Chapter 53 - Exocrine Pancreas: Michael L. Steer M.D

Use of this content is subject to the Terms and Conditions of the MD Consult web site.
Townsend: Sabiston Textbook of Surgery, 17th ed., Copyright 2004 Elsevier
1643
Chapter 53 - Exocrine Pancreas

Michael L. Steer M.D.
The pancreas was first mentioned in the writings of Eristratos (310250 BC) and given its name by Rufus of Ephesus (c.100 AD). The name pancreas (Greek pan: all, kreas: flesh or meat) was used because the organ contains neither cartilage nor bone. Its main duct was described by Wirsung in 1642, whereas the enlargement of that duct at its junction with the common bile duct and its projection into the duodenum as a papilla were first described by Vater in 1720. Santorini, in 1734, described the accessory duct that bears his name.
Copyright 2004 Elsevier Inc. All rights reserved. www.mdconsult.com
Bookmark URL: /das/book/view/41787982-2/1235/526.html/top
ANATOMY
Location
The pancreas lies posterior to the stomach and lesser omentum in the retroperitoneum of the upper abdomen. It extends obliquely, rising slightly as it passes from the medial edge of the duodenal C loop to the hilum of the spleen. It lies anterior to the inferior vena cava, aorta, splenic vein, and left adrenal gland.
Regions
The pancreas is divided into four regions: the head/uncinate process, neck, body, and tail. The head lies within the duodenal C loop, and its uncinate process extends posteriorly and medially to lie behind the portal/superior mesenteric vein and superior mesenteric artery. The neck of the gland extends medially from the head to lie anterior to those vessels. The body extends laterally from the neck toward the spleen, whereas the tail extends into the splenic hilum.
Blood Supply and Lymph Nodes
Both the celiac trunk and the superior mesenteric artery provide the arterial supply to the pancreas. Variations are common but, for the most part, the body and tail are supplied by branches of the splenic artery while the head and uncinate process receive their supply via arcades originating from the hepatic/gastroduodenal branch of the celiac artery and from the first branch of the superior mesenteric artery (Fig.531A). Venous drainage is to the splenic, superior mesenteric, and portal veins ( Fig. 531 B). The pancreas is drained by multiple lymph node groups. The major drainage of the pancreatic head and uncinate process is to the subpyloric, portal, mesenteric, mesocolic, and aortocaval nodes. The pancreatic body and tail, for the most part, are drained via nodes in the celiac, aortocaval, mesenteric, and mesocolic groups and via nodes in the splenic hilum.
Innervation
The pancreas is innervated by both sympathetic and parasympathetic components of the autonomic nervous system. The principal, and possibly only, pathway for pancreatic pain involves nociceptive fibers arising in the pancreas. They pass through the celiac ganglia to form the greater, lesser, and least splanchnic nerves that pass to cell bodies in the thoracic sympathetic chain. Efferent visceral motor supply to the pancreas is provided by both the sympathetic and parasympathetic systems. The latter involves preganglionic fibers arising from cell bodies in the vagal nuclei that travel through the posterior vagal trunk to the celiac plexus. Postganglionic fibers then innervate pancreatic islets, acini, ducts, and blood vessels. In general,
1644
Figure 53-1 Arterial supply to the pancreas (A) and venous drainage of the pancreas (B). The pancreatic head is supplied by branches of the gastroduodenal and superior mesenteric
arteries while the body and tail are supplied by branches of the splenic artery. Venous drainage is to the splenic and superior mesenteric/portal veins. (From Skandalakis JE, Gray SW, Rowe JS Jr, et al: Anatomical complications of pancreatic surgery. Contemp Surg 15:1750, 1979.)
the nerves of the pancreas travel with the blood vessels supplying the organ.
Ducts
The main pancreatic duct, or duct of Wirsung, arises in the tail of the pancreas and terminates at the papilla of Vater in the duodenum. It crosses the vertebral column between T-12 and L-2. Within the body and tail of the pancreas, the duct lies slightly cephalad to a line drawn midway between the superior and inferior edges. The duct is also more posterior than anterior. In adults, the duct within the head measures 3.1 to 4.8 mm in diameter and it gradually tapers, to measure 0.9 to 2.4 mm in the tail. With age, the duct diameter can increase. The duct of Santorini (i.e., the minor, or accessory, pancreatic duct) is smaller than the main duct. It extends from the main duct to enter the duodenum at the lesser papilla. That papilla lies approximately 2 cm proximal and slightly anterior to the major papilla.
EMBRYOLOGY AND HISTOLOGY

Organogenesis
During the 4th week of gestation, two endodermal buds arise from the duodenumthe hepatic diverticulum, which is destined to form the liver, gallbladder, and bile ducts, and the dorsal pancreatic bud that forms the body
1645
Figure 53-2 Organogenesis of the pancreas. A, Formation of dorsal and ventral pancreatic buds. B, Rotation of the ventral pancreas, distal bile duct, and major papilla. C, Fusion of the dorsal and ventral pancreata to form the adult pancreas. (From Skandalakis JE, Gray SW, Rowe JS Jr et al: Anatomical complications of pancreatic surgery. Contemp Surg 15:17 50, 1979.)
and tail of the pancreas ( Fig. 532 ). On the 32nd day of gestation, this hepatic diverticulum gives rise to a ventral pancreatic bud that eventually develops into the uncinate process and inferior part of the head of the pancreas. The dorsal pancreatic bud extends transversely across the abdomen, to lie anterior to the portal and mesenteric vessels. With time, as the duodenum rotates to form a C-loop configuration, the ventral pancreas and distal bile duct undergo clockwise rotation around the back of the duodenum to, finally, lie on the medial side of the duodenum, inferior and slightly posterior to the dorsal pancreas and posterior to the portal and mesenteric vessels. On the 37th day of gestation, the two pancreatic buds fuse and, in 90% of individuals, their duct systems also join.
Histology
The mature pancreas is an endocrine organ made up of the islets of Langerhans and an exocrine organ consisting of acinar and ductal cells. The acinar cells, so named because they are clustered like grapes on the stem of a vine, discharge their secretions into a centrally located acinar space that communicates with the main pancreatic duct. Most of the cells in the pancreas are acinar cells, and duct cells make up only 5% of pancreatic mass. Histologically, acinar cells have a high content of endoplasmic reticulum and an abundance of apically located eosinophilic zymogen granules. The cells lining the main pancreatic duct are tall columnar cells, and many contain mucin granules. With progression from the large ducts to the smaller intralobular and interlobular ducts, the lining cells become flatter, assuming a cuboidal configuration, and mucin granules are no longer seen. Centroacinar cells, located at the junction between ducts and acini, resemble acinar cells in size and shape but lack zymogen granules.
Cell Differentiation
The cells comprising the pancreatic buds are homogeneous and indistinguishable from other endodermal cells of the primitive gut. These endodermal cells undergo step-wise differentiation, from an undifferentiated precursor into committed islet and exocrine cell precursors and then into either acinar cells or ductal cells. Some recently presented evidence has also suggested that transdifferentiation can occurthat differentiated duct cells may give rise to islet and/or acinar cells.
CONGENITAL ANOMALIES
Pancreas Divisum
Failure of the dorsal and ventral pancreatic duct systems to join during embryogenesis (see Fig. 532 ) is referred to as pancreas divisum. It results in a pancreas with divided
1646
drainage because the dorsal pancreas drains, via the duct of Santorini, to empty at the lesser papilla while the ventral pancreas, composed of the head and uncinate process, drains via Vaters papilla. Pancreas divisum has been noted in as many as 11% of autopsy cases. The significance of pancreas divisum remains controversial. Some have suggested that it may contribute to the development of pancreatitis by establishing a condition of relative outflow obstruction because the major fraction of pancreatic exocrine secretion is obliged to exit through the relatively small orifice of the lesser papilla. On the other hand, the presence of pancreas divisum and the development of pancreatitis are, in most patients, not related to each other in a cause-effect manner and the corollary of this may also be true; that is, attempts to widen the orifice of the dorsal duct at the lesser papilla in patients with pancreas divisum and pancreatitis are unlikely to be of benefit.
Ectopic and Accessory Pancreas
Pancreatic tissue at ectopic sites is not unusual, and most ectopic pancreatic tissue is functional. The most common sites are in the walls of the stomach, duodenum or ileum, in a Meckels diverticulum, or at the umbilicus. Less common sites include the colon, appendix, gallbladder, omentum, and mesentery. Islet tissue is frequently present when ectopic pancreas is located in the stomach and duodenum but not when it is present elsewhere. For the most part, ectopic pancreatic tissue is a submucosal, irregular nodule of firm, yellow tissue that may have a central umbilication. Pancreatic secretions often exit through this umbilication into the lumen of the stomach or intestine. Ulceration and, on occasion, bleeding can be associated with these lesions. They may also be associated with obstruction or be the lead point for intussusception. Resection or bypass is indicated in such cases.
Annular Pancreas
Annular pancreas refers to the presence of a band of normal pancreatic tissue that partially or completely encircles the second portion of the duodenum and extends into the head of the pancreas. It usually contains a duct that joins the main pancreatic duct. The basis for annular pancreas is uncertain. It may result from failure of normal clockwise rotation of the ventral pancreas, or it may result from expansion of ectopic pancreatic tissue in the duodenal wall. It presents with varying degrees of duodenal obstruction that, in children, is often associated with other congenital anomalies. It may be totally asymptomatic or present later in life with obstructive symptoms if pancreatitis develops in the annular segment. Treatment usually involves bypass, via duodenojejunostomy, rather than resection.
Developmental Pancreatic Cysts
Solitary (congenital, duplication, or dermoid) cysts of the pancreas are rare. In contrast, multiple pancreatic cysts, lined with cuboidal epithelium, are more common. They are frequently associated with polycystic disease of the liver and/or kidney, and they can be seen in up to half of patients with von Hippel-Lindau syndrome. Pancreatic cysts only rarely become symptomatic and, in general, no treatment is indicated.
PHYSIOLOGY Approximately 2.5 L of clear, colorless, bicarbonate-rich pancreatic juice, containing 6 to 20 g of protein, is secreted by the human pancreas each day. It plays a critical role in duodenal alkalinization and in food digestion.
Protein Secretion
With the possible exception of the lactating mammary gland, the exocrine pancreas synthesizes protein at a greater rate, per gram of tissue, than any other organ. More than 90% of that protein consists of digestive enzymes. Most of the digestive enzymes are synthesized and secreted by acinar cells as inactive proenzymes or zymogens that, in health, are activated only after they reach the duodenum where enterokinase activates trypsinogen and the trypsin catalyses the activation of the other zymogens. Some of the pancreatic digestive enzymes are synthesized and secreted in their active forms without the need for an activation step (e.g., amylase, lipase, ribonuclease). Acinar cells also synthesize proteins, including enzymes, that are not destined for secretion but, rather, are intended for use within the acinar cell itself. Examples of this latter group of proteins include the various structural proteins and lysosomal hydrolases. Newly synthesized proteins are assembled within the cisternae of the rough endoplasmic reticulum and transported to the Golgi where they are modified by glycosylation. Those destined for secretion pass through the Golgi stacks and are packaged within condensing vacuoles that evolve into zymogen granules as they migrate toward the luminal surface of the acinar cell. By a process involving membrane fusion and fission, the contents of the zymogen granules are then released into the acinar lumen.[1] Other proteins that are not destined for secretion are segregated away from the secretory pathway as they pass through the Golgi, and they are then targeted to their appropriate intracellular site.[2] Secretion of protein from acinar cells is a regulated process. At rest, secretion occurs at a low or basal rate, but this rate can be markedly increased by secretory stimulation that, in the pancreas, is both hormonal and neural. Pancreatic acinar cells can express receptors for acetylcholine, cholecystokinin, secretin, and vasoactive intestinal peptide. Stimulation of secretion by either acetylcholine or cholecystokinin has been shown to involve activation of phospholipase C, generation of inositol triphosphate and diacyl glycerol, and a rise in intracellular ionized calcium levels that, by yet unidentified mechanisms, upregulates the rate of secretory protein discharge at the apical cell membrane. In contrast, secretin and vasoactive intestinal peptide activate adenylate cyclase, increase cellular levels of cyclic adenosine
1647
Figure 53-3 Secretion of bicarbonate-rich juice by duct cells. CO2 diffuses into duct cells and is converted to H2 CO3 by carbonic anhydrase. H2 CO3 dissociates into HCO3 and H+ , which exits at the basal side of the cell. Secretin stimulation increases cyclic adenosine monophosphate (cAMP), which accelerates apical chloride secretion through the cystic fibrosis transmembrane regulator (CFTR) chloride channel. The exchange of luminal Cl for cellular HCO3 results in net HCO3 secretion.
monophosphate (AMP), and activate protein kinase A. This also leads to protein secretion at the apical pole. Recent studies indicate that human acinar cells may not possess receptors for cholecystokinin and that, in humans, cholecystokinin stimulation of secretion is mediated by intrapancreatic nerves that express cholecystokinin receptors.
Electrolyte Secretion
Although stimulation of acinar cells results in the secretion of a small amount of serum-like fluid, most of the fluid and electrolytes secreted from the pancreas arise from duct cells ( Fig. 533 ). [3] The earliest step in duct cell electrolyte secretion involves diffusion of circulating carbon dioxide into the duct cell and that carbon dioxide is hydrated by carbonic anhydrase to yield carbonic acid. Subsequently, the carbonic acid dissociates into protons and bicarbonate ions. The protons diffuse out of the cell and are carried away in the circulation while the bicarbonate remains inside the cell. The fluid and electrolyte secretagogue secretin acts, via a cyclic AMP mediated process, to stimulate chloride secretion, at the apical cell surface, via cystic fibrosis transmembrane regulator (chloride) channels. Then, via an apical chloride-bicarbonate exchanger, that actively secreted chloride is re-taken up by the duct cell in exchange for bicarbonate. Taken together, the result of these events is the secretion of a bicarbonate-rich fluid into the duct and the discharge, into the circulation, of protons (see Fig. 533 ). In the absence of secretin stimulation, pancreatic juice has a more plasma-like composition
since it is comprised primarily of acinar cell secretions and there is little duct cell secretion of chloride to permit exchange with bicarbonate. With secretin stimulation, chloride secretion is increased, flow rates rise, and chloride-bicarbonate exchange results in juice that is rich in bicarbonate and poor in chloride.
Integrated Physiology
During the resting (interdigestive) phase of gastrointestinal function, pancreatic secretion is minimal and may be as low as 2% of that noted with maximal stimulation. The pancreatic response to a meal is a three-phase process that includes a cephalic phase, a gastric phase, and an intestinal phase. The cephalic phase, accounting for 10% to 15% of meal-stimulated pancreatic secretion, reflects the response to the sight, smell, or taste of food. It is believed to be almost exclusively mediated by peripherally released acetylcholine, which directly stimulates pancreatic secretion of enzymes and gastric secretion of acid. The acid indirectly stimulates pancreatic secretion of fluid and electrolytes by causing duodenal acidification and secretin release. The gastric phase of pancreatic secretion, accounting for 10% to 15% of meal-stimulated pancreatic secretion, reflects the response to gastric distention and the entry of food into the stomach. These events can cause release of gastrin and stimulate vagal afferents. By binding to cholecystokinin receptors, gastrin is itself a weak stimulant of pancreatic enzyme secretion. Vagal stimulation also increases enzyme secretion. More important, however, gastrin and vagal stimulation cause gastric acid secretion, and this leads to duodenal acidification, release of secretin from the duodenum, and pancreatic secretion of fluid and electrolytes. The intestinal phase of pancreatic secretion reflects the response to food and gastric secretions entering the proximal intestine. Acidification of the duodenum and the presence of bile in the duodenum promote secretin release. In addition, in the duodenum and proximal small intestine, the presence of fat
1648
and protein, as well as their partial breakdown products, stimulates the release of cholecystokinin, and this cholecystokinin stimulates enzyme secretion from acinar cells. The intestinal phase of pancreatic secretion accounts for 70% to 75% of meal-stimulated pancreatic secretion.
Feedback Loop
Luminal proteins, referred to as releasing factors, have been described that can also stimulate cholecystokinin and secretin release. The most well characterized are the releasing factors for cholecystokinin.[4] Two forms are known, one apparently synthesized by duodenal cells (cholecystokininreleasing factor) and the other secreted by the pancreas (monitor peptide). Both forms are subject to degradation by trypsin. Thus, with high-protein meals that quench intraduodenal tryptic activity, the releasing factor remains intact, cholecystokinin release is increased, and pancreatic secretion is stimulated. In contrast, when food is absent from the duodenum, the proteolytic activity that remains unquenched within the lumen degrades the releasing factor and, as a result, cholecystokinin release and pancreatic secretion are reduced. Some have argued that this feedback loop may, at least in part, explain the pain of chronic pancreatitis since, with pancreatic insufficiency, intraluminal proteolytic activity would be low and cholecystokinin release would increase. Based on this concept, some have advocated administration of exogenous pancreatic enzymes as a treatment for the chronic pain of pancreatitis. Presumably, administration of exogenous enzymes to such patients would result in degradation of the releasing factor and reduce pancreatic stimulation. However, evidence supporting a physiologic role for these releasing factors comes almost exclusively from experiments using rodents, and the actual existence of a physiologic feedback loop in humans has not been established.
PANCREATITIS
Definition and Classification
Pancreatitis can be classified as either acute or chronic based on its clinical characteristics, pathologic changes, and natural history. Clinically, acute pancreatitis is usually characterized by the acute onset of symptoms in a previously healthy individual and the disappearance of those symptoms as the attack resolves. In contrast, patients with chronic pancreatitis may have had prior attacks or symptoms of either exocrine or endocrine insufficiency prior to the current attack, and their symptoms may persist even after resolution of the current attack. From a clinical standpoint, however, attacks of either acute or chronic pancreatitis can be characterized by the abrupt onset of symptoms that are often similar. Thus, without the test of time or a tissue sample, it may be difficult or impossible to determine if a first attack is one of acute or chronic pancreatitis.
Pathology
The pathologic changes of acute pancreatitis include parenchymal and peripancreatic fat necrosis and an associated inflammatory reaction. The extent of these changes is directly related to the severity of an attack. In mild pancreatitis, changes frequently include interstitial edema and infiltration of inflammatory cells with relatively little necrosis, whereas in severe pancreatitis, extensive necrosis, thrombosis of intrapancreatic vessels, vascular disruption, and intraparenchymal hemorrhage can be seen. With infection, intrapancreatic or peripancreatic abscesses involving areas of necrosis can also develop. The major changes of chronic pancreatitis include fibrosis and loss of both endocrine and exocrine elements. In addition, an acute inflammatory reaction may be superimposed on a background of chronic inflammation. There may be enlargement of pancreatic nerves, and perineural inflammation has also been described.
Etiology
Both acute and chronic pancreatitis are frequently associated with other disease entities collectively referred to as the etiologies of pancreatitis ( Box 531 ). In developed countries, roughly 70% to 80% of patients with pancreatitis have their pancreatitis in association with either biliary tract stone disease or ethanol abuse. For the most part, biliary tract stone disease is associated with acute pancreatitis, whereas chronic pancreatitis is associated with the intake of large amounts of ethanol over protracted
Box 53-1. Etiologies of Pancreatitis Acute Pancreatitis Biliary tract stones Drugs ERCP Ethanol abuse Hypercalcemia Hyperlipidemia Idiopathic Infections Ischemia Parasites Postoperative Scorpion sting Trauma Chronic Pancreatitis Autoimmune Duct obstruction Ethanol abuse Hereditary Hypercalcemia Hyperlipidemia Idiopathic
ERCP, endoscopic retrograde cholangiopancreatography.
1649
periods. In 10% to 15% of pancreatitis cases, no etiology can be identified, and those individuals are said to have idiopathic pancreatitis. In the remaining 10% to 15% of patients, pancreatitis is associated with one of many possible miscellaneous etiologies. In underdeveloped countries, particularly in Africa and Southeast Asia, pancreatitis is frequently termed either tropical or nutritional. Recent reports indicate that many patients with tropical pancreatitis have a form of hereditary pancreatitis caused by mutations of the genes that code for pancreatic secretory trypsin inhibitors.[5] Affected individuals often complain of painful attacks, and they frequently develop pancreatic calcifications as well as diabetes. Ketoacidosis is uncommon.
Biliary Tract Stones
The onset of acute pancreatitis is frequently associated with the passage of biliary tract stones through the terminal biliopancreatic duct into the duodenum. Stones can be retrieved from the stools of roughly 90% of patients with stone-induced pancreatitis. There has been much speculation regarding the mechanisms by which such stones might cause pancreatitis. In 1901, Opie, a pathologist at Johns Hopkins University, noted a stone lodged in the terminal biliopancreatic duct of a patient who had died of severe pancreatitis. He suggested that the stone might have caused outflow obstruction from a common biliopancreatic channel, allowing bile to reflux into the pancreatic duct.[6] In a second publication based on observations made at another autopsy, Opie suggested that biliary pancreatitis could also occur when a stone, or the edema and inflammation caused by its passage, caused outflow obstruction of the pancreatic duct even in the absence of bile reflux. Although the bile reflux theory, often referred to as the common channel theory, was originally favored, subsequent studies have cast doubt on its validity, and most observers now believe that it is stone-induced pancreatic duct obstruction and ductal hypertension, rather than bile reflux, that triggers acute pancreatitis. Recent data derived from experiments using a model of pancreatitis induced in opossums also support the duct obstruction theory. Those experiments indicate that pancreatic duct obstruction, without bile duct obstruction or bile reflux, can cause pancreatitis and that the severity of pancreatitis is not worsened by bile reflux into the pancreatic duct.[7]
Abuse of Ethanol
The most frequent cause of morphologically defined chronic pancreatitis is ethanol abuse, but occasionally ethanol can also induce acute pancreatitis. There is no threshold rate of consumption below which ethanol consumption is not associated with an increased incidence of pancreatitis. The mean ethanol consumption among patients with ethanol-induced pancreatitis is 150 to 175 g/day. The mean duration of ethanol abuse for men is 18 11 years and, for women, it is 11 8 years. Ethanol-induced pancreatitis, like ethanol abuse itself, is more common in men than in women. Dietary factors, such as consumption of a high-protein diet with either high- or low-fat content, may contribute to the development of pancreatitis. Most observers currently believe that the chronic pancreatitis that follows prolonged ethanol abuse reflects repeated, but subclinical, episodes of acute pancreatic injury. These repeated episodes of pancreatic injury with necrosis eventually lead to the fibrosis that characterizes chronic pancreatitis.[8] A number of theories have been advanced to explain the mechanism by which ethanol might cause pancreatic injury. According to one theory, ethanol consumption causes hypertriglyceridemia and the generation of fatty acids as well as their ethyl ester metabolites that can injure the pancreas. Another theory suggests that ethanol ingestion causes intrapancreatic generation of oxygen-derived free radicals that can injure the pancreas. Others believe that ethanol acts directly on pancreatic acinar cells to cause injury or that it promotes secretion of pancreatic juice that is high in proteolytic enzyme content but low in enzyme inhibitor content. Theoretically, enzyme activation could occur under these conditions and that activation could cause pancreatic injury. Secretion of an enzyme-rich fluid deficient in enzyme inhibitors could also lead to protein precipitation and the formation of intraductal plugs. Those plugs, by causing duct obstruction and ductal hypertension, could subsequently trigger pancreatic injury. Ethanol ingestion has also been reported to cause sphincter of Oddi spasm, and this could also contribute to ethanol-induced pancreatitis if it resulted in ductal hypertension. Each of these various theories has attractive features and its own proponents, but, at present, the actual mechanisms by which ethanol causes pancreatitis remain unclear.
Drugs
Exposure to certain drugs is, perhaps, the third most frequent cause of pancreatitis ( Box 532 ), but the mechanisms by which those drugs cause pancreatitis is not known. Although many different drugs have been implicated, the strength of the data supporting a cause-effect relationship in pancreatitis varies considerably. Drugs associated with pancreatitis can be divided into the following three groups: (1) those considered to be definite causes of pancreatitis because their use has been associated with the onset of pancreatitis and the disease has recurred when patients have been rechallenged with the drug; (2) those considered to be probable causes of pancreatitis because the incidence of disease is increased in individuals exposed to the drug; and (3) those whose relationship to pancreatitis is just suspected because only anecdotal evidence has been presented to support such a relationship. In the past, there have been claims that steroids as well as H2 -histamine antagonists can cause pancreatitis, but there is little evidence to support those claims.
Obstruction
Even in the absence of biliary tract stones, pancreatic duct obstruction can cause pancreatitis. Thus, pancreatitis has been associated with duodenal lesions such as duodenal
1650
Box 53-2. Drugs Associated with Pancreatitis Definite Cause 5-Aminosalicylate 6-Mercaptopurine Azathioprine
Cytosine arabinoside Dideoxyinosine Diuretics Estrogens Furosemide Metronidazole Pentamidine Tetracycline Thiazide Trimethoprim-sulfamethoxide Valproic acid Probable Cause Acetaminophen -Methyl-DOPA Isoniazid
L
-Asparaginase
Phenformin Procainamide Sulindac DOPA, dihydroxyphenylalanine.
ulcers, duodenal Crohns disease, and periampullary tumors. It can also be triggered by a periampullary diverticulum, particularly if that diverticulum is filled with debris or food particles. Pancreatitis can also be the result of a pancreatic duct stricture or disruption following blunt pancreatic trauma or duct obstruction caused by a pancreatic tumor. Most patients with obstruction-induced pancreatitis have chronic, rather than acute, pancreatitis. That chronic pancreatitis affects only the obstructed portion of the pancreas, and it can be cured by removing that part of the pancreas. Post-traumatic strictures, the result of blunt abdominal trauma, can cause pancreatitis. They usually occur where the pancreas passes over the vertebral column. Parasites such as Ascaris and Clonorchis can also cause pancreatitis by obstructing the pancreatic duct. Pancreas divisum has also been described as a cause of obstructive pancreatitis. Presumably, that pancreatitis results from the relative obstruction that might occur at the lesser papilla when the major fraction of pancreatic secretion is forced to exit via that orifice. Most observers believe that pancreas divisum, which occurs in roughly 10% of individuals, is rarely the cause of pancreatitis.
Hereditary and Autoimmune Pancreatitis
There has been considerable recent interest in the few patients who develop pancreatitis on a hereditary basis.[9] It is generally believed that spontaneous trypsinogen activation normally occurs to a slight degree within the pancreas but that, in health, the pancreas is protected from injury by the presence of trypsin inhibitors. In hereditary pancreatitis, genetic mutations are believed to cause this protective process to fail either because a trypsin that is resistant to inhibition is synthesized or because the trypsin inhibitors themselves are defective. In either case, the end result could be expected to be further intrapancreatic activation of trypsin and, possibly, other digestive enzymes, eventually leading to repeated episodes of pancreatitis. In hereditary pancreatitis, those attacks begin at a young age and lead to chronic changes including fibrosis, calcifications, and loss of both exocrine and endocrine function. The incidence of pancreatic cancer is also markedly increased in patients with hereditary pancreatitis. Hereditary pancreatitis is an autosomal dominant disease with incomplete penetrance. Pancreatic cancer most frequently develops in those with a paternal pattern of inheritance. Pancreatitis can be the result of an autoimmune process, and, in those patients, it is frequently associated with other autoimmune diseases such as primary sclerosing cholangitis, Sjgrens syndrome, and primary biliary cirrhosis. Recently, a distinct form of autoimmune pancreatitis has been described in which there is a severe, sclerosing process characterized by intense lymphocyte and plasmacyte infiltration. It is frequently associated with bile as well as pancreatic duct strictures, pancreatic inflammation, and a pancreatic mass. Some refer to the disease as lymphoplasmacytic autoimmune pancreatitis. Most patients with this form of pancreatitis have elevated circulating IgG levels, and that elevation is mostly due to an elevation in the levels of IgG4. [10] Perhaps the most important feature of this form of autoimmune pancreatitis is the fact that it frequently presents as an otherwise unexplained mass in the head of the pancreas and the sclerotic process can result in bile as well as pancreatic duct strictures (i.e., a double-duct sign). Thus, it can easily be confused with pancreatic cancer. If correctly diagnosed, the mass as well as the strictures can completely resolve with steroid treatment.
Other Miscellaneous Causes of Pancreatitis
Pancreatitis can result from pancreatic trauma even without major duct disruption or stricture. In those cases, the inflammatory process is usually related to contusion or laceration of the gland and possibly disruption of small ducts. Pancreatitis can occur during the postoperative period in patients undergoing procedures on or near the pancreas or procedures associated with either hypoperfusion or atheroembolism (cardiopulmonary bypass, cardiac transplantation, renal transplantation). The injection of the pancreatic duct that occurs during endoscopic retrograde pancreatography or during sphincter of Oddi manometry can also cause pancreatitis. Both acute and chronic pancreatitis can be caused by metabolic abnormalities, especially those leading
to hypercalcemia (i.e., hyperparathyroidism) and those leading to hyperlipidemia (type I, IV, or V hyperlipoproteinemias). Hypercalcemia
1651
may result in pancreatitis by facilitating intrapancreatic digestive enzyme activation. Hyperlipidemia may lead to pancreatitis if the accompanying hyperchylomicronemia interferes with the pancreatic microcirculation or results in release of free fatty acids in the pancreatic microcirculation. Both hypercalcemia-induced pancreatitis and hyperlipidemia-induced pancreatitis can be prevented if the underlying metabolic abnormality is corrected by either parathyroidectomy or by drug and dietary management of the hyperlipidemia. In places such as Trinidad, scorpion stings are a frequent cause of pancreatitis. Scorpion toxin contains a potent pancreatic secretagogue and, presumably, the excessive pancreatic stimulation that follows exposure to this toxin leads to pancreatic injury.
Idiopathic Pancreatitis
In most series, roughly 20% of patients have pancreatitis without an identifiable etiology. Some of those individuals have gallbladder sludge or microcrystals, and further attacks can be prevented by either cholecystectomy or biliary sphincterotomy. Other patients have been found to have sphincter of Oddi malfunction, sometimes associated with increased pressures in the pancreatic duct system, and they can be effectively treated by sphincterotomy with pancreatic septotomy. Therefore, these patients have forms of biliary pancreatitis rather than truly idiopathic pancreatitis. There remains, however, a significant group of patients with no identifiable cause for their pancreatitis. Recent studies by several independent groups have suggested that some of these patients may have subclinical mutations of the cystic fibrosis gene.[11] In its most severe form, cystic fibrosis can cause pancreatic fibrosis and the loss of both exocrine and endocrine function as a result of the blockade of ducts with inspissated secretions. However, the patients with idiopathic pancreatitis related to cystic fibrosis have subclinical mutations of the cystic fibrosis transmembrane regulator gene. They do not have inspissated secretion and their pancreatitis probably develops on another basis.
Pathophysiology of Acute and Chronic Pancreatitis
It is generally believed that acute pancreatitis is triggered by obstruction of the pancreatic duct and that the injury begins within pancreatic acinar cells. That injury is believed to include, and possibly be the result of, intra-acinar cell activation of digestive enzyme zymogens including trypsinogen. Chronic pancreatitis is believed to reflect repeated episodes of subclinical acute pancreatitis with unrecognized pancreatic necrosis evolving into pancreatic fibrosis. One of the central issues in our understanding of the cellular events leading to acute pancreatitis is how duct obstruction could result in intra-acinar cell enzyme activation. Perhaps one of the most widely accepted theories to explain this coupling is the so-called colocalization hypothesis.[12] This hypothesis is based on a number of studies that have used experimental models of pancreatitis induced in laboratory animals. In those studies, one of the earliest changes noted has been the colocalization of digestive enzyme zymogens such as trypsinogen with lysosomal hydrolases such as cathepsin B inside cytoplasmic vacuoles. Under these conditions, cathepsin B can activate trypsinogen and trypsin can activate the other zymogens. According to the colocalization hypothesis, cathepsin Bmediated intra-acinar cell activation of the digestive enzymes leads to acinar cell injury and triggers an intrapancreatic inflammatory response. The intensity of that inflammatory response appears to regulate the severity of the pancreatitis and to couple pancreatitis to extrapancreatic events such as lung and renal injury.
Presentation of an Acute Attack
The clinical presentation, diagnosis, and management of an acute attack of pancreatitis are similar regardless of whether that attack is acute or chronic pancreatitis. In fact, many describe patients with chronic pancreatitis who present with acute symptoms as having acute on chronic pancreatitis. On the other hand, the long-term management of patients with acute and chronic pancreatitis may differ considerably. The former primarily involves elimination of the inciting cause whereas, for chronic pancreatitis, irreversible changes have usually occurred prior to diagnosis and long-term management primarily involves treatment of pain and pancreatic exocrine/endocrine insufficiency. For these reasons, this discussion of clinical presentation focuses on issues relevant to an acute attack and does not make distinctions based on whether that is an attack of acute or chronic pancreatitis.
Symptoms
Abdominal pain, nausea, and vomiting are the dominant symptoms of pancreatitis. Typically, the pain is located in the epigastrium, but it may also involve both upper quadrants, the lower abdomen, or the lower chest. It may have a pleuritic component and be felt in one or both shoulders. Most patients describe the pain as being knifelike and radiating straight through to the mid-central back. It is usually abrupt in onset and slowly increases in magnitude to reach a maximal level. The pain is usually constant, although it may be somewhat relieved by leaning forward or lying on the side with the knees drawn upward. Patients with chronic pancreatitis frequently describe similar prior attacks that are often noted to occur within 12 to 24 hours of ethanol consumption. The nausea and vomiting of pancreatitis usually persists even after the stomach has been emptied. The vomiting may lead to gastroesophageal tears (i.e., Mallory-Weiss syndrome) and upper gastrointestinal bleeding. Although vomiting and retching may be relieved by passage of a nasogastric tube, the pain usually persists even after gastric decompression. Some patients, especially those with postoperative pancreatitis who are already receiving analgesic medications, may not experience abdominal pain and, therefore, the diagnosis of pancreatitis may be particularly difficult.
1652 Physical Findings
Pancreatitis patients are frequently noted to be rolling or moving around in search of a more comfortable position and, in this sense, they are unlike patients with a perforated viscus who often remain motionless because movement worsens their pain. Patients with severe pancreatitis usually appear ill and anxious. Hyperthermia is common and may be explained by the release of proinflammatory factors, including cytokines and chemokines, from the injured pancreas. Tachycardia, tachypnea, and hypotension caused by hypovolemia are common. Hypovolemia can also result in collapsed neck veins, dry skin, dry mucous membranes, and diminished subcutaneous elasticity. Because pleuritic and abdominal pain may make breathing difficult, breath sounds in the lower lung fields are usually diminished and atelectasis may be present. A pleural effusion can often be detected on either side, although it is more commonly found on the left. Patients with severe pancreatitis frequently develop an acute lung injury that can clinically present as the adult respiratory distress syndrome (ARDS). Occasionally, patients with pancreatitis have alterations in their mental status as a result of drug or ethanol exposure, hypotension, hypoxemia, or release of circulating toxic agents from the inflamed pancreas. Some degree of jaundice is common. In gallstoneinduced acute pancreatitis, the jaundice may reflect distal bile duct obstruction, but jaundice can also occur in nonbiliary pancreatitis either as a result of duct obstruction caused by the inflamed pancreas or as a result of cholestasis induced by the severe illness itself. As a result of ileus, bowel sounds are usually diminished during an attack of pancreatitis and the abdomen may become distended and tympanitic. Direct, percussion, and rebound abdominal tenderness as well as both voluntary and involuntary guarding are common. These findings may be localized to the epigastrium or they may be diffusely present throughout the abdomen. An epigastric mass, reflecting the inflamed pancreas and surrounding tissues, may be felt in the upper abdomen or left
upper quadrant. On rare occasions, flank ecchymoses (Grey Turners sign) or periumbilical ecchymoses (Cullens sign), which result from retroperitoneal hemorrhage, can be seen during severe pancreatitis. Occasionally, patients develop areas of tender subcutaneous induration and erythema that resemble erythema nodosum but that, in the case of pancreatitis, are caused by subcutaneous fat necrosis.
Diagnosis
Routine Blood Tests
Pancreatitis can induce a diffuse capillary leak syndrome that, when combined with vomiting, can result in significant fluid losses. The resulting hypovolemia can be marked. It usually leads to an increased hematocrit, hemoglobin, blood urea nitrogen, and creatinine. Serum albumin levels may be markedly depressed, particularly if fluid losses are corrected by administration of albumin-free crystalloid solutions. The serum electrolytes may be normal but, with significant vomiting, a hypochloremic metabolic alkalosis can develop. The white blood cell count is usually elevated with an associated left shift in the differential count. Blood glucose may be elevated either due to associated diabetes mellitus or as a result of increased glucagon and catecholamine release combined with diminished insulin release. Hyperbilirubinemia is relatively common during the early stages of pancreatitis. It can be caused by either a biliary tract stone or by the inflamed (and possibly fibrotic) pancreas causing bile duct obstruction and, in this setting, cholangitis with positive blood cultures can be superimposed on the pancreatitis. On the other hand, the hyperbilirubinemia of pancreatitis can also reflect the nonobstructive cholestasis that often accompanies any severe illness. Hypertriglyceridemia is routinely noted in patients who have hyperlipidemia-induced pancreatitis. Hypertriglyceridemia can also be induced by exposure to ethanol and therefore, the diagnosis of pancreatitis should always be suspected when lactescent serum is found when evaluating an alcoholic patient with abdominal pain. Many patients with pancreatitis appear to have hypocalcemia but, for the most part, that hypocalcemia can be explained by the hypoalbuminemia that accompanies pancreatitis. Occasionally however, patients with severe pancreatitis have a reduction in their free, ionized calcium that is not a reflection of hypoalbuminemia. This type of hypocalcemia is associated with a poor prognosis. Some of these patients manifest tetany and carpopedal spasm, making treatment with calcium mandatory. The mechanisms responsible for this type of pancreatitis-associated hypocalcemia are not clear. Most likely, it occurs because bone calcium stores do not respond to circulating parathormone. Patients with severe pancreatitis can also develop disseminated intravascular coagulation. In those cases, thrombocytopenia, elevated levels of fibrin degradation products, a decreased fibrinogen level, prolonged partial thromboplastin time, and a prolonged prothrombin time can be observed.
Amylase Measurement
Serum amylase activity is usually, but not always, elevated during pancreatitis, but the magnitude of that elevation does not parallel the severity of the attack. In fact, as many as 10% of patients with lethal pancreatitis may have near-normal or normal amylase levels. This could reflect the fact that pancreatitis-associated hyperamylasemia can be transient. Typically, amylase levels rise 2 to 12 hours after the onset of symptoms and then decline so that, 3 to 6 days after the onset of an attack, the serum amylase levels are usually normal. Elevations that persist beyond a week suggest either ongoing inflammation or the development of a complication such as pseudocyst, abscess, or pancreatic ascites. Urinary amylase levels remain elevated longer than serum amylase levels and, thus, measurement of urinary amylase levels may be of diagnostic help in patients who present long after the onset of symptoms. Although amylase can enter the circulation from nonpancreatic sites, including the salivary glands, lung, prostate, and ovary, it is pancreatic amylase that
1653
accounts for the rise in circulating amylase activity during pancreatitis. The mechanisms responsible for the hyperamylasemia of pancreatitis are not clear. Some have suggested that, during pancreatitis, amylase and other digestive enzymes may be secreted from the basolateral, as opposed to the apical, surface of acinar cells and, in this manner, they could gain access to the lymphatic and vascular system. On the other hand, some recent studies have indicated that cell-cell contacts are loosened during pancreatitis.[13] This could allow enzymes in the duct to reach periacinar, lymphatic, and intravascular spaces. The overall sensitivity and specificity of serum amylase determination in the diagnosis of pancreatitis depend on both the clinical presentation and the cut-off value chosen for the upper limit of normal. In some series, sensitivity and specificity values in the low-to-mid 90% range have been reported. Hyperamylasemia can be associated with acute cholecystitis, perforated viscus, bowel obstruction, and bowel infarction. These states can also be clinically confused with pancreatitis since they are also characterized by abdominal pain, nausea, vomiting, and abdominal tenderness. In most cases, patients with hyperamylasemia that is not due to pancreatitis have only mild elevations in the circulating amylase level (i.e., twofold to threefold elevations from the normal value), whereas those with pancreatitis usually have greater elevations. At one time, measurement of the clearance ratio between amylase and creatinine was advocated as a method by which pancreatitis-associated hyperamylasemia could be distinguished from non pancreatitis-associated elevations of amylase activity. It was claimed that an elevated clearance rate was diagnostic of pancreatitis. Unfortunately, the so-called amylase-to-creatinine clearance ratio has not proven to be clinically useful since changes in this ratio are not specific to pancreatitis and elevated clearance ratios can be seen in other diseases. Occasionally, serum amylase activity may be normal, even during the early stages of an attack. The basis for this phenomenon is not certain. In some cases, it may reflect overwhelming necrosis of the gland. Some patients with acute pancreatitis superimposed on advanced chronic pancreatitis do not develop hyperamylasemia because there is little residual and functional pancreatic exocrine tissue and, therefore, little pancreatic amylase to be released into the circulation. In some patients with hyperlipidemia-induced pancreatitis, hyperamylasemia may be masked by circulating amylase inhibitors. This is particularly true of patients with lactescent serum in whom the circulating amylase activity may appear to be normal. Macroamylasemia is a form of pancreatitis-independent hyperamylasemia that affects 0.5% of individuals. It occurs when amylase binds to an abnormal circulating albumin-like protein. Because of its large size, this protein prevents the normal clearance of amylase and, as a result, circulating levels of amylase rise. In some patients, episodes of abdominal pain can also occur raising the suspicion that they have pancreatitis. In this situation, macroamylasemia can be distinguished from the hyperamylasemia of pancreatitis by simple measurement of urinary amylase activity. In the former, urinary amylase levels are very low.
Other Blood Tests
In addition to amylase, other enzymes and inflammatory mediators are released into the circulation during pancreatitis, and many have been the target of diagnostic or prognostic tests. Circulating lipase levels usually increase during pancreatitis. That increase usually parallels the rise in amylase activity, but the elevations of lipase activity may persist even after amylase activity has returned to normal. Thus, serum lipase measurement may be particularly helpful when patients are first seen several days after the onset of symptoms. Circulating levels of other pancreatic enzymes including trypsinogen, phospholipase, elastase, and chymotrypsinogen increase during pancreatitis, but measurement of these circulating enzymes is usually not performed since they add little to the information gained by the easier and more straightforward measurement of amylase activity. The activation peptides released during either trypsinogen, procarboxypeptidase, or prophospholipase activation are increased in the urine of patients with acute pancreatitis, and several studies have indicated that measurement of those activation peptides may aid in predicting the severity of an attack.[14] Although methemalbumin levels sometimes rise during attacks of severe pancreatitis, and methemalbuminemia is indicative of a poor prognosis, methemalbumin levels are usually not measured. Circulating levels of several inflammatory mediators and acute-phase reactants (e.g., interleukin [IL]-1, IL-6, tumor necrosis factor-alpha,
and C-reactive protein) also increase during pancreatitis, and the magnitude of those increases can be used to predict the severity of an attack.
Imaging Studies
In general, the plain chest and abdominal radiographs are not particularly helpful in the diagnosis of pancreatitis, although they may be useful in patient management by revealing other causes for the patients symptoms (e.g., pneumonia, perforated hollow viscus, mechanical bowel obstruction). In patients with pancreatitis, radiographs of the chest frequently reveal basal atelectasis and elevation of the diaphragm caused by splinting of respiration. Pleural effusions, most common on the left, can also be seen. Plain abdominal films usually show the gas pattern of a paralytic ileus but, occasionally, retroperitoneal gas bubbles indicating infection with gas-forming organisms can be seen. Pancreatic calcifications that are pathognomonic of chronic pancreatitis and are caused by the formation of calcified intraductal protein plugs may be seen on the routine abdominal films. Transcutaneous ultrasound may be useful in demonstrating the presence of gallbladder stones and/or dilated bile ducts, but ultrasound examination has limited value because of the presence of intestinal gas in the upper abdomen. Computed tomography (CT) has been shown to be particularly helpful in the diagnosis and management of
1654
patients with pancreatitis. During the early stages of an attack, CT can image the upper abdomen and pancreas without being obscured by overlying or surrounding intestinal gas. When combined with bolus administration of intravenous contrast material, helical CT can detect the subtle changes of mild pancreatitis (i.e., pancreatic swelling and edema) as well as the changes of more severe pancreatitis (i.e., varying degrees of pancreatic necrosis and the presence of peripancreatic or intrapancreatic fluid collections). Both clinical and experimental studies have demonstrated the close parallel that exists between nonperfused pancreas seen on CT examination and necrosis seen on morphologic examination of the pancreas. At later times during the evolution of an attack, CT can be used to detect and follow pseudocysts and to permit fine-needle aspiration of areas suspected of harboring pancreatic infection. The timing of CT during an attack of pancreatitis is a matter of considerable controversy. One study, using an experimental model of pancreatitis in rodents, suggested that early CT with administration of intravenous contrast material could adversely affect the course of pancreatitis and worsen outcome,[15] but this conclusion has not been borne out by other studies,[16] and, at present, it is generally believed that early performance of contrast-enhanced CT does not worsen pancreatitis. On the other hand, there may be little or no value in obtaining a CT in patients with obvious pancreatitis since early CT is unlikely to alter treatment. Early CT may be particularly helpful when the diagnosis of pancreatitis is in doubt. A normal pancreas imaged in a patient thought to have severe pancreatitis would prompt further diagnostic studies. Magnetic resonance imaging (MRI), which has the same sensitivity and specificity as CT in pancreatitis, has also been used in these patients. It provides information that is similar to that obtained by CT but, because of its ease of interpretation and ready availability, most clinicians prefer to use CT, rather than MRI, for the diagnosis and management of patients with pancreatitis.
Differential Diagnosis
The differential diagnosis of acute pancreatitis includes any process that can cause upper abdominal pain and tenderness, nausea, and vomiting ( Box 533 ). Usually, but not always, the serum amylase and/or lipase levels are elevated in pancreatitis, but those enzymes can also be elevated in other conditions including cholecystitis/ cholangitis, perforated hollow viscus, bowel obstruction, and bowel infarction. In those patients, the CT does not Box 53-3. Differential Diagnosis of Acute Pancreatitis Bowel obstruction Cholecystitis/cholangitis Mesenteric ischemia/infarction Perforated hollow viscus
suggest pancreatitis and, for the most part, enzyme elevations in those conditions are usually only twofold or threefold above normal. Occasionally, however, it may be difficult or even impossible to be certain that the patient actually has pancreatitis and, in those cases, a diagnostic exploratory laparotomy may be indicated.
Prognosis of an Acute Attack
The ultimate severity of an attack appears to be determined by events that occur within the first 24 to 48 hours. Most patients experience only a mild self-limited illness that can be expected to resolve with only supportive care, but approximately 10% of patients experience a severe attack. Severe attacks are more common in acute pancreatitis, but they can also occur when an acute attack is superimposed on chronic pancreatitis, that is, the socalled acute on chronic pancreatitis. Severe attacks are also more common in patients older than 60 years of age; those experiencing a first attack; those with postoperative pancreatitis; and those with methemalbuminemia, hypocalcemia, Grey Turners sign, or Cullens sign. The observation that the ultimate severity is determined by events that occur during the early stages of pancreatitis has prompted several groups of investigators to undertake studies designed to determine which clinical, chemical, or radiologic parameters might be used to identify those patients destined to experience a severe illness. As a result, a number of prognostic schemes have been developed. Among the clinical scoring systems, the most widely used are those developed in New York by Ransons group[17] ( Table 531 ), and, in Glasgow, by Imries group.[18] Patients with fewer than three of the prognostic criteria can be expected to have a mild attack with little morbidity and a mortality rate of less than 1%. On the other hand, TABLE 53-1 -- Ransons Prognostic Signs Admission Gallstone Pancreatitis Age > 70 yr WBC > 18,000/mm3 Glucose > 220 mg/100 mL LDH > 40 IU/L AST > 250 U/100 mL Non-Gallstone Pancreatitis Hct fall > 10 BUN elevation > 2 mg/100 mL Ca2+ > 8 mg/100 mL Base deficit > 5 mEq/L Fluid sequestration > 4 L Initial 48 Hours
Age > 55 yr WBC > 16,000/mm3 Glucose > 200 mg/100 mL LDH > 350 IU/L AST > 250 U/100 mL
Hct fall > 10 BUN elevation > 5 mg/100 mL Ca2+ > 8 mg/100 mL Pao2 > 55 mm Hg Base deficit > 4 mEq/L Fluid sequestration > 6 L
WBC, white blood count; LDH, lactic dehydrogenase; AST, aspartate transaminase; HCT, hematocrit; BUN, blood urea nitrogen; Ca2+ , calcium; Pao2 , arterial oxygen. Adapted from Ranson JHC, Rifkind KM, Roses DF, et al: Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet 139:6981, 1974 and Ranson JHC. Etiological and prognostic factors in human acute pancreatitis: A review. Am J Gastroenterol 77:633, 1982.
1655
with the presence of more prognostic factors, increased morbidity and mortality can be expected so that, with three or four of Ransons criteria, the mortality rate may reach 15%, and 50% of patients may need to be treated in an intensive care unit. Most patients with five or six signs will require intensive care and, with seven or eight of Ransons signs, the mortality rate may reach 90%. As an alternative to using clinical criteria, Balthazar and coworkers developed radiologic criteria for predicting a severe attack. In a prospective study employing contrast-enhanced CT examination,[19] they noted that the severity of an attack was related to the number of pancreatic fluid collections and the extent of pancreatic nonperfusion (i.e., necrosis) seen on CT examination. In addition to clinical and radiologic criteria, high levels of certain circulating factors can also be used to predict the evolution of a severe attack. Those factors include the following: Creactive protein, phospholipase A2 , polymorphonuclear elastase, immunoreactive trypsin, IL6, and pancreatitis-associated protein. High urinary levels of the activation peptides for trypsinogen, procarboxypeptidase, and prophospholipase also indicate a severe attack. The second version of the Acute Physiology and Chronic Health Evaluation (APACHE-II) scoring system has also been used to predict the severity of a pancreatitis attack. An APACHE-II score of 8 or more is generally indicative of a severe attack. The APACHE-II scoring system has the advantage of continually quantifying disease severity. Although the APACHE-II system can be used at the time of admission, recent studies have suggested that an admission score that worsens over the initial 48 hours of hospitalization in spite of aggressive treatment, or the score itself 48 hours after admission, may be particularly accurate in predicting the severity of the attack and a poor outcome.[20] Although each of these scoring systems may predict the severity of an attack, there is also evidence that a good examination by an experienced clinician can accurately discriminate between mild and severe pancreatitis. Furthermore, none of the prognostic schemes are intended for use as a diagnostic tool in pancreatitis. Their ultimate value is in triaging patients to their appropriate care settings. In addition, they may be useful in clinical studies by permitting comparison of therapeutic outcomes for comparable patients stratified to different treatments.
Treatment of an Acute Attack
An acute attack of pancreatitis evolves in two, somewhat overlapping, phases. The initial phase, which lasts for 1 to 2 weeks, involves an acute inflammatory and autodigestive process that takes place within and around the pancreas. It may have systemic effects as well. In patients with severe pancreatitis, this initial phase of pancreatitis seamlessly evolves into a later phase that may last for weeks or months. This later phase of pancreatitis is primarily characterized by the development of local complications that are, themselves, the result of necrosis, infection, and pancreatic duct rupture.
Initial Treatment
The initial management of patients with pancreatitis should focus on establishing the diagnosis, estimating its severity, addressing the major symptoms (i.e., pain, nausea, vomiting, and hypovolemia), and limiting its progression. Ideally, the diagnosis should be established without exploratory surgery since exploration may increase the incidence of later pancreatic infection. On occasion, however, exploration may be required to establish the diagnosis with certainty, especially when the diagnosis is uncertain and the patient has not responded favorably to aggressive nonoperative treatment. For the most part, patients with predicted severe pancreatitis should be treated in an intensive care setting since it is in this group that fluid and respiratory management may be particularly challenging and both morbidity and mortality are, essentially, confined to this group.
Management of Pain
The pain of pancreatitis may be severe and difficult to control. Most patients require narcotic medications. Meperidine or its analogues are probably preferable to morphine in this setting since morphine can induce spasm of the sphincter of Oddi and that could, at least theoretically, worsen biliary pancreatitis.
Fluid and Electrolyte Management
Aggressive fluid and electrolyte repletion is the most important element in the initial management of pancreatitis. Fluid losses can be enormous and can lead to marked hemoconcentration as well as hypovolemia. Inadequate fluid resuscitation during the early stages of pancreatitis can worsen the severity of an attack and lead to subsequent complications. The fluid depletion that occurs in pancreatitis results from the additive effects of losing fluid both externally and internally. The external fluid losses are caused by repeated episodes of vomiting and worsen by nausea that limits fluid intake. Repeated vomiting can result in a hypochloremic alkalosis. Internal fluid losses, which are usually even greater than the external losses, are caused by fluid sequestration into areas of inflammation (i.e., the peripancreatic retroperitoneum) and into the pulmonary parenchyma and soft tissues elsewhere in the body. These latter losses result from the diffuse capillary leak phenomenon that is triggered by proinflammatory factors released during pancreatitis. Total fluid losses may be so great that they lead to hypovolemia and hypoperfusion and, as a result, a metabolic acidosis can develop. Many of the patients with chronic pancreatitis are alcoholics who, even before the onset of pancreatitis, had hypoalbuminemia and hypomagnesemia. Those problems are exacerbated by the losses of pancreatitis. The measured values for serum albumin may be even further depressed as fluid losses are treated with albumin-free crystalloid solutions. Although hypocalcemia is common particularly during a severe attack, the low total serum calcium is usually attributable to the low levels of circulating albumin and no treatment is needed when ionized calcium is normal. Occasionally, however, ionized
1656
calcium levels may also be depressed and tetany as well as carpopedal spasm can occur. Under those circumstances, aggressive calcium repletion is indicated.
During the first several days of a severe attack, circulating levels of many proinflammatory factors, including cytokines and chemokines, are elevated. This cytokine storm, in many cases, triggers the systemic immune response syndrome and, as a result, the hemodynamic parameters of these patients may resemble those of sepsis associated with other disease states. Heart rate, cardiac output, and cardiac index usually rise and total peripheral resistance falls. Hypoxemia can also occur as a result of the combined effects of increased intrapulmonary shunting and a pancreatitis-associated lung injury that closely resembles that seen in other forms of the ARDS. Fluid management, although critical, may be particularly difficult when hypovolemia is combined with the respiratory failure of ARDS. Treatment requires meticulous replacement of fluid and electrolyte losses. A fluid balance flow sheet is helpful, but parameters such as pulse rate, blood pressure, oxygen saturation, and urine output are notoriously unreliable for determining fluid needs in this setting. The hematocrit, however, can be quite useful as increased levels usually are accurate indicators of the magnitude of extracellular fluid loss. However, in a setting of blood loss or hemolysis, hematocrit measurements may lose their value in fluid management. Measurement of central filling pressures, using a Swan-Ganz or central venous pressure catheter, can be helpful in guiding fluid management, particularly when hypovolemia is combined with lung injury.
Role of Nasogastric Decompression
The nausea and vomiting of pancreatitis can result in significant fluid as well as electrolyte losses. Furthermore, retching can lead to gastroesophageal mucosal tears and result in upper gastrointestinal bleeding (i.e., the Mallory-Weiss syndrome). To increase patient comfort, nasogastric decompression may be needed, although the institution of nasogastric drainage has not been shown to alter the eventual outcome of an attack.
Role of Prophylactic Antibiotics
Over the past decade, three separate studies have indicated that prophylactic antibiotics are useful in the management of patients with severe pancreatitis [21] [22] [23] but no benefit was observed when prophylactic antibiotics were given to patients with mild pancreatitis. They postulated that patients with mild pancreatitis almost invariably recover quickly and without infectious complications. In patients with severe pancreatitis, benefit was observed with regimens that included imipenem alone, imipenem with cilastatin, and cefuroxime. Selective gut decontamination with a combination of norfloxacin, colistin, and amphotericin has also been found to be beneficial,[24] although that approach is labor intensive and not readily available. Although these recent studies argue strongly for administration of prophylactic antibiotics to patients with severe pancreatitis, there is an opposing view that is becoming increasingly widespread. According to that school of thought, administration of prophylactic antibiotics favors emergence of resistant organisms in the area of pancreatic injury. That may be particularly true for fungal strains such as Candida. Some have advocated adding antifungal agents such as fluconazole to the prophylactic antibiotic regimen, whereas others have argued that, because of the risk of infection with resistant organisms or fungi, antibiotics should not be prophylactically used in the management of a severe attack.
Nutritional Support
Patients with severe pancreatitis may be unable to eat for prolonged periods and an alternative route for providing nutrition is required. Traditionally, these patients have been given parenteral nutrition administered via a central venous catheter. Widely differing opinions exist regarding the time that total parenteral nutrition should be started. Some advocate starting within the first day or two, whereas others delay starting total parenteral nutrition until the early phase of pancreatitis, characterized by extensive fluid shifts and high fluid requirements, has been completed. I favor the latter approach. Several investigative groups have recently demonstrated that most patients with pancreatitis, including those with severe pancreatitis, can actually tolerate small amounts of enterally administered nutrients. They have shown that those nutrients can be tolerated if given either into the stomach (via a nasogastric tube) or into the small intestine (via a nasojejunal tube). Pancreatic infections are believed to occur because gut bacteria are translocated across the injured bowel wall adjacent to areas of pancreatic injury. Theoretically, enteral nutrition exerts a trophic effect on the injured bowel wall that could reduce this translocation and, thus, reduce the incidence of pancreatic infections. Studies evaluating this concept are currently underway, but, even in the absence of definitive results, I favor administration of trophic amounts of nutrients to patients with severe pancreatitis and begin that treatment within 72 hours of hospitalization.
Treatments of Limited or Unproved Value
Peritoneal dialysis, designed to eliminate the proinflammatory factors released into the abdomen during pancreatitis, might theoretically be expected to reduce the severity of pancreatitis. Indeed, early anecdotal studies did support the use of peritoneal dialysis in patients with severe pancreatitis, but a more recent, prospective, randomized multi-institutional study showed that peritoneal dialysis was of no benefit. Nasogastric decompression does not appear to alter the course or outcome of a pancreatitis attack, although it may provide for greater patient comfort during the early stages when nausea and vomiting are common. Other attempts to reduce gastrointestinal and/or pancreatic secretion (i.e., H2 blockers, proton-pump inhibitors, antacids, atropine, somatostatin, glucagon, calcitonin) have not been shown to be beneficial in the treatment of pancreatitis. Similarly, the use of antiinflammatory agents (i.e., steroids, prostaglandins, and indomethacin) has not been helpful although recent experimental studies have suggested that specific inhibition
1657
of cyclooxygenase-2 might be beneficial. Many attempts to treat pancreatitis with agents designed to inhibit activated proteolytic enzymes (e.g., aprotinin, gabexate mesylate) have failed to alter the course of pancreatitis unless their use is begun prior to the onset of the attack. Hypothermia, thoracic duct drainage, and plasmapheresis have been evaluated in experimental models of pancreatitis but, to date, there is little evidence that these modes of therapy are clinically useful. Many other approaches have also been tried (e.g., procainamide, isoproterenol, heparin, dextran, vasopressin). Although these forms of treatment have been supported by experimental animal studies, particularly when the treatment is begun prior to the onset of pancreatitis, human clinical trials have failed to show a beneficial effect on the course of patients with established pancreatitis and, at present, none of these treatments are commonly employed. Platelet-activating factor (PAF) is a proinflammatory factor that has been shown to promote worsening of experimental pancreatitis in animal models. Recently, several clinical trials have evaluated the effects of interfering with PAF action during severe clinical pancreatitis, but they have failed to show these anti-PAF agents beneficially alter the outcome of patients with severe pancreatitis. Therefore, anti-PAF agents are not currently used.
Treatment of Early Systemic Complications of Pancreatitis
The pathogenesis and management of the cardiovascular collapse, respiratory failure, renal failure, metabolic encephalopathy, gastrointestinal bleeding, and disseminated intravascular coagulation that complicate severe pancreatitis appear to be identical to those involved when these processes are superimposed on other disease states that are characterized by peritonitis and hypovolemia. Cardiovascular collapse is largely caused by hypovolemia, and its management requires aggressive fluid and electrolyte repletion. This may necessitate placement of a central venous or Swan-Ganz monitoring catheter. Changes in hematocrit, filling pressures, and cardiac output can be used to monitor the adequacy of treatment, but changes in blood pressure, pulse, and urine output do not accurately and reliably reflect the adequacy of fluid replacement. The pulmonary manifestations of pancreatitis include atelectasis and acute lung injury. The latter appears to be similar to the acute lung injury caused by
other systemic processes including septic shock, ischemia/reperfusion, and massive blood transfusion. Management includes good pulmonary toilet combined with close monitoring of pulmonary function. For many patients, intubation and respiratory support may be required. Renal failure in pancreatitis is usually prerenal and is associated with a poor prognosis. In severe cases, dialysis, usually hemodialysis, may be required. Stress-induced gastroduodenal erosions account for most of the gastrointestinal bleeding in pancreatitis and prophylaxis with antacids, H2 receptor antagonists, or proton-pump inhibitors may be appropriate. Rarely, massive bleeding can result from injury to peripancreatic vascular structures leading to hemorrhage into the retroperitoneum. The peripancreatic inflammatory process can also cause thrombosis of major gastrointestinal vessels and result in ischemic lesions involving the stomach, small intestine, or colon that can cause bleeding. Management of these complications of pancreatitis is similar to that involved when they occur in the absence of pancreatitis. Some patients with severe pancreatitis develop disseminated intravascular coagulation, but it rarely causes bleeding and prophylactic heparinization is usually not indicated.
Role of Early Endoscopy and Stone Extraction
Patients with mild pancreatitis may ultimately require endoscopic duct clearance to prevent recurrent attacks, but they rarely benefit from early endoscopy because their pancreatitis generally resolves spontaneously within several days. On the other hand, the role of early endoscopic duct clearance in the initial management of patients with severe biliary pancreatitis is more controversial. Three randomized, controlled, prospective studies have differing results.[25] [26] [27] One study indicated that early stone clearance reduced the severity and mortality of biliary pancreatitis, whereas a second study indicated that early duct clearance reduced the incidence of infectious complications. The third study concluded that early endoscopy and duct clearance actually adversely affected the course of pancreatitis because it was associated with a high incidence of complications. At the present time, most experts would favor early (i.e., < 48 hours after the onset of symptoms) endoscopic intervention in severe biliary pancreatitis, but further studies are needed.
Role and Timing of Cholecystectomy in Patients with Gallstone Pancreatitis
In general, patients with gallstone pancreatitis should undergo some form of definitive treatment prior to discharge from the hospital, and that intervention should take place as soon as possible after resolution of their attack. Further delaying the intervention would increase the chances that additional stones might be passed and another attack of pancreatitis might be triggered. Intervening sooner, on the other hand, could introduce infection into the inflamed peripancreatic area and/or worsen the pancreatitis. For purposes of therapeutic decision making, patients with gallstone pancreatitis can be divided into two groups: those who have or have had gallbladder-derived problems (cholecystitis or biliary colic) and those whose only problems are purely related to stones in the biliary ductal system (i.e., cholangitis and pancreatitis). Patients in the first group should undergo cholecystectomy since that operation will prevent additional gallbladder attacks as well as eliminating the source of stones that might trigger another attack of pancreatitis. Patients in the second group, however, do not necessarily require cholecystectomy since their problem relates only to ductal stones. Theoretically they could be treated simply by endoscopic stone clearance combined with endoscopic sphincterotomy so that future stones are passed without becoming impacted in the ampulla and triggering either
1658
pancreatitis or cholangitis. Indeed, for poor surgical risk patients, the endoscopic approach is generally recommended. On the other hand, roughly 25% of patients treated in this manner will go on to develop gallbladder symptoms over the next 3 to 5 years. Thus, good surgical risk patients are better managed by cholecystectomy.
Treatment of Later Complications Definitions
In 1992, an international symposium was held to resolve the confusion that had arisen concerning the terminology used to describe the local complications of pancreatitis and the value of specific treatments for those complications.[28] The following definitions were agreed on at that conference: 1. Acute fluid collections. These occur during the early stages of severe pancreatitis in 30% to 50% of patients, lack a wall of granulation or fibrous tissue, and more than half regress spontaneously. Most are peripancreatic, but some are intrapancreatic. Those that do not regress may evolve into pseudocysts or involve areas of necrosis. 2. Pancreatic and peripancreatic necrosis. These are areas of nonviable pancreatic or peripancreatic tissue that may be either sterile or infected. They typically include areas of fat necrosis, and the necrotic tissue has a putty-like or paste-like consistency. Some necrotic regions may evolve into pseudocysts, whereas others may be replaced by fibrous tissue. 3. Pancreatic pseudocyst. These are collections of pancreatic juice, usually rich in digestive enzymes, that are enclosed by a nonepithelialized wall composed of fibrous and granulation tissue ( Fig. 534 ). Pseudocysts can be intrapancreatic but are more commonly extrapancreatic and occupy the lesser peritoneal sac. Pseudocysts are usually round or oval in shape and are not present before 4 to 6 weeks after the onset of an attack. Prior to that time, the fluid collection lacks a defined wall and is usually either an acute fluid collection or a localized area of necrosis (see earlier). Pseudocysts may be colonized by microorganisms, but infection, as evidenced by the presence of pus, is less common. When pus is present, the infected pseudocyst is referred to as a pancreatic abscess. Leakage or rupture of a pseudocyst into the peritoneal cavity results in pancreatic ascites. A pancreaticopleural fistula results from erosion of a pseudocyst into the pleural space. 4. Pancreatic abscess. These are circumscribed intra-abdominal collections of pus, usually in proximity to the pancreas, which contain little or no necrotic tissue but arise as a consequence of pancreatitis. An infected pseudocyst should be considered a pancreatic abscess. Pancreatic abscess and infected pancreatic necrosis represent the extremes of a spectrum that include lesions with varying amounts of necrosis. Thus, in a pancreatic abscess, there is little necrosis and the material has a liquid consistency, whereas in infected pancreatic necrosis, necrosis predominates and the material is paste or putty-like.
Figure 53-4 CT scan of pancreatic pseudocyst.
Diagnosis
Contrast-enhanced CT is particularly valuable as a means of quantifying the extent of pancreatic necrosis (i.e., nonenhancement). The maturation of a pseudocyst can be followed by both contrast-enhanced CT and endoscopic ultrasound. Management of local pancreatitis complications is dependent on whether the lesion is sterile or infected (see later). Occasionally, infection can be diagnosed when plain abdominal films or CT scans reveal extraintestinal gas bubbles or air either within the area of inflammation or elsewhere in the retroperitoneum. When the clinical suspicion of infection is high, fine-needle aspiration of peripancreatic or intrapancreatic fluid for culture and Gram stain analysis may be particularly helpful.[29] The procedure is most frequently done with CT guidance, and it is safe when performed by experienced radiologists.
Management of Sterile and Infected Acute Fluid Collections
Sterile acute fluid collections usually resolve spontaneously and no specific treatment is indicated. Attempts to drain acute fluid collections, either by using percutaneously placed drains or by intervening surgically, should be discouraged as they are usually unnecessary, and, furthermore, they are likely to lead to infection. Even without instrumentation, these fluid collections can become infected but, since they contain liquid pus with little or no necrotic tissue, they are amenable to transcutaneous catheter drainage along with antibiotic therapy. It is generally believed that aspirating fluid from any site near the pancreas yields information that is relevant to all of the fluid collections and that sampling multiple sites is unnecessary.
Management of Sterile and Infected Necrosis
The role of surgical intervention in the management of patients with sterile pancreatic or peripancreatic necrosis
1659
has been the subject of considerable controversy. Opinions range from those advocating aggressive dbridement for patients with sterile necrosis who fail to rapidly improve on nonoperative treatment to those who claim that surgical intervention is virtually never indicated when the necrosis is sterile. Those taking the former position claim that removing the necrotic tissue (i.e., necrosectomy) reduces morbidity and speeds recovery, whereas those taking the latter position, including me, base their position on the fact that most people treated nonoperatively will eventually recover and some who undergo operation may actually be made worse by the operation. There is, however, a general consensus that patients with infected necrosis require some form of intervention. Prospective studies have indicated that infection of areas of necrosis can occur at any time but that it usually occurs during the initial 3 to 4 weeks of an attack. Although some recent reports have indicated that highly selected patients might be adequately treated with antibiotics alone,[30] simple antibiotic therapy is generally considered to be inadequate because the necrotic tissue acts as a foreign body, making it impossible to sterilize the area with antibiotics alone. Combining antibiotic therapy with percutaneous catheter drainage may also not be adequate treatment because the pastelike necrotic tissue does not pass through the smallbore drainage catheters and, therefore, drainage is usually incomplete. Other methods of removing the necrotic tissue, either via a transpapillary endoscopic route or using minimally invasive surgical approaches with an operating nephroscope, have been tried, but experience with these techniques has been limited and essentially anecdotal. The conventional approach to managing infected necrosis involves laparotomy and surgical dbridement of the infected, devitalized tissue. Repeated operations and dbridement may be needed. The timing of the initial dbridement appears to be closely related to the outcome; that is, those undergoing later operations do better and require fewer repeat operations than those undergoing early operation. The goal of operation in patients with infected necrosis is to remove as much as possible of the infected, necrotic tissue and to provide drainage for the remaining viable exocrine tissue. Many different ways of achieving these goals have been described ( Box 534 ) and, although each has its advocates, none has been proven to be superior to the others. My practice is to perform repeated operations, each of which involves dbridement and abdominal wall closure. At the time of the final dbridement, drains and a feeding jejunostomy are placed. For the most part, the repeated laparotomies are performed every 2 to 3 days until no further dbridement is possible or necessary.
Management of Pancreatic Pseudocysts
Most pseudocysts communicate with the pancreatic ductal system and contain a watery fluid that is rich in pancreatic digestive enzymes. Typically, patients with pseudocysts have persistent elevations of circulating pancreatic enzymes. Recent reports have shown that many pseudocysts eventually resolve without complications and Box 53-4. Management Options for Infected Pancreatic Necrosis Conventional Approach Dbridement with reoperation when clinically indicated or at planned intervals Dbridement with open or closed packing and reoperation when clinically indicated or at planned intervals Dbridement with continuous lavage Unconventional Approach Antibiotics alone Antibiotics with percutaneous drainage Antibiotics with endoscopic drainage Antibiotics with surgical drainage but not dbridement Antibiotics with dbridement via minimally invasive surgery
that intervention is not mandatory in all cases unless the pseudocysts are symptomatic, enlarging, or associated with complications. The likelihood that a pseudocyst will resolve spontaneously, however, is dependent on its size. Large pseudocysts (i.e., > 6 cm in diameter) are more likely to become symptomatic either because they are tender or because of their mass effect on adjacent organs. Those that compress the stomach or duodenum may cause gastric outlet obstruction with nausea and vomiting. Those that reduce the capacity of the stomach frequently cause early satiety, whereas those impinging on the bile duct can cause obstructive jaundice. Pancreatic pseudocysts that erode into a neighboring vessel can result in formation of a pseudoaneurysm with hemosuccus pancreaticus and upper gastrointestinal bleeding. Symptomatic or enlarging pseudocysts can be treated by several methods. Those in the tail can be treated by excision (i.e., distal pancreatectomy) but excision in the setting of recent acute inflammation may be hazardous. Most patients who develop symptomatic pseudocysts are best managed by pseudocyst drainage. In poor surgical risk patients, percutaneous catheter drainage can be considered, but in my experience, that approach leads to considerable morbidity because of catheter-induced infection and the development of a prolonged external pancreatic fistula. Internal drainage can avoid these problems and seems preferable. Internal drainage can be accomplished either endoscopically (via transpapillary drainage, cyst-gastrostomy or cyst-duodenostomy) or surgically (via cyst-gastrostomy, cyst-duodenostomy, or Roux-en-Y cyst-jejunostomy). The approach chosen depends primarily on the locally available expertise as well as the location of the pseudocyst, but endoscopic drainage may be preferable in poor surgical risk patients. Pseudocysts that are directly adjacent to either the stomach or duodenum can be safely drained endoscopically if there are no intervening vessels. After endoscopic ultrasound and preliminary aspiration of the cyst fluid to confirm the diagnosis and exclude intervening vessels, endoscopic drainage is achieved by making an incision into the pseudocyst through the wall of the stomach or
1660
duodenum. To facilitate decompression, the opening should be relatively large and a pigtail catheter may be placed. Transpapillary drainage might be more appropriate for patients with pancreatic head pseudocysts whose CT and endoscopic ultrasound suggest that incising into the pseudocyst could be hazardous. At the time of endoscopic retrograde cholangiopancreatography (ERCP), a stent is passed into the pseudocyst through the papilla of Vater. Unfortunately, transpapillary drainage, particularly when incomplete, can allow bacteria to enter the pseudocyst and lead to development of an infected pseudocyst. Another transpapillary approach involves placing a stent across the duct defect rather than into the cyst through the defect. By excluding pancreatic juice from the pseudocyst, this bridging intraductal stent may permit the duct disruption to heal and the pseudocyst to resolve without drainage. Further experience with this technique will be needed before its ultimate use can be determined. Surgical internal drainage of pseudocysts is usually accomplished by creating either a Roux-en-Y cyst-jejunostomy, a side-to-side cyst-gastrostomy, or a side-to-side cyst duodenostomy. The former is usually accomplished by directly anastomosing a defunctionalized Roux-en-Y limb of jejunum to the opened pseudocyst. Surgical cystgastrostomy (or cyst-duodenostomy) has traditionally been accomplished by laparotomy and anterior gastrotomy (or lateral duodenotomy). A generous incision is then made through the posterior wall of the stomach (or medial wall of the duodenum) into the pseudocyst. Some surgeons now perform cyst-gastrostomy using a laparoscopic approach.
Management of Pancreatic Ascites and Pancreaticopleural Fistulas
Pancreatic ascites occurs when pancreatic juice gains entry into the peritoneal cavity either from a pancreatic duct disruption or from a leaking pseudocyst. The diagnosis can usually be made when high amylase levels are found in the ascitic fluid. The initial treatment usually is nonoperative and involves attempts to decrease pancreatic secretion by elimination of enteral feeding, institution of nasogastric drainage, and administration of the antisecretory hormone somatostatin. Repeated paracentesis may also be helpful. Roughly 50% to 60% of patients can be expected to respond to this treatment with resolution of pancreatic ascites within 2 to 3 weeks. Persistent or recurrent ascites can be treated either endoscopically or surgically. Endoscopic treatment involves endoscopic pancreatic sphincterotomy with or without placement of a transpapillary pancreatic duct stent. By reducing the resistance to drainage into the duodenum, and by bridging the site of duct disruption, this approach is designed to allow the site of leakage to seal. Surgical treatment of pancreatic ascites, usually preceded by performance of an ERCP to identify the site of duct disruption, involves either resection (for leaks in the pancreatic tail) or internal Roux-en-Y drainage (for leaks in the head/neck region). It seems most appropriate to attempt endoscopic treatment initially and to reserve surgical treatment for those patients who do not respond to endoscopic therapy. The genesis of pancreaticopleural fistula is similar to that of pancreatic ascites, but in this case the duct disruption is usually posterior and the extravasated juice travels in a cephalad direction through the retroperitoneum to reach the thoracic cavity. Although the incidence of pancreaticopleural fistula is lower than that of pancreatic ascites, the management of both is similar.
Management of Pancreatitis-Induced False Aneurysms
Rarely, pancreatic pseudocysts or areas of pancreatic necrosis can erode into pancreatic or peripancreatic vascular structures. This results in the formation of a false aneurysm since the vessel communicates with the pseudocyst. That false aneurysm may either communicate with the ductal system or rupture into the free peritoneal cavity. The former leads to bleeding into the pancreatic duct (hemosuccus pancreaticus) and presents as transpapillary upper gastrointestinal bleeding. Rupture into the peritoneal cavity can lead to hemoperitoneum. Therapeutic angiographic embolization is most appropriate for the unstable patient, and this approach may also provide definitive treatment, particularly for those patients whose false aneurysm is in the pancreatic head. For those whose false aneurysm is in the tail of the pancreas, subsequent distal pancreatectomy, once the patient is stabilized, may provide more secure hemostasis.
Management of Pancreaticoenteric Fistulas
Pancreatic pseudocysts or areas of pancreatic necrosis can erode into the small intestine, duodenum, stomach, bile duct, or splenic flexure of the colon. Occasionally, this results in resolution of the pseudocyst and no further treatment is needed. More often, however, such an event is accompanied by significant bleeding and/or signs of sepsis, and surgical intervention is usually required. Management of these fistulas is determined by the gastrointestinal organ involved.
Management of Pancreatitis-Induced Splenic Vein Thrombosis
Because of the close proximity of the splenic vein to the pancreas, splenic vein thrombosis is not unusual in cases of severe pancreatitis. For the most part, it does not result in early symptoms, but it may eventually result in the formation of gastroesophageal varices. Splenectomy provides effective and definitive treatment when these sinistral varices bleed, but because bleeding occurs in fewer than 10% of these patients, prophylactic splenectomy is not generally performed.
CHRONIC PANCREATITIS
Pathology and Etiology of Chronic Pancreatitis
Chronic pancreatitis is characterized by irreversible changes including pancreatic fibrosis and the loss of functional pancreatic exocrine and/or endocrine tissue. Most patients develop chronic pancreatitis as a result of prolonged
1661
ethanol abuse. It is generally believed that, in its earliest stages, chronic pancreatitis is an acute inflammatory process and that repeated episodes of subclinical acute pancreatic injury and necrosis lead to the fibrosis of chronic pancreatitis.
Diagnosis of Chronic Pancreatitis
There has been considerable confusion concerning the clinical distinction between chronic and acute pancreatitis. To a great extent, this confusion results from the fact that, from a clinical standpoint, attacks of chronic pancreatitis may be indistinguishable from those of acute pancreatitis. Fortunately, the initial management of acute or chronic pancreatitis attacks is identical, as is the management of complications such as infection, necrosis, and pseudocyst (see Treatment section, earlier). On the other hand, the two forms of pancreatitis have natural histories that differ considerably, and the long-term management of chronic pancreatitis presents challenges that are not inherent to the management of acute pancreatitis.
History
Patients with chronic pancreatitis may describe prior episodes of pancreatic-type abdominal pain, and 60% to 80% of patients have a long history of ethanol abuse. There may be a family history of pancreatitis suggestive of the presence of hereditary pancreatitis or a history of autoimmune diseases including primary sclerosing cholangitis and Sjgrens syndrome that should raise suspicion of pancreatitis on an autoimmune basis. Diabetes mellitus and/or a history suggestive of malabsorption (i.e., steatorrhea) indicate that significant pancreatic endocrine and/or exocrine function has been lost and this is most compatible with the diagnosis of chronic pancreatitis. Typically, patients with chronic pancreatitis have upper abdominal pain radiating to the back. It can be constant or episodic and triggered by drinking alcohol or eating. Repeated use of heating pads or hot water bottles to treat the chronic pain may result in skin lesions (erythema ab igne) that define the distribution of the pain ( Fig. 535 ). Some patients experience no pain.
Imaging Studies
Radiographs or CT scans showing pancreatic calcifications are diagnostic of chronic pancreatitis ( Fig. 536 ). Those calcifications reflect the deposition of calcium carbonate in the intraductal protein plugs that frequently, but not invariably, occur in chronic pancreatitis. Thus, the absence of pancreatic calcifications does not rule out a diagnosis of chronic pancreatitis. Perhaps the most sensitive methods for diagnosing chronic pancreatitis are those that provide images of the pancreatic ductal system. ERCP, CT cholangiopancreatography, or MR cholangiopancreatography may be particularly valuable in the diagnosis of chronic pancreatitis. Chronic pancreatitis is characterized by irregularities of the pancreatic ducts, ductal strictures, and areas of duct dilation ( Fig. 537 ). The
Figure 53-5 Erythema ab igne. Skin injury, characterized by keratinocyte injury and melanocyte activation, is induced by mild and repeated exposure to infrared sources. This patient with chronic pancreatitis repeatedly applied a heating pad to the painful area on his back.
Figure 53-6 Pancreatic calcifications. CT scan showing multiple, calcified, intraductal stones in a patient with hereditary chronic pancreatitis.
major as well as the side-branch ducts may be involved. For unexplained reasons, some patients with chronic pancreatitis develop dilated main pancreatic ducts (large duct disease), whereas others retain ducts of normal or even smaller than normal caliber (small duct disease). Some patients with chronic pancreatitis can be shown to
1662
Figure 53-7 Endoscopic retrograde cholangiopancreatography (ERCP) in chronic pancreatitis. The pancreatic duct and its side branches are irregularly dilated.
Box 53-5. Pancreatic Function Tests Tubeless Tests Fecal tests Fat stain 72-hour fat content Chymotrypsin, trypsin, elastase content Indirect tests Bentiromide test Pancreolauryl test Breath tests Tube Tests Lundh test meal Secretin/cholecystokinin test
have major ducts that have the appearance of a chain of lakes or a string of pearls that is the result of segments of dilated duct separated by areas of ductal stricture. Transcutaneous and endoscopic ultrasound can also be used to diagnose chronic pancreatitis if duct dilation, calcifications, pseudocysts, or parenchymal fibrosis are seen. Ultrasound examination is more operator dependent and perhaps less sensitive than either CT or MRI.
Pancreatic Function Tests
The pancreas has considerable functional reserve, and more than 90% of exocrine function must be lost before steatorrhea develops. More subtle losses may be identified by performance of pancreatic function tests that can be either noninvasive (tubeless tests) or invasive (tube tests) ( Box 535 ). Tubeless tests involve measuring the stool content of fat, measuring stool content of digestive enzymes, or orally administering a pancreatic digestive enzyme substrate and quantitating enzyme activity in the gut by measuring metabolic product in either the urine or exhaled gases. These tests, although nonintrusive, are notoriously insensitive and, therefore, normal results are not too helpful. The more invasive tube tests involve placement of a collecting tube in the duodenum and measuring pancreatic bicarbonate or enzyme output after meal or hormone stimulation of the pancreas. These tests are more specific and sensitive than the tubeless tests, but they are still relatively insensitive and they have a relatively high rate of false-negative results
Natural History
Some patients with chronic pancreatitis have a painless disease that remains unrecognized until complications or loss of pancreatic function lead to the diagnosis. Most patients, however, have intermittent or constant pain that may limit lifestyle and/or mandate repeated hospitalizations. Ammann and coworkers[31] have suggested that the painful pancreatitis experienced by many of these patients evolves into a painless disease as pancreatic function is lost, but the existence of this burn-out phenomenon is highly controversial. More often, the disease remains painful, addicting doses of narcotics are required, and loss of function results in diabetes, steatorrhea, and profound weight loss.
Treatment of Pancreatic Malabsorption
The loss of pancreatic exocrine function in chronic pancreatitis affects the output of all pancreatic digestive enzymes, but it is mostly fat absorption that is abnormal, and it is the delivery of lipolytic enzyme activity to the small intestine that determines the success of treatment. In health, roughly 300,000 IU of lipase is secreted by the pancreas within 4 hours of ingesting a typical meal, but only 10% (30,000 IU) of secreted lipase is needed to allow for normal fat digestion/absorption. Theoretically, pancreatic malabsorption of fat should be corrected by oral administration of exogenous lipase. Unfortunately, most orally administered lipase is inactivated as it traverses the acidic environment of the stomach, allowing only 8% to 15% of ingested lipase activity to reach the duodenum. Some of that lipase may be ineffective, either because of low duodenal pH (caused by inadequate pancreatic secretion of bicarbonate) or because the exogenously administered lipase arrives in the duodenum before or after the ingested fat. The use of acidinhibiting agents (e.g., proton-pump inhibitors) and enterically coated microsphere delivery systems can partially compensate for these problems. Thus, treatment involves acid suppression, a low-fat diet, and lipase doses of 90 to 150,000 IU per meal, although control of steatorrhea is often incomplete
even with this treatment.

Treatment of Pain in Chronic Pancreatitis
Medical Management
Complete abstinence from ethanol is advised for patients with alcohol-induced pancreatitis, but symptoms may persist even after complete abstinence. Attacks of hyperlipidemia-induced pancreatitis can be prevented by normalizing
1663
lipid levels with medication and/or dietary changes. Some patients with autoimmune pancreatitis are cured by administration of steroids. For most patients with painful chronic pancreatitis, intermittent or persistent pain remains a major issue and analgesics of increasing potency are needed. Toskes and coworkers[32] have noted that some of their patients with painful chronic pancreatitis have diminished pain if pancreatic secretion is reduced either by oral administration of pancreatic enzymes or by administration of the inhibitory hormone somatostatin. However, the clinical results achieved using exogenous pancreatic enzymes to reduce the pain of chronic pancreatitis have been variable and, at this time, the role of enzyme administration for pain relief in these patients is highly controversial.
Endoscopic Management
The endoscopic treatment of chronic pancreatitis has not been tested by well-designed prospective, randomized trials; therefore, the ultimate value of these treatments remains to be established. Several endoscopic approaches have been described. Endoscopic pancreatic sphincterotomy has been reported to benefit some patients with elevated sphincter of Oddi pressures. Endoscopic minor pancreatic sphincterotomy has been used to treat patients with pancreatitis and pancreas divisum. Pancreatic duct stones have also been removed or fragmented using an endoscopic approach with reported benefits. Finally, some patients with pancreatic duct strictures have been treated with endoscopically placed stents that pass across the stricture, but the value of this treatment is unclear since the stents themselves can cause strictures.
Neuroablative Procedures
Pain from the pancreas is carried in sympathetic fibers that traverse the celiac ganglia, reach the sympathetic chain through the splanchnic nerves, and then ascend to the cortex. Celiac plexus nerve blocks performed either percutaneously or endoscopically have been employed to abolish this pain with inconsistent results. Recently, splanchnicectomy performed in the chest via a thoracoscopic approach has been used with reports of transient improvement in 70% of patients and long-lasting pain control in 50%.[33] Experience with thoracoscopic splanchnicectomy has been only anecdotal, and randomized, prospective trials will be needed to determine its ultimate value.
Surgical Treatment of Chronic Pancreatitis
The two indications for surgical intervention are pain and concern over the possible presence of cancer. After the diagnosis of chronic pancreatitis has been established, surgical intervention should be considered when (1) the pain is severe enough to limit the patients lifestyle and/or reduce productivity and (2) the pain persists in spite of complete abstinence from alcohol and administration of non-narcotic analgesics. Imaging studies should be performed to define pancreatic and ductal anatomy since that will determine the surgical options. Finally, the risks and benefits of planned procedures should be clearly explained to the patient because, even with a technically successful operation, the pain may persist and further deterioration in exocrine and endocrine function can still occur.
Drainage Procedures for Patients With Small Ducts
Patients with small (>4- to 5-mm) pancreatic ducts, particularly those whose pancreatitis is caused by obstruction at the ampullary level, may benefit from transduodenal sphincteroplasty of the common bile duct with division of the septum that lies between the pancreatic duct and bile duct (pancreatic septotomy). Sphincteroplasty of the lesser papilla might be appropriate for patients with pancreas divisum. On the other hand, most patients with chronic pancreatitis have multiple areas of duct stricture throughout the pancreas and are unlikely to benefit from these transduodenal procedures.
Drainage Procedures for Patients With Dilated Ducts
The ideal treatment for these patients involves creating an anastomotic connection between the dilated duct and the intestinal lumen. There is little agreement concerning the minimum duct size needed to perform these anastomoses. Ducts larger than 1 cm in diameter are, clearly, large enough, but many surgeons perform duct-to-intestine drainage procedures with ducts as small as 5 mm. Duct drainage operations were pioneered by Duval, who described a procedure that involved splenectomy, resection of the pancreatic tail, and then creation of an end-to-end anastomosis between the transected end of the pancreas and a Roux-en-Y limb of jejunum. This procedure often failed because the presence of multiple pancreatic duct strictures interfered with complete ductal decompression. Puestow and Gillesby, in 1958, described an operation that involved longitudinally opening the entire duct and then invaginating the opened pancreas into a Roux-en-Y loop of jejunum. This allowed for more complete decompression but still required splenectomy. Later, Partington and Rochelle[34] modified the Puestow procedure by creating a side-to-side anastomosis between the opened duct and jejunum, thus eliminating the need for splenectomy ( Fig. 538 ). In appropriately selected patients (i.e., those with large ducts and those with intraductal stones), longitudinal pancreaticojejunostomy, performed according to the Partington and Rochelle modification of the Puestow procedure, has been reported to result in immediate pain relief in more than 80% of patients and long-term pain relief in roughly 60% of patients. More recently, Ho and Frey[35] further modified the procedure by including removal of part of the pancreatic head, thereby marsupializing the duct as it dives deeply in the pancreas to reach the ampulla of Vater. This allows for an even more complete duct decompression and a longer longitudinal pancreaticojejunostomy. Both short- and long-term pain relief appear to be improved, and the procedure can be performed when the duct is only moderately dilated.
1664
Figure 53-8 Partington and Rochelle modification of the Puestow procedure. The pancreatic duct is opened from the tail of the pancreas to the edge of the duodenum and a side-toside anastomosis is created joining a Roux-en-Y limb of jejunum to the opened pancreatic duct. (From Carey LC: Pancreaticojejunostomy with cystoduodenostomy. In Malt RA [ed]: Surgical Techniques Illustrated. New York, WB Saunders, 1985, 396405.)
Resective Procedures
Painful chronic pancreatitis can be treated with resection of the body and tail of the pancreas (distal pancreatectomy), resection of the head and uncinate process of the pancreas (Whipple procedure), with subtotal pancreatectomy that spares a rim of pancreas along the inner curve of the duodenum, and with total pancreatectomy. Each of these procedures can either cause or worsen pancreatic exocrine and endocrine insufficiency and, in the case of total pancreatectomy, a brittle form of diabetes can occur. Most experts believe that it is the inflammatory process in the pancreatic head that controls both the severity of symptoms and the further progression of the disease in the remainder of the gland. Perhaps because of this, resection of the pancreatic head has been shown to completely relieve the pain of chronic pancreatitis in 70% to 80% of patients. Resection of the pancreatic head can be accomplished by a standard pancreaticoduodenectomy (Whipple procedure) or by its pylorus-preserving modification (pylorus-preserving Whipple procedure) ( Fig. 539 ).[36] Relief of symptoms by either procedure is comparable, but some claim that the quality of life and gastrointestinal function are better after the pyloruspreserving operation. Beger and colleagues have modified the Whipple procedure even further by coring out the head of the pancreas and preserving the duodenum and distal bile duct.[37] They claim that this duodenum-preserving pancreatic head resection yields results that are as good as or better than those achieved with the standard Whipple procedure. Distal pancreatectomy is the ideal surgical procedure for patients whose chronic pancreatitis is confined to the pancreatic tail. This occurs in patients who develop a mid-duct stricture either as a result of necrotizing acute pancreatitis or as a result of trauma that injures the gland and duct as they cross the spine. Usually, distal pancreatectomy is combined with splenectomy for technical reasons but, in fact, the spleen can be preserved if its vascular supply is secure. Distal pancreatectomy should not be performed for patients with diffuse, chronic pancreatitis that involves the entire gland, even if the pancreatic tail is the area most severely involved, since recurrence of pancreatitis in the head can be anticipated and further resection
1665
Figure 53-9 Standard and pylorus-preserving Whipple procedure. The standard Whipple (A) involves resection of the gastric antrum, head of pancreas, distal bile duct, and entire duodenum with reconstruction as shown. The pylorus-preserving Whipple (B) does not include resection of the distal stomach, pylorus, or proximal duodenum. Gl. bl., gallbladder; duod., duodenum; stom., stomach; trans., transverse. (From Cameron, JL: Current status of the Whipple operation for periampullary carcinoma. Surg Rounds 7787, 1988.)
of the pancreas, in that case, would leave the patient without functioning pancreatic endocrine tissue. The role of total or near-total pancreatectomy in the treatment of patients with chronic pancreatitis is not clear. These procedures may represent the only surgical option for patients who have failed drainage procedures or those with small ducts who have already undergone distal pancreatectomy. Some patients continue to experience severe pancreatic pain even after total pancreatectomy, and, for this reason, the effects of total pancreatectomy on pain in chronic pancreatitis cannot be accurately predicted. On the other hand, it can be expected that patients undergoing total or near-total pancreatectomy will have brittle diabetes and severe steatorrhea. In combination with ongoing ethanol or drug abuse, the brittle diabetes and malnutrition may be unmanageable problems, and a high late mortality rate in these patients has been reported. In an attempt to avoid brittle diabetes in these patients, some surgeons have advocated harvesting and autotransplanting islets of Langerhans from the resected specimen. Modest success at insulin independence has been achieved using this approach, but the ultimate role of islet reimplantation remains to be established. In the past, some surgeons have reimplanted
the entire resected pancreas with mixed results, and the procedure is performed only rarely now.
BENIGN EXOCRINE TUMORS Most benign pancreatic exocrine tumors are cystic, but not all cystic tumors are benign. Benign cystic tumors, which account for 10% to 15% of pancreatic tumors, are usually asymptomatic but, when symptoms develop, they are usually related to pressure or obstruction of an adjacent organ.
Serous Cystadenoma
These tumors account for 20% to 40% of cystic pancreatic neoplasms. They are lined by a flattened epithelium with glycogen-rich cytoplasm that does not stain for mucins, and the finding of glycogen-rich cells on cytologic examination is diagnostic of a serous cystadenoma. Rare cases of malignant serous cystic lesions have been reported, but most are benign and have no malignant potential. Typically, they are large, spherical masses that contain a watery fluid and have a central, calcified stellate scar ( Fig. 5310 ). Oligocystic varieties, with large cystic spaces, can occur, but most are microcystic. They usually occur in the body or tail and are asymptomatic but, when located in the head, even benign serous cystadenomas can become symptomatic if they enlarge and compress adjacent structures. Resection is indicated when the diagnosis is in doubt or when they become symptomatic.
Mucinous Tumors
These tumors account for 20% to 40% of cystic tumors. Even if benign at the time of diagnosis, they are usually considered to have malignant potential. Two types have been described, but neither type usually communicates with the pancreatic duct. One form contains areas of ovarian-like stroma, is almost always found in women, and
1666
Figure 53-10 CT scan of serous (microcystic) cystadenoma. Note scar with central calcification (arrow).
is almost always found in the pancreatic tail ( Fig. 5311 ). The more common type, however, lacks ovarian stroma and it can be found anywhere in the pancreas. It occurs equally in both sexes. For both types, imaging studies usually indicate that the lesion is composed of one or more very large cysts (i.e., macrocystic), although microcystic mucinous tumors can also occur. The cysts are lined by a columnar, mucin-producing, and sometimes papillary epithelium. Prolonged survival (i.e., >5 years) can be anticipated in more than 50% of patients if these tumors are resected prior to the development of invasive malignancy, but even after the development of malignant changes and invasion, long-term survival is still better than for ductal adenocarcinoma.
Intraductal Papillary Mucinous Tumor
Intraductal papillary mucinous tumor (IPMT), also known as intraductal papillary mucinous neoplasm, is another type of cystic pancreatic neoplasm. Although first described in Japan in the 1980s, it is now being recognized worldwide and its incidence appears to be rapidly increasing. Both men and women are equally affected. IPMT can involve the major ducts (main duct variety), the smaller ducts (branch duct variety) or both types of ducts. It can be located in any or all parts of the pancreas, although involvement of the head appears to be its most common form. IPMT patients can experience pancreatitis when mucus, secreted by the tumor, transiently obstructs the orifice of the pancreatic duct. The diagnosis of IPMT can be made with near certainty if mucus is seen extruding through a large, fish-mouthlike papillary orifice at the time of endoscopy. The main or side-branch ducts involved with IPMT are usually lined by columnar mucin-producing cells that develop papillary projections. Other areas of the duct, although lined by normal epithelium, may be dilated as a result of prior obstructive episodes. IPMT is believed to follow an adenoma-carcinoma sequence, and it can be classified according to the Pan-IN classification scheme ( Fig. 5312 ) that categorizes tumors as having minimal or no dysplasia (PanIN-1), moderate
Figure 53-11 CT scan of mucinous (macrocystic) cystadenoma. Note multiple large cystic spaces. Microscopic examination revealed multiple areas of ovarian-like stroma.
dysplasia (PanIN-2), or severe dysplasia/carcinoma in situ (PanIN-3).[38] The natural history of tumors with only mild or no dysplasia is not known, but
those with severe dysplasia and/or carcinoma in situ are likely to become locally invasive and metastasize if left unresected. Resection with, at worst, PanIN-1 changes at the margin, prior to development of invasive malignancy, is usually curative. This may mandate total pancreatectomy. When resection is performed after development of invasive malignancy, cure rates are relatively low.
Management of Cystic Pancreatic Neoplasms
A recent history of pancreatitis, enzyme-rich fluid in the cyst, and communication between the cyst and the pancreatic duct suggests that the cystic lesion is a postpancreatitic (i.e., inflammatory) pseudocyst. Patients with asymptomatic pseudocysts can be left untreated. However, occasionally these features can also be associated with neoplastic cystic lesions; therefore, these patients should be closely followed. Serous cystadenomas, diagnosed by finding glycogen-rich cells on fine-needle aspiration biopsy, can also be followed without resection if they are not symptomatic. The major challenge is distinguishing pseudocysts and serous cystadenomas from the other, malignant or potentially malignant, types of cystic pancreatic tumors. The finding of mucin in the cyst, mucin-secreting cells on biopsy, a high cyst fluid viscosity, or a high cyst fluid carcinoembryonic antigen (CEA) is suggestive, but not diagnostic, of a potentially malignant tumor.[39] Failure to observe these changes, however, does not exclude a malignant or potentially malignant mass. Because of these uncertainties, all neoplastic cysts should be resected unless the diagnosis of a serous cystadenoma can be made with certainty. For those located in the pancreatic tail, a distal pancreatectomy is ideal, whereas cystic lesions in the pancreatic head present a greater
1667
Figure 53-12 Progression model for pancreatic cancer. The progression from histologically normal ductal epithelium to low-grade pancreatic intraepithelial neoplasia (PanIN) to highgrade PanIN (left to right) is associated with the accumulation of specific genetic changes. Early changes include Her-2/neu and K-ras mutations; intermediate changes include p16 mutations; and changes associated with either in situ or early invasive cancer include p53, BRCA2, and DPC4 mutations. (From Wilentz RE, Iacobuzio-Donahoe CA, Argani P, et al: Loss of expression of DPC4 in pancreatic intraepithelial neoplasia: Evidence that DPC4 inactivation occurs late in neoplastic progression. Cancer Res 60:20022006, 2000.)
decision-making challenge because resection would require pancreaticoduodenectomy. Cytologic examination of aspirated tissue, removed either transcutaneously or with endoscopic ultrasonographic guidance, may aid in the difficult management decision, but not infrequently, doubt persists, and in good surgical risk patients, pancreaticoduodenectomy is probably indicated.
Solid-Pseudopapillary Tumor of the Pancreas
Solid-pseudopapillary tumor, a relatively uncommon tumor, often occurs in young women and follows a benign course. Local resection is usually curative, although incomplete removal can result in local recurrence, and malignant varieties have been described. The tumors are usually large, round, and well-demarcated masses that can occur in any part of the pancreas. Histologically, the tumor is mostly solid and consists of monomorphous eosinophilic or clear cells that demonstrate a pseudopapillary architecture. In suitable operative candidates, solid-pseudopapillary tumors of the pancreas should be resected.
MALIGNANT PANCREATIC TUMORS

Incidence and Epidemiology
Pancreatic cancer affects 25,000 to 30,000 people in the United States each year and is the fourth or fifth leading cause of cancer-related death in this country. It occurs more frequently in men than in women and it is more common among blacks than whites. Roughly 80% of cases occur between 60 and 80 years of age, whereas less than 2% occur in people younger than 40. Other risk factors include a history of hereditary or chronic pancreatitis, cigarette smoking, and occupational exposure to carcinogens. The incidence of diabetes mellitus is increased in patients with pancreatic cancer, but the relationship of diabetes to pancreas cancer is controversial. Some studies have indicated that diabetes is a risk factor for the development of pancreas cancer, whereas others have argued that diabetes may be a manifestation of the cancer. Coffee drinking, which was once considered a risk factor, is no longer thought to play a role in the development of pancreatic cancer.
Pathology
Ductal adenocarcinoma and its variants account for 80% to 90% of all pancreatic neoplasms and for an even greater fraction of the malignant tumors. Roughly 70% of ductal cancers arise in the pancreatic head or uncinate process. Grossly, they appear as hard, irregular, gritty masses that are yellowgray and are usually poorly demarcated. At the time of diagnosis, they are usually larger than 3 cm in diameter and both nodal and distant metastases are also frequently present. Those originating in the body or tail of the pancreas are often larger and more likely to have spread before their presence is known. Microscopically, the degree of differentiation, mitotic index, and amount of mucous may vary considerably. They are frequently associated with an intense desmoplastic stromal reaction and fibrosis. A halo of chronic pancreatitis frequently surrounds the tumor, presumably caused by tumorinduced obstruction of neighboring ducts. Areas of vascular and lymphatic invasion within and around the tumor are commonly seen. In addition, perineural growth of the tumor is highly characteristic of this cancer and may
1668
account for the propensity of pancreatic cancer to extend into neighboring neural plexus causing both upper abdominal and back pain. Cancers such as mucinous noncystic carcinoma (colloid carcinoma), signet ring cell carcinoma, adenosquamous carcinoma, anaplastic carcinoma, giant cell carcinoma, and sarcomatoid carcinoma are considered to be variants of ductal adenocarcinoma that differ primarily in their degree of differentiation and their morphologic appearance. In addition to ductal adenocarcinoma, other cancers of the pancreas are known. Some of those cancers represent previously benign tumors that have undergone malignant change. For example, carcinoma in situ and invasive carcinoma can develop in areas of IPMT or within areas of mucinous cystic tumors. Other forms of pancreas cancer include acinar cell carcinomas, which are most common in the 5th to 7th decade of life and most frequently present as large masses in the body or tail of the pancreas. Pancreatoblastoma is a rare form of pancreas cancer that usually presents in young children. Nonepithelial cancers of all types, including leiomyosarcomas, liposarcomas, plasmacytomas, and lymphomas, can also develop within the pancreas, but these tumors are quite rare. Lymphomas of the pancreas, if diagnosed prior to resection, are usually best treated with chemoradiation.
Development and Molecular Biology
The following three types of genetic abnormalities have been observed in pancreatic cancers: activation of growth-promoting oncogenes, mutations that result in the inactivation of tumor suppressor genes, and excessive expression of growth factors and/or their receptors.[40] Almost all pancreatic cancers and most of the precursor lesions demonstrate codon 12 mutations of the K-ras oncogene. These and other mutations of K-ras are believed to be early events in pancreatic tumorigenesis. K-ras normally plays a key role in regulating many cellular events, including growth. Its activation is a guanosine triphosphate (GTP)-dependent event, terminated by GTP hydrolysis and dissociation of GDP. Oncogenic point mutations of K-ras interfere with that termination event, and, as a result, K-ras becomes permanently activated and a continuous growth signal is transmitted to the nucleus. Mutation of the p53 tumor suppressor gene is the most common genetic event in all human cancers, and it is observed in 75% of pancreatic cancers. The gene encodes a 53-kD nuclear phosphoprotein that plays an important role in regulating the cell cycle, DNA synthesis and repair, apoptosis, and cell differentiation. Although mutation of p53 results in the loss of its function, p53 mutations alone are not believed to be sufficient to cause malignant transformation. Mutations resulting in the functional loss of other tumor suppressor genes, including p16, SMAD-4, DPC, and DCC are also commonly observed in pancreatic cancer. Other much less frequent tumor suppressor gene deletions that have been observed in pancreatic cancers result in the loss of the retinoblastoma gene and the adenomatous polyposis coli (APC) gene. DNA mismatch repair genes have also been reported to be mutated and functionally deleted in pancreatic cancers. These genes encode for enzymes that function to repair potentially pathologic DNA changes that occur during DNA replication. Several growth factor systems are commonly upregulated in pancreatic cancer either by increased expression of their receptors or by increases in the relevant ligands. The most commonly observed changes involve the epidermal growth factor (EGF) receptor family that includes the EGF receptor (which responds to EGF and to transforming growth factor-) and the HER2, HER3, and HER4 receptors. Overexpression of EGF receptors or its ligands is correlated with tumor invasiveness, enhanced potential for metastasis, and a poorer prognosis. Increased expression of HER2 is associated with a better differentiated phenotype. Other growth factor systems implicated in pancreatic cancer development include those for hepatocyte growth factor and for transforming growth factor-. It is now generally believed that pancreas cancer evolves in a progressive, step-wise fashion, much like that observed for colon cancer. Precursor ductal lesions (see Fig. 5312 ) have been identified, and the step-wise progression toward invasive cancer and metastasis has been related to the accumulated presence of multiple genetic abnormalities. K-ras mutations and HER2/neu overexpression are the earliest changes to occur. Alterations in p16 are found primarily in PanIN-2 and PanIN-3. DPC4, BRCA2, and p53 are inactivated during the later stages of cancer progression and are found almost exclusively in invasive lesions.
Hereditary Pancreatic Cancer Syndromes
Pancreatic cancer has been observed to be increased in families with hereditary nonpolyposis colon cancer (HNPCC), those with familial breast cancer (associated with the BRCA2 mutation), those with Peutz-Jeghers syndrome, those with ataxia-telangiectasia, and those with the familial atypical multiple mole melanoma (FAMMM) syndrome. Patients with hereditary pancreatitis are also at increased risk of developing pancreatic cancer. This is particularly true in those with a paternal pattern of inheritance who may have a 75% risk of developing pancreatic cancer.[41] Even in the absence of one of these familial cancer syndromes or hereditary pancreatitis, individuals with a family history of pancreatic cancer, especially those with two or more
pancreatic canceraffected first-degree relatives, have an increased risk of developing pancreatic cancer.
Symptoms and Signs
Pancreatic cancers are insidious tumors that can be present for long periods and grow extensively before they produce symptoms. The symptoms, once they develop, are determined by the location of the tumor in the
1669
TABLE 53-2 -- Signs and Symptoms of Pancreatic Cancer Frequent Pancreatic Head Cancers Weight loss (92%) Pain (72%) Jaundice (82%) Dark Urine (63%) Light stools (62%) Anorexia (64%) Pancreatic Body/Tail Cancers Weight loss (100%) Pain (87%) Weakness (43%) Nausea (45%) Anorexia (33%) Vomiting (37%) pancreas ( Table 532 ). Those in the head or uncinate process of the pancreas make their presence known by causing bile duct, duodenal, or pancreatic duct obstruction. Symptoms include unexplained episodes of pancreatitis, painless jaundice, nausea, vomiting, steatorrhea, and unexplained weight loss. With further spread beyond the pancreas, these patients may note upper abdominal and/or back pain when peripancreatic nerve plexus are involved and ascites when peritoneal carcinomatosis or portal vein occlusion develops. Patients with tumors arising in the neck, body, or tail of the pancreas usually do not develop jaundice or gastric outlet obstruction. Their symptoms may be limited to unexplained weight loss and vague upper abdominal pain until the tumor has grown extensively and spread beyond the pancreas. New-onset diabetes mellitus is occasionally the first symptom of an otherwise occult pancreatic cancer. Recent studies have suggested that this form of diabetes may be mediated by a factor released from the tumor that either inhibits insulin release from islets or induces peripheral insulin resistance. Unexplained migratory thrombophlebitis (Trousseaus syndrome) may be associated with pancreatic and other types of malignancy. It is probably a paraneoplastic phenomenon that results from a tumor-induced hypercoagulable state. The physical findings in patients with pancreas cancer are also dependent on the location, size, and extent of the tumor. Liver nodules indicative of metastases can sometimes be felt. Metastatic subumbilical (Sister Mary Joseph node) and pelvic peritoneal (Blummers shelf) deposits as well as left supraclavicular lymphadenopathy (Virchows node) indicate the presence of distant metastases. Malignant ascites, caused by peritoneal carcinomatosis, may also be present. With portal, splenic, or superior mesenteric vein occlusion, mesenteric venous pressures may be increased and collateral channels, including gastroesophageal varices and a caput medusae, may develop. Distal common bile duct obstruction caused by the tumor often leads to bile duct and gallbladder distention. Thus, a palpable gallbladder in a patient with painless jaundice (i.e., Courvoisiers sign) should suggest the presence of a periampullary neoplasm.
Blood Tests
Infrequent Nausea (37%) Weakness (35%) Pruritus (24%) Vomiting (37%)
Jaundice (7%) Dark urine (5%) Light stool (6%) Pruritus (4%)
Patients with pancreatic head lesions frequently have elevated bilirubin and alkaline phosphatase levels suggestive of obstructive jaundice. Other routine laboratory studies are usually normal. The two most widely used pancreatic cancer serum markers are the CEA and the Lewis blood group carbohydrate antigen CA 199. Both are frequently elevated in patients with advanced disease, but, unfortunately, the circulating levels of these tumor markers are often normal in patients with early, potentially curable, tumors. Thus, using these tumor markers to screen patients with vague symptoms or those in high-risk groups has not been shown to be useful in detecting early disease. With a cutoff value of 37 U/mL, CA 199 has been reported to have a sensitivity of 86% and a specificity of 87%.[42] CA 199 can also be elevated in patients with cholangitis and jaundice not caused by pancreatic cancer. Extremely high levels of either the CA 199 or CEA usually indicate unresectable and/or metastatic disease.
Imaging Studies
For most patients, the initial imaging study is a transcutaneous ultrasound examination. It may reveal a pancreatic mass and indicate whether that mass is solid or cystic. It can also aid in the diagnosis of jaundiced patients by revealing the presence of extrahepatic ductal dilation in the absence of demonstrable biliary tract stones. Transcutaneous ultrasound, regardless of its findings, is usually followed by helical contrast-enhanced CT, performed in conjunction with intravenous infusion of contrast material. Timed images are taken to permit visualization of the pancreas during the arterial phase, the parenchymal phase, and finally the venous phase of contrast perfusion. Pancreatic cancer usually appears as a hypodense mass with poorly demarcated edges. It may have a more hypodense center, indicating either central necrosis or cystic change, and the pancreatic duct to the left of the lesion may be dilated. When performed and interpreted by experienced radiologists, CT has a specificity for diagnosing pancreatic tumors of 95% or better. Its sensitivity depends on the size of the tumor exceeding 95% for tumors larger than 2 cm in diameter. Much, if not all, of the information provided by CT can also be obtained with high-quality MRI. Although the sensitivity and specificity of MRI appear to equal those of CT, CT is currently more widely employed perhaps because it is cheaper, more user friendly, and more easily interpreted by clinicians. Positron emission tomography (PET) may be of value in diagnosing small pancreatic tumors that escape CT or MRI detection, but the sensitivity and specificity of PET scanning remain to be established. ERCP may be particularly helpful in evaluating patients with obstructive jaundice without a detectable mass on CT or MRI. It can identify stones or other nonmalignant
1670
causes of obstructive jaundice, define the location of the bile duct obstruction, identify ampullary and periampullary lesions, and establish the diagnosis of IPMT if mucous is seen extruding through a fish-mouth papillary opening. The finding of superimposable bile duct and pancreatic duct strictures (i.e., the double-duct sign) on ERCP is highly suggestive of a pancreatic head cancer ( Fig. 5313 ) but benign processes, such as chronic pancreatitis or autoimmune pancreatitis, can also produce a double-duct sign. The role of ERCP in the management of patients with a mass on CT is more controversial. Many surgeons (including me) would argue that ERCP is not helpful since a malignant lesion cannot be entirely excluded by ERCP. Thus, resection is indicated, regardless of the ERCP findings.
The Role of Biopsy
Biopsy to confirm the presence and identify the type of cancer is usually required before chemoradiation therapy of unresectable pancreatic tumors or neoadjuvant treatment of resectable tumors. Percutaneous biopsy, performed with either CT or ultrasound guidance, or transduodenal biopsy, performed with endoscopic ultrasound guidance, is routinely employed in these situations. Considerable controversy, however, surrounds the question of whether all patients with potentially resectable tumors should undergo preoperative biopsy. That biopsy might yield false-negative results, and, at least theoretically, transcutaneous biopsy could promote intraperitoneal dissemination of the tumor. Performing the biopsy via a transduodenal approach would eliminate the concern of tumor dissemination, but even with this approach a positive biopsy merely confirms the decision
Figure 53-13 Endoscopic retrograde cholangiopancreatography (ERCP) showing double-duct sign. A stent traverses the bile duct stricture. Note closely adjacent bile and pancreatic duct strictures in patient with adenocarcinoma of the pancreatic head.
for resection and a negative biopsy is inconclusive. For these reasons, most surgeons would not recommend routine preoperative biopsy for confirmation of the diagnosis in the management of patients with potentially resectable lesions. However, if this policy of not performing preoperative biopsy is followed, 5% to 10% of patients undergoing resection for suspected cancer will be found to have benign lesions.
Staging of Pancreatic Cancer
The American Joint Committee on Cancer (AJCC) staging system is widely used to stage pancreatic cancer ( Table 533 ). This system uses the TNM classification to define the tumor extent, nodal metastases, and distant metastases. [43] Tis, which denotes in situ cancer, corresponds to PanIN-3, that is, the most advanced of the ductal cancer precursor lesions (see Fig. 5312 ). T1 and T2 cancers are confined to the pancreas and either less than 2 cm or greater than 2 cm in diameter. T3 and T4 lesions extend beyond the pancreas. T3 lesions are considered to be potentially resectable because they do not involve the celiac axis or superior mesenteric artery. They may involve the portal/superior mesenteric vein and resection of the tumor may mandate venous resection and reconstruction. Long-term survival of patients undergoing major venous resection is poor, and it is not clear that those patients actually benefit from resection. T4 lesions are considered to be unresectable because they involve the critical peripancreatic arteries. N1 lesions have positive regional nodes and M1 lesions have distant metastases, whereas N0 and M0 lesions lack both of these features. Distant metastases are common, and they are most frequently located in the liver or lung or on the peritoneal surfaces of the abdomen. Stages 1 and 2 cancers are amenable to resection. Poor prognostic signs include aneuploidy, large tumor size (T2), the presence of positive regional nodes (N1), and an incomplete resection at the pancreatic or retroperitoneal margin. The latter is most closely associated with poor TABLE 53-3 -- American Joint Committee on Cancer: TNM System for Staging of Pancreatic Cancer Stage Stage 0 Stage IA Stage IB Stage IIA Stage IIB Tis T1 T2 T3 T1 T2 T3 Stage III Stage IV T4 Any T T Status N0 N0 N0 N0 N1 N1 N1 Any N Any N N Status M0 M0 M0 M0 M0 M0 M0 M0 M1 M Status
Adapted from AJCC Cancer Staging Handbook, 6th ed. New York, Springer, 2002, pp 179188.
1671
survival. Stages 3 and 4 cancers are considered to be unresectable, either because of distant metastases (stage 4) or because of major arterial involvement (stage 3). Mean survival for patients with stage 3 tumors ranges from 8 to 12 months, whereas that for patients with stage 4 tumors is only 3 to 6 months. Preoperative staging of pancreatic cancers is usually accomplished using one or more of the standard imaging techniques, such as CT, MRI, or ultrasonography. High-resolution helical CT, with phased imaging for visualization of the pancreas and major peripancreatic vessels, is the most widely used method of evaluating tumor resectability. Circumferential encasement, invasion, or occlusion of the portal vein/superior mesenteric vein and/or the superior mesenteric artery is generally considered to be a sign of unresectability although, strictly speaking, resection is still technically possible if only the venous structures are involved. Partial encasement (i.e., involvement of 30% to 60% of the circumference of the vein) may result in distortion of the vein and the appearance of a teardrop-shaped rather than a round structure. In our experience, this is associated with a low resectability rate. Other CT changes suggestive of unresectability include extension beyond the pancreatic capsule and into the retroperitoneum, involvement of neural or nodal structures surrounding the origin of either the celiac axis or superior mesenteric artery, and extension of the tumor along the hepatoduodenal ligament. MRI provides information that is similar to that obtained by CT, but combining MRI with CT does not seem to provide additional information. There have been recent claims that endoscopic ultrasound may provide staging information that is superior to that obtained by either CT or MRI[44] since it permits better visualization of nodal structures and allows for determination of the depth of invasion for tumors involving the duodenal wall. Unfortunately, endoscopic ultrasound is operator dependent, and its applicability to preoperative staging is dependent on the locally available expertise. Theoretically, PET scanning might be useful in staging pancreatic cancers by permitting detection of small metastatic lesions that escape detection by standard cross-sectional imaging techniques, but the sensitivity and specificity of PET scans in this setting are not currently known.
The Role of Laparoscopy in Staging
The role of staging laparoscopy in the management of pancreatic cancer is controversial. Proponents claim that 20% to 40% of patients believed to have stage 1 or 2 disease have unrecognized small metastases to peritoneal surfaces (such as diaphragm, liver) and that those metastases can be laparoscopically detected, thus preventing a needless laparotomy. Our own experience indicates that, with high-quality modern imaging techniques, few patients with pancreatic head tumors are deemed unresectable at operation merely because they are found to have metastases that might have been found by laparoscopy. More commonly, unappreciated vascular involvement is the operative finding that prevents resection and, in those patients, benefit can still be provided by performance of bilioenteric, and possibly a gastroenteric, bypass. The argument for preliminary laparoscopy would be more compelling in the case of patients with body or tail lesions since, with left-sided lesions, neither obstructive jaundice nor gastric outlet obstruction are likely to develop and, therefore, there is no role for surgical bypass in those patients. For these reasons, it is my practice to perform staging laparoscopy for potentially resectable body or tail lesions. For pancreatic head lesions that are deemed resectable by preoperative routine staging, laparoscopy is not advocated.
Resectional Surgery for Pancreatic Head and Uncinate Process Tumors
Tumors of the head, neck, and uncinate process of the pancreas account for approximately 70% of pancreatic tumors. They are generally resected by pancreaticoduodenectomy, with or without preservation of the pylorus and proximal duodenum (see Fig. 5310 ). The two procedures yield similar survival rates and have similar morbidity. The pylorus-preserving operation is technically easier and faster, but it may be associated with a higher incidence of and more prolonged delayed gastric emptying. Both operations can be performed through a midline, bilateral subcostal, or right subcostal incision. After a preliminary search for metastases or other reasons to abort resection, the retroperitoneal area behind the pancreatic head, the hepatoduodenal ligament, and the base of the transverse mesocolon should be carefully examined. The finding of involved peripancreatic nodes, even along the cephalad border of the pancreas near the portal vein, does not preclude resection although it worsens prognosis. A tunnel (the tunnel of love) should be developed behind the neck of the pancreas and anterior to the underlying visceral vessels before concluding that the lesion can be resected. Once it is deemed resectable, irreversible steps can be taken. The gallbladder is usually removed and the common bile duct is divided above the duodenum. The proximal gastrointestinal tract is divided at the level of the gastric antrum (standard Whipple) or first part of the duodenum (pyloruspreserving Whipple). The proximal jejunum is divided and the neck of the pancreas is transected. Finally, the uncinate process of the pancreas is resected from the retroperitoneum along the lateral surface of the superior mesenteric artery and the specimen is removed. Reconstruction can be accomplished in a variety of ways but the most common involves a pancreaticojejunostomy (as an end-to-end or as an endto-side), an end-to-side hepaticojejunostomy, and then an antecolic end-to-side gastrojejunostomy (standard Whipple) or duodenojejunostomy (pylorus-preserving Whipple) (see Fig. 5310 ). Frequently, drains and a feeding jejunal tube are placed, the latter in anticipation of the delayed gastric emptying that frequently occurs, and the abdomen is closed. Several reports, mostly from Japan, have suggested extending the Whipple procedure to include an extensive retroperitoneal lymphadenectomy and, in some cases,
1672
total pancreatectomy along with possible major venous resection, can result in increased long-term patient survival. This so-called extended radical pancreaticoduodenectomy is frequently followed by a prolonged delay in the return of gastrointestinal function and, on occasion, with incapacitating and persistent diarrhea. A recent prospective, randomized study in the United States has failed to show a survival benefit for this extended radical operation and it is rarely performed in the United States.[45]
Complications of Pancreaticoduodenectomy
When performed by experienced surgeons in high-volume centers, the operative mortality of pancreaticoduodenectomy is 2% to 4%. Anastomotic leaks, intra-abdominal abscesses, and delayed gastric emptying account for most of the perioperative complications after pancreaticoduodenectomy. Leakage from the pancreatic anastomosis, resulting in a pancreatic fistula, occurs in 15% to 20% of patients. With adequate drainage, these fistulas usually heal within several weeks. Some pancreatic surgeons have claimed that administration of a somatostatin analogue to patients undergoing pancreaticoduodenectomy for cancer could reduce the incidence and/or duration of pancreatic fistula, but a recent randomized trial addressing this subject concluded that prophylactic administration of a somatostatin analogue in this setting is not beneficial. Biliary fistulas are much less common than pancreatic fistulas after pancreaticoduodenectomy, but they also usually heal if adequate drainage is achieved. Delayed gastric emptying occurs in 15% to 40% of patients and almost always resolves with time. Its occurrence is unpredictable and it may persist for several weeks or even months. The basis for delayed gastric emptying after pancreaticoduodenectomy is not clear. Some have suggested that it is the result of removing the cells in the duodenum that secrete the promotility hormone motilin. Erythromycin, which has a structure that resembles motilin, acts as an agonist at motilin receptors and has been used to treat these patients. Success has been reported using this approach, but my experience with this use of erythromycin has been disappointing. The endocrine pancreas has considerable functional reserve and, for that reason, most patients do not develop diabetes after resection of the pancreatic head. In fact, some patients with tumor-induced diabetes may have resolution of their diabetes after resection. In contrast pancreatic malabsorption and steatorrhea are relatively common long-term problems. It may reflect exocrine secretory insufficiency, an obstruction at the pancreatic-jejunal anastomosis, or poor postoperative mixing of secreted enzymes with food. Although distressing, postoperative malabsorption and steatorrhea are rarely
incapacitating but require exogenously administered pancreatic enzymes.

Results of Pancreaticoduodenectomy for Pancreatic Cancer
Long-term survival, when the operation is performed for ductal adenocarcinoma, is unusual. Five-year survival rates of 15% to 20% have been reported from some centers, but rates of 10% to 15% are more common, and most of those patients who survive for 5 years succumb over the subsequent 5 years. A number of factors affect the survival rate. Perhaps the most influential is the presence or absence of tumor at the resection margin. Negative margins, in one study, were associated with a 26% 5-year survival, whereas positive margins were associated with an 8% 5-year survival. Extending the resection to a total pancreatectomy or to include more radical retroperitoneal lymph node resection does not appear to increase long-term survival. Other factors that affect long-term survival include tumor diameter, diploid/aneuploid DNA content, and lymph node status. The 5-year survival of nodepositive patients, in one series, was reported to be 14%, whereas 5-year survival of node-negative patients was 36%.
Resectional Surgery for Pancreatic Body and Tail Tumors
Most body and tail cancers have already metastasized to distant sites or extended locally to involve nodes, nerves, or major vessels by the time of diagnosis. Splenic vein involvement and/or occlusion is not uncommon and, by itself, is not a sign of nonresectability. On the other hand, involvement of the splenic/superior mesenteric vein confluence generally precludes resection. Resection involves a distal pancreatectomy either with or without concomitant splenectomy. Splenectomy should be performed for malignant tumors, but splenic preservation is not contraindicated when benign tumors are being removed. After a thorough search for metastatic disease, the operation starts by dividing the gastrocolic omentum, short gastric vessels, and peritoneal reflections around the body and tail of the pancreas to elevate the spleen and pancreatic tail out of the retroperitoneum. The splenic artery and vein, in that order, are ligated and the pancreas is transected, to the right of the tumor leaving an uninvolved margin. The resection margin can be either closed with a stapler or imbricated with sutures. Ideally, the transected duct should be suture ligated separately. Complications of distal pancreatectomy include subphrenic abscess, which may occur in 5% to 10% of patients, and pancreatic duct leak, which has been reported to occur in up to 20% of patients. Both complications can usually be managed nonoperatively by percutaneous drainage and reoperation is rarely required. Pancreatic duct leak usually results in the formation of a fluid collection that resembles a pseudocyst and that, following drainage, becomes a pancreaticocutaneous fistula. Some surgeons have advocated prophylactic administration of a somatostatin analogue to reduce the incidence or duration of pancreatic fistula after distal pancreatectomy, but there are no convincing data to support this practice. Somatostatin administration frequently reduces the output of these pancreatic fistulas, but it does not appear to alter the time of fistula closure. Only 10% of cancers involving the tail or body of the pancreas are resectable at the time of diagnosis. The 5-year survival of patients who are deemed resectable is
1673
somewhat lower than that of patients with resectable cancer of the pancreatic head (8% to 14%). Factors affecting long-term survival are similar to those for pancreatic head cancers.
Palliative Nonsurgical Treatment for Pancreatic Cancers
Establishing the diagnosis and relieving symptoms of jaundice, gastric outlet obstruction, and pain are the goals of palliative nonsurgical treatment. Tissue diagnosis can usually be made by CT- or ultrasound-guided percutaneous fine-needle aspiration of either the tumor or its metastases. Transduodenal fine-needle aspiration of the tumor with endoscopic ultrasound guidance and duct cytology obtained by brushings are alternative, but less frequently successful, methods of establishing the diagnosis. Decompression of the obstructed biliary tract can be achieved using either an endoscopic or a percutaneous-transhepatic approach. The former has been shown, in randomized trials, to yield better results with fewer associated complications. At the time of ERCP, a transpapillary stent is placed across the obstructed segment of bile duct. Either plastic or expandable metal stents can be used, but metal stents give more complete and more long-lasting relief of jaundice. Plastic stents can become obstructed by tumor or debris and, as a result, they must be changed every 2 to 3 months. The percutaneous-transhepatic approach to duct decompression is usually reserved for patients in whom, for technical reasons, a stent cannot be placed endoscopically. Pancreatic tumors can extend into and obstruct the duodenum leading to gastric outlet obstruction. This commonly occurs in the second portion of the duodenum in patients with pancreatic head cancers. Pancreatic body tumors can invade the third or fourth portion of the duodenum and also cause obstruction. Recent reports indicate that many of these patients can be palliated by endoscopic placement of expandable endoluminal metal stents into the duodenum.[46] For lesions that are not amenable to stents, surgical gastrojejunostomy may be required. Pain, which is a common symptom of pancreatic cancer, is usually caused by tumor invasion of the peripancreatic neural plexus. Most patients can be adequately treated with orally or transcutaneously administered analgesics. Narcotic medications may be required. When or if this fails, percutaneous CT-guided or endoscopic ultrasoundguided celiac plexus block may be helpful.
Palliative Surgical Management of Pancreatic Cancer
Most of the symptoms experienced by patients with unresectable pancreatic cancer can be relieved by nonsurgical means. Surgical palliation is, for the most part, employed for patients who are undergoing laparotomy for anticipated resectable disease and found to be unresectable at the time of surgery. In that situation, biliary tract decompression can be achieved by creating either a cholecystojejunostomy or a choledochojejunostomy. The former is most appropriate for patients with nondilated ducts in whom the cystic ductcommon bile duct junction is far from the pancreatic tumor. Choledochojejunostomy, on the other hand, should be performed when the tumor is close to or at the cystic ductcommon bile duct junction and the bile duct is dilated. Either a loop or a Roux-en-Y segment of jejunum can be used with similar results. Duodenal obstruction can be managed by creation of a side-to-side gastrojejunostomy in which an antecolic jejunal loop is anastomosed to the posterior wall of the gastric antrum. Duodenal obstruction, even in advanced pancreatic cancer, occurs in less than 25% of patients and, therefore, considerable controversy surrounds whether a prophylactic gastrojejunostomy should be performed prior to the development of gastric outlet obstruction. The issue is further complicated by the fact that a gastrojejunostomy can cause delayed gastric emptying and result in symptoms identical to those experienced by patients with duodenal obstruction. It is my practice to perform gastrojejunostomy selectively (i.e., for patients whose tumor is locally advanced but without distant metastases [stage 3 lesions]) since they have an expected survival of 8 to 12 months. I would not perform prophylactic gastrojejunostomy for patients with distant metastasis (stage 4 lesions) since their expected survival is only 3 to 6 months and they are less likely to develop duodenal obstruction prior to death. Palliation of pain can be achieved, intraoperatively, by injecting alcohol into the celiac plexus and some surgeons routinely perform operative celiac plexus block at the time of surgical palliation. This is usually accomplished by injecting 15 to 20 mL of 50% ethanol into the celiac plexus on either side of the aorta and, in one randomized, prospective trial, this treatment has been reported to reduce postoperative pain and the need for postoperative analgesics in patients with unresectable pancreatic cancers. However, many patients with unresectable pancreatic cancer can be successfully managed with minimal or no narcotic analgesics and, when more severe pain occurs, results similar to those achieved by intraoperative chemical splanchnicectomy can be achieved using a percutaneous approach. For these reasons, many surgeons do not routinely perform intraoperative celiac plexus blocks.
Chemoradiation Therapy
Many different protocols for chemoradiation treatment of recurrent or unresectable pancreatic cancer have been described. The body of literature on this subject is quite large and beyond the scope of this review. For the most part, the best results have been achieved using radiation therapy combined with either 5-fluorouracil or gemcitabine. Patients undergoing resection may also benefit from adjuvant chemoradiation therapy. The frequently quoted Gastrointestinal Tumor Study Group (GITSG) report suggested that the combination of 5-fluorouracil with radiation therapy could increase the 2-year survival rate for patients with tumor-free resection margins from 18% to 43%.[47] This approach, although based on the experience with a very small group of patients, has been generally accepted and is widely employed. Recently, the European Study Group for Pancreatic Cancer (ESPAC) has
1674
reported a different experience.[48] They conducted a large study evaluating adjuvant chemotherapy (5-fluorouracil), radiation therapy, both, or neither. A survival benefit was observed for those undergoing chemotherapy but not for those receiving either radiation therapy alone or combined chemoradiotherapy. Intraoperative radiation therapy, as adjuvant treatment, has been evaluated but no benefit has been found. Neoadjuvant chemoradiation has also been used for patients believed to have resectable lesions, but good prospective, randomized studies have not been reported. Some have claimed up to 15% of selected patients with locally advanced lesions deemed to be unresectable can be made resectable by aggressive administration of chemoradiation therapy. Although the specimens ultimately resected in these cases have frequently contained relatively little viable tumor, the long-term benefits achieved by resecting these advanced tumors are not known.
PANCREATIC AND PANCREATICODUODENAL TRAUMA Three percent to 12% of patients with severe abdominal trauma have pancreatic injury. On average, these patients have 3.5 other organs injured and isolated pancreatic injury is uncommon. Roughly two thirds of pancreatic injuries are the result of penetrating trauma, and the remaining one third are due to blunt trauma. Blunt abdominal trauma can cause damage that ranges from a mild contusion to a severe crushing injury. Classically, blunt injury is the result of midline upper abdominal trauma inflicted by objects as diverse as automobile seatbelts and bicycle handlebars. In these cases, the neck and body of the pancreas can be injured as they pass over the vertebral column. The mortality rate for pancreatic trauma is closely related to the nature of the injury. Thus, the mortality rate associated with blunt trauma is 17% to 19%, whereas, with stab wounds, it is 3% to 5%, with gunshot wounds it is 15% to 22%, and with shotgun wounds it is 46% to 56%.
Diagnosis
Serum amylase levels are elevated in most patients with significant pancreatic trauma, but they are also increased in up to 90% of severe abdominal trauma patients who do not have pancreatic injury. Thus, measurement of amylase at the time of hospital admission is not helpful in identifying those with pancreatic injury. On the other hand, a progressive rise in serum amylase activity is a more specific indicator of pancreatic injury. Contrastenhanced CT, which is increasingly being used as a screening test of patients with major abdominal trauma, is the most useful noninvasive method of evaluating patients with suspected pancreatic injury, but even high-quality CT examinations may be either falsely negative or falsely positive. Changes suggestive of pancreatic injury include the presence of peripancreatic fluid collections, focal or diffuse enlargement of the gland, parenchymal disruption, and areas of diminished contrast perfusion. Frequently, serial CT examinations are needed because the changes indicative of pancreatic injury may not be present on examinations performed very soon after injury. Ultimately, the management of pancreatic injuries depends on the presence of pancreatic duct disruption, major associated vascular injury, and/or significant injury to peripancreatic organs, especially the duodenum. Vascular and duodenal injury can be evaluated at the time of exploratory laparotomy, but the integrity of the pancreatic duct may be difficult to determine at the time of operation unless the pancreas is transected or pancreatic juice is seen to be extravasating from the region of the duct. In the absence of these findings, pancreatography is required to localize or exclude a duct injury. Intraoperative ductography can be performed either transduodenally after duodenotomy or, in a prograde fashion, from the cut end of the duct after resection of the pancreatic tail. Both of these approaches are difficult and dangerous since, if the suture line fails, they can result in either a pancreatic or a duodenal fistula. Thus, if the patient is stable and duct injury is suspected, the surgeon should consider preoperative ERCP to determine whether or not the duct is intact. ERCP may have an even more important role in the evaluation of patients managed by delayed exploratory laparotomy after abdominal trauma since, in this setting, delineation of the site and extent of duct injury helps plan a definitive procedure.
Management
The Pancreas Organ Injury Scale,[49] developed by the American Association for the Surgery of Trauma, can be used to grade pancreatic injuries ( Box 536 ). Management of pancreatic injuries is determined by the site and grade of that injury. Grade I injuries, which involve minor contusions or lacerations of the gland without duct injury, Box 53-6. Grading and Treatment of Pancreatic Injuries * Grade 1: Minor contusion or laceration without duct injury Treatment: Observation alone Grade II: Major contusion or laceration without duct injury Treatment: Dbridement, drainage, possible repair Grade III: Distal transection or injury with duct injury Treatment: Distal resection, possible Roux-en-Y drainage Grade IV: Proximal transection or injury involving ampulla or Grade V: Massive disruption of the pancreatic head Treatment: Damage control, hemostasis/drainage; resection and possible Roux-en-Y drainage; triple-tube decompression; duodenal diverticularization; pancreaticoduodenectomy
* Adapted from Moore EE, Cogbill TH, Malangoni MA, et al: Organ injury scaling: II. Pancreas, duodenum, small bowel, colon, and rectum. J Trauma 30:14271429, 1990.
1675
should be treated expectantly. Drainage is usually not needed. On the other hand, grade II injuries, which involve major contusions or lacerations of the gland without duct injury, are traditionally treated by dbridement, adequate hemostasis, and placement of sump drains. Some surgeons advocate repairing the disrupted capsule and placing an omental patch into areas of devitalized parenchyma. Transection of the neck/body/tail of the pancreas or parenchymal injury to those areas accompanied by duct injury is considered a grade III injury. The critical feature of these injuries is violation of the major pancreatic duct. Preoperative ERCP in stable patients may identify the presence and location of the duct injury in some patients but, because most patients are not stable enough to undergo preoperative ERCP, the status of the duct is uncertain at the time of operation. Occasionally, in such patients,
the intraoperative finding of a transected gland, fat necrosis, and/or extravasation of clear fluid from the region of the duct establishes the presence of a duct injury. The classic treatment of grade III injuries involves resection of the pancreas to the left of the injury along with splenectomy. Because of the fear of overwhelming postsplenectomy sepsis, splenic conservation might be considered for children with grade III injuries. When the location of a grade III injury is to the right of the pancreatic neck, the surgeon should consider dbriding the site of injury and conserving, rather than resecting, the uninjured body and tail by anastomosing it to a Roux-en-Y jejunal limb. Grades IV and V injuries present the greatest surgical challenge because they involve the head of the pancreas and, in addition, they also usually involve the adjacent duodenum and/or papilla of Vater. Because of their location, grades IV and V injuries are also the ones that are most likely to involve major adjacent vascular structures. The initial goal in managing these multiply injured patients is that of obtaining hemostasis, minimizing contamination by repairing torn bowel, and repairing associated injuries. In unstable patients or in those with major associated injuries, this damage control approach may be all that is done at the initial operation, and definitive repair may be postponed until the patients condition has improved. On the other hand, some patients may require emergent pancreaticoduodenectomy as the initial procedure if they have uncontrollable hemorrhage from the head of the pancreas, injury to adjacent major vessels that can only be controlled by removing the pancreatic head, or devitalization of the duodenal C loop and pancreatic head. Short of radical resection, however, other options designed to divert gastric, pancreatic, and biliary secretions away from the duodenum should be considered for the management of patients with grades IV and V injuries. These include duodenal diverticularization, pyloric exclusion/gastrojejunostomy, and triple-tube decompression. Duodenal diverticularization is accomplished by performing antrectomy and gastrojejunostomy to achieve gastric diversion, choledochostomy to divert bile if the ampulla is injured, tube duodenostomy for decompression of the duodenum, suture repair of any duodenal injuries, and extensive periduodenal and peripancreatic drainage. Pyloric exclusion with gastrojejunostomy ( Fig. 5314 ) is a simpler method of protecting the injured duodenum, and it does not permanently alter function.[50] Through a gastrotomy, the pylorus is closed with a purse-string suture and an antecolic gastrojejunostomy is performed at the site of the gastrotomy. As with duodenal diverticularization, duodenal injuries should be closed and the area should be extensively dbrided. Use of a slowly absorbable suture for pyloric closure results in diversion of gastric secretions from the duodenum for approximately 2 to 3 weeks but, after 21 days (i.e. dissolution of the suture), more than 90% of patients are found to have a patent and functioning pylorus. Triple-tube decompression involves placement of a gastrostomy tube for drainage of the stomach, drainage of the duodenum via a tube passed retrogradely through a jejunostomy, and antegrade passage of a jejunostomy tube for provision of enteral nutrition. The rapidity and ease of this procedure make it an attractive choice, but subsequent dislodgment of the duodenal tube and/or inadequate diversion provided by the gastrostomy tube may limit its effectiveness.
Figure 53-14 Pyloric exclusion with gastrojejunostomy. A, After repairing the duodenal injury, dbriding the area of pancreatic injury, and achieving adequate hemostasis, a gastrotomy is performed and the pylorus is closed with a slowly reabsorbable suture. B, A gastrojejunostomy is performed and the area of injury is widely drained. (1978 Baylor College of Medicine.)
1676
Delayed Complications
The major delayed complications of severe pancreatic injuries involve the development of sterile or infected pseudocysts or other fluid collections and the ultimate development of a pancreatic duct stricture at a site of duct injury or transection. Classically, the duct stricture usually develops in the portion of the pancreas that overlies the spinal column. In many patients, the initial injury may go undetected, and even the trauma itself may be forgotten until the patient returns months or years later with recurrent episodes of obstructive chronic pancreatitis involving the pancreatic tail but sparing the pancreatic head. Management involves distal pancreatectomy.
Selected References
Bradley EL III: A clinically based classification system for acute pancreatitis: Summary of the International Symposium on Acute Pancreatitis. Atlanta, GA, September 1113, 1992. Arch Surg 128:586590, 1993. A conference summary article that outlines the definitions of terms currently used to describe the complications of pancreatitis. Burch JM, Moore EE: Duodenal and pancreatic trauma. In Taylor MB, Gollan JL, Steer ML, Wolfe MM (eds): Gastrointestinal Emergencies. Baltimore, Williams & Wilkins, 1997, pp 9951001. A recent review of current surgical treatment for pancreatic and duodenal injuries. Case RM: Pancreatic exocrine secretion mechanisms and control. In Beger HG, Warshaw AW, Buchler MW, et al (eds): The Pancreas. Oxford, Blackwell Science, 1998, pp 63100. An up-to-date summary of current concepts regarding the physiology of pancreatic exocrine secretion. Chey WY, Chang TM: Neural hormonal regulation of exocrine pancreatic secretion. Pancreatology 1:320335, 2001. An up-to-date review of pancreatic exocrine physiology. Hamano H, Kawa S, Akira H, et al: High serum IgG4 concentrations in patients with sclerosing pancreatitis. N Engl J Med 344:732738, 2001. Report describing elevated IgG4 levels in patients with autoimmune pancreatitis and their favorable response to treatment with steroids. Hruban RH, Adsay NV, Albores-Saavedra J, et al: Pancreatic intraepithelial neoplasia: A new nomenclature and classification system for pancreatic duct lesions. Am J Surg Pathol 25:579586, 2001. A detailed description of criteria used to define ductal cancer precursor lesions in the pancreas. Kloppel G, Kosmahl M: Cystic lesions and neoplasms of the pancreas: The features are becoming clearer. Pancreatology 1:648655, 2001. A review of cystic exocrine pancreatic lesions that focuses on their pathology and growth characteristics. Opie EL: The etiology of acute hemorrhagic pancreatitis. Bull Johns Hopkins Hosp 12:182192, 1901. The classic paper by Opie that led to the common channel theory as an explanation for gallstone pancreatitis. Skandalakis LJ, Rowe JS Jr, Gray SW, et al: Surgical embryology and anatomy of the pancreas. Surg Clin North Am 73:661697, 1993. A comprehensive overview of embryology and anatomy focused on issues relevant to pancreatic disease and pancreatic surgery. Steer ML: The early intra-acinar cell events which occur during acute pancreatitis: The Frank Brooks Memorial Lecture. Pancreas 17:3137, 1998. A summary of recent experimental data defining the early events that occur within pancreatic acinar cells and that ultimately lead to pancreatitis. Steer ML, Waxman I, Freedman S: Chronic pancreatitis. N Engl J Med 332:14821490, 1995. A review of current concepts regarding the pathophysiology, diagnosis, and treatment of chronic pancreatitis. Whipple AO, Parsons WB, Mullens CR: Treatment of carcinoma of the ampulla of Vater. Ann Surg 102:763769, 1935. The first successful two-stage radical pancreaticoduodenectomy for periampullary carcinoma. Whitcomb DC, Gorry MC, Preston RA, et al: Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet 14:141145, 1996. Initial report defining the genetic abnormality underlying the most common form of hereditary pancreatitis. Yeo CJ, Cameron JL, Lillemoe KD, et al: Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma: II. Randomized controlled trial evaluating survival, morbidity, and mortality. Ann Surg 236:355368, 2002. Randomized, prospective trial of conventional Whipple resection versus extended Whipple resection for pancreatic cancer showing no benefit for the more radical procedure.
Copyright 2004 Elsevier Inc. All rights reserved.
www.mdconsult.com
References
1. Palade
G: Intracellular aspects of the process of protein secretion. Science 189:347358, 1975. S: Trafficking of lysosomal enzymes in normal and disease states. J Clin Invest 77:16, 1986.
2. Kornfeld
3. Case RM Argent BE: Pancreatic duct cell secretion: Control and mechanisms of transport. In Go VLW, DiMagno EP, Gardner JD, et al (eds): The Pancreas: Biology, Pathobiology, and Disease, 2nd ed. New York, Raven Press, 1993, pp 301350. 4. Case
RM: Pancreatic exocrine secretion: Mechanisms and control. In Beger HG, Warshaw AW, Buchler MW, et al (eds): The Pancreas. Oxford, Blackwell Science, 1998, pp 63
100.
5. Rossi
L, Pfutzer RH, Parvin S, et al: SPINK/PSTI mutations are associated with tropical pancreatitis in Bangladesh: A preliminary report. Pancreatology 1:242245, 2001.
1677
6. Opie
EL: The etiology of acute hemorrhagic pancreatitis. Bull Johns Hopkins Hosp 12:182192, 1901. MM, Saluja AK, Runzi M, et al: Pancreatic duct obstruction triggers acute necrotizing pancreatitis in the opossum. Gastroenterology 104:853861, 1993. G: Progression from acute to chronic pancreatitis: A pathologists view. Surg Clin North Am 79:801814, 1999.
7. Lerch
8. Kloppel 9. Gates
LD, Ulrich CD, Whitcomb DC, et al: Hereditary pancreatitis gene defects and their implications. Surg Clin North Am 79:711722, 1999. H, Kawa S, Horiuchi A, et al: High serum IgG4 concentrations in patients with sclerosing pancreatitis. N Engl J Med 344:732738, 2001.
10. Hamano 11. Cohn 12. Steer
JA, Friedman KJ, Noone PG, et al: Relation between mutations of the cystic fibrosis gene and idiopathic pancreatitis. N Engl J Med 339:653658, 1998. ML: The early intra-acinar cell events which occur during acute pancreatitis: The Frank Brooks Memorial Lecture. Pancreas 17:3137, 1998. M, Gorelick F, Anderson JM, et al: Effect of caerulein hyperstimulation on the paracellular barrier of rat exocrine pancreas. Gastroenterology 108:18631873, 1995. AM, Heath DI, Hurley P, et al: Trypsinogen activation peptide assays in the early prediction of severity of acute pancreatitis. Lancet 335:48, 1990.
13. Fallon
14. Gudgeon 15. Foitzik
T, Bassi DG, Schmidt J, et al: Intravenous contrast medium accentuates the severity of acute necrotizing pancreatitis in the rat. Gastroenterology 106:207214, 1994.
16. Kaiser AM, Grady T, Gerdes D, et al: Intravenous contrast medium does not increase the severity of acute necrotizing pancreatitis in the opossum. Dig Dis Sci 40:15471553, 1995. 17. Ranson 18. Imrie
JHC, Rifkind KM, Roses DF, et al: Prognostic signs and the role of operative management in acute pancreatitis. Surg Gynecol Obstet 139:6981, 1974.
CW, Benjamin IS, Ferguson JC, et al: A single-centre double-blind trial of Trasylol therapy in primary acute pancreatitis. Br J Surg 65:337341, 1978. EJ, Ranson JHC, Naidich DP, et al: Acute pancreatitis: Prognostic value of CT. Radiology 156:767772, 1985.
19. Balthazar 20. Khan
AA, Parekh D, Young C, et al: Improved prediction of outcome in patients with severe acute pancreatitis by the APACHE II score at 48 hours after hospital admission compared with the APACHE II score at admission. Arch Surg 137:11361140, 2002. P, Bassi C, Vesentini S, et al: A randomized multicenter clinical trial of antibiotic prophylaxis of septic complications in acute necrotizing pancreatitis with imipenem. Surg Gynecol Obstet 176:480483, 1993. V, Kemppainen E, Puolakkainen P, et al: Early antibiotic treatment in acute necroitizing pancreatitis. Lancet 346:663667, 1995. R, Siddiqi F, Pohl D: Role of antibiotics in acute pancreatitis: A meta-analysis. J Gastrointest Surg 115:15131517, 1998. EJ, Hop WC, Lange JF, et al: Controlled clinical trial of selective decontamination for the treatment of severe acute pancreatitis. Ann Surg 222:5765, 1995.
21. Pederzoli
22. Sainio 23. Golub 24. Luiten
25. 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 2:979983, 1988. OR, Nitsche R, Ludtke R, et al: Early ERCP and papillotomy compared with conservative treatment for acute biliary pancreatitis. The German Study Group on Acute Biliary Pancreatitis. N Engl J Med 336:237242, 1997. ST, Lai CS, Mok FPT, et al: Early treatment of acute biliary pancreatitis by endoscopic papillotomy. N Engl J Med 328:228232, 1993.
26. Folsch
27. Fan
28. Bradley EL: A clinically based classification system for acute pancreatitis: Summary of the International Symposium on Acute Pancreatitis. Atlanta, GA, September 1113, 1992. Arch Surg 128:586590, 1993. 29. Gerzof 30. Runzi
SG, Banks PA, Robbins AH, et al: Early diagnosis of pancreatic infection by computed tomographyguided aspiration. Gastroenterology 93:13151320, 1987.
M, Layer P: Nonsurgical management of acute pancreatitis: Use of antibiotics. Surg Clin North Am 79:759765, 1999. RW, Muellhaupt B: The natural history of pain in alcoholic chronic pancreatitis. Gastroenterology 116:12521257, 1999.
31. Ammann
32. Toskes
PP: Medical management of chronic pancreatitis. Scand J Gastroenterol Suppl 208:7480, 1995. HC, Jansen JB, van Dongen R: Long-term results of bilateral thoracoscopic splanchnicectomy in patients with chronic pancreatitis. Br J Surg 89:158162, 2002. PF, Rochelle EL: Modified Puestow procedure for retrograde drainage of pancreatic duct. Ann Surg 152:10371043, 1960.
33. Buscher
34. Partington 35. Ho
HS, Frey CF: The Frey procedure: Local resection of pancreatic head combined with lateral pancreaticojejunostomy. Arch Surg 136:13531358, 2001. LW, Longmire WP: Preservation of the pylorus during pancreaticoduodenectomy. Surg Gynecol Obstet 146:959962, 1978.
36. Traverso 37. Beger
HG, Schlosser W, Friess H, et al: Duodenumpreserving head resection in chronic pancreatitis changes the normal course of the disease: Single-center 26-year experience. Ann Surg 230:512519, 1999.
38. Hruban RH, Adsay V, Albores-Saavedra J, et al: Pancreatic intraepithelial neoplasia: A new nomenclature and classification system for pancreatic duct lesions. Am J Surg Pathol 25:579586, 2001. 39. Lewandrowski K, Lee J, Southern J: Cyst fluid analysis of the differential diagnosis of pancreatic cysts: A new approach to the preoperative assessment of pancreatic cystic lesions. AJR Am J Roentgenol 164:815819, 1995. 40. McCormick CSF, Lemoine NR: Molecular biological events in the development of pancreatic cancer. In Beger HG, Warshaw AL, Buchler MW, et al (eds) : The Pancreas. Oxford, Blackwell Science, 1998, pp 907921. 41. Lowenfels
AB, Maisonneuve P, Whitcomb DC: Risk factors for cancer in hereditary pancreatitis. International Hereditary Pancreatitis Study Group. Med Clin North Am 84:565
573, 2000.
42. Safi
F, Schlosser W, Falkenreck S, et al: Ca 199 serum course and prognosis of pancreatic cancer. Int J Pancreatol 20:155162, 1996. Cancer Staging Handbook, 6th ed. New York, Springer, 2002, pp 179188. MJ: Accuracy of endoscopic ultrasound in diagnosing and staging pancreatic carcinoma. Pancreatology 1:625632, 2001.
43. AJCC
44. Wiersema 45. Yeo
CJ, Cameron JL, Lillemoe KD, et al: Pancreaticoduodenectomy with or without distal gastrectomy and extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma: II. Randomized controlled trial evaluating survival, morbidity, and mortality. Ann Surg 236:355368, 2002. DG, Baron TH: Endoscopic palliation of malignant gastric outlet obstruction using self-expanding metal stents: Experience in 36 patients. Am J Gastroenterol 97:7280,
46. Adler
2002.
1678
47. Kalser
NH, Ellenberg SS: Pancreatic cancer: Adjuvant combined radiation and chemotherapy following curative resection. Arch Surg 120:899903, 1985.
48. Neoptolemos JP, Dunn JA, Stocken DD, et al: Adjuvant chemoradiotherapy and chemotherapy in resectable pancreatic cancer: A randomized controlled trial. Lancet 358:1576 1585, 2001. 49. Moore 50. Burch
EE, Cogbill TH, Malangoni MA, et al: Organ injury scaling: II. Pancreas, duodenum, small bowel, colon, and rectum. J Trauma 30:14271429, 1990.
JM, Moore EE: Duodenal and pancreatic trauma. In Taylor MB, Gollan JL, Steer ML, Wolfe MM (eds): Gastrointestinal Emergencies. Baltimore, Williams & Wilkins, 1997, pp 9951001.
www.mdconsult.com
1679
Chapter 54 - Spleen
R. Daniel Beauchamp M.D. Michael D. Holzman M.D., M.P.H. Timothy C. Fabian M.D.
SPLENIC ANATOMY The spleen develops from mesenchymal cells in the dorsal mesogastrium during the 5th week of gestation. The spleen is located in the posterior left upper quadrant of the abdomen. The convex smooth surface of the spleen faces superiorly, posteriorly, and to the left in relation to the abdominal surface of the diaphragm. The diaphragm separates the spleen from the pleura; the left lower lobe of the lung; and the adjacent 9th, 10th, and 11th ribs. The costodiaphragmatic recess of the pleura extends down as far as the inferior border of the normal-sized spleen. The normal size and weight vary somewhat; in adults, the approximate size of the spleen is 12 cm in length, 7 cm in width, and 3 to 4 cm in thickness. The average spleen weight in an adult is 150 g, with a range of 80 to 300 g. The visceral relationships of the spleen are with the proximal greater curvature of the stomach, the tail of the pancreas, the left kidney, and the splenic flexure of the colon ( Fig. 541 ). The parietal peritoneum adheres firmly to the splenic capsule, except at the splenic hilum. The peritoneum extends superiorly, laterally, and inferiorly, creating folds that form the suspensory ligaments of the spleen. The splenophrenic and splenocolic ligaments are usually relatively avascular. The splenorenal ligament extends from the anterior left kidney to the hilum of the spleen as a two-layered fold in which the splenic vessels and the tail of the pancreas are invested. These two layers continue anteriorly and superiorly to the greater curvature of the stomach to form the two leaves of the gastrosplenic ligament through which the short gastric arteries and veins course. A fibroelastic capsule commonly known as the splenic capsule invests the organ, and from it trabeculae pass into the parenchyma, branching to form a trabecular network that subdivides the organ into small compartments. The splenic artery is a tortuous vessel that arises from the celiac trunk; it courses along the superior border of the pancreas ( Fig. 542 ). The branches of the splenic artery include the numerous pancreatic branches, the short gastric arteries, the left gastroepiploic artery, and the terminal splenic branches. The splenic artery divides into several branches within the splenorenal ligament before entering the splenic hilum, where they branch again into these trabeculae as they enter the splenic pulp. Small arteriolar branches leave the trabeculae, and their adventitial coat becomes replaced by a sheath of lymphatic tissue that accompanies the vessels and their branches until they divide into capillaries. It is these lymphatic sheaths that make up the white pulp of the spleen and that are interspersed along the arteriolar vessels as lymphatic follicles. The interface between the white pulp and the red pulp is known as the marginal zone. As the arterioles lose their sheaths of lymphatic tissue, they traverse the marginal zone and enter the red pulp, which is composed of large branching, thin-walled blood vessels called splenic sinuses and sinusoids, and thin plates of cellular tissue comprising the splenic chords. The venous sinusoids empty into the veins of the red pulp, and these veins drain back along the trabecular veins that empty into at least five major tributaries, ultimately joining to form the splenic vein in the splenorenal ligament. The splenic vein runs inferior to the artery and posterior to the pancreatic tail and body. It receives several
1680
Figure 54-1 A, Spleen, from the front: (1) diaphragm, (2) stomach, (3) gastrosplenic ligament, (4) gastric impression, (5) superior border, (6) notch, (7) diaphragmatic surface, (8) inferior border, (9) left colic flexure, (10) costodiaphragmatic recess, (11) thoracic wall. The left upper abdominal and lower anterior thoracic walls have been removed and part of the diaphragm (1) turned upward to show the spleen in its normal position, lying adjacent to the stomach (2) and colon (9), with the lower part against the kidney (B9 and 10, opposite). B, Spleen, in a transverse section of the left upper abdomen: (1) left lobe of liver, (2) stomach, (3) diaphragm, (4) gastrosplenic ligament, (5) costodiaphragmatic recess of pleura, (6) 9th
rib, (7) 10th rib, (8) peritoneum of greater sac, (9) spleen, (10) left kidney, (11) posterior layer of lienorenal ligament, (12) tail of pancreas, (13) splenic artery, (14) splenic vein, (15) anterior layer of lienorenal ligament, (16) lesser sac, (17) left suprarenal gland, (18) intervertebral disc, (19) abdominal aorta, (20) coeliac trunk, (21) left gastric artery. The section is at the level of the disc (18) between the 12th thoracic and 1st lumbar vertebrae and is viewed from below looking toward the thorax. The spleen (9) lies against the diaphragm (3) and left kidney (10) but separated from them by peritoneum of the greater sac (8). The peritoneum behind the stomach (2) forming part of the gastrosplenic (4) and ileorenal (15) ligaments belongs to the lesser sac (16). (A and B, From McMinn RMH, Hutchings RT, Pegington J, Abrahams PH: Color Atlas of Human Anatomy, 3rd ed. St. Louis, MosbyYear Book, 1993, pp 230231.)
short tributaries from the pancreas. The splenic vein joins the superior mesenteric vein at a right angle behind the neck of the pancreas to form the portal vein. The inferior mesenteric vein often empties into the splenic vein; it may also empty into the superior mesenteric vein at or near the confluence of the splenic vein and superior mesenteric vein.
SPLENIC FUNCTION The spleen has important hematopoietic functions during early fetal development, with both red and white blood cell production. By the 5th month of gestation, the bone marrow assumes the predominant role in hematopoiesis, and normally there is no significant hematopoietic function left in the spleen. Under certain pathologic conditions, however, such as myelodysplasia, the spleen can reacquire its hematopoietic function. Removal of the spleen does not usually result in anemia or leukopenia in an otherwise healthy person. Although the hematopoietic function is usually lost during fetal development, the spleen continues to function as a sophisticated filter because of the unique circulatory system and lymphoid organization, and it has both blood cell monitoring and management functions as well as important immune functions throughout life. The functions of the spleen are closely linked to splenic structure and its unique circulatory system. The arteries flow through the white pulp (lymphoid tissues), after which part of the blood flow goes directly through endothelial celllined capillaries into the venous system (closed theory). Most of the blood flow, however, enters the macrophage-lined reticular meshwork, and the blood flows slowly back to the venous circulation through the venous sinuses (open theory). The formed blood elements must pass through slits in the lining of the venous sinuses; if they cannot pass, they are trapped in the spleen and ingested by splenic phagocytes[1] ( Fig. 543 ). Experimental animal studies have demonstrated that an intact
1681
Figure 54-2 The anatomic relationships of the splenic vasculature. (From Economou SG, Economou TS: Atlas of Surgical Techniques. Philadelphia, WB Saunders, 1966, p 562.)
Figure 54-3 Structure of the sinusoidal spleen shows the open and closed blood flow routes. (From Bellanti JA: Immunology: Basic Processes. Philadelphia, WB Saunders, 1979.)
1682
splenic arterial system is necessary for optimal control of infection.[2] Removal of the spleen results in loss of both the immunologic and filtering functions. The most important function of the spleen is probably its mechanical filtration, which removes senescent erythrocytes and likely contributes to control of infection. The spleen is important in clearing circulating pathogens that reside within erythrocytes, for example, malarial parasites or bacteria such as Bartonella species. Mechanical filtration by the spleen may also be important for removal of unopsonized, noningested bacteria from the circulation. It may be particularly important for clearing microorganisms for which the host has no specific antibody. Splenic-filtering function is important for maintaining normal erythrocyte morphology and function. Normal red blood cells are biconcave and deform relatively easily to facilitate both passages through the microvasculature and optimal oxygen and carbon dioxide exchange. The spleen is an important site for the processing of immature erythrocytes and for repair or destruction of deformed or aged erythrocytes. As immature red blood cells pass through the spleen, they may undergo several types of repair, including removal of nuclei and excessive cell membrane from immature cells to convert them from a spherical nucleated to a biconcave anucleated mature morphology. Erythrocytes may also undergo repair by having surface abnormalities such as pits or spurs removed. In the asplenic condition, there are several characteristic alterations in the morphologic appearance of the peripheral red blood cells, with the presence of target cells (immature cells), Howell-Jolly bodies (nuclear remnant), Heinz bodies (denatured hemoglobin), Pappenheimer bodies (iron granules), stippling, and spur cells. Aged red blood cells (120 days) that have lost enzymatic activity and membrane plasticity are trapped and destroyed in the spleen ( Box 541 ).
Box 54-1. Biologic Substances Removed by the Spleen * In Normal Subjects Red blood cell membrane Red blood cell surface pits and craters Howell-Jolly bodies Heinz bodies Pappenheimer bodies Acanthocytes Senescent red blood cells Particulate antigen In Patients with Disease
Spherocytes (hereditary spherocytosis) Sickle cells, hemoglobin C cells Antibody-coated red blood cells Antibody-coated platelets Antibody-coated white blood cells
* Modified from Eichner ER: Splenic function: Normal, too much and too little. Am J Med 66:311, 1979.
The filtering function of the spleen is also an important factor in anemic conditions associated with abnormal red blood cell morphology. Abnormal erythrocytes that result from hereditary spherocytosis, sickle cell anemia, thalassemia, or pyruvate kinase deficiency are trapped by the splenic-filtering mechanism, resulting in worsening anemia, symptomatic splenomegaly, and occasionally splenic infarction. In autoimmune hemolytic anemia, immunoglobulin G (IgG) bound to the cell membrane targets the red blood cells for splenic destruction by splenic macrophages. A similar IgGdependent mechanism is involved in splenic platelet destruction in immune thrombocytopenic purpura (ITP). Another major function of the spleen is the maintenance of normal immune function and host defenses against certain types of infectious agents. It is well established that people lacking a spleen are at a significantly higher risk for overwhelming postsplenectomy infection (OPSI) with fulminant bacteremia, pneumonia, or meningitis, as compared with those with normal splenic function. Major pathogens in OPSI are organisms such as Streptococcus pneumoniae, in which polysaccharide capsules requiring both antibody and complement are important in host defense against these organisms. Asplenic subjects have defective activation of complement by the alternative pathway, leaving them more susceptible to infection. Asplenic patients have a normal response to reimmunization to an antigen first encountered before splenectomy but do not have an optimal response to new antigen exposure, especially if the antigen is administered intravenously. For organisms such as the encapsulated bacteria, much higher quantities of antibody are necessary for effective clearance. The spleen, with its specialized circulatory system and large supply of macrophages that are capable of ingestion of organisms not optimally opsonized with antibody, greatly enhances their clearance. Asplenic subjects have been found to have subnormal IgM levels, and their peripheral blood mononuclear cells exhibit a suppressed immunoglobulin response.[3] The spleen is a major site of production for the opsonins properdin and tuftsin, and removal of the spleen results in decreased serum levels of these factors. Properdin can initiate the alternative pathway of complement activation to produce destruction of bacteria as well as foreign and abnormal cells. Tuftsin is a tetrapeptide that enhances the phagocytic activity of both polymorphonuclear leukocytes and mononuclear phagocytes. The spleen is the major site of cleavage of tuftsin from the heavy chain of IgG, and circulating levels of tuftsin are suppressed in asplenic subjects.[4] Neutrophil function is decreased in asplenic patients, and the defect appears to result from the absence of a circulating mediator.[5]
SPLENECTOMY FOR BENIGN HEMATOLOGIC CONDITIONS

Immune Thrombocytopenic Purpura
Immune thrombocytopenic purpura is also referred to as idiopathic thrombocytopenic purpura. A low platelet count, a normal bone marrow, and the absence of other
1683
causes of thrombocytopenia characterize this disease. ITP is principally a disorder of increased platelet destruction mediated by autoantibodies to platelet membrane antigens that results in platelet phagocytosis by the reticuloendothelial system.[6] Bone marrow megakaryocytes are present in normal or sometimes increased numbers; however, there is relative marrow failure in that the marrow does not increase production sufficiently to compensate for platelet destruction in the spleen. In adults, ITP is more common in young women than men. Seventy-two percent of patients older than 10 years of age are women, and 70% of affected women are younger than 40 years of age. In children, ITP is manifested somewhat differently. It affects both sexes equally, and its onset is typically abrupt, with severe thrombocytopenia; however, spontaneous permanent remissions are the rule, occurring in about 80% of affected children. Children who develop chronic thrombocytopenia are usually girls older than 10 years of age and present with a longer history of purpura. Patients with ITP often present with a history of purpura, epistaxis, and gingival bleeding. Hematuria and gastrointestinal bleeding occur less commonly, and intracerebral hemorrhage is a rare but sometimes fatal event. The diagnosis of ITP requires exclusion of other causes of thrombocytopenia ( Box 542 ).[6] Apparent thrombocytopenia may be an artifactual report on a complete blood count because of in vitro platelet clumping or the presence of giant platelets. Mild thrombocytopenia may occur in 6% to 8% of otherwise normal pregnant women and in up to one fourth of women with preeclampsia. Several drugs are known to induce thrombocytopenia, including heparin, quinidine, quinine, and sulfonamides. Human immunodeficiency virus (HIV) infection and other viral infections may cause thrombocytopenia that may be mistaken for ITP. Other conditions that may be associated with thrombocytopenia include myelodysplasia, congenital thrombocytopenia, thrombotic thrombocytopenic purpura, chronic disseminated intravascular coagulation, autoimmune diseases such as systemic lupus erythematosus, and lymphoproliferative disorders such as chronic lymphocytic leukemia (CLL) and non-Hodgkins lymphoma (NHL). Management of patients with ITP varies according to the severity of the thrombocytopenia.[7] Patients with asymptomatic disease and platelet counts higher than 50,000/mm3 may simply be followed with no specific treatment. Platelet counts higher than 50,000/mm3 are seldom associated with spontaneous clinically important bleeding, even with invasive procedures. Patients with platelet counts between 30,000 and 50,000/mm3 who do not have symptoms may also be observed without treatment; however, careful follow-up is essential in these patients because they are at risk for more severe thrombocytopenia. The initial medical treatment is with glucocorticoids, usually prednisone (1 mg/kg body weight per day). About two thirds of patients treated in this manner experience an increase in their platelet count to more than 50,000/mm3 , usually within 1 week of treatment, although it sometimes requires as long as 3 weeks. As many as 26% of patients may have a complete response with glucocorticoid therapy. Patients with platelet counts
Box 54-2. Differential Diagnosis of Immune Thrombocytopenic Purpura Falsely Low Platelet Count In vitro platelet clumping caused by ethylenediamine tetra-acetic acid (EDTA)-dependent or cold-dependent agglutinins Giant platelets Common Causes of Thrombocytopenia Pregnancy (gestational thrombocytopenia, preeclampsia) Drug-induced thrombocytopenia (common drugs include heparin, quinidine, quinine, and sulfonamides) Viral infections, such as human immunodeficiency virus, rubella, infectious mononucleosis Hypersplenism due to chronic liver disease Other Causes of Thrombocytopenia That Have Been Mistaken for Immune Thrombocytopenic Purpura Myelodysplasia Congenital thrombocytopenias Thrombotic thrombocytopenic purpura and hemolytic-uremic syndrome Chronic disseminated intravascular coagulation
Thrombocytopenia Associated With Other Disorders Autoimmune diseases, such as systemic lupus erythematosus Lymphoproliferative disorders (chronic lymphocytic leukemia, non-Hodgkins lymphoma)
From George JN, El-Harake MA, Raskob GE: Chronic idiopathic thrombocytopenic purpura. N Engl J Med 331:12071211, 1994.
greater than 20,000/mm3 who are symptom free or who have only minor purpura do not require hospitalization. Treatment of ITP is indicated in patients with platelet counts of less than 20,000 to 30,000/mm3 or for those with platelet counts of less than 50,000/mm3 and significant mucus membrane bleeding or risk factors for bleeding, such as hypertension, peptic ulcer disease, or a vigorous lifestyle. Hospitalization is often required for patients with platelet counts of less than 20,000/mm3 who have significant mucus membrane bleeding and is necessary for all patients who experience severe life-threatening hemorrhage. Although platelet transfusions are necessary for controlling severe hemorrhage, they are seldom indicated in patients with ITP in the absence of severe hemorrhage. Intravenous immunoglobulin is important in the management of acute bleeding and for the preparation of patients for operation or delivery in the case of pregnancy. The usual dose is 1 g/kg per day for 2 days. This dose increases the platelet count in most patients within 3 days. It also increases the efficacy of transfused platelets. Administration of intravenous immunoglobulin is also appropriate in
1684
patients with platelet counts of less than 20,000/mm3 who are being prepared for splenectomy. Splenectomy was the first effective treatment described for ITP and was an established therapeutic modality long before glucocorticoid therapy was introduced in 1950.[7] About two thirds of patients achieve a complete response with normalization of platelet counts after splenectomy and require no further therapy. Splenectomy is indicated in patients with refractory severe symptomatic thrombocytopenia, in patients requiring toxic doses of steroids to achieve remission, and in patients with a relapse of thrombocytopenia after initial glucocorticoid treatment. Splenectomy is an appropriate consideration for patients who have had the diagnosis of ITP for 6 weeks and continue to have a platelet count of less than 10,000/mm3 whether or not bleeding symptoms are present. Splenectomy is also indicated for patients who have had the diagnosis of ITP for as long as 3 months and have experienced a transient or incomplete response to primary therapy and have a platelet count of less than 30,000/mm3 . Splenectomy should be considered for women in the second trimester of pregnancy who have failed glucocorticoid and intravenous immunoglobulin therapy and have platelet counts of less than 10,000/mm3 or who have platelet counts of less than 30,000/mm3 and bleeding problems. Splenectomy is probably not indicated in nonbleeding patients who have had a diagnosis of ITP for 6 months, have platelet counts of greater than 50,000/mm3 , and are not engaged in high-risk activities. The success rate of splenectomy in achieving complete and permanent response was reported to be 65% in a cumulative study of 1761 patients.[7] The site of platelet destruction on preoperative indium 111 (111 In)-labeled platelet scintigraphy was also reported to be predictive of the efficacy of splenectomy. The best long-term cure rates have been when predominantly splenic sequestration is present. In a cumulative study of 564 patients with ITP, a predominant splenic sequestration site was associated with an 87% to 93% rate of good response after splenectomy. In comparison, in patients with hepatic sequestration, the response rate was significantly lower (7% to 30%).[8] Most patients who respond to splenectomy with increased platelet counts do so within the first 10 days after their operation. Durable platelet responses have been correlated with platelet counts greater than 150,000/mm3 by the 3rd postoperative day or greater than 500,000/mm3 on the 10th postoperative day. The immediate response rates among series collected between 1980 and 1998 range from 71% to 95%, with relapse rates of 4% to 12% ( Tables 541 and 542 ).[9] For patients with chronic ITP who fail to achieve complete response after splenectomy, the options range from simple observation in patients with no bleeding symptoms and platelet counts of more than 30,000/mm3 to continued long-term prednisone therapy ( Table 543 ). [6] Single-agent treatment with azathioprine or cyclophosphamide may be considered, but response to these agents may require up to 4 months of treatment. Patients who fail to respond to splenectomy or have relapsing disease after an initial response should be investigated for the presence of accessory spleen. Accessory spleen may be found in as many as 10% of these patients. The presence of an accessory spleen may be suggested by the absence of asplenic red blood cell morphologic features and may also be identified by radionuclide imaging. Identification of an accessory spleen in a patient who remains severely thrombocytopenic and is otherwise fit for operation warrants surgical excision of the accessory spleen. TABLE 54-1 -- Hematologic Response after Laparoscopic Treatment of Immune Thrombocytopenia Purpura Study Cadiere et al, 1994[118] Emmermann et al, 1995[119] Poulin et al, 1995[120] Yee and Akpata, 1995[96] Brunt et al, 1996[121] Flowers et al, 1996[97] Gigot et al, 1996[122] Smith et al, 1996[123] Friedman et al, 1997[82] Park et al, 1997[108] Tsiotos and Schlinkert, 1997[124] Katkhouda et al, 1998[9] Total Patients (n) 17 27 22 25 26 43 18 10 63 22 18 103 394 ITP (n) 8 20 22 14 17 22 16 8 28 8 18 67 237 AS (n [%]) 2 (11.8) 2 (7.4) 6 (27.2) 2 (8.0) 3 (11.5) 4 (9.3) 7 (39.0) 2 (20.0) 11 (17.5) 2 (9.0) 1 (5.6) 17 (16.5) 59/394 (15.0) NA 19 (95.0) NA 11 (76.0) 13 (76.0) 18 (82.0) NA NA NA NA 17 (94.0) 56 (83.6) 134/158 (85.0) IR (n [%]) NA 0 at 14 mo NA NA NA 0 at 21 mo 2 (12.5) at 14 mo 0 NA NA 0 4 (6.0) at 38 mo 6/151 (4.0) RR (n [%])
AS, total number of patients with accessory spleens; IR, immediate response in immune thrombocytopenic purpura (ITP); RR, relapse rate in ITP; NA, data not available. From Katkhouda N, Hurwitz MB, Rivera RT, et al: Laparoscopic splenectomy: Outcome and efficacy in 103 consecutive patients. Ann Surg 228:568 578, 1998.
1685
TABLE 54-2 -- Results in 749 Collected Cases of Open Splenectomy Study DiFino et al, 1980
[125]
Patients (n) Morbidity (n [%]) 37 66 306 102 100 138 749 9 (24.3) 13 (14.1) 118 (24.0) 15 (14.7) 8 (8.0) NA 163/611 (26.7)
Mortality (n [%]) ITP (n) 0 1 (1.4) 18 (6.0) 0 0 NA 19/611 (3.1) 37 66 65 102 100 138 508
AS (n [%]) 2 (5.4) 20 (28.2) 58 (19.0) NA 18 (18.0) NA 98/611 (16.0)
IR (n [%]) 27 (73.0) 56 (84.8) 50 (77.0) 95 (93.1) 71 (71.0) 114 (83.0) 413/508 (81.3)
RR (n [%]) 9 (24.3) 6 (9.0) NA 11 (10.7) 4 (4.0) 23 (17.0) 53/443 (12.0)
Mintz et al, 1981[126] Musser et al, 1984[89] Jacobs et al, 1986[127] Akwari et al, 1987
[128]
Julia et al, 1990[129] Total
ITP, immune thrombocytopenic purpura; AS, total number of patients with accessory spleens; IR, immediate response in ITP; RR relapse rate in ITP; NA, data not available. From Katkhouda N, Hurwitz MB, Rivera RT, et al: Laparoscopic splenectomy: Outcome and efficacy in 103 consecutive patients. Ann Surg 228:568 578, 1998.
TABLE 54-3 -- Treatment Options for Patients with Chronic Immune Thrombocytopenic Purpura Unresponsive to Initial Glucocorticoid Therapy and Splenectomy Intervention Observation Prednisone Indication No bleeding symptoms; platelet count of 30,00050,000/mm3 Symptomatic thrombocytopenia, with bleeding symptoms; platelet count of 30,00050,000/mm3 Symptomatic thrombocytopenia, with bleeding symptoms; platelet count of 30,00050,000 mm3 Symptomatic thrombocytopenia, with bleeding symptoms; platelet count of 30,00050,000 mm3 Symptomatic thrombocytopenia in a patient who is a good candidate for surgery Outcome Platelet count may remain stable, but the risk for more severe thrombocytopenia with serious bleeding is unknown Goal is a safe platelet count with a minimal dose, such as 10 mg every other day; steroid toxicity is the limiting factor Response may require 4 months of treatment; complete recovery may occur in 1040% of patients Some regimens promising in small, preliminary studies; in others, few patients recover completely Complete recoveries reported in a few patients; symptomatic improvement in some others Some patients require frequent supportive therapy; others have minimal symptoms despite severe thrombocytopenia; spontaneous remissions may occur
Azathioprine or cyclophosphamide
Other regimens
Resection of accessory spleen or spleens
Observation (with glucocorticoid or Unresponsive to treatment intravenous immune globulin as needed to treat or prevent bleeding)
From George JN, El-Harake MA, Raskob GE: Chronic idiopathic thrombocytopenic purpura. N Engl J Med 331:12071211, 1994. In a summary of splenectomy series for hematologic disease using either a laparoscopic or open approach, Katkhouda and colleagues[9] found that among 394 patients treated with laparoscopic splenectomy, 237 (60%) had ITP and 59 (15%) had an accessory spleen (see Tables 541 and 542 ). From the combined series of patients treated for ITP, there was an immediate response rate to splenectomy of 85% and a relapse rate of 4%. These series spanned from 1994 through 1998. The same authors also summarized another six series of open splenectomy for hematologic disease spanning from 1980 to 1990 and totaling 749 patients, of which 508 patients underwent splenectomy for ITP. The incidence of accessory spleen in these combined series was 16%. For treatment of ITP, the immediate response rate was 81.3%, and the relapse rate was 12%. About 10% to 20% of otherwise symptom-free patients infected with HIV develop ITP.[10] Splenectomy may be performed safely in this cohort of patients and produces sustained increases in platelet levels in more than 80%. Splenectomy does not increase the risk of progression to acquired immunodeficiency syndrome (AIDS), and a recent cohort study suggested that the absence of a spleen during the asymptomatic phase of HIV infection may delay disease progression.[10]
Hereditary Spherocytosis
Hereditary spherocytosis is an autosomal dominant disease that results from a deficiency of spectrin, a red blood cell cytoskeletal protein.[11] [12] This protein defect
1686
causes a membrane abnormality in the red blood cells resulting in small, spherical, and rigid erythrocytes. These cells have increased osmotic fragility. These spherocytes are more susceptible to becoming trapped in the spleen and destroyed. The clinical features of this disease include anemia,
occasionally with jaundice, and splenomegaly. The diagnosis is made by identification of spherocytes on the peripheral blood smear, an increased reticulocyte count, increased osmotic fragility, and a negative Coombs test. Splenectomy decreases the rate of hemolysis and usually leads to resolution of the anemia. Splenectomy is usually performed in childhood shortly after diagnosis. Although splenectomy does not normalize the morphology of the red blood cells, it does reduce the trapping and premature destruction of them. It is generally recommended that the operation be delayed until after the 4th year of life to preserve immunologic function of the spleen in young children who are most at risk for OPSI. There is a high incidence of pigmented gallstones among patients with spherocytosis, similar to other hemolytic anemias, and ultrasound should be performed before splenectomy. If gallstones are present, it is appropriate to perform cholecystectomy at the time of the splenectomy. Other anemias associated with erythrocyte structural abnormalities include hereditary elliptocytosis, hereditary pyropoikilocytosis, hereditary xerocytosis, and hereditary hydrocytosis. All of these conditions result in abnormalities of the erythrocyte cellular membrane and increased red blood cell destruction. Splenectomy is indicated for the severe hemolytic anemia that commonly occurs in these conditions. An exception to this is the mild anemia that is usually of limited clinical significance in the condition of hereditary xerocytosis, in which splenectomy is not usually indicated.
Hemolytic Anemia due to Erythrocyte Enzyme Deficiency
Glucose-6-phosphate dehydrogenase deficiency and pyruvate kinase deficiency are the two predominant hereditary conditions associated with hemolytic anemia. These deficiencies result in abnormal glucose use and metabolism, leading to increased hemolysis. Pyruvate kinase deficiency is an autosomal recessive condition in which there is decreased red blood cell deformability resulting in increased hemolysis. The spleen is the site of erythrocyte entrapment and destruction in patients with deficiency of pyruvate kinase. These patients often have splenomegaly, and splenectomy has been shown to decrease their transfusion requirements. For the previously mentioned reasons of preserving immunologic function, splenectomy is usually delayed until after 4 years of age in patients with this condition. Glucose-6-phosphate dehydrogenase deficiency is an X-linked hereditary condition that is most frequently seen in people of African, Middle Eastern, or Mediterranean ancestry. Hemolytic anemia occurs in most patients after exposure to certain drugs or chemicals. Splenectomy is rarely indicated in patients with glucose-6-phosphate dehydrogenase deficiency.[13]
Hemoglobinopathies
Thalassemia and sickle cell disease are hereditary hemolytic anemias that result from abnormal hemoglobin molecules. This results in abnormal shape of the erythrocyte, which may be subject to splenic sequestration and destruction. Sickle cell anemia is the result of the homozygous inheritance of hemoglobin S. Hemoglobin S has a single amino acid substitution of a valine for a glutamic acid in the sixth position of the beta chain of hemoglobin A. Sickling of erythrocytes may also occur in patients who coinherit hemoglobin S and other hemoglobin variants, such as hemoglobin C or sickle cell thalassemia.[14] Sickle cell disease results in patients who are homozygous for hemoglobin S. About 0.5% of African Americans are homozygous for hemoglobin S, whereas about 8% are heterozygous for hemoglobin S. The heterozygotes have the sickle cell trait. Under conditions of reduced oxygen tension, hemoglobin S molecules crystallize within the cell, which results in an elongated, distorted cell with a crescent shape. These altered erythrocytes are rigid and incapable of deforming in microvasculature. This lack of deformation results in capillary occlusion and thrombosis, ultimately leading to microinfarction. This occurs with particular frequency in the spleen. The spleen is enlarged during the 1st decade of life in most patients with sickle cell disease and then with progressive infarction caused by repeated attacks of vaso-occlusion and infarction, resulting in autosplenectomy. The spleen in patients with sickle cell disease usually atrophies by adulthood, although splenomegaly may occasionally persist into adult life. Thalassemias constitute a group of hemoglobin disorders that also result in hemolytic anemia. Thalassemias are inherited as autosomal dominant traits and occur as a result of a defect in hemoglobin synthesis. This results in variable degrees of hemolytic anemia. Splenic infarction, splenomegaly, and hypersplenism may be predominant features of either sickle cell disease or thalassemia. Hypersplenism and acute splenic sequestration are life-threatening disorders in children with sickle cell anemia and thalassemia. In these conditions, there may be rapid splenic enlargement, resulting in severe pain and requiring multiple blood transfusions.[14] [15] In addition to acute splenic sequestration crisis, these patients may suffer from symptomatic massive splenomegaly causing discomfort and interfering with daily activities. Indications for splenectomy in patients with sickle cell disease include acute splenic sequestration crisis, hypersplenism, and splenic abscess.[14] Children with sickle cell anemia often exhibit weight loss and poor growth; these conditions may be improved after splenectomy as the result of decreased whole-body total protein turnover and decreased resting metabolic rate.[16] Sickle cell diseaseassociated hypersplenism is characterized by anemiarequiring transfusions as well as leukopenia and thrombocytopenia. Splenectomy reduces the need for transfusions in these patients. Splenic abscesses are not uncommon in patients with sickle cell anemia and are characterized by fever, abdominal
1687
pain, and a tender, enlarged spleen. Many patients with splenic abscess have leukocytosis. Thrombocytosis and Howell-Jolly bodies also occur in these patients, indicating functional asplenia. Common organisms involved in splenic abscess in patients with sickle cell anemia are Salmonella species, Enterobacter species, and other enteric organisms. In acute splenic sequestration crisis, the patients have severe anemia, splenomegaly, and an acute bone marrow response with erythrocytosis. They may exhibit a dramatic decrease in their hemoglobin levels along with abdominal pain and may undergo circulatory collapse. These patients should be stabilized with hydration and transfusion and may require urgent splenectomy after stabilization.
SPLENECTOMY FOR MALIGNANCY

Lymphomas
Hodgkins Disease
Hodgkins disease is a malignant lymphoma that typically affects young adults in their 20s and 30s. Most patients have asymptomatic lymphadenopathy at the time of diagnosis, and most present with cervical node enlargement. A few patients, usually with more advanced disease, may present with constitutional symptoms such as night sweats, weight loss, and pruritus. Hodgkins disease is histologically classified as either lymphocytepredominant, nodular-sclerosing, mixed-cellularity, or lymphocyte-depleted Hodgkins disease. The disease is pathologically staged according to the Ann Arbor classification. Stage I represents disease in a single lymphatic site, whereas stage II indicates the presence of disease in two or more lymphatic sites on the same side of the diaphragm. Stage III indicates the presence of lymphatic disease (includes splenic involvement) on both sides of the diaphragm. Stage IV disease is disseminated into extralymphatic sites such as liver, lung, or bone marrow. Subscript E indicates single or contiguous extralymphatic involvement in stages I to III, and subscript S represents splenic involvement. Patients with constitutional symptoms are classified as B (presence) or A (absence), for example, stage IIA or IIB. Historically, staging laparotomy including splenectomy provided essential pathologic staging information that was necessary to select appropriate therapy for Hodgkins disease. The purpose of staging laparotomy is to pathologically stage the presence and extent of disease below the diaphragm. In one of the larger reported series from 1985, 38.9% of 825 splenectomy specimens tested positive for involvement with Hodgkins disease.[17] In half of the patients with splenic involvement, the spleen was the only site of intra-abdominal disease. The spleen was also involved in all of the 6.2% of patients with liver involvement. In this important series by Taylor and colleagues,[17] the clinical stage was changed in 43% of cases by laparotomy. Advances in imaging techniques, with widespread availability of dynamic helical computed tomography (CT) scan and lymphangiography and increasing availability of fluorodeoxyglucose positron-emission tomography imaging, have improved nonoperative staging of Hodgkins disease. The improved nonoperative staging, along with the use of less toxic systemic chemotherapeutics for earlier stages of Hodgkins disease, has led to a dramatic decrease in the numbers of patients requiring staging laparotomy. Patients at high risk for relapse, especially those with B symptoms and those with evidence of intra-abdominal involvement on one or more of the diagnostic imaging studies, require systemic chemotherapy and should not undergo a staging laparotomy. Staging laparotomy and splenectomy are appropriate for selected patients with an early clinical stage of disease (stage IA or IIA) in whom pathologic staging of the abdomen will significantly influence the therapeutic management. Early-stage Hodgkins disease is often cured with radiation therapy alone. When indicated, the staging laparotomy for Hodgkins disease should include a thorough abdominal exploration, splenectomy with splenic hilar lymphadenectomy, bilateral wedge and core-needle liver biopsies, retroperitoneal lymphadenectomy, iliac crest bone marrow biopsy, and in premenopausal women, oophoropexy. The perioperative mortality rate for staging laparotomy should be less than 1%, and the risk for major complications has been less than 10%.[17]
Non-Hodgkins Lymphomas
Splenomegaly or hypersplenism is a common occurrence during the course of NHL. Splenectomy is indicated for patients with NHL for treatment of massive splenomegaly when the bulk of the spleen contributes to abdominal pain, fullness, and early satiety. Splenectomy may also be effective in the treatment of patients who develop hypersplenism with associated anemia, thrombocytopenia, and neutropenia.[18] Splenectomy occasionally plays an important role in the diagnosis and staging of patients who present with isolated splenic disease. The most common primary splenic neoplasm is NHL. The spleen is involved in 50% to 80% of NHL patients, but fewer than 1% of patients present with splenomegaly without peripheral lymphadenopathy.[19] Disease that appears clinically confined to the spleen has been called malignant lymphoma with prominent splenic involvement. Most affected patients have low-grade NHL. There is frequent involvement of the splenic hilar lymph nodes, extrahilar nodes, bone marrow, and liver in these patients. About 75% exhibit clinical evidence of hypersplenism. In a series of 59 such patients reported by Morel and associates,[19] 40 underwent splenectomy, and 19 did not. Eighty-two percent of the cytopenic patients who underwent splenectomy had correction of their hematologic abnormalities postoperatively. For those patients with low-grade NHL who had spleenpredominant features, survival was significantly improved after splenectomy (median, 108 months) as compared with patients receiving similar treatment without splenectomy (median, 24 months).
1688
Leukemia
Hairy Cell Leukemia
Hairy cell leukemia is a rare disease that represents about 2% of adult leukemias. Splenomegaly, pancytopenia, and neoplastic mononuclear cells in the peripheral blood and bone marrow characterize the disease.[20] [21] The hairy cells are usually B lymphocytes that have cell membrane ruffling, which appears as cytoplasmic projections under the light microscope. The patients are usually elderly men with palpable splenomegaly. About 10% of cases have an indolent course requiring no specific therapy, but most require therapy for cytopenias such as symptomatic anemia, infectious complications from neutropenia, or hemorrhage from thrombocytopenia. The pancytopenia results from the combined effects of hypersplenism and bone marrow replacement by leukemic cells. Therapy may also be required for massive splenomegaly. Patients with hairy cell leukemia have a twofold to threefold increased risk for diagnosis of a second primary malignancy at a median interval of 40 months after the diagnosis of hairy cell leukemia.[20] Most of the second malignancies are solid tumors, and the types of tumors include prostate cancer, skin cancers, lung cancer, and gastrointestinal adenocarcinomas. Splenectomy and interferon-2 have been the standard treatment of hairy cell leukemia until recently; this approach is being replaced with systemic administration of purine analogues, such as 2-chlorodeoxyadenosine and deoxycoformycin, as initial treatment.[22] [23] Splenectomy is still indicated for some patients with massive enlargement of the spleen or with evidence of hypersplenism that is refractory to medical therapy. Splenectomy provides a highly effective and sustained palliation of these problems, and most patients show definite hematologic improvement after the procedure. About 40% of patients experience normalization of their blood counts after splenectomy. The responses to splenectomy usually last for 10 years or longer, and about half of patients require no further therapy. Patients with diffusely involved bone marrow who do not have significant splenomegaly are far less likely to achieve a significant benefit from splenectomy. The current 4-year survival rate after diagnosis of hairy cell leukemia is about 80%, as compared with 60% for patients diagnosed in the 1970s.[24]
Chronic Lymphocytic Leukemia
CLL is a B-cell leukemia that is characterized by the progressive accumulation of relatively mature, but functionally incompetent, lymphocytes. CLL occurs more frequently in men and usually occurs after 50 years of age. Staging of CLL is according to the Rai staging system, which correlates well with prognosis.[25] Stage 0 includes bone marrow and blood lymphocytosis only; stage I includes lymphocytosis and enlarged lymph nodes; stage II includes lymphocytosis and enlarged spleen, liver, or both; stage III includes lymphocytosis and anemia; and stage IV includes lymphocytosis with thrombocytopenia. Chlorambucil has long been the mainstay of medical therapy and was useful for the palliation of symptoms; however, there is increasing interest in using purine analogues, such as fludarabine, as first-line therapy, with some studies showing improved rates of remission.[24] Bone marrow transplantation has also become increasingly used in the treatment of CLL, and both allotransplantation and autotransplantation approaches are being investigated. The role of splenectomy in the treatment of CLL continues to be for palliation of symptomatic splenomegaly and for treatment of cytopenia related to hypersplenism. Relief of bulk symptoms from splenomegaly is virtually always successful, whereas the hematologic response rates for correction of anemia and thrombocytopenia are between 60% and 70%. [26] [27] Cusack and colleagues[26] reported a review of 77 consecutive patients with CLL (76% Rai stage III or IV) who underwent splenectomy at the University of Texas M. D. Anderson Cancer Center and who were compared with an age- and gender-matched cohort of CLL patients treated with fludarabine and no splenectomy. In this retrospective review, the patients with profound anemia and thrombocytopenia who underwent splenectomy had significantly better survival rates than the nonsplenectomy group.
Chronic Myelogenous Leukemia
Chronic myelogenous leukemia (CML) is a myeloproliferative disorder that results from neoplastic transformation of myeloid elements. CML was the first leukemic subset for which a chromosomal marker (the Philadelphia chromosome) was discovered.[28] The Philadelphia chromosome is caused by a fusion of fragments of chromosomes 22 and 9 and results in the expression of an abnormal chimeric oncogenic protein called p210bcr-abl . The disease is characterized by a progressive replacement of the normal diploid elements of the bone marrow with mature-appearing neoplastic myeloid cells. CML may occur from childhood to old age. CML usually presents with an indolent or chronic phase that is asymptomatic. Progression to the accelerated phase is marked by the onset of symptoms such as fever, night sweats, and progressive splenomegaly; however, this phase may also be asymptomatic and detectable only from changes in the peripheral blood or bone marrow. The accelerated phase may give rise to the blastic phase, which is characterized by the previously listed symptoms as well as anemia, infectious complications, and bleeding. Splenomegaly with splenic sequestration of blood elements often contributes to these symptoms. Treatment of CML is primarily medical and may include hydroxyurea, interferon-alfa, and high-dose chemotherapy with bone marrow transplantation. Symptomatic splenomegaly and hypersplenism in patients with CML may be effectively palliated by splenectomy.[29] Otherwise, the role of splenectomy in the treatment of CML has been controversial. Randomized studies of patients with CML have demonstrated no survival benefit when splenectomy is performed during the early chronic phase.[30] [31] Splenectomy has also not resulted in a survival benefit when performed before allogeneic bone marrow transplantation.[32] Thus, splenectomy before allogeneic bone marrow transplantation
1689
is recommended only for patients with significant splenomegaly.

Nonhematologic Tumors of the Spleen
The spleen is a site of metastatic tumor in up to 7% of autopsies of cancer patients. The primary solid tumors that most frequently metastasize to the spleen are carcinomas of the breast, lung, and melanoma; however, virtually any primary malignancy may metastasize to the spleen.[33] Metastases to the spleen are often asymptomatic but may be associated with symptomatic splenomegaly or even spontaneous splenic rupture. Splenectomy may provide effective palliation in carefully selected symptomatic patients with splenic metastasis. Vascular neoplasms are the most common primary splenic tumors that include both benign and malignant variants. Hemangiomas are usually incidental findings identified in spleens removed for other reasons. Angiosarcomas (or hemangiosarcomas) of the spleen have been associated with environmental exposure to thorium dioxide or monomeric vinyl chloride, but they most often occur spontaneously. Patients with these tumors may present with splenomegaly, hemolytic anemia, ascites, and pleural effusions or with spontaneous splenic rupture. These are highly aggressive tumors that have a poor prognosis. Lymphangiomas are usually benign endothelium-lined cysts that may become symptomatic by causing splenomegaly. Lymphangiosarcoma within a cystic lymphangioma has been reported.[34] Splenectomy is appropriate for diagnosis, treatment, or palliation of the conditions cited earlier.
SPLENECTOMY FOR MISCELLANEOUS BENIGN CONDITIONS

Splenic Cysts
Cystic lesions of the spleen have been recognized with increasing frequency since the advent of CT scanning and ultrasound imaging. Splenic cysts are classified as true cysts, which may be either nonparasitic or parasitic, and pseudocysts. Cystic-appearing tumors of the spleen include cystic lymphangiomas and cavernous hemangiomas, as discussed earlier.[35] [36] Primary true cysts of the spleen account for about 10% of all nonparasitic cysts of the spleen. Most nonparasitic cysts are pseudocysts and are secondary to trauma. The diagnosis of true splenic cysts is commonly made in the 2nd and 3rd decades of life. True cysts are characterized by a squamous epithelial lining, and many are considered congenital. These epithelial cells are often positive for CA 199 and carcinoembryonic antigen by immunohistochemistry, and patients with epidermoid cysts of the spleen may have elevated serum levels of one or both of these tumor-associated antigens.[35] Despite the presence of these tumor markers, these cysts are benign and apparently do not have malignant potential greater than any other native tissue. Often, true splenic cysts are asymptomatic and found incidentally. When symptomatic, patients may complain of vague upper abdominal fullness and discomfort, early satiety, pleuritic chest pain, shortness of breath, left back or shoulder pain, or renal symptoms from compression of the left kidney. A palpable abdominal mass may be present. The presence of symptoms is often related to the size of the cysts, and cysts smaller than 8 cm are rarely symptomatic.[36] Rarely, these cysts may present with acute symptoms related to rupture, hemorrhage, or infection. The diagnosis of splenic cysts is best made with CT imaging. Operative intervention is indicated for symptomatic cysts and for large cysts. Either total or partial splenectomy may provide successful treatment. The clear advantage of partial splenectomy is the preservation of splenic function. Preservation of at least 25% of the spleen appears sufficient to protect against pneumococcal pneumonia.[37] Most recent reports describe successful experience with partial splenectomy, cyst wall resection, or partial decapsulation, which may be accomplished with either an open or laparoscopic approach.[35] [36] [38] Most true splenic cysts are parasitic cysts in areas of endemic hydatid disease (Echinococcus species). Radiographic imaging may reveal cyst wall calcifications or daughter cysts. Although hydatid cysts are uncommon in North America, this diagnosis should always be excluded before the performance of invasive diagnostic or therapeutic procedures that may risk spillage of cyst contents. Serologic tests for Echinococcus species are often helpful in verifying the presence of parasites. As with hydatid cysts of the liver, spillage of cyst contents may precipitate an anaphylactic shock and risks intraperitoneal dissemination of infective scolices. Splenectomy is the treatment of choice, and great care should be taken to avoid rupture of the cysts intraoperatively. The cysts may be sterilized by injection of a 3% sodium chloride solution, alcohol, or 0.5% silver nitrate, as has been recommended for hydatid cysts of the liver.[39] Pseudocysts account for 70% to 80% of all nonparasitic cysts of the spleen. A history of prior trauma can usually be elicited. Splenic pseudocysts are not epithelial lined. Radiographic imaging may demonstrate focal calcifications in up to half of cases. Most splenic pseudocysts are unilocular, and the cysts are smooth and thick walled. Small asymptomatic splenic pseudocysts (<4 cm) do not require treatment and may undergo involution over time. When the pseudocysts are symptomatic, patients often present with left upper quadrant and referred left shoulder pain. Symptomatic pseudocysts should be treated surgically. If the spleen can be safely and completely mobilized and partial splenectomy accomplished to include the cystic portion of the spleen, this technique offers effective therapy that preserves splenic function.[40] Presented with less favorable conditions, the surgeon should not hesitate to perform total splenectomy. Successful percutaneous drainage has also been reported for splenic pseudocysts, although the success rate with this approach as compared with surgical intervention has not been determined. The 90% success rate of image-guided percutaneous drainage of unilocular splenic abscesses suggests that this may be a reasonable initial approach for the management of symptomatic splenic pseudocysts.[40]
1690
Splenic Abscess
Splenic abscess is an uncommon and potentially fatal illness. The incidence in autopsy series approximates 0.7%.[41] The mortality for splenic abscess ranges from about 80% for multiple abscesses in immunocompromised patients to about 15% to 20% in previously healthy patients with solitary unilocular lesions. Predisposing illnesses include malignancies, polycythemia vera, endocarditis, previous trauma, hemoglobinopathy (such as sickle cell disease), urinary tract infection, intravenous drug abuse, and AIDS.[42] [43] About 70% of splenic abscesses result from hematogenous spread of the infecting organism from another location, as occurs with endocarditis, osteomyelitis, and intravenous drug abuse. Splenic abscess may also occur as the result of infection of a contiguous structure, such as the colon, kidney, or pancreas. Gram-positive cocci, such as Staphylococcus, Streptococcus, or Enterococcus species, and gram-negative enteric organisms are often the infectious agents. Splenic abscesses may also be caused by other fastidious organisms, including Mycobacterium tuberculosis, Mycobacterium avium, and Actinomyces species. Immunosuppressed patients may develop multiple fungal abscesses, typically from Candida species infection. The clinical presentation of splenic abscess is often nonspecific and insidious, including abdominal pain, fever, peritonitis, and pleuritic chest pain. The abdominal pain is localized in the left upper quadrant less than half the time and is more often vague abdominal pain. Splenomegaly is present in a minority of patients. The diagnosis is made most accurately by CT; however, it may also be made with ultrasonography. Two thirds of splenic abscesses in adults are solitary, and the remaining one third are multiple. These ratios are reversed in children.[42] [44] [45] The initial approach to treatment of splenic abscess depends on whether it is unilocular or multilocular. Unilocular abscesses are amenable to CT-guided drainage, and this approach, along with systemic antibiotic administration, has a success rate that is in excess of 75%[46] and that may be as high as 90% when only patients with unilocular collections are considered.[47] Failure of a prompt clinical response to percutaneous drainage should lead to splenectomy without delay. Multilocular abscesses should usually be treated by splenectomy, with drainage of the left upper quadrant and antibiotic administration.
Wandering Spleen
Wandering spleen is a rare finding, accounting for only a fraction of a percentage of all splenectomies. The spleen normally has peritoneal attachments (called suspensory ligaments) that fix the spleen in its usual anatomic position. Failure to form these attachments has been postulated to result from failure of the dorsal mesogastrium to fuse to the posterior abdominal wall during embryonic development. The result is an unusually long splenic pedicle. It has also been postulated that an acquired defect in splenic attachment may occur in multiparous women secondary to hormonal changes during pregnancy and associated abdominal laxity. Wandering spleen is most often diagnosed in children or women between the ages of 20 and 40 years.
Most patients with wandering spleen are asymptomatic. Symptomatic patients often present with recurrent episodes of abdominal pain. This is likely related to tension on the vascular pedicle or intermittent torsion. Torsion of the splenic vessels may lead to venous congestion and splenomegaly. Severe and persistent pain is suggestive of splenic torsion and ischemia. On examination, a mobile abdominal mass may be present along with abdominal tenderness. The diagnosis may be most readily confirmed with a CT scan of the abdomen. The typical finding is the absence of a spleen in its normal position and the presence of a spleen in an ectopic location. Intravenous contrast injection during the CT scan provides valuable information. Lack of contrast enhancement of the spleen suggests splenic torsion, as does a whorled appearance to the splenic pedicle. Lack of splenic perfusion on the CT scan may be helpful in guiding the operative decision for splenectomy versus splenopexy.[48] [49]
SPLENIC TRAUMA The most significant contemporary issue in managing splenic trauma is the role of nonoperative management. Advances in diagnostic techniques that have occurred since the early 1980s have led to alternative approaches to managing these injuries. In recent years, abdominal ultrasound and CT have permitted observation of the location and relative quantification of the amount of intra-abdominal hemorrhage (ultrasound and CT), and CT has provided relatively good definition of the degree of anatomic splenic disruption. The main thrust of this section of the chapter is to trace those diagnostic developments in conjunction with the issues of clinical presentation and splenic anatomy to develop a logical platform for decision making.
General Considerations
Injury to the spleen is the most common indication for laparotomy after blunt mechanisms of injury. Motor vehicular crashes continue to be the major source of injury in industrialized nations. Other common mechanisms include motorcycle crashes, falls, pedestrian and vehicular incidents, bicycle accidents, and sports. Significant abdominal pain produced in the setting of nonvehicular blunt trauma is associated with a high incidence of significant intra-abdominal injury, with the spleen being the most commonly injured organ. Directly surrounding the spleen are the left hemidiaphragm, splenic flexure of the colon, kidney, distal pancreas, and stomach (see Fig. 541 ). The relatively avascular ligaments described earlier in this chapter (splenophrenic, splenorenal, splenocolic, and gastrosplenic) secure the spleen in this left upper quadrant niche somewhat protected by the lower costal margins (see Fig. 541 ). Splenic injuries are produced by rapid deceleration, compression,
1691
energy transmission through the posterolateral chest wall over the spleen, or puncture from an adjacent rib fracture. Rapid deceleration results in the spleen continuing in a forward motion while being tethered at the point of attachment. Injuries produced by deceleration forces result in capsular avulsion along the various ligamentous attachments and linear or stellate fractures of varying depths. Because of its solid structural characteristics and density, energy transfer to the spleen is relatively efficient. Injuries caused by assaults or falls are usually a result of direct blows over the lower chest wall, with transmission of energy resulting in splenic lacerations and fractures. The blood supply to the spleen is considerable. The spleen receives about 5% of the cardiac output. The blood supply is through the splenic artery and the short gastric vessels (see Fig. 542 ). The splenic arteries divide into several segmental vessels supplying the poles and midportion, and these arteries divide into second- and third-order vessels that course transversely within the spleen. Because of this extensive arterial supply, even superficial lacerations and capsular avulsions often yield substantial hemorrhage.
Diagnosis
The history and physical examination continue to be the basis from which splenic injury should be diagnosed. Details concerning the mechanism of injury delineated in the previous section should be sought. On physical examination, evidence of peritoneal irritation (tenderness, guarding, rebound) is sometimes apparent. Recently extravasated blood, however, is a fairly benign peritoneal irritant, and large amounts of blood may be contained in the free peritoneal cavity with minimal physical findings. In the era before diagnostic peritoneal lavage (DPL), physical examination was found to be accurate only 42% to 87% of the time.[50] [51] [52] [53] Perhaps more helpful than examination directed at the abdominal cavity is that focused in the left upper quadrant. Percussion tenderness or evidence of bruising and soft tissue contusion in the posterior left lower costal margin is usually present when direct blows have produced splenic injury. Complaints of left upper quadrant pain or of pain referred to the left shoulder (Kehrs sign) are highly correlated with injury. At least one fourth of patients with left lower rib fractures have associated injury to the spleen. Significant injury producing hemorrhage is indicated by the hemodynamic status of the patient. Hypotension or tachycardia should alert the clinician to the potential for splenic injury. At the initial trauma assessment, apparent injuries that may yield enough blood loss individually or in aggregate to produce physiologic changes in hemodynamics should be noted. If blood loss from long bone or pelvic fractures or from external losses from lacerations cannot be attributed, an intra-abdominal source must be assumed, and the spleen is the most common source. Indeed, since the classic study by West and associates[54] of preventable deaths that directly contributed to the development of trauma systems in the United States, mortality from missed or delayed recognition of splenic hemorrhage has remained near the top of the list of causes of preventable death. The diagnosis becomes more difficult in the presence of multiple injuries, which are both distracting to examination and confounding to interpretation of potential sources and volumes of blood loss. Furthermore, associated neurologic injury and substance abuse often add to the diagnostic difficulty. Closed-head injury is associated in 30% to 40% of cases and compromises or eliminates the reliability of the physical examination. Spinal cord injury also eliminates reliance on abdominal examination. Substance abuse has been documented in about 40% of patients involved in motor vehicle crashes. Because of the unreliability of physical examination accounted for by these several concerns, more objective means of diagnosis have evolved.
Diagnostic Peritoneal Lavage
DPL was introduced by Root and colleagues in 1965.[55] That modality remained the standard diagnostic procedure for blunt abdominal trauma evaluations for the subsequent 20 years. Initially, results were interpreted from a grossly positive examination or from quantification of red and white blood cells in the large effluent. A positive DPL consisted of either 10 mL of gross blood aspirated with catheter insertion or a microscopically positive examination. For microscopic examination in adults, 1 L of crystalloid solution is instilled through a periumbilical catheter inserted by either open or closed technique. Assuming complete instillation of the liter, positive examinations consisted of a red blood cell count higher than 100,000/mm3 or a white blood cell count higher than 500/mm3 in the completely mixed effluent. The dilution factor that produces a red blood cell count of more than 100,000/mm3 accounts for about 30 to 40 mL of blood in the peritoneal cavity. A microscopically positive examination by white blood cell count criteria is indicative of peritoneal inflammation generally produced by hollow viscus injury. After the introduction of DPL, multiple investigators demonstrated sensitivities approaching 99% and specificities in the range of 95% to 98%. Subsequent refinements included enzyme analysis of the effluent to enhance diagnostic accuracy of hollow viscus and pancreatic injury. An extraordinary amount of clinical research has been done on DPL-related issues. For several years, one could expect at least one presentation on DPL at any surgical meeting dealing with clinical trauma topics, and often multiple papers were presented. Reliance on DPL as a standard screening procedure, however, came into scrutiny from two fronts. As CT began to be applied for trauma diagnostics, it was observed that small splenic injuries had occurred and that the patient remained hemodynamically stable. It was further noted that enough blood was
present to have caused a positive lavage. Coincident with that development was the observation by surgeons that many of the positive DPL procedures led to laparotomies in which there was indeed a splenic injury, but often of a relatively trivial nature without active bleeding. Based on these parallel observations,
1692
the term nontherapeutic laparotomy began surfacing in the splenic injury literature, and it began to be appreciated that DPL was perhaps too sensitive.
Computed Tomography
Application of this technology for diagnosing abdominal injuries began in the early 1980s. The group at San Francisco General Hospital provided the early leadership in this area.[56] Initial observation consisted of examinations done for other injuries at some interval after initial trauma evaluation (e.g., chest or pelvic CT). Injuries to the spleen and liver were noted incidentally and, although initially unrecognized, produced no sequelae. This led to questioning of both the routine reliance on DPL for screening and the necessity of operating on all splenic injuries. As opposed to DPL, CT permitted not only identification of intraperitoneal blood but also definition of individual organ injuries. This actually revolutionized the management of splenic injury. Heretofore, it was assumed that once injured, the spleen invariably continued bleeding. It was generally taught in surgical training programs that if the spleen was incidentally injured from retraction during elective celiotomy, splenectomy was required because of the high risk for continued or recurrent hemorrhage. The early incidental CT observations of damaged spleens in stable patients ushered in the era of nonoperative management. The anatomic definition of injury provided objective criteria for classification of degrees of splenic injury. The American Association for the Surgery of Trauma developed a splenic injury grading scale through a consensus methodology ( Table 544 ). [57] This scale has been useful for comparison of data among institutions. It has also provided a structure for a logical approach to management decisions. Advances in CT technology have continued to increase the value of evaluating intra-abdominal and retroperitoneal injury. The current generation of helical and spiral scanning technology is both rapid and high resolution. Previous technology required 15 to 20 minutes for complete examination, but current technology requires 1 to 2 minutes. The resolution permits more precise delineation of organ fracture and intraparenchymal vascular disruption. The newest generation of CT technology is multislice scanning. That approach allows for scanning of the torso in less than 10 seconds and, owing to a much higher number of accumulated images compared to standard helical technology, an even higher resolution is obtained. This technology is highly promising to replace diagnostic angiography and may have profound importance in better defining solid-organ injury and further minimizing failures of nonoperative management.
Ultrasonography
During the 1990s, ultrasonography was introduced and firmly established as an important diagnostic tool for evaluating blunt abdominal trauma. It was first used in Europe in the early 1980s.[58] [59] [60] A few years later, it was adopted by physicians in the United States. Its advantages include noninvasiveness, rapidity, and low cost. Ultrasound provides similar but somewhat more information than DPL. The presence of free intraperitoneal fluid can be identified and semiquantitated. Acoustic windows are noted around solid interfaces. Those solid interfaces for trauma evaluation include the spleen, kidneys, liver, heart, and distended urinary bladder. The acronym FAST (focused abdominal sonography for trauma) has been applied to this quick survey, which takes about 3 minutes to complete in experienced hands. Significant bowel distention, obesity, and subcutaneous emphysema compromise the examination. Initial concerns were with the sensitivity and reproducibility of FAST. There was also a question of whether this technology could be grasped easily enough to permit widespread application by surgeons or whether the difficulty was such that only radiologists who had more extensive training could be relied on for accuracy. Tso and colleagues[61] reported in 1992 on 163 stable patients evaluated by ultrasound before either DPL or CT and found a 91% sensitivity, with all cases of clinically significant hemoperitoneum being identified. A subsequent study by Bode and associates[62] involved 353 blunt abdominal trauma patients evaluated with ultrasound by a radiologist. These investigators reported a 93% specificity rate, 99% accuracy rate, and no nontherapeutic laparotomies. Rothlin and coworkers[63] reported a study of 290 patients in which ultrasound was performed by a surgeon. They found a 90% sensitivity rate and a 99% specificity rate TABLE 54-4 -- American Association for the Surgery of Trauma Splenic Injury Scale (1994 Revision) Grade I II III Hematoma Laceration Hematoma Laceration Hematoma Type Subcapsular, < 10% surface area Capsular tear, < 1-cm parenchymal depth Subcapsular, 1050% surface area; intraparenchymal, < 5 cm in diameter 13 cm parenchymal depth, which does not involve a trabecular vessel Subcapsular, > 50% surface area or expanding; ruptured subcapsular or parenchymal hematoma Intraparenchymal hematoma > 5 cm or expanding Laceration IV Laceration Vascular >3 cm parenchymal depth or involving trabecular vessels Laceration involving segmental or hilar vessels producing major devascularization (>25% of spleen) Hilar vascular injury that devascularizes spleen Injury Description
Adapted from Moore EE, Cogbill TH, Jurkovich GJ, et al: Organ injury scaling: Spleen and liver (1994 revision). J Trauma 38:323, 1995.
1693
for intra-abdominal injury. They also noted the ease of repeatability for follow-up of these patients. The reported data would suggest that with a structured training format, FAST can be expeditiously taught to practicing surgeons. The American College of Surgeons is vigorously involved in evaluating new technologies and is developing structured courses for teaching surgeons the applications of ultrasound for evaluation of breast disease, intraabdominal pathology, and trauma. Ultrasound has emerged as a replacement for DPL. It appears to be as sensitive in detecting free intraperitoneal blood and is less invasive and quicker. The most important application is in evaluating the hemodynamically unstable patient with multiple injuries. A positive ultrasound would generally mandate expeditious exploratory laparotomy. The place for ultrasound use in the stable patient has been less clear. As nonoperative management has become so prominent, CT has become indispensable in defining the location and degree of organ injury. Ultrasound is not capable of accurately defining those anatomic characteristics. Technology is advancing rapidly, however, and it is likely that future generations of ultrasonography equipment will significantly improve resolution and anatomic definition. Furthermore, ultrasound may have a more positive impact on cost than is immediately
apparent. As ultrasound has become more widely adopted and confidence in the modality has developed, it has become apparent that ultrasound can be applied for screening as an alternative to the more costly CT evaluation. In a prospective study, Branney and associates[64] used ultrasound in the emergency department for screening stable patients. Their findings included a reduction in the number of CT examinations in patients with significant injury because those with negative examinations required no further studies. Sixty-five percent of the patients had no further studies, the number of admissions for observation decreased significantly, and no significant injuries were missed.
Issues Concerning Operation
Indications for Exploration
The clearest indication for urgent operation is hemodynamic instability. Unfortunately, this is not a binomial or discrete factor. The definition of stability is clearly associated with some degree of arbitrary assignment. This problem is underscored in the face of multiorgan system injury in which blood loss accumulates from external losses from lacerations, internal losses in fractures and soft tissues, and thoracic and abdominal cavities. Optimal decisions become apparent in retrospect. Nonetheless, when in doubt, abdominal exploration should be performed. The risks associated with nontherapeutic laparotomy are outweighed by the risks associated with shock secondary to prolonged intraperitoneal hemorrhage and the associated consequences of immunocompromise, multiorgan system dysfunction, and death. Because there can be no standard criteria for hemodynamic instability, a general guideline is to operate for a systolic blood pressure below 90 mm Hg or a pulse of more than 120 beats/min if there is not immediate response to 1 to 2 L of crystalloid resuscitation and when physical examination, ultrasound, or DPL indicates intra-abdominal blood loss. Indications for operation based on CT findings are delineated in a subsequent section of this chapter on nonoperative management.
Technical Issues
A midline incision is usually preferred for trauma exploration. This approach is expeditious and provides access to all areas of the abdominal cavity, including the retroperitoneum. A left subcostal approach may be preferred when laparotomy is directed by CT findings. The small bowel and lesser sac are easily evaluated from this incision. Extension to the right side provides outstanding exposure of the liver and access to all of the abdomen with the exception of the deepest portion of the pelvis. Both incisions are adequate, but the performance of a midline incision is quicker. After rapid evacuation of free blood and clots to assess other sources of injury, including the liver and mesentery, the spleen should be mobilized into the wound. Splenic mobilization should be accomplished by the fundamental operative principle of traction and countertraction. In the case of splenic mobilization, traction and countertraction are based on the spleen and its suspensory ligaments. The operating surgeon should apply dorsal and medial traction on the spleen with the hand splayed widely over the splenic surface to stretch and define clearly the splenorenal and splenophrenic ligaments. It appears that there is a natural tendency to place ventral traction on the spleen to lift it from the left upper quadrant, a maneuver that results in decapsulation over the posterior splenic surface and along the splenocolic ligament, producing iatrogenic trauma and increasing hemorrhage. After exposure of the splenorenal and splenophrenic ligaments, which is facilitated by the first assistant providing countertraction with tissue forceps, the ligaments can be divided under direct vision. The incision begins at the phrenocolic ligament, continuing through the ligaments to the stomach in the vicinity of the highest short gastric vessels. Division should occur 1 to 2 cm from the spleen to avoid injury to both spleen and diaphragmatic muscle. Continued tension on the tissues allows gradual mobilization anterior to the spleen as deeper layers of filmy connective tissue planes are placed under tension and easily visualized and divided. The dissection should progress such that the left adrenal gland is visualized and left undisturbed in its posterior location. As this dissection progresses through these thin connective tissue planes, the posterior surface of the pancreas and the splenic vein densely adherent to the pancreas are visualized. The spleen-pancreas complex can be mobilized over the top of the aorta, taking care to avoid injury of the superior mesenteric artery. After this mobilization, the spleen and distal pancreas are delivered to the level of the subcutaneous tissue. Laparotomy pads are placed in the left upper quadrant to maintain the spleen in the wound. At this point, the degree of injury can be
1694
clearly evaluated and the decision for extirpation or repair made. Splenectomy is usually indicated under the following circumstances: (1) the patient is unstable; (2) other injuries require prompt attention; (3) the spleen is extensively injured with continuous bleeding; and (4) bleeding is associated with hilar injury. There are both anterior (short gastric arteries) and posterior (splenic artery) blood supplies. After complete splenic mobilization, traction can be placed on the gastrosplenic ligament that places tension on and exposes the short gastric arteries. These are rapidly divided; with appropriate traction, the gastric wall is easily visualized, avoiding clamp injury. The spleen can then be grasped and elevated by the surgeon or assistant and the splenic artery identified at the superior border of the pancreas. The artery and vein are separately divided and ligated. Not infrequently, the tail of the pancreas extends right into the hilum, in which case it is generally more practical and safer to ligate the splenic vessels after they have divided to avoid the morbid complication of injury to the pancreatic tail. In the absence of pancreatic injury, drainage of the splenic fossa is not necessary. As noted earlier in this chapter, thrombocytosis occurs in about half of postoperative patients in the initial weeks after splenectomy. The thrombocytosis may increase the risk for deep venous thrombosis. When the platelet count rises higher than 750,000/mm3 , many surgeons treat the patient with antiplatelet therapy, low-dose heparin, or low-molecular-weight heparin. Pimpl and colleagues[65] reviewed 37,000 autopsies over 20 years of adults who died after splenectomy and compared them with a deceased population of 403 who did not have splenectomy. These investigators found higher incidence rates of lethal pneumonia, sepsis with multiorgan failure, purulent pyelonephritis, and pulmonary embolism in the splenectomy group. They concluded that splenectomy carries a considerable lifelong risk for severe infection and thromboembolism.
Splenorrhaphy
Attempts at splenic repair were initiated with appreciation of the entity of OPSI (see previous section). Additionally, splenic absence provides a relative dead space in the left upper quadrant, which often becomes occupied with blood clot or serum, creating a potential for subphrenic abscess. This occurrence is especially pronounced in the face of hollow viscus or pancreatic injuries. Those scenarios provide for bacterial colonization and culture media. Splenorrhaphy techniques became widely applied in the late 1970s, reaching a peak in the mid-1980s. Their application has gradually decreased since that time in association with the rise of nonoperative management. Splenorrhaphy was applied to nearly half of splenic injuries at the height of its use. A general rule is that if more than 1 unit of blood is required for salvage, splenectomy should be performed. Beyond that, the risks associated with transfusion generally outweigh the risks of OPSI. Four types of splenorrhaphy have been used: (1) superficial hemostatic agents (cautery, oxidized cellulose, absorbable gelatin sponge, topical thrombin); (2) suture repair; (3) absorbable mesh wrap; and (4) resectional dbridement.
Superficial hemostatic approaches are useful for American Association for the Surgery of Trauma grades I and II injuries (see Table 544 ). They may also be adjunctive in higher grades of injury. The argon-beam coagulator has received some support, but there is no clinical evidence that it is superior to other approaches. Suture repair of lacerations in grades II and III injuries has become common. When feasible, temporary occlusion of the splenic artery may reduce blood loss and facilitate repair. A problem with suture repair is the tendency for the sutures to tear the spleen further when tied. Pledgeted repairs reduce that occurrence. Many surgeons have used Teflon pledgets. The use of pledgets constructed of 2- to 3-cm absorbable gelatin sponge wrapped in oxidized cellulose and secured with suture ties to resemble a cigarette has been commonly applied ( Fig. 544 ). These are placed along each edge of the laceration and secured with a running polypropylene suture. That approach has proved effective for both splenic and hepatic lacerations. Mesh wrapping has been effectively used for grade III and some grade IV injuries. Investigators from Cook County Hospital provided some of the earliest data and description of this technique.[66] Disposable mesh, composed of either polyglycolic acid or polyglactin, has been used. If knitted mesh is used, a keyhole about 1 to 2 cm in diameter is cut in the middle of the mesh and is stretched, so that the spleen can be delivered through it, resulting in the keyhole around the splenic hilum. The edges of the mesh are then approximated with a running suture over the top, so that the spleen is effectively enclosed in a mesh sac. If woven mesh is used, it will not stretch; thus, the keyhole for surrounding the hilum is designed by dividing one side of a square piece of mesh to the center and constructing an appropriately sized circular hole. The mesh is tightened around the hilum by reapproximating the severed mesh. The mesh is secured
Figure 54-4 Pledgets constructed of gelatin sponge wrapped in oxidized cellulose are used to suture-repair a grade III splenic laceration.
1695
over the top of the spleen as described previously ( Fig. 545 ). This technique works surprisingly well, apparently through a tamponade effect. Resectional dbridement has been applied for major fractures, usually involving the upper or lower pole (grade II or IV). The raw surfaces are approximated. Pledgeted materials are of considerable benefit for reapproximating those edges. At least one third of the splenic mass is necessary to maintain immunocompetence. As noted previously, in the past, splenic conservation through splenorrhaphy was applied to nearly half of all splenic injuries. That percentage decreased substantially in the 1990s and probably now accounts for no more than 10%. A high percentage of splenorrhaphy was accounted for by simple techniques for repair of grades I and II injuries. Most of those types of injury are now being managed nonoperatively. Most patients who currently undergo operation for splenic injury have active bleeding or destructive injuries requiring splenectomy. Heightened awareness of risk for transmission of viral disease, especially hepatitis, with blood transfusion has also dampened enthusiasm for splenic repair.
Nonoperative Management
The basis for nonoperative management can be traced to OPSI and incidental notation of splenic injury in the early years of CT scanning for trauma. Nonoperative management originated in pediatric surgery. For several years, general surgeons questioned the judgment of their pediatric surgical colleagues. With the passage of time, it has become clear that nonoperative management is logical. Nonoperative management in adults is somewhat more controversial but is gaining continuously wider acceptance as data accumulate. Currently, 70% to 90% of children with splenic injury are successfully treated without operation, and 40% to 50% of adult patients with splenic injury are managed nonoperatively in large-volume trauma centers. The lower
Figure 54-5 Splenorrhaphy of grade IV splenic laceration is accomplished by wrapping with woven polyglactin.
percentage of nonoperatively managed splenic injury in adults than in children has been a source of speculation. It has been suggested that anatomic differences between adults and children are responsible, including a more elastic, cartilaginous rib cage providing protection and more elastin in the spleen producing contraction and some degree of hemostasis in children. Powell and colleagues,[67] however, have produced data demonstrating that the
differences in management of splenic injury between adults and children are most likely related to mechanisms of injury. Their analysis of 411 patients (293 adults and 118 children) found major differences in mechanism of injury in adults and children (P < .05), including the following: motor vehicle crash (67% in adults vs. 24% in children), motorcycle crash (9% vs. 1%), sports-related injury (2% vs. 17%), falls (9% vs. 25%), pedestrian and auto collision (4% vs. 11%), and bicycle crash (1% vs. 9%). Higher injury severity scores, lower Glasgow Coma Scale scores, and higher mortality rates indicated that adults were more severely injured. A fundamental rule for consideration of nonoperative management is the requirement that the patient be hemodynamically stable. Additionally, institutional resources should be such that the patient can be monitored in a critical care environment and that operating room facilities and personnel are available in the event of sudden bleeding that requires splenectomy. Most grades I and II injuries can be managed nonoperatively; these account for about 60% to 70% of cases of nonoperative management.[68] Although CT scanning is the fundamental measure for selecting nonoperative management, it has shortcomings that must be taken into consideration. A report by Sutyak and coworkers[69] involving CT in 49 patients with splenic injury correlated surgical with CT findings. These investigators found that CT matched surgical grading in 10 patients, underestimated it in 18, and overestimated it in 6. They also reported that 5 injuries were missed by CT and that radiologists disagreed in their interpretations of 20% of scans. Note, however, that the CT scan used in this study was not the current state-of-the-art helical CT scan. In early reports, most investigators expressed extreme caution regarding nonoperative management of grades III and IV, even with hemodynamic stability. As experience has accumulated, most feel comfortable with observing stable grade III injuries, and many have begun observing grades IV and V injuries.[68] [70] [71] Most of the early reports of nonoperative management in the 1980s were anecdotal, noting occasional successful management in highly selected cases. Since the 1990s, nonoperative management has become a more standardized approach. In a review in 1990, Shackford and Molin [72] reported on 1866 splenic injuries in which 13% were managed nonoperatively. Subsequently, Smith and colleagues[73] reported a 47% nonoperative management rate among 166 splenic injuries. These and most other reports have reserved nonoperative management for injury grades I through III. Analysis of failure rates is important in evaluating criteria for selecting appropriate patients for nonoperative management. Although age older than 55 years has been reported to be associated with high failure rates,[74] others have refuted that observation. [68] [70] [71] Nonoperative management
1696
TABLE 54-5 -- Comparison of Results of Nonoperative Management of Blunt Splenic Injuries from Published Series Study Splenic Injuries (n) 1866 309 166 135 524 204 368 558 Cases Planned Nonoperatively Managed (%) 13 25 47 18 61 68 57 77 Nonoperative Success (%) 69 87 97 52 94 93 86 92 31 13 3 48 6 7 14 8 Failure (%)
Shackford and Molin, 1990[72] Schurr et al, 1995[75] , * Smith et al, 1996[73] , * Morrell et al, 1996[74] Davis et al, 1998[68] Myers et al, 1999[76] Cocanour et al, 1999[70] Bee et al, 2001[77]
* Excluded grade IV and V injuries.
Figure 54-6 CT scan demonstrates vascular blush in an injured spleen.
failure rates from recent series are shown in Table 545 . An important finding that has been correlated with failed nonoperative management is the presence of a vascular blush on CT examination. Schurr and associates[75] reported on 309 blunt splenic injuries, of which 29% were managed nonoperatively. They noted a 13% failure rate; two thirds of the failures were associated with a vascular blush ( Fig. 546 ). Those vascular blushes were proved to represent false aneurysms of intraparenchymal branches of the splenic artery ( Fig. 547 ). The cause of failure from false aneurysms is gradual enlargement of the aneurysm, with rupture of the aneurysm and spleen. This pathophysiology most likely accounts for many of the instances of delayed splenic rupture noted from the past. Not all of these pseudoaneurysms rupture, however. It appears likely that 30% to 40% spontaneously thrombose.[75] A subsequent study from the same institution by Davis and colleagues[68] dealt with 524 blunt splenic injuries during a 4.5-year period. In that report, a protocol was followed in which vascular blushes identified by CT were followed by
Figure 54-7 After splenectomy, the sectioned spleen demonstrated a false aneurysm, which corresponds to the vascular blush in Figure 546 .
angiography with embolization of the false aneurysm ( Fig. 548 ). That approach yielded a failure rate of 5% for nonoperative management, a rate significantly lower (P < .03) than their prior experience. Controversy exists concerning the need for follow-up CT evaluations. Some institutions have suggested that follow-up scans are unnecessary.[70] [76] Davis and colleagues,[68] however, noted that 74% of vascular blushes were seen only in follow-up studies. Because a high rate of failure was noted with the finding, these investigators recommend follow-up scans within 2 to 3 days for all but the most trivial lesions. The reasons for absence of vascular blush on initial scans include missed 1-cm-cut protocol and lysis of initial clot. The cause of other failures is less clear. Investigators have become more aggressive by including patients with higher injury grades (IV and V). Fifty-six of 106 grades IV and V injuries were nonoperatively managed with a failure rate of 18%.[68] This and subsequent reports have also correlated a higher failure rate with large amounts of free intraperitoneal blood on the admission CT scan
1697
Figure 54-8 A and B, Angiographic embolization of a false aneurysm, which was demonstrated by helical CT.
(perisplenic and perihepatic and gutter and pelvis).[68] [70] [76] It seems intuitive that larger degrees of splenic injury as indicated by CT scan and by volume of intraperitoneal blood would be associated with higher failure rates. In a recent report published by Bee and coworkers, the published contraindications to nonoperative management (NOM) of blunt splenic injury were reviewed and compared with their experience with nonoperative failures.[77] The published indications that were reviewed included age of 55 years, Glasgow Coma Scale score of 13, admission blood pressure lower than 100 mm Hg, major (grades III through V) injuries, and large amounts of hemoperitoneum in their evaluation of 430 consecutive patients who were observed. The nonoperative management failure rate was 8%. Multivariate analysis identified age of 55 in grades III to V as independent predictors of failure. They noted the highest failure rate (30% to 40%) occurred in patients older than 55 years with major injury or moderate to large amounts of
hemoperitoneum. They noted a mortality rate for successful nonoperative management of 12% compared to 9% for failed nonoperative management and concluded that failed nonoperative management was not significantly associated with adverse outcome. An ideal situation would be for 70% to 80% of cases to be manageable nonoperatively, with a 1% to 2% failure rate. This may not be possible. Future clinical studies should be designed to address these important issues of selection, causes of failure, and use of adjunctive techniques, including embolization. We believe that advancing CT technology, which will better define the injuries combined with clinical trials, will ultimately approach those numbershigh nonoperative management rates coupled with very low failure rates. No objective data are available related to recommendations for return to activity after splenic injury. For grades I and II injuries, 2 to 3 weeks is probably adequate healing time. For higher-grade injuries, 6 to 8 weeks is probably more appropriate, and many surgeons would obtain a CT scan at that time to evaluate degree of healing before return to excessive physical activity.
Evidence-Based Medicine Evaluation
The use of evidence-based medicine (EBM) to address optimal management of clinical problems is gaining increased acceptance. One of the most important applications of EBM is to develop practice management guidelines. There are two basic methodologies. The first is through statistical methodology centered on metaanalysis. Rigorous meta-analytic techniques require randomized, controlled trials for analysis; however, there are relatively few randomized, controlled trials for addressing most questions in surgery. Therefore, a second methodology for EBM addresses questions through data classification and assessment of confidence levels for recommendations. Extensive work on data classification has been done by the Canadian and U.S. Preventive Task Forces and by the Agency for Health Care Policy and Research of the U.S. Department of Health and Human Services.[78] A classification system based on that work is outlined in Box 543 . The Agency for Health Care Policy and Research has also developed a system of assessing confidence levels for recommendations based on that data classification system, which is presented in Box 544 . Using those methodologies for data evaluation and assessment, the Eastern Association for the Surgery of Trauma (EAST) has published 11 patient management guidelines on their Internet website (www.east.org). One of those evidence-based guidelines addresses nonoperative management of splenic injury. To develop the guidelines for splenic injury, a study group consisting of seven surgeons identified 50 Englishlanguage clinical articles published between 1976 and 1996 addressing pertinent questions about blunt splenic injury management. The recommendations based on the classification and assessment of those published data are listed in
1698
Box 54-3. Evidence-Based Classification of Medical Literature Adapted from Recommendations by the Agency for Health Care Policy and Research Class I Evidence Prospective, randomized controlled trialsthe gold standard of clinical trials. Some may be poorly designed, have inadequate numbers, or suffer from other methodologic inadequacies and thus may not be clinically significant. Class II Evidence Clinical studies in which the data were collected prospectively and retrospective analyses that were based on clearly reliable data. These types of studies include observational studies, cohort studies, prevalence studies, and case-control studies. Class III Evidence Studies based on retrospectively collected data. Evidence used in this class includes clinical series, databases or registries, case reviews, case reports, and expert opinion.
Box 54-4. Categorization of Strengths of Recommendations Based on a Medical Literature Review of a Specified Topic * According to Methodologies Derived from the Agency for Health Care Policy and Research Level 1 This recommendation is convincingly justifiable based on the available scientific information alone. It is usually based on class I data; however, strong class II evidence may form the basis for a level 1 recommendation, especially if the issue does not lend itself to testing in a randomized format. Conversely, weak or contradictory class I data may not be able to support a level 1 recommendation.
Level 2 This recommendation is reasonably justifiable by available scientific evidence and strongly supported by expert critical care opinion. It is usually supported by class II data or a preponderance of class III evidence. Level 3 This recommendation is supported by available data, but adequate scientific evidence is lacking. It is generally supported by class III data. This type of recommendation is useful for educational purposes and in guiding future studies.
* See Table 541 .
Box 54-5. Eastern Association for the Surgery of Trauma * Recommended Patient Management Guidelines for the Nonoperative Management (NOM) of Blunt Injuries to the Liver and the Spleen Level I There are insufficient data to suggest NOM as a level I recommendation for the initial management of blunt injuries to the liver and/or spleen in the hemodynamically stable patient. Level II 1. There are class II and mostly class III data to suggest that NOM of blunt hepatic and/or splenic injuries in a hemodynamically stable patient is reasonable. 2. Severity of hepatic or splenic injury (as suggested by CT grade or degree of hemoperitoneum), neurologic status, and/or the presence of associated injuries are not contraindications to NOM. 3. Abdominal CT is the most reliable method to identify and assess the severity of the injury to the spleen or liver. Level III 1. The clinical status of the patient should dictate the frequency of follow-up scans. 2. Initial CT of the abdomen should be performed with oral and intravenous contrast agents to facilitate the diagnosis of hollowviscus injuries. 3. Medical clearance to resume normal activity status should be based on evidence of healing. 4. Angiographic embolization is an adjunct in the NOM of the hemodynamically stable patient with hepatic and splenic injuries and evidence of ongoing bleeding.
* See their website at www.east.org.
Box 545 . The EAST evidence-based analysis demonstrates that there are neither class I data nor level I recommendations. Those findings make it clear that splenic injury management must be addressed by prospective, multiinstitutional trials. A strong case for such trials can be made because splenic injury occurs at a high incidence and is associated with substantial cost and morbidity. The intent of EBM-derived guidelines is that they be adopted at the local level into management schemes through clinical pathways and algorithms that consider resource availability and practice patterns. An algorithm developed from the EAST recommendations and their accompanying literature review is represented in Figure 549 .
ELECTIVE LAPAROSCOPIC SPLENECTOMY The technique of open splenectomy was described in detail in the previous section on splenic trauma. Many surgeons now prefer to use the laparoscopic approach for
1699
Figure 54-9 Algorithm for nonoperative management of splenic injuries in the hemodynamically stable patient. (From Bee TK, Croce MA, Miller PR, et al: Failures of splenic nonoperative management: Is the glass half empty or half full? J Trauma 50:231, 2001.)
most elective splenectomies. The technique of laparoscopic splenectomy was first described in 1992.[79] [80] [81] In experienced hands, laparoscopic splenectomy can be performed as safely and effectively as open splenectomy, particularly for hematologic diseases in which the spleen size is normal or only slightly enlarged.[9] [82] [83] Early experiences with laparoscopic splenectomy have demonstrated many similarities to the early days of laparoscopic cholecystectomy. Operative time is longer for laparoscopic splenectomy, but the procedure offers the advantages of more rapid postoperative recovery and shorter duration of hospital stay, leading some to consider laparoscopic splenectomy the standard of care for some hematologic disorders requiring splenectomy.[83] Results of laparoscopic splenectomy for benign hematologic diseases, such as ITP, should be compared with the standard of open splenectomy, which is technically feasible in 100% of patients and is associated with a hospital mortality rate of less than 1% and a morbidity rate of 10% to 20%. In a retrospective literature review of 1358 patients who underwent open splenectomy, the rate of wound-related complications was estimated to be 3%. The mean postoperative hospital stay ranged from 7.5 to 11 days.[84] [85] [86] [87] [88] [89] [90] In patients with splenomegaly secondary to malignant hematologic disorders, however, the operative mortality rates are increased in the range of 0 to 18% and morbidity rates are increased in the range of 19% to 56%, respectively.[91] [92] [93] In comparison, laparoscopic splenectomy can be completed in about 90% of properly selected patients. The incidence of conversion to open splenectomy is between 0 and 20%. Most of the conversions are caused by intraoperative bleeding, but lack of surgical experience,[82] [94] [95] [96] extensive adhesions,[95] [97] large splenomegaly,[82] [94] [95] [98] [99] [100] [101] and obesity[80] [82] [94] [102] are also involved. A significant learning curve is observed with laparoscopic splenectomy, and with increasing experience, the conversion rate has been reported to decrease dramatically.[96] [103] Yee[96] and Glasgow[103] and their associates reported a conversion rate of 36% during the initial 11 laparoscopic splenectomies; during the subsequent operations, the conversion rate dropped to 0% to 5%. In two reviews[8] [9] of laparoscopic series that included 418 and 948 patients, respectively, the mean operative time for laparoscopic splenectomy ranged from 88 to 261 minutes, with an open conversion rate ranging from 0 to 30% ( Table 546 ). In a multivariate analysis by Friedman and coworkers,[82] operative time was significantly related to patient age, hematologic diagnosis, operative technique, and splenic weight. The perioperative morbidity rates averaged 8% and 12%, respectively (range, 0 to 30%), and the mortality rate was 0.7% (0 to 6%). Most deaths were attributable to the patients underlying diseases or hematologic disorder. In the review by Gigot and associates[8] of 984 patients, 119 had reported complications. Bleeding was the most frequent perioperative complication and has been significantly linked to the surgeons learning curve.[97] The mean number of patients requiring intraoperative or postoperative transfusions was 13%, ranging from 0 to 40%. Local complications, including wound-related complication (seroma, hematoma, infection, evisceration, or incisional hernia) occurred in 1.5%. Postoperative bleeding occurred in 1% of patients, and 73% of these cases required re-exploration. Pancreatic complications (pancreatitis or pancreatic fistula) occurred in 0.6% and subphrenic abscess in 0.5%. General postoperative complications occurred in 7.4%; most (3.2%) were pleuropulmonary. Postoperative recovery after laparoscopic splenectomy is surprisingly fast, as has previously been observed with laparoscopic cholecystectomy. The length of stay ranged from 1.8 to 6 days after laparoscopic splenectomy.[8] [9] Most patients are able to return to full activities within 1 week if their underlying hematologic disorder allows. In the study by Flowers and coworkers,[97] 9% of patients returned to work in 7 days, and all patients with uncomplicated
1700
TABLE 54-6 -- Retrospective Case-Control Series Comparing Open and Laparoscopic Splenectomies in Adults
Yee et al, 1995[130] OS Patients Operative time (min) Delay before regular diet (days) Blood loss (mL) Transfusion rate Complications (%) POHS Total hospital costs USD 103 Cure rate at FU in ITP (%) Mean FU (mo) AS detected (%) Return to full activity (days) 25 156 4.3 273 12 12 6.7 25
Rhodes Brunt et al, et al, Watson et 1995[131] 1996[121] al, 1997[132] LS 13 88 0 0 2 11 24 20 26
Diaz et al, 1997[133] OS 15 116 359 13 8.8 LS 15 196 385 6.7 2.3
Delaitre Friedman et & Pitre, al, 1997[82] 1997[102] OS 74 121 3.2 437 27 22 6.7 LS OS LS 63 28 28
*
Smith et al, 1996[123] OS 10 131 4.4 20 20 5.8 LS 10 261 * 1.9 * 40 0 3*
Hashizume Glasgow et al, et al, 1996 [134] 1997[103] OS 28 156 4.3 274 18 14 6.7 LS 52 196 2 320 27 10 4.8 * OS 41 LS 10
LS OS LS OS LS OS 47 84 13 19 10 198 75 120 134 202 2.1 4.1 1.4
153 127 183 1.5 * 259 3.5 12 36 32 29 11

*
249 100
319 222 376 16 8 5.1 * 8 7 27 15 23 10 23
512 176 * 46 20 0 8.2
3 5.8 2.5 75
3.5 * 8.6 5.1
13,433 9207 76 24 81 8
4224 2238 16,362 18,015 10,900 9700

*
13,196 17,071 17,876 20,295 9264 6438
75
83 60 6
92 14 15
75 20 23
80 29 12 *
13.5
19
86 18
93 11
30 6 wk
81 21
81 21 12 wk
74 16
63
80
6.5 6.5 5 5
FU, follow-up; ITP, immune thrombocytopenic purpura; LS, laparoscopic splenectomy; OS, open splenectomy; POHS, postoperative hospital stay; USD, U.S. dollars. From Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:159, 1998.
P < .001. * P < .05. P = .0005.
1701
laparoscopic splenectomy were fully recovered by 21 days, regardless of profession. Although a prospective, randomized comparison has not been conducted, several retrospective case-control series have compared the laparoscopic approach to an open splenectomy (see Table 546 ). Even though it is difficult to extrapolate definitive data from these series because of potential selection bias (patient age, diagnosis and indication, splenic weight, and major comorbid conditions), the results are consistently in favor of the laparoscopic approach. Each series has demonstrated an earlier resumption of diet, reduced postoperative analgesics, and decreased postoperative hospital stay. Despite the promising technical results of laparoscopic splenectomy, the hematologic response and long-term cure rates are the most important. Open splenectomy for the treatment of ITP achieves a long-term cure rate of 65% to 90%. In clinical series of laparoscopic splenectomy, the mean follow-up is usually limited to 1 to 2 years. During this short follow-up, reports of 76% to 100% success rates have been published. Similarly, Katkhouda and associates[9] reported that the treatment of ITP appears to be at least as good after laparoscopic surgery as after the open technique in comparisons of separate series (see Tables 541 and 542 ). Gigot and colleagues[8] reported on 279 patients with ITP where an absence of initial response and/or recurrent thrombocytopenia occurred in approximately 15% ( Table 547 ). The causes of recurrence in patients with ITP are multifactorial, but residual accessory spleens are well known to be one of the factors of recurrence. There have been several reported cases of recurrent thrombocytopenia secondary to retained accessory spleens (see Table 547 ). This issue of accessory spleens has been prominent in the discussion of the role of laparoscopic splenectomy. Autopsy studies have demonstrated that the incidence of accessory spleens in the normal population is about 10%. In clinical series, however, accessory spleens have been reported in 15% to 30% of patients. The incidence of accessory spleen in recently reviewed splenectomy series since 1980 is 15% to 16%.[9] The incidence probably correlates with the diligence of the search and is higher than commonly estimated because small accessory spleens are easily missed or mistaken for lymph nodes. Because accessory spleens are often easier felt than seen, it is likely that the incidence of missed splenunculi is higher at laparoscopic than open splenectomy. The reliability of laparoscopic exploration in dealing with the problem of accessory spleens is a key factor in establishing the long-term credibility of the laparoscopic approach.[8] Though the indications for a laparoscopic splenectomy remain the same as open splenectomy, some cases require caution in performing laparoscopically. Absolute contraindications to the laparoscopic approach include severe cardiopulmonary disease, cirrhosis, and pregnancy.[104] Variceal short gastric vessels compounded by TABLE 54-7 -- Long-Term Hematologic Cure Rate after Laparoscopic Treatment of Immune Thrombocytopenic Purpura Patients Absence of Initial Response and/or Residual Recurrent Accessory Spleen Thrombocytopenia Detected 1 4 1 1 ? 1 1 ?
Study Emmermann et al, 1995[119] Yee et al, 1995[130] Liew and Storey, 1995[135] Legrand et al, 1996[95]
Patients (n) 20 14 6 9
Mean Follow-Up (mo [range]) 14 (128) At discharge ? 12 (126)
Parent et al, 1995[136] Flowers et al, 1996[97] Dexter et al, 1996[137] Zamir et al, 1996 [138] Katkhouda et al, 1996[139] Watson et al, 1997[132] Tsiotos and Schinkert, 1997[124] Lee and Kim, 1997[140] Glasgow et al, 1997[103] Delaitre and Pitre, 1997[102] Decker et al, 1998[99] Trias et al, 1998[101] Gigot, 1998 *
*Unpublished data.
11 22 6 15 20 13 18 15 16 26 17 32 19
(69) 21 (336) 24 (1733) (236) 20 (146) 14 (521) 15 (130) ? ? (348) 12.5 (128) 12 (150) 45 (2263)
1 4 0 0 0 1 1 3 4 2 4 9 6
? ? 1 ? ? 1 ? ? 3 3
?, data not given despite treatment failuresnot assessed; , no treatment failures, therefore not assessed.
1702
the coagulopathy of liver disease present an unacceptable risk of operative hemorrhage and temper enthusiasm for the laparoscopic approach in patients with portal hypertension. Thrombocytopenia in pregnancy is frequently gestational. Surgery is reserved for the failure of medical management and associated with a fetal mortality of 31%.[105] Though laparoscopic cholecystectomy has been shown to be safe in the second trimester of pregnancy, results of laparoscopic splenectomy in this rare patient population have not been reported. Initially, splenomegaly was thought to be an absolute contraindication to the laparoscopic approach. Increasing experience and improvement of surgical devices have made this a relative contraindication. Splenic size and surgeon experience are the determining factors. Though technically feasible, laparoscopic splenectomy in the patient with splenomegaly can be a challenge. The introduction of hand-assisted laparoscopic surgery has led some surgeons to approach these larger spleens with outcomes similar to the totally laparoscopic approach.[106] The laparoscopic technique may be performed with the patient in the supine (or modified lithotomy) position[107] or in the right lateral decubitus position. [108] After induction of general anesthesia and endotracheal intubation, a nasogastric tube and a urinary catheter are inserted, and pneumatic compression stockings are applied. Appropriate patient positioning is of paramount importance to the successful completion of a laparoscopic splenectomy. With either the lateral or the supine approach, the patient is placed so that the kidney rest can be raised to maximize the space between the iliac crest and the costal margin. The patient is positioned so that the table may be flexed to create a wider working space. The patient is then placed in a reverse Trendelenburg position to facilitate gravity retraction of the viscera away from the left upper quadrant. In the supine approach, the surgeon stands to the left of the patient, and the first assistant and camera assistants stand to the right.[97] It is often easier for a right-handed surgeon to work from a position between the patients legs, with the patient in a modified lithotomy position. The scrub nurse stands to the patients left side near the foot of the table. Alternatively, the patient may be placed on a beanbag in a 60-degree right lateral decubitus position, with a right axillary roll. The left arm should be supported by a splint. In this approach, the surgeon and the scrub nurse stand to the patients right, and the assistants stand to the left. The spleen is suspended from its diaphragmatic attachments, and gravity retracts the stomach, transverse colon, and greater omentum, while placing the splenic hilum under tension. For both approaches, the video monitors are placed on each side of the table, at or above the level of the patients shoulders. Pneumoperitoneum is established to a pressure of 12 to 15 mm Hg, and three to five 2- to 12-mm diameter operating ports are used, with the camera port at the umbilicus or offset between the umbilicus and the left costal margin. The other port sites are arrayed in the positions depicted in Figure 54 10. The operation is begun with a thorough search of the abdominal cavity for the presence
Figure 54-10 Strict lateral position of the patient for laparoscopic splenectomy. The table is angulated, giving forced lateral flexion of the patient to open the costophrenic space. Trocars are inserted along the left costal margin more posteriorly. The spleen is hanged by its peritoneal attachments. The numbered lines show the position of laparoscopic ports. (From Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:149, 1998.)
of accessory splenic tissue ( Fig. 5411 ). The stomach is retracted to the right to facilitate inspection of the gastrosplenic ligament. The splenocolic ligament, the greater omentum, and the phrenosplenic ligament are then carefully inspected. The small and large bowel mesenteries, the pelvis, and the adnexal tissues all should be inspected. The gastrosplenic ligament should be opened and the area of the pancreatic tail inspected. Our preference has been to use the lateral decubitus approach.[108] The initial dissection is begun by mobilization of the splenic flexure of the colon. The splenocolic ligament is divided using sharp dissection. This mobilizes the inferior pole of the spleen and allows the spleen to be retracted cephalad. Great care is taken to avoid rupture of the splenic capsule during retraction. The lateral peritoneal attachments of the spleen are then incised using either sharp dissection or ultrasonic shears. A 1-cm cuff of peritoneum is left along the lateral aspect of the spleen to be grasped if the spleen needs to be drawn medially. The lesser sac is entered along the medial border of the spleen. With the spleen elevated, the
1703
Figure 54-11 Usual location of accessory spleens: (1) gastrosplenic ligament, (2) splenic hilum, (3) tail of the pancreas, (4) splenocolic ligament, (5) left transverse mesocolon, (6) greater omentum along the greater curvature of the stomach, (7) mesentery, (8) left mesocolon, (9) left ovary, (10) Douglas pouch, (11) left testis. (From Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:156, 1998.)
short gastric vessels and main vascular pedicle are visualized. The tail of the pancreas is also visualized and avoided at this point as it approaches the splenic hilum. The short gastric vessels are divided. This can be accomplished with several different modalities as long as the surgeon has familiarity with the device and its application. Currently available instrumentation for the control of splenic vessels include ultrasonic dissector, hemoclips, bipolar devices, Ligasure, or an endovascular stapling device. The use of hemoclips should be minimized throughout the procedure and especially around the hilum because the clips may interfere with future applications of a stapling device. The stapler will not function if a clip is caught within its jaws, and this can result in significant bleeding from hilar vessels. After the short gastric vessels have been divided, the splenic pedicle may be carefully dissected from both the medial and lateral aspects (an advantage over the anterior approach). After the artery and vein are dissected, the vessels are divided by application of endoscopic vascular staplers or suture ligatures. In the more common distributed mode, there are multiple vascular branches that enter
Figure 54-12 Extraction of the spleen within a heavy plastic bag. Instrumental morcellation of the organ with forceps. (From Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:154, 1998.)
the spleen; these arise from the main vascular trunks about 2 to 3 cm from the hilum.[9] For this reason, we generally try to keep our dissection 2 cm from the splenic capsule. Several branches may still be encountered, however, and should be controlled individually if necessary. A pedicle formed by the artery and vein that enters the hilum as a compact bundle is known as the magistral mode. In this circumstance, the pedicle should be transected en bloc using a linear vascular stapler. The tail of the pancreas is well visualized as the staplers are applied to avoid injury to this structure. The surgeon should be acutely aware of the position of the tail of the pancreas during the hilar division. The pancreatic tail lies within 1 cm of the splenic hilum in 75% of patients and touches the splenic hilum in 30%.[109]
After the hilar vessels have been controlled, the completely devascularized spleen is suspended by a small cuff of avascular superior pole splenophrenic attachments. This is left in place to facilitate entrapment of the spleen into the retrieval bag. To remove the detached spleen, a puncture-resistant nylon extraction bag is introduced through one of the trocar sites, typically the left lateral site. The bag is opened within the abdominal cavity, and the spleen is placed into the bag. The drawstring is grasped and the bag is closed, leaving only the superior pole attachments to be divided at this stage. The open end of the closed bag is brought outside the abdomen through the supraumbilical or epigastric trocar site. The spleen is then morcellated with ring forceps and with finger fracture and is removed in fragments ( Fig. 5412 ).
1704
In conditions requiring pathologic evaluation of an intact spleen, an incision adequate to enable removal of the bag containing the intact spleen is made. Care must be taken to avoid spillage of any fragments of spleen into either the abdominal cavity or the wound. The laparoscope is reinserted, and the splenic bed is assessed for hemostasis. At this point, drains may be placed if deemed necessary. The fascia of all trocar ports larger than 5 mm should be closed.
LATE MORBIDITY AFTER SPLENECTOMY Postsplenectomy thrombocytosis may be associated with both hemorrhagic and thromboembolic phenomena. This occurs particularly in patients with myeloproliferative disorders such as CML, agnogenic myeloid dysplasia, essential thrombocytosis, and polycythemia vera. Thrombosis of the mesenteric, portal, and renal veins may be a life-threatening sequela of postsplenectomy thrombocytosis.[110] The lifelong risk of deep venous thrombosis and pulmonary embolism has not been well defined but may be significant. In review of 37,012 autopsies over a 20-year period, Pimpl and colleagues[65] identified 202 deceased adults who had a history of splenectomy and matched them with a cohort of 403 deceased patients who had not undergone splenectomy. Pulmonary embolism was the major or contributory cause of death more often in the splenectomy group (35.6%) than in the control group (9.7%).[65] OPSI is among the more devastating sequelae of asplenia and is the most common fatal late complication of splenectomy.[111] Hyposplenism in the neonatal period has been suggested to contribute to the poor outcomes from neonatal sepsis. The exact incidence of OPSI has been difficult to determine. The incidence of infection in postsplenectomy patients is likely to be under-reported. In the same autopsy series by Pimpl and colleagues mentioned in the previous paragraph, lethal pneumonia was identified twice as often in autopsies of splenectomized patients than in controls (57.9% vs. 24.1%) and lethal sepsis with multiorgan failure occurred in 6.9% of splenectomized versus 1.5% of autopsies on controls.[65] One consistent observation is that the risk for OPSI is greater after splenectomy for malignancy or hematologic disease than for trauma. The risk also appears to be greater in young children (<4 years of age). The risk for fatal OPSI is estimated to be 1 per 300 to 350 patient-years follow-up for children and 1 per 800 to 1000 patientyears follow-up for adults.[112] The incidence of nonfatal infection and sepsis is likely to be significantly greater. A recent review of selected reported splenectomy series of 7872 total cases inclusive of both children and adults revealed 270 episodes of sepsis (3.5%), with 169 septic fatalities (2.1%).[113] Infection may occur at any time after splenectomy; in one recent series, most infections occurred more than 2 years after splenectomy, and 42% occurred more than 5 years after splenectomy.[114] OPSI typically begins with a prodromal phase characterized by fever and chills and nonspecific symptoms, including sore throat, malaise, myalgias, diarrhea, and vomiting. Patients may have had rigors for 1 to 2 days before seeking appropriate medical treatment. Pneumonia and meningitis may be present, but many cases have no identifiable focal site of infection and present with high-grade primary bacteremia. Progression of the illness is classically rapid, with the development of hypotension, disseminated intravascular coagulation, respiratory distress, coma, and death within hours of presentation. The mortality rate is between 50% and 70% for fully developed OPSI despite antibiotics and intensive care. Survivors often have a long and complicated hospital course with severe sequelae, such as peripheral gangrene requiring amputation, deafness from meningitis, mastoid osteomyelitis, bacterial endocarditis, and cardiac valvular destruction. S. pneumoniae is the most frequently involved organism in OPSI and is estimated to be responsible for between 50% and 90% of cases. Other organisms involved in OPSI include Haemophilus influenzae, Neisseria meningitidis, Streptococcus species and other than pneumococcal species, Salmonella species, and Capnocytophaga canimorsus (implicated in OPSI as a sequela of dog bites).
Prophylactic Treatment of Splenectomized Patients
The spleen is important for generating responses to thymus-independent antigens. In elective procedures, immunization should be administered before splenectomy whenever possible; the Advisory Committee on Immunization Practices has recommended that the immunization precede splenectomy by at least 2 weeks.[115] Presplenectomy immunization is not possible in cases of splenic trauma. The immunizations should be administered to these patients during the hospitalization in which their splenectomy occurred rather than waiting until they return for a follow-up visit. Many of these patients become lost to follow-up, and clinical studies have demonstrated adequate antibody response to immediate immunization.[116] [117] High-risk patients without spleens should be considered for revaccination if they received the earlier 14-valent vaccine rather than the more current 23-valent preparation or if more than 3 to 6 years have elapsed since primary immunization.[117] Simultaneous immunization with H. influenzae type b, meningococcal serogroup C, and polyvalent pneumococcal vaccine is both immunogenic and well tolerated. Unfortunately, rare cases of OPSI have been reported in vaccinated patients. Penicillin prophylaxis is commonly practiced in children during the first few years after splenectomy, and some authorities have advocated this form of prophylaxis in adults, although data showing the efficacy of this treatment are lacking. OPSI has been reported in both adults and children taking prophylactic penicillin, despite penicillin-sensitive pneumococci infection.[112] Available data do not support the practice of long-term penicillin prophylaxis in asplenic adults. Another approach that appears rational is to provide the asplenic patient with a supply of oral antibiotics, such as amoxicillin, with instructions to begin taking the medication at the onset of rigor or a
1705
febrile illness if appropriate medical evaluation is not immediately available. Fever and rigor in an asplenic patient should prompt immediate aggressive, empirical treatment with antibiotic coverage, even in the absence of culture data.
Selected References
Advisory Committee on Immunization Practices: Prevention of pneumococcal disease: Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 46:124, 1997. This report of the Advisory Committee on Immunization Practices provides a summary of the epidemiology of pneumococcal infections as well as treatment and prophylaxis recommendations. Groups at high risk are discussed in detail, including patients who have undergone splenectomy. Bohnsack JF, Brown EJ: The role of the spleen in resistance to infection. Annu Rev Med 37:4959, 1986. This is a comprehensive review of the roles played by the spleen in protection from infection. The roles of opsonins and splenic bacterial clearance are well covered. George JN, Woolf SH, Raskob GE, et al: Idiopathic thrombocytopenic purpura: A practice guideline developed by explicit methods for the American Society of Hematology. Blood 88:340, 1996. This report comprehensively summarizes the practice guideline recommendations for the treatment of idiopathic thrombocytopenic purpura (ITP) as established by the American Society of Hematology in 1994. This comprehensive review provides the evidence for the current treatment recommendations and the data for treatment outcomes for children and adults with ITP. Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:147167, 1998. Gigot and colleagues nicely summarized the published experience on the role of laparoscopic splenectomy in the management of hematologic disorders. The authors provide useful technical tips for the practicing surgeon. Katkhouda N, Hurwitz MB, Rivera RT, et al: Laparoscopic splenectomy: Outcome and efficacy in 103 consecutive patients. Ann Surg 228:568578, 1998. Katkhouda and colleagues show the safety and efficacy of laparoscopic splenectomy in a large series of patients. The discussion section of this article provides an extensive review of previously published series and provides helpful tables comparing the outcomes of open versus laparoscopic splenectomy. Schwartz SI: Role of splenectomy in hematologic disorders. World J Surg 20:11561159, 1996. This review is by one of the recognized surgical authorities on splenic diseases and their management. This is a comprehensive review of the role of splenectomy in hematologic disorders. Styrt B: Infection associated with asplenia: Risks, mechanisms, and prevention. Am J Med 88:533N, 1990. This is an excellent review of the risks, mechanisms, and prevention of infections that are associated with the condition of asplenia. The mechanisms by which the spleen protects from certain infectious organisms are well covered. Data regarding the level of risk to asplenic patients are provided. Specific recommendations regarding antibiotic and vaccine prophylaxis are also provided.
www.mdconsult.com
References
1. Groom 2. Horton
AC: The Microcirculatory Society Eugene M. Landis Award Lecture: Microcirculation of the spleennew concepts, new challenges. Microvasc Res 34:269289, 1987. J, Ogden ME, Williams S, et al: The importance of splenic blood flow in clearing pneumococcal organisms. Ann Surg 195:172176, 1982.
3. Drew PA, Kiroff GK, Ferrante A, et al: Alterations in immunoglobulin synthesis by peripheral blood mononuclear cells from splenectomized patients with and without splenic regrowth. J Immunol 132:191196, 1984. 4. Bohnsack 5. Foster
JF, Brown EJ: The role of the spleen in resistance to infection. Annu Rev Med 37:4959, 1986.
PN, Bolton RP, Cotter KL, et al: Defective activation of neutrophils after splenectomy. J Clin Pathol 38:11751178, 1985. JN, el-Harake MA, Raskob GE: Chronic idiopathic thrombocytopenic purpura. N Engl J Med 331:12071211, 1994.
6. George 7. George
JN, Woolf SH, Raskob GE, et al: Idiopathic thrombocytopenic purpura: A practice guideline developed by explicit methods for the American Society of Hematology. Blood 88:340, 1996.
8. Gigot JF, Jamar F, Ferrant A, et al: Inadequate detection of accessory spleens and splenosis with laparoscopic splenectomy: A shortcoming of the laparoscopic approach in hematologic diseases. Surg Endosc 12:101106, 1998. 9. Katkhouda 10. Tsoukas 11. Croom
N, Hurwitz MB, Rivera RT, et al: Laparoscopic splenectomy: Outcome and efficacy in 103 consecutive patients. Ann Surg 228:568578, 1998.
CM, Bernard NF, Abrahamowicz M, et al: Effect of splenectomy on slowing human immunodeficiency virus disease progression. Arch Surg 133:2531, 1998.
RD III, McMillan CW, Orringer EP, et al: Hereditary spherocytosis: Recent experience and current concepts of pathophysiology. Ann Surg 203:3439, 1986. SI: Role of splenectomy in hematologic disorders. World J Surg 20:11561159, 1996. LH, Ginsberg MS, Burt ME, et al: MRI as an alternative to CT-guided biopsy of adrenal masses in patients with lung cancer. Ann Thorac Surg 65:193197, 1998. AH, Qaisaruddin S, Nasserallah Z, et al: Splenectomy in patients with sickle cell disease. Am J Surg 172:254258, 1996. AH: The role of splenectomy in patients with sickle cell disease. Ann Saudi Med 17:1997.
12. Schwartz 13. Schwartz 14. al-Salem 15. al-Salem
16. Badaloo AV, Singhal A, Forrester TE, et al: The effect of splenectomy for hypersplenism on whole-body protein turnover, resting metabolic rate, and growth in sickle cell disease. Eur J Clin Nutr 50:672675, 1996. 17. Taylor
MA, Kaplan HS, Nelsen TS: Staging laparotomy with splenectomy for Hodgkins disease: The Stanford experience. World J Surg 9:449460, 1985.
1706
18. Brodsky
J, Abcar A, Styler M: Splenectomy for non-Hodgkins lymphoma. Am J Clin Oncol 19:558561, 1996.
19. Morel P, Dupriez B, Gosselin B, et al: Role of early splenectomy in malignant lymphomas with prominent splenic involvement (primary lymphomas of the spleen): A study of 59 cases. Cancer 71:207215, 1993. 20. Au
WY, Klasa RJ, Gallagher R, et al: Second malignancies in patients with hairy cell leukemia in British Columbia: A 20-year experience. Blood 92:11601164, 1998. HM, Ratain MJ: Recent advances in the treatment of hairy cell leukemia. N Engl J Med 316:870872, 1987. R, Lee TC, Piro LD, et al: The optimal management of hairy cell leukaemia. Drugs 49:921931, 1995.
21. Golomb 22. Gollard 23. Saven
A, Burian C, Koziol JA, et al: Long-term follow-up of patients with hairy cell leukemia after cladribine treatment. Blood 92:19181926, 1998. E: Chronic lymphoproliferative disorders. Curr Opin Oncol 9:3441, 1997.
24. Montserrat 25. Rai
KR, Sawitsky A, Cronkite EP, et al: Clinical staging of chronic lymphocytic leukemia. Blood 46:219234, 1975. JC Jr, Seymour JF, Lerner S, et al: Role of splenectomy in chronic lymphocytic leukemia. J Am Coll Surg 185:237243, 1997. JF, Cusack JD, Lerner SA, et al: Case/control study of the role of splenectomy in chronic lymphocytic leukemia. J Clin Oncol 15:5260, 1997.
26. Cusack
27. Seymour 28. Khouri
I, Sanchez FG, Deisseroth A: Leukemias. In DeVita VT Jr, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology, 5th ed. Philadelphia, LippincottRaven, 1997, pp 22852321.
29. Bouvet M, Babiera GV, Termuhlen PM, et al: Splenectomy in the accelerated or blastic phase of chronic myelogenous leukemia: A single-institution, 25-year experience. Surgery 122:2025, 1997. 30. Medical
Research Councils Working Part for Therapeutic Trials in Leukaemia: Randomized trial of splenectomy in PH 1-positive chronic granulocytic leukaemia, including an
analysis of prognostic features. Br J Haematol 54:415430, 1983.

31. The
Italian Cooperative Study Group on Chronic Myeloid Leukemia: Results of a prospective randomized trial of early splenectomy in chronic myeloid leukemia. Cancer 54:333
338, 1984.
32. Gratwohl
A, Goldman J, Gluckman E, et al: Effect of splenectomy before bone marrow transplantation on survival in chronic granulocytic leukaemia. Lancet 2:12901291, 1985. L, Rosenberg J, Geller SA: Tumors of the spleen. World J Surg 9:468476, 1985.
33. Morgenstern 34. Feigenberg
Z, Wysenbeek A, Avidor E, et al: Malignant lymphangioma of the spleen. Isr J Med Sci 19:202204, 1983.
35. Sardi A, Ojeda HF, King D Jr: Laparoscopic resection of a benign true cyst of the spleen with the harmonic scalpel producing high levels of CA 199 and carcinoembryonic antigen. Am Surg 64:11491154, 1998. 36. Tsakayannis 37. Goldthorn
DE, Mitchell K, Kozakewich HP, et al: Splenic preservation in the management of splenic epidermoid cysts in children. J Pediatr Surg 30:14681470, 1995.
JF, Schwartz AD, Swift AJ, et al: Protective effect of residual splenic tissue after subtotal splenectomy. J Pediatr Surg 13:587590, 1978. RJ, Maharaj A, Ghoussoub R, et al: Partial decapsulation of splenic epithelial cysts: Studies on etiology and outcome. J Pediatr Surg 32:272274, 1997.
38. Touloukian 39. Little
JM, Deane SA: Hydatid disease. In Blumgart LH (ed): Surgery of the Liver and Biliary Tract. Edinburgh, Churchill Livingstone, 1988, pp 955966.
40. Pachter
HL, Hofstetter SR, Elkowitz A, et al: Traumatic cysts of the spleenthe role of cystectomy and splenic preservation: Experience with seven consecutive patients. J Trauma 35:430436, 1993. TR: Splenic abscess. World J Surg 9:410415, 1985. Cohen MA, Galera MJ, Ruiz M, et al: Splenic abscess. World J Surg 14:513517, 1990.
41. Gadacz 42. Alonso 43. Wolff
MJ, Bitran J, Northland RG, et al: Splenic abscesses due to Mycobacterium tuberculosis in patients with AIDS. Rev Infect Dis 13:373375, 1991. WE, Gilbertson JJ, Nelson EW: Splenic abscess: Presentation, treatment options, and results. Am J Surg 158:612614, 1989.
44. Faught 45. Keidl 46. van
CM, Chusid MJ: Splenic abscesses in childhood. Pediatr Infect Dis J 8:368373, 1989.
der Laan RT, Verbeeten B Jr, Smits NJ, et al: Computed tomography in the diagnosis and treatment of solitary splenic abscesses. J Comput Assist Tomogr 13:7174, 1989. S, Wolin DA, Herbsman H: A review of percutaneous drainage in splenic abscess. Surg Gynecol Obstet 167:211216, 1988. G, Zissin R, Apter S, et al: CT findings in congenital anomalies of the spleen. Br J Radiol 74:767772, 2001. S, Koniaris LG, Kovach SJ, et al: Torsion of a wandering spleen. Surgery 132:535536, 2002. W, Polk HC Jr: Blunt abdominal trauma: A study of relationship between diagnosis and outcome. South Med J 66:11271130, 1973. LH, Benjamin CI, Strate RG, et al: Diagnostic peritoneal lavage in blunt abdominal trauma. J Trauma 15:854859, 1975.
47. Gleich 48. Gayer
49. Sayeed 50. Ahmad 51. Engrav 52. Olsen 53. Perry 54. West 55. Root
WR, Redman HC, Hildreth DH: Quantitative peritoneal lavage in blunt abdominal trauma. Arch Surg 104:536543, 1972. JF Jr, Strate RG: Diagnostic peritoneal lavage in blunt abdominal trauma: Indications and results. Surgery 71:898901, 1972.
JG, Trunkey DD, Lim RC: Systems of trauma care: A study of two counties. Arch Surg 114:455460, 1979. HD, Hauser CW, McKinely CR: Diagnostic peritoneal lavage. Surgery 57:633637, 1965. MP, Crass RA, Jeffrey RB, et al: Computed tomography in blunt abdominal trauma. Arch Surg 117:645650, 1982.
56. Federle 57. Moore
EE, Cogbill TH, Jurkovich GJ, et al: Organ injury scaling: Spleen and liver (1994 revision). J Trauma 38:323324, 1995. M, Kofler H: Sonographic acute diagnosis in polytrauma. Aktuelle Traumatol 13:5557, 1983.
58. Aufschnaiter 59. Halbfass
HJ, Wimmer B, Hauenstein K, et al: Ultrasonic diagnosis of blunt abdominal injuries. Fortschr Med 99:16811685, 1981. KH, Wimmer B, Billmann P, et al: Sonography of blunt abdominal trauma (authors transl). Radiologe 22:106111, 1982.
60. Hauenstein 61. Tso
P, Rodriguez A, Cooper C, et al: Sonography in blunt abdominal trauma: A preliminary progress report. J Trauma 33:3944, 1992. PJ, Niezen RA, van Vugt AB, et al: Abdominal ultrasound as a reliable indicator for conclusive laparotomy in blunt abdominal trauma. J Trauma 34:2731, 1993. MA, Naf R, Amgwerd M, et al: Ultrasound in blunt abdominal and thoracic trauma. J Trauma 34:488495, 1993.
62. Bode
63. Rothlin
1707
64. Branney SW, Moore EE, Cantrill SV, et al: Ultrasound-based key clinical pathway reduces the use of hospital resources for the evaluation of blunt abdominal trauma. J Trauma 42:10861090, 1997. 65. Pimpl
W, Dapunt O, Kaindl H, et al: Incidence of septic and thromboembolic-related deaths after splenectomy in adults. Br J Surg 76:517521, 1989.
66. Rogers
FB, Baumgartner NE, Robin AP, et al: Absorbable-mesh splenorrhaphy for severe splenic injuries: Functional studies in an animal model and an additional patient series. J Trauma 31:200204, 1991. M, Courcoulas A, Gardner M, et al: Management of blunt splenic trauma: Significant differences between adults and children. Surgery 122:654660, 1997.
67. Powell
68. Davis
KA, Fabian TC, Croce MA, et al: Improved success in nonoperative management of blunt splenic injuries: Embolization of splenic artery pseudoaneurysms. J Trauma 44:10081015, 1998. JP, Chiu WC, DAmelio LF, et al: Computed tomography is inaccurate in estimating the severity of adult splenic injury. J Trauma 39:514518, 1995. CS, Moore FA, Arteaga BS: Age should not be a consideration for nonoperative management of blunt splenic injury. J Trauma 47:220, 1999.
69. Sutyak
70. Cocanour
71. Myers JG, Dent DL, Stewart RM: Nonoperative management of blunt splenic injuries: Dedicated trauma surgeons can achieve a high rate of success in patients of all ages [Abstract]. J Trauma 47:220, 1999. 72. Shackford 73. Smith
SR, Molin M: Management of splenic injuries. Surg Clin North Am 70:595620, 1990.
JS Jr, Cooney RN, Mucha P Jr: Nonoperative management of the ruptured spleen: A revalidation of criteria. Surgery 120:745750; discussion 750741, 1996. DG, Chang FC, Helmer SD: Changing trends in the management of splenic injury. Am J Surg 170:686690, 1995.
74. Morrell
75. Schurr MJ, Fabian TC, Gavant M, et al: Management of blunt splenic trauma: Computed tomographic contrast blush predicts failure of nonoperative management. J Trauma 39:507513, 1995. 76. Myers
LK, Tang B, Rosloniec EF, et al: Characterization of a peptide analog of a determinant of type II collagen that suppresses collagen-induced arthritis. J Immunol 161:3589 3595, 1998. TK, Croce MA, Miller PR, et al: Failures of splenic nonoperative management: Is the glass half empty or half full? J Trauma 50:230236, 2001.
77. Bee
78. Agency
for Health Care Policy and Research: Interim Manual for Clinical Practice Guideline Development. Rockville, MD, U.S. Department of Health and Human Services, Public Health Service, 1991. BJ, Phillips EH, Semel CJ, et al: Laparoscopic splenectomy. Surg Endosc 6:183185, 1992. B, Maignien B: Laparoscopic splenectomytechnical aspects. Surg Endosc 6:305308, 1992. C, Mamazza J, Letourneau R, et al: Laparoscopic splenectomy: Operative technique and preliminary report. Surg Laparosc Endosc 2:248253, 1992. RL, Hiatt JR, Korman JL, et al: Laparoscopic or open splenectomy for hematologic disease: Which approach is superior? J Am Coll Surg 185:4954, 1997.
79. Carroll
80. Delaitre 81. Thibault
82. Friedman
83. Gigot JF, Lengele B, Gianello P, et al: Present status of laparoscopic splenectomy for hematologic diseases: Certitudes and unresolved issues. Semin Laparosc Surg 5:147167, 1998. 84. Berchtold 85. Chirletti 86. Coon
P, McMillan R: Therapy of chronic idiopathic thrombocytopenic purpura in adults. Blood 74:23092317, 1989.
P, Cardi M, Barillari P, et al: Surgical treatment of immune thrombocytopenic purpura. World J Surg 16: 10011005, 1992.
WW: Splenectomy for idiopathic thrombocytopenic purpura. Surg Gynecol Obstet 164:225229, 1987. AA, Jones PF, King DJ: Splenectomy in the management of haematological disease. Br J Surg 74:353357, 1987.
87. Dawson 88. Eraklis 89. Musser 90. Naouri
AJ, Filler RM: Splenectomy in childhood: A review of 1413 cases. J Pediatr Surg 7:382388, 1972. G, Lazar G, Hocking W, et al: Splenectomy for hematologic disease: The UCLA experience with 306 patients. Ann Surg 200:4045, 1984. A, Feghali B, Chabal J, et al: Results of splenectomy for idiopathic thrombocytopenic purpura: Review of 72 cases. Acta Haematol 89:200203, 1993. DN Jr, Fraker DL: Splenectomy for the massively enlarged spleen. Am Surg 57:108113, 1991. J: Splenectomy for massive splenomegaly. Am J Surg 135:385388, 1978.
91. Danforth
92. Goldstone 93. Schwartz 94. Friedman 95. Legrand 96. Yee
SI, Bernard RP, Adams JT, et al: Splenectomy for hematologic disorders. Arch Surg 101:338347, 1970. RL, Fallas MJ, Carroll BJ, et al: Laparoscopic splenectomy for ITP: The gold standard. Surg Endosc 10:991995, 1996.
MJ, Honore P, Joris J, et al: Techniques of laparoscopic morcellation of the spleen. Minim Invasive Ther Allied Tech 5:143146, 1996.
JC, Akpata MO: Laparoscopic splenectomy for congenital spherocytosis with splenomegaly: A case report. Can J Surg 38:7376, 1995. JL, Lefor AT, Steers J, et al: Laparoscopic splenectomy in patients with hematologic diseases. Ann Surg 224:1928, 1996.
97. Flowers 98. Bove
T, Delvaux G, Van Eijkelenburg P, et al: Laparoscopic-assisted surgery of the spleen: Clinical experience in expanding indications. J Laparoendosc Surg 6:213217, 1996. G, Millat B, Guillon F, et al: Laparoscopic splenectomy for benign and malignant hematologic diseases35 consecutive cases. World J Surg 22:6268, 1998.
99. Decker 100. Trias 101. Trias
M, Targarona EM, Balague C: Laparoscopic splenectomyan evolving technique: A comparison between anterior and lateral approaches. Surg Endosc 10:389392, 1996. M, Targarona EM, Espert JJ, et al: Laparoscopic surgery for splenic disorders: Lessons learned from a series of 64 cases. Surg Endosc 12:6672, 1998. B, Pitre J: Laparoscopic splenectomy versus open splenectomy: A comparative study. Hepatogastroenterology 44:4549, 1997. RE, Yee LF, Mulvihill SJ: Laparoscopic splenectomy: The emerging standard. Surg Endosc 11:108112, 1997. N, Mavor E: Laparoscopic splenectomy. Surg Clin North Am 80:12851297, 2000.
102. Delaitre
103. Glasgow
104. Katkhouda 105. Moise
KJ Jr: Autoimmune thrombocytopenic purpura in pregnancy. Clin Obstet Gynecol 34:5163, 1991.
106. Romanelli 107. Phillips 108. Park
JR, Kelly JJ, Litwin DE: Hand-assisted laparoscopic surgery in the United States: An overview. Semin Laparosc Surg 8:96103, 2001.
EH, Carroll BJ, Rosenthal RJ: Laparoscopic splenectomy. In Cameron J (ed): Current Surgical Therapy. St. Louis, CV Mosby, 1995.
A, Gagner M, Pomp A: The lateral approach to laparoscopic splenectomy. Am J Surg 173:126130, 1997.
1708
109. Michels 110. Gordon
NA: The variational anatomy of the spleen and the splenic artery. Am J Anat 70:2172, 1942.
DH, Schaffner D, Bennett JM, et al: Postsplenectomy thrombocytosis: Its association with mesenteric, portal, and/or renal vein thrombosis in patients with myeloproliferative disorders. Arch Surg 113:713715, 1978. J, Smith JL, Weber TK, et al: Postoperative complications after splenectomy for hematologic malignancies. Ann Surg 223:290296, 1996.
111. Horowitz 112. Styrt
B: Infection associated with asplenia: Risks, mechanisms, and prevention. Am J Med 88:33N-42N, 1990. K, Singer DB: Asplenic-hyposplenic overwhelming sepsis: Postsplenectomy sepsis revisited. Pediatr Dev Pathol 4:105121, 2001. GL, Watkins DN, Watts AD, et al: Severe late postsplenectomy infection. Br J Surg 78:716721, 1991.
113. Hansen
114. Cullingford
115. Advisory Committee on Immunization Practices: Prevention of pneumococcal disease: Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Morb Mortal Wkly Rep 46:124, 1997. 116. Caplan ES, Boltansky H, Snyder MJ, et al: Response of traumatized splenectomized patients to immediate vaccination with polyvalent pneumococcal vaccine. J Trauma 23:801 805, 1983. 117. Rutherford 118. Cadiere
EJ, Livengood J, Higginbotham M, et al: Efficacy and safety of pneumococcal revaccination after splenectomy for trauma. J Trauma 39:448452, 1995.
GB, Verroken R, Himpens J, et al: Operative strategy in laparoscopic splenectomy. J Am Coll Surg 179:668672, 1994. A, Zornig C, Peiper M, et al: Laparoscopic splenectomy: Technique and results in a series of 27 cases. Surg Endosc 9:924927, 1995.
119. Emmermann 120. Poulin 121. Brunt 122. Gigot 123. Smith
EC, Thibault C, Mamazza J: Laparoscopic splenectomy (see Discussion). Surg Endosc 9:172177, 1995.
LM, Langer JC, Quasebarth MA, et al: Comparative analysis of laparoscopic versus open splenectomy. Am J Surg 172:596601, 1996. JF, de Ville de Goyet J, Van Beers BE, et al: Laparoscopic splenectomy in adults and children: Experience with 31 patients. Surgery 119:384389, 1996.
CD, Meyer TA, Goretsky MJ, et al: Laparoscopic splenectomy by the lateral approach: A safe and effective alternative to open splenectomy for hematologic diseases. Surgery 120:789794, 1996.
124. Tsiotos 125. DiFino 126. Mintz
G, Schlinkert RT: Laparoscopic splenectomy for immune thrombocytopenic purpura. Arch Surg 132:642646, 1997. SM, Lachant NA, Kirshner JJ, et al: Adult idiopathic thrombocytopenic purpura: Clinical findings and response to therapy. Am J Med 69:430442, 1980.
SJ, Petersen SR, Cheson B, et al: Splenectomy for immune thrombocytopenic purpura. Arch Surg 116:645650, 1981. P, Wood L, Dent DM: Results of treatment in immune thrombocytopenia. Q J Med 58:153165, 1986.
127. Jacobs
OE, Itani KM, Coleman RE, et al: Splenectomy for primary and recurrent immune thrombocytopenic purpura (ITP): Current criteria for patient selection and results. Ann Surg 206:529541, 1987.
129. Julia
128. Akwari
A, Araguas C, Rossello J, et al: Lack of useful clinical predictors of response to splenectomy in patients with chronic idiopathic thrombocytopenic purpura. Br J Haematol 76:250255, 1990.
130. Yee
LF, Carvajal SH, de Lorimier AA, et al: Laparoscopic splenectomy: The initial experience at University of California, San Francisco (see Discussion). Arch Surg 130:874 879, 1995. M, Rudd M, ORourke N, et al: Laparoscopic splenectomy and lymph node biopsy for hematologic disorders. Ann Surg 222:4346, 1995. DI, Coventry BJ, Chin T, et al: Laparoscopic versus open splenectomy for immune thrombocytopenic purpura. Surgery 121:1822, 1997.
131. Rhodes 132. Watson 133. Diaz
J, Eisenstat M, Chung R: A case-controlled study of laparoscopic splenectomy. Am J Surg 173:348350, 1997.
134. Hashizume M, Ohta M, Kishihara F, et al: Laparoscopic splenectomy for idiopathic thrombocytopenic purpura: Comparison of laparoscopic surgery and conventional open surgery. Surg Laparosc Endosc 6:129135, 1996. 135. Liew
SC, Storey DW: Laparoscopic splenectomy. Aust N Z J Surg 65:743745, 1995. S, Bresler L, Tortuyaux JM, et al: Splenectomy under celioscopy. Ann Chir 49:477481, 1995. SP, Martin IG, Alao D, et al: Laparoscopic splenectomy: The suspended pedicle technique. Surg Endosc 10:393396, 1996.
136. Parent 137. Dexter 138. Zamir
O, Szold A, Matzner Y, et al: Laparoscopic splenectomy for immune thrombocytopenic purpura. J Laparoendosc Surg 6:301304, 1996. N, Waldrop DJ, Feinstein D, et al: Unresolved issues in laparoscopic splenectomy. Am J Surg 172:585590, 1996.
139. Katkhouda 140. Lee
WJ, Kim BR: Laparoscopic splenectomy for chronic idiopathic thrombocytopenic purpura. Surg Laparosc Endosc 7:209212, 1997.
www.mdconsult.com

Chapter 53 - Exocrine Pancreas: Michael L. Steer M.D

Uploaded by

Document Information

Original Description:

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

Chapter 53 - Exocrine Pancreas: Michael L. Steer M.D

Uploaded by

Copyright:

Available Formats

Use of this content is subject to the Terms and Conditions of the MD Consult web site.

Townsend: Sabiston Textbook of Surgery, 17th ed., Copyright 2004 Elsevier