gastrointest Endosc
1995;42:138.
105. Jan YY, Chen MF. Percutaneous transhepatic cholangioscopic lithotomy for hepatolithiasis:
Long-term results. Gastrointest Endosc 1995;42:15.
106. Saing H, Chan KL, Mya GH, et al. Cutaneous stoma in the roux limb of hepaticojejunostomy
(hepaticocutaneous jejunostomy): Useful access for intrahepatic stone extraction. J Pediatr
Surg 1996;31:24750.
107. Ker CG, Kuo KK, Tsai CC, et al. Evaluation of choledochojejunostomy with subcutaneous
jejunostomy for treatment of intrahepatic stones. Int Surg 1994;79:1103.
108. Kappas A, Alexander-Williams J, Keighley MRB, Watts GT. Operative choledochoscopy. Br J
Surg 1979;66:1779.
109.
Chen MF, Chou FF, Wang CS, Jang YY. Experiences and complications of post-operative
85.
choledochofiberscopy. J Formos Med Assoc 1981;80:51922.

110. Weller RM. The choledochoscopeTwo hazards. Anaesthesia 1982;37:606.
111. Griffin WT. Choledochoscopy. Am J Surg 1976;132:6978.
112. Chen MF, Jan YY. Bacteremia following postoperative choledochofiberscopyA prospective
study. Hepatogastroenterology 1996;43:5869.
Chapter 72 Congenital Anomalies of the Pancreas

(7478)
(7479)
GLEN A. LEHMAN, M.D.
By definition, congenital anomalies are present at birth but commonly are not suspected or detected
until adulthood. The term congenital anomaly indicates that during embryonic maturation, there was
atypical development and, in the final analysis, an abnormality occurred that by implication may cause
some disease, limitation, or disability. Developmental alterations that are generally of limited clinical
importance might best be termed congenital variations.
Most congenital anomalies and variants of the pancreas are of no clinical significance and are detected
coincidentally at endoscopy, surgery, or autopsy. In selected settings, however, these alterations have
clinical consequences; these are the focus of this chapter. This chapter is written largely from an adult
patient's viewpoint. Although significant congenital anomalies of the pancreas presenting in childhood
are generally detected by computed tomography (CT), barium studies, or surgery, endoscopic methods
of diagnosis and management are being employed with increasing frequency (see Chapter 79:
Endoscopic Retrograde Cholangiopancreatography in Infants and Children).14(7480) Other systemic
anomalies are also found with increased frequency in the population of patients with pancreatic
anomalies and variants, but these are not reviewed here.
A general classification of anomalous and variant pancreatic conditions is given in Table 721. The
embryology of the pancreas is reviewed in Chapter 58: Anatomy and Embryology of the Biliary Tract
and Pancreas. The embryonic dorsal pancreas and ventral pancreas come together medially as a
result of asymmetric rotation of the developing duodenum. The dorsal and ventral ductal systems
normally fuse during the sixth to eighth weeks of gestation, thereby resulting in normal pancreatic
ductal anatomy at birth (Figure 721A). However, a relatively large number of deviations from the
normal fused ductal pattern are known to exist (Figure 721B-H); some of these are significant in
terms of potential clinical consequences, whereas others are not.
Anatomic Categorization of
Congenital Pancreatic Anomalies and Variants
TABLE 721
1. Ventral and dorsal ductal malfusion

A. Pancreas divisum
B. Incomplete pancreas divisum
C. Isolated dorsal segment
2. Rotation and migration problems
A. Annular pancreas
B. Ectopic pancreas
C. Ectopic papillae
3. Quantitative underdevelopment
A. Agenesis
B. Hypoplasia
4. Duplication
A. Ductal
1. Total
Anatomic Categorization of
Congenital Pancreatic Anomalies and Variants
TABLE 721
2. Partialtail, body
B. Accessory papilla
5. Atypical ductal configuration
A. Ansa
B. Spiral
C. Horseshoe
D. Miscellaneous
5. Anomalous pancreatobiliary junction
6. Cystic malformations
A. Single
B. Polycystic
(7481)Figure 721. Variations of pancreatic ductal anatomy. A, Normal pancreatic duct with
a patent accessory duct that permits flow to the duodenum via the minor papilla. B, Complete
pancreas divisum with a small ventral duct draining to the major papilla and a large-caliber
dorsal duct draining via the minor papilla. C, Incomplete pancreas divisum with filamentous
communication between the ventral and the dorsal ducts. D, Ansa pancreatica with a looping
branch descending toward the uncinate process and back toward the minor papilla. Drainage
via the minor papilla is absent in this case. E, Typical pancreatic ductal system with an
accessory duct ending blindly with no connection to the minor papilla. F, Functional variant of
pancreas divisum in which the entire ductal system, including the uncinate process, drains via
the minor papilla. No pancreatic duct connects to the major papilla. G, Reverse pancreas
divisum with an isolated segment of the dorsal pancreas draining via the minor papilla that
does not connect with the main duct draining via the major papilla. H, Anomalous
pancreatobiliary junction with a long common channel measuring more than 15 mm.
Pancreas Divisum
Pancreas divisum is the most common congenital pancreatic anatomic variant. In autopsy series, it is
found in approximately 7% of cases (range 1 to 14%). The condition appears to be less frequent in
Asian5(7482) and black6(7483) populations. It has been described in two members of a kindred
(mother and son) with hereditary pancreatitis.7(7484)
The pancreatic ducts described by Wirsung, Santorini, and others8(7485) refer to the anatomically
normal ductal system without pancreas divisum. The eponyms for the pancreatic ducts associated with
their names cannot be applied readily to pancreas divisum and are not used in this chapter.
Pancreas divisum results from malfusion of the pancreatic ductal systems. Initially, the ventral
pancreas rotates posterior to the duodenum and comes to lie in contact with the dorsal pancreas.
Although parenchymal fusion nearly always occurs, the ducts may not unite. This results in a divided
pancreatic ductal system with a small ventral pancreas (approximately 10% of the pancreatic mass on
average) draining via the major papilla and a much larger dorsal pancreatic ductal system draining
through the generally small orifice of the minor papilla (Figure 722A; see also Figure 721B).
(7486)Figure 722. A, Typical small delicate ventral pancreas with a terminal branching
pattern. B, Adenocarcinoma of the pancreas (biopsy confirmed) with complete main duct
cutoff near the genu simulating a ventral pancreas.
Pancreas divisum has clinical relevance for three reasons:
1. At endoscopic retrograde cholangiopancreatography (ERCP), only the duct of the ventral
pancreas can be opacified via the major duodenal papilla. Thus, incomplete ductography results
and the dorsal pancreas remains unevaluated unless minor papilla cannulation is performed.
2. The ventral pancreatic duct is relatively short by comparison with the retrograde pancreatogram of
a normally fused ductal system. A wide range of pancreatic abnormalities can cause obstruction
of the pancreatic duct. Thus, the appearance of the seemingly short ventral duct must be
differentiated from conditions that result in a pancreatogram with a cutoff such as pancreatic
cancer (see Figure 722B).
3. In select patients, the minor papilla orifice appears to be critically small, such that excessive
intrapancreatic dorsal ductal pressure occurs during active secretion. This may result in ductal
distention, inadequate drainage, pain, or pancreatitis.
Because the vast majority of patients with pancreas divisum have no pancreatic symptoms throughout
their lifetime, it appears that this condition only predisposes individuals to the sequence of events
mentioned previously. This low frequency of symptoms has incited great controversy as to whether
pancreas divisum and an associated small minor papilla orifice are ever a cause of obstructive
pancreatitis.920(7487) However, an increased prevalence of pancreas divisum has been noted in
several series of patients with "idiopathic pancreatitis." In these patients, pancreatitis is generally of the
acute recurrent type and of mild to moderate severity, although pseudocysts and chronic calcific
pancreatitis have also been observed.2123(7488) In many of these reported cases of pancreas
divisum associated with chronic pancreatitis, however, other potential causes such as alcohol abuse
and trauma were also identified. For example, in the series of Barthet et al.,23(7489) 18 of 20 patients
were alcoholics. Staritz et al.24(7490) found basal intraductal (dorsal duct) pressure in these patients
to be twofold greater than that of patients without divisum. The predominance of available data
supports the view that patients with pancreas divisum are at increased risk for obstructive
pancreatopathy. Further studies are needed to determine whether the effects of known pancreatic
irritants such as alcohol and trauma are potentiated by low-grade ductal obstruction as occurs in
pancreas divisum.22(7491)
Failure of ventral and dorsal ductal fusion is not absolute in 15% of cases of pancreas
divisum.2527(7492) In such "incomplete" or "partial" cases, a filamentous branch of the ventral duct
communicates with the dorsal duct (see Figure 721C). The clinical implications of incomplete
pancreas divisum remain the same as for complete pancreas divisum except that partial to full
visualization of the dorsal duct may occur with vigorous injection of contrast material via the major
papilla.
The Ventral Pancreas

Characteristically, the main ventral duct is 1 to 4 cm long and tapers terminally into multiple small side
branches (see Figure 722A). This ductal system does not cross over the spine, and acinarization
(opacification to the level of the acinus) quickly occurs if the pancreatographic findings are not
interpreted correctly and contrast material injection is continued. Acinarization typically causes pain and
is viewed fluoroscopically as a focal, fluffy collection of contrast media with generally sharp peripheral
margins. Acinarization must be differentiated from submucosal injection (fuzzy peripheral margins),
acinarization as a result of cannulation of a side branch of the normal main pancreatic duct, and
opacification of a pseudocyst, diverticulum, or tumor cavity. After ventral pancreas acinarization,
contrast material will drain promptly in 1 to 2 minutes, and radiographs obtained during this interval will
identify the underlying ventral ductal system. Alternatively, the ventral duct can be demonstrated by
reinjection with lesser volumes of contrast material. Usually, no additional pathologic abnormalities are
found when the small, delicate ventral duct is opacified, but occasionally, there will be changes of
chronic pancreatitis (generally associated with similar changes in the dorsal duct) or tumor. Very rarely,
the ventral pancreas will be abnormal while the dorsal duct remains normal.2831(7493) In up to
one-third of cases of pancreas divisum, no pancreatic duct can be identified connecting to the major
papilla (see Figure 721F). In such cases, the entire ventral portion of the pancreas usually drains
cephalad through a branch of the dorsal duct.
At ERCP, differentiation of a ventral duct in pancreas divisum from pancreatic cancer or benign
stricture may be extremely difficult until complete dorsal ductography is obtained. Complementary
information from endoscopic ultrasonography (EUS) (see Chapter 78: Endoscopic Ultrasonography of
the Retroperitoneal Organs) or CT may be required.
Minor Papilla Cannulation and Dorsal Ductography

Cannulation of the minor papilla and dorsal ductography are essential to fully evaluate the pancreatic
ductal system in pancreas divisum. In general, minor papilla cannulation should be done when a
ventral duct is demonstrated via the major papilla or attempts to achieve pancreatography via this
structure have failed. The minor papilla is nearly always located in the right upper quadrant of the
visual endoscopic field when facing the major papilla. It may be as close as 10 mm from the major
papilla and at the cephalad end of a fold, but generally it is 2 to 3 cm more cephalad and anterior to the
major papilla.*(7494) There is a case report of termination of the dorsal duct at the major papilla with
an orifice separate from that for the ventral pancreatic and bile ducts.32(7495)
Cannulation of the minor papilla is usually best achieved with the endoscope pushed into the long
position along the greater curve of the stomach (see Chapter 57: Technique of Endoscopic Retrograde
Cholangiopancreatography). Various accessory devices are potentially useful for achieving cannulation
of the minor papilla (Figure 723). Rarely, a standard tapered-tip 5-French catheter will suffice. The
most helpful accessory for achieving cannulation is the highly tapered 5-French catheter with a blunt
needle protruding 1 to 2 mm beyond the catheter tip.3335(7496) Catheters with longer needle tips
and those without tapered tips are less helpful. The needle tip should be seated gently into the minor
papilla orifice to avoid tissue trauma and consequent blurring of landmarks. Contrast media, patient
positioning, and fluoroscopic techniques are the same as for conventional ERCP (see Chapter 57).
(7497)Figure 723. Accessory devices for cannulation of the minor papilla: 23-gauge blunted
needle-tipped catheters (top and middle); a 0.018-inch-diameter soft-tipped guidewire inside a
3-French catheter inside a 5-French catheter (bottom).
Newer-generation video endoscopes clearly facilitate minor papilla identification and cannulation;
however, in approximately one-third of cases, the orifice to the minor papilla will not be evident despite
careful study. To facilitate identification of the orifice, secretin can be given intravenously (0.25 to 1
U/kg) (Figure 724).34(7498) This usually results in vigorous flow of pancreatic exocrine secretions
and dilatio

Gastroenterologic Endoscopy (Vickey1214) VRG

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gastroenterologic Endoscopy (Vickey1214) VRG

Uploaded by

Copyright:

Available Formats

WILLIAM S. HAUBRICH, M.D.

JAMES M. EDMONSON, PH.D.

Inception of Gastrointestinal Endoscopy

Early Attempts at Gastroscopy

The Remarkable Career of Rudolf Schindler

employed by Dr. Rudolf Schindler. A, Cricopharyngeal sphincter in open and closed

(27)Figure 14. A, Schindler's semiflexible gastroscope with accessories. B, Proximal

only secondarily one who was skilled in technology.

Gastroscopy As It Was Performed During the Era of Rudolf Schindler

by Dr. Eddy Palmer.27(32)

(33)Figure 16. A, Eder-Palmer semiflexible gastroscope. Note side-viewing optics. B,

The Brief but Illuminating Era of the Gastrocamera

million subjects annually).

(47)Figure 111. Scheme of the prescribed sequence in which gastrocamera photographs

The Advent of Fiberoptics

advent of fiberoptic endoscopy.39(53)

Development of the Fiberoptic Endoscope

The New Era of Video Endoscopy

Proctosigmoidoscopy and Colonoscopy

(69)Figure 115. Early prototype fiberoptic sigmoidoscope developed at the University of

Endoscopic Retrograde Cholangiopancreatography

States by Herman Moersch92(105) and by Cecil Patterson and Milford Rouse.93(106)

Walk L. The history of gastroscopy. Clio Med 1965;1:20922.

Chapter 2 Flexible Endoscope Technology: The Fiberoptic

Mechanical Construction of the Flexible Endoscope

Bending Section and Angulation System

(113)Figure 24. Bending section and angulation system of a flexible endoscope.

(114)Figure 25. Cross-sectional view of the bending section of a flexible endoscope.

Air, Water, and Suction System

Light Guide Connector

Principle of Total Internal Reflection

n sin A = n' sin B

(122)Figure 213. Alignment of fibers on the face of a coherent fiber bundle.

Basic Optical System of a Fiberscope

(123)Figure 214. A basic fiberscope optical system. I.G.image guide.

where M0 is usually much less than 1.

where M1 typically ranges between 15 and 30.

where f is the focal length of the objective lens.

(124)Figure 215. Flexible endoscope illumination system. L.G.light guide.

Production and Quality of the Image Guide

(126)Figure 217. Master IG bundle.

(127)Figure 218. Process of pulling a master IG bundle to reduce its diameter.

(131)Figure 222. Example of various forms of fiber bundle imperfections.

Total Endoscopic System

Cleaning, Sterilization, and Maintenance of the Endoscope and

Chapter 3 Flexible Endoscope Technology: The Video Image

Solid-State Imaging Technology

light and the exposure time.

Line Transfer CCD

Frame Transfer CCD

Interline Transfer CCD

Shape of the Displayed Image

Theory of Color Video

RGB Sequential Imaging Technology

Color-Chip Video Imaging Technology

(152)Figure 314. A, CCD with an RGB-striped filter. B, Combining the brightness

Inherent Advantages of a Color-Chip Videoscope

therapy procedures (see later).

Laser Therapy Via the Videoscope

Advantages of the RGB Sequential Videoscope

Recent Advances in RGB Sequential Videoscopes

Designing for Compatibility

Functions of a Typical Video Processor