Clin Orthop Relat Res (2008) 466:1017–1019
DOI 10.1007/s11999-008-0220-9
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Surgical and Molecular Advances in Osteonecrosis
Editorial Comment
Quanjun Cui MD, Khaled J. Saleh MD, MSc, FRCSC
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Osteonecrosis (ON) continues to be a considerable chal-
lenge to orthopaedic surgeons and researchers. There has
been no major breakthrough in research of this devastating
disease at the basic science and clinical level. However, the
recent explosion of molecular technology, advancement of
biomaterials, state-of-the-art computing and diagnostic
tools, and better defined surgical techniques has increased
our contemporary knowledge base. This symposium con-
tains peer reviewed scientific manuscripts selected from the
work of international experts across the globe, representing
current developments in etiology, pathogenesis, diagnosis
and treatment of ON.
The pathogenesis of ON still is unknown. To some
extent, this is because there is no ideal animal model that
can reproduce the human form of ON. Chemically induced
and traumatic bipedal animal models have shown promise
but need further study and validation. Quadruped animal
models only showed marrow and bone cell death but no
collapse consistent with early stage ON of the human.
Many hypotheses regarding pathogenesis of ON have been
proposed, including intravascular coagulation, fat emboli,
fat cell hypertrophy, intraosseous hypertension, micro-
fractures, and vasculitis, however none of these are
conclusive. Due to the lack of understanding of patho-
genesis of the disease, many therapeutic options including
electromagnetic field are developed and evaluated. Further
studies are needed for their effectiveness and validation.
Genetic factors may play an important role in the
development of ON. A large body of work on inherited
Q. Cui, K. J. Saleh (&)
Department of Orthopaedic Surgery, University of Virginia
School of Medicine, P.O. Box 800159, Charlottesville, VA
22908-0159, USA
e-mail: saleh@virginia.edu
thrombophilia and hypofibrinolysis by Dr. Glueck and his
colleagues has indicated that patients’ genetic characteris-
tics do associate with ON. Similar findings were also
reported in the Korean population, although the genetic risk
profile of ON may differ among ethnic groups.
Adipogenesis by bone marrow stem cells also may be
critical for the development of ON. As reported by Wang
et al. and Yeh et al., both glucocorticoid steroids and
alcohol can enhance adipogenesis while inhibiting osteo-
genesis. In addition to changes in transcription factors,
Wnt/ß-catenin signaling pathway was also involved in this
process. Previous studies and reports by Wang et al. and
Nishida et al. in this symposium have demonstrated that
lipid lowering agents have therapeutic effects in the treat-
ment of ON in human and animal models.
Although many treatment options for ON are available,
total hip replacement (THA) is the most common for
postcollapse stages of the disease. Historically, THA had
poor outcomes in patients with ON when compared to
those with osteoarthritis (OA). However, recent studies
have demonstrated improved outcomes using the latest
surgical techniques and implant designs. Nevertheless,
there is still a subset of patients that have less than favor-
able results when compared to OA.
Dorr and colleagues reported their midterm outcome
study using a metal-on-metal bearing surface and found
that patients with ON (27 patients, 30 hips) or OA (80
patients, 82 hips) were similar in clinical and radiographic
performance at a mean followup of 5.5 years (range 2.2–
11.7 years). In another study by Koo et al., survivorship
analysis has revealed a survival rate of 93.3% in 36 patients
(46 hips) with a minimum followup of 10 years. However,
long term outcome studies in patients with ON regarding
prosthetic hip surgery, including hemi- and total hip
resurfacing arthroplasty, are still needed.
123
1018 Cui and Saleh
Treatment of patients with early stage ON continues to
be controversial. Core decompression is the most com-
monly applied form of precollapse treatment, but the
results are not variable. Free fibular graft, rotational os-
teotomies, and nonsurgical managements are less
commonly used. Osteotomies are primarily used in Asian
countries and continue to produce favorable results. Dr.
Sugioka and his associate reported their long term results
of transtrochanteric posterior rotational osteotomy for
ARCO (Association Research Circulation Osseous) stage
III and IV ON; Dr. Yoon and colleagues of Korea used a
modified transtrochanteric rotational osteotomy for ARCO
stage II and III disease with a short term followup. Their
results indicate osteotomy is effective not only for early
stage but also for advanced stages of ON. Free fibular
transfer also has generated relatively good outcomes in a
skillful surgeon’s hands, but its indications are limited
due to technical difficulties. In a study by Yoo et al., 110
patients (124 hips) had received free vascularized fibular
graft and were followed for a minimum 10 years (range
10 to 23.7 years). Thirteen hips (10.5%) failed and con-
verted to total hip arthroplasty. The remaining patients
had a mean Harris Hip Score of 88 at final followup.
Aldridge and Urbaniak reviewed 154 patients of various
subgroups treated with free vascularized fibular graft with
a minimum followup one year (range 1 to 19 years) and
found it highly successful in treating athletic or pregnant
patients and patients with pyarthritis or slipped capital
femoral epiphysis. A high failure rate was noted in
patients with organ transplant. It is generally accepted
patient selection is the key for a successful joint pre-
serving procedure. The size of the lesion, extent of
femoral head depression, and involvement of the acetab-
ulum are important parameters dictating outcomes of the
treatments.
Modern core decompression techniques, including
multiple small diameter drilling, in conjunction with use of
bone marrow stem cells, growth factors such as bone
morphogenetic proteins, and nonvascularized bone grafts
have produced improved outcomes in limited studies with
short to midterm followups. The major limitations of the
procedure include incomplete removal of necrotic bone and
lack of a mechanism to prevent subchondral collapse.
Endoscope-aided core decompressions and sophisticated
grafting techniques may surmount these shortcomings.
However, data from randomized, double-blinded, pro-
spective clinical trials evaluating modern core
decompression techniques, different bone grafts, and graft
substitutes are needed.
Many classifications for ON of the femoral head have
been used in the past. Ficat and Arlet, Steinberg (the
University of Pennsylvania), ARCO and Japanese Ortho-
paedic Association classifications are the most commonly
123
Clinical Orthopaedics and Related Research
used systems. Although there is no consensus supporting
the use of any one classification system, the Steinberg and
ARCO classification systems are comprehensive and
incorporate the most sensitive diagnostic techniques such
as magnetic resonance imaging and other important
parameters that allow for a uniform data collection and
cross-study comparison. We encourage use of these two
classification systems in reporting studies on ON of the
femoral head. Otherwise, the key parameters including size
of the lesion, presence or absence of collapse, amount of
depression of the femoral head, acetabular involvement,
and the location of the lesion should be collected and
reported. As further information becomes available, the
classification system will be further defined and one uni-
versal classification system should be used.
Magnetic resonance imaging (MRI) continues to be the
most sensitive and accurate diagnostic tool for ON of the
femoral head. Size and extent of the necrotic lesion, an
important prognostic parameter, can be better defined on
MRI than on radiographs. Measurement should be per-
formed on both coronal and sagittal planes. Min et al.
prospectively monitored 81 patients (81 hips) with
asymptomatic ON of the femoral head. The diagnosis
was established by MRI. Patients were followed for
8.3 years (range 5 to 16 years). Twenty-six hips (32%)
progressed to collapse. They found no correlation
between the collapse and patients’ age, gender, weight,
causes of the disease, and the length of followup. How-
ever, the size and location of the lesion were the factors
that predicted collapse. The concept also can be applied
to shoulder ON. Sakai and colleagues have found the
extent of the lesion measured as necrotic angle on mid-
oblique coronal and sagittal planes of the humeral head
can predict collapse of the humeral head. Newer tech-
nologies are evolving. As reported by Dasa et al., F-18
positron emission tomography (PET) seemed more sen-
sitive detecting early changes in the acetabulum than
traditional imaging systems but it is uncertain whether
and how these changes relate to the pathophysiology of
ON. But further studies are needed to validate the use of
newer diagnostic tools.
Challenges still lie ahead. We hope in the next few years
there will be more studies on the development of animal
models that can be used to evaluate various etiologies and
treatment options. Knowing the predisposition factors,
capability for early detection of the disease and available
effective therapeutic agents may prevent the disease pro-
cess in the future. With further defined surgical techniques
and availability of osteogenic and angiogenic growth fac-
tors, early stage ON will hopefully become treatable. This
approach can optimally decrease or eliminate the need for
total joint replacement for this young, productive patient
population.
Volume 466, Number 5, May 2008 Surgical and Molecular Advances in Osteonecrosis 1019

We conclude by thanking all the authors for taking time our sincere appreciation to Dr. Richard A. Brand, Editor in
to contribute to this special symposium on Surgical and Chief of Clinical Orthopaedics and Related Research, and
Molecular Advances in ON. We also would like to express his editorial staff for their support for this symposium.

123
Clin Orthop Relat Res (2008) 466:1020–1033
DOI 10.1007/s11999-008-0206-7

SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS

The Classic
Repair of Bone in the Presence of Aseptic Necrosis Resulting from Fractures,
Transplantations, and Vascular Obstruction*

Dallas Burton Phemister MD

Published online: 27 March 2008

Dallas Burton Phemister was born in 1882 in Carbondale,

Illinois [5]. After beginning a secondary education at the
Normal School of Northern Indiana, he studied medicine at
the Rush Medical College of the University of Chicago.
(Rush Medical College was chartered in 1837, by Dr. Daniel
Brainard, originally from New York and educated in
Philadelphia [1]. When Brainard moved to the Midwest, he
named his new proprietary medical college after the famous
Philadelphia physician, Benjamin Rush, who is considered a
founding father of American medicine. At the turn of the
century the College became affiliated with the University of
Chicago, itself founded in 1890. That affiliation lasted until
WWII, at which time Rush Medical College affiliated with
the University of Illinois’ College of Medicine. Then in 1956
the College accepted the invitation to merge with St. Luke’s
Hospital at which time they were no longer affiliated with
the University of Illinois.) Phemister graduated from Rush in
1904 and interned at Cook County Hospital. Orthopaedic
surgery was not well developed in this country, and as did so
many of his peers at the time, he went abroad for advanced
training after five years of practice. He spent two years
traveling, mostly spending time in Paris, Berlin, and Vienna
[8]. In Vienna, one of the world’s leading centers of ortho-
paedics at the time [6], he encountered the famous
pathologist Jakob Erdheim and undoubtedly there devel-
oped his lifelong interest in bone pathology.
Phemister returned to Chicago in 1911 to resume a
teaching position at Rush Medical College. During WW I he
served with the Presbyterian Hospital Unit in France. (Many
individual hospitals in the US organized such units to support
the war effort since the medical support of the US military
Dr. Dallas Burton Phemister is shown. Figure is Ó1952 by the
Journal of Bone and Joint Surgery, Inc. and is reprinted with
permission from Dallas Burton Phemister 1882–1951. J Bone Joint
Surg Am. 1952;34:746–747.
was insufficient at the time.) He remained at Rush until the
University of Chicago established a new medical school in
1927, and asked him to organize a department of surgery.
During his very productive years at the University of
Chicago he published many dozens of paper. Perhaps more
importantly, his influence was felt through his training of
Dr. Howard Hatcher, who in turn trained a generation of
orthopaedic surgeons with a primary interest in bone tumors
and pathology: Drs. Wayne H. Akeson, Michael Bonfiglio,
Thomas D. Brower, Crawford J. Campbell, William K.
Enneking, H. Relton McCarroll, James S. Miles, Eugene R.
Mindell, John A. Siegling, Mary Sherman.
The paper we reproduce here described in detail the
repair of osteonecrotic bone. Axhausen was perhaps the
first to systematically examine the repair of dead bone [3],

123
Volume 466, Number 5, May 2008
and described the process of ‘‘simultaneous absorption of
dead bone and incomplete, irregular replacement by new
bone.’’ [7] Phemister extended those observations and
coined the term, ‘‘creeping substitution.’’ ‘‘By the process
of creeping substitution the old bone is gradually absorbed
and replaced by new bone, so that in the course of months
or occasionally years the necrotic area is more or less
completely transformed into living bone ... the amount of
new bone formed ... depends largely on the extent of the
living bone with which it (the dead bone) is in contact.’’ At
the time, necrosis was believed mostly related to fractures,
transplants, or infection (septic necrosis). Axhausen had
earlier described aseptic necrosis [2, 3] but the cause and
process was not well understood. Phemister commented,
‘‘Aseptic necrosis of bone, the result of proven thrombosis
or embolism, is a condition that has been rarely reported in
man.’ He did not mention what we would today consider
idiopathic aseptic necrosis or osteonecrosis. (The term,
‘‘avascular necrosis’’ first appears in PubMed in 1949 in
relation to fractures [4]. It has been commonly used, but as
with ‘‘aseptic necrosis’’ appears to be diminishing in use
compared to ‘‘osteonecrosis.’’ In the past ten years, aseptic
necrosis appears in titles in PubMed only 21 times, avas-
cular necrosis 378, and osteonecrosis 1110. We use
osteonecrosis in this symposium because it is general and
avoids implications regarding etiology, which remains
unclear.) At the time Phemister was writing, steroids were
unknown, as was necrosis related to deep sea diving or
other causes known today. Nonetheless, he well described
the histologic processes we know today.
Richard A. Brand MD
Aseptic necrosis of bone may result from a number of
causes,—as vascular disturbances, injuries, the action
of chemicals, and the application of radium in the treatment
of malignant disease.
The process of repair of the damaged area differs with
the causative agent and with the amount of functional
stimulation to which the part is subjected during the
reparative period. It also differs from that seen in case of
necrosis produced by infection. Necrosis en masse produced
by severe infection is nearly always followed by a fibro-
blastic and fixed tissue phagocytic reaction which usually
results in complete absorption of the dead bone, if the area is
small, or in sequestration, if it is large. However, mild
inflammatory reactions, particularly when associated with
embolic or thrombotic processes and due to low grade
micro-organisms, may sometimes result in necrosis which
is followed by simultaneous absorption of dead bone and
incomplete, irregular replacement by new bone, as first
pointed out by Axhausen [1]. This same change may also be
Repair of Bone in Aseptic Necrosis 1021
References
1. About Rush Medical College. Rush University Web site. Available
at: http://www.rushu.rush.edu/medcol/history.html. Accessed
February 14, 2008.
2. Axhausen G. Über einfache, aseptische Knochen-und
Knorpelnekrose, Chondritis dissecans und Arthritis deformans.
Arch Klin Chir. 1912.
3. Axhausen G. Knochennekrose und Squesterbildung. Dtsch Med
Wochenschr. 1914;40:111–115.
4. Compere EL. Avascular necrosis of large segmental fracture
fragments of the long bones. J Bone Joint Surg Am. 1949;31:47–
54.
5. Dallas Burton Phemister, 1882–1951. J Bone Joint Surg Am.
1952;24:746–747.
6. Kotz R, Engel A, Schiller C, ed. 100 Jahre Orthopädie an der
Universität Wien. Vienna, Austria: Verlag der Wiener Medizi-
nischen Akademie; 1987.
7. Phemister DB. Repair of bone in the presence of aseptic necrosis
resulting from fractures, transplantations, and vascular obstruc-
tion. J Bone Joint Surg Am. 1930;12:769–787.
8. Phemister DB. Treatment of the necrotic head of the femur in
adults. Dallas Burton Phemister (1882–1951). Clin Orthop Relat
Res. 2000;381:4–8.
Ó The Association of Bone and Joint Surgeons 2008
This article is Ó1930 by the Journal of Bone and Joint Surgery, Inc.
and is reprinted with permission from Phemister DB. Repair of bone
in the presence of aseptic necrosis resulting from fractures,
transplantations, and vascular obstruction. J Bone Joint Surg Am.
1930;12:769–787.
*Read before the American Orthopaedic Association, Chatham,
Massachusetts, June 20, 1930.
Richard A. Brand MD &
Clinical Orthopaedics and Related Research,
1600 Spruce Street, Philadelphia,
PA 19103, USA
e-mail: dick.brand@clinorthop.org
observed in a bone graft when mild infection has occurred.
It is not uncommon to see a portion of a graft, which has
been exposed by opening of the wound with the escape of
exudate, heal in and subsequently undergo transformation
without sequestration. Epiphysitis may result in rather
extensive necrosis which is followed by absorption and a
certain amount of collapse and new bone formation with
eventual bony reconstruction of a deformed epiphysis.
The introduction of radium into or in contact with bone in
the treatment of malignant disease may lead to more or less
extensive bone necrosis without destruction of the peripheral
layer of the overlying soft parts, in which case the necrotic
bone remains free from infection. The necrotic bone produced
in this way may be sequestrated at a very slow rate or, if it
performs a supporting function, it may be very slowly
absorbed and replaced by new bone. If infection is present it is
always slowly sequestrated. The explanation of the slow
absorption is to be found in the fact that the tissues adjacent to
the dead bone are radium burnt and consequently bring about
123
1022 Phemister
repair very slowly. I implanted radium needles into an
undifferentiated round-cell sarcoma of the ischium, producing
aseptic bone necrosis, and after subsequent x-ray treatments
saw healing of the sarcoma with repair of the bony defect
without evidences of sequestration of the necrotic bone after
the lapse of seven years. On the other hand in a case of
undifferentiated round-cell sarcoma of the lower arm treated
by excision, radium implantation in contact with humerus, and
subsequent roentgen therapy, there was very slight infection in
the field. In this case a large portion of the cortex which was
killed by radium was very slowly sequestrated in the course of
six years and at the end of eight years was still present as a
loose piece which had worked out into the soft parts.
In experimental radium necrosis [2] the entire circum-
ference of the shaft of the femur of the dog was killed by the
insertion of radium into the medullary cavity downward
from the trochanter. The dead bone was very slowly
replaced but its fate was dependent to a considerable extent
on the amount of function which the limb performed. Some
animals walked on the limb for months without the occur-
rence of a fracture and without sequestration of the dead
bone. There was very slow creeping substitution of dead
bone by new bone from the two ends. After the lapse of
months fracture sometimes occurred through the dead bone.
If the fracture resulted in non-union so that the extremity
was no longer used, there was gradual sequestration of the
dead bone in the course of one to two years. If there was
overriding of fragments and if portions of the two fragments
that were undamaged by radium approximated each other,
new bone was formed and the fracture united. If a portion of
the necrotic bone was caught in the line of the new bone
formed by unburnt tissues, it was rapidly absorbed and
replaced by new bone. However, if the necrotic bone lay
outside the callus and the line of stress and strain, it was
very slowly absorbed without replacement by new bone.
Bancroft [3] produced chemical necrosis of bone
experimentally in dogs by the introduction of croton oil
into the medullary cavity. The inflammatory reaction set up
by the oil resulted in a marked absorptive and osteoplastic
process with sequestration and formation of an involucrum.
Owing to the rapid disappearance of the inflammatory
reaction, the reparative changes proceeded more rapidly
than in case of necrosis produced by infection and in the
course of a few weeks there was complete absorption of the
dead bone with reformation of a new shaft.
Interference with Circulation
The fate of bone which undergoes aseptic necrosis as a
result of circulatory disturbance varies considerably,
according to the environment of the necrotic area. If the
necrotic bone is attached to and directly continuous with
123
Clinical Orthopaedics and Related Research
living bone, there is ingrowth of vessels and of fibrous and
osteogenic tissues from the living into the necrotic area.
There is usually survival of unossified osteogenic elements
about the periphery and ends of the area that has had its
circulation interrupted. By the process of creeping substi-
tution the old bone is gradually absorbed and replaced by
new bone, so that in the course of months or occasionally
years the necrotic area is more or less completely trans-
formed into a living one. Whether the amount of new bone
formed is as great as the amount of bone which died
depends largely on the extent of living bone with which it
is in contact and the amount of osteogenic elements sur-
viving about its periphery. When these are both extensive,
there may be complete replacement of the dead bone by
new bone; but when limited, bone absorption may be
greatly in excess of bone formation, so that incomplete
replacement results with the formation of bone that is less
dense than the original and that shows varying degrees of
cavitation. Functional stimulation also plays an important
role in the rate and degree of transformation. In general the
greater the degree of function, the greater the rate and
degree of transformation into new bone. If the necrotic
bone is largely or wholly intra-articular and extensively
bordered by articular cartilage, the attachment to the sur-
rounding soft parts will be markedly limited and chances
for revascularization and transformation will be greatly
reduced. Thus, injury to vessels may be the cause of some
of the cases of Kienböck’s malacia of the carpal lunatum
and of Legg-Perthes’ disease,—as when it follows reduc-
tion of congenital dislocation of the hip.
Operative Interference with the Blood Supply
Operations which involve severence of bone or extensive
separation of soft parts from bone may occasionally
interfere with blood supply sufficiently to cause massive
necrosis. An example of this is the necrosis of the head and
neck of the femur in the following case in which arthro-
plasty of the hip was performed.
CASE 1. Male, age twenty-eight, entered the Univer-
sity of Chicago Clinics May 7, 1929, because of bony
ankylosis of the left hip resulting from acute hematoge-
nous pyogenic arthritis two and one-half years previously
(Fig. 1). At operation on May 9, through a goblet inci-
sion the greater trochanter was reflected, the soft parts
reflected from the neck, and the head detached from the
acetabulum with a chisel. After roughly rounding it off
and deepening the acetabulum, a cap of free fascia lata
was placed over the head and tied about the neck with a
purse-string suture of chromic catgut. The head was then
restored to acetabulum, the muscles and greater tro-
chanter were sutured in place, the wound closed, and a
Volume 466, Number 5, May 2008 Repair of Bone in Aseptic Necrosis 1023

Fig. 1 Case 1. Bony ankylosis following pyogenic arthritis. Fig. 3 Case 1. Thirty-six days after arthroplasty. Beginning reduction
in density in base of neck and shaft. No change in head.

Fig. 2 Case 1. Twenty days after arthroplasty. Fig. 4 Case 1. One hundred and fifty-one days after operation.
Density of necrotic head unchanged but shaft and pelvis show atrophy
of disuse.

123
1024 Phemister
body and leg cast applied. There was slight postoperative
febrile reaction, but the temperature returned to normal at
the end of ten days and the wound healed with only a
slight superficial serous discharge. The cast was removed
after two weeks and weight extension applied for ten
days. A roentgenogram taken twenty days postoperatively
(Fig. 2) showed the reduced head of the femur with its
sharply outlined chisel cuts resting in the acetabulum.
Movement was begun and after five weeks the patient
was allowed to walk with crutches. In a roentgenogram
taken thirty-six days postoperatively (Fig. 3) the head and
neck showed their outline and density unchanged, but
there was a slight reduction in density in hone at the
junction of neck and shaft and slight atrophy of disuse of
the shaft. A fair range of mobility persisted in the hip but
it remained painful so that the patient continued to walk
on crutches. A roentgenogram taken 151 days postoper-
atively (Fig. 4) showed the head retaining its original
density and sharp outline of its cut surface. The adjacent
innominate bone and the trochanters and shaft of the
femur showed marked atrophy of disuse. The zone of
reduced density at the base of the neck had broadened to
include almost the entire neck of the femur. A diagnosis
was made of necrosis of the head and neck of the femur
with retention of the original density of the head and
atrophy of disuse of the surrounding living bone. The
extension of the rarefaction in the neck was interpreted
Fig. 5 Case 1. Two hundred and fifty days after operation. Head
extensively reduced in density from invasion and replacement from
the living bone of base of neck.
123
Clinical Orthopaedics and Related Research
as evidence of beginning invasion of the necrotic portion
by tissue from the living portion with absorption of the
dead bone and replacement by a less dense living bone.
The hip improved slowly and the patient began to bear a
small amount of weight on it. A roentgenogram 250 days
postoperative (Fig. 5) showed persistence of the atrophy
of the shaft and innominate bone and a very marked
reduction in the density of the shadow cast by the head
as compared with the previous roentgenograms. The
cortex at the inferomesial and superolateral portions of
the base of the head retained its old density but the
cancellous bone of the interior and articular surfaces of
the head had been reduced to practically the same density
as the trochanters and upper end of the shaft. This was
interpreted as evidence of extensive invasion of the head
with absorption of the dead bone and a certain amount of
replacement by new bone. The patient then made greater
use of the limb in weight-bearing but continued to assist it
with a crutch. A roentgenogram taken 350 days after
operation showed slight increase in density of all of the
atrophic living bone in the region and slight further reduc-
tion in the dense areas of the head. The patient then walked
with the assistance of a cane, and a roentgenogram taken
404 days postoperatively (Fig. 6) showed little change
except for evidence of cavitation in the inferomesial part of
the head, indicating absorption of the necrotic bone with
failure of replacement by new bone.
Volume 466, Number 5, May 2008
Fig. 6 Case 1. Four hundred and four days
postoperative Cavitation at base of head and further
transformation.
The hip showed about forty degrees of flexion but
marked limitation of abduction and rotation. Its strength
had gradually improved and there was only slight dis-
comfort in walking unless it was prolonged.
The cause for this exceptional occurrence of necrosis of
head and neck without sequestration appeared to lie in the
traumatism of head and denudation of neck of its covering
of soft parts. However, it may have been due in part to
deep-seated mild infection in the region of the new joint,
despite the absence of discharge from that region. Failure
of the necrotic head to be sequestrated speaks decidedly in
favor of aseptic instead of septic necrosis although
sequestration may not occur in the presence of mild
infection of a necrotic area. In another case I observed
necrosis of the head and neck following an arthroplasty in
which there was infection of the wound with a purulent
discharge, but the dead bone was sequestrated in the course
of several weeks. In the case here reported functional
stimulation of the bones resulting from movement of the
joint, combined with protection of the necrotic head from
weight-bearing, was followed by creeping substitution of
the necrotic bone by new bone in the greater part of the
head without either erosion of dead bone or the develop-
ment of marked deformity.
Extensive separation of periosteum from bone, particu-
larly during the growing period, with resultant injury of
Repair of Bone in Aseptic Necrosis 1025
nutrient and metaphyseal vessels may result in extensive
aseptic necrosis of bone. Brunschwig of the University of
Chicago has produced areas of necrosis in the femur and
tibia of young dogs varying up to nearly the entire extent of
the shaft. There is absence of sequestration, and creeping
substitution of the necrotic bone takes place by new bone,
formed from the adjacent living bone and the surviving
osteogenic elements of the separated bone and periosteum
with only slight changes in size, shape, and density. Such
bone necrosis is no doubt occasionally produced by
extensively denuding operations, but it remains undetected
as the damaged area is repaired without appreciable alter-
ation in the clinical course of the disease.
Necrosis in Fractures
Fracture may cause necrosis by interference with the blood
supply of the bone in the immediate vicinity of the break or
by damage to larger vessels that furnish a large part or all
of the nutrition to one of the fragments.
Cornil and Coudray [5] found that in experimental
fractures there is necrosis of the cortex for a variable dis-
tance back from the fragment ends which is gradually
replaced by new bone ingrowing from the living cortex
with which it is continuous and to a less extent from the
123
1026 Phemister
adjacent callus. This occurrence does not interfere with the
process of healing. In fact, it has even been claimed that the
presence of the necrotic bone is a stimulus to bone repair.
Non-union and neck absorption in fractures of the femoral
neck have probably been wrongly attributed to the mar-
ginal necrosis of fragment ends. Clinically I have not seen
evidence of any appreciable amount of bone necrosis
extending back into the fragment ends from the fracture
line, either at operation or in roentgenograms. A narrow
strip of fragment end may be bare and white when a simple
fracture of some standing is exposed at operation; but I
have never seen the fragment ends appear appreciably
denser than the cortex some distance back from the fracture
as shown by roentgenograms of fractures, either united or
ununited, of several weeks’ standing, where the regional
bone had undergone atrophy of disuse. The indications are
that such bone as becomes necrotic at the ends of vascu-
larized fragments is rapidly revascularized by ingrowth
from the living bone and is absorbed and reduced in density
about as rapidly as is the adjacent living bone from atrophy
of disuse. In histological sections of resected ends of frag-
ments from cases of delayed union or nonunion, I have not
seen extensive amounts of dead bone and that which I have
found showed signs of creeping substitution by new bone.
In the case of fractures followed by bony union, mar-
ginal necrotic bone may rarely be absorbed more rapidly
than new bone is deposited, thereby creating a zone of
reduced density in the region of the fragment ends. It is
seen in pathological fractures of osteomalacia, osteogenesis
imperfecta, and in bone diseases associated with parathy-
roid adenoma and is known in the German literature as
Losser’s Umbauzone [6] (zone of transformation). I saw it
once in a patient who had no general bone disturbance but a
refracture of the ulna one year after the original fracture
had been successfully treated by an autogenous inlay bone
graft. Following the refracture a second bone graft was
inserted in the accompanying fractured radius which had
failed to unite following the first transplantation. Three and
onehalf months afterwards the radius was united, but a
zone of reduced density had developed along the oblique
fracture line in the ulna although no motion could be
elicited. Six and a half months later the zone was slightly
narrowed and increased in density and eleven months later
it had fully ossified. This zone apparently resulted from
absorption of marginal necrotic bone produced by the
fracture with marked delay in replacement by new bone for
some unexplained reason.
Bone Necrosis in Joint Fractures
Necrosis of bone may also occur in fractures bordering on
joints from injury to the blood vessels to a large part or all of
123
Clinical Orthopaedics and Related Research
the joint fragment. It undergoes either partial or total
necrosis. The condition may be met with in intracapsular
fractures of the neck of the femur, of the capitellum humeri,
of the carpal navicular, of the head of the radius and of the
neck of the astragalus, and it always follows complete
detachment of small fragments in intra-articular fractures.
Necrosis of the proximal fragment in complete intra-
capsular fracture of the neck of the femur occurs in a not
inconsiderable percentage of the cases. Santos [7] has
recently completed a detailed study which has thrown light
on many aspects of the condition. A summary of his report
follows: Necrosis may occur in fractures at any age, and
although relatively more frequent in older persons, it is often
absent in them. This may sometimes be because of blood
reaching it by way of vessels in the untorn periosteum.
The vessels of the ligamentum teres may supply adequate
blood to keep the head alive even in the aged.
Death of the proximal fragment predisposes to non-
union, but union occurs in a considerable percentage of the
cases with a dead head in which there is impaction of
fragments or in which there is efficient reduction and fixa-
tion of fragments, the best results apparently being obtained
from the closed abduction method of Whitman. At the end
of the period of immobilization the distal living fragment
usually shows roentgen-ray signs of reduced density due to
atrophy of disuse, but the necrotic head which cannot
atrophy casts practically as even and heavy a shadow as at
the time of fracture. This usually makes it possible to
diagnose necrosis roentgenologically at this stage.
If bony union occurs the head is gradually invaded by
the ingrowth of tissue from the distal fragment and some-
times from the hypertrophied round ligament. The necrotic
marrow is replaced by vascular connective tissue and
eventually by bone marrow, but the necrotic bone is much
more slowly absorbed and replaced by a variable amount of
irregularly arranged living bone. The necrotic cartilage is
slowly absorbed and replaced by fibrous tissue, fibro-
cartilage, and to some extent eventually by new hyaline
cartilage. If too much weight is borne on the head before it
has been transformed into new bone, its weight-bearing
portion may collapse with a resultant deformity and a poor
functional result, but this may be averted by prolonged
avoidance of weight-bearing. The functional stimulation
which comes from motion of the limb without weight-
bearing is sufficient to hasten transformation of the head.
If non-union results, the neck fragments are gradually
more or less completely eroded and the remaining surfaces
usually become more or less adherent to the capsule and to
each other. The marrow spaces of the dead and non-func-
tioning head are gradually invaded by vascularized
connective tissue by way of the round ligament and in some
cases by way of the adhesions. Absorption of the invaded
old bone and partial replacement by new bone either does
Volume 466, Number 5, May 2008
Fig. 7 Necrotic femoral head casting heavier
shadow than surrounding atrophied living bone.
Fig. 8 Roentgenogram of slice of
excised head in Fig. 7, showing old
necrotic bony trabeculae still
undisturbed.
not take place at all, at least for many months, or it goes on
very much slower than is the case when bony union takes
place and the head is subjected to functional stimulation.
The head casts a shadow for at least one to two years that is
denser than that of the neighboring living bone which is
atrophic from disuse. These points are illustrated by Fig. 7
showing a dense necrotic head in a case of ununited fracture
of the neck of the femur of sixteen months’ duration and by
Fig. 8 which is a roentgenogram of a slice of the excised
head. Although both bone and marrow were dead and about
Repair of Bone in Aseptic Necrosis 1027
one-half of the marrow space was reinvaded by connective
tissue from the round ligament, there was no absorption of
the dead bone except about the fovea, where there was also
a small amount of newly formed bone. Eventually absorp-
tion and partial replacement of the dead head by new bone is
brought about and the head may remain as a cavitous hull or
in rare cases it may be completely absorbed. The bearing of
these findings on prognosis and treatment are brought out in
Santos’s article.
Necrosis of Os Calcis
In fracture of the neck of the os calcis the body may rarely
have so much of its blood supply cut off that it undergoes
extensive necrosis. Its subsequent history is analagous to
that of a necrotic proximal fragment in intracapsular frac-
ture of the neck of the femur. If bony union between the
fragments follows, the necrotic body will be invaded
gradually by blood vessels, fibrous, osteogenic, and mye-
logenous tissue, and a gradual replacement of the necrotic
by living elements will be brought about. The replacement
may be incomplete and the necrotic portion located farthest
from the fracture may be broken down by weight-bearing
123
1028 Phemister
Fig. 9 Case 2. Seventy-four days after fracture of
neck of astragalus, posterior dislocation of body, and
reduction by operation. Body necrotic and retained
normal density. Atrophy of disuse of other bones.
Bony union of fracture.
before the ingrowing tissue reaches it, thereby leading to
deformity and permanent derangement of the bone.
The following cases are illustrative of this condition.
CASE 2. Male, aged twenty-six, sustained a fracture at
the junction of body and neck of the astragalus with
complete dorsal dislocation of the body, the posterior
capsule being the only remaining attachment. It was
reduced by open operation and a cast applied which was
worn for nine weeks. A roentgenogram taken seventy-four
days after injury (Fig. 9) revealed marked reduction in
density of the bones of the foot and the vicinity of the
ankle, except the body of the astragalus which cast a sha-
dow of normal density. The fracture appeared to he uniting,
but the fracture line on the body fragment was irregular and
worm-eaten, indicating invasion and absorption by tissue
from the distal fragment. The diagnosis was made of
aseptic necrosis of the body of the astragalus as indicated
by the preservation of its normal density while the sur-
rounding bone underwent atrophy of disuse. The patient
had stiffness and weakness in the ankle and walked with
crutches. A roentgenogram taken 160 days after injury
revealed appearance of further invasion and reduction in
density of the body of the astragalus with preservation of
its bony articular cortex (Fig. 10). The patient was then lost
sight of.
123
Clinical Orthopaedics and Related Research
CASE 3 is that of a five-year old boy who fell twenty-
two feet, sustaining a fracture at the junction of the neck
and body of the left astragalus with little displacement of
fragments and a fracture of the posterior superior portion of
the os calcis (Fig. 11). A foot and leg cast was worn for six
weeks, after which the patient gradually began to walk on
the limb; but stiffness and pain persisted, causing a con-
siderable amount of disability. A roentgenogram taken 274
days after the injury (Fig. 12) showed union of the frac-
tures in both astragalus and os calcis but the body of the
os calcis had undergone extensive change. There was
irregular increase in density of the distal and inferior por-
tion of the proximal fragment, which was due to newly
formed bone, but the superior and posterior portions bor-
dering on the ankle joint were reduced in volume and
irregularly in density,’ and the shadow of articular surface
was lost in almost its entire extent. There was a small island
of bone in the posterior articular portion of the body. The
body had apparently become necrotic; its distal portion had
been invaded and replaced by new hone, but the articular
portion had broken down from weight-bearing, before being
invaded and partially absorbed by vascularized connective
tissue. Walking was continued with little improvement of
symptoms. A roentgenogram taken fourteen and a half
months after the injury showed still further irregularity and
Volume 466, Number 5, May 2008 Repair of Bone in Aseptic Necrosis 1029

Fig. 10 Case 2. One hundred and sixty days after

fracture. Body slightly reduced in density near
fracture due to invasion from neck.

Fig. 11 Case 3. Fresh fracture of neck of astragalus and chip off Fig. 12 Case 3. Two hundred and seventy-four days after injury.
superior and posterior portion of os calcis. Fracture of astragalus united. Superior and posterior part of body
broken down and irregular while remaining portion dense and
transformed. Indicative of necrosis of body with secondary changes.

123
1030 Phemister
Fig. 13 Case 3. Eighteen months after injury, showing a defective
but gradually reforming articular portion of the body of the astragalus.
reduction in volume of the body of the os calcis. Eighteen
months after the injury (Fig. 13) there was little change in
appearance except that the island of bone posteriorly had
increased considerably in size. Another roentgenogram
twenty-three months after injury showed the changes in the
body had remained practically stationary, but the articular
surface was more regular in outline. The island of bone in
the posterior part of the body was interpreted as a trans-
formed necrotic portion that had become revascularized
with gradual growth of the bony center.
The blood supply of the astragalus is derived mainly
from a branch of the arteria dorsalis pedis which traverses
the sinus tarsi lateral to the neck and breaks up to enter the
bone near the junction of the neck and body along the
lateral and inferior surfaces. There are very small branches
entering the bone mesially and posteriorly at points of
ligamentous and capsular attachments. Apparently when
there is a fracture along the junction of body and neck the
important vessels to the body are interrupted and there may
be insufficient circulation through the remaining vessels, so
that aseptic necrosis of a large part or all of the fragment
follows. It is evident from the partial collapse which
occurred in Case 3 that when necrosis of the body is
diagnosed, the limb should be protected from weight-
bearing for at least several months,—until union, revas-
cularization, and transformation of necrotic area has been
largely brought about. It seems probable that some of the
bad results that have been reported in fracture of the neck
of the astragalus, either united or ununited, have been due
to overlooked aseptic necrosis of the body.
123
Clinical Orthopaedics and Related Research
Bone Transplants
Nearly all of the more recent studies tend to confirm the
view that when bone is transplanted the bone cells within
the lacunae practically all become necrotic as a result of
cutting off of the circulation. The unossified osteogenic
elements along the periphery and in the vascular canals
may survive. The extent to which they proliferate and take
part in the transformation of the necrotic bone depends to a
considerable extent upon the location of the transplant and
the extent to which it is subjected to functional stimulation.
If the transplant is in the course of a bone, with the frag-
ments of which it should become united in order that the
continuity of the bone be restored, and as a part of which it
should function, the surviving osteogenic elements prolif-
erate and take an active part in osteosynthesis and
transformation of the necrotic portion of the graft. Osteo-
genic elements of adjacent bone also participate and more
recent studies as those of Baschkirzew and Petrow [8],
Leriche and Policard [9], and Kartaschew [10], indicate
that fibrous tissue invading aseptic bone may be gradually
transformed into bone. That this is not the source of any
considerable amount of new bone replacing the dead bone
is shown by the observations of Santos that a necrotic head
in ununited fracture of the femoral neck is very little
replaced by new bone by metaplasia of invading connec-
tive tissue from the round ligament. If the transplant is in
the soft parts, disconnected with the skeleton and serving
no supporting function, the surviving unossified osteogenic
elements about the ends may proliferate temporarily,—as
in the healing of a fracture of a bone graft as shown by
Haas, or in closing the open end of the medullary cavity of
the graft as shown by Phemister. But soon resorptive
activities are found to exceed proliferative activities and
the fragment begins to decrease in volume. It may even-
tually be entirely removed in the course of months or years,
depending on the size of the fragment, or the remnants of
the mass after years may consist largely if not entirely of
spongy new bone. Thus, a large splinter of cortex in a
fracture of the tibia displaced dorsally into the soft parts is
shown in a roentgenogram taken shortly after the injury
(Fig. 14). The fracture, united in malposition, was suc-
cessfully treated by open operation, but the fragment was
left undisturbed. Two years later its shadow was about one
third the original size. Seven years later it was reduced to a
small spongy oval mass (Fig. 15) which roentgenologically
has the appearance of living bone.
In case a bone graft of the whole thickness is trans-
planted into a defect in a bone and anchored by insertion of
one end into the medullary cavity, the subsequent changes
in the intramedullary and extramedullary portions are dif-
ferent, due to some extent to their differences in function.
After union occurs the extramedullary portion may
Volume 466, Number 5, May 2008
Fig. 14 Case 4. Large splinter (a) displaced into soft parts from fresh
fracture of tibia.
Fig. 15 Case 4. The same (a) seven years later, showing reduction in
size of splinter and apparent replacement by very spongy new bone.
hypertrophy and gradually be completely transformed into
living bone which may approach the shape and size of the
missing portion. In adults the rate of transformation is
Repair of Bone in Aseptic Necrosis 1031
Fig. 16 Case 5. Tibial graft thirty-nine days after insertion in
humerus.
slower and the extent less than in children. A striking
finding is the relatively small amount of absorption which
goes on in the non-functioning portion of transplant which
lies within the medullary canal beyond the point of union
with the end of the cortex of the host. Such bone may
present practically the same roentgenological appearance
for many years. Figure 16, Case 5, shows a broad graft of
tibial cortex seen on edge, thirtynine days after transplan-
tation for replacement of the upper end of the humerus,
excised because of a small chondrosarcoma, in a man
twenty-two years old. Figure 17 shows the same ten years
and eight months later, there being no recurrence of tumor
in the meantime. The extramedullary portion of the graft
has become slightly enlarged and rounded off and pos-
sessed of a small medullary cavity in its upper third. New
lamellae have formed in its cortex. The intramedullary
portion below the level of union with the end of the hu-
merous is almost unchanged. However, I have examined a
specimen of fibula transplanted eighteen months previously
in precisely the same way to replace the upper end of the
123
1032 Phemister
Fig. 17 Case 5. Ten years and eight months after operation. Shows
non-functioning portion of graft in medullary cavity practically
unchanged, but the functioning portion above hypertrophied and
transformed.
humerus excised for osteogenic sarcoma. Amputation was
performed because of recurrence of the tumor. The extra-
medullary portion which was largely surrounded by tumor
was about one third replaced by living bone. The intra-
medullary portion was porous and on histological
examination showed some absorption with very little
replacement by new bone, suggesting that eventually it
would have been transformed. In contrast with these slow
changes in non-functioning intramedullary grafts is the
rapid absorption which a non-functioning piece of bone in
soft parts undergoes, as shown by Case 4.
Embolism and Necrosis
The etiology of the lesions dealt with up to this point has
been clear. The röle of embolism in the production of both
septic and aseptic bone necrosis is still imperfectly
understood. Ritter, Winkebauer and Axhausen [11] have
expressed the belief that in pyogenic osteomyelitis the
123
Clinical Orthopaedics and Related Research
extensive necrosis of bone is the result of septic embolism
cutting off the blood supply of large areas at the onset of
the process with subsequent invasion and infection of the
infarcted area by bacteria from the embolus. This view is
inconsistent with our knowledge of the blood supply of the
cortex of the shaft of bones. The experiments of Johnson
[12] showed that the periosteal vessels supply approxi-
mately the external half, while the internal half is supplied
by the nutrient artery and metaphyseal vessels. Injury of
either set of vessels results in necrosis of much of the bone
which they supply. Now the dead bone in osteomyelitis is
not limited to the region supplied by one or the other of
these sets of vessels. While an embolus might account for
extensive necrosis in the region supplied by the nutrient
artery, this would be impossible in the regions supplied by
the periosteal vessels since they are numerous and small.
Usually the entire thickness of cortex of the involved
region becomes necrotic. Also in primary and especially in
recurrent osteomyelitis the infection may be observed,
starting at one limited point and spreading more or less
extensively to the rest of the bone.
Aseptic necrosis of bone, the result of proven throm-
bosis or embolism, is a condition that has been rarely
reported in man. Axhausen [11] reported a case which was
classed as multiple anaemic infarction of bones, but
streptococcus longus anhemolyticus was cultured from the
lesions. Consequently it is incorrect to speak of the con-
dition as aseptic necrosis of bone.
Aseptic necrosis has been produced experimentally by
injection of small aseptic emboli into the femoral artery.
The infarcts formed in the ends of the shafts and rarely
comprised a large part of the diaphysis, but epiphyseal
necrosis was not produced. A marked osteoblastic repara-
tive reaction developed about the necrotic bone which was
rapidly replaced by new bone by the process of creeping
substitution. The so called necrosing lesions of the center
of ossification of certain epiphyses in children and the os
lunatum in young adults have been variously attributed to
injury to the blood vessels, vascular obstruction from
embolism or thrombosis, and infection.
Thus Legg-Perthes’ disease sometimes follows reduc-
tion of congenital dislocation of the hip which points to
vascular injury and the experiments of Nussbaum [13]
show that cutting epiphyseal vessels results in necrosis
followed by reorganization and deformity of the bony
center. Phemister, Brunschwig, and Day [14] have cultured
biopsied specimens from two cases each of Köhler’s dis-
ease of the tarsal navicular bone, Legg-Perthes’ disease,
and Kienbock’s disease of the os lunatum. They found that
streptococci grew in cultures of four, and the indications
are strong that this organism is the etiological factor in
some cases. In one case of Köhler’s disease and in one of
Legg-Perthes’ disease the cultures remained sterile,
Volume 466, Number 5, May 2008
although the case of Köhler’s disease was associated with
multiple osseous and lymph glandular tuberculosis. Histo-
logically there was necrosis but no sign of tuberculosis in
the navicular; consequently the case supports the theory of
Axhausen that a caseous tuberculous embolus may be a
cause of the necrosing lesions.
Histologically the majority of these lesions appear to
have something more back of them than a simple bland
embolus or injury cutting off the circulation’and producing
aseptic necrosis. As previously stated, aseptic necrotic
bone in continuity with living bone is gradually invaded
and more or less completely replaced by new bone through
the process of creeping substitution, unless it is too inac-
cessible and is broken down by traumatism. In these
diseases there is a marked fibroblastic and fixed tissue
phagocytic reaction and, in rare cases, an infiltrative
reaction which result in absorption of necrotic bone with-
out bony replacement by creeping substitution except in
occasional instances. New bone formed from surviving
osteogenic elements replaces more or less completely the
absorptive tissues in the course of time.
Bibliography
1. AXHAUSEN, G.: Knochennekrose and Sequesterbildung. Deutsche
Med. Wchnschr., XL, 1, 111, 1914.
2. PHEMISTER, D. B.: Radium Necrosis of Bone. Am. J. Roentgenol.,
XVI, 340, 1926.
Repair of Bone in Aseptic Necrosis 1033
3. BANCROFT, F. W.: Bone Repair Following Injury and Infection.
Arch. Surg., V, 646, 1922.
4. BRUNSCHWIG, A.: Experimental Infarction of Bone Marrow. Proc.
Soc. Exper. Biol. and Med., XXVII, 1049, 1930.
5. CORNIL, V. AND COUDRAY, P.: Du Cal. Au Point de Vue Expéri-
mental et Histologique. J. de l’Anatomie et de la Physiologie, XL,
113, 1904.
6. LOOSER, E.: Über Spätrachitis and Osteomalacie. Klinische,
röntgenologische and pathologisch-anatomische Untersuchungen.
Deutsche Ztschr. f. Chir., CLII, 210, 1920.
7. SANTOS, J. V.: Changes in the Head of the Femur after Complete
Intracapsular Fracture of the Neck: Their Bearing on Non-union
and Treatment. Arch. Surg., XXI, 470, Sept. 1930.
8. BASCHKIRZEW, N. J., UND PETROW, N. N.: Beiträge zur freien
Knochenüberpflanzung. Deutsche Ztschr. f. Chir., CXIII, 490,
1911–1912.
9. LERICHE ET POLICARD. Les Problems de la Physiologie normale et
pathologique de l’Os. Paris, Masson et Cie, 1926.
10. KARTASCHEW, S. I.: Beiträge zur Frage der freien autoplastichen
Knochentransplantation. Experimentelle Untersuchungen mit beson-
derer Berücksichtigung der Transplantation feiner Knochenstückchen
und -splitter. Arch. f. klin. Chir., CLVI, 758, 1930.
11. AXHAUSEN, G.: Über anämische Infarkte am Knochensystem und
ihre Bedeutung für die Lehre von den primären Epiphyseonek-
rosen. Arch. f. klin. Chir., CLI, 72, 1928.
12. JOHNSON, R. W.: A Physiological Study of the Blood Supply of
the Diaphysis. J. Bone and Joint Surg., IX, 153, Jan. 1927.
13. NUSSBAUM, A.: Die arteriellen Gefässe der Epiphysen des
Oherschenkels und ihre Beziehungen zu normalen und patho-
logischen Vorgängen. Bruns’ Beiträge z. klin. Chir., CXXX, 495,
1923–1924.
14. PHEMISTER, D. B., BRUNSCHWIO, A., AND DAY, L.: Streptococcus
Infections of Epiphyses and Short Bones and their Relation to
Legg-Perthes’ Disease, Kienböcks Disease and Köhler’s Disease
of the Tarsal Navicular. J. Am. Med. Assn., XCV, 1930, (in press).
123
Clin Orthop Relat Res (2008) 466:1034–1040
DOI 10.1007/s11999-008-0148-0
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Heritable Thrombophilia-Hypofibrinolysis and Osteonecrosis
of the Femoral Head
Charles J. Glueck MD, Richard A. Freiberg MD,
Ping Wang PhD
Published online: 19 March 2008

The Association of Bone and Joint Surgeons 2008
Abstract We hypothesized that inherited thrombophilia
and hypofibrinolysis were risk factors for osteonecrosis of
the femoral head. We compared measures of thrombophilia
and hypofibrinolysis in referred new adult patients with
idiopathic osteonecrosis (n = 71) or secondary osteonecro-
sis (n = 62) with the same measures in sex- and race-
matched healthy control subjects. Heritable thrombophilic
Factor VIII and hypofibrinolytic Lp(a) were more frequently
high in the 71 patients with idiopathic osteonecrosis than in
control subjects. High Factor VIII, Factor V Leiden hetero-
zygosity, and resistance to activated protein C, all heritable
thrombophilias, were more frequently present in the 62
patients with secondary osteonecrosis than in control sub-
jects. Our data suggest inherited thrombophilia and
hypofibrinolysis are risk factors for both idiopathic and
secondary osteonecrosis of the head of the femur.
Level of Evidence: Level IV, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
One or more of the authors (CJG) has received funding from the
Jewish Hospital Medical Research Council.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent was obtained.
C. J. Glueck (&), P. Wang
Cholesterol Center, Jewish Hospital, ABC Building, 3200 Burnet
Avenue, Cincinnati, OH 45229, USA
e-mail: glueckch@healthall.com
R. A. Freiberg
Cincinnati Veteran’s Administration Hospital, Cincinnati, OH,
USA
123
Introduction
Osteonecrosis (ON) of the femoral head is either
secondary, associated with various factors such as corti-
costeroids, alcoholism, lupus erythematosus, hip trauma
(dislocation, fracture), chemotherapy, HIV-AIDS, dysba-
ria, and others, or considered idiopathic when there is no
known etiology or risk factors [3, 20, 58]. In 1993, we
reported two brothers with idiopathic bilateral hip ON
who were homozygous for the hypofibrinolytic 4G/4G
polymorphism of the plasminogen activator inhibitor-1
gene (PAI-1 gene) and had very high levels of the PAI-1
gene product, hypofibrinolytic plasminogen activator
inhibitor (PAI-Fx) [34]. This kindred [34] and four similar
patients reported by Van Veldhuisen in 1993 [76] led us
to speculate that some cases of ON of the hip are caused
by familial or acquired hypofibrinolysis-thrombophilia
with resultant pathoetiologic venous thrombosis in the
femoral head [5, 24, 32].
Initially we [31], and subsequently others [5, 12, 19, 41,
44, 46, 49, 51, 59, 65, 76, 79, 80], reported data suggesting
venous thrombosis in the femoral head, mediated in many
cases by thrombophilia and hypofibrinolysis, leads to
increased intraosseous venous pressure and thence to
impaired arterial flow, osseous hypoxia, and bone death.
Beyond hip ON, thrombophilia-hypofibrinolysis also
appears associated with some cases of Legg-Calve-Perthes
disease [5, 19, 23, 28, 33, 41, 72].
To confirm and extend previous data, we asked whether
thrombophilia and hypofibrinolysis were risk factors for
ON of the femoral head in patients with idiopathic ON or
ON associated with high-dose, long-term corticosteroids.
We further asked whether age, race, diabetes, hypertension,
and cigarette smoking influenced any association between
Factor VIII levels and ON.
Volume 466, Number 5, May 2008
Materials and Methods
We compared measures of thrombophilia and hypofibrin-
olysis in 133 previously unreported patients with idiopathic
and secondary (corticosteroid-acquired) ON of the head of
the femur with healthy race- and sex-matched normal
control subjects. Alcoholism, postchemotherapy, HIV-
AIDS, and/or a history of fracture or dislocation of the hip
excluded patients from the study. After excluding patients
for these reasons, we identified 133 previously unreported
patients who had measures of thrombophilia and hypofi-
brinolysis in the course of their evaluations. All patients
meeting these criteria were included in the study. The
research protocol was approved by the FDA and the
Institutional Review Board at the Jewish Hospital; signed
informed consent was obtained.
ON was documented by anteroposterior and frog-leg
lateral radiographs of both hips and by MRI evaluation
[31]. MRI was used to confirm the clinical diagnosis of ON
[31]. We made no attempt to quantify the extent of femoral
head involvement by MRI. A consensus diagnosis was
made from the imaging studies by a four-person committee
of radiologists-orthopaedists blinded to patients’ clinical
status, age, and hip symptoms [31]. No selection bias was
used beyond these exclusion criteria. This evaluation pro-
vided a new, previously unreported cohort (Table 1) of 71
patients with idiopathic ON and 62 with ON associated
with corticosteroids (approximately 3000–4000 mg pred-
nisone or its equivalent) [20, 58]. Seven of 25 women
(28%) with idiopathic ON and three of 30 (10%) women
with secondary ON developed ON while taking exogenous
estrogens.
We estimated sample size based on patient-control dif-
ferences in key measures of thrombophilia (Factor V
Table 1. Differences between patients with idiopathic and secondary osteonecrosis and race- and sex-matched healthy normal control subjects
for heritable thrombophilia and hypofibrinolysis
Variable High factor VIII
(greater than 150%)
Idiopathic osteonecrosis 19/71 (27%)
(n = 71; 46 men, 25 women)
Race- and sex-matched control subjects 3/66 (5%)
(n = 69; 44 men, 25 women)
p 0.0004
Variable High factor VIII
(greater than 150%)
Secondary osteonecrosis 16/62 (26%)
(n = 62; 32 men, 30 women)
Race- and sex-matched control subjects 5/60 (8%)
(n = 62; 32 men, 30 women)
p 0.011
Thrombophilia, Hypofibrinolysis, and ON 1035
Leiden mutation [10], resistance to activated protein C
[22], Factor VIII [38]) and hypofibrinolysis (Lp[a]) [64].
Sample size analyses revealed that with alpha = 0.05 and
power 80%, based on patient-control differences in Factor
V Leiden [10], Factor VIII [38], resistance to activated
protein C [22], and Lp(a) [64], there should be 33 or more,
32 or more, 25 or more, and 32 or more subjects in both
patient and control groups, respectively. Our sample size
was adequate to ascertain patient-control differences in our
key measures of thrombophilia and hypofibrinolysis with
alpha = 0.05 and beta = 0.20, because there were 71
patients with idiopathic ON versus 69 control subjects and
62 patients with secondary ON versus 62 control subjects.
The 71 patients with idiopathic ON included 46 men (42
white, four black) and 25 women (23 white, two black).
The 62 patients with secondary ON included 32 men (26
white, six black) and 30 women (26 white, four black). As
control subjects, 44 adult men were available, including 17
previously described healthy male hospital personnel [5]
and 27 new healthy men from family studies (39 white, one
black, and four Asian). Fifty-seven healthy adult female
control subjects included 23 previously described healthy
female hospital personnel [5] and 34 new healthy women
from family studies (49 white, five black, and three Asian).
Subjects were excluded from the control groups if they
were pregnant or taking estrogens, raloxifene, tamoxifen,
corticosteroids, or anticoagulants, all of which might affect
serologic measures of thrombophilia-hypofibrinolysis [4,
30].
We compared thrombophilia and hypofibrinolysis in
patients with 1:1 race- and sex-matched healthy normal
control subjects. Mean age in the 46 men and 25 women
with idiopathic ON (48 ± 12 years and 49 ± 9 years,
respectively) was greater than the race- and sex-matched
High Lp(a) (35 mg/dL or greater)
25/69 (36%)
12/67 (18%)
0.016
Factor V Leiden Resistance to activated
heterozygosity protein C
6/61 (10%) 8/51 (16%)
0/61 (0%) 2/59 (3%)
0.028 0.042
123
1036 Glueck et al.
adult male and female control subjects (43 ± 12 years and
43 ± 9 years, respectively). Mean age in the 32 men and in
the 30 women with secondary ON was 46 ± 10 years and
44 ± 11 years, respectively, similar to the race- and sex-
matched male and female control subjects (41 ± 10 years
and 45 ± 14 years, respectively).
Of the 71 patients with idiopathic ON, 19 (27%)
smoked, similar to 11 of 69 (16%) race- and sex-matched
control subjects (p = 0.12). Of the 62 patients with sec-
ondary ON, 13 (21%) smoked, similar to eight of 62 (13%)
race- and sex-matched control subjects (p = 0.23).
As previously described [4, 5], blood was collected in
3.2% buffered sodium citrate (one part citrate:nine parts
blood). The samples were immediately transported and
centrifuged at 2600 9 g for 15 minutes to obtain platelet-
poor plasma. The samples were run in batches. The plasma
was frozen in aliquots and stored at -70 C. Blood for
polymerase chain reaction (PCR) analysis was drawn in
tubes containing the appropriate anticoagulant (ethylene
diamine tetra-acetic acid).
PCR analysis was used to study four heritable thromb-
ophilic gene mutations: heterohomozygosity for the
G1691A Factor V Leiden, G20210A prothrombin gene, the
platelet glycoprotein PL A1/A2 mutation, homozygosity
for the C677T MTHFR mutation [4, 5, 24, 30, 35], and the
heritable hypofibrinolytic 4G/4G mutation of the PAI-1
gene [4].
Serologic tests used to study thrombophilia were
anticardiolipin antibodies IgG and IgM, the lupus antico-
agulant, deficiency in proteins C and S (total and free),
antithrombin III, homocysteine, and Factors VIII and
XI [4, 5, 29, 30, 35]. Protein C, total and free protein S, and
antithrombin III levels below the fifth percentile for normal
control subjects were considered abnormal [30]. Homo-
cysteine, anticardiolipin antibodies IgG and IgM, Factor
VIII, and Factor XI equal to or over the ninety-fifth
percentile for normal control subjects were considered
abnormal [30, 35].
Hypofibrinolysis studied by serologic tests included
PAI-Fx and Lp(a) [4, 5, 29, 30, 35]. Plasma PAI-Fx and
Lp(a) levels equal to or over the ninety-fifth percentile for
normal control subjects were considered abnormal [30].
Differences in PCR and serologic measures of throm-
bophilia-hypofibrinolysis between patients with idiopathic
or secondary ON and control subjects were assessed using
chi square tests (Table 1). Wilcoxon tests were used to
compare age in patients versus control subjects. Because
Factor VIII might be influenced by age, race, diabetes,
hypertension, and cigarette smoking [17], stepwise logistic
regression analysis was used with the dependent variable
being Factor VIII (level high/normal) and explanatory
variables group, age, race, diabetes, hypertension, and
cigarette smoking. These models were run separately for
123
Clinical Orthopaedics and Related Research
idiopathic and secondary ON. All statistical evaluations
were performed using SAS (SAS/STAT software, 9.1.3,
2002; SAS Institute, Cary, NC).
Results
Heritable thrombophilia and hypofibrinolysis were more
common in patients with ON than in control subjects
(Table 1). Patients with idiopathic ON were more likely
(p = 0.0004) than sex- and race-matched healthy control
subjects to have high (greater than 150%) levels of heri-
table thrombophilic Factor VIII and were also more likely
(p = 0.016) to have inherited high levels of hypofibrino-
lytic Lp(a) (Table 1). Patients with secondary ON were
more likely than sex- and race-matched healthy control
subjects to have high (greater than 150%) levels of Factor
VIII (p = 0.011), to be heterozygous for the Factor V
Leiden mutation (p = 0.028), and to have heritable
thrombophilic resistance to activated protein C (p = 0.042)
(Table 1).
Factor VIII levels in patients with idiopathic and sec-
ondary ON were higher than in control subjects
(p = 0.0017, p = 0.02, respectively) independent of age,
race, hypertension, diabetes, and cigarette smoking.
Discussion
We [31] and others [5, 12, 19, 41, 44, 46, 49, 51, 59, 65, 76,
79, 80] have suggested venous thrombosis in the femoral
head, mediated in many cases by thrombophilia and hyp-
ofibrinolysis, leads to increased intraosseous venous
pressure and subsequent impaired arterial flow, osseous
hypoxia, and bone death. To confirm these suggestions we
asked whether thrombophilia and hypofibrinolysis
increased the risk for ON of the femoral head in patients
with idiopathic ON or ON associated with high-dose, long-
term corticosteroids.
Our study had the following limitations. We did not
have a second control group who received comparable
doses of corticosteroids but who did not develop ON of the
femoral head on prospective followup. We did not measure
intraosseous pressure or do intramedullary venography [55,
73] to document reduction in venous return, venous stasis,
intraosseous hypertension, or decreased arterial inflow. To
optimally further explore the hypothesis that inherited or
acquired thrombophilia-hypofibrinolysis mediates osseous
venous thrombosis [24, 25, 27, 29–34, 39, 43, 44, 51, 59,
62, 65, 72, 76, 79, 80], a placebo-controlled, double-blind
clinical trial [31] would be needed.
We found heritable, thrombophilic high Factor VIII
was much more common in both idiopathic and
Volume 466, Number 5, May 2008
secondary ON than in healthy control subjects. Heritable,
hypofibrinolytic high Lp(a) was more common in idio-
pathic ON than in control subjects. In addition,
heterozygosity for the thrombophilic Factor V Leiden
mutation and thrombophilic resistance to activated pro-
tein C were more common in ON associated with
corticosteroids than in normal control subjects. These
findings, congruent with the amelioration of idiopathic
ON with low-molecular-weight heparin therapy [31],
suggested to us and other authors [1, 5, 8, 11, 41, 49, 50,
62, 65, 76, 79, 80] that thrombophilia-hypofibrinolysis-
mediated thrombosis is a potentially reversible cause of
ON of the head of the femur. Experimental models of
ON [12, 13, 52, 63] and Legg-Calve-Perthes disease [57]
implicate venous occlusion as a precipitating event with
subsequent increased intraosseous pressure, reduced
arterial inflow, ischemia, and infarction. We believe
thrombophilia-hypofibrinolysis, promoting deep osseous
venous thrombosis, initiates this cascade [5, 12, 29, 31,
32, 44, 46, 57, 63].
There are many associations between genetic mutations-
polymorphisms and ON. Single nucleotide polymorphisms
in the multidrug resistance gene have been associated with
corticosteroid-induced ON [2]. Genetic variation in alcohol-
metabolizing enzyme genes is related to alcoholism-
induced ON [15]. In two families with autosomal-dominant
multigenerational idiopathic ON, Chen et al. [16] mapped a
candidate gene to a 15cM region between D12S1663 and
D12S1632 on chromosome 12q13. Another genetic mech-
anism for development of idiopathic ON appears to involve
mutations in the endothelial nitric oxide synthase gene that
controls nitric oxide release [26, 48]. Endothelial nitric
oxide synthase polymorphisms can act alone, synergisti-
cally with cigarette smoking as a genetic risk factor for
idiopathic ON, or in concert with thrombophilia-
hypofibrinolysis [26].
In the current report, high levels of thrombophilic Factor
VIII [6, 53, 54] were more common in patients with idio-
pathic ON or secondary ON than in control subjects (27%
versus 5%, 26% versus 8%). High Factor VIII can be
inherited [6, 53, 54, 69] or acquired related to smoking-
diabetes-hypertension-mediated inflammation [17]. In the
current study, the higher Factor VIII in patients with idi-
opathic and secondary ON versus control subjects could
not be attributed to race, age, smoking, diabetes, or
hypertension, suggesting high Factor VIII in subjects with
ON is not an acute phase reactant, but a contributor to
thrombosis [45]. Our finding of high Factor VIII associated
with idiopathic and secondary ON is congruent with
associations of familial thrombophilias (V Leiden, pro-
thrombin gene, low protein S) [39, 65] with idiopathic ON
in adults and with Legg-Calve-Perthes disease (V Leiden
[5], low protein S [23, 33]).
Thrombophilia, Hypofibrinolysis, and ON 1037
Familial hypofibrinolytic high Lp(a) was associated with
idiopathic ON in the current study and has previously been
reported as a risk factor for idiopathic ON [66]. High Lp(a)
has been associated with familial clustering of bone mar-
row edema of the hip in three sisters [7].
The G1691A Factor V Leiden mutation [75] was more
common in patients with secondary ON than in healthy
control subjects (10% versus 0%). Bjorkman et al. [10]
reported a higher prevalence of the V Leiden mutation in
patients with idiopathic (but not secondary) ON than in the
Swedish population. Zalavras et al. [80] reported that the V
Leiden mutation was present in 18% of 72 patients (23
idiopathic, 49 secondary ON) versus 4.6% of control sub-
jects. Bjorkman et al. [9] also reported that the V Leiden
mutation was associated with ON of the knee. Celik et al.
[14] did not find an excess of the V Leiden mutation in
patients who developed corticosteroid-associated second-
ary ON after renal transplant. The V Leiden mutation and/
or thrombophilic resistance to activated protein C are also
risk factors for Legg-Calve-Perthes disease [5, 19, 22, 41].
Resistance to activated protein C, a heritable risk factor
for venous thrombosis with and without the V Leiden
mutation [74], was more commonly present in patients in
the current report with secondary ON than in control sub-
jects (16% versus 3%). In secondary ON occurring after
corticosteroid-treated severe acute respiratory syndrome,
resistance to activated protein C was more common in
patients than in control subjects [71].
We found endogenous and exogenous hyperestrogen-
emia were common in female patients with ON (28%
idiopathic, 10% secondary) consistent with our previous
studies [27, 30, 37, 39] and those of Montella et al. [61]
When estrogen-induced thrombophilia [40] is superimposed
on familial thrombophilia-hypofibrinolysis, intraosseous
thrombosis is promoted, facilitating development of ON
[30, 32, 36, 37, 39].
Preservation of the femoral head is the ultimate goal of
treatment of ON, but, as summarized by Lieberman [56],
‘‘…development of successful strategies to treat this disease
has been difficult to do because ON is associated with
numerous different diseases and neither the etiology nor the
natural history have been delineated clearly.’’ Assouline-
Dayan et al. [3] concluded ‘‘…management of osteonecrosis
is primarily palliative and does not necessarily halt or retard
the progression of the disease. Treatment options focus on
repairing the secondary changes that develop in the femoral
head and not on reversing the idiopathic pathology.’’ Cur-
rently, strategies for treatment of ON are difficult to develop
[56], do not reverse ON pathologies [3], do not halt pro-
gression to segmental collapse [3], and all have certain
limitations [18, 21, 42, 47, 70, 71]. Having outlined the
strong association between thrombophilia-hypofibrinolysis
with ON of the hip here and elsewhere [24, 25, 27, 29–34, 39,
123
1038 Glueck et al.
43, 44, 51, 72, 79, 80], we have studied the use of 3 months of
low-molecular-weight heparin in patients with Ficat Stage I
or II ON of the hip and one or more thrombophilias or hyp-
ofibrinolyses [31]. Anticoagulation [29, 31, 32] with low-
molecular-weight heparin [67] can stop the progression of
idiopathic hip ON [31] in patients with thrombophilia-
hypofibrinolysis, decreasing the frequency of THA [31].
The diagnosis of thrombophilia-hypofibrinolysis is also
important in patients with ON because of associations with
other venous thromboses, as a stimulus for coagulation
screening in first-degree relatives of affected probands, and
in identifying patients at high risk for deep venous
thrombosis- pulmonary emboli after hip-knee arthroplasty
for whom longer-term postoperative thromboprophylaxis
may be warranted [60, 68, 77, 78].
References
1. Aksoy MC, Aksoy DY, Haznedaroglu IC, Sayinalp N, Kirazli S,
Alpaslan M. Enhanced tissue factor pathway inhibitor response as
a defense mechanism against ongoing local microvascular events
of Legg-Calve-Perthes disease. Pediatr Hematol Oncol.
2005;22:391–399.
2. Asano T, Takahashi KA, Fujioka M, Inoue S, Okamoto M,
Sugioka N, Nishino H, Tanaka T, Hirota Y, Kubo T. ABCB1
C3435T and G2677T/A polymorphism decreased the risk for
steroid-induced osteonecrosis of the femoral head after kidney
transplantation. Pharmacogenetics. 2003;13:675–682.
3. Assouline-Dayan Y, Chang C, Greenspan A, Shoenfeld Y,
Gershwin ME. Pathogenesis and natural history of osteonecrosis.
Semin Arthritis Rheum. 2002;32:94–124.
4. Balasa VV, Gruppo RA, Glueck CJ, Stroop D, Becker A, Pillow
A, Wang P. The relationship of mutations in the MTHFR, pro-
thrombin, and PAI-1 genes to plasma levels of homocysteine,
prothrombin, and PAI-1 in children and adults. Thromb Haemost.
1999;81:739–744.
5. Balasa VV, Gruppo RA, Glueck CJ, Wang P, Roy DR, Wall EJ,
Mehlman CT, Crawford AH. Legg-Calve-Perthes disease and
thrombophilia. J Bone Joint Surg Am. 2004;86:2642–2647.
6. Bank I, Libourel EJ, Middeldorp S, Hamulyak K, van Pampus
EC, Koopman MM, Prins MH, van der Meer J, Buller HR. Ele-
vated levels of FVIII:C within families are associated with an
increased risk for venous and arterial thrombosis. J Thromb
Haemost. 2005;3:79–84.
7. Berger CE, Kluger R, Urban M, Kowalski J, Haas OA, Engel A.
Elevated levels of lipoprotein(a) in familial bone marrow edema
syndrome of the hip. Clin Orthop Relat Res. 2000;377:126–131.
8. Berger CE, Kroner A, Stiegler H, Erdel M, Haas OA, Engel A.
Hypofibrinolysis, lipoprotein(a), and plasminogen activator
inhibitor. Clin Orthop Relat Res. 2002;397:342–349.
9. Bjorkman A, Burtscher IM, Svensson PJ, Hillarp A, Besjakov J,
Benoni G. Factor V Leiden and the prothrombin 20210A gene
mutation and osteonecrosis of the knee. Arch Orthop Trauma
Surg. 2005;125:51–55.
10. Bjorkman A, Svensson PJ, Hillarp A, Burtscher IM, Runow A,
Benoni G. Factor V Leiden and prothrombin gene mutation: risk
factors for osteonecrosis of the femoral head in adults. Clin
Orthop Relat Res. 2004;425:168–172.
11. Blanche P, Si-Larbi AG, Jouve P. Femoral head necrosis and
hyperhomocysteinemia. J Rheumatol. 2001;28:1469.
123
Clinical Orthopaedics and Related Research
12. Boss JH, Misselevich I. Osteonecrosis of the femoral head of
laboratory animals: the lessons learned from a comparative study
of osteonecrosis in man and experimental animals. Vet Pathol.
2003;40:345–354.
13. Boss JH, Misselevich I, Bejar J, Norman D, Zinman C, Reis DN.
Experimentally gained insight-based proposal apropos the treat-
ment of osteonecrosis of the femoral head. Med Hypotheses.
2004;62:958–965.
14. Celik A, Tekis D, Saglam F, Tunali S, Kabakci N, Ozaksoy D,
Manisali M, Ozcan MA, Meral M, Gulay H, Camsari T. Asso-
ciation of corticosteroids and Factor V, prothrombin, and
MTHFR gene mutations with avascular osteonecrosis in renal
allograft recipients. Transplant Proc. 2006;38:512–516.
15. Chao YC, Wang SJ, Chu HC, Chang WK, Hsieh TY. Investi-
gation of alcohol metabolizing enzyme genes in Chinese
alcoholics with avascular necrosis of hip joint, pancreatitis and
cirrhosis of the liver. Alcohol Alcohol. 2003;38:431–436.
16. Chen WM, Liu YF, Lin MW, Chen IC, Lin PY, Lin GL, Jou YS,
Lin YT, Fann CS, Wu JY, Hsiao KJ, Tsai SF. Autosomal dom-
inant avascular necrosis of femoral head in two Taiwanese
pedigrees and linkage to chromosome 12q13. Am J Hum Genet.
2004;75:310–317.
17. Coca M, Cucuianu M, Hancu N. Effect of abdominal obesity on
prothrombotic tendency in type 2 diabetes. Behavior of clotting
Factors VII and VIII, fibrinogen and von Willebrand factor. Rom
J Intern Med. 2005;43:115–126.
18. Coogan PG, Urbaniak JR. Multicenter experience with free vas-
cularized fibular grafts for osteonecrosis of the femoral head. In:
Urbaniak JR, Jones JP Jr, eds. Osteonecrosis: Etiology, Diag-
nosis, and Treatment. Rosemont, IL: American Academy of
Orthopedic Surgeons; 1997:327–352.
19. Eldridge J, Dilley A, Austin H, El-Jamil M, Wolstein L, Doris J,
Hooper WC, Meehan PL, Evatt B. The role of protein C, protein
S, and resistance to activated protein C in Legg-Perthes disease.
Pediatrics. 2001;107:1329–1334.
20. Ficat P. Cortisone associated necrosis of bone. In: Hungerford D,
ed. Ischemia and Necrosis of Bone. Baltimore, MD: Williams &
Wilkins; 1980:171–176.
21. Garino JP, Steinberg ME. Total hip arthroplasty in patients with
avascular necrosis of the femoral head: a 2- to 10-year follow-up.
Clin Orthop Relat Res. 1997;334:108–115.
22. Glueck CJ, Brandt G, Gruppo R, Crawford A, Roy D, Tracy T,
Stroop D, Wang P, Becker A. Resistance to activated protein C
and Legg-Perthes disease. Clin Orthop Relat Res. 1997;338:
139–152.
23. Glueck CJ, Crawford A, Roy D, Freiberg R, Glueck H, Stroop D.
Association of antithrombotic factor deficiencies and hypofi-
brinolysis with Legg-Perthes disease. J Bone Joint Surg Am.
1996;78:3–13.
24. Glueck CJ, Fontaine RN, Gruppo R, Stroop D, Sieve-Smith L,
Tracy T, Wang P. The plasminogen activator inhibitor-1 gene,
hypofibrinolysis, and osteonecrosis. Clin Orthop Relat Res.
1999;366:133–146.
25. Glueck CJ, Freiberg R, Glueck HI, Tracy T, Stroop D, Wang Y.
Idiopathic osteonecrosis, hypofibrinolysis, high plasminogen
activator inhibitor, high lipoprotein(a), and therapy with Stano-
zolol. Am J Hematol. 1995;48:213–220.
26. Glueck CJ, Freiberg R, Oghene J, Fontaine RN, Wang P. Asso-
ciation between the T-786C eNOS polymorphism and idiopathic
osteonecrosis of the head of the femur. J Bone Joint Surg Am.
2007;89:1–26.
27. Glueck CJ, Freiberg R, Tracy T, Stroop D, Wang P. Thrombo-
philia and hypofibrinolysis: pathophysiologies of osteonecrosis.
Clin Orthop Relat Res. 1997;334:43–56.
28. Glueck CJ, Freiberg RA, Crawford A, Gruppo R, Roy D, Tracy
T, Sieve-Smith L, Wang P. Secondhand smoke, hypofibrinolysis,
Volume 466, Number 5, May 2008
and Legg-Perthes disease. Clin Orthop Relat Res. 1998;352:
159–167.
29. Glueck CJ, Freiberg RA, Fontaine RN, Sieve-Smith L, Wang P.
Anticoagulant therapy for osteonecrosis associated with heritable
hypofibrinolysis and thrombophilia. Exp Opin Investig Drugs.
2001;10:1309–1316.
30. Glueck CJ, Freiberg RA, Fontaine RN, Tracy T, Wang P. Hyp-
ofibrinolysis, thrombophilia, osteonecrosis. Clin Orthop Relat
Res. 2001;386:19–33.
31. Glueck CJ, Freiberg RA, Sieve L, Wang P. Enoxaparin prevents
progression of stages I and II osteonecrosis of the hip. Clin
Orthop Relat Res. 2005;435:164–170.
32. Glueck CJ, Freiberg RA, Wang P. Role of thrombosis in osteo-
necrosis. Curr Hematol Rep. 2003;2:417–422.
33. Glueck CJ, Glueck HI, Greenfield D, Freiberg R, Kahn A, Hamer
T, Stroop D, Tracy T. Protein C and S deficiency, thrombophilia,
and hypofibrinolysis: pathophysiologic causes of Legg-Perthes
disease. Pediatr Res. 1994;35:383–388.
34. Glueck CJ, Glueck HI, Mieczkowski L, Tracy T, Speirs J, Stroop
D. Familial high plasminogen activator inhibitor with hypofi-
brinolysis, a new pathophysiologic cause of osteonecrosis?
Thromb Haemost. 1993;69:460–465.
35. Glueck CJ, Iyengar S, Goldenberg N, Smith LS, Wang P. Idio-
pathic intracranial hypertension: associations with coagulation
disorders and polycystic-ovary syndrome. J Lab Clin Med.
2003;142:35–45.
36. Glueck CJ, McMahon RE, Bouquot JE, Triplett D. Exogenous
estrogen may exacerbate thrombophilia, impair bone healing and
contribute to development of chronic facial pain. Cranio.
1998;16:143–153.
37. Glueck CJ, McMahon RE, Bouquot JE, Triplett D, Gruppo R,
Wang P. Heterozygosity for the Leiden mutation of the Factor V
gene, a common pathoetiology for osteonecrosis of the jaw, with
thrombophilia augmented by exogenous estrogens. J Lab Clin
Med. 1997;130:540–543.
38. Glueck CJ, Munjal J, Aregawi D, Agloria M, Winiarska M,
Khalil Q, Wang P. Thrombophilia-hypofibrinolysis and athero-
thrombotic cardiovascular disease \age 45. Transl Res. 2007;
150:93–100.
39. Glueck CJ, Phillips HG, Cameron D, Wang P. Estrogen
replacement in a protein S deficient patient leads to diarrhea,
hyperglucagonemia, and osteonecrosis. JOP. 2001;2:323–329.
40. Gomes MP, Deitcher SR. Risk of venous thromboembolic disease
associated with hormonal contraceptives and hormone replacement
therapy: a clinical review. Arch Intern Med. 2004;164:1965–1976.
41. Gruppo R, Glueck CJ, Wall E, Roy D, Wang P. Legg-Perthes
disease in three siblings, two heterozygous and one homozygous
for the Factor V Leiden mutation. J Pediatr. 1998;132:885–888.
42. Hofmann S, Mazieres B. Osteonecrosis: natural course and con-
servative therapy [in German]. Orthopade. 2000;29:403–410.
43. Jones JP. Risk factors potentially activating intravascular coag-
ulation and causing nontraumatic osteonecrosis. In: Urbaniak JR,
Jones JP Jr, eds. Osteonecrosis: Etiology, Diagnosis, and Treat-
ment. Rosemont, IL: American Academy of Orthopedic
Surgeons; 1997:89–97.
44. Jones LC, Mont MA, Le TB, Petri M, Hungerford DS, Wang P,
Glueck CJ. Procoagulants and osteonecrosis. J Rheumatol.
2003;30:783.
45. Kamphuisen PW, Eikenboom JC, Vos HL, Pablo R, Sturk A,
Bertina RM, Rosendaal FR. Increased levels of Factor VIII and
fibrinogen in patients with venous thrombosis are not caused by
acute phase reactions. Thromb Haemost. 1999;81:680–683.
46. Koo KH, Kim R, Cho SH, Song HR, Lee G, Ko GH. Angiog-
raphy, scintigraphy, intraosseous pressure, and histologic findings
in high-risk osteonecrotic femoral heads with negative magnetic
resonance images. Clin Orthop Relat Res. 1994;308:127–138.
Thrombophilia, Hypofibrinolysis, and ON 1039
47. Koo KH, Kim R, Ko GH, Song HR, Jeong ST, Cho SH.
Preventing collapse in early osteonecrosis of the femoral head. A
randomised clinical trial of core decompression. J Bone Joint
Surg Br. 1995;77:870–874.
48. Koo KH, Lee JS, Lee YJ, Kim KJ, Yoo JJ, Kim HJ. Endothelial
nitric oxide synthase gene polymorphisms in patients with non-
traumatic femoral head osteonecrosis. J Orthop Res. 2006;24:
1722–1728.
49. Korompilias AV, Gilkeson GS, Ortel TL, Seaber AV, Urbaniak
JR. Anticardiolipin antibodies and osteonecrosis of the femoral
head. Clin Orthop Relat Res. 1997;345:174–180.
50. Korompilias AV, Ortel TL, Urbaniak JR. Coagulation abnor-
malities in patients with hip osteonecrosis. Orthop Clin North
Am. 2004;35:265–271, vii.
51. Kubo T, Tsuji H, Yamamoto T, Nakahara H, Nakagawa M,
Hirasawa Y. Antithrombin III deficiency in a patient with mul-
tifocal osteonecrosis. Clin Orthop Relat Res. 2000;378:306–311.
52. Laroche M. Intraosseous circulation from physiology to disease.
Joint Bone Spine. 2002;69:262–269.
53. Lavigne G, Mercier E, Quere I, Dauzat M, Gris JC. Thrombo-
philic families with inheritably associated high levels of
coagulation Factors VIII, IX and XI. J Thromb Haemost.
2003;1:2134–2139.
54. Legnani C, Cini M, Cosmi B, Poggi M, Boggian O, Palareti G.
Risk of deep vein thrombosis: interaction between oral contra-
ceptives and high Factor VIII levels. Haematologica.
2004;89:1347–1351.
55. Li YZ, Yue T. Intraosseous pressure measurement and bone
marrow venography in idiopathic necrosis of femoral head [in
Chinese]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao. 1989;11:
338–343.
56. Lieberman JR. Core decompression for osteonecrosis of the hip.
Clin Orthop Relat Res. 2004;418:29–33.
57. Liu SL, Ho TC. The role of venous hypertension in the patho-
genesis of Legg-Perthes disease. A clinical and experimental
study. J Bone Joint Surg Am. 1991;73:194–200.
58. McKee MD, Waddell JP, Kudo PA, Schemitsch EH, Richards
RR. Osteonecrosis of the femoral head in men following short-
course corticosteroid therapy: a report of 15 cases. CMAJ.
2001;164:205–206.
59. Mont MA, Glueck CJ, Pacheco IH, Wang P, Hungerford DS,
Petri M. Risk factors for osteonecrosis in systemic lupus eryth-
ematosus. J Rheumatol. 1997;24:654–662.
60. Mont MA, Jones LC, Rajadhyaksha AD, Shuler MS, Hungerford
DS, Sieve-Smith L, Wang P, Cordista AG, Glueck CJ. Risk
factors for pulmonary emboli after total hip or knee arthroplasty.
Clin Orthop Relat Res. 2004;422:154–163.
61. Montella BJ, Nunley JA, Urbaniak JR. Osteonecrosis of the
femoral head associated with pregnancy. A preliminary report.
J Bone Joint Surg Am. 1999;81:790–798.
62. Nilsson IM, Krook H, Sternby NH, Soderberg E, Soderstrom N.
Severe thrombotic disease in a young man with bone marrow and
skeletal changes and with a high content of an inhibitor in the
fibrinolytic system. Acta Med Scand. 1961;169:323–337.
63. Norman D, Miller Y, Sabo E, Misselevich I, Peskin B, Zinman C,
Levin D, Reis DN, Boss JH. The effects of enoxaparin on the
reparative processes in experimental osteonecrosis of the femoral
head of the rat. APMIS. 2002;110:221–228.
64. Petersen NH, Schmied AB, Zeller JA, Plendl H, Deuschl G,
Zunker P. Lp(a) lipoprotein and plasminogen activity in patients
with different etiology of ischemic stroke. Cerebrovasc Dis.
2007;23:188–193.
65. Pierre-Jacques H, Glueck CJ, Mont MA, Hungerford DS.
Familial heterozygous protein-S deficiency in a patient who had
multifocal osteonecrosis. A case report. J Bone Joint Surg Am.
1997;79:1079–1084.
123
1040 Glueck et al.
66. Posan E, Szepesi K, Gaspar L, Csernatony Z, Harsfalvi J, Ajzner
E, Toth A, Udvardy M. Thrombotic and fibrinolytic alterations in
the aseptic necrosis of femoral head. Blood Coagul Fibrinolysis.
2003;14:243–248.
67. Research Committee of the British Thoracic Society. Optimum
duration of anticoagulation for deep-vein thrombosis and pul-
monary embolism. Lancet. 1992;340:873–876.
68. Salvati EA, Della Valle AG, Westrich GH, Rana AJ, Specht L,
Weksler BB, Wang P, Glueck CJ. The John Charnley Award: heri-
table thrombophilia and development of thromboembolic disease
after total hip arthroplasty. Clin Orthop Relat Res. 2005;441:40–55.
69. Schambeck CM, Hinney K, Haubitz I, Mansouri Taleghani B,
Wahler D, Keller F. Familial clustering of high Factor VIII levels
in patients with venous thromboembolism. Arterioscler Thromb
Vasc Biol. 2001;21:289–292.
70. Stulberg BN, Davis AW, Bauer TW, Levine M, Easley K.
Osteonecrosis of the femoral head. A prospective randomized
treatment protocol. Clin Orthop Relat Res. 1991;268:140–151.
71. Sun W, Li ZR, Shi ZC, Zhang NF, Zhang YC. Changes in
coagulation and fibrinolysis of post-SARS osteonecrosis in a
Chinese population. Int Orthop. 2006;30:143–146.
72. Szepesi K, Posan E, Harsfalvi J, Ajzner E, Szucs G, Gaspar L,
Csernatony Z, Udvardy M. The most severe forms of Perthes’
disease associated with the homozygous Factor V Leiden muta-
tion. J Bone Joint Surg Br. 2004;86:426–429.
73. Tao SN. Hemodynamics changes in proximal femur of patients
with femoral head necrosis [in Chinese]. Zhonghua Wai Ke Za
Zhi. 1991;29:452–454, 464.
123
Clinical Orthopaedics and Related Research
74. Tosetto A, Castaman G, Cappellari A, Rodeghiero F. The VITA
Project: heritability of resistance to activated protein C. Vincenza
Thrombophilia and Arteriosclerosis. Thromb Haemost. 2000;84:
811–814.
75. Tripodi A. Issues concerning the laboratory investigation of
inherited thrombophilia. Mol Diagn. 2005;9:181–186.
76. Van Veldhuizen PJ, Neff J, Murphey MD, Bodensteiner D,
Skikne BS. Decreased fibrinolytic potential in patients with idi-
opathic avascular necrosis and transient osteoporosis of the hip.
Am J Hematol. 1993;44:243–248.
77. Wahlander K, Larson G, Lindahl TL, Andersson C, Frison L,
Gustafsson D, Bylock A, Eriksson BI. Factor V Leiden
(G1691A) and prothrombin gene G20210A mutations as
potential risk factors for venous thromboembolism after total
hip or total knee replacement surgery. Thromb Haemost. 2002;
87:580–585.
78. Westrich GH, Weksler BB, Glueck CJ, Blumenthal BF, Salvati
EA. Correlation of thrombophilia and hypofibrinolysis with
pulmonary embolism following total hip arthroplasty: an anal-
ysis of genetic factors. J Bone Joint Surg Am. 2002;84:
2161–2167.
79. Zalavras C, Dailiana Z, Elisaf M, Bairaktari E, Vlachogiannop-
oulos P, Katsaraki A, Malizos KN. Potential aetiological factors
concerning the development of osteonecrosis of the femoral head.
Eur J Clin Invest. 2000;30:215–221.
80. Zalavras CG, Vartholomatos G, Dokou E, Malizos KN. Genetic
background of osteonecrosis: associated with thrombophilic
mutations? Clin Orthop Relat Res. 2004;422:251–255.
Clin Orthop Relat Res (2008) 466:1041–1046
DOI 10.1007/s11999-008-0147-1
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Genetic Background of Nontraumatic Osteonecrosis
of the Femoral Head in the Korean Population
Jun-Dong Chang MD, PhD, Mina Hur MD, PhD,
Sang-Soo Lee MD, PhD, Je-Hyun Yoo MD,
Kyu Man Lee MD, PhD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Major thrombophilic mutations have been
identified as risk factors for nontraumatic osteonecrosis of
the femoral head (ONFH) in Caucasians. We asked
whether the genetic background of patients with ONFH
in the Korean population was similar. We analyzed factor
V G1691A mutation (factor V Leiden), prothrombin
G20210A mutation, and methylenetetrahydrofolate reduc-
tase C677T and A1298C polymorphisms in 71 patients (53
men, 18 women) with ONFH. We classified these patients
as 51 alcohol-induced, 18 idiopathic, one steroid-induced,
and one dysbaric. We recruited 200 normal control subjects
(128 men, 72 women). We used multiplex PCR/restriction
fragment length polymorphism for each genotyping. We
observed neither factor V Leiden nor prothrombin
G20210A mutation. Although methylenetetrahydrofolate
reductase A1298C genotypes were not associated with
osteonecrosis, methylenetetrahydrofolate reductase C677T
variant genotypes increased the risk of ONFH compared
with 677CC. Odds ratios of 677CT and 677CT+TT were
One or more of the authors (J-DC) have received funding from a
research grant at Hallym University Medical Center (01-2006-01).
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent was obtained.
J.-D. Chang, S.-S. Lee, J.-H. Yoo
Department of Orthopedic Surgery, Hallym University College
of Medicine, Seoul, Korea
M. Hur (&), K. M. Lee
Department of Laboratory Medicine, Hangang Sacred Heart
Hospital, Hallym University College of Medicine, 94-200,
Youngdeungpo-dong, Youngdeungpo-gu, Seoul 150-719, Korea
e-mail: dearmina@hanmail.net
2.00 (95% confidence interval, 1.05–3.81) and 1.96 (95%
confidence interval, 1.07–3.59), respectively, compared
with 677CC. Our data suggest methylenetetrahydrofolate
reductase C677T polymorphism plays a role in the patho-
genesis of osteonecrosis in the Korean population. It also
implies the genetic risk profile of ONFH may differ among
ethnic populations.
Level of Evidence: Level II, diagnostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis of the femoral head (ONFH) is an ischemic
injury, which results in necrosis of the subchondral bone,
collapse of the femoral head, and degeneration of the hip
[28]. The incidence of ONFH in Korea is relatively higher
compared with that in other countries [11, 17]. ONFH is
one of the most common diseases of the hip in Korea,
occupying more than a half of the underlying causes of
total hip arthroplasty, whereas it is relatively rare in the
United States. Furthermore, the incidence of alcohol-
induced ONFH is also higher in Korea [5].
Although ONFH can be caused by various conditions
such as trauma, glucocorticoid therapy, alcoholism, storage
diseases, and diseases resulting in vasculitis, except for
traumatic conditions, the pathogenesis of osseous ischemia
is not yet completely understood [3, 4, 28]. Recently, a
number of authors have proposed intravascular coagulation
as a pathogenetic mechanism of ONFH [10], with resulting
interruption of the fine osseous blood supply. Several
studies suggest a higher prevalence of coagulation abnor-
malities in patients with ONFH compared with control
subjects [10, 15, 25].
123
1042 Chang et al.
Major genetic mutations related to coagulation abnor-
malities are factor V G1691A mutation (factor V Leiden),
prothrombin G20210A mutation, and 5, 10-methylenetet-
rahydrofolate reductase (MTHFR) C677T and A1298C
polymorphisms [4, 9, 22]. The presence of these genetic
variations is associated with a hypercoagulable state, and
increases the risk of thromboembolic events. Such phe-
nomena, however, have been observed mainly in
Caucasians, and an association between genetic predispo-
sition and thrombotic tendency may differ between ethnic
groups [1, 13, 14, 16, 20]. Recent studies suggest these
genetic predispositions play a role in the risk of ONFH [26,
27], but it is unclear whether they apply to non-Caucasian
populations.
We therefore asked whether there was an association
between major thrombophilic mutations (factor V Leiden,
prothrombin G20210A mutation, and MTHFR C677T and
A1298C polymorphisms) and the risk of ONFH in the
Korean population, and whether the risk was related to the
presumed etiology.
Materials and Methods
We recruited 271 individuals (181 men, 90 women): 71
consecutive patients with nontraumatic ONFH (53 men, 18
women) and 200 normal control subjects (128 men, 72
women) between September 2005 and November 2006.
If the true change in the dependent variables is 0.20 stan-
dard deviations per one standard deviation change in the
independent variable, this sample size would have approxi-
mately 90% power (alpha = 0.05, two-tail). We enrolled
healthy individuals without any clinical disorders or history
of thrombosis as normal control subjects. Their health
status was decided by a physician by routine physical
checkup consisting of clinical, radiographic, and laboratory
evaluations. Alcohol or medication history was checked by
questionnaire and interview, and the presence of pelvic or
hip lesions by plain radiograph. The control group was
used for estimation of the frequency of factor V Leiden,
Table 1. Demographic data of patients with osteonecrosis of the femoral head
Cause Number of patients Age* (years, range)
Idiopathic 18 54 (18–76)
Alcohol-induced 51 55 (29–80)
Steroid-induced 1 41
Dysbarism 1 46
Total 71 55 (18–80)
*Age expressed in median value.
123
Clinical Orthopaedics and Related Research
prothrombin G20210A mutation, and MTHFR C677T and
A1298C polymorphisms in the general population. The
median ages of patients with ONFH and normal control
subjects were 55 years (range, 18–80 years) and 34 years
(range, 21–63 years), respectively. We obtained prior
approval of our Institutional Review Board, and informed
consent on genetic analysis from the subjects at enrollment.
Each patient underwent complete clinical, radiographic
and laboratory evaluations. In all the 71 patients (102 hips),
the diagnosis of ONFH depended on the combination of
clinical symptoms and both plain radiographs and MRI.
According to the Association Research Circulation Osse-
ous (ARCO) classification [8], 23 hips were classified as
stage II, 34 hips as stage III, and 45 hips as stage IV. In the
clinical evaluation, we paid special attention to cortico-
steroid medication, alcohol abuse, and history of
thromboembolic events, and established a presumed etiol-
ogy for the osteonecrosis in each patient. To be considered
as having alcohol-induced osteonecrosis, patients had to
have had an estimated regular alcohol consumption of
more than 400 mL per week before developing symptoms
[5, 12]. To be considered as having corticosteroid-induced
osteonecrosis, patients had to have been taking continuous
corticosteroid medication of more than 20 mg prednisone
per day for a minimum of two months before developing
symptoms [9]. We considered patients with no obvious
underlying etiology as having idiopathic osteonecrosis.
Fifty-one patients had alcohol-induced osteonecrosis, one
had steroid-induced, one had dysbaric osteonecrosis, and
18 had idiopathic (Table 1).
Peripheral blood samples were used for molecular
analysis. DNA was extracted using proteinase K treatment
followed by phenol-chloroform extraction and ethanol
precipitation [21].
For factor V Leiden and prothrombin G20210A muta-
tions, the previously described multiplex polymerase chain
reaction (PCR) method was modified [23]. Primers for
factor V Leiden were FV1 (50 -tgcccagtgcttaacaagacca-30 )
and FV2 (50 -tgttatcacactggtgctaa-30 ), and those for pro-
thrombin G20210A mutation were PT1 (50 -tctagaaacagtt
Gender Involvement
Male Female Unilateral Bilateral
3 15 12 6
49 2 28 23
0 1 0 1
1 0 0 1
53 18 40 31
Volume 466, Number 5, May 2008
gcctggc-30 ) and PT2 (50 -atagcactgggagcattgaagc-30 ). PCR
temperature cycling parameters were: 94° C for 15 minutes
followed by 40 cycles of 94° C for 30 seconds, 55° C for
30 seconds, and 72° C for 30 seconds. After final extension
at 72° C for 10 minutes, two amplification products (345-
bp band for factor V Leiden and 267-bp band for pro-
thrombin G20210A) were simultaneously digested with
Mnl I and Hind III restriction enzymes. Informative bands
for both factor V Leiden (wild type, 272-bp band; hetero-
zygote, 272- and 249-bp bands; mutant type, 249-bp band)
and prothrombin G20210A (wild type, 163-bp band; het-
erozygote, 200- and 163-bp bands; mutant type, 200-bp
band) were identified by electrophoresis on a 2.5% agarose
gel.
For simultaneous detection of MTHFR C677T and
A1298C polymorphisms, another multiplex PCR was per-
formed following the previously described method [24].
Primers for C677T genotyping were 677F (50 -tgaagga
gaaggtgtctgcggga-30 ) and 677R (50 -aggacggtgcggtgagag
tg-30 ), and those for A1298C were 1298F (50 -caaggagga
gctgctgaaga-30 ) and 1298R (50 -ccactccagcatcactcact-30 ).
After 40 cycles of PCR, amplification products were indi-
vidually digested with Hinf I and Mbo II restriction
enzymes for C677T and A1298C genotypings, respec-
tively. Informative bands for both C677T genotyping (wild
type, single 198-bp band; heterozygote, 198- and 175-bp
bands; homozygote variant, single 175-bp band) and
A1298C genotyping (wild type, single 72-bp band; het-
erozygote, 72- and 100-bp bands; homozygote variant,
single 100-bp band) were identified by electrophoresis on a
4% NuSieve GTG agarose gel (Cambrex BioScience
Rockland, Rockland, ME) (Fig. 1).
Fig. 1A–B Restriction fragment length polymorphism analysis of
MTHFR polymorphisms is shown. (A) Shown are gels for Hinf I
digestion for C677T genotyping. Informative bands were 175-bp and
198-bp bands. A, molecular marker; B, CC type; C, TT type; D, CT
type. (B) Shown are gels for Mbo II digestion for A1298C
genotyping. Informative bands were 72-bp and 100-bp bands. A,
molecular marker; B, AA type; C, CC type; D, AC type. MTHFR,
methylenetetrahydrofolate reductase.
Genetic Background of Nontraumatic ONFH in Koreans 1043
Two sets of multiplex PCR were conducted in a Peltier
Thermal Cycler-200 (MJ Research, Inc, Waltham, MA),
and all restriction enzymes were purchased from Promega
(Madison, WI).
Odds ratio (OR) as an estimate of relative risk and 95%
confidence interval (CI) were calculated using MedCalc
software (MedCalc Software, version 9.30; Mariakerke,
Belgium). If the value 1 was not in the range of CI, we
determined there was an increased relative risk in one
group compared with the other.
Results
For MTHFR C677T polymorphism, MTHFR C677T vari-
ant genotypes increased the risk of ONFH compared with
677CC genotype. The presence of the 677CT genotype was
associated with twofold increase (95% CI, 1.05–3.81) in
risk of ONFH compared with the CC genotype. The pres-
ence of combined CT and TT genotypes also increased the
risk of ONFH (OR, 1.96; 95% CI, 1.07–3.59) compared
with the CC genotype. None of the 271 subjects (71
patients with ONFH, 200 normal control subjects) was a
carrier of factor V Leiden and prothrombin G20210A
mutations. For MTHFR A1298C, the presence of variant
genotypes, 1298AC, 1298CC, or both, showed no differ-
ence in the risk of ONFH when using 1298AA as a
reference (Table 2).
When patients were divided into idiopathic (n = 18)
and alcohol-induced groups (n = 51), only in the alcohol-
induced group was the presence of the 677CT genotype
related to the 2.15-fold increase (95% CI, 1.04–4.46) in the
risk of ONFH compared with the CC genotype (Table 2).
Discussion
Major thrombophilic mutations are apparent risk factors for
nontraumatic osteonecrosis of the femoral head (ONFH) in
Caucasians but these have not been confirmed in other
populations. We asked whether there was any association
between major thrombophilic mutations (Factor V Leiden,
prothrombin G20210A mutation and MTHFR polymor-
phisms) and the occurrence of ONFH in the Korean
population. Because current knowledge on the thrombo-
philic genetic background of ONFH was mainly obtained
from the Caucasian populations, we questioned whether
any difference exists in other ethnic groups. Our second
question was whether the risk was related to the presumed
etiology, considering the higher prevalence of an alcohol-
induced ONFH in Korea [5].
Our data are, however, limited in that they were
obtained from a relatively small study population. In
123
1044 Chang et al. Clinical Orthopaedics and Related Research

Table 2. MTHFR polymorphisms and the risk of ONFH

Genotypes Number of patients with ONFH (%) Control subjects Odds ratio (95% CI)
(%) (N = 200)
Total Idiopathic Alcohol-induced Total versus Idiopathic versus Alcohol-induced
(N = 71) (N = 18) (N = 51) control subjects control subjects versus control subjects

MTHFR C677T
CC 18 (25.4) 4 (22.2) 13 (25.5) 80 (40) 1* 1* 1*
CT 36 (50.7) 8 (44.4) 28 (54.9) 80 (40) 2.00 (1.05–3.81)  2.00 (0.58– 6.91) 2.15 (1.04–4.46) 
TT 17 (23.9) 6 (33.3) 10 (19.6) 40 (20) 1.89 (0.88–4.05) 3.00 (0.80–11.24) 1.54 (0.62–3.81)
CT + TT 53 (74.6) 14 (77.8) 38 (74.5) 120 (60) 1.96 (1.07–3.59)  2.33 (0.74–7.34) 1.95 (0.98–3.89)
MTHFR A1298C
AA 49 (69) 10 (55.6) 37 (72.5) 116 (58) 1* 1* 1*
AC 22 (31) 8 (44.4) 14 (27.5) 78 (39) 0.67 (0.37–1.19) 1.19 (0.45–3.15) 0.56 (0.29–1.11)
CC 0 (0) 0 (0) 0 (0) 6 (3) ND ND ND
AC + CC 22 (31) 8 (44.4) 14 (27.5) 84 (42) 0.62 (0.35–1.10) 1.10 (0.42–2.92) 0.53 (0.27–1.03)
Two patients with steroid therapy (n = 1) and dysbarism (n = 1) were included in total patients with ONFH; *reference category (odds ratio,
1.0);  statistically significant; MTHFR = methylenetetrahydrofolate reductase; ONFH = osteonecrosis of the femoral head; CI = confidence
interval; ND = not determined.

particular, the tendency of increased risk in the idiopathic
ONFH group (n = 18) might have been significant with a
larger sample size (Table 2). When allele frequencies were
analyzed in MTHFR C677T polymorphism, although the
presence of T allele tended to increase the risk of ONFH, it
did not reach a statistical significance (OR, 1.46; 95% CI,
0.99–2.14). Accordingly, the modest statistical significance
of MTHFR C677T variant genotypes with about twofold
increased risk of ONFH should be verified in the future
large-scale studies. The possibility of underestimation of
alcohol abuse in both patients and control subjects may be
another limitation, considering the difficulties in detecting
and confirming the degree of alcohol consumption.
Although the exact pathophysiology of nontraumatic
ONFH is still unclear, an increased tendency for intravas-
cular coagulation has been recently proposed as a
pathogenetic mechanism leading to the interruption of the
osseous blood supply and osteonecrosis. Several authors
propose that if thrombosis occurs, it is followed by a
sequential process of obstruction of the venous drainage,
progressive rise of venous pressure, impairment of arterial
perfusion, and osseous necrosis [9, 10, 28].
Factor V Leiden generates coagulation factor V, which
is less effectively degraded by activated protein C resulting
in a hypercoagulable state, and has been established as an
important and unequivocal risk factor for venous throm-
bosis [2, 7]. Prothrombin G20210A mutation leads to
higher levels of prothrombin, increased generation of
thrombin, and thrombophilia. Factor V Leiden and pro-
thrombin G20210A mutations are the most common
genetic risk factors predisposing to thrombosis in Cauca-
sians, and their role in the occurrence of ONFH has been
also identified in several studies. One study reported that
either mutation was present in 22.2% of patients and in
7.3% of control subjects, and their presence was related to
osteonecrosis with an OR of 3.6 [28]. Such a higher
prevalence of factor V Leiden and prothrombin G20210A
mutations in patients with ONFH than in control subjects,
however, was observed only in Caucasians, leaving room
for further investigation in the other ethnic groups.
We found neither factor V Leiden nor prothrombin
G20210A mutation occurred in the study population. This
implies factor V Leiden and prothrombin G20210A
mutations are not genetic risk factors for ONFH in Asians,
or at least in the Korean population. Our finding confirms
previous reports, demonstrating the absence of these
mutations in the Korean and Chinese populations [16, 18].
It provides more evidence that the genetic risk profile of
ONFH, likewise those of the other hypercoagulable dis-
eases including deep vein thrombosis or pulmonary
embolism, may be variable in different ethnic populations.
Hyperhomocysteinemia is an established risk factor for
thrombosis. MTHFR polymorphisms, C677T and A1298C,
decrease the enzyme activity regulating the intracellular
metabolism of homocysteine and thereby mildly elevate
the plasma homocysteine level [6, 22]. In the present study,
MTHFR A1298C genotypes were not associated with the
risk of ONFH. However, MTHFR C677T variant geno-
types increased the risk of ONFH compared with 677CC.
When patients were further divided into etiologic groups,
such a statistical significance was observed only in the
alcohol-induced group (Table 2). Most of our patients (51
of 71 patients, 71.8%) had alcohol-induced ONFH in
contrast to studies of Caucasians, in which most patients
had idiopathic or steroid-induced ONFH [4, 26–28]. Ours
may be the largest genetic study on alcohol-induced

123
Volume 466, Number 5, May 2008
ONFH. Our data suggests alcohol-induced ONFH is also
related to coagulation abnormalities likewise the other
etiologic types of ONFH, although the genetic risk profile
may be different. Taken together, variable incidences and
causes of ONFH may be implicated with an ethnic differ-
ence or a sociocultural difference, which may affect the
pathogenesis of ONFH.
In one small case-control study of Korean subjects, sev-
eral thrombotic (protein C activity, protein S activity,
antithrombin, anticardiolipin antibody, and immunoglobu-
lins), and fibrinolytic factors (tissue plasminogen activator,
plasminogen activator inhibitor-1, lipoprotein[a], and plas-
minogen) were compared among 24 patients with
nontraumatic ONFH and their age- and gender-matched
control subjects [19]. There were no differences in the levels
of these factors, and data could not confirm an etiologic role
for thrombotic and fibrinolytic disorders in East Asian
patients with nontraumatic ONFH. However, homocysteine
levels were not measured in that study, and the role of
thrombophilia in the risk of ONFH cannot be exactly elu-
cidated with the measurement of coagulation profile alone,
which may be affected by a variety of acquired conditions.
We found neither factor V Leiden nor prothrombin
G20210A mutation in our Korean study population.
Although MTHFR A1298C genotypes were not associated
with the risk of ONFH, MTHFR C677T variant genotypes
increased the risk of ONFH compared with the 677CC wild
genotype. The data suggest the MTHFR C677T polymor-
phism may play a role in the pathogenesis of ONFH in the
Korean population, especially in that of alcohol-induced
ONFH. It also implies the genetic risk profile of ONFH
may be variable in different ethnic populations. Further
studies in various ethnic groups are awaited to support the
present findings.
Acknowledgments We thank Tae Young Kang, MT, for his
excellent technical assistance.
References
1. Angchaisuksiri P, Pingsuthiwong S, Aryuchai K, Busabaratana
M, Sura T, Atichartakarn V, Sritara P. Prevalence of the G1691A
mutation in the factor V gene (factor V Leiden) and the G21210A
prothrombin gene mutation in the Thai population. Am J Hema-
tol. 2000;65:119–122.
2. Bertina RM, Koeleman BP, Koster T, Rosendaal FR, Dirven RJ,
de Ronde H, van der Velden PA, Reitsma PH. Mutation in blood
coagulation factor V associated with resistance to activated pro-
tein C. Nature. 1994;369:64–67.
3. Bjorkman A, Burtscher IM, Svensson PJ, Hillarp A, Besjakov J,
Benoni G. Factor V Leiden and the prothrombin 20210A gene
mutation and osteonecrosis of the knee. Arch Orthop Trauma
Surg. 2005;125:51–55.
4. Bjorkman A, Svensson PJ, Hillarp A, Burtscher IM, Rünow A,
Benoni G. Factor V Leiden and prothrombin gene mutation: risk
Genetic Background of Nontraumatic ONFH in Koreans 1045
factors for osteonecrosis of the femoral head in adults. Clin
Orthop Relat Res. 2004;425:168–172.
5. Chang JD, Lee SH, Oh SY, Wi YH, Lee JS. The risk factors
associated with alcohol-induced osteonecrosis of the femoral
head [in Korean]. J Korean Orthop Assoc. 2004;39:692–699.
6. den Heijer M, Koster T, Blom HJ, Bos GM, Briet E, Reitsma PH,
Vandenbroucke JP, Rosendaal FR. Hyperhomocysteinemia as a
risk factor for deep-vein thrombosis. N Engl J Med. 1996;334:
759–762.
7. De Stefano V, Chiusolo P, Paciaroni K, Leone G. Epidemiology
of factor V Leiden: clinical implications. Semin Thromb Hemost.
1998;24:367–379.
8. Gardeniers JWM. A new international classification of osteone-
crosis of the ARCO Committee on terminology and classification.
J Jpn Orthop Assoc. 1992;66:18–20.
9. Glueck CJ, Freiberg RA, Fontaine RN, Tracy T, Wang P. Hyp-
ofibrinolysis, thrombophilia, osteonecrosis. Clin Orthop Relat
Res. 2001;386:19–33.
10. Glueck CJ, Freiberg RA, Tracy T, Stroop D, Wang P. Throm-
bophilia and hypofibrinolysis: pathophysiologies of osteone-
crosis. Clin Orthop Relat Res. 1997;334:43–56.
11. Han CD, Choe WS, Yoo JH. Effect of polyethylene wear on
osteolysis in cementless primary total hip arthroplasty: minimal
5-year follow-up study. J Arthroplasty. 1999;14:714–723.
12. Hirota Y, Hotokebuchi T, Sugioka Y. Idiopathic osteonecrosis of
the femoral head: nationwide epidemiologic studies in Japan. In:
Urbiniak JR, Jones JP Jr, eds. Osteonecrosis: Etiology, Diagno-
sis, Treatment. Rosemont IL: American Academy of Orthopaedic
Surgeons; 1997:51–58.
13. Hong SH, Song J, Kim JQ. Genetic variation of the methylene-
tetrahydrofolate reductase and cystathionine beta-synthase genes
in Korean patients with coronary artery disease and a new
polymorphism in intron 7. Mol Cell Probes. 2001;15:119–123.
14. Hsu LA, Ko YL, Wang SM, Chang CJ, Hsu TS, Chiang CW, Lee
YS. The C677T mutation of the methylenetetrahydrofolate
reductase gene is not associated with the risk of coronary artery
disease or venous thrombosis among Chinese in Taiwan. Hum
Hered. 2001;51:41–45.
15. Jones LC, Mont MA, Le TB, Petri M, Hungerford DS, Wang P,
Glueck CJ. Procoagulants and osteonecrosis. J Rheumatol.
2003;30:783–791.
16. Jun ZJ, Ping T, Lei Y, Li L, Ming SY, Jing W. Prevalence of
factor V Leiden and prothrombin G20210A mutations in Chinese
patients with deep vein thrombosis and pulmonary embolism.
Clin Lab Haematol. 2006;28:111–116.
17. Kim YH, Kim VE. Uncemented porous-coated anatomic total hip
replacement: results at six years in a consecutive series. J Bone
Joint Surg Br. 1993;75:6–13.
18. Kim YW, Yoon KY, Park S, Shim YS, Cho HI, Park SS. Absence
of factor V Leiden mutation in Koreans. Thromb Res.
1997;86:181–182.
19. Lee JS, Koo KH, Ha YC, Koh KK, Kim SJ, Kim JR, Song HR,
Cho SH. Role of thrombotic and fibrinolytic disorders in osteo-
necrosis of the femoral head. Clin Orthop Relat Res.
2003;417:270–276.
20. Lu Y, Zhao Y, Liu G, Wang X, Liu Z, Chen B, Hui R. Factor V
gene G1691A mutation, prothrombin gene G20210A mutation,
and MTHFR gene C677T mutation are not risk factors for pul-
monary thromboembolism in Chinese population. Thromb Res.
2002;106:7–12.
21. Miller SA, Dykes DD, Polesky HF. A simple salting out proce-
dure for extracting DNA from human nucleated cells. Nucleic
Acids Res. 1988;16:1215–1218.
22. Salvati EA, Della Valle AG, Westrich GH, Rana AJ, Specht L,
Weksler BB, Wang P, Glueck CJ. The John Charnley Award:
heritable thrombophilia and development of thromboembolic
123
1046 Chang et al.
disease after total hip arthroplasty. Clin Orthop Relat Res.
2005;441:40–55.
23. Wisotzkey JD, Bell T, Monk JS. Simultaneous polymerase chain
reaction restriction fragment length polymorphism identification
of the factor V Leiden allele and the prothrombin 20210A
mutation. Diagn Mol Pathol. 1998;7:180–183.
24. Yi P, Pogribny IP, James SJ. Multiplex PCR for simultaneous
detection of 677 C?T and 1298 A?C polymorphisms in
methylenetetrahydrofolate reductase gene for population studies
of cancer risk. Cancer Lett. 2002;181:209–213.
25. Zalavras C, Dailiana Z, Elisaf M, Bairaktari E, Vlachogiannopoulos
P, Katsaraki A, Malizos KN. Potential aetiological factors
123
Clinical Orthopaedics and Related Research
concerning the development of osteonecrosis of the femoral head.
Eur J Clin Invest. 2000;30:215–221.
26. Zalavras CG, Malizos KN, Dokou E, Vartholomatos G. The
677C?T mutation of the methylene-tetrahydrofolate reductase
gene in the pathogenesis of osteonecrosis of the femoral head.
Haematologica. 2002;87:111–112.
27. Zalavras CG, Vartholomatos G, Dokou E, Malizos KN. Factor V
Leiden and prothrombin gene mutations in femoral head osteo-
necrosis. Thromb Haemost. 2002;87:1079–1080.
28. Zalavras CG, Vartholomatos G, Dokou E, Malizos KN. Genetic
background of osteonecrosis. Associated with thrombophilic
mutations? Clin Orthop Relat Res. 2004;422:251–255.
Clin Orthop Relat Res (2008) 466:1047–1053
DOI 10.1007/s11999-008-0171-1
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Ethanol May Suppress Wnt/b-catenin Signaling on Human Bone
Marrow Stroma Cells
A Preliminary Study
Ching-Hua Yeh MS, Je-Ken Chang MD,
Yan-Hsiung Wang PhD, Mei-Ling Ho PhD,
Gwo-Jaw Wang MD
Published online: 21 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Ethanol and glucocorticoids are risk factors
associated with osteonecrosis. Previous reports suggest
ethanol and glucocorticoids induce adipogenesis, decrease
osteogenesis in bone marrow stroma cells, and produce
intracellular lipid deposits resulting in death of osteocytes.
The Wnt/b-catenin signal pathway is involved in the
Ching-Hua Yeh and Je-Ken Chang contributed equally to this
manuscript.
One or more of the authors (GJW, JKC, MLH) have received funding
from the National Health Research Institute of Taiwan (NHRI-EX94-
9316EP and NHRI-EX96-9615EP), the Hip Society, Technology
Development Program for Academia in Taiwan (96-EC-17-A-17-S1-
041), and Zimmer, Inc.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
C.-H. Yeh, Y.-H. Wang, M.-L. Ho
Department of Physiology, College of Medicine, Kaohsiung
Medical University, Kaohsiung, Taiwan
C.-H. Yeh, J.-K. Chang, Y.-H. Wang, M.-L. Ho, G.-J. Wang
Orthopaedic Research Center, Kaohsiung Medical University,
Kaohsiung, Taiwan
C.-H. Yeh, M.-L. Ho
Graduate Institute of Medicine, Kaohsiung Medical University,
Kaohsiung, Taiwan
J.-K. Chang, G.-J. Wang (&)
Department of Orthopaedics, College of Medicine, Kaohsiung
Medical University, No 100 Shi-Chuan 1st Road, San Ming
District, Kaohsiung City, Taiwan
e-mail: gwojaw@cc.kmu.edu.tw
J.-K. Chang, G.-J. Wang
Department of Orthopaedics, Kaohsiung Medical University
Hospital, Kaohsiung Medical University, Kaohsiung, Taiwan
regulation of homeostasis of bone and we presume gluco-
corticoids and ethanol may induce osteonecrosis in humans
through a similar mechanism as in rodents. We hypothe-
sized (1) ethanol, like glucocorticoids, decreases
osteogenesis and increases adipogenesis through the Wnt/
b-catenin signaling pathway in human bone marrow stro-
mal cells; and (2) ethanol decreases intranuclear
translocation of b-catenin. We found both dexamethasone
and ethanol decrease the gene and protein expression of
osteogenesis and increase that of adipogenesis through Wnt
signaling-related genes by semiquantitative and quantita-
tive polymerase chain reaction and Western blot. Ethanol
hampered intranuclear translocation of b-catenin by
immunofluorescence analysis. The data suggest the Wnt/b-
catenin signaling pathway may be associated with ethanol-
induced osteonecrosis.
Introduction
Osteonecrosis (ON) is a pathologic process resulting from
direct and indirect injury to the osteoblasts [23]. Numerous
risk factors associated with nontraumatic ON include
corticosteroid treatment, alcoholism, smoking, hyperlipid-
emia, and hyperviscosity [1, 11, 23, 28]. However, the
pathogenesis of nontraumatic ON remains controversial
and no clear connection between adipogenesis and ON has
been established as yet. In our previous studies, we dem-
onstrated chickens treated with steroids developed fat cell
hypertrophy and eventual ON in the femoral head [7, 8].
Ethanol, on the other hand, induces adipogenesis and also
produces intracellular lipid deposits resulting in the death
of osteocytes, which may be associated with the develop-
ment of ON, especially in patients with long-term and
excessive consumption of alcohol [28].
123
1048 Yeh et al.
Wnt signaling pathway and the various Wnt family
members are involved in morphogenesis, organogenesis,
oncogenesis, cell fate determination, regulation of cell
proliferation, and differentiation during embryogenesis [14,
22]. Recent studies suggest the Wnt signaling pathway is
involved in the regulation of homeostasis of bone mass.
The Wnt proteins activate two types of signaling pathways:
canonic and noncanonic. The canonic Wnt proteins bind to
a member of the Frizzled family receptor and its core-
ceptors, LRP5/LRP6, at the cell membrane leading to
glycogen synthase kinase-3b (GSK-3b) inactivation and
the nuclear accumulation of b-catenin by inhibiting phos-
phorylation of b-catenin [20]. Nuclear b-catenin acts as a
transcriptional coactivator by interacting with transcription
factors of the T-cell factor (Tcf)/lymphoid enhancer factor
(Lef) family to regulate gene expression [20]. Low-density-
lipoprotein receptor-related proteins 5 and 6 (LRP5/LRP6)
are indispensable transmembrane proteins for Wnt/b-cate-
nin signaling and are likely to act as Wnt coreceptors [14].
Knockout of LRP5 in mice leads to osteopenia [29].
Overexpression of LRP5 increases bone mass and reduces
osteoblast apoptosis [2]. Therefore, the Wnt signaling
pathway likely plays important roles in the development of
bone.
Glucocorticoids may affect the Wnt signaling pathway
and reduce bone formation by inhibiting the activity of b-
catenin and regulating the expression of Wnt signal-related
molecules in osteoblasts [27]. Because of the similar
pathologic changes between steroid-induced and alcohol-
induced ON [28], we proposed the mechanism between
these two situations might be similar. In previous experi-
ments, we demonstrated ethanol decreased the mRNA
expression of osteogenic genes and increased the mRNA
expression of adipogenic genes [28] similar to the effects
of dexamethasone in human bone marrow stroma cells
from the patients with ON [4]. Based on these previous
studies, we inferred glucocorticoids and ethanol may
induce ON in humans through a similar mechanism.
We therefore hypothesized (1) ethanol, similar to glu-
cocorticoids, decreases osteogenesis and increases
adipogenesis through the regulation of the Wnt signal
pathway on human marrow cells by mRNA and protein
expression; and (2) ethanol decreases intranuclear translo-
cation of b-catenin on human marrow cells.
Materials and Methods
We monitored the responses of multipotent human marrow
cells treated with dexamethasone and ethanol to evaluate
the pathologic change of ON by monitoring gene
expressions related to osteogenesis and adipogenesis
and that of Wnt signaling-related genes. To confirm the
123
Clinical Orthopaedics and Related Research
posttranscriptional effects of ethanol on adipogenesis and
the Wnt signaling pathway, we analyzed the protein
expression of b-catenin and PPARc by Western blot
analysis. b-catenin, the pivotal protein in the Wnt signaling
pathway, was also observed by confocal microscopy for
intranuclear translocation.
We enrolled 13 patients with ethanol-induced ON of the
femoral head and nine patients without ON. Patients with
impaired renal or liver function, patients receiving hor-
mone therapy, and those with malignancy or diabetes
mellitus were excluded. Age, gender, and body mass index
were similar between groups. The non-ON group included
three patients who had THA for dysplastic arthritis of the
hip and six patients who had internal fixation for a fresh
fracture at the acetabulum, pelvis, or femoral shaft within
2 days after injury. Frequent alcohol consumption of more
than 400 mL ethanol per week was reported for all patients
with alcohol-induced ON.
We previously reported the detailed procedures of HMC
culture [4]. Cells from 22 patients were cultured and tested
separately. The cells were separated by a PercollTM gra-
dient (Amersham Pharmacia, Piscataway, NJ); the
nucleated stroma cells were then collected for primary cell
culture [13]. The third passage of culture was used for
experiments. Donor cells (104 cells/cm2) were seeded on a
plate and when they reached 80% confluence were treated
with either 10 or 30 mmol/L ethanol (Sigma, St Louis,
MO) or 100 nmol/L dexamethasone. For immunohisto-
chemistry assay, the cells were pretreated with 25 mmol/L
LiCl (Sigma) as provocative treatment and then further
treated with and without ethanol. Messenger RNA
expression of all target genes was evaluated by semi-
quantitative reverse transcriptase–polymerase chain
reaction (RT-PCR) and quantitative real-time PCR after
treatment for 3 days. Total protein was isolated for Wes-
tern blotting after treatment for 3 days. All independent
experiments containing at least three tests were repeated at
least twice (Table 1).
All 22 cell lines were tested for the surface markers and
the ability of osteogenic, adipogenic, and chondrogenic
differentiation. The surface markers in all cell lines are
compatible with those of mesenchymal stem cells. All cell
lines showed good osteogenic, chondrogenic, and adipo-
genic differentiation after proper induction as previously
reported [19, 30].
We examined the mRNA expression of the Wnt sig-
naling ligand, Wnt 3a; the Wnt protein antagonist, SFRP2;
membrane coreceptor, LRP5; and osteogenic-related genes,
including BMP2, Runx2, and osteocalcin. The adipogenic-
related genes, including PPARc and adipsin, were exam-
ined using RT-PCR and quantitative real-time PCR. For
each gene, the quantitative RT-PCR experiments were
performed with at least three independent batches of
Volume 466, Number 5, May 2008 Ethanol Inhibits Wnt Signal on HMCs 1049

Table 1. List of tests performed

Case ON
ON#1 ON#2 ON#3 ON#4 ON#5 ON#6 ON#7 ON#8 ON#9 ON#10 ON#11 ON#12 ON#13

PCR N=3 N=3 N=3 N=3

n=3 n=3 n=3 n=3
Real-time PCR N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3
n=3 n=2 n=2 n=3 n=2 n=2 n=2 n=2 n=3 n=2 n=2 n=3 n=2
Western blotting N=3 N=3 N=3 N=3
n=3 n=3 n=3 n=3
b-catenin intranuclear translocation N=3 N=3 N=3
n=3 n=3 n=3

Case Non-ON
TA#1 TA#2 TA#3 TA#4 TA#5 TA#6 OA#1 OA#2 OA#3

PCR N=3 N=3 N=3

n=3 n=3 n=3
Real-time PCR N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3 N=3
n=3 n=3 n=2 n=2 n=2 n=3 n=3 n=3 n=2
Western blotting N=3 N=3 N=3 N=3
n=3 n=3 n=3 n=3
b-catenin intranuclear N=3 N=3 N=3
translocation n=3 n=3 n=3

HMCs were isolated from 22 patients (13 ON and nine non-ON cases; non-ON cases include six trauma and three osteoarthritis cases); we have
RT-PCR data from seven patients, real-time PCR data from 22 patients, Western blotting data from eight patients, and b-catenin intranuclear
translocation data from six patients; all independent experiments contain at least three repeat tests (n); each individual experiment was tested at
least twice (N); ON = osteonecrosis; RT-PCR = reverse transcriptase–polymerase chain reaction; TA = Trauma; HMC = Human marrow cell.

cDNAs. Changes (x-fold) in gene expression level were
calculated by the 2-DDct method [21]. Analysis of variance
was performed using Excel software (Microsoft Corp,
Cupertino, CA) as in previous studies [4, 5]. HMCs were
isolated from 22 patients (13 ON and nine non-ON cases;
non-ON cases included six trauma and three osteoarthritis
cases). In total, we have RT-PCR data from seven patients,
real-time PCR data from 22 patients, Western blot data
from eight patients, and b-catenin intranuclear transloca-
tion data from six patients. All independent experiments
containing at least three tests were repeated at least twice
(Table 1).
We performed Western blots on cell extracts in eight
different cell lines, including four patients with ON and
four patients without ON, separated on a 10% sodium
dodecyl sulfate-polyacrylamide gel and blotted onto Hy-
bond-C membrane (Amersham Pharmacia). The
membranes were blocked by 5% nonfat milk and probed
with b-catenin, PPARc, and b-actin. Blots were incubated
with a horseradish peroxidase-conjugated goat anti-mouse
or anti-rabbit IgG (Santa Cruz Biotechnology, Santa Cruz,
CA) and visualized by the enhanced chemiluminescence
system (Amersham Biosciences). The optical densities of
the resolved bands were then semiquantified using Image-
Pro Plus1 analysis software (Media Cybernetics, Bethesda,
MD). All independent experiments containing at least three
tests were repeated at least twice (Table 1).
To further understand whether ethanol suppresses the
canonic Wnt signaling pathway, human marrow cells from
six different cell lines, three from patients with ON and
three from patients without ON, were treated with LiCl, an
inhibitor of GSK-3b, to activate b-catenin. LiCl treatment
was used as a provocative treatment in the intranuclear
translocation of b-catenin by immunofluorescence analysis.
Human marrow cells were cultured on glass coverslips.
After drug treatments, the cells were fixed and permeated
with 0.2% Triton X-100 (Sigma) and then blocked. Cells
were incubated with anti-b catenin antibody for 1 hour and
goat anti-mouse IgG coupled to FITC for 40 minutes.
Meanwhile, the cells were counterstained with DAPI
(Sigma) to highlight the nuclei. All images were observed
with a Fluoview FV 500 (Olympus, Tokyo, Japan) confocal
microscope and processed by a Fluoview FV 500 analysis
system (Olympus). All independent experiments contain-
ing at least three tests were repeated at least twice
(Table 1).

123
1050 Yeh et al.
In the first hypothesis, data from RT-PCR, real-time
PCR, and Western blotting were presented as
mean ± standard deviation and evaluated by one-way
analysis of variance (ANOVA) and Post Hoc test by
Scheffe’s method. The independent variables in the test of
RT-PCR, real-time PCR, and Western blotting are dexa-
methasone and ethanol. A p value less than 0.05 was
considered significant. In the second hypothesis, the
intranuclear translocation of b-catenin was not evaluated
statistically because of the difficulty in the quantitation of
b-catenin on the image. The variables in the test of intra-
nuclear translocation of b-catenin were LiCl and ethanol.
Results
Ethanol suppressed the mRNA expression of osteogenic
genes, including BMP2 (p = 0.005 and 0.009), Runx2
(p = 0.01 and 0.002), and osteocalcin (p = 0.003 and
0.002) (Fig. 1). On the other hand, mRNA expression of
PPARc (p = 0.005 and 0.03) and adipsin (p = 0.03 and
Fig. 1 The human bone marrow cells were treated with dexameth-
asone and ethanol for 3 days and the expression of osteogenic and
adipogenic genes were evaluated by reverse transcriptase–polymerase
chain reaction. Ethanol, 10 and 30 mmol/L, and dexamethasone
Fig. 2 Dexamethasone decreased
SFRP2 expression, but ethanol,
10 and 30 mmol/L, increased
SFRP2 gene expression in
a dose-dependent manner by
reverse transcriptase–polymer-
ase chain reaction. Dexameth-
asone and ethanol, 10 and 30
mmol/L, decreased LRP5 and
Wnt 3a gene expression. The
data are consistent with our first
hypothesis.
123
Clinical Orthopaedics and Related Research
0.02) was upregulated after treatment with ethanol
(Fig. 1). Ethanol suppressed osteogenic gene expression
and induced adipogenic gene expression, similar to dexa-
methasone-treated cells. Like with dexamethasone
treatment, ethanol, 10 and 30 mmol/L, decreased mRNA
expression of LRP5 (p = 0.002 and \ 0.001) and Wnt3a
(p \ 0.001 and \ 0.001) genes. On the other hand, the
expression of Wnt3a was suppressed (p \ 0.001) in a
dose-dependent manner by ethanol treatment (Fig. 2). The
expression of SFRP2 was increased by ethanol treatment,
10 and 30 mmol/L, (p = 0.01 and 0.003) but suppressed
(p = 0.02) by dexamethasone treatment (Fig. 2). The
trends of gene expression were the same in RT-PCR and
real-time PCR. Both dexamethasone and ethanol increased
the mRNA expression of PPARc and decreased that of
BMP2, osteocalcin, Runx2, Wnt 3a, and LRP5 (Table 2).
The ethanol, 10 and 30 mmol/L, suppressed (p = 0.002
and 0.006) b-catenin expression and increased PPARc
(p \ 0.001 and 0.01) expression in a dose-dependent
manner similar to mRNA expression by Western blot
analysis (Fig. 3).
inhibited the mRNA expression of osteogenic genes, including Runx2,
osteocalcin, and BMP2, in a dose-dependent manner and both increased
mRNA expression of adipogenic genes, including PPARc and adipsin,
in a dose-dependent manner. The data support our first hypothesis.
Volume 466, Number 5, May 2008 Ethanol Inhibits Wnt Signal on HMCs 1051

Table 2. Result of osteogenic, adipogenic, and Wnts signal pathway-specific mRNA expression after drug treatment by real-time polymerase
chain reaction measurement in 22 samples of human bone marrow stroma cells
Gene Control Dexamethasone 0.1 lM Ethanol 10 mM Ethanol 30 mM
Mean ± SD Mean ± SD P Mean ± SD P Mean ± SD P

BMP2 100 ± 10 22 ± 9 0.018 24 ± 12 0.012 22 ± 10 0.021

Osteocalcin 100 ± 10 27 ± 8 0.013 27 ± 6 0.036 17 ± 6 0.016
Runx2 100 ± 8 42 ± 13 0.022 34 ± 12 0.011 22 ± 11 0.0041
PPARc 100 ± 7 150 ± 6 0.049 186 ± 3 0.020 272 ± 5 0.031
SFRP2 100 ± 11 47 ± 9 0.022 134 ± 12 0.041 174 ± 12 0.021
Wnt 3a 100 ± 14 42 ± 12 0.035 59 ± 12 0.036 44 ± 13 0.045
LRP5 100 ± 6 60 ± 10 0.038 64.7 ± 8 0.006 59 ± 7 0.014
p \ 0.05 was considered significant; SD = standard deviation.

Addition of the GSK-3b inhibitor, LiCl, at 25 mmol/L
intensified the green light in the nucleus, which meant the
increase of b-catenin–FITC complexes in the nucleus.
After ethanol treatment, immunofluorescence staining
showed b-catenin–FITC complexes were reduced in the
nucleus (Fig. 4). Decrease of b-catenin in the nucleus may
hamper the transcriptional activity.
Discussion
Previous reports indicate an association of prolonged eth-
anol intake and ON [1, 9, 15–18, 24, 28]. Ethanol-induced
ON is associated with marrow cell changes on histologic
sections similar to those from prolonged glucocorticoid
administration [28]. Because of the similar pathologic
changes between glucocorticoid-induced and ethanol-
induced ON, we presumed the mechanism between these
two situations might also be similar. Thus, we presumed
glucocorticoids and ethanol induce ON in humans through
a similar mechanism such as that in the murine cells. We
proposed two hypotheses: (1) ethanol, like glucocorticoids,
decreases osteogenesis and increases adipogenesis through
the regulation of the Wnt signal pathway on human marrow
cells; (2) ethanol diminishes intranuclear translocation of
b-catenin.
There are some limitations to this study. First, the
number of patients enrolled in this study is not large
enough to provide a conclusive answer, but rather a pre-
liminary one. Second, although we demonstrated ethanol
suppresses Wnt/b-catenin signaling in mRNA expression,
protein expression, and intranuclear translocation of b-
catenin, we cannot clearly identify the protein expression
of GSK3b, especially the phosphorylated form. Third, the
activity of GSK3b is not confirmed in this study, although
we clearly observed the changes of intranuclear translo-
cation of b-catenin. The important role of GSK3b is not
clearly linked in this study. Fourth, the control osteoar-
thritis subjects in this study were not exposed to a
substantial amount of ethanol; we do not know whether
they will develop ON if they drink excessively. Further-
more, patients with osteoarthritis may have abnormal bone
remodeling, and therefore we cannot presume their data are
normal.
Gaur et al. [12] reported canonic Wnt signaling pro-
moted osteogenesis by directly stimulating Runx2 gene
expression. We found ethanol suppressed the mRNA
expression of BMP2, Runx2, and osteocalcin, whereas it

Fig. 3 Cell extracts were sub-

jected to immunoblotting using
antibodies against b-catenin and
PPARc The amounts of loading
control were determined by b-
actin. Dexamethasone and etha-
nol, 10 and 30 mmol/L, decreased
b-catenin protein expression but
increased PPARc protein expres-
sion in a dose-dependent pattern.
The results support our first
hypothesis.

123
1052 Yeh et al.
Fig. 4 Human bone marrow cells
were stained with b-catenin anti-
body (green) and the cells were
costained with DAPI to make the
cell nuclei (blue). Immunofluo-
rescent images were observed
with a confocal microscope. After
the treatments with ethanol,
immunofluorescence staining
showed b-catenin–FITC com-
plexes decreased in the nucleus.
The results support our second
hypothesis.
activated the mRNA expression of adipogenic genes,
including PPARc and adipsin. Therefore, ethanol causes
human marrow cells toward adipogenic differentiation
rather than osteogenic differentiation. This may be one of
the possible mechanisms of ethanol-induced ON.
Wang et al. reported death of osteocytes in alcohol-
induced ON [28]. Calder et al. [3] reported steroid-induced
and alcohol-induced ON was accompanied by widespread
apoptosis of osteoblasts and osteocytes. b-catenin was
reported to modulate cell proliferation and survival [29] We
found ethanol decreased the quantity of b-catenin and
hampered the intranuclear translocation of b-catenin in
human marrow cells. It is worthwhile to delineate the rela-
tionship between b-catenin and alcohol-induced apoptosis.
Wang et al. [27] suggested dexamethasone increases the
mRNA expression of SFRP1 in primary mesenchymal cells
from male Sprague-Dawley rats. Our data suggest dexa-
methasone decreases the mRNA expression of SFRP2,
whereas ethanol increases that of SFPR2. Ethanol and
glucocorticoids may act on different SFRPs to modulate
Wnt signaling.
Bone morphogenetic proteins (BMPs) have emerged as
key regulators of stem cell fate commitment [26]. Wnts and
TGF-b superfamily members interact to regulate the tran-
scription of a number of genes [6]. BMPs and Wnt are
important signals determining the fate of immature cells
into cells of the osteoblastic lineage [10]. However,
Nakashima et al. [25] reported BMP-2 did not induce
canonic Wnt expression and Tcf/Lef1-dependent tran-
scriptional activation in C2C12 cells. Our data suggest
123
Clinical Orthopaedics and Related Research
ethanol decreases the gene expression of BMP2 and oste-
ocalcin. Meanwhile, it also decreased gene expression of
LRP5 and Wnt 3a and enhanced the gene expression of
SFRP2 and diminished b-catenin protein level. Crosstalk
between BMP and the Wnt signal pathway may coregulate
osteogenic gene expression, but the relation between Wnt
and the BMP signal pathway requires further investigation.
Acknowledgments We thank Yi-Jen Chen for helping in the
experimental process and Dr Chung-Hwan Chen for the preparation
of the manuscript. We also thank Chihuei Wang, PhD, for help with
the experiment and discussion.
References
1. Arlet J. Nontraumatic avascular necrosis of the femoral head.
Past, present, and future. Clin Orthop Relat Res. 1992;277:12–21.
2. Babij P, Zhao W, Small C, Kharode Y, Yaworsky PJ, Bouxsein
ML, Reddy PS, Bodine PV, Robinson JA, Bhat B, Marzolf J,
Moran RA, Bex F. High bone mass in mice expressing a mutant
LRP5 gene. J Bone Miner Res. 2003;18:960–974.
3. Calder JD, Buttery L, Revell PA, Pearse M, Polak JM. Apopto-
sis—a significant cause of bone cell death in osteonecrosis of the
femoral head. J Bone Joint Surg Br. 2004;86:1209–1213.
4. Chang JK, Ho ML, Yeh CH, Chen CH, Wang GJ. Osteogenic
gene expression decreases in stromal cells of patients with oste-
onecrosis. Clin Orthop Relat Res. 2006;453:286–292.
5. Chen CH, Ho ML, Chang JK, Hung SH, Wang GJ. Green tea
catechin enhances osteogenesis in a bone marrow mesenchymal
stem cell line. Osteoporos Int. 2005;16:2039–2045.
6. Crease DJ, Dyson S, Gurdon JB. Cooperation between the activin
and Wnt pathways in the spatial control of organizer gene
expression. Proc Natl Acad Sci USA. 1998;95:4398–4403.
Volume 466, Number 5, May 2008
7. Cui Q, Wang GJ, Balian G. Steroid-induced adipogenesis in a
pluripotential cell line from bone marrow. J Bone Joint Surg Am.
1997;79:1054–1063.
8. Cui Q, Wang GJ, Su CC, Balian G. The Otto Aufranc Award.
Lovastatin prevents steroid induced adipogenesis and osteone-
crosis. Clin Orthop Relat Res. 1997;344:8–19.
9. Cui Q, Wang Y, Saleh KJ, Wang GJ, Balian G. Alcohol-induced
adipogenesis in a cloned bone-marrow stem cell. J Bone Joint
Surg Am. 2006;88(Suppl 3):148–154.
10. Deregowski V, Gazzerro E, Priest L, Rydziel S, Canalis E. Notch
1 overexpression inhibits osteoblastogenesis by suppressing Wnt/
beta-catenin but not bone morphogenetic protein signaling. J Biol
Chem. 2006;281:6203–6210.
11. Felson DT, Anderson JJ. Across-study evaluation of association
between steroid dose and bolus steroids and avascular necrosis of
bone. Lancet. 1987;1:902–906.
12. Gaur T, Lengner CJ, Hovhannisyan H, Bhat RA, Bodine PV, Komm
BS, Javed A, van Wijnen AJ, Stein JL, Stein GS, Lian JB. Canonical
WNT signaling promotes osteogenesis by directly stimulating
Runx2 gene expression. J Biol Chem. 2005;280:33132–33140.
13. Haynesworth SE, Baber MA, Caplan AI. Cell surface antigens on
human marrow-derived mesenchymal cells are detected by
monoclonal antibodies. Bone. 1992;13:69–80.
14. He X, Semenov M, Tamai K, Zeng X. LDL receptor-related
proteins 5 and 6 in Wnt/beta-catenin signaling: arrows point the
way. Development. 2004;131:1663–1677.
15. Hirota Y, Hirohata T, Fukuda K, Mori M, Yanagawa H, Ohno Y,
Sugioka Y. Association of alcohol intake, cigarette smoking, and
occupational status with the risk of idiopathic osteonecrosis of the
femoral head. Am J Epidemiol. 1993;137:530–538.
16. Hungerford DS, Lennox DW. The importance of increased
intraosseous pressure in the development of osteonecrosis of the
femoral head: implications for treatment. Orthop Clin North Am.
1985;16:635–654.
17. Hungerford DS, Zizic TM. Alcoholism associated ischemic
necrosis of the femoral head. Early diagnosis and treatment. Clin
Orthop Relat Res. 1978;130:144–153.
Ethanol Inhibits Wnt Signal on HMCs 1053
18. Jacobs B. Epidemiology of traumatic and nontraumatic osteone-
crosis. Clin Orthop Relat Res. 1978;130:51–67.
19. Jaiswal N, Haynesworth SE, Caplan AI, Bruder SP. Osteogenic
differentiation of purified, culture-expanded human mesenchymal
stem cells in vitro. J Cell Biochem. 1997;64:295–312.
20. Kengaku M, Capdevila J, Rodriguez-Esteban C, De La Pena J,
Johnson RL, Belmonte JC, Tabin CJ. Distinct WNT pathways
regulating AER formation and dorsoventral polarity in the chick
limb bud. Science. 1998;280:1274–1277.
21. Livak KJ, Schmittgen TD. Analysis of relative gene expression
data using real-time quantitative PCR and the 2(-Delta Delta
C[T]) method. Methods. 2001;25:402–408.
22. Logan CY, Nusse R. The Wnt signaling pathway in development
and disease. Annu Rev Cell Dev Biol. 2004;20:781–810.
23. Mankin HJ. Nontraumatic necrosis of bone (osteonecrosis).
N Engl J Med. 1992;326:1473–1479.
24. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of
the femoral head. J Bone Joint Surg Am. 1995;77:459–474.
25. Nakashima A, Katagiri T, Tamura M. Cross-talk between Wnt
and bone morphogenetic protein 2 (BMP-2) signaling in differ-
entiation pathway of C2C12 myoblasts. J Biol Chem.
2005;280:37660–37668.
26. Varga AC, Wrana JL. The disparate role of BMP in stem cell
biology. Oncogene. 2005;24:5713–5721.
27. Wang FS, Lin CL, Chen YJ, Wang CJ, Yang KD, Huang YT, Sun
YC, Huang HC. Secreted frizzled-related protein 1 modulates
glucocorticoid attenuation of osteogenic activities and bone mass.
Endocrinology. 2005;146:2415–2423.
28. Wang Y, Li Y, Mao K, Li J, Cui Q, Wang GJ. Alcohol-induced
adipogenesis in bone and marrow: a possible mechanism for
osteonecrosis. Clin Orthop Relat Res. 2003;410:213–224.
29. Westendorf JJ, Kahler RA, Schroeder TM. Wnt signaling in
osteoblasts and bone diseases. Gene. 2004;341:19–39.
30. Zuk PA, Zhu M, Mizuno H, Huang J, Futrell JW, Katz AJ,
Benhaim P, Lorenz HP, Hedrick MH. Multilineage cells from
human adipose tissue: implications for cell-based therapies.
Tissue Eng. 2001;7:211–228.
123
Clin Orthop Relat Res (2008) 466:1054–1058
DOI 10.1007/s11999-008-0189-4
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Pitavastatin may Reduce Risk of Steroid-induced Osteonecrosis
in Rabbits
A Preliminary Histological Study
Kenjiro Nishida MD, Takuaki Yamamoto MD, PhD,
Goro Motomura MD, PhD, Seiya Jingushi MD, PhD,
Yukihide Iwamoto MD, PhD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Several animal and human studies suggest
pharmacological approaches may prevent steroid-induced
osteonecrosis (ON). We asked whether the newly
developed 3-hydroxymethyl-3-glutaryl-CoA (HMG-CoA)
reductase inhibitor, pitavastatin, could prevent steroid-
induced ON in rabbits. We injected 65 adult male Japanese
white rabbits once with 20 mg/kg of methylprednisolone
acetate into the right gluteus medius muscle. The rabbits
were divided into two groups; one group of 35 rabbits
received pitavastatins (PS), and the other group of 30
rabbits received no prophylaxis (CTR). Hematological
examinations were performed just before the steroid
injection (0 weeks) and at 1 and 2 weeks after steroid
injection; both the femora and the humeri were histologi-
cally examined 2 weeks postinjection. The incidence of
histologic changes consistent with early ON in the PS
group (13 of 35; 37%) was lower in comparison to the CTR
group (21 of 30; 70%). The size of the bone marrow fat
cells in the PS group (56.6 ± 10 lm) was smaller than
those in the CTR group (60 ± 4 lm). The data suggest
pitavastatin has the potential to lower the incidence of
steroid-induced ON in rabbits.
One of more of the authors (TY) received funding from a Grant-in-
Aid in Scientific Research (No.18591665) from the Japan Society for
the Promotion of Science.
Each author certifies that his or her institution has approved the
animal protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
K. Nishida, T. Yamamoto (&), G. Motomura, S. Jingushi,
Y. Iwamoto
Department of Orthopaedic Surgery, Graduate School of
Medical Sciences, Kyushu University, 3-1-1 Maidashi,
Higashi-ku, Fukuoka 812-8582, Japan
e-mail: yamataku@ortho.med.kyushu-u.ac.jp
123
Introduction
Osteonecrosis (ON) of the femoral head frequently occurs
(3% to 40%) in patients who receive corticosteroids as a
treatment for underlying diseases such as systemic
erythematosus (SLE), nephrotic syndrome, and renal
transplantation [1, 13, 14, 19]. Once ON collapses the
femoral head, most patients undergo surgery [13, 22, 27].
Therefore, preventing ON would be an ideal strategy for
the treatment of this disease.
Several possible factors in the pathogenesis of ON have
been suggested based on both human and animal studies,
including coagulation abnormalities [11], hyperlipidemia
[4, 10, 17, 18, 20, 31–33], and oxidative stress [8, 9].
Human studies suggest vascular occlusion may occur
because of mechanical interruption by the thrombi or lipid
emboli in the nutrient vessels [4, 31]. In the rabbit ON
model, hyperlipidemia with associated abnormal thromb-
ophilic coagulopathy has been linked to the development of
ON [32, 33]. Based on these findings, several recent clin-
ical and experimental studies have explored the effects of
lipid-lowering agents on preventing ON [20, 23, 30]. Wang
et al. [30] reported lovastatin prevented steroid-induced
osteonecrosis in a chicken model.
3-hydroxymethyl-3-glutaryl-CoA (HMG-CoA) reduc-
tase inhibitors (statins) are potent inhibitors of cholesterol
biosynthesis in the liver by blocking the conversion of
HMG-CoA to mevalonate [6]. They have been widely used
for the treatment of hyperlipidemia as well as preventing
coronary artery diseases [25, 26]. Pitavastatin, a newly
developed statin, is apparently a potent and prolonged
inhibitor of sterol synthesis, lowering total cholesterol
(TC), and affecting triglycerides (TG) by enhancing the
hepatic low-density lipoprotein (LDL) receptor and sup-
pressing very-low-density lipoprotein (VLDL) secretion.
Volume 466, Number 5, May 2008
The cholesterol-reducing effect of pitavastatin is greater
than that of the other statins [7].
We asked whether pitavastatin could reduce the risk of
early histologic changes (bone marrow fat cell size) and the
associated hyperlipidemia of steroid-induced ON in
rabbits.
Materials and Methods
We compared the incidence of early histological changes
of ON and the average sizes of the bone marrow fat cells in
two groups, a steroid-only group (control group) and a
group also treated with pitavastatin. The rabbit model of
steroid-induced ON has been previously reported [33]. We
studied 65 adult (with closed growth plates) male Japanese
white rabbits (Kyudo, Tosu, Japan) ranging in age from 28
to 32 weeks. The rabbits were injected once with 20 mg/kg
body weight of methylprednisolone acetate (MPSL,
Upjohn, Tokyo, Japan) intramuscularly into the right glu-
teus medius muscle before the start of the investigation
(week 0) [33]. The rabbits were divided into two groups,
consisting of a pitavastatin group (PS group, n = 35), and
a control group (CTR group, n = 30). Pitavastatin (Kowa
Pharmaceutical, Nagoya, Japan) at a dose of 0.7 mg/kg
body weight per day was intravenously administered in the
PS group once daily for 4 weeks, starting from 2 weeks
before the MPSL injection until 2 weeks after the injection.
In the previous study we reported the incidence of ON in
the rabbits which were injected once with 20 mg/kg body
weight of MPSL was 70% and the incidence of ON in those
with both MPSL injection and probucol treatment was 37%
[20]. In order to detect whether probucol reduced the
incidence of steroid-induced ON in rabbits, with a signifi-
cance level (alpha) of 5% and a power of 80%, sample
sizes of 31 were required for both groups. Because we
expected pitavastatin would be more effective on the pre-
vention of ON than probucol, we decided the sample size is
over 30. The animals were housed at the Animal Center of
Kyushu University. All of the experiments were conducted
in accordance with the Guidelines for Animal Experiments
of Kyushu University, the Law (no. 105), and the notifi-
cation (no. 6) of the government and the Committee on
Ethics in Japan.
The body weight of each rabbit was measured before the
experiment and at 1 and 2 weeks after the MPSL injection.
Two weeks after the MPSL injection, both the femora and
humeri were histologically examined for the presence of
ON and the sizes of the bone marrow fat cells were
examined morphologically. We used a camera that sent
electronic images of the sections to an image processor.
The diameter of bone marrow fat cells displayed on the
video monitor was measured using an interactive
Prevention of Osteonecrosis in Rabbits 1055
mousepad-tracing instrument (NIH image software pro-
gram) [18]. We (KN, TY, GM) determined the size of the
bone marrow fat cells as the average of the greatest
diameters of 100 fat cells in four randomly selected fields
(1 field = 4 9 10-8 m2) from the viable areas. Based on
previous reports, the diagnosis of ON was histologically
confirmed 2 weeks after the steroid administration [9, 20].
Whole areas of the proximal third and distal condyles of
both the femora and humeri (eight regions) were examined
histologically for the presence of ON. A diagnosis of ON
was made blindly and independently by three authors (KN,
TY, GM). Based on the previously published criteria of
ON, a diagnosis of ON was determined based on the
presence of the accumulation of bone marrow cell debris,
and the bone trabeculae with empty lacunae or pyknotic
nuclei of osteocytes within the bone trabeculae, accompa-
nied by surrounding bone marrow cell necrosis. The
presence of repair tissue, comprising granulation tissue,
infiltration of inflammatory cells and appositional bone
formation, was also examined [32, 33]. If the diagnoses
differed among the three examiners, a consensus was
reached by discussing the histologic findings without
knowledge of the group from which the sample was
obtained. Rabbits with at least one osteonecrotic lesion
among the eight areas examined were considered to have
ON.
To evaluate the effect of pitavastatin as a lipid-lowering
agent, we examined plasma lipid levels and the plasma
LDL:HDL cholesterol ratio which is considered a potential
risk factor for corticosteroid-induced ON in rabbits [17].
We collected blood samples from the auricular arteries
while the animals were in a fasting state. The samples were
obtained in the early morning just before the MPSL
injection (week 0) and at weeks 1 and 2 after the MPSL
injection. We measured total cholesterol, low-density
lipoprotein, very-low-density lipoprotein, and triglyceride.
Data were expressed as the mean ± standard deviation.
The size of the bone marrow fat cells in the two groups was
compared using one-way analysis of variance (ANOVA)
with Scheffe’s post hoc test. The hematologic data were
compared by repeated-measures ANOVA with Scheffe’s
post hoc test. Statistical analyses were performed using the
Stat View j-0.5 software program (SAS Institute, Cary,
NC).
Results
The incidence of early histological changes of ON in the
PS group was lower (p = 0.008) in comparison to the CTR
group: 21 of 30 (70%) rabbits in the control and 13 of 35
(37%) in the PS group (Fig. 1). We observed yellowish
areas in the metaphysis and diaphysis. Histologically, the
123
1056 Nishida et al.
Fig. 1 The pitavastatin group had a lower (p = 0.008) incidence of
ON (37%) than the control group (70%).
study rabbits demonstrated an accumulation of bone mar-
row cell debris, and the bone trabeculae had empty lacunae
(Fig. 2A–B). These findings were consistent for all of the
ON-positive rabbits.
The average sizes of the bone marrow fat cells were
smaller (p = 0.002) in the PS group (56.6 ± 10 lm) than
in the CTR group (60 ± 4 lm). In the CTR group, the
average size of bone marrow fat cells was larger
(p = 0.0001) in rabbits with early histological alterations
(61.2 ± 2.7 lm) than in those (56.5 ± 2.5 lm).
The levels of TC in the PS group were lower
(p = 0.001) than those in the CTR group throughout the
experimental period (Fig. 3A). The LDL cholesterol levels
in the PS group remained at lower levels (p \ 0.0001) than
those in the CTR group throughout the experimental period
Fig. 2A–B Histology suggesting early osteonecrotic lesions in the
pitavastatin group is shown. (A) A lower magnified view exhibits an
eosinophilic ON lesion (arrows) as compared with the normal area
2 weeks after the steroid injection (stain, hematoxylin and eosin;
123
Clinical Orthopaedics and Related Research
(Fig. 3B). The average of the plasma LDL:HDL choles-
terol ratio across the experimental period was lower
(p \ 0.0001) in the PS group than in the CTR group,
although the plasma LDL:HDL cholesterol ratio was lower
(p = 0.0002) in the CTR group than in the PS group at
week 0 when the steroid had just been injected (Fig. 3C).
However, the plasma lipid levels (VLDL, TG) did not
differ between the two groups.
Discussion
Several animal and human studies suggest pharmacological
approaches may prevent steroid-induced osteonecrosis
(ON). We therefore tested the hypothesis that pitavastatin
would prevent the development of steroid-induced ON in
rabbits.
The major limitation of this study is that the duration of
steroid treatment may be too short to confirm a diagnosis
of ON. Previous studies suggest histopathologic occur
2–20 weeks after steroid administration [5, 8–12, 17, 18,
20, 31, 32, 33]. In this study, we examined the histological
changes at 2 weeks, before any collapse and confirmation
of ON could be observed. We presume our therefore
represent the early changes of ON.
We measured only serum markers of adipogenesis, total
cholesterol, low-density lipoprotein, very-low-density
lipoprotein, and triglyceride, but we were not able to detect
the direct mechanism of pitavastatin for the prevention of
steroid-induced ON. It has been reported statins have
favorable effects on the progression of atherosclerosis and
plaque instability, independent of their lipid-lowering
activity [16]. These pleiotropic effects of statins also
include improvement of the endothelial function, anti-
thrombotic actions, plaque stabilization, reduction of the
vascular inflammatory process, and antioxidant function
original magnification, 940). (B) A higher magnified view demon-
strates an accumulation of bone marrow cell debris and the bone
trabeculae showing empty lacunae (stain, hematoxylin and eosin;
original magnification, 9100).
Volume 466, Number 5, May 2008
Fig. 3A–C Levels of total cholesterol (TC) and LDL and the ratio of
LDL cholesterol to HDL cholesterol in the control and pitavastatin
groups are shown. The timing just before the MPSL injection is indicated
as 0-week. (A) The TC levels in the pitavastatin group were lower
(p = 0.001) than those in the control group throughout the experimental
[28, 29]. Furthermore, in a recent study, statin was reported
to inhibit both adipogenic and stimulated osteogenic dif-
ferentiation [15]. These various effects of statin seem to
have play an important role in the prevention of steroid-
induced ON.
3-hydroxymethyl-3-glutaryl-CoA (HMG-CoA) reduc-
tase inhibitors (statins) are widely used for the treatment of
hyperlipidemia as well as for the prevention of coronary
artery disease [25, 26]. We selected pitavastatin among the
various statins because pitavastatin has a stronger effect on
LDL-cholesterol reduction than any of the other new stat-
ins, such as pravastatin, simvastatin, or atorvastatin [7, 24].
Another reason why pitavastatin was selected in this study
is that the metabolism of pitavastatin by the cytochrome
P450 (CYP) system is minimal, principally through
CYP2C9, with little involvement of the CYP3A4 isoen-
zyme. Other new statins, simvastatin, lovastatin,
atorvastatin, and cerivastatin are inhibitors of the CYP3A4
isoenzyme [21]. Therefore, the risk of a drug-drug inter-
action between statin and a steroid that is metabolized by
the CYP 3A4 could be reduced by using pitavastatin.
Wang et al. [30] suggested lovastatin prevented steroid-
induced osteonecrosis using a chicken model. We used a
single high dose (20 mg/kg) which was higher than that
used in several previous studies [4, 9, 30]. Abeles et al. [1]
reported the high initial corticosteroid dosage in patients
with systemic lupus erythematosus might induce ON of the
femoral head.
Dexamethasone reportedly stimulates the differentiation
of bone marrow stromal cells into adipocytes as well as the
accumulation of fat in the marrow at the expense of
Prevention of Osteonecrosis in Rabbits 1057
period. (B) The levels of LDL cholesterol in the pitavastatin group
remained lower (p \ 0.0001) than those in the control group throughout
the experimental period. (C) The pitavastatin group exhibited a lower
(p \ 0.0001) than average plasma LDL:HDL cholesterol ratio across the
experimental period in comparison to the control group.
expression of Type-1 collagen and osteocalcin mRNA [2].
This mechanism may explain steroid-induced hypertrophy
and hyperplasia of fat cells in the bone marrow. On the
other hand, a high LDL:HDL cholesterol ratio apparently
reflects prominent lipid transport to the peripheral tissue, a
potential risk factor for corticosteroid-induced osteone-
crosis in rabbits [3, 17]. We observed a decrease in the
LDL:HDL cholesterol ratio. In addition, pitavastatin
reduced the size of the bone marrow fat cells. We therefore
speculate the inhibitory effects of pitavastatin on the
development of ON may be partly explained by the
decrease in lipid deposition in the bone marrow fat cells.
Recent animal experimental studies suggest a combi-
nation treatment with warfarin plus probucol prevents the
development of ON in steroid-treated rabbits [20]. In
addition, Ichiseki et al. [9] reported oxidative stress
plays a crucial role in the development of steroid-induced
ON. These studies suggested steroid-induced ON has a
multifactorial pathogenesis including hyperlipidemia,
coagulation abnormalities, and oxidative stress. In a recent
study, statins preserve endothelial integrity, reduce ische-
mia/reperfusion injury, and depress the interdependent
inflammatory and coagulation cascades via pleiotropic
properties [28]. We thus suppose that not only the LDL-
lowering effects of pitavastatin but also these nonlipid
effects contributed to the prevention of ON by reducing the
formation of the thrombi and lipid emboli in the blood
vessels.
Our preliminary data in this rabbit model suggest pita-
vastatin, a new HMG-CoA reductase inhibitor, may be
useful to prevent steroid-induced ON.
123
1058 Nishida et al.
Acknowledgments We thank Naoko Kinukawa, Department of
Medical Informatics, Kyushu University, Fukuoka, Japan, for her help
and advice on the statistical analysis.
References
1. Abeles M, Urman JD, Rothfield NF. Aseptic necrosis of bone in
systemic lupus erythematosus. Relationship to corticosteroid
therapy. Arch Intern Med. 1978;138:750–754.
2. Cui Q, Wang GJ, Balian G. Steroid-induced adipogenesis in a
pluripotential cell line from bone marrow. J Bone Joint Surg Am.
1997;79:1054–1063.
3. Dobiasova M, Frohlich J. Understanding the mechanism of
LCAT reaction may help to explain the high predictive value of
LDL/HDL cholesterol ratio. Physiol Res. 1998;47:387–397.
4. Fisher DE. The role of fat embolism in the etiology of cortico-
steroid-induced avascular necrosis: clinical and experimental
results. Clin Orthop Relat Res. 1978;130:68–80.
5. Gold EW, Fox OD, Weissfeld S, Curtiss PH. Corticosteroid-
induced avascular necrosis: an experimental study in rabbits. Clin
Orthop Relat Res. 1978;135:272–280.
6. Goldstein JL, Brown MS. Regulation of the mevalonate pathway.
Nature. 1990;343:425–430.
7. Hayashi T, Rani P JA, Fukatsu A, Matsui-Hirai H, Osawa M,
Miyazaki A, Tsunekawa T, Kano-Hayashi H, Iguchi A, Sumi D,
Ignarro LJ. A new HMG-CoA reductase inhibitor, pitavastatin
remarkably retards the progression of high cholesterol induced
atherosclerosis in rabbits. Atherosclerosis. 2004;176:255–263.
8. Ichiseki T, Kaneuji A, Katsuda S, Ueda Y, Sugimori T,
Matsumoto T. DNA oxidation injury in bone early after steroid
administration is involved in the pathogenesis of steroid-induced
osteonecrosis. Rheumatology (Oxford). 2005;44:456–460.
9. Ichiseki T, Matsumoto T, Nishino M, Kaneuji A, Katsuda S.
Oxidative stress and vascular permeability in steroid-induced
osteonecrosis model. J Orthop Sci. 2004;9:509–515.
10. Irisa T, Yamamoto T, Miyanishi K, Yamashita A, Iwamoto Y,
Sugioka Y, Sueishi K. Osteonecrosis induced by a single
administration of low-dose lipopolysaccharide in rabbits. Bone.
2001;28:641–649.
11. Jones JP Jr. Intravascular coagulation and osteonecrosis. Clin
Orthop Relat Res. 1992;277:41–53.
12. Kawai K, Tamaki A, Hirohata K. Steroid-induced accumulation
of lipid in the osteocytes of the rabbit femoral head. A histo-
chemical and electron microscopic study. J Bone Joint Surg Am.
1985 Jun;67:755–763.
13. Koo KH, Kim R. Quantifying the extent of osteonecrosis of the
femoral head. A new method using MRI. J Bone Joint Surg Br.
1995;77:875–880.
14. Landmann J, Renner N, Gächter A, Thiel G, Harder F. Cyclo-
sporin A and osteonecrosis of the femoral head. J Bone Joint Surg
Am. 1987;69:1226–1228.
15. Li X, Cui Q, Kao C, Wang GJ, Balian G. Lovastatin inhibits adi-
pogenic and stimulates osteogenic differentiation by suppressing
PPARgamma2 and increasing Cbfa1/Runx2 expression in bone
marrow mesenchymal cell cultures. Bone. 2003;33:652–659.
16. Liao JK. Isoprenoids as mediators of the biological effects of
statins. J Clin Invest. 2002;110:285–288.
17. Miyanishi K, Yamamoto T, Irisa T, Yamashita A, Jingushi S,
Noguchi Y, Iwamoto Y. A high low-density lipoprotein
123
Clinical Orthopaedics and Related Research
cholesterol to high-density lipoprotein cholesterol ratio as a
potential risk factor for corticosteroid-induced osteonecrosis in
rabbits. Rheumatology (Oxford). 2001;40:196–201.
18. Miyanishi K, Yamamoto T, Irisa T, Yamashita A, Jingushi S,
Noguchi Y, Iwamoto Y. Bone marrow fat cell enlargement and a
rise in intraosseous pressure in steroid-treated rabbits with oste-
onecrosis. Bone. 2002;30:185–190.
19. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of
the femoral head. J Bone Joint Surg Am. 1995;77:459–474.
20. Motomura G, Yamamoto T, Miyanishi K, Jingushi S, Iwamoto Y.
Combined effects of an anticoagulant and a lipid-lowering agent
on the prevention of steroid-induced osteonecrosis in rabbits.
Arthritis Rheum. 2004;50:3387–3391.
21. Mukhtar RY, Reid J, Reckless JP. Pitavastatin. Int J Clin Pract.
2005;59:239–252.
22. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
23. Pritchett JW. Statin therapy decreases the risk of osteonecrosis in
patients receiving steroids. Clin Orthop Relat Res. 2001;386:173–
178.
24. Saito Y, Yamada N, Teramoto T, Itakura H, Hata Y, Nakaya N,
Mabuchi H, Tushima M, Sasaki J, Goto Y, Ogawa N. Clinical
efficacy of pitavastatin, a new 3-hydroxy-3-methylglutaryl
coenzyme A reductase inhibitor, in patients with hyperlipidemia.
Dose-finding study using the double-blind, three-group parallel
comparison. Arzneimittelforschung. 2002;52:251–255.
25. Scandinavian Simvastatin Survival Study Group. Randomised
trial of cholesterol lowering in 4444 patients with coronary heart
disease: the Scandinavian Simvastatin Survival Study (4S).
Lancet. 1994;344:1383–1389.
26. Shepherd J, Cobbe SM, Ford I, Isles CG, Lorimer AR, Macfarlane
PW, McKillop JH, Packard CJ. Prevention of coronary heart dis-
ease with pravastatin in men with hypercholesterolemia. West of
Scotland Coronary Prevention Study Group. N Engl J Med.
1995;333:1301–1307.
27. Shimizu K, Moriya H, Akita T, Sakamoto M, Suguro T. Pre-
diction of collapse with magnetic resonance imaging of avascular
necrosis of the femoral head. J Bone Joint Surg Am.
1994;76:215–223.
28. Sowers J. Effects of statins on the vasculature: implications for
aggressive lipid management in the cardiovascular metabolic
syndrome. Am J Cardiol. 2003;91:14–22.
29. Takemoto M, Liao JK. Pleiotropic effects of 3-hydroxy-3-meth-
ylglutaryl coenzyme a reductase inhibitors. Atheroscler Thromb
Vasc Biol. 2001;21:1712–1719.
30. Wang GJ, Cui Q, Balian G. The pathogenesis and prevention of
steroid-induced osteonecrosis. Clin Orthop Relat Res.
2000;370:295–310.
31. Wang GJ, Sweet DE, Reger SI, Thompson RC. Fat-cell changes
as a mechanism of avascular necrosis of the femoral head in
cortisone-treated rabbits. J Bone Joint Surg Am. 1977;59:729–
735.
32. Yamamoto T, Hirano K, Tsutsui H, Sugioka Y, Sueishi K. Cor-
ticosteroid enhances the experimental induction of osteonecrosis
in rabbits with Shwartzman reaction. Clin Orthop Relat Res.
1995;316:235–243.
33. Yamamoto T, Irisa T, Sugioka Y, Sueishi K. Effects of pulse
methylprednisolone on bone and marrow tissues: corticosteroid-
induced osteonecrosis in rabbits. Arthritis Rheum. 1997;40:2055–
2064.
Clin Orthop Relat Res (2008) 466:1059–1067
DOI 10.1007/s11999-008-0178-7
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Preventive Effects of Puerarin on Alcohol-induced Osteonecrosis
Yisheng Wang MD, Li Yin MD, Yuebai Li PhD,
Peilin Liu MD, Quanjun Cui MD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Alcohol can induce adipogenesis by bone
marrow stromal cells and may cause osteonecrosis of the
femoral head. Currently, there are no medications available
to prevent alcohol-induced osteonecrosis. We hypothesized
puerarin, a Chinese herbal medicine with antioxidative and
antithrombotic effects, can prevent alcohol-induced adi-
pogenesis and osteonecrosis. Both bone marrow stromal
cells (in vitro) and mice (in vivo) were treated either with
ethanol or with ethanol and puerarin, with an untreated
group serving as a control. In the in vitro study, the number
of adipocytes, contents of triglycerides, and levels of
PPARc mRNA expression were decreased and alkaline
phosphatase activity, contents of osteocalcin, and levels of
osteocalcin mRNA expression were increased in cells
Investigation performed at Orthopaedic Institute of Zhengzhou
University, Zhengzhou, Henan, China.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
animal protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
Y. Wang, L. Yin, P. Liu
Department of Orthopaedic Surgery, 1st Affiliated Hospital,
Zhengzhou University, Zhengzhou, China
Y. Li
Department of Biochemistry and Molecular Biology, Basic
Medical College, Zhengzhou University, Zhengzhou, China
Q. Cui (&)
Department of Orthopaedic Surgery, University of Virginia
School of Medicine, P.O. Box 800159, Charlottesville,
VA 22908-0159, USA
e-mail: qc4q@hscmail.mcc.virginia.edu
treated with both alcohol and puerarin, compared with cells
treated with alcohol only. In the in vivo study, marrow
necrosis, fat cell hypertrophy and proliferation, thinner and
sparse trabeculae, diminished hematopoiesis, and increased
empty osteocyte lacunae in the subchondral region of the
femoral head were observed in mice treated with alcohol.
However, no such changes were seen in femoral heads of
mice treated with alcohol and puerarin. The data suggest
puerarin can inhibit adipogenic differentiation by bone
marrow stromal cells both in vitro and in vivo and prevents
alcohol-induced osteonecrosis in this model.
Introduction
Approximately one-third of patients with nontraumatic
osteonecrosis (ON) have associated alcohol abuse [1, 5,
16–19, 36, 45]. Animal studies suggest alcohol can cause
hypertrophy and proliferation of fat cells, fatty degenera-
tion of osteocytes, and thinner and sparse trabeculae [20].
Several reports suggest primary marrow mesenchymal
stromal cells (MSCs) and cloned bone marrow stem cells
treated with alcohol resulted in adipogenic differentiation
and decreased levels of alkaline phosphatase (ALP) and
osteocalcin [10, 20, 28]. Adipose-specific gene 422(aP2)
expression is enhanced and osteogenic gene Type I colla-
gen expression decreased in MSCs exposed to alcohol [28].
These findings suggest the mechanism of alcohol-induced
ON may be related to alcohol directly inducing adipogen-
esis and inhibiting osteogenesis in MSCs.
For thousands of years the traditional Chinese herbal
medicine pueraria has been used to treat alcoholism [2, 25,
26, 32, 38, 41]. Recently, it was found puerarin, extracted
from pueraria, is antioxidative, antithrombotic, and
decreases cell injuries secondary to lipid peroxidation by
123
1060 Wang et al.
protecting the stability of the cell membrane [4, 6, 7, 11,
12, 53, 55]. The findings indicate puerarin may have the
potential to prevent alcohol-induced ON.
We therefore hypothesized puerarin inhibits alcohol-
induced adipocytic differentiation of MSCs and fatty
degeneration of osteocytes and may be able to prevent
alcohol-induced osteonecrosis. We further hypothesized
puerarin would reduce expression of PPARc, a transcrip-
tion factor important in adipogenesis.
Materials and Methods
In order to test our hypotheses, we conducted both in vitro
and in vivo studies. Bone marrow stromal cells and mice
were treated either with alcohol only or with both alcohol
and puerarin, or received no treatment to serve as control.
Preventive effects of puerarin on alcohol-induced adipo-
genesis and osteonecrosis were analyzed by examining the
morphologic changes, specific gene expression, serology,
and histology.
We obtained bone marrow cells from the midshafts of
6- to 8-week-old male and female mice femurs and plated
at a density of 1.5 9 106 cells/cm2. The cells were main-
tained in Dulbecco’s modified Eagle medium (Gibco BRL,
Gaithersburg, MD) containing 10% fetal bovine serum
(Hyclone Laboratories, Logan, UT), 50 mg/mL sodium
ascorbate, and antibiotics (100 U/mL penicillin G and
100 mg/mL streptomycin) in a humidified atmosphere of
5% CO2 at 37° C. Medium was added to the wells at the
same time as the cells were seeded, and was first changed
after 72 hours, then every 48 hours thereafter. At the same
time when cells were seeded, we established three groups:
(1) cells treated with 0.09 mol/L ethanol, (2) cells treated
with 0.09 mol/L ethanol and 0.01 mg/mL puerarin, and (3)
cells without treatment as controls. Four replicates of each
group were obtained.
After 21 days of culture, cells of each group were fixed
in 75% ethanol for 8 minutes and stained with Sudan III for
30 minutes at 60° C and counterstained with hematoxylin.
The number of adipocytes in 100-mm2 on each well in 24-
well plates was counted under a microscope (Leitz, Al-
lendale, NJ, USA) equipped with a color video camera and
frame grabber. The average number of adipocytes in 24-
well plates was calculated.
We assayed triglycerides using a kit (Sino-America,
Luoyang, China). Confluent cells cultured for 21 days in
each group were removed mechanically into 2 mL phos-
phate-buffered saline. The cells were lysed by freezing and
thawing repeatedly and centrifuged at 700 9 g for
10 minutes by Sorvall RMC 14 (Kendro, Newtown, CT,
USA). The supernatants were used for the assay. A solution
of chloroform and methanol (2:1, volume per volume) was
123
Clinical Orthopaedics and Related Research
mixed with the supernatants, placed at 37° C for 5 minutes,
and then the chloroform-methanol phase was removed.
Using the kit, the levels of the triglycerides were deter-
mined with a Hitachi 7150 Biochemistry analyzer (Hitachi,
Tokyo, Japan).
Alkaline phosphatase activity was determined using a
kit for ALP (Changzheng Technological Inc, Shanghai,
China). Confluent cells cultured for 12 days in each group
were suspended mechanically in 1 mL phosphate-buffered
saline for each well. The cells were lysed by freezing and
thawing repeatedly and centrifuged at 700 9 g for
10 minutes by Sorvall RMC 14 (Kendro, Newtown, CT).
The supernatants were used for the assay. Absorption was
measured at 410 nm on a spectrophotometer (Spectronic
Instruments, Rochester, NY). The values were standardized
by determining the total protein in the cell layers with the
use of a Coomassie brilliant blue method. Culture media
from cells cultured for 14 days in each group were col-
lected and used to determine the levels of osteocalcin by
radioimmunoassay.
We randomly divided 216 4-week-old Kunming mice
(experimental animal center, Henan Province) into three
groups: (1) model group: mice received spirits (20 mL/kg
body weight) containing 46% ethanol intragastrically and
normal saline (10 mL/kg) by intramuscular injection; (2)
experimental group: received spirits (20 mL/kg body
weight) containing 46% ethanol intragastrically and puera-
rin (0.5 g/kg body weight) by intramuscular injection; and
(3) control group: received water (20 mL/kg) intragastri-
cally and normal saline (10 mL/kg) by intramuscular
injection daily. In addition, the animals were free to receive
food and water at all times. Fifteen animals from each group
were sacrificed using overdose anesthesia 4, 6, 8, and
10 months after treatment and specimens were processed for
histology, and additional 12 animals were used at 10 months
to detect gene expression. The study protocol was approved
by the Animal Review Board of the University.
We collected blood samples after a 12-hour fasting
period 4, 6, 8, and 10 months after treatment. Serum levels
of total cholesterol (CHO), triglyceride (TG), and alkaline
phosphatase activity (ALP) were determined by a method
of biochemical assay using kits (Sino-America, Luoyang,
China).
Liver specimens harvested from the right lobe each
measuring about 5 9 5 9 3 mm were processed for frozen
sectioning, and 5-lm sections were cut, stained with
hematoxylin and eosin (HE) or Sudan III (Sigma Chemical,
St. Louis, MO), and examined by light microscopy. Fem-
oral head specimens were cut symmetrically along the
coronal plane into two parts. Half of the specimens were
fixed in 10% formalin for 24 hours and then decalcified in
10% ethylenediaminetetraacetic acid (EDTA) in Tris-HCl
buffer. The tissues were embedded in paraffin. Five-lm
Volume 466, Number 5, May 2008 Alcohol-induced Osteonecrosis Prevention 1061

sections were cut and stained with hematoxylin and eosin. Osteocalcin:
The remaining specimens were sent for frozen section and Forward primer-50 GAGCAGAGCTCCCTGAACTG30
stained with Sudan III. We examined five sections from Reverse primer-50 GGTCGCCCTAGAGACAAGAA30
each animal using a modification of the method described
by Warner et al. [51]. Briefly, five fields within the zone of b-actin:
the subchondral area of the femoral head on each section Lower-50 CGACCAGAGGCATACAGG30
were chosen. The first field was located at the approximate Upper-50 GGTGTGATGGTGGGAATG30
center of the subchondral bone at the weight-bearing area,
Products of b-actin, PPARc and osteocalcin were
and the remaining four fields were located at both sides of
408 bp, 205 bp, and 200 bp oligonucleotides respectively.
the first field with two fields at each side. The mean of the
Reverse transcription of extracted RNA was performed to
five fields from each section was determined to represent
synthesize cDNA. The synthesized cDNA was subse-
that section. The mean of the five sections from each ani-
quently amplified by the polymerase chain reaction (PCR)
mal was taken as the value for that animal. The following
with specific primers. PCR reaction products were obtained
parameters were assessed: (1) 200 osteocyte lacunae in
by electrophoresis using 1.2% agarose gel. Semiquantita-
each established field were counted under a light micro-
tive analysis of PCR products was performed using a gel
scope at 9200 magnification, and then the percentage of
imaging scanning system. The ratios of osteocalcin or
the empty osteocyte lacunae was determined; and (2) the
PPARc absorbance value to b-actin absorbance value were
average diameter of the largest fat cell was measured in
viewed as absolute value of osteocalcin or PPARc product
each field using an ocular micrometer under the light
respectively.
microscope at 9200 magnification [49], and the average
The estimate of sample size was based on being able to
diameter for fat cells in each animal was determined.
detect at Month 10 a significant increase in largest fat cell
After MSCs of each group were treated 6 days in vitro
size, number of empty osteocyte lacunae, and serum
or animals of each group for 6 months in vivo, using
chemistry parameters in animals treated with alcohol only
b-actin as endocontrol, the expression levels of PPARc
compared to mice that received both alcohol and puerarin
mRNA and osteocalcin mRNA were analyzed by reverse
and to mice that received no treatments. In computing
transcription polymerase chain reaction. Bone marrow cells
sample size we assumed that each outcome represented a
from all three groups were maintained in media for
continuous random variable, taking on a value between 0
10 days, treated 6 days with 0.09 mol/L ethanol, 0.09 mol/
and 100%. We assumed the outcomes of study groups were
L ethanol and 0.01 mg/ml puerarin, or neither. The cells
distributed normally with means, and had a common
were digested with solution of 0.05% trypsin/0.02%
standard deviation (r) of 9.49. In our computations we also
EDTA, transferred into a centrifuge tube, and centrifuged
assumed a Type I error rate (a) of 0.05, and power (1-b) of
at 1000 rpm for 10 minutes. The supernatants were
0.80 to detect a 30% change in the parameters. Based on
removed. The remaining cells were lysed. Total RNA was
these specifications we estimated that a total of 216 mice
isolated from cells using a Flash UNIQ-10 Spin Column
were required with an estimated drop of rate of 10%, with
Total RNA isolation kit (Sino-America, Luoyang, China).
mice being randomly assigned in equal number to the three
Animals from each group were sacrificed at 6 months after
study groups. Data are presented as mean ± standard
treatment. The femoral head specimens were placed into
deviation (SD). Means of cholesterol, triglyceride, ALP,
1 mL TRIzol solution, ground completely, and rested for
largest fat cell size, and gene expression were compared
5 minutes. Total RNA was isolated from ground femoral
using one-way ANOVA, followed by the SNK multiple
head using TRIzol methods. The expression levels of adi-
comparison procedure. Rates of empty osteocyte lacunae
pogenic transcription factor PPARc (peroxisome
were compared using the multisample Kruskal-Wallis test.
proliferator-activated receptor-c) mRNA and osteocalcin
mRNA in cells and in animals were examined by reverse
transcription polymerase chain reaction. The PPARc and
Results
osteocalcin primer were designed by the biomolecular
research center of University of Virginia. The b-actin pri-
In Vitro Study
mer was designed by Jikang Biotechnology Ltd. Co. of
Shanghai. All primers were synthesized by Dingan Bio-
Puerarin inhibited adipogenesis while maintaining osteo-
technology Ltd. Co. of Shanghai. The sequences were:
genesis. In MSCs treated with alcohol for 2 weeks,
PPARc: cytoplasmic lipid droplets were observed under the inver-
Forward primer-50 CTGGCCTCCCTGATGAATAA30 ted phase-contrast microscope. The size and number of
Reverse primer-50 GGCGGTCTCCACTGAGAATA30 fatty droplets increased with longer duration of culture.

123
1062 Wang et al. Clinical Orthopaedics and Related Research

Few fatty drops were observed in the experimental and
control group. At Day 21, staining with Sudan III showed a
number of adipocytes in the MSCs of the model group
were filled with reddish-orange lipid droplets (Fig. 1A),
while fewer triglyceride vesicles appeared in the MSCs of
the experimental group (Fig. 1B) and the control group
(Fig. 1C). Adipocytes in the model group were 8.9-times
(p \ 0.001) and 15.6-times (p \ 0.001) of that in the
experimental group and control group respectively. The
levels of triglycerides in the experimental group were
decreased (p \ 0.001) compared to the model group. As
compared to the model group, the levels of ALP in the
experimental group and control group were increased
(p \ 0.001). The levels of osteocalcin in the experimental
group and control group were 2.2- and 5.29- times of that
in the model group respectively (p \ 0.001). No differ-
ences (p [ 0.05) were seen between the experimental and
control groups (Table 1).
In Vivo Study
Similar to the in vitro results, puerarin inhibited adipo-
genesis while maintaining a relative normal level of lipid
metabolism and osteogenesis, and thus prevented osteo-
necrosis. After 6 months of treatment, levels of total
cholesterol in serum in mice treated with alcohol was
increased (p \ 0.01) as well as the triglyceride level
(p \ 0.001) compared to the experimental and control
groups. In contrast, ALP values were decreased (p \ 0.01)
(Table 2). Fatty livers were found in animals that were
treated with alcohol but not seen in the experimental or
control group mice. Marrow and bone necrosis, diminished
hematopoiesis, increased fat, and an increased percentage
of empty osteocyte lacunae occurred in the subchondral
region of the femoral head 6 months after treatment with
alcohol (Fig. 2A–C). But no such changes were observed
in the experimental group. While the percentage of empty

Fig. 1A–C A number of red-

dish-orange triglyceride vesicles
in the bone marrow stromal cells
of (A) the model group treated
with alcohol only for 21 days,
while fewer triglyceride vesicles
appear in the MSCs of (B) the
experimental group treated with
both alcohol and puerarin and in
(C) the control group (Sudan III
stain, original magnification,
9250).

Table 1. Changes of triglyceride, ALP activity, and osteocalcin in MSCs of mice

Groups Adipocytes Triglyceride ALP activity Osteocalcin
(lg/well) (U/100 mg protein) (U/100 mg protein)

Model 319.17 ± 19.92 11.55 ± 4.42 29.02 ± 13.37 4.95 ± 2.31

Experimental 335.92 ± 23.77 4.15 ± 1.92 57.06 ± 17.73 11.11 ± 4.57
Control 20.42 ± 12.15 3.42 ± 1.60 67.08 ± 18.64 13.43 ± 5.29
MSCs = marrow mesenchymal stromal cells. Analysis of variance: p \ 0.001, model versus experimental or control; p [ 0.05, experimental
versus control. Model: cells treated with alcohol only. Experimental: cells treated with both alcohol and puerarin. Control: no treatment.

123
Volume 466, Number 5, May 2008 Alcohol-induced Osteonecrosis Prevention 1063

Table 2. Changes of cholesterol, triglyceride, and ALP activity Table 3. Changes of empty osteocyte lacuna in the femoral heads of
levels in serum of mice 6 months after treatment mice
Groups Cholesterol Triglyceride ALP activity Time Empty osteocyte lacuna (%)
(mmol/L) (mmol/L) (International Units) (months)
Model group Experimental group Control group
Model 6.39 ± 0.49c 1.01 ± 0.15c 161.6 ± 32.44b
4 11.2 ± 3.2a 10.5 ± 1.8a 10.8 ± 2.3
Experimental 3.19 ± 0.11a 0.71 ± 0.13a 196.5 ± 31.52a b
6 13.5 ± 1.6 9.8 ± 2.2a 10.3 ± 2.7
Control 2.83 ± 0.36 0.68 ± 0.22 203.4 ± 22.83
8 15.8 ± 3.4c 10.7 ± 3.1a 11.1 ± 2.9
ALP = alkaline phosphatase. Analysis of variance = a: p [ 0.05, b: 10 19.5 ± 4.1c 11.6 ± 3.1a 12.0 ± 3.1
p \ 0.05, c: p \ 0.01. Model: animals treated with alcohol only.
Experimental: animals treated with both alcohol and puerarin. Con- Analysis of variance = a: p [ 0.05, b: p \ 0.05, c: p \ 0.01.
trol: no treatment.

osteocyte lacunae increased (p \ 0.05) in animals treated Table 4. Changes of largest fat cell diameter in the femoral heads of
mice
with alcohol, the percentage of empty osteocyte lacunae in
animals treated with both alcohol and puerarin did not Time Largest fat cell diameter (lm)
(months)
increase (p [ 0.05) compared to the control group Model group Experimental group Control group
(Table 3). The average diameter of the largest fat cells in
4 38.69 ± 4.14a 38.89 ± 2.25a 38.65 ± 3.26
the model group was increased (p \ 0.05) compared to the b
6 40.02 ± 3.25 39.15 ± 3.67a 38.51 ± 3.09
other two groups (Table 4).
8 42.67 ± 2.66c 38.76 ± 3.41a 39.12 ± 2.85
10 45.38 ± 3.02c 39.61 ± 3.94a 40.13 ± 2.63
Gene Expression Analysis of variance = a: p [ 0.05, b: p \ 0.05, c: p \ 0.01.

The expression of PPARc mRNA in the cells or animals of
both the experimental group and control group was lower
(p \ 0.05) than that in the cells or animals of model group,
and there was no difference (p [ 0.05) between the
experimental and control groups. The expression of
osteocalcin mRNA in the cells or animals of both the
experimental group and control group was higher
(p \ 0.01) than that in the cells or animals of the model
groups, and there were no differences (p [ 0.05) between
the experimental and control groups (Figs. 3, 4, 5).

Fig. 2A–C Fat cell hypertrophy, empty osteocyte lacunae, bone and and no marrow necrosis were found in (B) the experimental group
marrow necrosis in subchondral area of the femoral head noticed in treated with both alcohol and puerarin, compared to normal in (C) the
(A) the model group treated with alcohol only, while less adipocytes control group (original magnification 9100).

123
1064 Wang et al.
Fig. 3 Expression of PPARc mRNA in the cells of both the
experimental group and control group was lower than that in the
cells of the model group, and there was no difference between
the experimental and control group.
Fig. 4 Expression of osteocalcin mRNA in the cells of both
experimental group and control group was higher (p \ 0.01) than
of the model group, and there was no difference (p [ 0.05) between
the experimental and control group.
Discussion
Although a number of studies document alcoholism in 10%
to 74% of patients with nontraumatic ON of the femoral
head, the mechanisms of alcohol-induced ON remain
unknown and presently there is no effective treatment for
the disease [1, 5, 16–19, 21, 23, 36, 45, 47]. Hypertrophy
and proliferation of fat cells, diminished hematopoiesis,
lipid deposition in osteocytes, fatty degeneration of
osteocytes, marrow necrosis, and thinner and sparse tra-
beculae are histopathologic changes occurring in the early
stages of steroid- and alcohol-induced ON of the femoral
head [9, 19, 22, 23, 36, 47, 50, 57]. Several hypotheses
have been proposed in the literature, including the
increased size and number of fatty cells, increased
123
Clinical Orthopaedics and Related Research
intraosseous pressure, fatty degeneration of osteocytes, fat
embolism, and extraosseous arterial occlusion due to
abnormal changes in histologic features, hemodynamics,
metabolism, and biochemical features within the femoral
head [1, 10, 17, 23, 24, 35, 39, 48]. Studies on primary and
cloned MSCs have demonstrated differentiation into a
large number of adipocytes increased while levels of ALP
activity and osteocalcin decreased with longer durations of
exposure and with higher concentrations of steroid or
ethanol [8, 10, 28, 56]. These findings indicated steroids or
alcohol can directly induce adipogenesis and reduce
osteogenesis in bone marrow stroma, and produce intra-
cellular lipid deposits leading to death of osteocytes, which
may be associated with the development of ON [20, 23, 27,
44], especially in patients with long-term and excessive use
of steroids or alcohol. Puerarin, a Chinese herbal medicine,
has antioxidative and antithrombotic effects. We therefore
hypothesized puerarin inhibits alcohol-induced adipocytic
differentiation of MSCs and fatty degeneration of osteo-
cytes and may be able to prevent alcohol-induced
osteonecrosis. We further hypothesized puerarin would
reduce expression of PPARc, a transcription factor
important in adipogenesis.
A major limitation of the study is that while marrow
adipocyte proliferation and necrosis were observed, no
femoral head collapse was noted during the study. These
pathologic changes simulate early stage human form of ON
but do not reflect the late stage of ON. Therefore, even if
the treatment is effective in this animal model, it may not
be useful in humans.
At present, there are no effective treatments for early-
stage ON. To some extent, this is because the pathogenesis
of ON is unknown. However, recent advances in cell and
molecular biology have enabled researchers to identify
some of the key factors contributing to the development of
ON. It is now well known MSCs can differentiate into
osteoblasts, adipocytes, chondrocytes, and even myoblasts
[3, 40]. There is evidence for a considerable degree of
plasticity in the differentiation of these stromal cell lines
[14, 52]. Most MSCs can differentiate into osteoblasts and
osteocytes, and a few into adipocytes, under normal culture
conditions in vitro [8, 20, 56]. Alcohol can induce the
differentiation of MSCs into adipocytes and inhibit their
osteogenic differentiation, which may be a triggering step
causing the onset of ON and the pathogenesis of alcohol-
induced ON [10, 20, 28, 50]. Because the differentiation of
MSCs into adipocytes might be one of the most important
reasons leading to steroid- or alcohol-induced ON, the
proper therapy should be directed at inhibiting the differ-
entiation of MSCs into fat cells. Lovastatin prevents
steroid-induced adipogenesis and onset of ON [9, 29, 30].
However, few studies are associated with treatment of
alcohol-induced ON.
Volume 466, Number 5, May 2008
Fig. 5 Electrophoresis of
PPARc and osteocalcin mRNA
in the femoral head of mice
shows expression of PPARc
mRNA in the tissue of both
(B) the experimental group and
(C) the control group was lower
(p \ 0.05) than that in the tissue
of (A) the model group. Expres-
sion of osteocalcin mRNA in
the tissue of both the experi-
mental group and control group
was higher (p \ 0.01) than the
model group, and there was no
difference (p [ 0.05) between
the experimental and control
group.
Consistent with previous observations [10, 20, 24, 28,
46, 50], our data suggest alcohol can induce adipogenesis
both in vitro and in vivo, a change found in early stage
ON. Adipogenic differentiation is a complex process
regulated by many factors. PPARc is closely involved in
the induction of fatty differention [29, 33]. It is an
adipogenic regulator and belongs to the nucleus hormone
receptor subgroup. Its activity is regulated by ligands
like most members of nucleus receptor families [33].
PPARc mRNA appears before activation of many other
adipocyte genes in adipogenic differentiation of 3T3-L1
and 3T3-F442A cells [43]. Our study suggests the high
level of expression of PPARc occurred only in model
groups treated with alcohol, indicating alcohol may be a
factor for these bone marrow cells to change from a
primarily osteogenic nature to adipogenic phenotype.
Although the regulation of adipogenesis may involve
complex mechanisms, the data suggest fat cell hyper-
trophy and hyperplasia in bone marrow may be a direct
result of treatment with alcohol. Other mechanisms may
also be involved in the development of alcohol-induced
ON. Consumption of alcohol can elevate serum lipid
peroxides and reduce superoxide dismutase activity [25,
26]. Lipid peroxidation may cause cytomembrane injury
and induce degeneration of arterioles and arterioloscle-
rosis, which eventually leads to ischemia in the target
organ, including the femoral head. Furthermore, the
direct cytotoxicity of lipid peroxidation, caused by
alcohol and its metabolites, might further insult ischemic
osteocytes, resulting in an irreversible state of injury
leading to cell death and finally ON. In addition, alcohol
Alcohol-induced Osteonecrosis Prevention 1065
consumption produces a fatty liver and alcoholic liver
injury [17, 31, 37]. Serum triglyceride and cholesterol
levels increase, fat degeneration is induced, and liver
cells are injured. These changes, as our data indicate,
may produce fatty vesicles in the circulation, leading to
embolism and ischemia in the subchondral region of the
femoral head.
Puerarin inhibited alcohol-induced osteonecrosis in our
preliminary studies. The traditional Chinese medicine
pueraria has been used to treat alcoholism for thousands of
years. Recently, many observations demonstrate puerarin
extracted from pueraria is antioxidative, can prevent inju-
ries to cells due to lipid peroxidation, and can maintain the
stability of the cell membrane structure by removing the
oxyradicals from tissues and cells [4, 6, 12, 13]. The pro-
tective effects of puerarin on cells are also related to
calcium ion rivalry and b-receptor blockage [54]. In
addition, puerarin can decompose acetaldehyde, suppress
alcohol absorption, accelerate metabolism and excretion of
alcohol in blood, decrease blood viscosity, and inhibit
blood platelet aggregation [7, 15, 34, 42]. Thus, these
effects of puerarin can prevent the development of a fatty
liver, hyperlipidemia, and capillary vessel wall injuries,
and hypercoagulation. Most importantly, the data suggest
puerarin can decrease alcohol-induced adipogenic gene
expression and thus diminish fat marrow changes, while
maintaining the cell’s osteogenic differentiation. The
inhibitive effects of puerarin on bone-marrow adipogenesis
and its concomitant enhancement of osteogenesis may
provide a novel approach to the prevention and treatment
of alcohol-induced ON.
123
1066 Wang et al.
References
1. Arlet J. Nontraumatic avascular necrosis of the femoral head.
Past, present, and future. Clin Orthop Relat Res. 1992;277:12–21.
2. Benlhabib E, Baker JI, Keyler DE, Singh AK. Effects of purified
puerarin on voluntary alcohol intake and alcohol withdrawal
symptoms in P rats receiving free access to water and alcohol.
J Med Food. 2004;7:180–186.
3. Bianco P, Gehron Robey P. Marrow stromal stem cells. J Clin
Invest. 2000;105:1663–1668.
4. Cervellati R, Renzulli C, Guerra MC, Speroni E. Evaluation of
antioxidant activity of some natural polyphenolic compounds
using the Briggs-Rauscher reaction method. J Agric Food Chem.
2002;50:7504–7509.
5. Chakkalakal DA. Alcohol-induced bone loss and deficient bone
repair. Alcohol Clin Exp Res. 2005;29:2077–2090.
6. Cherdshewasart W, Sutjit W. Correlation of antioxidant activity
and major isoflavonoid contents of the phytoestrogen-rich Pue-
raria mirifica and Pueraria lobata tubers. Phytomedicine. 2007
Sept 21; [Epub ahead of print].
7. Choo MK, Park EK, Yoon HK, Kim DH. Antithrombotic and
antiallergic activities of daidzein, a metabolite of puerarin and
daidzin produced by human intestinal microflora. Biol Pharm
Bull. 2002;25:1328–1332.
8. Cui Q, Wang GJ, Balian G. Steroid-induced adipogenesis in a
pluripotential cell line from bone marrow. J Bone Joint Surg Am.
1997;79:1054–1063.
9. Cui Q, Wang GJ, Su CC, Balian G. The Otto Aufranc Award.
Lovastatin prevents steroid induced adipogenesis and osteone-
crosis. Clin Orthop Relat Res. 1997;344:8–19.
10. Cui Q, Wang Y, Saleh KJ, Wang GJ, Balian G. Alcohol-induced
adipogenesis in a cloned bone-marrow stem cell. J Bone Joint
Surg Am. 2006;88 Suppl 3:148–154.
11. Fan LL, Sun LH, Li J, Yue XH, Yu HX, Wang SY. The pro-
tective effect of puerarin against myocardial reperfusion injury.
Study on cardiac function. Chin Med J (Engl). 1992;105:11–17.
12. Guerra MC, Speroni E, Broccoli M, Cangini M, Pasini P,
Minghett A, Crespi-Perellino N, Mirasoli M, Cantelli-Forti G,
Paolini M. Comparison between chinese medical herb Pueraria
lobata crude extract and its main isoflavone puerarin antioxidant
properties and effects on rat liver CYP-catalysed drug metabo-
lism. Life Sci. 2000;67:2997–3006.
13. Han RM, Tian YX, Becker EM, Andersen ML, Zhang JP,
Skibsted LH. Puerarin and conjugate bases as radical scavengers
and antioxidants: molecular mechanism and synergism with beta-
carotene. J Agric Food Chem. 2007;55:2384–2391.
14. Herzog EL, Chai L, Krause DS. Plasticity of marrow-derived
stem cells. Blood. 2003;102:3483–3493.
15. Heyman GM, Keung WM, Vallee BL. Daidzin decreases ethanol
consumption in rats. Alcohol Clin Exp Res. 1996;20:1083–1087.
16. Hirota Y, Hirohata T, Fukuda K, Mori M, Yanagawa H, Ohno Y,
Sugioka Y. Association of alcohol intake, cigarette smoking, and
occupational status with the risk of idiopathic osteonecrosis of the
femoral head. Am J Epidemiol. 1993;137:530–538.
17. Hungerford DS, Lennox DW. The importance of increased
intraosseous pressure in the development of osteonecrosis of the
femoral head: implications for treatment. Orthop Clin North Am.
1985;16:635–654.
18. Hungerford DS, Zizic TM. Alcoholism associated ischemic
necrosis of the femoral head. Early diagnosis and treatment. Clin
Orthop Relat Res. 1978:144–153.
19. Jacobs B. Epidemiology of traumatic and nontraumatic osteone-
crosis. Clin Orthop Relat Res. 1978;130:51–67.
20. Jin WS, Tan YY, Chen YG, Wang Y. Determination of puerarin,
daidzin and daidzein in root of Pueraria lobata of different origin
123
Clinical Orthopaedics and Related Research
by HPLC [in Chinese]. Zhongguo Zhong Yao Za Zhi.
2003;28:49–51.
21. Jones JP Jr. Intravascular coagulation and osteonecrosis. Clin
Orthop Relat Res. 1992;277:41–53.
22. Jones JP Jr. Fat embolism, intravascular coagulation, and osteo-
necrosis. Clin Orthop Relat Res. 1993;292:294–308.
23. Jones JP Jr. Alcoholism, hypercortisonism, fat embolism and
osseous avascular necrosis. 1971. Clin Orthop Relat Res.
2001;393:4–12.
24. Kawai K, Tamaki A, Hirohata K. Steroid-induced accumulation
of lipid in the osteocytes of the rabbit femoral head. A histo-
chemical and electron microscopic study. J Bone Joint Surg Am.
1985;67:755–763.
25. Keung WM. Biochemical studies of a new class of alcohol
dehydrogenase inhibitors from Radix puerariae. Alcohol Clin Exp
Res. 1993;17:1254–1260.
26. Keung WM, Vallee BL. Daidzin and its antidipsotropic analogs
inhibit serotonin and dopamine metabolism in isolated mito-
chondria. Proc Natl Acad Sci U S A. 1998;95:2198–2203.
27. Leach RE, Baskies A. Alcoholism and its effect on the human
hip. Clin Orthop Relat Res. 1973;90:95–99.
28. Li J, Wang Y, Li Y, Xu J, Xiong T. Alcohol induced regulation of
adipogenic and osteogenic gene expression of marrow stromal
cells. Chin J Orthop. 2003;23:493–495.
29. Li X, Cui Q, Kao C, Wang GJ, Balian G. Lovastatin inhibits adi-
pogenic and stimulates osteogenic differentiation by suppressing
PPARgamma2 and increasing Cbfa1/Runx2 expression in bone
marrow mesenchymal cell cultures. Bone. 2003;33:652–659.
30. Li X, Jin L, Cui Q, Wang GJ, Balian G. Steroid effects on
osteogenesis through mesenchymal cell gene expression. Osteo-
poros Int. 2005;16:101–108.
31. Lieber CS, Savolainen M. Ethanol and lipids. Alcohol Clin Exp
Res. 1984;8:409–423.
32. Lin RC, Li TK. Effects of isoflavones on alcohol pharmacoki-
netics and alcohol-drinking behavior in rats. Am J Clin Nutr.
1998;68:1512S–1515S.
33. Lowell BB. PPARgamma: an essential regulator of adipogenesis
and modulator of fat cell function. Cell. 1999;99:239–242.
34. Luo ZR, Zheng B. Effect of Puerarin on platelet activating factors
CD63 and CD62P, plasminogen activator inhibitor and C-reac-
tive protein in patients with unstable angia pectoris [in Chinese].
Zhongguo Zhong Xi Yi Jie He Za Zhi. 2001;21:31–33.
35. Mankin HJ. Nontraumatic necrosis of bone (osteonecrosis).
N Engl J Med. 1992;326:1473–1479.
36. Mont MA, Jones LC, Hungerford DS. Nontraumatic osteone-
crosis of the femoral head: ten years later. J Bone Joint Surg Am.
2006;88:1117–1132.
37. Niemela O, Parkkila S, Pasanen M, Iimuro Y, Bradford B,
Thurman RG. Early alcoholic liver injury: formation of protein
adducts with acetaldehyde and lipid peroxidation products, and
expression of CYP2E1 and CYP3A. Alcohol Clin Exp Res.
1998;22:2118–2124.
38. Overstreet DH, Lee YW, Rezvani AH, Pei YH, Criswell HE,
Janowsky DS. Suppression of alcohol intake after administration
of the Chinese herbal medicine, NPI-028, and its derivatives.
Alcohol Clin Exp Res. 1996;20:221–227.
39. Prasain JK, Jones K, Brissie N, Moore R, Wyss JM, Barnes S.
Identification of puerarin and its metabolites in rats by liquid
chromatography-tandem mass spectrometry. J Agric Food Chem.
2004;52:3708–3712.
40. Prockop DJ. Marrow stromal cells as stem cells for nonhemato-
poietic tissues. Science. 1997;276:71–74.
41. Rezvani AH, Overstreet DH, Perfumi M, Massi M. Plant deriv-
atives in the treatment of alcohol dependency. Pharmacol
Biochem Behav. 2003;75:593–606.
Volume 466, Number 5, May 2008
42. Shi RL, Zhang JJ. Protective effect of puerarin on vascular
endothelial cell apoptosis induced by chemical hypoxia in vitro
[in Chinese]. Yao Xue Xue Bao. 2003;38:103–107.
43. Smas CM, Sul HS. Control of adipocyte differentiation. Biochem
J. 1995;309 (Pt 3):697–710.
44. Solomon L. Drug-induced arthropathy and necrosis of the fem-
oral head. J Bone Joint Surg Br. 1973;55:246–261.
45. Solomon L. Mechanisms of idiopathic osteonecrosis. Orthop Clin
North Am. 1985;16:655–667.
46. Suh KT, Kim SW, Roh HL, Youn MS, Jung JS. Decreased
osteogenic differentiation of mesenchymal stem cells in alcohol-
induced osteonecrosis. Clin Orthop Relat Res. 2005;431:220–
225.
47. Turner RT, Evans GL, Zhang M, Sibonga JD. Effects of
parathyroid hormone on bone formation in a rat model
for chronic alcohol abuse. Alcohol Clin Exp Res. 2001;25:
667–671.
48. Wang GJ, Dughman SS, Reger SI, Stamp WG. The effect of core
decompression on femoral head blood flow in steroid-induced
avascular necrosis of the femoral head. J Bone Joint Surg Am.
1985;67:121–124.
49. Wang GJ, Sweet DE, Reger SI, Thompson RC. Fat-cell changes
as a mechanism of avascular necrosis of the femoral head in
cortisone-treated rabbits. J Bone Joint Surg Am. 1977;59:729–
735.
Alcohol-induced Osteonecrosis Prevention 1067
50. Wang Y, Mao K, Li Y, Xu Z. Alcohol induced animal model
of osteonecrosis of the femoral head. Chin Exper Surg. 1998;
15:182–183.
51. Warner JJ, Philip JH, Brodsky GL, Thornhill TS. Studies of
nontraumatic osteonecrosis. Manometric and histologic studies of
the femoral head after chronic steroid treatment: an experimental
study in rabbits. Clin Orthop Relat Res. 1987;225:128–140.
52. Wolf NS, Penn PE, Rao D, McKee MD. Intraclonal plasticity for
bone, smooth muscle, and adipocyte lineages in bone marrow
stroma fibroblastoid cells. Exp Cell Res. 2003;290:346–357.
53. Wu L, Qiao H, Li Y, Li L. Protective roles of puerarin and
Danshensu on acute ischemic myocardial injury in rats. Phyto-
medicine. 2007;14:652–658.
54. Xiao L, Luo W, Su H, Cheng X, Wu Q, Ye C. The influence of
puerarin on the NO and ACE activity secreted by blood vessel
endothelial cells cultured under hyperpressure condition. New J
Traditional Chin Med. 2002;42:31–33.
55. Xu X, Zhang S, Zhang L, Yan W, Zheng X. The Neuroprotection
of puerarin against cerebral ischemia is associated with the pre-
vention of apoptosis in rats. Planta Med. 2005;71:585–591.
56. Yin L, Li YB, Wang YS. Dexamethasone-induced adipogenesis
in primary marrow stromal cell cultures: mechanism of steroid-
induced osteonecrosis. Chin Med J (Engl). 2006;119:581–588.
57. Zhang W, Yu S, Zheng L. Experimental study on IL-1 regulating
differentiation of osteoblast. Chin J Stomatol. 1996;31:288–291.
123
Clin Orthop Relat Res (2008) 466:1068–1073
DOI 10.1007/s11999-008-0182-y

SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS

Electromagnetic Fields
A Novel Prophylaxis for Steroid-induced Osteonecrosis

Masashi Ishida MD, Mikihiro Fujioka MD, PhD,

Kenji A. Takahashi MD, PhD, Yuji Arai MD, PhD,
Toshikazu Kubo MD, PhD

Published online: 19 March 2008

Ó The Association of Bone and Joint Surgeons 2008

Abstract Establishing a means to prevent osteonecrosis
after corticosteroid administration is an important theme.
We asked whether pulsed electromagnetic field stimulation,
a noninvasive treatment, could prevent osteonecrosis.
Ninety rabbits were divided into four treatment groups: (1)
exposure of 10 hours per day to electromagnetic stimulation
for 1 week, followed by injection of methylprednisolone
(20 mg/kg), and exposure of 10 hours per day to electro-
magnetism for a further 4 weeks (n = 40); (2) methyl-
prednisolone injection only (n = 40); (3) no treatment
(n = 5); and (4) exposure of 10 hours per day to electro-
magnetism for 5 weeks (n = 5). After 5 weeks, we
harvested and histologically examined femurs bilaterally.
The frequency of osteonecrosis was lower in the steroid-
electromagnetism group (15/40) than in the steroid-only
group (26/40). No necrotic lesions were found in the two
control groups. We observed no clear effects of electro-
magnetism on the number, location, extent, and repair of
Each author certifies that he has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
One or more of the authors (MF, TK) received funding from the
Japanese Investigation Committee for Osteonecrosis of the Femoral
Head, under the auspices of the Ministry of Health, Labor and
Welfare of Japan; and one of the authors (MF) from the Hip Joint
Foundation of Japan, Inc.
Each author certifies that his institution has approved the animal
protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
M. Ishida, M. Fujioka (&), K. A. Takahashi, Y. Arai, T. Kubo
Department of Orthopaedics, Graduate School of Medical
Science, Kyoto Prefectural University of Medicine,
465 Kawaramachi-Hirokoji, Kamigyo-ku, Kyoto 602-8566,
Japan
e-mail: fujioka@koto.kpu-m.ac.jp
necrotic lesions and intramedullary fat cell size in affected
rabbits. Pulsed electromagnetic field stimulation reportedly
augments angiogenesis factors and dilates blood vessels;
these effects may lower the frequency of osteonecrosis.
Exposure to pulsed electromagnetic field stimulation before
corticosteroid administration could be an effective means to
reduce the risk of osteonecrosis.
Introduction
Osteonecrosis (ON) has been recognized as a complication
of high-dose corticosteroid administration. If the necrotic
area is large, spontaneous healing is unlikely, and surgery is
usually necessary if the necrotic area collapses. Arthroplasty
effectively improves quality of life for patients, but the costs
and risks of surgery are not insignificant; furthermore, an
acceptably durable prosthesis is not yet available for the
relatively young population in which this complication
typically occurs [4, 5, 20, 21]. Therefore, ON must be pre-
vented in patients who require high-dose corticosteroid
therapy, but there is no established prophylactic measure.
Ischemia in the bone has long been believed one of the
causes of ON after corticosteroid administration [7, 8, 28].
The pathogenesis of the ischemic necrosis remains elusive
but is thought to involve hypercoagulable conditions [8],
vasoconstriction [7], and disorders of lipid metabolism [28]
after corticosteroid administration. Therefore, prevention
of bone ischemia might also prevent steroid-induced ON.
We considered pulsed electromagnetic field stimulation
could be a preventive therapy because it promotes angio-
genesis [11, 19, 27] and dilates blood vessels [23]. We
speculated it might also influence the enlargement of
intramedullary fat cells since that is reportedly a potential
mechanism in bone necrosis [14, 15, 28].

123
Volume 466, Number 5, May 2008
We first asked whether electromagnetism would reduce
the frequency of ON. We then asked whether electro-
magnetism reduces the severity (number of lesions) of ON.
Finally, we asked whether electromagnetism influenced the
enlargement of intramedullary fat cells.
Materials and Methods
We divided 90 male Japanese white rabbits (Kitayama
Labes Co, Ltd, Nagano, Japan) (body weight, 3.2–4.3 kg;
age, 28–32 weeks) into four groups: (1) a steroid-electro-
magnetism group (n = 40), which received 1 week of
exposure to pulsed electromagnetic field stimulation
(10 hours/day) over the gluteofemoral area, followed by
methylprednisolone (20 mg/kg body weight) intramuscu-
larly into the gluteus medius muscle, and 4 subsequent
weeks of exposure to electromagnetism (10 hours/day;
total exposure period, 5 weeks); (2) a steroid-only group
(n = 40), which received an injection of methylpredniso-
lone (20 mg/kg body weight) intramuscularly into the
gluteus medius muscle after 1 week; (3) an untreated group
(n = 5); and (4) an electromagnetism-only group (n = 5),
which received 5 weeks of electromagnetic exposure
(10 hours/day). Steroid-induced ON was produced by an
intramuscular injection of methylprednisolone (20 mg/kg
body weight). This experimental model induces ON in the
humeral and femoral bones of approximately 70% of
treated rabbits [13–16]. The animals were housed at the
Animal Center of Kyoto Prefectural University of Medi-
cine, fed nutritionally adequate food daily, and had free
access to clean drinking water. This study followed the
guidelines of the Kyoto Prefectural University of Medicine
Animal Care and Use Committee fully.
A study using the same rabbit model reported a reduc-
tion in the frequency of ON to 30% to 40% with an
anticoagulant drug (warfarin) or a lipid-lowering agent
(probucol) [16]. The sample size calculation was based on
an estimated effect size of 42% for electromagnetism with
steroid administration and 70% for steroid administration
alone. With an alpha level of 0.05 and a test power of 0.8,
the calculated sample size (ie, the minimum number of
rabbits required) was 39 for each group.
The protocol of pulsed electromagnetic field stimulation
in this study was chosen according to that reported effec-
tive for bone fracture healing [2]. One study reported
acceleration of angiogenesis within 1 week of commencing
electromagnetic therapy [27]. Another study reported
ischemia in the bone occurs within 5 to 6 days after cor-
ticosteroid administration [10]. Therefore, we began pulsed
electromagnetic field stimulation 1 week before the
administration of corticosteroid. The pulsed electromag-
netic fields were generated by an EBI Bone Healing
Electromagnetic Fields and Osteonecrosis 1069
Fig. 1 The pulsed electromagnetic field was generated using the EBI
Bone Healing System1. It consists of a portable pulse-shaping circuit
(right) and a coil that is affixed to the cast surface (left).
System1 (Biomet Osteobiologics, Parsippany, NJ), which
delivered uniform time-varying fields consisting of asym-
metric 4.5-ms pulses repeated at 15 Hz (Fig. 1). A coil was
installed in each cage to generate an electromagnetic field
on the gluteofemoral area of the rabbit. During the elec-
tromagnetic treatment, the rabbits had free access to water
and food.
All rabbits were euthanized after 5 weeks via intrave-
nous injection of a large dose of pentobarbital sodium. We
excised bilateral femurs of all rabbits and obtained a tissue
section at the proximal third and distal third of the femur in
the coronal plane (total four sections). The specimens were
stained with hematoxylin and eosin.
Osteonecrosis and all other histologic measures were
assessed blindly by two independent authors (MI, MF). A
positive diagnosis was based on the presence of empty
lacunae or pyknotic nuclei of osteocytes in the bone tra-
beculae, accompanied by surrounding bone marrow cell
necrosis. Only bone marrow cell necrosis showing tissue
debris consisting of both hematopoietic cell necrosis and
fat cell necrosis and included no bone trabeculae qualified
as ON. Lesions consisting of only empty lacunae in normal
bone trabeculae and/or fat cell necrosis without bone
marrow cell necrosis were excluded from the assessment of
ON in this study. If the diagnoses of the two examiners
differed, consensus was reached after discussing the his-
tologic findings without unblinding of the group. Rabbits
with at least one osteonecrotic lesion in the four areas
examined were considered to have ON [29]. The effect of
electromagnetic treatment on the prevention of steroid-
induced ON was evaluated as the difference between the
proportions of rabbits that acquired ON in the steroid-only
group and the steroid-electromagnetism group.
123
1070 Ishida et al.
In this rabbit model, ON does not occur later than
4 weeks after the methylprednisolone injection [9], and the
repair process of necrosis is incomplete at 4 weeks [29].
Therefore, histopathologic evidence of ON is obtainable
4 weeks after the methylprednisolone injection. We histo-
logically evaluated the effects of methylprednisolone
administration and pulsed electromagnetic field stimulation
on steroid-induced ON 4 weeks after the injection. To
determine whether electromagnetism could suppress mul-
tiple occurrences of ON, the necrotic lesions in the
proximal third and distal third of the femurs of affected
rabbits were counted in each group, and the average
number was compared among the groups. The effects of
electromagnetism on the location of ON were assessed by
counting and comparing the number of lesions in the
epiphyseal area, proximal third of the femur, and distal
third of the femur. The influence of electromagnetic field
stimulation on the size of osteonecrotic lesions was
examined by measuring the affected area. The ON-affected
area was expressed as the ratio of the affected area to the
total area in the proximal third of the femur. The area was
calculated from computerized microscope images using
NIH image software (US National Institutes of Health,
Bethesda, MD). The osteonecrotic area was defined as the
area containing necrosis of osteocytes or bone marrow
cells. Areas undergoing any repair process were not
regarded as necrotic [29]. To determine the influence of
pulsed electromagnetic field stimulation on the repair
process, we examined the bone samples for the presence of
histologic features of repair, including granulation tissue,
fibrosis, and/or appositional bone formation against
necrotic tissue, and findings in the steroid-only and steroid-
electromagnetism groups were compared.
To evaluate bone marrow fat cell enlargement caused by
corticosteroid administration, the size of the fat cells was
calculated and compared between the steroid-only group
and the untreated group. The effect of pulsed electromag-
netic field stimulation on corticosteroid-induced bone
marrow fat cell enlargement was assessed by comparing fat
cell size in the steroid-electromagnetism group and the
steroid-only group. Bone marrow fat cell size was calcu-
lated as the average of the largest diameters of 25 fat cells
in each of four randomly selected fields (one field = 25 9
10-8 m2) in nonnecrotic regions in the proximal third of
the right femur using NIH imaging software [22, 28].
All values were expressed as the mean ± standard
deviation. Each variable was compared using parametric or
nonparametric methods with or without the normality of
distribution of data. The proportions of rabbits in the
steroid-only and steroid-electromagnetism groups that
acquired ON were compared nonparametrically using
the chi-square test. The number of necrotic lesions per
rabbit that acquired lesions was compared between the
123
Clinical Orthopaedics and Related Research
steroid-only and steroid-electromagnetism groups. The
difference between groups was evaluated with the Mann-
Whitney U test. The unpaired t test was used to test for a
difference between the mean area of lesions in the proximal
femur in the steroid-only group and the steroid-electro-
magnetism group. The mean diameter of 100 fat cells from
each of five rabbits from each group was compared among
the four groups. One-way analysis of variance was used to
determine the overall difference among the groups. If the
difference was significant at p \ 0.05, Scheffe’s post hoc
tests were performed to test the significance of individual
comparisons. Values with p \ 0.05 were considered
significant.
Results
Electromagnetism reduced (p = 0.01) the frequency of ON
in the steroid-electromagnetism rabbits compared to the
steroid-only rabbits (15 of 40, or 37.5% versus 26 of 40, or
65%, respectively) (Figs. 2A–C, 3A–B). None of the rab-
bits in the untreated group and the electromagnetism-only
group had features of ON.
Rabbits treated without and with electromagnetism had
similar severities of steroid-induced ON (Table 1). The
number of osteonecrotic lesions per affected rabbit was
1.1 ± 0.3 and 1.2 ± 0.4, respectively. Lesions in the
proximal third of the femur occurred in 15 of 26 and 10 of
15 affected rabbits in the steroid-only and steroid-electro-
magnetism groups, respectively; lesions occurred in the
distal third of the femur in 14 of 26 and eight of 15 affected
rabbits in each group, respectively. The epiphysis did not
display bone necrosis. The extent of ON was similar in the
two groups (2.8 ± 1.9% in the treated and 2.7 ± 1.8% in
the untreated). In the steroid-only group and the steroid-
electromagnetism group, we observed similar amounts of
fibrous and cellular-rich granulation tissue (reparative tis-
sues) and they were found on the margins of osteonecrotic
lesions. There was no appositional bone formation.
Mean bone marrow fat cell size in the steroid-only group
was larger (p \ 0.01) than that in the untreated group. We
observed no difference in fat cell size after corticosteroid
administration with or without electromagnetism (p =
0.43), ie, 61 ± 4.5 lm in the steroid-only group,
56.9 ± 5.4 lm in the steroid-electromagnetism group,
47.6 ± 3 lm in the untreated group, and 46.1 ± 3.4 lm in
the electromagnetism-only group.
Discussion
Pulsed electromagnetic field stimulation induces angio-
genesis [11, 19, 27] and vasodilatation [23], but its clinical
Volume 466, Number 5, May 2008 Electromagnetic Fields and Osteonecrosis 1071

Fig. 2A–C (A) The osteonecrotic

lesions (arrows) in this rabbit from
the steroid-only group showed
pyknosis (stain, hematoxylin and
eosin; original magnification,
9100). (B) Bone marrow cells
had necrosis and stained acido-
philic (arrowheads). The nuclei of
bone marrow cells display pykno-
sis and karyorrhexis. Some cells
went through karyolysis and
uniformly stained acidophilic
(arrows). The cellular structure of
fat cells also collapsed (stain,
hematoxylin and eosin; original
magnification, 9200). (C) Bone
cells in the bone trabeculae
showed pyknosis and empty lacu-
nae (arrows) that were associated
with necrotic changes of the sur-
rounding bone marrow cells (stain,
hematoxylin and eosin; original
magnification, 9200).

Fig. 3A–B (A) The osteonecrotic lesions (arrows) in this rabbit from marrow cells showed cytolysis and pyknosis, and there was bone
the steroid-electromagnetism group stained less intensively (stain, marrow cell necrosis (stain, hematoxylin and eosin; original magnifi-
hematoxylin and eosin; original magnification, 9100). (B) Bone cation, 9200). Nec = necrotic zone; Liv = living bone marrow tissue.

Table 1. Number, location, and extent of osteonecrosis in affected rabbits of the steroid only and steroid-electromagnetism groups
Group Number with Proximal third Distal third Ratio of osteonecrosis
osteonecrosis* of femur of femur area (%)*

Steroid only (n = 26) 1.1 ± 0.3 15 (58%) 14 (54%) 2.8 ± 1.9

Steroid-electromagnetism (n = 15) 1.2 ± 0.4 10 (67%) 8 (53%) 2.7 ± 1.8
* Values are expressed as mean ± standard deviation.

123
1072 Ishida et al.
effects on ON have not been clarified [3, 12, 24–26]. We
therefore asked whether electromagnetism would influence
the risk and severity of ON and if so whether it would
influence the size of fat cells since enlargement has been
proposed as a mechanism to induce ischemia.
The major limitation of the current study is that osteo-
necrosis in the rabbit model is different from steroid-
induced bone necrosis in humans [29]. The rabbit model is
different from human ON in that (1) bone necrosis in
rabbits frequently occurs in the metaphysis, not in the
epiphysis, and (2) the model does not lead to collapse and
the lesions resolve spontaneously. However, the patho-
physiology of ON in the rabbit model is characterized by
empty lacunae accompanied by surrounding bone marrow
cell necrosis, and that of reparative changes as granulation
tissue and appositional bone formation are thought to be
closer to steroid-induced ON in humans [29]. Therefore,
the rabbit model has been widely used in various studies [9,
10, 13–16].
Histopathologic evidence of ON is obtainable 4 weeks
after the steroid injection as described in Materials and
Methods. We examined bone necrosis at one time point,
and did not evaluate chronological effects of electromag-
netism on steroid-induced ON, therefore bone necrosis
should be examined in earlier and later stages of the
experiment in another study. We used a protocol (dosage
and frequency) for electromagnetic field stimulation used
to treat adult bone nonunion (ie, exposure to 15-Hz elec-
tromagnetic stimuli for 10 hours/day) [2], but the optimal
electrical stimulation for the augmentation of growth fac-
tors varies according to cell type [11]. Thus, the optimal
protocol of pulsed electromagnetic stimulation for the
prevention of ON should be investigated in another study.
Our data suggest the number of rabbits with ON
4 weeks after steroid administration was lower in the group
that received pulsed electromagnetic field stimulation than
the group that did not. We did not determine the mecha-
nism of the pulsed electromagnetic field benefit, but we
speculate acceleration of angiogenesis and vasodilatation
caused by the electromagnetic waves could suppress
ischemia in the bone after corticosteroid administration.
Fat cells were not affected by electromagnetism, so we
presume that is not the mechanism.
We found no influence of electromagnetism on histo-
logic factors relating to the severity of necrosis (ie, number
of osteonecrotic lesions per rabbit), size, location, and
repair level of bone necrosis. This suggests this treatment
can help prevent ON, but if it occurs, the treatment does not
reduce its severity. This same conclusion has been reached
for the prevention of steroid-induced ON with chemicals
[16].
Several researchers have shown steroid-induced ON can
be prevented by medications [6, 17, 18], and a study that
123
Clinical Orthopaedics and Related Research
used the same rabbit model as ours reported a reduction in
ON frequency to 30% to 40% [16]. Our study showed
similar suppressive effects, and no rabbits developed tissue
damage due to electromagnetism. This point also indicates
electromagnetism could be a useful prophylactic therapy
for steroid-induced ON.
Miyanishi et al. [15] reported a larger bone marrow fat
cell size in rabbits that developed ON than in those without
ON after corticosteroid administration. They hypothesized
corticosteroids become a cause of ON through enlargement
of bone marrow fat cells and an increase in intraosseous
pressure, resulting in a disturbance of intraosseous circu-
lation. Another study proposed prevention of fat cell
enlargement could also prevent steroid-induced ON [16].
We found bone marrow fat cell size was larger in rabbits
that received corticosteroids than in those that did not, but
electromagnetism did not affect bone marrow fat cell size.
Therefore, we assumed its preventative effect on steroid-
induced ON occurs via a mechanism independent of lipid
metabolism.
Our preliminary data suggest electromagnetic fields
could be a preventative method, particularly since it is
noninvasive [1, 2]. Further studies are necessary to clarify
the optimal protocol of pulsed electromagnetic fields for
the prevention of ON and to apply this treatment in clinical
trials.
Acknowledgments We thank Drs. Takuaki Yamamoto and Kenjiro
Nishida for their valuable advice and kind help on the evaluation of
histologic features.
References
1. Akai M, Hayashi K. Effect of electrical stimulation on muscu-
loskeletal systems: a meta-analysis of controlled clinical trials.
Bioelectromagnetics. 2002;23:132–143.
2. Bassett CA. Fundamental and practical aspects of therapeutic
uses of pulsed electromagnetic fields (PEMFs). Crit Rev Biomed
Eng. 1989;17:451–529.
3. Bassett CA, Schink-Ascani M, Lewis SM. Effects of pulsed
electromagnetic fields on Steinberg ratings of femoral head
osteonecrosis. Clin Orthop Relat Res. 1989;246:172–185.
4. Brinker MR, Rosenberg AG, Kull L, Galante JO. Primary total
hip arthroplasty using noncemented porous-coated femoral
components in patients with osteonecrosis of the femoral head.
J Arthroplasty. 1994;9:457–468.
5. Chiu KH, Shen WY, Ko CK, Chan KM. Osteonecrosis of
the femoral head treated with cementless total hip arthroplasty:
a comparison with other diagnoses. J Arthroplasty. 1997;12:
683–688.
6. Cui Q, Wang GJ, Su CC, Balian G. The Otto Aufranc Award:
Lovastatin prevents steroid induced adipogenesis and osteone-
crosis. Clin Orthop Relat Res. 1997;344:8–19.
7. Drescher W, Bunger MH, Weigert K, Bunger C, Hansen ES.
Methylprednisolone enhances contraction of porcine femoral
head epiphyseal arteries. Clin Orthop Relat Res. 2004;423:112–
117.
Volume 466, Number 5, May 2008
8. Drescher W, Weigert KP, Bunger MH, Ingerslev J, Bunger C,
Hansen ES. Femoral head blood flow reduction and hypercoag-
ulability under 24 h megadose steroid treatment in pigs. J Orthop
Res. 2004;22:501–508.
9. Kabata T, Kubo T, Matsumoto T, Hirata T, Fujioka M, Takahashi
KA, Yagishita S, Kobayashi M, Tomita K. Onset of steroid-
induced osteonecrosis in rabbits and its relationship to hyperli-
paemia and increased free fatty acids. Rheumatology (Oxford).
2005;44:1233–1237.
10. Kabata T, Kubo T, Matsumoto T, Nishino M, Tomita K, Katsuda
S, Horii T, Uto N, Kitajima I. Apoptotic cell death in ste-
roid induced osteonecrosis: an experimental study in rabbits.
J Rheumatol. 2000;27:2166–2171.
11. Kanno S, Oda N, Abe M, Saito S, Hori K, Handa Y, Tabayashi K,
Sato Y. Establishment of a simple and practical procedure applicable
to therapeutic angiogenesis. Circulation. 1999;99:2682–2687.
12. Massari L, Fini M, Cadossi R, Setti S, Traina GC. Biophysical
stimulation with pulsed electromagnetic fields in osteonecrosis of
the femoral head. J Bone Joint Surg Am. 2006;88:56–60.
13. Miyanishi K, Yamamoto T, Irisa T, Motomura G, Jingushi S,
Sueishi K, Iwamoto Y. Effects of different corticosteroids on the
development of osteonecrosis in rabbits. Rheumatology (Oxford).
2005;44:332–336.
14. Miyanishi K, Yamamoto T, Irisa T, Yamashita A, Jingushi S,
Noguchi Y, Iwamoto Y. A high low-density lipoprotein choles-
terol to high-density lipoprotein cholesterol ratio as a potential
risk factor for corticosteroid-induced osteonecrosis in rabbits.
Rheumatology (Oxford). 2001;40:196–201.
15. Miyanishi K, Yamamoto T, Irisa T, Yamashita A, Jingushi S,
Noguchi Y, Iwamoto Y. Bone marrow fat cell enlargement and a
rise in intraosseous pressure in steroid-treated rabbits with oste-
onecrosis. Bone. 2002;30:185–190.
16. Motomura G, Yamamoto T, Miyanishi K, Jingushi S, Iwamoto Y.
Combined effects of an anticoagulant and a lipid-lowering agent
on the prevention of steroid-induced osteonecrosis in rabbits.
Arthritis Rheum. 2004;50:3387–3391.
17. Nagasawa K, Tada Y, Koarada S, Tsukamoto H, Horiuchi T,
Yoshizawa S, Murai K, Ueda A, Haruta Y, Ohta A. Prevention of
steroid-induced osteonecrosis of femoral head in systemic lupus
erythematosus by anti-coagulant. Lupus. 2006;15:354–357.
18. Pritchett JW. Statin therapy decreases the risk of osteonecrosis in
patients receiving steroids. Clin Orthop Relat Res. 2001;386:173–178.
Electromagnetic Fields and Osteonecrosis 1073
19. Roland D, Ferder M, Kothuru R, Faierman T, Strauch B. Effects
of pulsed magnetic energy on a microsurgically transferred ves-
sel. Plast Reconstr Surg. 2000;105:1371–1374.
20. Saito S, Saito M, Nishina T, Ohzono K, Ono K. Long-term results
of total hip arthroplasty for osteonecrosis of the femoral head: a
comparison with osteoarthritis. Clin Orthop Relat Res. 1989;244:
198–207.
21. Sarmiento A, Ebramzadeh E, Gogan WJ, McKellop HA. Total
hip arthroplasty with cement: a long-term radiographic analysis
in patients who are older than fifty and younger than fifty years.
J Bone Joint Surg Am. 1990;72:1470–1476.
22. Shaw SL, Salmon ED, Quatrano RS. Digital photography for the
light microscope: results with a gated, video-rate CCD camera
and NIH-image software. Biotechniques. 1995;19:946–955.
23. Smith TL, Wong-Gibbons D, Maultsby J. Microcirculatory
effects of pulsed electromagnetic fields. J Orthop Res. 2004;22:
80–84.
24. Steinberg ME, Brighton CT, Bands RE, Hartman KM. Capacitive
coupling as an adjunctive treatment for avascular necrosis. Clin
Orthop Relat Res. 1990;261:11–18.
25. Steinberg ME, Brighton CT, Corces A, Hayken GD, Steinberg
DR, Strafford B, Tooze SE, Fallon M. Osteonecrosis of the
femoral head: results of core decompression and grafting with
and without electrical stimulation. Clin Orthop Relat Res.
1989;249:199–208.
26. Steinberg ME, Brighton CT, Steinberg DR, Tooze SE, Hayken
GD. Treatment of avascular necrosis of the femoral head by a
combination of bone grafting, decompression, and electrical
stimulation. Clin Orthop Relat Res. 1984;186:137–153.
27. Tepper OM, Callaghan MJ, Chang EI, Galiano RD, Bhatt KA,
Baharestani S, Gan J, Simon B, Hopper RA, Levine JP,
Gurtner GC. Electromagnetic fields increase in vitro and in vivo
angiogenesis through endothelial release of FGF-2. FASEB J.
2004;18:1231–1233.
28. Wang GJ, Sweet DE, Reger SI, Thompson RC. Fat-cell
changes as a mechanism of avascular necrosis of the femoral
head in cortisone-treated rabbits. J Bone Joint Surg Am. 1977;
59:729–735.
29. Yamamoto T, Irisa T, Sugioka Y, Sueishi K. Effects of pulse
methylprednisolone on bone and marrow tissues: corticosteroid-
induced osteonecrosis in rabbits. Arthritis Rheum. 1997;40:
2055–2064.
123
Clin Orthop Relat Res (2008) 466:1074–1080
DOI 10.1007/s11999-008-0179-6
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Extent of Osteonecrosis on MRI Predicts Humeral Head Collapse
Takashi Sakai MD, PhD, Nobuhiko Sugano MD, PhD,
Takashi Nishii MD, PhD, Takehito Hananouchi MD,
Hideki Yoshikawa MD, PhD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Although MRI is useful for predicting pro-
gression of osteonecrosis (ON) of the femoral head or
femoral condyle, predicting outcome of atraumatic osteo-
necrosis of the humeral head using MRI has not been
previously examined. We asked whether the prognosis was
related to the extent and location of necrotic lesions on
MRI. We investigated 46 radiographically noncollapsed
humeral heads in 27 patients, 24 steroid-related and three
alcohol-related, using MRI and serial radiographs. The
minimum followup was 24 months (mean, 84.9 months;
range, 24–166 months). The necrotic lesion was typically
located at the medial and superior aspect of the humeral
head. The necrotic angle, which expressed the extent of the
necrotic lesion, was measured on midoblique-coronal plane
(range; 0°–134.7°) and on midoblique-sagittal plane
(range; 0°–150.6°). Of the 46 lesions, 34 were less than 90°
and did not collapse, whereas 11 of the other 12 lesions of
more than 90° (92%) collapsed within 4 years. Of these 11
collapsed lesions, four of less than 100° did not progress,
followed by reparative reaction on plain radiographs,
whereas the other seven of more than 100° progressed to
osteoarthritis. The extent of a necrotic lesion on MRI is
useful to predict collapse of the humeral head.
The institution of one of the authors (NS) has received funding from a
grant from the Japanese Investigation Committee under the auspices
of the Ministry of Health and Welfare.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
T. Sakai (&), N. Sugano, T. Nishii, T. Hananouchi,
H. Yoshikawa
Department of Orthopedic Surgery, Osaka University Medical
School, 2-2, Yamadaoka, 565-0871 Suita, Osaka, Japan
e-mail: tsakai-osk@umin.ac.jp
123
Level of Evidence: Level IV, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Atraumatic osteonecrosis of the humeral head (ONHH)
generally develops in the subchondral region [1, 2, 4–7, 12,
13, 15–17]. In some patients, osteonecrosis (ON) can lead
to collapse of the necrotic subchondral bone, development
of an irregular joint surface, and subsequent joint degen-
eration requiring nonoperative [1, 2] or operative treatment
[1–7, 12, 13, 15–17]. The rate of patients with ONHH
undergoing surgery varies from 22% to 78% [1, 15, 17] in
part because the natural history of corticosteroid- or alco-
hol-related ONHH is still not well understood. Although
one study [5] suggests replacement surgery is related to the
radiographic stage of humeral head involvement we did not
know which cases would collapse and which would not.
Several studies suggest the extent of a necrotic lesion on
radiographs correlates with the prognosis and to the results
of surgical treatment in patients with corticosteroid-related
ONHH or traumatic ONHH [5, 15]. These two studies
reported large lesions in ONHH and patients with collapse
on plain radiographs tend to have poor prognoses. How-
ever, accurate evaluation of a lesion on radiographs is
difficult at best, and impossible in Stage 1 disease, which
is not visualized by plain radiograph or CT scan [24]. MRI
is superior to radiographs in the early detection of necrotic
lesions as well as in evaluating the extent and location of
lesions in hip and knee ON [22, 25]. Early diagnosis and
understanding of the natural history are important for
treatment planning including core decompression [12],
statin therapy [20], and alendronate therapy [18].
Volume 466, Number 5, May 2008
Given MRI is a more sensitive tool, we asked whether
the likelihood of collapse of corticosteroid-related or
alcohol-related ONHH is related to the extent and location
of necrotic lesions on MRI. We also asked whether pro-
gression or cessation of collapse in ONHH depends on the
extent of necrotic lesion visible on MRI.
Materials and Methods
Between April 1993 and May 2004, we diagnosed 220
patients with osteonecrosis of the femoral head by MRI. In
these patients, we performed MRI screening of both
shoulders, knees, and ankles to identify multiple regions of
ON. Atraumatic ONHH was diagnosed by MRI in 29 of
220 patients (13.2%). We reviewed the MRIs of 50
shoulders of these 29 patients with atraumatic ONHH. We
included symptomatic and asymptomatic patients. We
diagnosed ON as an area of normal fat intensity surrounded
by a low-intensity band, band pattern, on spoiled gradient-
recalled echo pulse sequence (SPGR) [21, 25]. We used
MRI as the gold standard since we could not obtain a
histologic diagnosis in asymptomatic patients or those for
whom surgery was not indicated. We excluded two patients
(four shoulders) because both humeral heads of one patient
had already collapsed and the other patient was lost to
followup prior to 2 years. Therefore, we included 46
shoulders of 27 patients with noncollapsed, atraumatic
ONHH in this study. Thirty four of the 46 shoulders were
initially asymptomatic and 12 were symptomatic. Nineteen
of the 27 (70%) patients also had knee ON in addition to
that of the hip. There were nine men and 18 women. The
mean age at diagnosis was 39 years (range, 17–64 years).
The minimum followup was 24 months (mean,
84.9 months; range, 24–166 months). The followup period
was defined as the period from the day of initial exami-
nation to the day of the latest examination or to the day of
surgical treatment. Of the 27 patients, 24 had a history of
corticosteroid therapy and the remaining three patients had
a history of alcohol abuse without any history of trauma
around the shoulder. The history of alcohol abuse was
defined as more than 400 mL ethanol intake per week [8].
Of the 24 corticosteroid-related patients, 15 also had a
history of pulse corticosteroid administration. The under-
lying diseases in the corticosteroid-related patients were
systemic lupus erythematosus in 12 patients and nephritic
syndrome in two patients. Four patients underwent bone
marrow transplantation with subsequent graft versus host
disease. The reasons for corticosteroid administration in the
other six patients were Cushing syndrome, pure red cell
aplasia, sudden deafness, sarcoidosis, pemphoid, and idi-
opathic sclerotic pruritis. The mean total dose of
prednisolone before the diagnosis of ON was 15.6 g (range,
Humeral Head Osteonecrosis on MRI 1075
1.4–35 g). The mean maximum daily dose of prednisolone
was 51.4 mg (range, 20–100 mg). Nineteen patients had
affected bilateral shoulders.
We (TS and NS) evaluated clinical status at diagnosis
and at the latest followup using the UCLA shoulder rating
system [9]. Patients were followed every six months.
We examined all shoulders by anteroposterior radio-
graphs at each visit with the arm in external rotation and
internal rotation. We (TS and NS) radiographically staged
the disease using the Association Research Circulation
Osseous international classification staging system [24].
This staging system has been used for the femoral head
ON, and we applied it to ONHH. Stage 1 is not apparent on
plain radiographs or CT, but is detectable on MRI, bone
scintigraphy, and biopsy. Stage 2 disease is apparent on
radiographs but without collapse; typically radiographs
reveal cysts or areas of increased density in the humeral
head. Stage 3 exhibits collapse. Stage 4 has flattening of the
humeral head with joint space narrowing (osteoarthritis). In
order to investigate the intraobserver reproducibility and
the interobserver reliability, we staged the disease based on
plain radiographs of the initial examination and the latest
followup, and then reviewed them again at 6-month
intervals. The kappa intraobserver reproducibility coeffi-
cient was 0.90 (TS) and 0.90 (NS), whereas the
interobserver reliability coefficient was 0.90 for the initial
examination and 0.89 for the latest followup [23].
We performed imaging on patients within 4 months of
the diagnosis of femoral head ON using a 1.0-Tesla
superconducting magnet (Magnetom Impact; Siemens,
Erlangen, Germany, or Signa; General Electric, Milwau-
kee, WI). To eliminate misdiagnosis, we performed not
only SPGR images, but also T2-weighted or fat-saturation
SPGR images. T2-weighted spin-echo images (TR/TE =
3600/105 msec), SPGR (TR/TE = 14/2.3 msec), and fat-
suppression SPGR were obtained with 1.5-mm slice
thickness. The image matrix was 256 9 256. The SPGR
and fat-suppression SPGR image data were reconstructed
using the Virtual Place-M software (Medical Imaging
Laboratory, Tokyo, Japan). We resliced the humeral head
on the axial view that included the humeral head center,
parallel to the scapula body, and produced a midoblique-
coronal image (Fig. 1) because an oblique-coronal image is
common for shoulder disease [14]. Then, we also sliced the
humeral head on the axial view that included the humeral
head center, perpendicular to the scapular body, and pro-
duced a midoblique-sagittal image (Fig. 1). MRI images
were reviewed for diagnosis by one of the authors (TN)
without benefit of clinical or radiographic data.
We (TS, TH) independently estimated the extent of
ONHH lesions on midoblique-coronal and on midoblique-
sagittal images of the humeral head on SPGR by the
necrotic angle (\ AOB) of the lesion using modified
123
1076 Sakai et al.
Fig. 1A–D (A, B) The necrotic angle (\ AOB) and the median angle (\ COD) on midoblique-coronal spoiled gradient-recalled echo (SPGR)
images and (C, D) on midoblique-sagittal SPGR images of the humeral head were evaluated as the extent and location of the necrotic lesion.
Kerboul’s estimation [10] (Fig. 1). In order to estimate the
extent of ONHH lesions in two cross-sectional directions,
the necrotic angle was not a sum of angles measured on
the midoblique-coronal plane and midoblique-sagittal
plane as in Kerboul’s original definition. These two angles
were recorded independently. The center of the humeral
head was defined as point O, and two end points of the
necrotic lesion on the joint surface were defined as points
A and B. Even if ONHH was present in the humeral head,
we estimated the necrotic angle as 0° when a necrotic
lesion was not detected on the midoblique-coronal image
or on the midoblique-sagittal image. The location of
ONHH lesions on midoblique-coronal and on midoblique-
sagittal SPGR images was expressed by the median point
of each necrotic lesion using the median angle (\ COD)
(Fig. 1), because the range of angle of the necrotic lesion
was so complicated as the location analysis. One medial
basal point on a midoblique-coronal image or one anterior
basal point on a midoblique-sagittal image was defined as
Point C. The median point of the necrotic lesion was
defined as Point D. The necrotic angle and the median
angle were calculated on a computer using the Virtual
Place-M software. The intraobserver error (r = 0.97) and
the interobserver error (r = 0.80) were calculated using
Spearman’s rank test.
We compared the necrotic angle and the median angle
among the four latest radiographic stages (stage 1, 2, 3, and
123
Clinical Orthopaedics and Related Research
4 disease) using the Kruskal-Wallis test. We used SPSS for
Windows 13.0 (SPSS Incorporated, Chicago, IL).
Results
The lesions with initial necrotic angles of more than 90° on
midoblique-coronal SPGR images and on midoblique-
sagittal SPGR images more often (p \ 0.0001) progressed
to collapsed radiographic stages than those with smaller
angles (Table 1). Higher latest stages had greater necrotic
angles (Fig. 2). Analysis of the necrotic angle on mid-
oblique-coronal SPGR images revealed 11 of 12 lesions
(92%) with more than 90° collapsed within 4 years
(Fig. 3A-G), whereas 34 of 34 (100%) lesions of less than
90° did not collapse. In other words, of 46 shoulders with
ONHH lesions, including 31 shoulders in Stage 1 and 15 in
Stage 2 at the initial examination, 11 humeral heads col-
lapsed (24%) within 4 years (Table 2). The range of the
necrotic angle was 0° to 134.7° on the midoblique-coronal
plane, and 0°–150.6° on the midoblique-sagittal plane.
The median angle did not predict progression of ONHH
since there were no differences in the median angle among
the latest radiographic stages (Table 1). The mean median
angle was 51.5° (range, 13.2°–110°) on the midoblique-
coronal plane and 100.7° (range, 69.1°–131.9°) on the
midoblique-sagittal plane. The necrotic lesion was typically
Volume 466, Number 5, May 2008 Humeral Head Osteonecrosis on MRI 1077

Table 1. Relationship between the MRI angles at initial examination and latest radiographic stage
Angles at initial examination Stage 1* (22 shoulders) Stage 2*(13 shoulders) Stage 3*(4 shoulders) Stage 4* (7 shoulders) p Value

Necrotic angle (degrees)

On midoblique-coronal plane 25.3 (0–94.7) 54.4 (13.6–85.4) 94.7 (90.8–99.2) 113.9 (96.6–134.7) \ 0.0001 
On midoblique-sagittal plane 12.3 (0–102.1) 35.8 (0–72.3) 85.4 (64.5–98.2) 121.8 (100–150.6) \ 0.0001 
Median angle (degrees)
On midoblique-coronal plane 54.7 (13.2–110) 49.4 (32.6–77.7) 50.8 (39.2–57) 50.1 (22.3–61.8) NS 
On midoblique-sagittal plane 95.5 (86.3–105.1) 106.4 (69.1–131.9) 103.6 (92.8–113.3) 95.6 (76.2–113.7) NS 
*Association Research Circulation Osseous (ARCO) international classification staging system; Stage 1 = not apparent on plain radiographs or
CT but are detectable on MRI, bone scintigraphy, and biopsy; Stage 2 = abnormal radiograph with cystic or sclerotic changes in the humeral
head without joint collapse; Stage 3 = collapsed stage; Stage 4 = flattening of the humeral head with joint space narrowing;  Kruskal-Wallis
test; NS = not significant.

asymptomatic at the latest followup (Table 3). Four

shoulders of the three patients with alcohol-related ONHH
remained in Stage 1.

Discussion

Several studies suggest the extent of a necrotic lesion on

Fig. 2 Scattergram of the necrotic angle on spoiled gradient-recalled
echo (SPGR) image showed 11 of 12 necrotic lesions (92%) with
more than 90° on the midoblique-coronal plane collapsed. Higher
latest stages had greater necrotic angles.
located at the medial and superior aspect of the humeral
head. In all cases, the necrotic lesion was located within the
proximal half of the humeral head on both the midoblique-
coronal and midoblique-sagittal SPGR images.
All seven shoulders with a necrotic angle of more than
100° progressed to osteoarthritis. The four shoulders with
initially collapsed lesions of less than 100° did not pro-
gress, but rather developed a reparative reaction reflected
by increased density at the demarcation area on plain
radiographs (Fig. 4A-B). Two Stage 4 shoulders underwent
hemiarthroplasty at 38 months and at 65 months,
respectively.
We treated all 12 initially symptomatic shoulders with
nonsteroidal antiinflammatory drugs and the avoidance of
overhead shoulder motions. Twenty nine shoulders were
radiographs correlates with the prognosis and to the results
of surgical treatment in patients with corticosteroid-related
ONHH or traumatic ONHH [5, 15] but these studies were
based on plain radiographs. Given MRI is a more sensitive
tool, we asked whether the likelihood of collapse of corti-
costeroid-related or alcohol-related ONHH is related to the
extent and location of necrotic lesions on MRI. We also
asked whether progression or cessation of collapse in ONHH
depends on the extent of necrotic lesion visible on MRI.
Our study has several limitations. First, the number of
patients was small. To estimate the extent and location of
necrotic lesions accurately, we excluded patients with
ONHH who were not examined by SPGR MRI and patients
whose humeral head had already collapsed because we
could not estimate the lesion size and location on MRI
exactly as in the noncollapsed humeral heads. Although the
sample size was small, this limitation is compensated for
by including only one diagnosis and by ensuring accurate
digital measurements of the necrotic lesion. Second, this
was not a prospective study. The shoulder SPGR MRI was
not always examined at initial examination because we
performed shoulder MRI as screening for multiple ON after
femoral head ON was diagnosed. However, the present
results were not likely influenced by the timing of MRI
because all patients were examined on SPGR MRI before
the necrotic lesion had collapsed.
The likelihood of collapse of corticosteroid-related
ONHH was related to the extent of necrotic lesion on MRI.
The lesion with a necrotic angle of more than 90° on
midoblique-coronal SPGR images and on midoblique-
sagittal SPGR images more often progressed to the

123
1078 Sakai et al. Clinical Orthopaedics and Related Research

Fig. 3A–G (A, B) A midoblique-coronal spoiled gradient-recalled showed the necrotic angle was 103.6° and the median angle was
echo (SPGR) image of a 33-year-old woman with steroid-related ON 82.4°. (E) An initial anteroposterior radiograph showed no demarca-
of the left humeral head showed the necrotic angle was 107.3° and the tion (Stage 1), (F) Stage 3 at 27 months, and (G) Stage 4 at
median angle was 51.5°. (C, D) A midoblique-sagittal SPGR image 42 months after the initial examination.

Table 2. Radiographic stage change (shoulders)

Initial examination Latest followup

Stage 1* 31 Stage 1* 22
Stage 2* 2
Stage 3* 1
Stage 4* 6
Stage 2* 15 Stage 2* 11
Stage 3* 3
Stage 4* 1
*Association Research Circulation Osseous (ARCO) international
classification staging system; Stage 1 = not apparent on plain
radiographs or CT but are detectable on MRI, bone scintigraphy, and
biopsy; Stage 2 = abnormal radiograph with cystic or sclerotic
changes in the humeral head without joint collapse; Stage 3 = col-
lapsed stage; Stage 4 = flattening of the humeral head with joint
space narrowing.

collapsed stage. Collapsed lesions of less than 100° did not

progress to osteoarthritis, followed by reparative reaction
on plain radiographs. The median angle, which expressed
location of the lesion, was not useful to prognosticate
ONHH.
Large lesions in ONHH and collapsed-stage cases on
plain radiographs tend to have poor prognoses [5, 15].
However, predicting outcome of atraumatic ONHH using
MRI has not been previously reported, and we did not
know which cases would collapse. We found a necrotic
Fig. 4A–B (A) An anteroposterior radiograph of a 17-year-old
woman with corticosteroid-related ON of the left humeral head
showed Stage 3 at 6 months after the initial examination, and (B)
extensive reparative reaction at 1 year.
angle of more than 90° on midoblique-coronal SPGR
images and on midoblique-sagittal SPGR images, whereas
the location of ONHH on MRI was not related to the
radiographic stage progression.

123
Volume 466, Number 5, May 2008
Table 3. Clinical status change
Parameters Initial Latest
examination followup
Pain UCLA score (points) 9.4 (5–10) 8.7 (2–10)
Total UCLA score (points) 28.9 (21–30) 27.3 (10–30)
Number of symptomatic shoulders 12* 17*
Number of patients 10 15
with pain on motion
Number of patients 3 10
with limited range of motion
Number of patients 8 8
with muscle weakness
Number of asymptomatic shoulders 34 29
*Total number of symptomatic shoulders does not equal the sum of
the numbers of patients with each symptom since some patients had
more than one symptom.
The necrotic lesion was typically located at the medial
and superior aspect of the humeral head, namely the site of
glenohumeral contact in approximately 90° of shoulder
abduction, like in previous reports [1, 13, 15]. With respect
to the anteroposterior location on MRI, the necrotic lesion
affected the middle posterior portion to a greater extent
than the anterior portion. These findings may be related to
the distribution of vessels supplying nutrition. In the
humeral head, the blood supply is derived from the anterior
and posterior circumflex arteries [3, 11]. The anterolateral
ascending branch of the anterior circumflex artery supplies
almost all of the area of the humeral head, whereas the
posterior circumflex artery vascularizes only the posterior
portion of the greater tuberosity and a small posteroinferior
part of the head [3].
The goal of treatment for ONHH should be to prevent
both collapse of the humeral head and osteoarthritic
change. Although most patients were using steroids, the
incidence of collapse was relatively low (24%) compared
with an incidence of ON of the femoral head of 77% in
one study [25]. The lower collapse rate may be related to
the locations of the lesions or distribution of stresses in the
humeral head. Thus, the course of most patients with
corticosteroid-related or alcohol-related ONHH may be
better than ON of the femoral head [25]. Nonoperative
treatment may be sufficient for small necrotic lesions of
less than 90° on the midoblique-coronal images because
they will not collapse. Necrotic lesions of more than 90°
and less than 100° showed cessation of collapse and
extensive reparative reaction. In femoral head ON, even if
a necrotic lesion begins to collapse, some cases stabilize
without further collapse [19]. In these patients, aggressive
surgery may not be necessary. On the other hand, surgery
may be inevitable for large necrotic lesions of more than
100° on the midoblique-coronal images. Because patients
Humeral Head Osteonecrosis on MRI 1079
with atraumatic ONHH are relatively young [1, 17], total
shoulder arthroplasty [1, 2, 6] should be avoided if pos-
sible. For young patients, arthroscopic débridement [4, 7]
or core decompression [12, 16, 17] may be indicated. The
efficacy of these treatment modalities should be assessed
after evaluating the extent of necrotic lesions on MRI in
the future.
Acknowledgments We thank Kazuomi Sugamoto, MD, PhD, Akira
Goto, MD, PhD, and Masaki Takao, MD,PhD, for technical support.
References
1. Cruess RL. Steroid-induced avascular necrosis of the head of the
humerus. Natural history and management. J Bone Joint Surg Br.
1976;58:313–317.
2. Cruess RL. Experience with steroid-induced avascular necrosis of
the shoulder and etiologic considerations regarding osteonecrosis
of the hip. Clin Orthop Relat Res. 1978;130:86–93.
3. Gerber C, Schneeberger AG, Vinh T. The arterial vascularization
of the humeral head. An anatomical study. J Bone Joint Surg Am.
1990;72:1486–1494.
4. Hardy P, Decrette E, Jeanrot C, Colom A, Lortat-Jacob A, Benoit
J. Arthroscopic treatment of bilateral humeral head osteonecrosis.
A case report. Arthroscopy. 2000;16:332–335.
5. Hattrup SJ, Cofield RH. Osteonecrosis of the humeral head:
relationship of disease stage, extent, and cause to natural history.
J Shoulder Elbow Surg. 1999;8:559–564.
6. Hattrup SJ, Cofield RH. Osteonecrosis of the humeral head:
results of replacement. J Shoulder Elbow Surg. 2000;9:177–182.
7. Hayes JM. Arthroscopic treatment of steroid-induced osteone-
crosis of the humeral head. A case report. Arthroscopy.
1989;5:218–221.
8. Hirota Y, Hirohata Y, Fukuda K, Mori M, Yanagawa H, Ohno Y,
Sugioka Y. Association of alcohol intake, cigarette smoking, and
occupational status with the risk of idiopathic osteonecrosis of the
femoral head. Am J Epidemiol. 1993;137:530–538.
9. Kay SP, Amstutz HC. Shoulder hemiarthroplasty at UCLA. Clin
Orthop Relat Res. 1988;228:42–48.
10. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. The
conservative surgical treatment of idiopathic aseptic necrosis of
the femoral head. J Bone Joint Surg Br. 1974;56:291–296.
11. Laing PG. The arterial supply of the adult humerus. J Bone Joint
Surg Am. 1956;38:1105–1116.
12. LaPorte DM, Mont MA, Mohan V, Pierre-Jacques H, Jones LC,
Hungerford DS. Osteonecrosis of the humeral head treated by
core decompression. Clin Orthop Relat Res. 1998;355:254–260.
13. Lee JA, Farooki S, Ashman CJ, Yu JS. MR patterns of
involvement of human head osteonecrosis. J Comput Assist
Tomogr. 2002;26:839–842.
14. Lee SY, Lee JK. Horizontal component of partial-thickness tears
of rotator cuff: imaging characteristics and comparison of ABER
view with oblique coronal view at MR arthrography – initial
results. Radiology. 2002;224:470–476.
15. L’Insalata JC, Pagnani MJ, Warren RF, Dines DM. Humeral head
osteonecrosis: clinical course and radiographic predictors of
outcome. J Shoulder Elbow Surg. 1996;5:355–361.
16. Mont MA, Maar DC, Urquhart MW, Lennox D, Hungerford DS.
Avascular necrosis of the humeral head treated by core decom-
pression. A retrospective review. J Bone Joint Surg Br.
1993;75:785–788.
123
1080 Sakai et al.
17. Mont MA, Payman RK, LaPorte DM, Petri M, Jones LC,
Hungerford DS. Atraumatic osteonecrosis of the humeral head.
J Rheumatol. 2000;27:1766–1773.
18. Nishii T, Sugano N, Miki H, Hashimoto J, Yoshikawa H. Does
Alendronate prevent collapse in osteonecrosis of the femoral
head? Clin Orthop Relat Res. 2006;443:273–279.
19. Nishii T, Sugano N, Ohzono K, Sakai T, Haraguchi K,
Yoshikawa H. Progression and cessation of collapse in osteo-
necrosis of the femoral head. Clin Orthop Relat Res. 2002;
400:149–157.
20. Pritchett JW. Statin therapy decreases the risk of osteonecrosis in
patients receiving steroids. Clin Orthop Relat Res. 2001;386:173–
178.
21. Sakai T, Sugano N, Nishii T, Haraguchi K, Ochi T, Ohzono K.
MR findings of necrotic lesions and the extralesional area of
123
Clinical Orthopaedics and Related Research
osteonecrosis of the femoral head. Skeletal Radiol. 2000;29:
133–141.
22. Sakai T, Sugano N, Ohzono K, Matsui M, Hiroshima K, Ochi T.
MRI evaluation of steroid- or alcohol-related osteonecrosis of the
femoral condyle. Acta Orthop Scand. 1998;69:598–602.
23. Smith SW, Meyer RA, Connor PM, Smith SE, Hanley EN.
Interobserver reliability and intraobserver reproducibility of the
modified Ficat classification system of osteonecrosis of the
femoral head. J Bone Joint Surg Am. 1996;78:1702–1706.
24. Stulberg BN. Editorial comment. Clin Orthop Relat Res.
1997;334:2–5.
25. Sugano N, Ohzono K, Masuhara K, Takaoka K, Ono K. Prog-
nostication of osteonecrosis of the femoral head in patients with
systemic lupus erythematosus by MRI. Clin Orthop Relat Res.
1994;305:190–199.
Clin Orthop Relat Res (2008) 466:1081–1086
DOI 10.1007/s11999-008-0219-2
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
F-18 Fluoride Positron Emission Tomography of the Hip for
Osteonecrosis
Vinod Dasa MD, Hani Adbel-Nabi MD, PhD,
Mark J. Anders MD, William M. Mihalko MD, PhD
Published online: 24 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Osteonecrosis (ON) of the femoral head con-
tinues to be a devastating disorder for young patients. We
evaluated the F-18 fluoride positron emission tomography
(PET) imaging modality for use in detection of the bone
involved in ON of the hip. We retrospectively reviewed the
records of 60 consecutive patients diagnosed with ON and
interviewed all by phone. Eleven patients (17 hips) of those
interviewed agreed to participate in the study. We
classified the ON using the University of Pennsylvania
classification system and compared each patient’s plain AP
bone scan, single photon emission 3-D computed tomog-
raphy, and MRI. ON was associated with HIV, alcohol,
steroid use, and polycythemia vera in this group. Nine of
17 hips (8 patients) had acetabular increased uptake when
using the F-18 fluoride PET scans that were not seen on
MRI, single photon emission computed tomography, or
William M. Mihalko MD PhD is a consultant for Stryker Inc, Smith &
Nephew Inc., Ethicon and Aesculap. He also receives research
support from Stryker and Aesculap. No research or consulting is
related to this study.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
V. Dasa, M. J. Anders
Department of Orthopaedic Surgery, University of Buffalo,
Buffalo, NY, USA
H. Adbel-Nabi
Department of Nuclear Medicine, University of Buffalo,
Buffalo, NY, USA
W. M. Mihalko (&)
Department of Orthopaedic Surgery, University of Virginia,
PO Box 800159, Charlottesville, VA 22908-0159, USA
e-mail: Wmm4n@virginia.edu
bone scans. These data suggest earlier acetabular changes
in osteonecrosis may exist that traditional imaging
modalities do not reveal.
Level of Evidence: Level III, diagnostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis (ON) of the hip can be a devastating prob-
lem, especially in young patients, and may result in
irreversible changes of the hip [6, 8, 12]. Once collapse and
substantial degenerative changes occur, the patient usually
undergoes reconstruction of the hip with an arthroplasty
[10, 15, 17, 18]. In young patients, this can lead to con-
siderable future difficulties, including multiple revision
arthroplasties and loss of income or considerable career
changes, which can have a major socioeconomic impact
throughout the patient’s life [10, 15].
The pathogenesis and etiology of ON remain unclear.
Known associated factors include traumatic dislocation or
injury, steroid use, and alcohol abuse; some patients have
no identifiable risk factors. Most data point to a micro-
vascular insult or hyperlipidemia in nontraumatic cases
[10, 13, 15, 18]. Even though there is evidence that dis-
tinguishes a specific cause-and-effect relationship between
certain risk factors (steroid use, hyperlipidemia and sickle
cell disease) and ON, standard diagnostic techniques (MRI
or technetium bone scan) do not always provide prognostic
information. Given the array of potential risk factors, from
steroid use and alcohol abuse to HIV, an analysis which
reflected the metabolic activity of the bone might be useful.
In cases of osteonecrosis, an infarct region on the femoral
head may have a proprioceptive impact on the joint that
123
1082 Dasa et al. Clinical Orthopaedics and Related Research

may start to overload the acetabular side of the joint. If this trauma and without surgical intervention in at least one hip
finding is discovered on an image modality then it aids in if bilateral disease was present. Sixty patients were iden-
predicting which patients may go on to progression of tified with the diagnosis of ON of the femoral head at this
disease. time.
Positron emission tomography (PET) scans provide a The study was designed to identify hips with atraumatic
real-time image of physiology based on the type of radio- etiologies of ON to determine if any differences between
labeled marker used. Traditionally, PET scans, in addition F-18 fluoride PET scan imaging and the traditional MRI
to MRI and SPECT scans, have been utilized to determine and SPECT scan imaging were found. Sixty patients were
vascularity and uptake changes in patients with tumor identified with the diagnosis of ON of the femoral head. All
progression; however, PET scan may be more sensitive in of the patients were then contacted by phone for inclusion
detecting early changes compared to MRI and these in the study. Out of the original 60 patients 11 (18%) with
changes might predict subsequent progression. PET imag- aseptic and atraumatic ON of one or both femoral heads
ing has been utilized extensively in orthopaedic skeletal agreed to participate. Each patient underwent staging by
disease assessment as well as in cases where interference the University of Pennsylvania classification system
from implants inhibits the use of other imaging modalities (Table 1) [19]. In this study group two hips were excluded
[2, 4]. F-18 FDG accumulates in cancer cells due an due to previous unilateral core decompression (these two
increased glucose metabolism. The process, however, is hips were in patients with bilateral ON, leaving them in the
not specific to tumors. FDG-18 also accumulates in study for the untreated side) and three hips in three patients
inflammatory cells, such as lymphocytes, neutrophils, and
macrophages which have elevated glucose requirements, Table 1. Comparison of the University of Pennsylvania osteone-
and therefore the process may be useful in ON [3, 5, 11, crosis classification system to Ficat
14–17]. As suggested above, it is possible PET scans will Stage Ficat University of Pennsylvania
detect ON earlier than MRI and single-photon emission
computed tomography (SPECT) scans or that some early I No changes on Normal radiograph
radiograph
uptake or vascular changes might predict the lesions that
Clinical symptoms Abnormal MR or bone scan
are going to progress to changes on both sides of the joint suspicious
and eventual arthroplasty. PET scan is a powerful tool in A = mild, \ 15% of femoral
oncology and it may also play a role in diagnosing ON head affected;
[17]. In a pilot study, Schiepers et al. [17] determined a B = moderate, 15–30%
flow ratio could be established and used to predict a suc- affected;
cessful outcome with a conservative regimen in patients C = severe, [ 30% affected
with ON of the femoral head. The authors suggested this II Bone remodeling Cystic and sclerotic change
type of image modality could be used in clinical practice in femoral head
and would permit prediction of the outcome depending No changes in the shape A = mild, \ 15% of femoral
of the femoral head head affected;
upon regional skeletal flow measurements [17].
Subchondral sclerosis B = moderate, 15–30%
We hypothesized F-18 fluoride PET scan imaging would
affected;
match the traditional ‘‘gold standard’’ imaging studies of
Cysts C = severe, [ 30% affected
MRI and SPECT modalities but would also provide further
III Crescent sign Crescent sign without flattening
information not seen with standard imaging modalities. If
A = mild, \ 15% of femoral
PET scan imaging can be determined to give further head affected;
information concerning areas of activity in the hip itself, B = moderate, 15–30% affected
then it may potentially be utilized as a prognostic study in C = severe, [ 30% affected
the future. IV Joint space narrowing Flattening of the femoral head
Degenerative changes on A = mild, \ 15% of femoral
both sides of the joint head affected and \ 2 mm
Materials and Methods of depression;
Femoral head deformation B = moderate, 15–30% affected
Utilizing the ICD-9 code for osteonecrosis of the femoral and 2–4 mm of depression;
head and neck in a county-based hospital clinic in April C = severe, [ 30% affected)
2003, a list of active patients with this diagnosis were V — Joint narrowing or acetabular
identified and recruited until December 2003 for inclusion changes (can be graded
according to severity)
in this pilot study. Inclusion criterion was simply a diag-
VI — Advanced degenerative changes
nosis of ON of the femoral head without a history of

123
Volume 466, Number 5, May 2008 F-18 Fluoride PET Scan Analysis of ON 1083

Table 2. Demographics of patients who met the inclusion criteria injected intravenously. For the dynamic study, 12 frames,
and agreed to participate in the study each for 5 seconds, were acquired during the first minute
Patient Age Gender Etiology U of Penn. Stage using the Minimize Wait Program (ADAC C-PET PLUS
250, Philips Medical Systems, Netherlands), with NaI
Right Left
scintillation crystal transmission imaging of Cs-137 trans-
1 58 M steroids IB IB mission scan 45 sec/rotation (2 rotations = 1.5 min) and
2 36 F steroids IA NA* small sinograms (2-mm slice thickness). Then, a 1-minute
3 78 M ETOH VA VI frame was acquired for the next 4 minutes followed by
4 31 F HIV NA  IVB 2-minute frames 10 times. Static images, 5 minutes per
5 58 M ETOH NA* IIIB frame, were acquired over the next 15 minutes. Total
6 51 F polycythemia IIB IIA acquisition time, including transmission, was approxi-
7 34 F steroids IVA IA mately 55 minutes.
8 54 M ETOH IA VB One board-certified nuclear medicine physician (HN)
9 48 M steroids IB IVC evaluated PET scans and bone scans and provided descrip-
10 53 M steroids IA NA  tive results. The scan area of the bone in the ilium was
11 41 M ETOH IIIA NA* utilized as a baseline normal for each PET scan and the
difference in the femoral head and surrounding bone of the
* = not applicable due to a previous THA;   = not applicable due to
a previous core decompression; ETOH = alcohol abuse.
hip was then determined to have more uptake in each PET
scan image window. This provided an objective means of
determining differences in each of the image windows
analyzed for the PET scans. Two attending physicians (WM,
for the presence of a total hip arthroplasty opposite an MA) and one resident (VD) in the orthopaedic department
active case of ON of the femoral head (Table 2). This left reviewed all MRI images and staged all hips using the
11 patients with 17 osteonecrotic hips for inclusion. Oste- University of Pennsylvania classification system [10].
onecrosis was associated with steroid use (five patients, The data were organized by presence of changes on each
eight hips), alcohol abuse (four patients, six hips), and with imaging modality and compared utilizing a Kappa score to
HIV (one patient, two hips), polycythemia vera (one determine agreement between the PET scan findings and
patient, one hip). All patients underwent PET scanning, each of the traditional imaging modalities. Any differences
SPECT and bone scans, and MRI imaging over 2 days. where PET scan images revealed changes not recorded on
Patients received no financial incentive other than free the traditional imaging modalities were recorded as well
transportation to the imaging center provided by the for analysis. Confidence intervals were also calculated and
nuclear medicine department. This study was approved by reported to reveal the effects of the small sample size. All
the Investigational Review Board at our institution. data were calculated using SPSS software (Chicago, IL).
For MRI, we obtained both T1- and T2-weighted images
without contrast during one of the imaging days. Images of
each hip and pelvis were in the axial, coronal, and sagittal Results
planes. No patients had contraindications to obtaining an
MRI. Three-phase bone and SPECT scans were obtained In general, all patients had MRI findings of the femoral
on a dual head Vertex camera (ADAC Laboratories, Mil- head consistent with ON. The patient with HIV (Stage IV)
pitas, CA). For three-phase bone and SPECT scans, showed consistent findings between the PET scan and bone
approximately 20 to 25 mCi of Tc-99m MDP was injected scans. Both showed increased uptake of the acetabulum;
intravenously while the subject was lying in the SPECT however, the left femoral head showed no changes on PET
camera gantry. Twenty-five-second dynamic acquisition but did show increased uptake in the neck and greater
frames were acquired over the hip regions followed by a trochanter. The second patient with steroid-induced ON
blood pool image. Approximately 2 hours later, anterior who underwent decompression of a Stage II lesion which
and posterior planar views of the hips were obtained in a was excluded from the study showed no evidence of
static mode (256 9 256 matrix) with high-resolution col- increased uptake on PET but still showed MRI changes
limators. Next, SPECT of the pelvic region was acquired consistent with Stage II disease.
and the images reconstructed in 3-D mode. All patients had MRI findings of ON in the femoral
For F-18 fluoride PET scans, on a separate day, a head. (Table 3). Bone scans did not reveal ON in either
transmission scan of the pelvis overlapping the full axial femoral head in five of six hips when compared with the
field of view covered by the dynamic emission data was MRI. The findings were comparable to (Kappa = 0.88) the
acquired. Next, 2.5 to 3.5 mCi of F-18 fluoride was PET imaging findings. When the PET and MRI scans for

123
1084 Dasa et al. Clinical Orthopaedics and Related Research

Table 3. Presence of imaging findings in femoral head and acetab- Discussion

ulum in MRI and PET scan*
Patient MRI femoral PET scan MRI PET scan In this study we sought to determine if agreement between
head right/left femoral head acetabulum acetabulum more traditional imaging modalities and PET F-18 fluoride
right/Left right/left right/left scans existed in patient with atraumatic ON of the femoral
1 1/1 0/0 0/0 0/0 head and also whether the PET scan modality provided any
2 1/NA 1/NA 0/NA 0/NA further areas about the hip joint where activity was recorded
3 1/1 1/1 0/0 1/0 that MRI and SPECT scan did not detect. Standard diag-
4 NA/1 NA/1 NA/0 NA/1 nostic techniques (MRI or technetium bone scan) for ON do
5 NA/1 NA/1 NA/0 NA/1
not always provide prognostic information. Traditionally,
6 1/1 1/1 0/0 1/1
PET scans, in addition to MRI and SPECT scans, have been
utilized to determine vascularity and uptake changes in
7 1/1 1/1 0/0 0/1
patients with tumor progression; however, PET scans may be
8 1/1 1/1 0/0 1/0
more sensitive in detecting early changes compared to MRI
9 1/1 1/1 0/0 1/0
and these changes might predict subsequent progression. We
10 1/NA 1/NA 0/NA 1/NA
hypothesized F-18 fluoride PET scan imaging would match
11 1/NA 1/NA 0/NA 0/NA
the traditional ‘‘gold standard’’ imaging studies of MRI and
*Presence is indicated by a 1 and absence by a 0. SPECT modalities but would also provide further informa-
NA = Not Applicable due to previous surgery. tion not seen with standard imaging modalities.
Our study is limited by the sample size and the possible
bias in the sample population. We did record good agree-
right-sided ON of the femoral head were compared, an ment of the PET scan with MRI findings in the femoral
agreement of 88.9% (95% CI = 51.8 to 99.7) was deter- head and also identified increased activity in the acetabu-
mined. PET versus MRI for the left side of the femoral lum of several patients with the PET scan modality which
head revealed an agreement of 87.5% (95% CI = 47.4 to was not apparent in the MRI or bone scan. This resulted in
99.7). poor agreement on the acetabular side of the joint and a
PET F-18 fluoride imaging studies in 9 hips (8 patients) low Kappa score. In this small pilot study it may be diffi-
revealed signal changes in the region of the acetabulum not cult to determine the implication of the acetabular findings.
detected on MRI (Figs. 1 and 2, Table 3). Comparison of With a large number of patients identified who met the
these findings resulted in poor agreement with a Kappa of inclusion criteria (60), but a small fraction (11) agreeing to
0.36. participate, we experienced a major limitation that is

Fig. 1A–E (A) A female patient with polycythemia vera remains involvement. Positron emission tomography scan results at (C) 5 to
asymptomatic with early-onset ON visible in this radiograph. (B) 10 minutes, (D) 15 to 20 minutes, and (E) 35 to 40 minutes reveal
MRI reveals changes in the femoral head with no acetabular increased acetabular uptake outlined by arrows.

123
Volume 466, Number 5, May 2008
Fig. 2A–D A female patient
with ON secondary to steroid
use had undergone left THA. She
remains (A) asymptomatic on the
right side as shown in this radio-
graph. (B) MRI reveals changes
in the femoral head with no
acetabular involvement. Positron
emission tomography scan
results at (C) 25 to 30 minutes
and (D) 35 to 40 minutes reveal
increased acetabular uptake out-
lined by the arrow.
difficult to overcome. Because the PET scan image read-
ings were performed by a single reader, this may be viewed
as a limitation as well, despite utilizing an objective cri-
terion. The ilium of each patient acted as a normal baseline,
allowing the different timelines to be more objectively
reported. It is difficult to rate the sensitivity and specificity
of our findings due to the fact it is uncertain if they indicate
a pathological process or adaptation.
Osteonecrosis of the femoral head is a disease process for
which progression is difficult to predict. Given the various
etiologies this should not be surprising. However, despite
the various causes the end result is often the same [14, 18].
In our limited series we found a portion of patients with
good agreement of findings compared to MR imaging, but a
subset of patients having acetabular changes in early stages
of ON was identified only by PET imaging. These findings
had no correlation to findings in this region on MRI or
SPECT. In the later stages, degenerative arthritis confounds
the accuracy of all the imaging modalities, especially at the
articular surface [3, 9, 10]. However, in early disease, the
presence of changes on both sides of the joint may lead to
greater acceleration of degenerative changes. The acetab-
ular blood supply comes primarily from the superior
gluteal, inferior gluteal, and obturator arteries. The blood
supply for the femoral head arises from branches of the
profunda femoris [1]. These vessels are clearly from dif-
ferent sources, but the changes seen on PET were still
recorded in both in the femoral head and acetabulum.
Fink et al. [3] described acetabular findings on MRI in
9.5% of patients with femoral head ON. We observed
acetabular changes on PET scans in patients with ON
without corresponding MRI changes. All bone scans
showed increased uptake of both the acetabulum and
femur, but we were unable to distinguish between the two
F-18 Fluoride PET Scan Analysis of ON 1085
given the poor resolution of the images. Interestingly, the
patient with HIV who had undergone core decompression
3 weeks previously had no improvement in flow relative to
the nonoperative side included in the study, whereas the
patient who underwent core decompression 2 years previ-
ously did not show PET evidence of ON but continues to
have MRI changes on the hip that was excluded from the
study due to previous surgery.
Further study is needed to ascertain whether acetabular
ON predicts progression of the disease. Furthermore, if the
early presence of acetabular metabolic changes as indicated
on F-18 fluoride PET can predict outcome in nonoperative
treatments, then it may serve as an important role in the
future [7, 11]. With the advent of PET computed tomog-
raphy combined with labels such as fluorodeoxyglucose,
we may gain considerable insight into the pathophysiology
of ON as well. Drawing conclusions from this small sample
size is difficult, and our observations have raised more
questions. However, the results serve as a starting point for
future investigation to determine if the presence of a
positive signal on the acetabular side of the joint is a
prognostic indicator for success or failure of progression of
the disease. The results, however, may suggest an expan-
ded vascular and anatomic role of ON extending into the
acetabulum than previously believed.
Acknowledgments We thank Wendy Novicoff, PhD, for her
assistance on statistical analysis of this study.
References
1. Beck M, Leunig M, Ellis T, Sledge JB, Ganz R. The acetabular
blood supply: implications for periacetabular osteotomies. Surg
Radiol Anat. 2003;25:361–367.
123
1086 Dasa et al.
2. Cook GJ, Fogelman I. The role of positron emission tomography
in skeletal disease. Semin Nucl Med. 2001;31:50–61.
3. Fink B, Assheuer J, Enderle A, Schneider T, Ruther W. Avas-
cular osteonecrosis of the acetabulum. Skeletal Radiol.
1997;26:509–516.
4. Forrest N, Welch A, Murray AD, Schweiger L, Hutchison J,
Ashcroft GP. Femoral head viability after Birmingham resur-
facing hip arthroplasty: assessment with use of [18F] fluoride
positron emission tomography. J Bone Joint Surg Am. 2006;
88(Suppl 3):84–89.
5. Grigolon MV, Delbeke D. F-18 FDG uptake in a bone infarct: a
case report. Clin Nucl Med. 2001;26:613–614.
6. Hernigou P, Habibi A, Bachir D, Galacteros F. The natural
history of asymptomatic osteonecrosis of the femoral head in
adults with sickle cell disease. J Bone Joint Surg Am.
2006;88:2565–2572.
7. Hung GU, Tsai SC, Lin WY. Extraordinarily high F-18 FDG
uptake caused by radiation necrosis in a patient with nasopha-
ryngeal carcinoma. Clin Nucl Med. 2005;30:558–559.
8. Hungerford DS, Mont MA. The natural history of untreated
asymptomatic hips in patients who have non-traumatic osteone-
crosis. J Bone Joint Surg Am. 1998;80(5):765–6.
9. Lee MJ, Corrigan J, Stack JP, Ennis JT. A comparison of modern
imaging modalities in osteonecrosis of the femoral head. Clin
Radiol. 1990;42:427–432.
10. Lieberman JR, Berry DJ, Mont MA, Aaron RK, Callaghan JJ,
Rajadhyaksha AD, Urbaniak JR. Osteonecrosis of the hip: man-
agement in the 21st century. Instr Course Lect. 2003;52:337–355.
11. Liu SH, Chang JT, Ng SH, Chan SC, Yen TC. False positive
fluorine-18 fluorodeoxy-D glucose positron emission tomography
123
Clinical Orthopaedics and Related Research
finding caused by osteoradionecrosis in a nasopharyngeal carci-
noma patient. Br J Radiol. 2004;77:257–260.
12. Malizos KN, Karantanas AH, Varitimidis SE, Dailiana ZH,
Bargiotas K, Maris T. Osteonecrosis of the femoral head: etiol-
ogy, imaging and treatment. Eur J Radiol. 2007;63:16–28.
13. Matsui M, Saito S, Ohzono K, Sugano N, Saito M, Takaoka K,
Ono K. Experimental steroid-induced osteonecrosis in adult
rabbits with hypersensitivity vasculitis. Clin Orthop Relat Res.
1992;277:61–72.
14. Mont MA, Carbone JJ, Fairbank AC. Core decompression versus
nonoperative management for osteonecrosis of the hip. Clin
Orthop Relat Res. 1996;324:169–178.
15. Mont MA, Jones LC, Sotereanos DG, Amstutz HC, Hungerford
DS. Understanding and treating osteonecrosis of the femoral
head. Instr Course Lect. 2000;49:169–185.
16. Nguyen BD, Roarke MC. F-18 FDG PET/CT incidental finding
of large ischiogluteal bursitis. Clin Nucl Med. 2007;32:
535–537.
17. Schiepers C, Broos P, Miserez M, Bormans G, De Roo M.
Measurement of skeletal flow with positron emission tomography
and 18F-fluoride in femoral head osteonecrosis. Arch Orthop
Trauma Surg. 1998;118:131–135.
18. Steinberg ME, Brighton CT, Corces A, Hayken GD, Steinberg
DR, Strafford B, Tooze SE, Fallon M. Osteonecrosis of the
femoral head. Results of core decompression and grafting with
and without electrical stimulation. Clin Orthop Relat Res.
1989;249:199–208.
19. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system
for staging avascular necrosis. J Bone Joint Surg Br. 1995;77:
34–41.
Clin Orthop Relat Res (2008) 466:1087–1092
DOI 10.1007/s11999-008-0191-x
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Untreated Asymptomatic Hips in Patients With Osteonecrosis
of the Femoral Head
Byung-Woo Min MD, Kwang-Soon Song MD,
Chul-Hyun Cho MD, Sung-Moon Lee MD,
Kyung-Jae Lee MD
Published online: 8 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Because there is no consensus regarding the
factors predicting femoral head collapse in asymptomatic
osteonecrosis of the hip, we studied the risk factors for
collapse. Between 1990 and 2000, we used MRI to confirm
asymptomatic osteonecrosis of the femoral head in 81
patients (81 hips) whose other hip had nontraumatic
symptomatic osteonecrosis and we monitored them pro-
spectively. The minimum followup was 5 years (mean,
8.3 years; range, 5–16 years). At the latest followup, 31
hips (38%) were symptomatic and 26 hips (32%) had
collapsed. The mean interval between diagnosis and col-
lapse was 4.1 years. We observed no correlation between
femoral head collapse and patients’ age, gender, weight,
presumed cause of osteonecrosis, or length of followup.
With combined factors, only extent of large necrotic lesion
(hazard ratio, 4.06; 95% confidence interval, 1.29–12.77)
and location of Type C2 necrotic lesion (hazard ratio, 6.35;
95% confidence interval, 1.18–34.11) predicted collapse.
Level of Evidence: Level I, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Each author certifies that he has no commercial associations
(eg, consultancies, stock ownership, equity interest, patient/
licensing arrangements, etc) that might pose a conflict of interest
in connection with the submitted article.
Each author certifies that his institution has approved the human
protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
B.-W. Min (&), K.-S. Song, C.-H. Cho, S.-M. Lee, K.-J. Lee
Department of Orthopaedic Surgery, Keimyung University,
School of Medicine, Dongsan Medical Center,
194 Dongsan-dong, Joong-gu, Daegu 700-712, Korea
e-mail: min@dsmc.or.kr
Introduction
In patients with symptomatic osteonecrosis (ON) of the
femoral head, the natural history has been well documented
to progress to collapse and painful degenerative arthritis in
almost all patients [1, 2, 4–6, 9, 11, 13, 14, 18, 21, 23, 26,
28, 29]. However, there is no consensus regarding the
natural history in patients with asymptomatic ON of the
femoral head. The reported rates of progression in these
hips have ranged from as low as 17% to as high as 100%
[4, 5, 7, 9, 11, 14, 18, 29].
Historically, investigators had difficulty evaluating the
fate of asymptomatic ON because the disease could not be
clearly detected in asymptomatic hips by simple radio-
graphs. However, with the advent of MRI, it became easier
to diagnose asymptomatic disease and to more accurately
ascertain the extent, location, and stage of the necrotic
lesion [2, 17, 28].
We asked whether the radiographic stage, extent of the
necrotic lesion, and lesion location influenced the rate of
collapse of the asymptomatic ON of the femoral head and
if so, how long after symptom onset and initial diagnosis of
ON. We then asked whether a number of potential demo-
graphic risk factors would predict collapse.
Materials and Methods
We identified 152 consecutive patients with established
symptomatic atraumatic ON of the femoral head in one hip
treated with THA or other surgical interventions such as
osteotomy or core decompression between 1990 and 2000.
Because of the high probability of bilateral involvement,
both hips were evaluated with radiographs and MRI even
when the contralateral hip was asymptomatic. Among the
123
1088 Min et al.
152 patients, we identified 91 (59.9%) with asymptomatic
ON of the femoral head in the contralateral hip. We
monitored these 91 patients prospectively to detect col-
lapse. We recorded the following potential risk factors:
age, length of followup, weight, gender, presumed cause of
necrosis, stage, extent of involvement, and location of the
necrotic segment. We followed all patients a minimum of
5 years. All patients gave informed consent for participa-
tion, and our Institutional Review Board approved the
protocol.
The necrotic lesions were confirmed by the band-like
homogeneous and nonhomogeneous patterns shown on
MRI [2]. Four patients (four hips) died from causes unre-
lated to surgery, and six patients (six hips) were lost to
followup before the end of the minimum 5-year followup
period; this left 81 patients (81 hips) as the subjects of our
study. None of the 10 patients (12.3%) who died or were
lost to followup monitoring showed collapse of the femoral
head by the time of their final evaluation. There were 68
men and 13 women with a mean age of 50.5 years (range,
22–77 years). The left hip was studied in 36 patients and
the right hip in 45. The presumed cause of ON was alcohol
in 39 patients (48.1%), idiopathic in 30 (37%), and steroids
in 12 (14.9%). Cases of posttraumatic ON were excluded.
The minimum followup was 5 years (mean, 8.3 years;
range, 5–16 years).
All patients underwent clinical and radiographic fol-
lowup evaluation at 3-month intervals for the first 2 years
and at yearly intervals thereafter until the time of collapse
or for a minimum of 5 years. The initial examination
included clinical evaluation, radiographs, and MRI images.
Clinical information was obtained by means of an inter-
view and physical examination conducted by two of the
authors (BWM, KSS). Patients were asked to indicate the
grade of hip pain as none, slight, mild, moderate, or severe
using the same criteria as for the pain categories in the
Harris hip score [8].
Radiographs and MRI images of all patients were ana-
lyzed by two independent readers (CHC, SML) who did
not know the patients’ clinical and radiographic histories. If
Table 1. Modified classification system of Steinberg et al. [25]
Stage Radiographic findings
Stage I Normal
Stage II* Abnormal
Sclerotic type Sclerotic change in necrotic segments
Cystic type Cystic change in necrotic segments
* Stage II was subdivided as Stage II sclerotic type and Stage II cystic type according to a method of Bozic et al. [3].
123
Clinical Orthopaedics and Related Research
there was a disagreement, a third observer (KJL)
interpreted the films and a unanimous decision was reached
regarding the parameter. The coefficient of variations of
estimation between the observers was 4.2%. Our analysis
included various parameters, including Steinberg stage,
extent of involvement, and location of the necrotic seg-
ment. Radiographic stage was assessed according to the
classification system of Steinberg et al. [25]. Stage II was
subdivided as Stage II sclerotic type and Stage II cystic
type according to the method of Bozic et al. [3] (Table 1).
The extent of the necrotic lesion was quantified using
the method of Steinberg et al. [25] in which hips were
evaluated by simple visual estimate of lesion size. Hips
evaluated by visual estimate were grouped by lesion size:
small (less than 15% of head involvement); medium (15%
to 30% involvement); and large (greater than 30%).
The location of necrosis on T1-weighted midcoronal
images was classified as Type A, B, C1, or C2 using the
criteria described by Sugano et al. [27]. Type A lesions
occupy the medial third or less of the weightbearing area of
the femoral head; Type B lesions the medial two-thirds or
less; Type C1 lesions more than the medial two-thirds but
not extending laterally to the acetabular edge; and Type C2
lesions more than the medial two-thirds and extending
laterally to the acetabular edge.
We determined differences in clinical data between hips
in which collapse developed and those in which it did not
using the Kruskal-Wallis test (age, length of followup,
weight) and chi square test (gender, presumed cause of
necrosis). Clinical failure was defined as the occurrence
of pain, and radiographic failure was defined as occurrence
of segmental collapse. With the collapse of the femoral
head as seen on a radiograph as the end point, survival for
all enrolled patients was calculated using the Kaplan-Meier
method and plotted. The differences in the survival distri-
butions were tested with the log-rank test. We performed
multivariate regression analysis using the Cox proportional
hazards model to identify the independent factors with
regard to the collapse of the head. We used SPSS software
(version 12.0; SPSS Inc, Chicago, IL).
MRI findings
Abnormal
Abnormal
Low signal intensity in necrotic segment
in T1-weighted and T2-weighted images
Low signal intensity in necrotic segment
in T1 and high signal intensity in necrotic
segment in T2
Volume 466, Number 5, May 2008 The Fate of Asymptomatic Osteonecrosis 1089

Results Table 2. Fate of asymptomatic hips according to Steinberg stage

[25]
At the latest followup examination, 31 of 81 hips (38%) Stage Hips Pain Collapse
became symptomatic and 26 hips (32%) demonstrated
collapse. With occurrence of collapse as the end point, the Stage I 58 (68%) 23 (42%) 20 (36%)
cumulative rates of survival (with 95% confidence inter- Stage II*
vals) were 91.4% at 5 years, 76.5% at 8 years, and 70% at Cystic type 18 (22%) 8 (44%) 6 (33%)
10 years (Fig. 1A). Five hips that were symptomatic had Sclerotic type 8 (10%) 0 (0%) 0 (0%)
only slight pain and did not collapse. The mean interval * Stage II was subdivided as Stage II sclerotic type and Stage II cystic
between diagnosis and collapse was 4.1 years (range, 1.2– type according to a method of Bozic et al. [3].
11.9 years). The mean interval between diagnosis and pain
onset was 3.4 years (range, 0.7–8.9 years). Twelve of the 47%). Only eight hips (10%) were classified as Stage II
26 hips (46%) demonstrated collapse after 4 years; the sclerotic type and neither had pain nor demonstrated col-
longest delay between diagnosis and collapse was lapse. The survival rate for those hips was 100% (Table 2).
11.9 years. When collapse occurred, it was always Lesion size predicted survival likelihood and duration.
preceded by pain of an average 8 months’ duration (range, The survival rate for hips with small lesions was 100%.
1–36 months). The survival rate for hips with medium lesions was 36.4%
We observed no differences in survival (p = 0.060, log- (95% CI, 0%–73%). However, the survival rate for hips
rank test) among hips classified as Steinberg stages with large lesions was 0% (Table 3). The log-rank test
(Fig. 1B). The survival rate for Stage I hips was 21.9% showed longer durations (p = 0.000) of survival for hips
(95% confidence interval [CI], 10%–54%). The survival with small or medium lesions than for hips with large
rate for Stage II cystic-type hips was 17% (95% CI, 13%– lesions (Fig. 1C).

Fig. 1A–D The Kaplan-Meier sur-

vivorship curve. (A) The cumulative
rates of survival (with 95% confi-
dence intervals) are 91.4% at
5 years, 76.5% at 8 years, and 70%
at 10 years with collapse of the
femoral head as the end point. Error
bars show the 95% confidence inter-
val. (B) Survival rates according to
initial Steinberg stages of osteone-
crosis. There were no differences
(p = 0.060, log-rank test) in sur-
vival among hips classified as
Steinberg stages. Error bars show
the 95% confidence interval. (C)
Survival rates according to initial
extent of the osteonecrosis by MRI
using the method of Steinberg et al.
[25]. The times to collapse were
different (p = 0.000, log-rank test)
among the three groups. Error bars
show the 95% confidence interval.
(D) Survival rates according to
the location of osteonecrosis. The
times to collapse were different
(p = 0.000, log-rank test) among
the three groups. Error bars show the
95% confidence interval.

123
1090 Min et al. Clinical Orthopaedics and Related Research

Table 3. Fate of asymptomatic hips according to extent of the Table 5. Relationship between the final status of femoral head and
necrotic lesion* various clinical and radiographic parameters for patients with
asymptomatic osteonecrosis of the femoral head
Extent Hips Pain Collapse
Patient characteristics No collapse Collapse p Value
Small (greater than 15%) 30 (37%) 0 (0%) 0 (0%)
Medium (15% to 30%) 29 (36%) 9 (31%) 6 (21%) Number of hips: 81 55 hips (68%) 26 hips (32%)
Large (greater than 30%) 22 (27%) 22 (100%) 20 (91%) Age at surgery (years) 49.2 ± 12.3 53.2 ± 10.2 0.154
Gender 0.592
* The extent of the necrotic lesion was quantified with using the
method of Steinberg et al. [25] in which hips were evaluated by Male 47 21
simple visual estimate of lesion size. Female 8 5
Length of followup (years) 8.4 ± 3.0 8.3 ± 3.6 0.167
Weight (kg) 63.5 ± 6.7 60.4 ± 8.8 0.086
Table 4. Fate of asymptomatic hips according to location of the
necrotic lesion* Presumed cause of necrosis 0.146
Alcohol 24 15
Location Hips Pain Collapse
Idiopathic 20 10
Type A 3 (3%) 0 (0%) 0 (0%) Steroid 11 1
Type B 35 (43%) 1 (3%) 0 (0%) All values are presented as the mean ± standard deviation.
Type C1 15 (19%) 2 (13%) 2 (13%)
Type C2 28 (35%) 28 (100%) 24 (86%) stage, extent of the necrotic lesion, and lesion location
* The location of the necrosis on T1-weighted midcoronal images was influenced the rate of collapse of the asymptomatic ON of
classified as Type A, B, C1, or C2 using the criteria described by the femoral head and if so, how long after symptom onset
Sugano et al. [27]. and initial diagnosis of ON. Because of the potential
influence of demographic factors on collapse, we then
The log-rank test showed longer durations (p = 0.000) asked whether these would predict collapse.
of survival for the hips with Type A, B, or C1 necrotic Our study had certain limitations, including the small
lesions than for those with Type C2 lesion (Fig. 1D). The size of certain radiographic subgroups and the heteroge-
survival rate for hips with Type A or B (lesions occupying neity of patient demographics other than presumed causes
less than the medial two-thirds) was 100%. The survival of ON. We also did not use quantitative volumetric mea-
rate for hips with Type C1 (lesions occupying more than surements to determine lesion size by digital image
the medial two-thirds but not extending laterally to the analysis, a method that has been proven more accurate than
acetabular edge) was 77.8% (95% CI, 48%–100%). The angular measurement [15, 24, 25]. In addition, we exclu-
extent of necrosis in these two hips was classified as large. sively included patients with asymptomatic hips on one
The survival rate for Type C2 lesions (lesions occupying side and symptomatic hips on the other side. The outcome
more than the medial two-thirds and extending laterally to of patients with both hips asymptomatic might be different
the acetabular edge) was 4.2% (95% CI, 0%–12%) from what we observed. Asymptomatic hips will more
(Table 4). likely have progressive collapse or symptoms when there
We observed no correlation between femoral head col- the have symptoms in the contralateral hip [10].
lapse and patients’ age, gender, weight, presumed cause of We found 31 of 81 (38%) hips became symptomatic and
ON, or length of followup (Table 5). When all possible 26 hips (32%) demonstrated collapse. The collapse was
factors were analyzed together using a Cox model, the related to the extent and location of the necrotic lesion on
extent of large necrotic lesions (hazard ratio, 4.06; 95% CI, MRI. The mean interval between diagnosis of the asymp-
1.29–12.77; p = 0.016) and location of Type C2 necrotic tomatic ON and collapse was 4.1 years. The clinical and
lesions (hazard ratio, 6.35; 95% CI, 1.18–34.11; radiographic failure rates in our study are lower than those
p = 0.031) were risk factors for collapse. of previous reports (Table 6). Discrepancies in disease
prognosis between previous studies and our study arise
from several sources such as differences in study popula-
Discussion tions, in length of followup periods, and in techniques for
measuring the extent or location of necrotic lesions.
There is no consensus regarding the factors predicting Jergessen and Khan [14] reported 14 of 19 asymptomatic
femoral head collapse in asymptomatic ON of the hip, hips with Steinberg Stage II disease had progression of the
although considerable evidence relates predictive factors disease and collapse. However, they did not use MRI
(eg, size and location of lesion) to outcomes in symptom- images for diagnosis and measuring the extent of necrotic
atic ON. We therefore asked whether the radiographic lesions and may have excluded hips that in fact had ON

123
Volume 466, Number 5, May 2008 The Fate of Asymptomatic Osteonecrosis 1091

Table 6. Review of the literature regarding outcomes in patients occurrence of collapse between Stage I and Stage II (36%
with untreated asymptomatic osteonecrosis of the femoral head and 23%, respectively). However, six of 18 hips with Stage
Authors Number Length of Symptomatic Collapse II cystic type demonstrated collapse, and none of eight hips
of hips followup progression (%) with Stage II sclerotic type showed collapse when we
(years)* (%) subdivided Stage II. This finding is consistent with the
Hernigou et al. [9] 121 14 91 77 observations of a previous report [3]. Bozic et al. [3]
Kopecky et al. [17] 25 1.3 28 28 reported cystic changes in the femoral head as seen on
Takatori et al. [29] 32 1.7 Not 43.8 plain radiographs were associated with a more than four-
available fold increase in the rate of failure after core decompression.
Jergessen and 19 [5 73.7 71.4 Previous studies demonstrate hips with large necrotic
Khan [14] lesions have a high possibility of collapse [13, 16, 23, 24,
Hernigou et al. [11] 40 11 88 73 29]. In our series large necrotic lesions were not observed
Bradway and 15 Not 100 100 in eight hips with Stage II sclerotic disease but were in five
Morrey [4] available of 18 hips with Stage II cystic disease. Some investigators
* All values are presented as the mean. have suggested sclerotic changes in necrotic lesions indi-
cate a repair process that inhibits bone resorption
that subsequently did not collapse, thus biasing their group. mechanism and provides structural integrity in the femoral
This throws into question the older literature using only head [13, 18]. On the other hand, cystic changes within
radiographs. Bradway and Morrey [4] also reported all 15 necrotic lesions may represent poor trabeculae and exten-
hips in their study of asymptomatic hips eventually col- sive bone resorption by osteoclastic resorption with greater
lapsed. In a prospective study of 40 asymptomatic hips potential for subsequent collapse [7, 13].
with very small osteonecrotic lesions, Hernigou et al. [11] The location of the necrotic lesion also influences the
reported 88% became symptomatic and 73% showed rate of collapse [22, 24–29]. Sugano et al. [28] originally
collapse. Recently, Hernigou et al. [9] reported pain proposed an MRI classification system that includes Types
developed in 91% and collapse had occurred in 77% in A, B, and C lesions; they reported Type A lesions are
asymptomatic ON associated with sickle cell disease. medial and rarely progress, Type B lesions are central and
However, several investigators [5, 17, 21] have shown have intermediate progression, and Type C lesions have
using MRI certain small lesions spontaneously heal and lateral involvement of the head and the worst prognosis.
decrease in size over time. However, Ito et al. [13] reported the location of the
A number of factors may influence the rate of progres- necrotic lesion was not a major factor. In 2001, Japanese
sion of ON of the femoral head [11, 13, 14–17, 19, 23–26]. investigators [27] revised the classification system of
The extent of the necrotic lesion is generally considered the location and subdivided Type C lesions into Subtypes C1
principal factor in determining the risk of collapse [11, 16, and C2 because the incidence of progressive collapse of the
29]. Takatori et al. [29] reported a close relationship femoral head varies considerably between them [20, 27].
between the size of the necrotic lesion and collapse. Koo Our data also suggest the risk of collapse is higher in Type
and Kim [16] reported collapse in 97% of the medium and C2 lesions (86%) than in Type C1 lesions (14%). We
large lesions but in only 12.5% of smaller lesions. Our believe this explains the lack of observed difference in the
findings support the finding that the more extensive the risk of collapse between Type B and Type C lesions in
necrotic lesion, the higher the risk of collapse. previous reports [13, 21] that do not differentiate C1 and
Surprisingly, we found a higher survival rate for Stage II C2 lesions.
sclerotic hips than for Stage I and Stage II cystic hips. Most Our study confirmed two important prognostic factors
researchers have reported ON of the femoral head gener- for collapse: extent and location of the necrotic lesion.
ally has a progressive course of collapse once changes are Although it was more likely to occur in hips with large
apparent radiographically [1, 12, 24]. However, asymp- lesions, collapse also occurred in medium-sized lesions if
tomatic hips in the early stage of ON have variable paths of they were localized to the lateral area of the femoral head.
disease progression [13, 14, 17, 22, 24, 29]. Ito et al. [13] Six of eight hips with medium-sized lesions but with C2
observed there was no difference in occurrence of pro- location (lateral area of the femoral head) developed
gression between hips with Steinberg Stage I disease and collapse.
hips with Stage II disease. Shimizu et al. [23] reported A question frequently asked by our patients with ON of
survival rates for the asymptomatic hips with normal the femoral head is how long the surgery on their symp-
radiographic findings were better than for hips with tomatic hip will last until they need a second operation
abnormal radiographic findings. When we classified for the asymptomatic contralateral hip. We found it took
hips by radiographic stage, there was no difference in an average of 4.1 years (but with a wide range of

123
1092 Min et al.
1.2–11.9 years) from MRI diagnosis to collapse with
nearly 50% of the collapse occurring as soon as 4 years
after diagnosis.
The occurrence of pain in the asymptomatic hip was a
good predictor of collapse in our study. We recommend
evaluating these patients immediately after pain onset
because pain usually precedes collapse by an average of
8 months.
References
1. Aaron RK, Lennox D, Bunce GE, Ebert T. The conservative
treatment of osteonecrosis of the femoral head. A comparison of
core decompression and pulsing electromagnetic fields. Clin
Orthop Relat Res. 1989;249:209–218.
2. Bassett LW, Gold RH, Reicher M, Bennett LR, Tooke SM.
Magnetic resonance imaging in the early diagnosis of ischemic
necrosis of the femoral head. Clin Orthop Relat Res. 1987;214:
237–248.
3. Bozic KJ, Zurakowski D, Thornhill TS. Survivorship analysis of hips
treated with core decompression for nontraumatic osteonecrosis of
the femoral head. J Bone Joint Surg Am. 1999;81:200–209.
4. Bradway JK, Morrey BF. The natural history of the silent hip in
bilateral atraumatic osteonecrosis. J Arthroplasty. 1993;8:383–387.
5. Cheng EY, Thongtrangan I, Laorr A, Saleh KJ. Spontaneous
resolution of osteonecrosis of the femoral head. J Bone Joint Surg
Am. 2004;86:2594–2599.
6. Ficat RP. Idiopathic bone necrosis of the femoral head. Early
diagnosis and treatment. J Bone Joint Surg Br. 1985;67:3–9.
7. Glimcher MJ, Kenzora JE. The biology of osteonecrosis of the
human femoral head and its clinical implications: II. The path-
ologic changes in the femoral head as an organ and in the hip
joint. Clin Orthop Relat Res. 1979;139:283–312.
8. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
9. Hernigou P, Habibi A, Bachir D, Galacteros F. The natural history
of asymptomatic osteonecrosis of the femoral head in adults with
sickle cell disease. J Bone Joint Surg Am. 2006;88:2565–2572.
10. Hernigou P, Lambotte JC. Bilateral hip osteonecrosis: influence
of hip size on outcome. Ann Rheum Dis. 2000;59:817–821.
11. Hernigou P, Poigrand A, Nogier A, Manicom O. Fate of very
small asymptomatic stage-I osteonecrotic lesions of the hip.
J Bone Joint Surg Am. 2004;86:2589–2593.
12. Hungerford DS, Zizic TM. The treatment of ischemic necrosis of
bone in systemic lupus erythematosus. Medicine. 1980;59:143–148.
13. Ito H, Matsuno T, Kaneda K. Prognosis of early stage avascular
necrosis of the femoral head. Clin Orthop Relat Res.
1999;358:149–157.
123
Clinical Orthopaedics and Related Research
14. Jergessen HE, Khan AS. The natural history of untreated
asymptomatic hips in patients who have non-traumatic osteone-
crosis. J Bone Joint Surg Am. 1997;79:359–363.
15. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. The
conservative surgical treatment of idiopathic aseptic necrosis of
the femoral head. J Bone Joint Surg Br. 1974;56:291–296.
16. Koo KH, Kim R. Quantifying the extent of osteonecrosis of the
femoral head. A new methods using MRI. J Bone Joint Surg Br.
1995;77:875–880.
17. Kopecky KK, Braunstein EM, Brandt KD, Filo RS, Leapman SB,
Capello WN, Klatte EC. Apparent avascular necrosis of the hip:
appearance and spontaneous resolution of MR findings in renal
allograft recipients. Radiology. 1991;179:523–527.
18. Lee CK, Hansen HT, Weiss AB. The ‘silent hip’ of idiopathic
ischemic necrosis of the femoral head in adults. J Bone Joint Surg
Am. 1980;62:795–800.
19. Mont MA, Jones LC, Pacheo I, Hungerford DS. Radiographic
predictors of outcome of core decompression for hips with
osteonecrosis stage III. Clin Orthop Relat Res. 1998;354:159–
168.
20. Nishi T, Sugano N, Miki H, Hashimoto J, Yoshiokawa H. Does
alendronate prevent collapse in osteonecrosis of the femoral
head? Clin Orthop Relat Res. 2006;443:273–279.
21. Nishi T, Sugano N, Ohzono K, Sakai T, Haraguchi K, Yoshikawa
H. Progression and cessation of collapse in osteonecrosis of the
femoral head. Clin Orthop Relat Res. 2002;400:149–157.
22. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
23. Shimizu K, Moriya H, Sakamoto M, Suguro T. Prediction of
collapse with magnetic resonance imaging of avascular necrosis.
J Bone Joint Surg Am. 1994;76:215–223.
24. Steinberg DR, Steinberg ME, Garino JP, Dalinka M, Udupa JK.
Determining lesion size in osteonecrosis of the femoral head.
J Bone Joint Surg Am. 2006;88(Suppl 3):27–34.
25. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system
for staging avascular necrosis. J Bone Joint Surg Br. 1995;77:
34–41.
26. Stulberg BN, Davis AW, Bauer TW, Levine M, Easley K.
Osteonecrosis of the femoral head. A prospective randomized
treatment protocol. Clin Orthop Relat Res. 1991;268:140–151.
27. Sugano N, Atsumi T, Ohzono K, Kubo T, Hotokebuchi T,
Takaoka K. The 2001 revised criteria for diagnosis, classification,
and staging of idiopathic osteonecrosis of the femoral head.
J Orthop Sci. 2002;7:601–605.
28. Sugano N, Ohzono K, Masuhara K, Takaoka K, Ono K. Prog-
nostication of osteonecrosis of the femoral head with systemic
lupus erythematosus by magnetic resonance imaging. Clin Ort-
hop Relat Res. 1994;305:190–199.
29. Takatori Y, Kokubo T, Hinomiya S, Nakamura S, Morimoto S,
Kusaba I. Avascular necrosis of the femoral head. Natural history
and magnetic resonance imaging. J Bone Joint Surg Br.
1993;75:217–221.
Clin Orthop Relat Res (2008) 466:1093–1103
DOI 10.1007/s11999-008-0184-9
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Do Modern Techniques Improve Core Decompression Outcomes
for Hip Osteonecrosis?
David R. Marker BS, Thorsten M. Seyler MD,
Slif D. Ulrich MD, Siddharth Srivastava BA,
Michael A. Mont MD
Ó The Association of Bone and Joint Surgeons 2008
Abstract Core decompression procedures have been used
in osteonecrosis of the femoral head to attempt to delay the
joint destruction that may necessitate hip arthroplasty. The
efficacy of core decompressions has been variable with
many variations of technique described. To determine
whether the efficacy of this procedure has improved during
the last 15 years using modern techniques, we compared
recently reported radiographic and clinical success rates to
results of surgeries performed before 1992. Additionally,
we evaluated the outcomes of our cohort of 52 patients
(79 hips) who were treated with multiple small-diameter
drillings. There was a decrease in the proportion of patients
undergoing additional surgeries and an increase in radio-
graphic success when comparing pre-1992 results to
patients treated in the last 15 years. However, there were
fewer Stage III hips in the more recent reports, suggesting
that patient selection was an important reason for this
improvement. The results of the small-diameter drilling
cohort were similar to other recent reports. Patients who had
small lesions and were Ficat Stage I had the best results with
79% showing no radiographic progression. Our study
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
D. R. Marker, T. M. Seyler, S. D. Ulrich, S. Srivastava,
M. A. Mont (&)
Rubin Institute of Advanced Orthopedics, Center for Joint
Preservation and Reconstruction, Sinai Hospital of Baltimore,
2401 West Belvedere Avenue, Baltimore, MD 21215, USA
e-mail: mmont@lifebridgehealth.org; Rhondamont@aol.com
confirms core decompression is a safe and effective pro-
cedure for treating early stage femoral head osteonecrosis.
Level of Evidence: Level IV, therapeutic study (see the
Guidelines for Authors for a complete description of levels
of evidence).
Introduction
Various techniques for performing core decompression
have been used to save the osteonecrotic femoral head.
There is also considerable disagreement as to the degree of
efficacy of this procedure, how it might help, and the level
of influence of various patient factors (such as a history of
alcohol abuse or smoking, corticosteroid use, as well as
underlying diagnoses such as systemic lupus erythematosus
or sickle cell anemia) and radiographic lesion character-
izations (such as presence or degree of collapse, lesion size
or location).
The technique of performing core decompression has
varied in terms of surgical approaches, number of drillings,
and the diameter of the trephines. A number of authors
have advocated the use of small-diameter percutaneous
drilling and believe that it as effective as large-diameter
core decompression procedures [56, 73, 95]. Some authors
have supplemented core decompression with electrical
stimulation [79] or growth and differentiation factors [19,
24, 82]. Other studies have reported adjunctive vascular-
ized [96] and/or nonvascularized bone grafting [35, 63].
Vascularized fibular grafting is essentially a large core
decompression procedure with the introduction of a vas-
cularized fibula, ilium, or trochanteric bone on a more local
pedicle. While vascularized and nonvascularized long
cortical strut bone grafting approaches could be considered
variations of core decompression procedures, we believe
123
1094 Marker et al.
these procedures are sufficiently different that they should
be considered as alternate approaches, rather than varia-
tions of core decompression and will not be considered in
this study.
The primary question we asked was whether the efficacy
of core decompression, measured in terms of decreased
proportion of patients having additional surgeries or
showing radiographic progression to collapse, has
improved during the last 15 years using modern tech-
niques. Using these same measures of efficacy, we also
asked whether modern core decompression techniques
provide better outcomes than those reported in studies
using non-operative treatment. Secondary questions were:
(1) whether the clinical and radiographic outcomes of hip
osteonecrosis patients who were treated using a recently
developed small-diameter drilling core decompression
technique were similar to other modern studies; and (2)
whether patients who had less radiographic progression and
smaller lesion sizes at the time of treatment using small-
diameter drilling would be less likely to have poor out-
comes with subsequent collapse and the need for additional
more invasive surgeries.
Materials and Methods
We systematically reviewed the literature on the Medline
and EMBASE bibliographic databases that were related to
core decompression and osteonecrosis of the hip. The ini-
tial search parameters used to identify potentially relevant
articles were ‘‘necrosis and hip and decompression.’’ We
then searched bibliographies of review articles for any
additional relevant studies. Two of us (DRM, TMS)
screened all articles according to a previously defined
protocol [94]. The following inclusion/exclusion criteria
were used: (1) The report provided radiographic outcomes
and/or indicated whether patients underwent additional
surgeries following an initial core decompression for the
treatment of osteonecrosis of the hip; (2) We excluded
reports that did not provide sufficient data to analyze out-
comes or involved fewer than 10 patients, for example a
report of a single patient treated with a powered core
decompression [50]; (3) Only the most recent studies were
included for patient cohorts reported at multiple times at
different followups; (4) Although some reports included
patients who were younger than 18 years old, we excluded
studies that focused only on adolescent patients [84]; (5)
We did not include reports that used long cortical strut
bone grafting or vascularized bone grafting. We did include
studies that reported the use of ancillary cancellous bone
grafting such as the technique reported by Steinberg et al.
[82]; (6) Studies with a mean followup of less than
18 months were excluded (see below for this exclusion
123
Clinical Orthopaedics and Related Research
rationale) [10, 40, 44, 65, 91]; (7) We also included the
previously unpublished results of patients at our institution
that were treated using a small-diameter drilling technique.
The criteria, which required a minimum mean 18-month
followup for study inclusion, were used because it was
believed unreasonable to consider shorter term followups
when trying to assess efficacy and ‘‘failure’’ of these pro-
cedures. Eighteen months was utilized as approximately
one standard deviation above the mean time to collapse of
multiple studies (11 months). It can be difficult to deter-
mine the exact time to femoral head collapse, which may
predict when a patient needs a hip replacement. This could
occur fairly soon or months after head collapse when the
patients’ hips become more symptomatic. An example of a
study with data for mean time to collapse was from our
patients who had percutaneous drilling. In this study
patients had a mean time to detected femoral head collapse
of 11 months which led to needing a total hip replacement
at a mean of 14 months. For the purpose of this report, we
used the mean of 11 months plus one standard deviation
(6.9 months) to determine the previously noted minimum
mean followup of 18 months for the studies in our litera-
ture review.
We made an attempt to stratify all studies that met our
inclusion/exclusion criteria into two groups according to
when the reported procedures were performed: before
1992, and from 1992 to 2007. When the dates of surgery
were not specifically noted in the study, the followup and
year the study was published were used to estimate the
period in which the surgeries were performed. Some
studies reported procedures both before and after 1992. For
these studies, attempts were made to subgroup each patient
according to when the procedure was performed. However,
because it was impossible to stratify the patients for some
reports, we categorized these studies by when the majority
of the patients were treated. There were five studies clas-
sified as pre-1992 based on these criteria [7, 52, 54, 70, 82].
For each report included in this study, the level of evi-
dence was determined using the Clinical Orthopaedics and
Related Research guidelines [14]. The demographic data
fields analyzed included: etiology/associated risk factors,
age, followup, and preoperative stage of the disease as
defined by Ficat [18]. The outcome parameters collected
for each report were the number and percentage of addi-
tional surgeries and radiographic failures. Additional
surgeries were only included if they were directly related to
progression of the osteonecrosis. For example, if a patient
had an evacuation of a hematoma it would not have been
included as a case that underwent additional surgery. Due
to the variability in the modalities used in the studies to
assess radiographic outcomes (Fig. 1), progression to col-
lapse or advancement after collapse was defined as
radiographic failure for this study (Table 1). Radiographic
Volume 466, Number 5, May 2008 Modern Core Decompression Techniques 1095

Fig. 1 The Ficat and Arlet system [18] has historically been the most
frequently used staging modality. However, as noted in this graph, a
large percentage of recent core decompression studies have reported
using various other radiographic staging systems such as the
Pennsylvania [81], ARCO [55], and Ohzono classifications [60].
Fig. 2 This figure provides the percentage of studies that used
various clinical assessment modalities. The Harris hip score [22] and
the Merle d’Aubigné-Postel scale [49] were the two most common
evaluation methods used to assess clinical outcomes.

outcomes were excluded for studies that did not indicate

whether radiographic progression was to collapse [15, 41,
42, 70, 71] or if success was defined only in terms of a
combination of radiographic and clinical failure without
stratification [88, 97]. An attempt was made to also
compare reported clinical outcomes. However, it was
determined that the question of whether there were any
differences was unanswerable using the literature given the
variability and the inconsistency in clinical evaluation
criteria used by the studies (Fig. 2).
We identified 47 studies that reported on the outcome of
core decompression in hip osteonecrosis and met our
inclusion criteria. Approximately half (25 of 47, 53%) of Fig. 3 The studies reviewed in our meta-analysis were grouped
according to their levels of evidence [14], and the proportion of
these reports were Level of Evidence IV, and 6% (n = 3) studies for each level is presented in this chart. There have been
were conducted at Level I (Fig. 3). Alcohol abuse and relatively few randomized, prospective studies concerning osteone-
corticosteroid usage were the most frequently cited risk crosis of the hip, and the majority of the reports have been level of
factors (Fig. 4). Overall, there were 2,605 hips treated with evidence IV.
core decompression. From studies reporting relevant
demographic data, the mean age for patients was 39 years
(range, 12–83 years), and the minimum followup was physicians continue to utilize nonoperative treatment
1 month (mean, 64 months; range, 1–216 months). methods. To compare the results of core decompression to
While we do not consider withholding surgery an a baseline of natural progression, we conducted a separate
appropriate option based on previous studies showing literature search using the same criteria to identify a group
outcomes that are less efficacious than interventional pro- of patients who were treated by nonoperative measures.
cedures used at our institution [51], we recognize that some Because the purpose of this review was to assess natural

Table 1. Criteria for assessing effectiveness of core decompressions

Measure Inclusion/exclusion criteria Examples

Additional surgery 1. Include additional surgeries associated with Total hip arthroplasty, vascularized bone
progression of osteonecrosis. grafting, osteotomy
2. Exclude surgeries not directly related to long-term Evacuation of a hematoma
failure of core decompression.
Radiographic failure 1. Include progression to collapse. Progression from Ficat II to III.
2. Include progression from collapse to further Progression from Steinberg IV to V.
stage of degeneration.
3. Exclude progression without collapse.  Progression from ARCO I to II.
 
Studies that only indicated ‘‘progression’’ in stage without indicating whether the progression was to collapse were excluded from our analysis.

123
1096 Marker et al.
Fig. 4 The most frequently reported etiology/risk factors are listed
and the number of studies in our meta-analysis that reported the
outcomes of patients who were diagnosed with each of these factors is
noted.
progression, we excluded nonoperative treatment modali-
ties using external electrical therapy, ultrasound therapy, or
pharmacological agents [39, 78, 90]. The mean age for
these studies was 38 years (range, 13–79 years) and the
minimum followup was 3 months (mean, 54 months;
range, 3–240 months). The same outcome data was col-
lected for these studies as for the review of core
decompression reports.
From our institution, we identified 52 consecutive
patients (79 hips) who had a core decompression utilizing a
multiple small-diameter drilling (3.2–3.4 mm) technique
with a minimum followup of 36 months (mean, 65 months;
range, 36–81 months). The surgical technique used for
these patients and the initial short-term followup of the first
45 hips was previously reported [56]. The most common
risk factors in this cohort of patients were corticosteroids
(n = 47 hips), tobacco abuse (n = 26 hips), and systemic
lupus erythematosus (n = 20 hips) with some hips having
multiple risk factors. Patients were assessed preoperatively
and at final followup using the Harris hip score [22] and the
Ficat and Arlet staging system [18] for clinical and radio-
graphic evaluations, respectively. Additionally, lesion size
was measured using the combined necrotic angle as
described by Kerboul et al. [34]. For Stage I hips or patients
in whom the lesion was not seen on radiographs, magnetic
resonance imaging was used to determine the lesion size.
Patients with collapse (Ficat Stage III or greater) were not
candidates for this procedure. The radiographic evaluations
were conducted by two of the authors (TMS, SDU). We
evaluated the overall effectiveness of the small-diameter
core decompression technique by combining the results of
our study with those of a previously published small-
diameter drilling study by Song et al. [73] and compared the
proportions of patients who had radiographic failures or
underwent additional surgeries to the outcomes of the other
modern studies published since 1992.
123
Clinical Orthopaedics and Related Research
To address the specific questions asked in this study, we
compared the following groups: (1) procedures before
1992; (2) procedures from 1992 forward; (3) reports of
nonoperative treatment; and (4) reports using the multiple
small-diameter drilling technique. The number and per-
centage of additional surgeries and radiographic failures
were stratified by Ficat stage when possible. For our per-
cutaneous multiple small-diameter drilling cohort we also
stratified the results by lesion size. A chi-square analysis
was used to compare the differences in outcomes for all the
groups that were evaluated. The key variable used for the
power analysis was the difference in proportions of patients
who underwent additional surgery in the pre-1992 studies
compared to the studies from 1992 to 2007. A power
analysis was conducted to ensure the comparison of failure
rates was sufficiently powered (p \ 0.05; power: 80%) to
reveal the p values necessary to answer the primary
research questions in this study. Prior studies that reported
on comparisons of core decompression techniques were
assessed to determine a clinically justifiable and appropri-
ate effect size [1, 20]. Based on these studies and the
success rates of core decompression that we have seen at
our institution, we determined that we would need a min-
imum proportions sample size of 186 hips to identify an
improvement from 60 percent to 45 percent of patients
undergoing additional surgery. All comparisons were
conducted using 95% confidence intervals where a p value
of less than 0.05 was considered significant. We used SPSS
version 13.0 software (SPSS Inc, Chicago, IL) for all
analyses.
Results
Overall, the success rates were higher for the studies that
reported core decompressions performed during the last
15 years (Table 2) compared to procedures performed
before 1992 (Table 3). From these reports, there were 1337
hips treated before 1992 and 1268 hips since 1992. The
proportion of patients surviving without additional surgery
increased (p \ 0.001) from 59% (range, 29%–85%) in the
earlier studies to 70% (range, 39%–100%) in the more
recent reports. Similarly, the radiographic success also
increased (p = 0.027) from 56% (range, 0–94%) for the
pre-1992 cohort to 63% (range, 22%–90%). Stratification
by Ficat stage (Table 4) showed there were fewer
(p \ 0.001) patients who were Ficat Stage III after 1992.
The reports of nonoperative treatment (Table 5) had
higher proportions of failures compared to the core
decompression studies from 1992 to 2007. There were 791
hips in 18 studies between 1960 and 2007. In the studies
that reported relevant data, the proportion of patients who
underwent surgery by final followup at a mean of 67%
Volume 466, Number 5, May 2008 Modern Core Decompression Techniques 1097

Table 2. Literature review of core decompression outcomes for 1992 to 2007 patient cohort studies
Author/Year Number of hips Months followup (range) Additional surgery (%) Radiographic failure (%)

Kane et al./1996 [33] 19 (24–60) 11 (58) 11 (58)

Markel et al./1996 [47] 54 (2–53) 26 (48) –
Chang et al./1997 [11] 84 57 (24 to 165) 22 (26) 59 (70)
Mazieres et al./1997 [48] 20 24 9 (45) 9 (45)
Powell et al./1997 [64] 34 48 9 (26) –
Iorio et al./1998 [30] 33 64 (24–120) 11 (33) 18 (55)
Scully et al./1998 [68] 98 (21–50) 52 (53) –
Chen et al./2000 [12] 27 28 (12–128) – 10 (37)
Lavernia and Sierra/2000 [41] 67 41 16 (24) –
Maniwa et al./2000 [46] 26 94 (53–164) 8 (31) –
Specchiulli et al./2000 [74] 20 67 4 (20) 4 (20)
Piperkovski/2001 [62] 39 48 4 (10) –
Yoon et al./2001 [97] 39 61 (24–118) 19 (49) –
Aigner et al./2002 [2] 45 69 (31–120) 7 (16) 12 (27)
Hernigou et al./2003a [23] 189 84 (60–132) 34 (18) 39 (21)
Wirtz et al./2003  [93] 51 (36–132) 18 (35) –
Gangji et al./2004a [20] 10 24 0 (0) 1 (10)
Gangji et al./2004 [20] 8 24 2 (25) 5 (63)
Lieberman et al./2004a [45] 17 53 (26–94) 3 (18) 3 (18)
Bellot et al./2005 [4] 31 (1–176) 19 (61) 19 (61)
Ha et al./2006 [21] 18 (50–96) – 14 (78)
Neumayr et al./2006 [59] 17 36 3 (18) –
Veillette et al./2006c [89] 58 24 (6–52) 9 (16) 16 (28)
  
Marker et al./2007b, 79 24 (20–39) 27 (34) 27 (34)
Shuler et al./2007c [69] 22 39 (27–59) 3 (14) 3 (14)
b
Song et al./2007 [73] 163 87 (60–134) 50 (31) –
Total 1268 63 (1–176)à 366 (30)àà 250 (37)¥
 
Previous study not listed includes Wirtz et al. [92];    Results of the present study. Previous study not listed includes Mont et al. [56];
à
Weighted average follow-up; àà Data for total of 1223 hips; ¥ Data for total of 680 hips; a biologics; b multiple small diameter drilling;
c
tantalum; – = Data meeting our definition of additional surgery or radiographic failure was not available.

(range, 14% to 91%) was statistically higher than the
modern reports (p \ 0.001). Similarly, the mean reported
radiographic failure rates at 72% (range, 41% to 100%)
were considerably higher (p \ 0.001). Only 164 natural
history patients were reported between 1992 and 2007,
although the clinical and radiographic failure rates were
similar between this group of patients and those evaluated
before 1992.
The results using the small-diameter drilling technique
at our institution combined with those reported by Song
et al. [73] were similar to other studies of the last 15 years
(Table 6). At our institution, there were 21 patients (27
hips, 34%) who underwent additional surgery. The distri-
bution of Harris hip scores by number of hips were: 25 (90
points or greater), 24 (80–89 points), seven (70–79 points),
and 23 (less than 70 points). Excluding the patients who
underwent additional surgery, the mean Harris hip score
was 89 points (range, 72–100 points). Two patients (three
hips) both had scores of 72 points but did not receive
additional treatment. The patient who had bilateral osteo-
necrosis reported moderate pain in both hips. The other
patient progressed from Ficat stage I to Ficat Stage II and
his reported pain scores increased from mild (30 points)
preoperatively to moderate (20 points) at final followup.
There were 27 hips (34%) that showed radiographic
progression of the disease to collapse following core
decompression.
Patients in our small-diameter drilling cohort with
higher Ficat stages and larger lesion sizes had increased
failure rates. The proportion of hips (n = 13, 59%) with a
large lesion (combined necrotic angle C 200°) that
underwent additional surgery was greater (p = 0.008) than
the proportion of hips (n = 14, 25%) that had small lesions
(a combined necrotic angle \ 200°) and underwent addi-
tional surgery. Similarly, the rate of additional surgery was
higher (p = 0.044) for hips that were Ficat Stage II (52%)

123
1098 Marker et al. Clinical Orthopaedics and Related Research

Table 3. Literature review of core decompression outcomes for pre-1992 patient cohort studies
Author/Year Number of hips Months follow-up (range) Additional surgery (%) Radiographic failure (%)

Solomon/1981 [72] 22 24 (6–48) 5 (23) –

Ficat/1985 [18] 133 114 (60–204) – 28 (21)
Camp and Colwell/1986 [9] 40 18 (3–40) 6 (15) 8 (20)
Hopson and Siverhus/1988 [28] 20 39 (12–78) 12 (57) –
Saito et al./1988 [67] 17 48 (24–168) – 9 (53)
Tooke et al./1988 [86] 45 36 (12–84) 16 (36) 16 (36)
Aaron et al./1989 [1] 50 26 28 (56) 32 (64)
Aaron et al./1989a [1] 56 27 18 (32) 22 (39)
Beltran et al./1990 [5] 34 23 (11–47) – 16 (47)
a
Trancik et al./1990 [87] 11 45 (24–60) 5 (45) 11 (100)
Kristensen et al./1991 [37] 18 39 (12–60) – 3 (17)
Stulberg et al./1991 [83] 28 27 8 (29) 21 (75)
Robinson and Springer/1993 [66] 19 48 3 (16) 4 (21)
Lafforgue et al./1993 [38] 27 46 – 17 (63)
Leder and Knahr/1993 [43] 47 44 (24–100) 9 (19) 11 (23)
Holman et al./1995 [27] 31 (18–67) 14 (45) 8 (40)*
Koo et al./1995 [36] 18 (minimum 24) – 14 (78)
Smith et al./1995 [71] 114 40 (24–78) 64 (56) –
Mont et al./1997  [52] 79 144 (48–216) 37 (47) –
Mont et al./1998 [53] 68 144 (48–216) 48 (71) 48 (71)
Bozic et al./1999 [7] 54 120 (24–196) 28 (52) 34 (62)
Simank et al./1999 [70] 94 72 (18–180) 32 (34) –
Steinberg et al./2001  ,a [82] 312 48 (3–155) 113 (36) –
Total 1337 65 (3–216)à 446 (41)àà 302 (44)¥
* Radiographic outcomes were only provided for 20 hips;   Previous studies not listed include Hungerford and Zizic [29] and Fairbank et al. [17];
  
Other studies not listed include Steinberg et al. [75, 77–80] and Israelite et al. [31]; à Weighted average follow-up; àà Data for total of 1090
hips; ¥ Data for total of 685 hips; a core decompression combined with electrical stimulation; – = Data meeting our definition of additional
surgery or radiographic failure was not available.

preoperatively, compared to Ficat Stage I (26%). The best
results were seen in patients who had small lesions and
Ficat Stage I prior to treatment with 79% of these hips
showing no radiographic stage progression.
Discussion
While core decompression is relatively commonly per-
formed for ON of the femoral head, the variations in
reported techniques and drilling procedures make it diffi-
cult to interpret the efficacy of these procedures. Some
recent reports using innovative techniques such as growth
and differentiation factors to fill the core decompression
tract suggest excellent results, although the literature con-
tains a wide variety of results. Because of the relatively
small number of procedures reported for each of these
studies reporting on varied techniques, we analyzed recent
techniques by comparing studies that reported procedures
that were performed before 1992 to reports that had
procedures between 1992 and 2007. The primary question
of our study was whether the outcomes reported in the
recent studies were better than those prior to 1992 in terms
of reduced proportions of patients having additional sur-
geries and/or showing radiographic signs of femoral head
collapse. Additionally, using these same measures, we
asked whether modern core decompression techniques
provided better outcomes than non-operative treatment.
One of the limitations of this study was the small
numbers of patients in many of the reports reviewed.
Another limitation was that in some cases it was difficult to
determine when the core decompressions were performed
in order to stratify the study as pre-1992 or 1992 to 2007.
However, we believe our approach of using the publication
date and the mean followup to estimate when procedures
were performed would correctly stratify the majority of the
studies that were close to our 1992 cutoff. In addition, there
were only midterm mean followups (range, 18 months to
144 months) for many studies, and the long-term outcome
of core decompression is unclear. Another limitation was

123
Volume 466, Number 5, May 2008 Modern Core Decompression Techniques 1099

Table 4. Comparison of historical and modern core decompression suggests that patient selection may have been the primary
studies reason for this gain as there were fewer Ficat Stage III
Data* Studies prior Studies from p-Value patients in the later studies. However, based on the
to 1992 1992 to 2007 improvements in clinical outcomes for Ficat Stage II hips,
it appears that modern core decompression techniques did
Demographic variables
provide improved outcomes for some subsets of patients.
Mean age (range) 39 (15–83) 39 (13–72) –
years years The literature review (Table 2) suggests patients who
Mean followup (range) 65 (3–216) 63 (1–176) – have hips with Ficat Stage III disease are more likely to
months months have radiographic progression, clinical failure, and have
Preoperative ficat stage additional surgeries, suggesting these patients may not be
Ficat Stage I 32% 29% 0.302 appropriate candidates for this procedure. Although there
Ficat Stage II 42% 52% \ 0.001  appears to have been increased patient selectivity in the
Ficat Stage III 27% 19% \ 0.001  past 15 years in terms of fewer Ficat Stage III hips being
Outcomes treated with core decompressions, a number of surgeons
Additional surgery continue to use this procedure. Based on the literature
Overall 41% 30% \ 0.001  review, there were 132 patients (18% of all patients in
Ficat Stage I 15% 20% 0.413 studies after 1992 that stratified hips by Ficat stage) who
Ficat Stage II 44% 35% 0.056 were Stage III and treated using core decompression. These
Ficat Stage III 67% 66% 0.939
patients were included in 9 of the 35 studies (26%) after
Radiographic failure
1992. A recent study by Tingart et al. [85] reported similar
Overall 44% 37% \ 0.001 
results. They reported 11% of surgeons they surveyed used
core decompression for patients who were Ficat Stage III
Ficat Stage I 22% 21% 0.919
or IV. While some surgeons may be using core decom-
Ficat Stage II 47% 48% 0.887
pression only as a pain-relieving procedure or assessing the
Ficat Stage III 66% 50% 0.708
potential efficacy of modern techniques in Stage III hips,
* Some studies did not stratify by Ficat stage and/or report both we continue to recommend that other treatment options
outcome measures.
 
such as total hip arthroplasty or resurfacing be used for
Values were statistically significant.
these difficult to treat patients.
Our own data from patients in whom we used small-
the level of evidence for the scientific literature reviewed. diameter multiple drilling also confirms that the prognosis
As previously noted, most of the studies were Level IV and is influenced by the extent of the lesion size (Table 3).
there were few Level I studies. There is a need for more These results are similar to a prospective study of 73 hips
prospective randomized multicenter studies that further by Steinberg et al. [76] which evaluated the effect of lesion
analyze some of these newer techniques which will need size on the outcome of core decompression. They defined
longer followup and larger patient numbers in the future. three groups based on lesion size: small, less than 15% of
Additionally, if standardized clinical and radiographic femoral head involvement; medium, 15% to 30%; and
evaluation criteria were adopted, future meta-analyses large, greater than 30%. The difference between the per-
could provide more valid comparisons across studies. The centage of patients who had small lesions and later
limitations of our assessment of the percutaneous multiple underwent total hip arthroplasty (7%) was lower than
small-diameter drilling technique were similar to those of patients with large lesions (33%) who received a total hip
other studies: a limited number of patients from a single arthroplasty.
center, no long-term followup, and lack of a randomized The overall success rate of our cohort of small-diameter
control group. Nevertheless, we do not believe these lim- multiple drilling patients was similar to two other recent
itations detract from the overall results of the present study, studies that used a similar technique. In one of these
as in general, the results of all of the different techniques studies, Yan et al. [95] reported an improvement in Harris
were somewhat comparable and appear better than the hip score from a mean of 58 points (range, 46–89 points)
natural history. preoperatively to a mean of 86 points (range, 70–94 points)
The meta-analysis and our cohort of multiple small- at a minimum 2-year followup. In the other study by Song
diameter drilling patients suggest that core decompression et al. [73], 79% of patients who had Ficat Stage I disease
provides fewer treatment failures than nonoperative treat- had no additional surgery at a minimum 5-year followup.
ment. Although there are improvements in overall success The rationale and advantages for the small-diameter dril-
rates for the procedures performed from 1992 to present, ling presented in these prior studies were that: (1) the small
the stratification of the meta-analysis data by Ficat stage diameter drill can more easily reach the anterior portion of

123
1100 Marker et al. Clinical Orthopaedics and Related Research

Table 5. Literature review of nonoperative treatment outcomes

Author/Year Number of hips Months followup (range) Additional surgery (%) Radiographic failure (%)

Coste et al./1965 [16] 100 24 – 73 (73)

Merle d’Aubigne et al./1965 [49] 90 36 (12–48) – 61 (68)
Boettcher et al./1970 [6] 5 (minimum 24) – 4 (80)
Zizic and Hungerford/1985 [98] 15 44 – 13 (87)
Musso et al./1986  [58] 50 30 34 (68) 41 (82)
Steinberg et al./1989 [79] 55 21 (6–120) 46 (84) –
Churchill and Spencer/1991 [13] 18 60 9 (50) 8 (44)
Ohzono et al./1991 [61] 115 63 – 78 (68)
Stulberg et al./1991 [83] 22 27 20 (91) 11 (50)
Robinson and Springer/1993 [66] 16 39 (24–61) 7 (44) 9 (56)
Bradway and Morrey/1993 [8] 15 23 (3–66) 13 (87) 15 (100)
Koo et al./1995 [36] 19 (minimum 24) – 15 (79)
Jergesen et al./1997 [32] 19 111 (51–81) 11 (58) 7 (41)*
Lai et al./2005 [39] 25 24 17 (68) 19 (76)
Ha et al./2006 [21] 19 (50–96) – 15 (79)
Hernigou et al./2006   [26] 121 168 (120–240) 91 (75) 93 (77)
Neumayr et al./2006 [59] 21 36 3 (14) –
Morse et al./2007 [57] 67 23 (17–31) 20 (30) –
Total 792 53 (3–240)à 271 (63)àà 455 (72)¥
* Radiographs were only available for 17 patients;   Previous study not listed includes Bassett et al. [3];    Previous studies not listed include
Hernigou et al. [23, 25]; à Weighted average follow-up; àà Data for total of 429 hips; ¥ Data for total of 630 hips; – = Data meeting our definition
of additional surgery or radiographic failure was not available.

Table 6. Multiple small-diameter drilling compared to other modern reports in the last decade confirms that core decompression
studies is a safe and effective procedure for the treatment of early
Data Small-diameter Other 1992–2007 p Value stages of osteonecrosis of the femoral head. Based on the
technique studies results of our experience as well as other studies, we will
use core decompression to treat patients who have early
Demographics
small- and medium-sized lesions and are Ficat Stage I or II.
Number of hips 242 1026 –
Additionally, the midterm followup of the multiple small-
Mean age (range) 39 (18–72) 39 (12–71) –
diameter core decompression patients at our institution
Mean followup (range) 80 (36–134) 58 (1–176) –
was longer than most studies, and had a success rate similar
Outcomes
to or higher than other reports, which makes this technique
Additional surgery 32% 29% 0.520
the authors’ preferred modality. However, prospective,
Radiographic failure 34% 37% 0.437 randomized studies are recommended to verify these
observations before this technique can be recommended as
the femoral head, an area frequently involved in osteone- a standard for practicing surgeons.
crosis; (2) there is minimal morbidity; (3) the risk of
weakening or penetrating the femoral head and injuring the
articular cartilage when using a large-diameter trephine for References
multiple drillings is potentially reduced; and (4) the risk of
stress risers that can lead to a subtrochanteric fracture is 1. Aaron RK, Lennox D, Bunce GE, Ebert T. The conservative
treatment of osteonecrosis of the femoral head. A comparison of
also reduced. core decompression and pulsing electromagnetic fields. Clin
The literature review and our data suggest recent Orthop Relat Res. 1989;249:209–218.
techniques provide better clinical scores or radiographic 2. Aigner N, Schneider W, Eberl V, Knahr K. Core decompression
outcomes than pre-1992 studies of core decompression. in early stages of femoral head osteonecrosis–an MRI-controlled
study. Int Orthop. 2002;26:31–35.
However, it is unclear whether this improvement is due 3. Bassett CA, Schink MM, Mitchell SN. Treatment of osteone-
to improved patient selection or surgical technique. At a crosis of the hip with specific, pulsed electromagnetic fields
minimum, the additional accumulation of successful (PEMFs): a preliminary clinical report. In: Arlet J, Ficat RP,

123
Volume 466, Number 5, May 2008
Hungerford DS (eds). Bone Circulation. Baltimore, MD: William
and Wilkins; 1984:343–354.
4. Bellot F, Havet E, Gabrion A, Meunier W, Mertl P, de Lestang
M. Core decompression of the femoral head for avascular
necrosis [in French]. Rev Chir Orthop Reparatrice Appar Mot.
2005;91:114–123.
5. Beltran J, Knight CT, Zuelzer WA, Morgan JP, Shwendeman LJ,
Chandnani VP, Mosure JC, Shaffer PB. Core decompression for
avascular necrosis of the femoral head: correlation between long-
term results and preoperative MR staging. Radiology.
1990;175:533–536.
6. Boettcher WG, Bonfiglio M, Smith K. Non-traumatic necrosis of
the femoral head. II. Experiences in treatment. J Bone Joint Surg
Am. 1970;52:322–329.
7. Bozic KJ, Zurakowski D, Thornhill TS. Survivorship analysis of
hips treated with core decompression for nontraumatic osteone-
crosis of the femoral head. J Bone Joint Surg Am. 1999;81:200–209.
8. Bradway JK, Morrey BF. The natural history of the silent hip
in bilateral atraumatic osteonecrosis. J Arthroplasty. 1993;8:
383–387.
9. Camp JF, Colwell CW Jr. Core decompression of the femoral head
for osteonecrosis. J Bone Joint Surg Am. 1986;68:1313–1319.
10. Chan TW, Dalinka MK, Steinberg ME, Kressel HY. MRI
appearance of femoral head osteonecrosis following core decom-
pression and bone grafting. Skeletal Radiol. 1991;20:103–107.
11. Chang MC, Chen TH, Lo WH. Core decompression in treating
ischemic necrosis of the femoral head. Zhonghua Yi Xue Za Zhi
(Taipei). 1997;60:130–136.
12. Chen CH, Chang JK, Huang KY, Hung SH, Lin GT, Lin SY.
Core decompression for osteonecrosis of the femoral head at pre-
collapse stage. Kaohsiung J Med Sci. 2000;16:76–82.
13. Churchill MA, Spencer JD. End-stage avascular necrosis of bone
in renal transplant patients. The natural history. J Bone Joint Surg
Br. 1991;73:618–620.
14. Clinical Orthopaedics and Related Research. Author Resources.
Available at: http://edmgr.ovid.com/corr/accounts/ifauth.htm.
Accessed April 1, 2007.
15. Coleman BG, Kressel HY, Dalinka MK, Scheibler ML, Burk DL,
Cohen EK. Radiographically negative avascular necrosis: detec-
tion with MR imaging. Radiology. 1988;168:525–528.
16. Coste F, Merle DAR, Postel M, Massias P, Gueguen J, Grellat P.
Course of primary osteonecrosis of the femoral head (pon) and
therapeutic prospects [in French]. Presse Med. 1965;73:263–267.
17. Fairbank AC, Bhatia D, Jinnah RH, Hungerford DS. Long-term
results of core decompression for ischaemic necrosis of the
femoral head. J Bone Joint Surg Br. 1995;77:42–49.
18. Ficat RP. Idiopathic bone necrosis of the femoral head. Early
diagnosis and treatment. J Bone Joint Surg Br. 1985;67:3–9.
19. Gangji V, Hauzeur JP. Treatment of osteonecrosis of the femoral
head with implantation of autologous bone-marrow cells. Surgi-
cal technique. J Bone Joint Surg Am. 2005;87 Suppl 1:106–112.
20. Gangji V, Hauzeur JP, Matos C, De Maertelaer V, Toungouz M,
Lambermont M. Treatment of osteonecrosis of the femoral head
with implantation of autologous bone-marrow cells. A pilot
study. J Bone Joint Surg Am. 2004;86:1153–1160.
21. Ha YC, Jung WH, Kim JR, Seong NH, Kim SY, Koo KH. Pre-
diction of collapse in femoral head osteonecrosis: a modified
Kerboul method with use of magnetic resonance images. J Bone
Joint Surg Am. 2006;88 Suppl 3:35–40.
22. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
23. Hernigou P, Bachir D, Galacteros F. The natural history of
symptomatic osteonecrosis in adults with sickle-cell disease.
J Bone Joint Surg Am. 2003;85:500–504.
Modern Core Decompression Techniques 1101
24. Hernigou P, Beaujean F. Treatment of osteonecrosis with autol-
ogous bone marrow grafting. Clin Orthop Relat Res. 2002;405:
14–23.
25. Hernigou P, Galacteros F, Bachir D, Goutallier D. Deformities of
the hip in adults who have sickle-cell disease and had avascular
necrosis in childhood. A natural history of fifty-two patients.
J Bone Joint Surg Am. 1991;73:81–92.
26. Hernigou P, Habibi A, Bachir D, Galacteros F. The natural history
of asymptomatic osteonecrosis of the femoral head in adults with
sickle cell disease. J Bone Joint Surg Am. 2006;88:2565–2572.
27. Holman AJ, Gardner GC, Richardson ML, Simkin PA. Quanti-
tative magnetic resonance imaging predicts clinical outcome of
core decompression for osteonecrosis of the femoral head.
J Rheumatol. 1995;22:1929–1933.
28. Hopson CN, Siverhus SW. Ischemic necrosis of the femoral head.
Treatment by core decompression. J Bone Joint Surg Am. 1988;
70:1048–1051.
29. Hungerford DS, Zizic TM. Alcoholism associated ischemic
necrosis of the femoral head. Early diagnosis and treatment. Clin
Orthop Relat Res. 1978;130:144–153.
30. Iorio R, Healy WL, Abramowitz AJ, Pfeifer BA. Clinical outcome
and survivorship analysis of core decompression for early osteo-
necrosis of the femoral head. J Arthroplasty. 1998;13:34–41.
31. Israelite C, Nelson CL, Ziarani CF, Abboud JA, Landa J, Stein-
berg ME. Bilateral core decompression for osteonecrosis of the
femoral head. Clin Orthop Relat Res. 2005;441:285–290.
32. Jergesen HE, Khan AS. The natural history of untreated asymp-
tomatic hips in patients who have non-traumatic osteonecrosis.
J Bone Joint Surg Am. 1997;79:359–363.
33. Kane SM, Ward WA, Jordan LC, Guilford WB, Hanley EN Jr.
Vascularized fibular grafting compared with core decompression
in the treatment of femoral head osteonecrosis. Orthopedics.
1996;19:869–872.
34. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. The
conservative surgical treatment of idiopathic aseptic necrosis of
the femoral head. J Bone Joint Surg Br. 1974;56:291–296.
35. Kim SY, Kim YG, Kim PT, Ihn JC, Cho BC, Koo KH. Vascu-
larized compared with nonvascularized fibular grafts for large
osteonecrotic lesions of the femoral head. J Bone Joint Surg Am.
2005;87:2012–2018.
36. Koo KH, Kim R, Ko GH, Song HR, Jeong ST, Cho SH. Pre-
venting collapse in early osteonecrosis of the femoral head. A
randomised clinical trial of core decompression. J Bone Joint
Surg Br. 1995;77:870–874.
37. Kristensen KD, Pedersen NW, Kiaer T, Starklint H. Core
decompression in femoral head osteonecrosis. 18 Stage I hips
followed up for 1–5 years. Acta Orthop Scand. 1991;62:113–114.
38. Lafforgue P, Dahan E, Chagnaud C, Schiano A, Kasbarian M,
Acquaviva PC. Early-stage avascular necrosis of the femoral
head: MR imaging for prognosis in 31 cases with at least 2 years
of followup. Radiology. 1993;187:199–204.
39. Lai KA, Shen WJ, Yang CY, Shao CJ, Hsu JT, Lin RM. The use
of alendronate to prevent early collapse of the femoral head in
patients with nontraumatic osteonecrosis. A randomized clinical
study. J Bone Joint Surg Am. 2005;87:2155–2159.
40. Lausten GS, Mathiesen B. Core decompression for femoral head
necrosis. Prospective study of 28 patients. Acta Orthop Scand.
1990;61:507–511.
41. Lavernia CJ, Sierra RJ. Core decompression in atraumatic oste-
onecrosis of the hip. J Arthroplasty. 2000;15:171–178.
42. Learmonth ID, Maloon S, Dall G. Core decompression for early
atraumatic osteonecrosis of the femoral head. J Bone Joint Surg
Br. 1990;72:387–390.
43. Leder K, Knahr K. Results of medullary space decompression in
the early stage of so-called idiopathic femur head necrosis [in
German]. Z Orthop Ihre Grenzgeb. 1993;131:113–119.
123
1102 Marker et al.
44. Li YZ, Yue T. Diagnosis and treatment of idiopathic necrosis of
the femoral head. Proc Chin Acad Med Sci Peking Union Med
Coll. 1990;5:88–92.
45. Lieberman JR, Conduah A, Urist MR. Treatment of osteonecrosis
of the femoral head with core decompression and human bone
morphogenetic protein. Clin Orthop Relat Res. 2004;429:139–145.
46. Maniwa S, Nishikori T, Furukawa S, Kajitani K, Iwata A,
Nishikawa U, Ochi M. Evaluation of core decompression for
early osteonecrosis of the femoral head. Arch Orthop Trauma
Surg. 2000;120:241–244.
47. Markel DC, Miskovsky C, Sculco TP, Pellicci PM, Salvati EA.
Core decompression for osteonecrosis of the femoral head. Clin
Orthop Relat Res. 1996;323:226–233.
48. Mazieres B, Marin F, Chiron P, Moulinier L, Amigues JM,
Laroche M, Cantagrel A. Influence of the volume of osteone-
crosis on the outcome of core decompression of the femoral head.
Ann Rheum Dis. 1997;56:747–750.
49. Merle D’Aubigne R, Postel M, Mazabraud A, Massias P,
Gueguen J, France P. Idiopathic necrosis of the femoral head in
adults. J Bone Joint Surg Br. 1965;47:612–633.
50. Mihalko WM, Balos L, Santilli M, Mindell ER. Osteonecrosis
after powered core decompression. Clin Orthop Relat Res.
2003;412:77–83.
51. Mont MA, Carbone JJ, Fairbank AC. Core decompression versus
nonoperative management for osteonecrosis of the hip. Clin
Orthop Relat Res. 1996;324:169–178.
52. Mont MA, Fairbank AC, Petri M, Hungerford DS. Core
decompression for osteonecrosis of the femoral head in systemic
lupus erythematosus. Clin Orthop Relat Res. 1997;334:91–97.
53. Mont MA, Jones LC, Einhorn TA, Hungerford DS, Reddi AH.
Osteonecrosis of the femoral head. Potential treatment with
growth and differentiation factors. Clin Orthop Relat Res.
1998;355 Suppl:S314–335.
54. Mont MA, Jones LC, Pacheco I, Hungerford DS. Radiographic
predictors of outcome of core decompression for hips with oste-
onecrosis stage III. Clin Orthop Relat Res. 1998;354:159–168.
55. Mont MA, Marulanda GA, Jones LC, Saleh KJ, Gordon N,
Hungerford DS, Steinberg ME. Systematic analysis of classifi-
cation systems for osteonecrosis of the femoral head. J Bone Joint
Surg Am. 2006;88 Suppl 3:16–26.
56. Mont MA, Ragland PS, Etienne G. Core decompression of the
femoral head for osteonecrosis using percutaneous multiple
small-diameter drilling. Clin Orthop Relat Res. 2004;429:131–
138.
57. Morse CG, Mican JM, Jones EC, Joe GO, Rick ME, Formentini
E, Kovacs JA. The incidence and natural history of osteonecrosis
in HIV-infected adults. Clin Infect Dis. 2007;44:739–748.
58. Musso ES, Mitchell SN, Schink-Ascani M, Bassett CA. Results
of conservative management of osteonecrosis of the femoral
head. A retrospective review. Clin Orthop Relat Res.
1986;207:209–215.
59. Neumayr LD, Aguilar C, Earles AN, Jergesen HE, Haberkern
CM, Kammen BF, Nancarrow PA, Padua E, Milet M, Stulberg
BN, Williams RA, Orringer EP, Graber N, Robertson SM,
Vichinsky EP. Physical therapy alone compared with core
decompression and physical therapy for femoral head osteone-
crosis in sickle cell disease. Results of a multicenter study at a
mean of three years after treatment. J Bone Joint Surg Am.
2006;88:2573–2582.
60. Ohzono K, Saito M, Sugano N, Takaoka K, Ono K. The fate of
nontraumatic avascular necrosis of the femoral head. A radiologic
classification to formulate prognosis. Clin Orthop Relat Res.
1992:73–78.
61. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
123
Clinical Orthopaedics and Related Research
62. Piperkovski T. Results of treatment in patients with nontraumatic
avascular necrosis of the femoral head by monitor assisted core
decompression. Rentgenol. Radiol. 2001;40:281–284.
63. Plakseychuk AY, Kim SY, Park BC, Varitimidis SE, Rubash HE,
Sotereanos DG. Vascularized compared with nonvascularized
fibular grafting for the treatment of osteonecrosis of the femoral
head. J Bone Joint Surg Am. 2003;85:589–596.
64. Powell ET, Lanzer WL, Mankey MG. Core decompression for
early osteonecrosis of the hip in high risk patients. Clin Orthop
Relat Res. 1997;335:181–189.
65. Radke S, Kirschner S, Seipel V, Rader C, Eulert J. Treatment of
transient bone marrow oedema of the hip–a comparative study.
Int Orthop. 2003;27:149–152.
66. Robinson Jr. HJ, Springer JA. Success of core decompression in
the management of early stages of avascular necrosis: A four year
prospective study. Orthop Trans. 1993;16:707.
67. Saito S, Ohzono K, Ono K. Joint-preserving operations for idi-
opathic avascular necrosis of the femoral head. Results of core
decompression, grafting and osteotomy. J Bone Joint Surg Br.
1988;70:78–84.
68. Scully SP, Aaron RK, Urbaniak JR. Survival analysis of hips
treated with core decompression or vascularized fibular grafting
because of avascular necrosis. J Bone Joint Surg Am.
1998;80:1270–1275.
69. Shuler MS, Rooks MD, Roberson JR. Porous tantalum implant in
early osteonecrosis of the hip: preliminary report on operative,
survival, and outcomes results. J Arthroplasty. 2007;22:26–31.
70. Simank HG, Brocai DR, Strauch K, Lukoschek M. Core
decompression in osteonecrosis of the femoral head: risk-factor-
dependent outcome evaluation using survivorship analysis. Int
Orthop. 1999;23:154–159.
71. Smith SW, Fehring TK, Griffin WL, Beaver WB. Core decom-
pression of the osteonecrotic femoral head. J Bone Joint Surg Am.
1995;77:674–680.
72. Solomon L. Idiopathic necrosis of the femoral head: pathogenesis
and treatment. Can J Surg. 1981;24:573–578.
73. Song WS, Yoo JJ, Kim YM, Kim HJ. Results of multiple drilling
compared with those of conventional methods of core decom-
pression. Clin Orthop Relat Res. 2007;454:139–146.
74. Specchiulli F. Core decompression in the treatment of necrosis of
the femoral head. Long-term results. Chir Organi Mov.
2000;85:395–402.
75. Steinberg ME. Core decompression of the femoral head for
avascular necrosis: indications and results. Can J Surg. 1995;38
Suppl 1:S18–24.
76. Steinberg ME, Bands RE, Parry S, Hoffman E, Chan T, Hartman
KM. Does lesion size affect the outcome in avascular necrosis?
Clin Orthop Relat Res. 1999;367:262–271.
77. Steinberg ME, Belmar CJ. Role of core decompression in the
treatment of avascular necrosis of the femoral head. Current
Orthopaedics. 1997;11:173–178.
78. Steinberg ME, Brighton CT, Bands RE, Hartman KM. Capacitive
coupling as an adjunctive treatment for avascular necrosis. Clin
Orthop Relat Res. 1990;261:11–18.
79. Steinberg ME, Brighton CT, Corces A, Hayken GD, Steinberg
DR, Strafford B, Tooze SE, Fallon M. Osteonecrosis of the
femoral head. Results of core decompression and grafting with
and without electrical stimulation. Clin Orthop Relat Res.
1989;249:199–208.
80. Steinberg ME, Brighton CT, Steinberg DR, Tooze SE, Hayken
GD. Treatment of avascular necrosis of the femoral head by a
combination of bone grafting, decompression, and electrical
stimulation. Clin Orthop Relat Res. 1984;186:137–153.
81. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system
for staging avascular necrosis. J Bone Joint Surg Br. 1995;77:
34–41.
Volume 466, Number 5, May 2008
82. Steinberg ME, Larcom PG, Strafford B, Hosick WB, Corces A,
Bands RE, Hartman KE. Core decompression with bone grafting
for osteonecrosis of the femoral head. Clin Orthop Relat Res.
2001;386:71–78.
83. Stulberg BN, Davis AW, Bauer TW, Levine M, Easley K.
Osteonecrosis of the femoral head. A prospective randomized
treatment protocol. Clin Orthop Relat Res. 1991;268:140–151.
84. Styles LA, Vichinsky EP. Core decompression in avascular
necrosis of the hip in sickle-cell disease. Am J Hematol.
1996;52:103–107.
85. Tingart M, Bathis H, Perlick L, Lerch K, Luring C, Grifka J.
Therapy of femoral head osteonecrosis: results of a national
survey [in German]. Z Orthop Ihre Grenzgeb. 2004;142:553–558.
86. Tooke SM, Nugent PJ, Bassett LW, Nottingham P, Mirra J,
Jinnah R. Results of core decompression for femoral head oste-
onecrosis. Clin Orthop Relat Res. 1988;228:99–104.
87. Trancik T, Lunceford E, Strum D. The effect of electrical stim-
ulation on osteonecrosis of the femoral head. Clin Orthop Relat
Res. 1990;256:120–124.
88. Van Laere C, Mulier M, Simon JP, Stuyck J, Fabry G. Core
decompression for avascular necrosis of the femoral head. Acta
Orthop Belg. 1998;64:269–272.
89. Veillette CJ, Mehdian H, Schemitsch EH, McKee MD. Survi-
vorship analysis and radiographic outcome following tantalum
rod insertion for osteonecrosis of the femoral head. J Bone Joint
Surg Am. 2006;88 Suppl 3:48–55.
90. Wang CJ, Wang FS, Huang CC, Yang KD, Weng LH, Huang
HY. Treatment for osteonecrosis of the femoral head: comparison
Modern Core Decompression Techniques 1103
of extracorporeal shock waves with core decompression and
bone-grafting. J Bone Joint Surg Am. 2005;87:2380–2387.
91. Warner JJ, Philip JH, Brodsky GL, Thornhill TS. Studies of
nontraumatic osteonecrosis. The role of core decompression in
the treatment of nontraumatic osteonecrosis of the femoral head.
Clin Orthop Relat Res. 1987;225:104–127.
92. Wirtz C, Zilkens KW, Adam G, Niethard FU. MRI-controlled
outcome after core decompression of the femur head in aseptic
osteonecrosis and transient bone marrow edema [in German].
Z Orthop Ihre Grenzgeb. 1998;136:138–146.
93. Wirtz DC, Rohrig H, Neuss M. Core decompression for avascular
necrosis of the femoral head. Oper Orthop Traumatol. 2003;15:
288–303.
94. Wright RW, Brand RA, Dunn W, Spindler KP. How to write a
systematic review. Clin Orthop Relat Res. 2007;455:23–29.
95. Yan ZQ, Chen YS, Li WJ, Yang Y, Huo JZ, Chen ZR, Shi JH, Ge
JB. Treatment of osteonecrosis of the femoral head by percuta-
neous decompression and autologous bone marrow mononuclear
cell infusion. Chin J Traumatol. 2006;9:3–7.
96. Yoo MC, Chung DW, Hahn CS. Free vascularized fibula grafting
for the treatment of osteonecrosis of the femoral head. Clin
Orthop Relat Res. 1992;277:128–138.
97. Yoon TR, Song EK, Rowe SM, Park CH. Failure after core
decompression in osteonecrosis of the femoral head. Int Orthop.
2001;24:316–318.
98. Zizic TM, Hungerford DS. Avascular necrosis of bone. In: Kelley
WN, Harris ED, Ruddy S, Sledge CB (eds). Textbook of Rheuma-
tology. Ed 2. Philadelphia, PA: WB Saunders Co; 1985:1689–1710.
123
Clin Orthop Relat Res (2008) 466:1104–1109
DOI 10.1007/s11999-008-0192-9
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Transtrochanteric Posterior Rotational Osteotomy
for Osteonecrosis
Yoichi Sugioka MD, PhD, Takuaki Yamamoto MD, PhD
Published online: 8 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract When osteonecrosis is located in the mid- to
posterior region, we generally perform a transtrochanteric
posterior rotational osteotomy. We retrospectively
reviewed the clinical and radiographic results in 47 con-
secutive patients (51 hips) in whom we performed posterior
rotational osteotomies. The average age was 37 years at the
time of surgery. There were 30 male and 17 female patients.
Thirty-six hips were ARCO Stage III, and 15 were Stage IV.
Conversion to THA was defined as the failure end point.
Three patients died and one was lost to followup. We were
therefore able to follow 43 patients (46 of the 51 hips, or
90%) a minimum of 1.2 years (average, 12 years; range,
1.2–21 years). We used the Harris hip score for preopera-
tive and most recent followup. The average preoperative
Harris hip score of 52 points improved to an average of 84 at
the latest followup. Radiographically, the osteonecrosis in
30 hips (65%) had no progressive collapse, and 13 (28%)
showed osteoarthritic changes, but no patients underwent
THA. A posterior rotational osteotomy appears useful for
patients with extensive necrosis and advanced collapse.
Each author certifies that he has no commercial associations (e.g.,
consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc.) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved or waived
approval for the human protocol for this investigation and that all
investigations were conducted in conformity with ethical principles of
research.
Y. Sugioka (&), T. Yamamoto
Department of Orthopaedic Surgery, Graduate School of
Medical Sciences, Kyushu University, 3-1-1 Maidashi,
Higashi-ku, Fukuoka 812-8582, Japan
e-mail: sugioka@mub.biglobe.ne.jp
123
Level of Evidence: Level IV, therapeutic study. See
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Once subchondral collapse occurs in osteonecrosis (ON) of
the femoral head, osteoarthritis subsequently develops [8,
14, 15]. ON occurs predominantly in the middle-aged
population. A higher risk of mechanical failure is reported
following THA in patients with osteonecrosis who are
younger than 50 years old than in age-matched osteoar-
thritis patients [11]. The best surgical treatment for
osteonecrosis thus remains to be determined in these
patients with collapse.
The natural history of ON depends on the location and
extent of the necrotic region [6, 10]. If the necrotic region
is located in the nonweightbearing middle part of the
femoral head, it will not collapse [6, 10]. Therefore, we
believe the principles of the treatment of ON are (1) to
eliminate the shear stress from the necrotic region and to
prevent a progression of collapse, and (2) to obtain joint
realignment of the femoral head subluxated due to col-
lapse. To satisfy these principles, transtrochanteric
rotational osteotomy was developed to transpose the
necrotic area to the nonweightbearing portion [15].
ON of the femoral head is generally located in the
anterosuperior aspect of the femoral head, in which the
posterior aspect of the femoral head is normal and has
smooth healthy articular cartilage. In these patients, several
authors recommend an anterior rotational osteotomy.
Clinical followup studies from several institutions suggest
this procedure has merit to preserve the joint [2, 3, 13, 15].
On the other hand, when ON is located in the mid- to
Volume 466, Number 5, May 2008
posterior portion, in which the intact area is located in the
anterior portion of the femoral head, a transtrochanteric
posterior rotational osteotomy is logical because this pro-
cedure transposes the necrotic area to the posterior
nonweightbearing portion [1, 13, 14].
We therefore asked whether posterior rotational osteot-
omy would preserve the joint as reflected in survival (with
conversion to THA as an endpoint), Harris hip scores, and
radiographic collapse or osteoarthritis. We further asked
whether the ARCO staging system would predict pro-
gressive osteoarthritis.
Materials and Methods
We retrospectively reviewed 47 consecutive patients (51
hips) with clinical and radiographic evidence of ON of the
femoral head treated by a posterior rotational osteotomy
from 1981 to 1996. We used a posterior rotational osteot-
omy in patients with: (1) the necrotic area in the posterior
portion of the femoral head; (2) an intact healthy area in the
anterior portion; and (3) a postoperative intact area ratio
over 34% after the posterior rotational osteotomy (Fig. 1)
[1, 8, 13]. During this same period we treated a total of 581
Fig. 1 Postoperative intact area ratio is the most important factor
influencing the results of a rotational osteotomy. This ratio is
expressed as C-D/A-B and should be over 34% after the rotational
osteotomy. Point A is an edge of the acetabulum and Point E is the
lowest point of the teardrop. Point B is determined by drawing a
perpendicular line from the midpoint of line of AE. Point C is the
lateral edge of the loaded portion and point D represents the medial
edge of the intact articular surface of the femoral head.
Rotational Osteotomy for Osteonecrosis 1105
patients with ON with either other osteotomies (412) or
THA/hemiarthroplasty (169). Thirty patients were male
and 17 were female. Nine hips were ARCO Stage III-A, 13
Stage III-B, 14 Stage III-C, and 15 Stage IV [9]. One hip
was categorized as Type B, 15 Type C1 and 35 Type C2
[12]. The average age was 37 years at the time of surgery
(range, 9 to 58 years). The etiology of ON was idiopathic
in four hips, alcohol abuse in 12, trauma in 14, and corti-
costeroids in 21. Four patients (four hips) had bilateral
disease and underwent bilateral posterior rotational oste-
otomy. Three patients died and one was lost to followup.
This left 43 patients (46 hips) in the study. The minimum
followup was 1.2 years (mean, 12 years; range, 1.2 to
21 years).
The basic surgical procedure has remained unchanged
since it was described by one of the authors (YS) [13–15]
and all operations were performed by that surgeon. Briefly,
this procedure contains three osteotomies: (1) an osteotomy
of the greater trochanter; (2) an intertrochanteric osteotomy
which passes from superolateral to inferomedial on the AP
view; and (3) an osteotomy which passes from the proxi-
mal flare of the lesser trochanter inferolaterally towards the
inferomedial extent of the primary osteotomy. During these
osteotomies, preservation of the nutrient artery located in
the adipose tissue just beneath the quadratus femoris is
important. The degree of posterior rotation ranged from 90°
to 180°, with an average of 105°, which was determined to
obtain a sufficient postoperative intact area (more than
34%). In three cases, 180° of posterior rotation was per-
formed because the intact area was very small and 180° of
posterior rotation was necessary to obtain a sufficient intact
area. In these three cases the postoperative scintigram
showed no uptake indicating insufficiency of blood supply
of the femoral head and these three were converted to THA
secondary to subcapital fracture (not osteoarthritic chan-
ges) within 1.6 years of posterior rotational osteotomy. We
therefore excluded these three cases from further analysis.
Postoperatively, the patients were nonweightbearing for
5 weeks, after which gradual weightbearing was permitted.
Full weightbearing was permitted after 6 months.
Followup radiographic and clinical examinations are
performed at 3 months, 6 months, and yearly thereafter. A
bone scintigram was routinely performed 5 weeks post-
operatively to confirm the blood supply to the rotated
femoral head. At each exam we (YS, TY) determined the
Harris hip score [4] including prior to the posterior rota-
tional osteotomy, before conversion to THA, and at the
most recent followup examination.
We (YS, TY) reviewed the preoperative and postoper-
ative radiographs to establish the size and location of the
necrotic lesion [12] and evidence of the progression of
collapse and/or development of osteoarthritis. Osteoar-
thritic changes were diagnosed based on the presence of
123
1106 Sugioka and Yamamoto Clinical Orthopaedics and Related Research

joint space narrowing, osteophyte formation, and/or scle-

rotic changes in both the femoral head and acetabulum.
We defined conversion to THA as the end point of
failure of the posterior rotational osteotomy in the survival
analysis including the patients who died or were lost to
followup. We compared the preoperative Harris hip scores
with the last postoperative scores using a paired t-test. The
correlation between the preoperative ARCO stage and a
progression of osteoarthritis was assessed using the Mann-
Whitney U test. Among the etiologies of ON (steroids,
alcohol, trauma and idiopathic), both the postoperative
Harris hip score and a progression of osteoarthritis were
compared using one-way ANOVA and Fisher’s exact
probability test, respectively. Statistical analyses were
performed using the Stat View J-5.0 software package
(SAS Institute Inc, Cary, NC, USA).

Results

The overall rate of survival was 93%. Three patients

underwent THA, all for subcapital fracture in patients
having an osteotomy with 180 degrees rotation.
The average preoperative Harris hip score of 52.4 points
increased (p \ 0.0001) to an average of 83.7 points at the Fig. 2A–D Posterior rotational osteotomy. (A) The preoperative
latest followup. The average preoperative Harris hip score anteroposterior radiograph of a thirty-six-year-old man with preop-
erative ARCO Stage III-A is shown. The osteosclerotic line is seen in
for range of motion was 4.5 points, which was unchanged the femoral head. (B) The preoperative lateral view indicates the
at the latest followup. However, the other two factors of intact area in the anterior portion is wider than that in the posterior
pain and function improved after posterior rotational portion. The arrows indicate the margin of the necrotic area. (C) A
osteotomy: pain from 15 to 36 points (p \ 0.0001) and radiograph obtained 1 year after the 90° posterior rotational osteot-
omy is shown. The postoperative intact area ratio is about 70%. (D) A
function from 29 to 39 points (p \ 0.0001). radiograph obtained 13 years after the operation shows no evidence of
We observed no progressive radiographic collapse in 30 a progression of collapse or the development of osteoarthritis. The
hips (69.8%) (Fig. 2A–D). Thirteen hips (30.2%) developed size of the necrotic area, which is surrounded by a sclerotic line, has
osteoarthritic changes. However, no patient underwent decreased (arrows), thus indicating a sufficient repair was obtained by
the rotational osteotomy.
THA due to progressive osteoarthritic changes.
The progression of osteoarthritis correlated (p = 0.0004)
with the preoperative ARCO stage. In the six early ARCO
Stage III-A cases there was no progression of osteoarthritis,
while in the advanced ARCO Stage IV cases, eight of the 12 Table 1. Summary of data according to the ARCO stage
cases progressed to osteoarthritis (Table 1). We observed
ARCO No. of hips Progression of osteoarthritis**
progression of osteoarthritis in three of 16 cases with ste- stage(pre-operation)
roid-induced ON, three of 10 cases with alcohol-related - +
disease, six of 13 with traumatic ON, and one of four with III-A 6 6 0 (0%)
idiopathic ON. There was no difference in the rate among III-B 13 12 1 (8%)
etiologies (Table 2). We observed no differences in post- III-C 12 8 4 (33%)
operative Harris hip scores based on etiology or location of
IV 12 4 8 (67%)
ON (Type of ON).
Total 43* 30 (70%) 13 (30%)
* Three cases in which 180° posterior rotation was performed are not
included, because all of them underwent a subcapital fracture within
Discussion 1.6 years.
** Based on the Mann-Whitney U test, a significant correlation was
ON of the femoral head is generally located in the anter- seen between the preoperative ARCO stage and the progression of
osuperior aspect of the femoral head, in which the posterior osteoarthritis.

123
Volume 466, Number 5, May 2008
Table 2. Summary of data according to the etiology of osteonecrosis
Etiology No. of hips Progression of
osteoarthritis**
Steroids 16 3 (19%)
Alcohol 10 3 (30%)
Trauma 13 6 (46%)
Idiopathic 4 1 (25%)
Total 43* 13 (30%)
* Three cases in which 180° posterior rotation was performed are not included, because all of them underwent a subcapital fracture within
1.6 years.
** Based on the Fisher’s exact probability test, no significant difference was seen in the progression of osteoarthritis among the etiologies of
osteonecrosis.
*** Based on the one-way analysis of variance, no significant difference was seen in the postoperative Harris hip score among the etiologies of
osteonecrosis.
aspect of the femoral head is relatively normal. In these
patients, some authors recommend an anterior rotational
osteotomy. On the other hand when ON is located in the
mid- to posterior portion, in which the intact area is located
in the anterior portion of the femoral head, a transtro-
chanteric posterior rotational osteotomy is logical. We
therefore asked whether posterior rotational osteotomy
would preserve the joint as reflected in survival (with
conversion to THA as an endpoint), Harris hip scores, and
radiographic collapse or osteoarthritis. We further asked
whether the ARCO staging system would predict pro-
gressive osteoarthritis.
The major limitation of this study is the small number of
patients, since most ON lesions are located in the anterior
portion of the femoral head, in which anterior rotational
osteotomy is considered. The large number of confounding
variables (e.g., stages, etiologies) and relatively small
numbers of patients preclude more complete analysis of
predictive variables and thus more clear indications. The
followup was limited to a minimum of 1.2 years but the
mean was 12 years with a range of 1.2–21 years. Many
patients who collapse and develop progressive osteoar-
thritis will not do so for a number of years and not by
1.2 years. However, the mean followup was 12 years, at
which time most patients with collapse would have been
detected.
Serial studies of patients administered corticosteroids
suggest ON can be asymptomatic (silent hip) and can heal
spontaneously if the lesion is located in the nonweight-
bearing area and if the articular surface is prevented from
collapsing [6, 7]. In addition, healing of ON has been
demonstrated in histopathologic investigations of femoral
heads removed from patients who have undergone rota-
tional osteotomy which transposed the necrotic area into
the nonweightbearing portion [13]. On the other hand,
if the osteonecrotic lesion is located in the major
Rotational Osteotomy for Osteonecrosis 1107
Harris hip score (points) ARCO stage
Preop. Postop.*** III-A III-B III-C IV
50.8 80.8 3 6 3 4
55.6 84.5 3 3 1 3
49.5 85.1 0 2 7 4
60.9 88.8 0 2 1 1
52.4 83.7 6 13 12 12
weightbearing area, it will eventually undergo collapse,
and a collapsed femoral head tends to be subluxated an-
terolaterally, thus resulting in a further collapse. Once
collapse progresses, the reparative tissue is exposed to
greater mechanical stress which we suspect impairs the
reparative process [6, 10, 13].
Based on our experience we derive two important
principles in the treatment of ON of the femoral head. First
is to eliminate the shear forces from the necrotic focus,
which leads reduces the risk of progressive collapse and
enhances the repair process. The second is to realign the
femoral head subluxation that occurs due to the collapse of
the articular surface. A transtrochanteric rotational osteot-
omy, which transposes the necrotic area into a
nonweightbearing area, appears an effective method for
preventing collapse by diverting mechanical stress from the
lesion [8, 13, 14].
We have previously reported the outcomes of anterior
rotational osteotomy in 136 hips [5]. In that study, patients in
the early stages such as ARCO Stage II or III-A had a low
rate of conversion to THA (6.8%) over the long-term. On the
other hand, patients with ARCO Stage III-C and IV had
conversion rates of 21% and 24%, respectively. Thus, to
prevent long-term osteoarthritis, we believe an anterior
rotational osteotomy should be performed in the early stages
with minimum collapse such as ARCO Stage II or III-A.
In our current study, 30 hips (70%) had no progressive
collapse after posterior rotational osteotomy. Thirteen
(30%) had osteoarthritic changes but the preoperative
stages were advanced ARCO Stage III-C and IV in 12 of
these 13 hips. Nevertheless, no patient had conversion to
THA due to the progression of osteoarthritic changes even
in such advanced preoperative stages. These results seem
better than those after an anterior rotational osteotomy. An
anterior rotational osteotomy transposes the necrotic area
to the anterior portion. After the anterior rotation, the
123
1108 Sugioka and Yamamoto
Fig. 3 Posterior rotational osteotomy for the treatment of ON of the
femoral head. When ON is located mainly in the mid- to posterior
portion, a transtrochanteric posterior rotational osteotomy, which
transposes the necrotic area to the posterior nonweightbearing
portion, is considered. In this procedure, the collapsed joint surface
of the femoral head was transposed to the acetabulum, and thus joint
stability was obtained. In addition, because the tension to the nutrient
vessels decreases after the posterior rotation, the femoral head can be
rotated by more than 100°. The upper limit is around 140° to 150°.
collapsed lesion still exists in the anterior portion of the
joint. Therefore, hip instability, especially anterior to pos-
terior direction, cannot be solved, especially in patients
with advanced collapse. In addition, this collapsed lesion
may cause crepitus to the anterior edge of the acetabulum
when taking a flexion position. However, with a posterior
rotational osteotomy the collapsed joint surface is trans-
posed into the acetabulum, joint stability can be restored
(Fig. 3). The original anterior acetabular coverage, which
is generally difficult to ascertain on the AP view, can
enhance the weightbearing area, especially in patients with
posterior rotation (Fig. 4).
The most important factor affecting the results of a
rotational osteotomy is the postoperative intact area ratio,
which should be over 34% [8]. In order to evaluate the
exact intact area preoperatively, it is important to take a
precise lateral view of the femoral head with 90° of flexion
and 45° of abduction. The more flexion of the hip, the more
the inferior part of the femoral head can be seen on this
lateral view. By taking a lateral view over 90° of flexion,
one can identify the intact area located in the anteroinferior
part of the femoral head which will be transposed to the
weightbearing portion after a posterior rotational osteot-
omy over 90°.
We believe the indications for posterior rotational
osteotomy can be derived from our data. In every ARCO
stage, if the intact area is located in either the anterior
portion or located equally in the anterior or posterior por-
tion (Fig. 2B), then we consider a posterior rotational
123
Clinical Orthopaedics and Related Research
Fig. 4 The sagittal CT image shows original anterior acetabular
coverage (arrow line), which is not generally visible on the AP
radiograph. This original anterior acetabular coverage can enhance
the intact weightbearing area, especially in patients with posterior
rotation.
osteotomy. In advanced ARCO stages, the indications for
posterior rotational osteotomy can be expanded. If the
intact area is located primarily in the posterior portion at
90° of hip flexion, then an anterior rotational osteotomy
may be considered. However, if the anterior intact area is
much wider with 120° of flexion, then we recommend a
120° posterior rotational osteotomy rather than a 90°
anterior rotational osteotomy, in order to prevent osteoar-
thritic changes after the osteotomy.
Because the tension to the nutrient vessels decreases
after posterior rotation, the femoral head can be rotated
more than 100° (Fig. 3). However, the upper limit is
around 140° to 150°, because patients with 180° of pos-
terior rotation had cold lesions on postoperative scintigram
indicating insufficient blood supply to the femoral head,
and all later had a subcapital fracture necessitating THA. In
the posterior rotational osteotomy an anteversion angle
yields a valgus position after the osteotomy if the osteot-
omy line was determined without reducing the anteversion
angle. In addition, postoperative treatment should be
approached carefully because the posterior part of the bone,
not originally strong, is transposed to the calcar. Since this
procedure needs to obtain such bone remodeling at the
weight bearing portion as well as bone union at the oste-
otomy site, full weightbearing should not be started until at
least 6 months after the operation. This may cause an
extension of the hospitalization period. Atsumi et al. [1]
reported similarly encouraging results for posterior rota-
tional osteotomy in ON patients, in which the survival rate
was 78% at an average of 5 years.
Volume 466, Number 5, May 2008
In summary, when ON is located in the mid- to posterior
portion and the intact area is located in the anterior portion
of the femoral head a transtrochanteric rotational osteot-
omy appears to delay the need for THA in the majority of
patients.
References
1. Atsumi T, Muraki M, Yoshihara S, Kajihara T. Posterior rota-
tional osteotomy for the treatment of femoral head osteonecrosis.
Arch Orthop Trauma Surg. 1999;119:388–393.
2. Dean MT, Cabanela ME. Transtrochanteric anterior rotational
osteotomy for avascular necrosis of the femoral head. Long-term
results. J Bone Joint Surg Br. 1993;75:597–601.
3. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
4. Hasegawa Y, Sakano S, Iwase S, Iwasada S, Torii H, Iwata H.
Pedicle bone grafting versus transtrochanteric rotational osteot-
omy for avascular necrosis of the femoral head. J Bone Joint Surg
Br. 2003;85:191–198.
5. Hosokawa A, Mohtai M, Hotokebuchi T, Jingushi S, Sugioka Y.
Transtrochanteric rotational osteotomy for idiopathic and steroid-
induced osteonecrosis of the femoral head: Indications and long-
term follow-up. In: Urbaniak JR, Jones JP, eds. Osteonecrosis:
Etiology, Diagnosis, and Treatment. Rosemont, IL: American
Academy of Orthopaedic Surgeons; 1998:309–314.
6. Kokubo T, Takatori Y, Ninomiya S, Nakamura T, Kamogawa M.
Magnetic resonance imaging and scintigraphy of avascular
Rotational Osteotomy for Osteonecrosis 1109
necrosis of the femoral head. Prediction of subsequent segmental
collapse. Clin Orthop Relat Res. 1992;277:54–60.
7. Kubo T, Yamazoe N, Sugano N, Fujioka M, Naruse S,
Yoshimura N, Oka T, Hirasawa Y. Initial MRI findings of non-
traumatic osteonecrosis of the femoral head in renal allograft
recipients. Magn Reson Imag. 1997;15:1017–1023.
8. Miyanishi K, Noguchi Y, Yamamoto T, Irisa T, Suenaga E,
Jingushi S, Sugioka Y, Iwamoto Y. Prediction of the outcome of
transtrochanteric rotational osteotomy for osteonecrosis based on
the postoperative intact ratio. J Bone Joint Surg Br. 2000;82:512–
516.
9. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of
the femoral head. J Bone Joint Surg Am. 1995;77:459–474.
10. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
11. Ortiguera CJ, Pulliam IT, Cabanela ME. Total hip arthroplasty
for osteonecrosis: matched-pair analysis of 188 hips with long-
term follow-up. J Arthroplasty. 1999;14:21–28.
12. Sugano N, Atsumi T, Ohzono K, Kubo T, Hotokebuchi T,
Takaoka K. The 2001 revised criteria for diagnosis, classification,
and staging of idiopathic osteonecrosis of the femoral head.
J Orthop Sci. 2002;7:601–605.
13. Sugioka Y. Transtrochanteric anterior rotational osteotomy of the
femoral head in the treatment of osteonecrosis affecting the hip; a
new osteotomy operation. Clin Orthop Relat Res. 1978;130:191–
201.
14. Sugioka Y. Transtrochanteric rotational osteotomy of the femoral
head. In: Macnicol MF, ed. Color Atlas, Text of Osteotomy of the
Hip. London, UK: Mosby-Wolfe; 1996:145–156.
15. Sugioka Y, Hotokebuchi T, Tsutsui H. Transtrochanteric anterior
osteotomy for idiopathic and steroid-induced necrosis of the
femoral head. Clin Orthop Relat Res. 1992;277:111–120.
123
Clin Orthop Relat Res (2008) 466:1110–1116
DOI 10.1007/s11999-008-0188-5
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Modified Transtrochanteric Rotational Osteotomy for Femoral
Head Osteonecrosis
Taek Rim Yoon MD, Azlina Amir Abbas MD,
Chang Ich Hur MD, Sang Gwon Cho MD,
Jin Ho Lee MD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Osteonecrosis of the femoral head is a disabling
condition affecting young patients and treatment of the
disease in these patients is variable. We retrospectively
reviewed 39 patients (43 hips) in whom a modified tran-
strochanteric rotational osteotomy was performed for
osteonecrosis. The minimum followup was 24 months
(mean, 36.6 months; range, 24–52 months). The mean
patient age was 34.3 years (range, 20–51 years). Based on
the ARCO classification, 17 hips were classified as Stage II
and 26 as Stage III. We performed rotational osteotomy
alone in 15 cases, in combination with simple bone grafting
in three, and in combination with muscle-pedicle-bone
grafting in 25. Sixteen of 17 ARCO Stage II cases and 24
of 26 ARCO Stage III cases had no progression of collapse
or lesion size; three hips progressively collapsed. Of the 40
hips without progression the Harris hip score improved
from a mean 70 to 92 points at final followup, as did the
range of motion of the hip. Modified transtrochanteric
rotational osteotomy is an effective method for delaying
the progression of collapse in the treatment of selected
cases of osteonecrosis of the femoral head.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
T. R. Yoon, C. I. Hur, S. G. Cho, J. H. Lee
Department of Orthopedics, Chonnam National University
School of Medicine, Gwangju, Korea
T. R. Yoon (&), A. A. Abbas
Center for Joint Disease, Chonnam National University Hwasun
Hospital, 160 Ilsimri, Hwasuneup, Hwasungun, 519-809
Jeonnam, Korea
e-mail: tryoon@chonnam.ac.kr
123
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis of the femoral head is a disabling disease
involving the hips of young people [6]. Despite many
efforts, the etiology of osteonecrosis still has not been
identified and the pathogenesis is not fully understood.
Therefore, treatment methods are various and are often
chosen according to stage, location, and size of the necrotic
area. Surgery usually involves head-preserving procedures
or total hip arthroplasty (THA). In general, the head-
preserving treatments are used in Ficat-Arlet stage I or II
and THA is used in Ficat-Arlet stage III or IV [25]. How-
ever, when performing THA in young patients, high rates of
failure have been reported despite continuous improvement
in the design and technique [4, 6, 15, 23, 24]. The etiology is
unknown but the disease is characterized by the death of
bone. Although in some cases the disease is static or pro-
gresses very slowly [20], in others it typically progresses to
collapse of the subchondral bone and articular cartilage of
the joint, ultimately leading to secondary degenerative
change causing pain and limitation of joint motion.
Various head preservation procedures have been repor-
ted, such as core decompression which reduces bone-
marrow pressure [8, 21], proximal femoral osteotomy [9],
nonvascularized or vascularized bone graft [3, 5], and
trochanteric or transtrochanteric rotational osteotomy [20,
27]. Sugioka’s transtrochanteric rotational osteotomy [27]
especially is known as an effective head preservation
procedure in younger patients. Despite some reports of
good results, not all have reported success with this
Volume 466, Number 5, May 2008
procedure [7, 26, 29]. Most complications and failure of
this procedure are related to its technically demanding
nature and the poor fixation stability provided by the
screws, causing delayed union and increased varus [7, 14,
22]. This led several authors to recommend fixation with
side plates and blade plates rather than with screws as
originally described by Sugioka [2, 14, 26].
We modified Sugioka’s osteotomy technique so the
greater trochanter was not detached, presuming our modi-
fication of the transtrochanteric rotational osteotomy
(TRO) produces less tissue trauma, and provides better hip
function. This method of TRO allows muscle-pedicle-bone
grafting to be performed at the same time. We believe
subsequent total hip arthroplasty would be easier in the
event of failure of this procedure.
We asked whether: (1) this modified technique of TRO
could prevent further collapse of most hips with ARCO II
and III osteonecrosis; (2) blood flow to the femoral head
would be restored following this procedure; (3) patients
would have improved functional abilities postoperatively;
and (4) there would be no increase in complication rates.
Materials and Methods
We retrospectively reviewed 39 patients (43 hips) with
osteonecrosis of the femoral head treated by a modified
TRO, and in whom followup was possible for at least
2 years. Association Research Circulation Osseous
(ARCO) stages [1] were used for classification of osteo-
necrosis (Table 1). The study population included 33 men
and six women, and the average patient age was 34.3 years
(range, 20–51 years). The causes of osteonecrosis were
excessive alcohol consumption in 15, steroid use in 11,
idiopathic in 12, and posttraumatic in five. The indications
for this surgery were (1) ARCO Stage II and ARCO Stage
III with less than 50% involvement, and (2) intact area of
the posterior or anterior femoral head greater than a third of
Table 1. Classification of ARCO [1] stage and treatment method
Stage Number of cases MTRO MTRO with simple bone graft
2-A-central 4 3 1
2-B-central 2 2
2-B-lateral 4 2
2-C-central 3 1
2-C-lateral 4 4
3-A-lateral 2 2
3-B-lateral 5 2 1 2
3-C-central 1 1
3-C-lateral 18 2 1
MTRO = modified transtrochanteric rotational osteotomy; MPBG = muscle-pedicle-bone graft.
Modified Transtrochanteric Rotational Osteotomy 1111
the total articular surface. We performed MTRO alone in
15 cases, a combination of MTRO and simple bone
grafting in three, and a combination of MTRO and muscle-
pedicle-bone grafting in 25. Anterior rotation was per-
formed in 40 cases and posterior rotation in three cases.
Bilateral surgery was performed in four patients. The
minimum followup duration was 24 months (average,
36.6 months; range, 24–52 months). No patients were lost
to followup.
Preoperative assessment of the patients included anter-
oposterior and lateral radiographs of the affected hip, MRI,
and bone scans. The extent of necrosis was determined by
the method described by Kerboul et al. [13] and the exact
location of the lesion was determined by MRI. This
information was used to determine the rotation angle
required for the modified transtrochanteric rotational oste-
otomy. Anterior rotated osteotomy was indicated if the
lesion involved less than the posterior third of the entire
femoral head on true lateral radiograph. Posterior rotated
osteotomy was indicated if the lesion involved more than
the posterior third of the entire femoral head but the
anterior portion of the head was still intact.
The surgery was performed by one surgeon (TRY). The
patient was placed in a lateral decubitus position, and a
lateral, straight, longitudinal skin incision of 12 to 17 cm
was made, depending on the size and weight of the patient.
The incision was centered over the greater trochanter and
extended from 5 cm below the vastus lateralis ridge of the
greater trochanter to 5 to 7 cm proximal to the greater
trochanter. We used a posterior approach to the hip joint.
The presence of the posterior branch of the medial cir-
cumflex artery was confirmed using Doppler (Minidop
ES-100VX; Habeco, Tokyo, Japan), and protected from
injury by not detaching the quadratus femoris. The short
external rotator muscles, except the quadratus femoris,
were completely transected. The joint capsule was circum-
ferentially incised close to the acetabular rim. A 3.5-mm
Steinman pin was driven centrally along the axis of the
MTRO with MPBG Cases with progression
2
2
1
1
15 1
123
1112 Yoon et al.
Fig. 1 The osteotomy sparing the greater trochanter is shown.
femoral neck from the lateral aspect of the greater
trochanter. Using a k-wire bent to 90° as a guide, the
transtrochanteric osteotomy site was determined perpen-
dicular to the wire just proximal to the intertrochanteric
line. In contrast to Sugioka’s technique [27], the greater
trochanter was not detached (Fig. 1). With mobilization of
the proximal fragment, any remnants of the capsule
attachment to the proximal fragment was confirmed and
detached. A Steinman pin was inserted into the proximal
fragment and was used as a joystick to rotate the proximal
fragment 70° to 90° in the anterior direction or 90° to 100°
in the posterior direction. (Fig. 2). Two or three Steinman
pins were inserted to fix the transtrochanteric osteotomy
site temporarily. The position and rotation were confirmed
by fluoroscopy. We used two or three 6.5-mm cancellous
screws to fix the osteotomy site and then removed the
temporary Steinman pins (Fig. 3). If we performed anterior
rotation of the proximal fragment, one of the cancellous
screws was inserted from the anteroinferior aspect of the
trochanteric area to the posterosuperior area of the femoral
neck to obtain fixation through the cortical bone, providing
added stability and vice-versa in a case of posterior rotation of
the proximal fragment. The remaining one or two cancellous
screws were fixed from the lateral side of the greater tro-
chanter to the femoral head. In the cases of ARCO Stage III
ON, bone grafting was performed. In a small collapsed lesion,
we performed simple bone grafting. However, if the lesion
was relatively large, muscle-pedicle-bone grafting from the
distal gluteus medius muscle and its attachment to the
anterior greater trochanter was performed. We created a
123
Clinical Orthopaedics and Related Research
Fig. 2 The femoral head was rotated anteriorly depending on the
location of the necrotic area.
window in the neck-head junction for insertion of this
muscle-pedicle-bone graft. Two cannulated cancellous
screws were used for fixation in 29 cases, and three
cannulated cancellous screws were used in 14 cases.
The average operation time was 159.3 minutes (range,
125–220 minutes). Hemoglobin level changed from a
preoperative mean of 12.4 gm/dL (range, 9.3–15.6 gm/dL)
to 10.2 gm/dL (range, 7.6–13.2 gm/dL) postoperatively.
Hematocrit levels dropped from 36.1% (range, 27.2%–45.1
%) to 26.6 % (range, 19.8%–32.3%).
We reconfirmed the presence of the artery with Doppler
(Minidop ES-100VX; Habeco, Tokyo, Japan) after confir-
mation of the fixation by fluoroscopy again and prior to
repair of the deep fascia, subcutaneous layer, and skin.
Postoperatively, patients began nonweightbearing crutch
walking within 1 week, depending on the general condition
of the patient. Range-of-motion exercises were recom-
mended during the postoperative period. The crutch
walking with nonweightbearing was continued until the
radiological bone union was confirmed, usually for 3 to
4 months. Radiological bone union was defined as absence
of the radiolucent transtrochanteric osteotomy line. Post-
operative scintigram was performed at 2 to 3 weeks and
3 months postoperatively.
We (TRY, SGC, JHL) performed a clinical evaluation
using the Harris hip score (HHS) [11]. Other hip functions
such as squatting and range of motion were also
evaluated.
Volume 466, Number 5, May 2008
Fig. 3 Fixation of the osteotomy with two cannulated screws is
shown.
We (TRY, AAA, CIH) used regular anteroposterior and
lateral radiographs to monitor femoral head collapse or
degenerative changes. Radiographs were taken at 3 weeks,
6 months, and 1 year, then annually after the operation.
Bone scan with pin-hole spectrometry was performed at 2
to 3 weeks and 3 months postoperatively.
If there was progression of osteonecrosis, if THAs were
performed for any reason, or if collapse occurred in the
followup period, the results were considered a ‘‘failure.’’
We considered the surgery successful if there was no
further collapse or increase in apparent necrotic area within
the minimum 24-month followup.
Results
Sixteen of 17 ARCO Stage II cases and 24 of 26 ARCO
Stage III cases had no progression of collapse or lesion
size. The location of the lesion in all three cases with
collapse was lateral and the size of lesion in two was over
30%. All three cases with progression were treated by
MTRO in combination with muscle-pedicle-bone graft or
simple bone graft. One of the three cases underwent THA.
The early postoperative scintigrams revealed increased
blood flow that gradually increased during the 3-month
followup studies.
Modified Transtrochanteric Rotational Osteotomy 1113
Table 2. Average range of motion of hip at last followup versus
preoperative values
Motion Hip range of motion
Preoperative Postoperative
Flexion 90.9° 98.7°
Internal rotation 8° 13.3°
External rotation 36.1° 48.4°
Abduction 24.3° 27°
Adduction 13.1° 14.8°
Excluding the three failed cases, the average preopera-
tive Harris hip score was 70 points (range, 59–82 points),
and it improved to 92 points (range, 77–100 points) at the
last followup. The average range of motion of hips at last
followup improved compared to that of preoperative values
(Table 2). Thirty-nine cases (90.7%) could squat down and
40 cases (93%) could sit cross-legged.
Bone union occurred at the osteotomy site in all cases.
The average union time was 14 weeks (range, 11–
22 weeks). The complications included delayed union in
one case, valgus angulation over 140° in two cases, and
osteophyte formation in three cases. In the two cases with
valgus angulation, the average postoperative neck-shaft
angle was 146°, but no additional treatment was performed.
We had no cases of infection or deep vein thrombosis.
Three patients had subcapital osteophyte formation on
radiograph, but no additional treatment was performed
because these patients were free of pain.
Discussion
We asked whether our technique of modified transtro-
chanteric rotational osteotomy could prevent further
collapse of most hips with ARCO II and III osteonecrosis,
whether blood flow to the femoral head would be restored
following this procedure, whether patients would have
improved functional abilities postoperatively and whether
these can be achieved without increased complication
rates.
Our study is limited by the short-term followup. Longer
followup would be necessary in this cohort of patients to
determine the true survivability of this procedure. This
study also lacked controls and the patients were not ran-
domized to any other treatment methods.
Osteonecrosis of the femoral head is a devastating
condition because of its propensity to affect young people
and its often unrelenting progression despite treatment
[30]. Treatments for osteonecrosis of the femoral head are
varied and can be largely categorized into joint-preserving
procedures and THA. Most patients are treated by THA.
123
1114 Yoon et al.
Fig. 4A–E (A) Radiographs from
a 42-year-old man with osteone-
crosis of the femoral head (Ficat-
Arlet Stage II) are shown. (B) An
axial MRI shows medial collapse.
(C) A coronal MRI shows necrosis
extending into the femoral neck.
(D) An immediate radiograph after
transtrochanteric rotational osteot-
omy with bone graft is shown. (E)
Radiographs at 14 months postop-
eratively showed good union of the
osteotomy site with good incorpo-
ration of grafted bone at the
necrotic area.
However, THA is not a definitive treatment because new
problems such as liner wear, osteolysis, and loosening have
developed, requiring later revision surgery. For these rea-
sons we believe THA is not the best choice for young
patients. Therefore, we prefer head-preserving procedures
for those who are young and diagnosed early.
A report on 474 patients treated with the Sugioka oste-
otomy [27] revealed a success rate of 78%, with higher
success rates seen in cases in earlier stages (stage II had an
89% success rate, stage III 73%, and stage IV 70%), and in
cases involving smaller lesions. A comparison of several
joint-preserving techniques was made by Saito et al. [22].
In their study, 54 hips were classified according to the
classification of Inoue and Ono [12] and treated with core
decompression (stage I, 17 cases), bone grafting (stage II,
18 cases), or osteotomy (stage III, 15 rotation, 4 varus).
Overall, good or excellent results were seen in 67% of
cases. For Stage II and III lesions, necrosis involving less
than 50% of the femoral head resulted in a success rate of
91% compared to 27% success in cases involving larger
lesions.
Sugioka [27] proposed a transtrochanteric rotational
osteotomy (TRO) to preserve the femoral head in young
patients when the posterior portion of the femoral head was
not involved. By rotating the femoral head, the diseased
weight-bearing surface is reestablished by repositioning the
123
Clinical Orthopaedics and Related Research
femoral head, reducing the forces on the diseased part to
prevent collapse and allow healing [16, 28]. Sugioka [28]
reported successful outcomes with rotation angles of
55°–70° in anterior rotated TRO and Atsumi and Kuroki
[2] reported successful outcomes with rotation angles of up
to 180° when performing a TRO with posterior rotation. As
with THA, TRO is not permanent but it can delay THA for
those young patients who have neither metabolic bone
disease nor articular destruction.
We do not recommend TRO in Ficat Stage I hips in
which core decompression seems more appropriate, or in
Stage IV hips in which head-preserving surgery is no
longer effective. For Stages II and III, although successful
results have been reported on short-term followup, long-
term followup results are variable and we believe unsatis-
factory [10, 17]. We judged TRO had a high success rate
for at least 24 months in Stage II, when success was
defined by absence of progression of collapse or size of the
osteonecrotic region (Table 1). Sugioka [28] reported 89%
success in hips followed 2 to 11 years postoperatively and
Masuda et al. [18] reported 82% success at average
5.1 years. In our series, ‘‘success’’ was defined as no evi-
dence of further progression of osteonecrosis. The success
rate for Stage II ON in our series was 94% (Fig. 4A–E).
For Stage III, there is no consensus in the literature con-
cerning results. Dean and Cabanela [7] had only 17%
Volume 466, Number 5, May 2008
satisfactory results, but Saito et al. [23] had 45%, Sugano
et al. [26] had 56%, and Sugioka [28] reported 73% sat-
isfactory results. Atsumi and Kuroki [2] reported
successful outcomes in 17 of 18 hips with FICAT III and
IV treated with posterior rotated transtrochanteric osteot-
omy at a minimum of 24 months of followup (mean,
42 months; range, 24–94 months). However, our study
revealed a 92% success rate in Stage III. The success rate
in our study may be related to a more limited indication for
the osteotomy and the use of supplementary bone and
muscle pedicle-bone-grafting. Muscle pedicle-bone-grafts
were indicated for ARCO Stage III hips where there was
relatively large extent of necrosis and collapse of the
femoral head. We believe the high success rate is partly
attributable to this muscle pedicle-bone-grafting, which
increased the vascularity of the femoral head. The location
of the lesions in all three cases with progression was lateral
and the size of the lesions in two cases was large; caution
should be taken if a lesion is lateral and large as in all other
preserving surgeries.
Saito et al. [22] reported a high postoperative compli-
cation rate of 33% in the rotational osteotomy group of that
series, consisting of femoral neck fracture in three patients
with renecrosis of the femoral head in two and two late
varus deformities. Sugioka [27] also reported complica-
tions related to his procedure, including lesser trochanteric
fracture in one, femoral neck fracture in four, delayed
union in five leading to increased varus deformities, and
progression of osteonecrosis in two. Ohzono et al. [20]
proposed a lack of skilled surgical technique or inappro-
priate patient selection or fixation causes a high failure rate.
Our complications were not greatly different from those
reported in the literature. There was one case of delayed
union at 22.2 weeks, two cases had valgus angulation of
the osteotomy, and three cases had osteophyte formation,
all of which did not require any surgical intervention.
Valgus position was one of our complications that devel-
oped during our early experience with the modified TRO
technique, as part of the learning curve. As we gained more
experience, we utilized intraoperative fluoroscopy and took
extra care to avoid valgus positioning of the osteotomy.
Varus positioning has the disadvantage of limb shortening
so we did not perform intentional varus positioning to
avoid limb length discrepancy. We believe the results of
TRO depend on the indications for the surgery and the skill
of the surgeons, both of which may vary.
In comparison with the original technique of TRO, a
MTRO has several advantages. There is no need to detach
and reattach the greater trochanter. Therefore, there is no
concern about nonunion of the greater trochanter. The
quadratus femoris is also not detached. By redirecting one
of the screws for fixation of the osteotomy, we have had no
cases of nonunion of the osteotomy site. Additional
Modified Transtrochanteric Rotational Osteotomy 1115
procedures such as muscle-pedicle-bone graft can be
combined with MTRO. The postoperative rehabilitation as
described by Sugioka [27] includes continuous skin trac-
tion for 1 week, followed by nightly skin traction for a
further 2 weeks. Active hip exercises commenced within
2 weeks postoperatively. In our series, early rehabilitation
is possible as skin traction is not required postoperatively
and our patients begin hip mobilization exercises the day
after surgery. Subsequent THA may be easier because the
anatomy of the greater trochanter is not altered and
the quadratus femoris, which may be important to preserve
the circulation, is not detached. When the posterior branch
of the medial circumflex artery is injured during surgery,
the circulation through the quadratus femoris may be
maintained. One study suggested quadratus femoris mus-
cle-pedicle-bone grafting for treating displaced femoral
neck fractures since the microcirculation in the muscle
would theoretically provide additional blood flow through
the quadratus femoris muscle [19].
Most of our patients had rotational osteotomy without a
greater trochanteric osteotomy. With this method, we were
able to perform anterior and posterior transtrochanteric
rotation. Bone grafting was a supplementary procedure and
was performed with a small window through the head-neck
junction. We believe by adding this small procedure, the
success rate could be improved. With the combination of
rotational osteotomy and muscle pedicle bone grafting the
indication for this MTRO could be further extended.
This procedure prevented progression (no collapse or
increase regions of ON) in 93% by the minimum 24-month
followup. In the short term the MTRO seems effective in
Stages II and III with less than 50% involvement in either
anterior or posterior portions of the femoral head. We
therefore believe this technique is promising in young
patients with Stage II or III osteonecrosis of the femoral
head.
References
1. ARCO (Association Research Circulation Osseous). Committee
on Terminology and Classification. ARCO News. 1992;4:41–46.
2. Atsumi T, Kuroki Y. Modified Sugioka’s osteotomy: More than
130° posterior rotation for osteonecrosis of the femoral head with
large lesion. Clin Orthop Relat Res. 1997;334:98–107.
3. Bonfiglio M, Bardenstein MB. Treatment by bone-grafting of
aseptic necrosis of the femoral head and non-union of the femoral
neck (Phemister technique). J Bone Joint Surg Am. 1958;40:
1329–1346.
4. Brinker MR, Rosenberg AG, Kull L, Galante JO. Primary total
hip arthroplasty using noncemented porous coated femoral
components in patients with osteonecrosis of the femoral head.
J Arthroplasty. 1994;9:457–468.
5. Buckley PD, Gearen PF, Petty RW. Structural bone-grafting for
early atraumatic avascular necrosis of the femoral head. J Bone
Joint Surg Am. 1991;73:1357–1364.
123
1116 Yoon et al.
6. Coventry MB, Beckenbaugh RD, Nolan DR, Ilstrup DM. 2,012
total hip arthroplasties: A study of postoperative course
and early complications. J Bone Joint Surg Am. 1974;56:
273–284.
7. Dean MT, Cabanela ME. Transtrochanteric anterior rotational
osteotomy for avascular necrosis of the femoral head. Long-term
results. J Bone Joint Surg Br. 1993;75:597–601.
8. Fairbank AC, Bhatia D, Jinnah RH, Hungerford DS. Long-term
results of core decompression for ischemic necrosis of the fem-
oral head. J Bone Joint Surg Br. 1995;77:42–49.
9. Fuchs B, Knothe U, Hertel R, Ganz R. Femoral osteotomy and
iliac graft vascularization for femoral head osteonecrosis. Clin
Orthop Relat Res. 2003;412:84–93.
10. Ganz R, Buchler U. Overview of attempts to revitalize the dead
head in aseptic necrosis of the femoral head–osteotomy and
revascularization. In: Hungerford DS, ed. Hip. St Louis, MO:
Mosby;1983:96–305.
11. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
12. Inoue A, Ono K. A histological study of idiopathic avascular
necrosis of the head of the femur. J Bone Joint Surg Br. 1979;61-
B:138–143.
13. Kerboul M, Thomine J, Postel M, D’Aubigne RM. The conser-
vative surgical treatment of idiopathic aseptic necrosis of the
femoral head. J Bone Joint Surg Br. 1974;56:291–296.
14. Kinnard P, Lirette R. The Borden and Gearen modification of the
Sugioka transtrochanteric rotational osteotomy in avascular
necrosis: A preliminary report. Clin Orthop Relat Res.
1990;255:194–197.
15. Krackow KA, Mont MA, Maar DC. Limited femoral endopros-
thesis for avascular necrosis of the femoral head. Orthop Rev.
1993;22:457–463.
16. Lafforgue P, Dahan E, Chagnaud C, Schiano A, Kasbarian M,
Acquaviva PC. Early-stage avascular necrosis of the femoral
head: MR imaging for prognosis in 31 cases with at least 2 years
of follow-up. Radiology. 1993;187:199–204.
17. Maistrelli G, Fusco U, Avai A, Bombelli R. Osteonecrosis of the
hip treated by intertrochanteric osteotomy. A four- to 15-year
follow-up. J Bone Joint Surg Br. 1988;70:761–766.
18. Masuda T, Matsuno T, Hasegawa I, Kanno T, Ichioka Y, Kaneda K.
Results of transtrochanteric rotational osteotomy for nontraumatic
123
Clinical Orthopaedics and Related Research
osteonecrosis of the femoral head. Clin Orthop Relat Res.
1988;228:69–74.
19. Morewessel R, Evarts CM. The use of quadratus femoris muscle
pedicle bone graft for the treatment of displaced femoral neck
fractures. Orthopedics. 1985;8:972–976.
20. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
21. Rosenwasser MP, Garino JP, Kiernan HA, Michelsen CB. Long
term followup of thorough debridement and cancellous bone
grafting of the femoral head for avascular necrosis. Clin Orthop
Relat Res. 1994;306:17–27.
22. Saito S, Ohzono K, Ono K. Joint-preserving operations for idi-
opathic avascular necrosis of the femoral head: Results of core
decompression, grafting and osteotomy. J Bone Joint Surg Br.
1988;70:78–84.
23. Saito S, Saito M, Nishina T, Ohzono K, Ono K. Long-term results
of total hip arthroplasty for osteonecrosis of the femoral head. A
comparison with osteoarthritis. Clin Orthop Relat Res.
1989;244:198–207.
24. Stauffer RN. Ten-year follow-up study of total hip replacement.
J Bone Joint Surg Am. 1982;64:983–990.
25. Steinberg ME. Management of avascular necrosis of the femoral
head–an overview. Instr Course Lect. 1988;37:41–50.
26. Sugano N, Takaoka K, Ohzono K, Matsui M, Saito M, Saito S.
Rotational osteotomy for non-traumatic avascular necrosis of the
femoral head. J Bone Joint Surg Br. 1992;74:734–739.
27. Sugioka Y. Transtrochanteric anterior rotational osteotomy of the
femoral head in the treatment of osteonecrosis affecting the hip: a
new osteotomy operation. Clin Orthop Relat Res. 1978;130:
191–201.
28. Sugioka Y. Transtrochanteric rotational osteotomy in the treat-
ment of idiopathic and steroid induced femoral head necrosis,
Perthes’ disease, slipped capital femoral epiphysis, and osteoar-
thritis of the hip. Indications and results. Clin Orthop Relat Res.
1984;184:12–23.
29. Tooke SMT, Amstutz HC, Hedley AK. Results of transtrochan-
teric rotational osteotomy for femoral head osteonecrosis. Clin
Orthop Relat Res. 1987;224:150–157.
30. Vail TP, Covington DB. The incidence of osteonecrosis. In
Urbaniak JR, Jones JP, eds. Osteonecrosis : Etiology, Diagnosis,
and Treatment. Chicago, Ill: American Academy of Orthopaedic
Surgeons; 1997:43–49.
Clin Orthop Relat Res (2008) 466:1117–1124
DOI 10.1007/s11999-008-0201-z
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Vascularized Fibular Grafting for Osteonecrosis of the Femoral
Head With Unusual Indications
J. Mack Aldridge III MD, James R. Urbaniak MD
Published online: 21 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract We retrospectively reviewed the charts of 154
patients of various subgroups treated with the free vascu-
larized fibular graft procedure for osteonecrosis of the
femoral head (ONFH), evaluating pre- and postoperative
Harris hip scores, hip range of motion, radiographs, and
number of conversions to total hip arthroplasty (THA).
Patients were followed a minimum of 1 year (mean,
6.8 years, range, 1 to 19 years). Athletes and patients with
pyarthrosis-related osteonecrosis had high Harris hip scores
at final review with scores of 94 and 97, respectively. Patients
with ONFH after a slipped capital femoral epiphysis or fol-
lowing pregnancy had a low conversion rate to THA at 6%
and 8%, respectively. Twenty-five percent of patients with
transplant-related osteonecrosis of the femoral head were
converted to THA at an average of 2.7 years. However, with
select subsets of patients (athletes, pregnancy, organ trans-
plant, femoral neck non-union, slipped capital femoral
epiphysis, infection) the FVFG can result in a high rate of
success.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
J. M. Aldridge III (&), J. R. Urbaniak
Division of Orthopaedic Surgery, Duke University Medical
Center, 3116 North Duke Street, Durham, NC 27704, USA
e-mail: aldri004@mc.duke.edu; julian.aldridge@duke.edu
Introduction
Free vascularized fibular grafting (FVFG) for the treatment of
osteonecrosis of the femoral head (ONFH) was first described
in the 1970s concurrent with the emergence of microsurgical
techniques and during a time when durable and predictable
hip arthroplasty components were not readily available.
Because ONFH tends to affect younger patients, most
surgeons have traditionally believed any procedure aimed at
preserving the patient’s native hip preferable. That remains
true today, although our enthusiasm for the FVFG procedure
in patients between the ages of 40 and 50 years with pre-
operative femoral head collapse is tempered with our
acknowledgment of time-proven hip arthroplasty as a rea-
sonable surgical alternative. We recognize the FVFG
procedure is not a panacea for ONFH nor is it indicated for all
patients with ONFH; however, drawing from our experience
with over 7000 cases of ONFH and performing over 2800
FVFG procedures, certain trends have emerged with respect
to patient selection, outcomes, and technical refinements.
We have reviewed select subsets of patients with ONFH
treated with the FVFG and have updated reviews of previ-
ously reviewed groups [11, 16, 18] in hopes of identifying in
which patients this procedure is beneficial. We provide
evidence for our belief that the FVFG procedure remains a
reasonable and effective option in patients with ONFH and
will and continue to have a role in the foreseeable future.
Materials and Methods
For this article we retrospectively reviewed the charts of
selected nonconsecutive subsets of over 2,800 patients who
underwent FVFG and whom we presumed had performed
better postoperatively than our general population with
123
1118 Aldridge and Urbaniak
ONFH. These subsets included: athletes (n = 15, 19 hips),
pregnancy (n = 39 patients, 41 hips), organ transplant
(n = 36 patients, 47 hips), femoral neck nonunion (n = 22
patients, 22 hips), slipped capital femoral epiphysis (n = 31
patients, 36 hips), infection (n = 11 patients, 11 hips). We
prospectively contacted all patients directly in these subsets
with followup questionnaires (Harris hip score and SF-12).
For the physical examination portion of the Harris Hip
evaluation, we either brought patients back to the clinic for
an examination or, for patients who lived too far away to
return directly to us, we used hip range of motion measure-
ments from their local orthopaedic surgeon. All patients were
diagnosed with osteonecrosis of one or both femoral heads
and were referred to our institution where either author
(JMA, JRU) performed a vascularized fibular graft. The
surgical technique has been described previously [1] and has
changed little in the 30 years since it was first introduced. A
few minor and recent additions to the procedure are detailed
in the Discussion section of this paper. Both authors perform
the procedure identically. We obtained prior approval from
the Committee on Clinical Investigation of our Institutional
Review Board. We offer the procedure to patients under
50 years of age, with symptomatic ONFH, and with little or
no articular stepoff in the femoral head.
Patients were followed at the following postoperative
times: six weeks, 3 months, 6 months, one year, then yearly
thereafter. Preoperatively and on each postoperative visit we
(JRU, JMA) determined hip range of motion and the Harris
hip score. We determined patient satisfaction by the SF-12
form. We consider an outcome as successful if the patient has
an improvement in or absence of pain, an improved Harris
hip score, and preservation of their native hip (ie, no hip
arthroplasty) at last followup.
We also obtained an anteroposterior (AP) pelvis and
frog leg lateral radiographs of the operative hip at each
visit. We (JRU, JMA) determined on each film the size and
location of the necrotic lesion and the presence or absence
of collapse and/or joint space narrowing. Size of the lesion
was determined by the authors’ (JRU, JMA) estimation of
percentage involvement on the AP and lateral radiograph
and the coronal and axial magnetic resonance imaging.
From this, a number is then selected that represents a
percentage of a sphere.
Athletes with Osteonecrosis of the Femoral Head
Athletes with their drive to succeed and disciplined approach
to overcoming physical challenges appear to have particular
benefit from the FVFG procedure. Coincidentally, this is a
group of patients in whom a biology-sparing procedure is
particularly attractive for allowing unrestricted return to
sporting activities. The obligate restrictions of hip
123
Clinical Orthopaedics and Related Research
resurfacing or THA frequently herald the end of competitive
sports participation and, as such, are typically viewed as last
options for this group of patients.
We have performed the FVFG procedure on 15 athletes
(19 hips). The various sports included baseball, basketball
(one patient was on a national women’s basketball team),
ballet, football, roping horses, downhill skiing, triathlons,
gymnastics, and volleyball. Not included was the typical
‘‘weekend athlete,’’ but rather these were all competitive
athletes on the high school, intercollegiate, or professional
level. The average age of the patients was 27.8 years (range,
11–43 years). There were 10 male and five female patients.
The hips were classified as Stage II in six patients, Stage III in
two patients, Stage IV in six patients, and Stage V in one
patient, using the Steinberg classification. Etiologies of
ONFH included idiopathic in eight, trauma in two, alcohol in
two, corticosteroids in one, Legg-Calvé-Perthes in one, and
hemangioma of the femoral head with osteonecrosis in one.
Followup was available for all 15 athletes at a mean of
8 years (range, 1.8–19 years).
The mean Harris hip score improved from 73.6 (range,
17–96) preoperatively to 94 postoperatively. Average final
hip flexion for 12 hips in which range of motion data were
available was 113° (range, 75°–130°). Radiographic changes
were noted in three patients (two joint space narrowing; one
femoral head collapse), whereas the remaining 12 patients
showed no further collapse or change in their hip joint space.
Eleven patients (73%) postoperatively returned to athletic
activities and eight of these patients believed they could
perform at the same level of competition as before devel-
oping ONFH. One such patient returned to competitive
football within 9 months following the FVFG procedure
(Fig. 1A–D). Three hips were converted to THA at an
average of 10 years (one hip at 3 years, one at 9 years, and
one at 17 years after FVFG). Patients rated their overall
satisfaction (SF-12) with the procedure as extremely satis-
fied in 10, moderately in four, and slightly satisfied in one.
Pregnancy-related Osteonecrosis of the Femoral Head
We earlier reported treating female patients with the FVFG
procedure for pregnancy-related ONFH [14]. The average
age in this cohort of 39 patients was 33.1 years. The stage at
time of presentation for the group was Steinberg Stage I
(n = 1), Stage II (n = 8), Stage III (n = 3), Stage IV
(n = 25), Stage V (n = 1), and Stage VI (n = 1). We fol-
lowed these patients a minimum of 24 months (mean,
46 months; range, 24 to 95 months). Of these 39 patients,
one presented with unilateral disease of the right femoral
head, whereas two had bilateral involvement and the
remainder (n = 36) had ONFH of the left femoral head. Our
results with the FVFG procedure for pregnancy-related
Volume 466, Number 5, May 2008 Vascularized Fibular Grafting for Hip ON 1119

Fig. 1A–D (A) A pelvic MRI of a

sixteen year old with septic-related
ONFH is shown. (B) This lateral
radiograph shows increased den-
sity of the epiphysis and
irregularity and collapse of the
subchondral bone. (C) The one-
year radiograph demonstrates
good fibular incorporation and
preservation of the femoral head
preoperative state. In fact, it
appears there has been some
improvement. (D) Three years
after the FVFG procedure, the
patient played quarterback for his
high school football team.

ONFH have been consistently good with a 94% success rate

at 5 years. Average preoperative Harris hip score was 44.2
(range, 26–93) for the group, improving to 80.2 (range,
22–100) at final followup. Hip motion remained the same or
improved for every patient. Two patients’ hips were con-
verted to THA at 2 and 14 years after the FVFG procedure.
Both of these patients had collapse of the femoral head prior
to surgery with approximately 40% involvement of the
femoral heads.
When evaluating a pregnant patient with new-onset hip
pain, it is important to distinguish ONFH from bone marrow
edema syndrome (BME), a condition known to occur with Fig. 2 A T1-weighted coronal MRI image shows a patient with
greater frequency in pregnant women. BME is a poorly osteonecrosis of the left femoral head. The arrow is pointing to the
understood entity but is easily distinguished from ONFH characteristic serpiginous border between nonviable and viable bone.
because the latter has a characteristic serpiginous border
Followup was available for all 36 patients at a minimum of
between viable and nonviable bone seen on T1- and T2-
1 year (mean, 4.3 years; range, 1–16 years). Eleven patients
weighted MRI images (Fig. 2). BME will not have this dis-
(23%) have had conversions to THA. The average time to
tinct serpiginous border and, unlike ONFH, BME will have
conversion was 2.7 years (range, 1.5–4.0 years). Preopera-
bone marrow signal changes extending down the femoral neck
tively, eight of those 11 patients had at least 50% estimated
into the trochanteric region. Furthermore the MRI changes
involvement of the femoral head. We now consider the
seen in BME syndrome ultimately reverse to normal, whereas
percentage of femoral head involvement in our decision to
those changes from ONFH persist or enlarge with time.
offer the FVFG procedure. We are continually challenged in
deciding who would benefit most from this procedure,
Patients Who Have Undergone Transplantation especially because patients having undergone renal trans-
plantation are also not ideal recipients of THA. Although a
We have performed the FVFG procedure for transplant- 25% THA conversion rate might imply to some a contrain-
dependent steroid-induced ONFH in 36 patients (47 hips). dication to the FVFG procedure, we believe the many more

123
1120 Aldridge and Urbaniak
patients who continue to function well with their native hips
is rather an indication. For example, a 32-year-old man
underwent renal transplantation for glomerulosclerosis and
2 years afterward developed left-sided groin pain. Sub-
sequent workup revealed Steinberg et al. [16] Stage IVB
ONFH and he was referred to our institution for consider-
ation of the FVFG procedure. Because of his young age,
preserved joint space and, equally important, his hip motion,
we performed FVFG. He is now 8 years postsurgery and
walking without a limp or hip discomfort. His Harris hip
score improved from 73 to 100 at final followup. Harris hip
scores for this cohort improved from 58.6 (range, 18–83)
preoperatively to 81.3 (range, 33–100) postoperatively.
Femoral Neck Nonunion
A unique group of patients to whom we offer FVFG are
those patients with concomitant ONFH and femoral neck
nonunion. The risk of developing posttraumatic ONFH in
patients with femoral neck fractures is well established [3,
18]. The concern for nonunion in this particular fracture is
also well recognized in the literature [6]. Despite an
awareness of these potential complications and a timely
and anatomic reduction of the fracture, not infrequently,
either femoral neck nonunion and/or ONFH still occur.
This is a particularly devastating problem for the younger
Fig. 3 An 18-year-old patient is shown with Stage V osteonecrosis of
the femoral head and femoral neck nonunion before surgery,
4 months after, and 10 years after the free vascularized fibular
grafting procedure. The femoral neck has healed and there has been
no interval collapse of the femoral head with preservation of the joint
space. The reduction, screw fixation and FVFG are performed
123
Clinical Orthopaedics and Related Research
patient (\ 40 years of age) in whom THA is considered a
salvage procedure and imparts certain lifestyle restrictions,
potentially interfering with avocational and vocational
interests.
We recently reported our experience with vascularized
fibular bone grafting in this particular set of patients [10].
We have performed the FVFG in 23 patients with com-
bined femoral neck nonunion and ONFH. The average age
of these patients was 28.7 years (range,10–49 years). Of
these 23 patients, four presented with Steinberg Stage I, 12
with Stage II, 2 with Stage III, four with Stage IV, and one
with Stage V ONFH. The minimum followup was
24 months (mean, 82 months; range, 24–195 months). One
such case involved an 18-year-old girl who presented with
a 2-year-old femoral neck nonunion with ONFH (Fig. 3).
We treated her hip with open reduction, internal fixation,
and placement of a FVFG. She ultimately healed and
reported ‘‘occasional aches’’ in her hip at long-term fol-
lowup. She was able to deliver two children subsequent to
the FVFG procedure. Twenty of the 22 nonunions healed at
an average time of 9.9 months (range, 3–23 months). Two
underwent additional procedures to facilitate healing.
Ultimately all fractures healed. The final median Harris hip
score for the group was 78.9 (range, 60–98). Two hips were
converted to THA an average of 1 year after surgery
because of femoral head collapse and accelerated hip
arthrosis. These two patients presented with 40%
contemporaneously with placement of the screws preceding the
insertion of the fibula, thus ensuring the screw threads do not disrupt
the vascular pedicle. Visual inspection of the core with no screw
threads visible or palpable, confirms safe placement of the screws.
Two screws and one FVFG have proven adequate in our experience
treating this entity.
Volume 466, Number 5, May 2008
involvement and no preoperative collapse of the femoral
head; however, both patients were noncompliant and dis-
carded their crutches prematurely 2 months after the
surgery.
Slipped Capital Femoral Epiphysis
The development of ONFH is a well recognized compli-
cation of a slipped capital femoral epiphysis [9]. We have
performed FVFG on 31 patients (36 hips; 26 unilateral, five
bilateral) with slipped capital femoral epiphysis compli-
cated by ONFH with an average patient age of 13.4 years
(range, 9–17 years) [5]. There were 20 male and 11 female
patients. All patients in this particular group presented with
some degree of femoral head collapse (Fig. 4A–B): one
patient with Duke Stage 3 (3.6%), 24 patients with Stage 4
(75%), and six patients with Stage 5 (21.4%) disease. We
followed these patients a minimum of 26 months (mean,
75 months; range, 26 months to 120 months). Followup
was possible for all patients. Three hips in two patients had
been converted to THA at the time of this review. One
additional patient’s hip had been converted to fusion. The
overall hip survival rate was 91.9% at 5 years. Reported
pain and functional activity improved in all patients and
Harris hip scores increased from an average preoperative
score of 55 (range, 11–90) to an average postoperative
score of 82.8 (range, 28–100). Twenty-seven of 28 patients
were extremely or very satisfied at final followup.
Infection
ONFH after pyarthrosis of the hip can be difficult to treat
because there is often bony destruction at the capsular
reflection (head/neck junction) in addition to the osteone-
crosis within the femoral head. Furthermore, many surgeons
consider a history of infection a contraindication to the use of
large metal implants if there is femoral head-neck disconti-
nuity. For these patients we routinely use FVFG and have
found the structural integrity of the fibular graft is sufficient
to allow rigid fixation of the neck while also addressing the
Fig. 4A–B (A) Preoperative ant-
eroposterior radiograph of a 13-
year-old boy shows a slipped
capital femoral epiphysis on the
left. Notice the large cyst and
increased density of the epiphysis,
both indicative of osteonecrosis of
the femoral head. (B) The same
patient is shown at 10-year
followup.
Vascularized Fibular Grafting for Hip ON 1121
ONFH (Fig. 5A–C). We augment the FVFG with one or two
screws, with less concern about implanting nonbiologic
materials in a previously infected area, because we believe
the extra delivery of blood flow diminishes this concern for
reinfection.
We have treated 11 such patients with an average age of
12.5 years (range, 9–20 years) for ONFH after pyarthrosis
of the hip [17]. This cohort included eight male and three
female patients. Followup was possible for nine patients at
a minimum of 2 years (mean, 4 years; range, 2–5 years)
after surgery. The mean preoperative Harris hip score for
the group was 68 (range, 62–79), which improved to 97
(range, 90–100) postoperatively. All patients presented
with femoral head collapse of some degree, ranging from
flattening of the femoral head up to 3 mm of articular
stepoff. Despite this advanced presentation, none of these
patients had subsequent or conversion surgery, and as a
subgroup of patients, they have the highest Harris hip
scores at most recent followup.
Discussion
We believe it important to review the context in which the
concept for vascularized bone grafting for the treatment of
ONFH was conceived. In the late 1960s and early 1970s,
when so many of the biologic-preserving procedures were
developed, a durable and reliable artificial hip prosthesis
largely evaded orthopaedic surgeons. Today, technologi-
cally advanced metals and polymers provide excellent wear
characteristics that have allowed surgeons greater freedom
of implanting such components in much younger patients
who, only years ago, would have not been offered that
option. Similarly, refinements in the surgical technique
have provided greater longevity in total hip arthroplasty
(THA), as have improvements in component fixation
(biologic ingrowth, third-generation cement technique).
For these reasons, THA has become a more durable and
reliable procedure. However; these implants do have a
finite lifespan inversely proportional to the patient’s
activity level [4, 7, 8, 11, 15], which in younger patients
can be very demanding. Because a younger patient will
123
1122 Aldridge and Urbaniak Clinical Orthopaedics and Related Research

Fig. 5A–C (A) Lateral radio-

graph shows the proximal femur
of a 14-year-old girl. The infection
resulted in femoral head-neck dis-
continuity and femoral head
osteonecrosis. (B) One year after
the free vascularized fibular graft-
ing procedure. Screws are placed
before inserting the fibula to pre-
vent the head from spinning during
the reaming portion of the case.
Notice the femoral head has healed
to the neck and there is continued
remodeling of the anterior portion
of the femoral head. (C) Two years
after free vascularized fibular
grafting, the cortices of the fibular
graft are less noticeable indicating
good incorporation. Femoral head
shape and joint space are relatively
well preserved.

likely outlive his or her hip prosthesis, a scenario requiring
one or more revision surgeries, preserving the native hip of
a patient with some disabling hip abnormality for as long as
possible has obvious benefit.
We acknowledge the lack of uniform longer-term
followup among the various groups studied. The senior
author (JRU) has been performing the FVFG procedure for
30 years and a full-time research analyst has been engaged
in clinical data collection on these patients for nearly
20 years. Consistent long-term followup is often difficult
because the majority of patients are referred from all
regions of the United States and other countries, which
challenges consistent acquisition of yearly followup. Also,
there are small numbers of patients in all groups because of
the unusual circumstances leading to the development of
the ONFH. We also acknowledge that since our patients are
all relatively young they and their surgeon jointly want to
delay arthroplasty as long as possible. Such deferment of
arthroplasty can lead to bias in interpreting the benefit and/
or longevity of the FVFG procedure. Finally, the estimates
of involvement of the femoral head were based on the
subjective evaluations, rather than a formal attempt to
quantify involvement. These evaluations limit any inter-
pretation regarding lesion size.
It is well established ONFH, if left untreated, progresses
to femoral head collapse with subsequent hip degeneration
in the majority of cases [13]. Despite recent technologic
advances in THA, an excellent option for the older patient
with an arthritic hip joint, it remains difficult for the
arthroplasty surgeon to feel justified in proceeding with
THA in patients younger than 50 years of age simply for
ONFH-related pain. This is particularly true for the
symptomatic patient without femoral head collapse, a
preserved joint space, and no acetabular involvement. As
we have contended since the original procedure was per-
formed over 30 years ago by the senior author (JRU), the

123
Volume 466, Number 5, May 2008
ideal patient is young with a small lesion, no femoral head
collapse, and good hip motion. Any deviation from this
ideal candidate must be the result of careful consideration
of numerous variables by the operative surgeon, which
often comes only with experience.
Several approaches to treatment for osteonecrosis of the
femoral head have been described, including weight
restriction and observation, core decompression, various
osteotomies, bone grafting (structural or nonstructural and
vascularized [2] or nonvascularized), and arthroplasty
(hemiarthroplasty, resurfacing techniques, and THAs). Few
centers around the world perform the FVFG with great
frequency. From China, Zhang and colleagues [21] repor-
ted their experience with 48 patients (56 hips) undergoing
the FVFG procedure with an average patient age of
37.7 years and an average followup of 16 months. Etiolo-
gies included trauma, steroids, alcohol, and idiopathic.
Patients who had Steinberg Grade [16] II ONFH had better
Harris hip scores at last followup than did patients who had
Steinberg Stage [16] III or IV disease. The preoperative
Harris hip scores for patients who had Steinberg Grades
[16] II, III, and IV ONFH were 78.5, 69.3, and 58.4,
respectively. At the most recent followup, the Harris hip
scores improved to 94.4, 85.7, and 76.4, respectively.
Shaffer [12] reported his series of 101 hips treated with the
FVFG procedure with a minimum followup of 5 years.
Sixty-one percent of the hips had not been converted to
THA at the 5-year mark and 42% survived until the 8-year
postoperative mark. The average Harris hip score for the
cohort improved from 58 ± 13 preoperatively to 80 ± 15
at the 5-year mark. The majority of these patients had
preoperative femoral head collapse [12].
A recent level IV study [20] reported the use of the new
commercially available tantalum implant from Zimmer
(Warsaw, IN). The authors contend the early results are
equal to those of patients treated with the FVFG. The
24-month survival rate for the tantalum dowel was 81.7%,
which decreased to 68.1% at 48 months. The survivorship
improved to 92% at 48 months if patients with systemic
disease were excluded. However, it is typically these
excluded patients who comprise our series and whom we
believe would be underserved with a nonvascularized,
nonbiologic implant. We also avoid implanting metal in
women who are pregnant or of child-bearing age, and those
patients with a history of a hip infection, or renal com-
promise (i.e., renal transplants). We are further concerned
about the presence of metal in the femur at the time of
THA should that be needed in the future.
Some modifications to the FVFG surgical technique are
worthy of mention. We have recently begun obtaining an
intraoperative arteriogram after the fibula has been placed
within the femoral head but before the anastomosis is
completed. This ensures the pedicle is not kinked,
Vascularized Fibular Grafting for Hip ON 1123
strangulated, or otherwise not flowing properly. Renogr-
affin can be visualized flowing up the main pedicle into the
femoral head and diffusing into the periosteal and endos-
teal vessels. If this is not confirmed, the fibula is removed,
the pedicle is checked, and the core may be enlarged
slightly. Also, in 2007 we began using, in certain cases,
computer navigation with standard fluoroscopy and/or CT
for exact localization of the necrotic lesion. This serves to
increase accuracy of the core placement into the necrotic
lesion while decreasing the total radiation exposure. We
are scientifically studying the benefit of this modality.
There are several unique benefits to this procedure,
namely its ability to preserve the native hip in young
patients. In addition, there are no bridges burned should
this procedure fail [5]. Lastly there is the potential to return
to full-time activity in heavy work or athletics. The greatest
drawback is prolonged protective weightbearing approxi-
mately 3 to 4 months if no preoperative collapse, or
6 months if preoperative collapse was present. The pro-
cedure is criticized for prolonged operative time; however
in our hands it is usually 3 hours ± 30 minutes. The donor
site morbidity has been minimal [19]. The most common
complication, occurring in 3% of patients, is a flexion
contracture of the flexor hallucis longus (FHL). If this
becomes bothersome to the patient it can easily be cor-
rected with ‘‘Z’’ lengthening of the tendon at the ankle
level. Only one patient has had a persistent distal tibio-
fibular syndesmotic problem advanced enough to warrant
surgery. Expense is also often cited as a negative factor;
however, there is no cost for the implant (which may be
considerable in the lifetime of a young patient) and the
duration of hospital stay is only 3 days.
We believe the left femoral head is uniquely involved in
pregnancy-related osteonecrosis because the growing fetus
and uterus compress the left common iliac vein, which is
directly posterior to the uterus on the left side. Such com-
pression decreases venous egress from the femoral head, in
turn increasing vascular capacitance, thus decreasing arte-
rial perfusion, and ultimately resulting in an intraosseous
compartment syndrome. This local anoxia leads to the
demise of bone-forming and supporting cells. Perhaps there
is also some synergy between the hypercoagulability of
pregnancy and this femoral head venous congestion.
Patients who have undergone either solid organ or bone
marrow transplantation are, with few exceptions, commit-
ted to a prolonged period of corticosteroid administration
to curtail immunologic-mediated rejection. Although the
benefits of steroids in this context are unquestioned for
prolonging life, such chronic exposure to high levels of
steroids exposes this group of patients to a high risk of
developing osteonecrosis of the femoral head(s). We have
observed several unique features within this particular
group of patients. Patients who have undergone
123
1124 Aldridge and Urbaniak
transplantation and who have ONFH tend to have larger
necrotic areas within the femoral heads; bilateral involve-
ment is more common than unilateral disease, and these
patients tend to present later in the disease process than
patients with ONFH from other causes. We believe this is
primarily because the requisite steroids suppress the oste-
onecrosis-related synovitis, which often alerts the patients
to the presence of the disease process.
While we continue to have success with the FVFG in
younger patients with idiopathic ONFH, we have also used
this procedure to treat special subsets of patients repre-
sented in this paper. We present our data for these groups
and contend the FVFG remains a reasonable option in the
treatment of ONFH.
Acknowledgments We thank Dawn Pedrotty, PhD, MS, and
Jennifer Friend for their assistance in gathering the data for this paper.
References
1. Aldridge JM 3rd, Berend KR, Gunneson EE, Urbaniak JR. Free
vascularized fibular grafting for the treatment of postcollapse
osteonecrosis of the femoral head. Surgical Technique. J Bone
Joint Surg Am. 2004;86(Suppl 1):87–101.
2. Aldridge JM 3rd, Urbaniak JR. Avascular necrosis of the femoral
head: role of vascularized bone grafts. Orthop Clin North Am.
2007;38:13–22.
3. Bachiller FG, Caballer AP, Portal LF. Avascular necrosis of the
femoral head after femoral neck fracture. Clin Orthop Relat Res.
2002;399:87–109.
4. Berend KR, Gunneson E, Urbaniak JR, Vail TP. Hip arthroplasty
after failed free vascularized fibular grafting for osteonecrosis in
young patients. J Arthroplasty. 2003;18:411–419.
5. Dean GS, Kine RC, Fithc RD, Gunneson EE, Urbaniak JR. Treatment
of osteonecrosis in the hip of pediatric patients by free vascularized
fibular graft. Clin Orthop Relat Res. 2001;386:106–113.
6. Dedrick DK, Mackenzie JR, Burney RE. Complications of fem-
oral neck fracture in young adults. J Trauma. 1986;26:932–937.
7. Herberts P, Malchau H. How outcome studies have changed total
hip arthroplasty practices in Sweden. Clin Orthop Relat Res.
1997;344:44–60.
8. Johnsen SP, Sorensen HT, Lucht U, Soballe K, Overgaard S,
Pedersen AB. Patient-related predictors of implant failure after
123
Clinical Orthopaedics and Related Research
primary total hip replacement in the initial, short- and long-terms.
A nationwide Danish follow-up study including 36,984 patients.
J Bone Joint Surg Br. 2006;88:1303–1308.
9. Krahn TH, Canale ST, Beatty JH, Warner WC, Lourenço P. Long
term follow-up of patients with avascular necrosis after treatment
of slipped capital femoral epiphysis. J Pediatr Orthop.
1993;13:154–158.
10. LeCroy CM, Rizzo M, Gunneson EE, Urbaniak JR. Free vascu-
larized fibular bone grafting in the management of femoral neck
nonunion in patients younger than fifty years. J Orthop Trauma.
2002;16:464–472.
11. Manley MT, Capello WN, D’Antonio JA, Edidin AA, Geesink
RG. Fixation of acetabular cups without cement in total hip
arthroplasty. A comparision of three different implant surfaces at
a minimum duration of follow-up of five years. J Bone Joint Surg
Am. 1998;80:1175–1185.
12. Marciniak D, Furey C, Shaffer JW. Osteonecrosis of the femoral
head. A study of 101 hips treated with vascularized fibular
grafting. J Bone Joint Surg Am. 2005;87:742–747.
13. Merle D’Aubigne R, Postel M, Mazabraud A, Massias P,
Gueguen J, France P. Idiopathic necrosis of the femoral head in
adults. J Bone Joint Surg Br. 1965;47:612–633.
14. Montella BJ, Nunley JA, Urbaniak JR. Osteonecrosis of the
femoral head associated with pregnancy. A preliminary report.
J Bone Joint Surg Am. 1999;81:790–798.
15. Prudhommeaux F, Hamadouche M, Nevelos J, Doyle C,
Meunier A, Sedel L. Wear of alumina-on-alumina total hip
arthroplasties at a mean 11-year followup. Clin Orthop Relat
Res. 2000;379:113–122.
16. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system
for staging avascular necrosis. J Bone Joint Surg Br. 1995;77:34–
41.
17. Stubbs AJ, Gunneson EE, Urbaniak JR. Pediatric femoral avas-
cular necrosis after pyarthrosis. Use of free vascularized fibular
grafting. Clin Orthop Relat Res. 2005;439:193–200.
18. Tooke SM, Favero KJ. Femoral neck fractures in skeletally
mature patients, fifty years old or less. J Bone Joint Surg Am.
1985;67:1255–1260.
19. Vail TP, Urbaniak JR. Donor-site morbidity with use of vascu-
larized autogenous fibular grafts. J Bone Joint Surg
Am.1996;78:204–211.
20. Veillette CJ, Mehdian H, Schemitsch EH, McKee MD. Survi-
vorship analysis and radiographic outcome following tantalum
rod insertion for osteonecrosis of the femoral head. J Bone Joint
Surg Am. 2006;88:48–55.
21. Zhang C, Zeng B, Xu Z, Song W, Shao L, Jing D, Sui S.
Treatment of femoral head necrosis with free vascularized fibular
grafting: a preliminary report. Microsurgery. 2000;25:305–309.
Clin Orthop Relat Res (2008) 466:1125–1132
DOI 10.1007/s11999-008-0211-x
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Nonvascularized Bone Grafting Defers Joint Arthroplasty in Hip
Osteonecrosis
Thorsten M. Seyler MD, David R. Marker BS,
Slif D. Ulrich MD, Tobias Fatscher BS,
Michael A. Mont MD
Published online: 20 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract A variety of nonvascularized bone grafting
techniques have been proposed with varying degrees of
success as treatment alternatives for osteonecrosis of the
femoral head. The success of these procedures may be
enhanced using ancillary growth and differentiation fac-
tors. We retrospectively reviewed 33 patients (39 hips)
with osteonecrosis of the hip who had nonvascularized
bone grafting procedures with supplemental OP-1. We
compared the outcomes in this cohort to similar patients
treated nonoperatively or with other nonvascularized bone
grafting procedures. We used a trapdoor to make a window
at the head-neck junction to remove necrotic bone and
packed the excavated area with autogenous cancellous
bone graft, marrow, and OP-1. The minimum followup was
24 months (mean, 36 months; range, 24–50 months). We
Each author certifies that he or she has no commercial associations
(e.g., consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc.) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
Electronic supplementary material The online version of this
article (doi:10.1007/s11999-008-0211-x) contains supplementary
material, which is available to authorized users.
T. M. Seyler, D. R. Marker, S. D. Ulrich, T. Fatscher,
M. A. Mont (&)
Rubin Institute for Advanced Orthopedics, Center for Joint
Preservation and Reconstruction, Sinai Hospital of Baltimore,
2401 West Belvedere Avenue, Baltimore, MD 21215, USA
e-mail: mmont@lifebridgehealth.org; Rhondamont@aol.com
T. M. Seyler
Department of Orthopaedic Surgery, Wake Forest University,
Winston-Salem, NC, USA
performed no further surgery in 25 of 30 small- and
medium-sized lesions (80%) but did in two of nine large
lesions. Hips with Ficat Stage II disease were not reoper-
ated in 18 of 22 cases during the followup periods. Our
short-term results compare similarly to nonoperative
treatment and other reports of nonvascularized bone
grafting. With the addition of ancillary growth factors,
these procedures effectively reduce donor site morbidity
and may defer joint arthroplasty in selected patients.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis of the femoral head is a devastating disease
that often leads to destruction of the hip and the need for
total hip arthroplasty [32, 35]. Annual reports from various
joint registries such as the Canadian Joint Replacement
Registry, the Australian National Joint Replacement Reg-
istry, and the Swedish Hip Arthroplasty Register have
demonstrated that the diagnosis of osteonecrosis accounts
for between 2.8% to 6% of all primary total hip arthro-
plasties performed [1, 7, 60]. In early stages of the disease,
head-preserving treatment modalities such as core decom-
pression, osteotomy, and vascularized or nonvascularized
bone grafting are often utilized to defer head-replacing
options such as total hip arthroplasty [32, 35].
The rationale for the use of nonvascularized bone grafting
is to remove necrotic bone and replace it with cancellous and
cortical autografts that support the subchondral bone and
articular cartilage of the femoral head and may stimulate
bone formation [11, 43]. Three different surgical techniques
have been popularized for nonvascularized bone grafting:
123
1126 Seyler et al.
Fig. 1 The five key steps for vas-
cularized bone grafting through a
window made in the femoral neck
or femoral head-neck junction are
illustrated.
(1) grafting through a core decompression tract (Phemister
technique) (Appendix 1 -Supplemental Website Materials;
supplemental materials are available with the online version
of CORR ) [2–4, 24, 43, 55]; (2) grafting through a window
or trapdoor in the articular cartilage (Appendix 2 - Supple-
mental Website Materials; supplemental materials are
available with the online version of CORR) [21, 28–30]; and
(3) grafting through a window made in the femoral neck or
femoral head-neck junction (Fig. 1) [31, 48]. Each of these
techniques has its advantages and its limitations. While
earlier studies of nonvascularized bone grafting through a
core tract or cartilage window reported promising clinical
results [2, 6, 24, 29, 48], later studies using this technique
reported less favorable outcomes [9, 39].
The use of OP-1 (BMP 7) in combination with allograft
has been applied in various bone healing applications
(nonunions, trauma, spine fusion) [36]. For osteonecrosis,
it was used in a canine model in which defects treated with
OP-1 and bone grafting healed faster radiographically than
defects simply treated with bone grafting [34]. This study
provided a rationale for the possible use of OP-1 in com-
bination with allograft to heal human osteonecrotic defects.
We describe the principles, indications, and surgical
techniques for nonvascularized bone grafting through a
window made at the femoral head-neck junction. We asked
whether this technique effectively and similarly deferred
further surgical treatment options when compared to those
reported in studies using nonoperative treatment. In addi-
tion, we questioned how these outcomes compared to other
studies of nonvascularized bone grafting.
Materials and Methods
We retrospectively reviewed 33 patients (39 hips) with
osteonecrosis of the femoral head who had nonvascular-
ized bone grafting procedures with supplemental OP-1
performed consecutively for the appropriate indications
123
Clinical Orthopaedics and Related Research
between December 1, 2002, and January 1, 2004. Indi-
cations for the procedure were Ficat and Arlet Stage II or
III lesions that met various intraoperative criteria (descri-
bed later). There were 15 women (16 hips) and 18 men
(23 hips) who had a mean age of 35 years (range, 18–
52 years). The mean body mass index was 27.2 kg/m2
(range, 19.4–36.0 kg/m2). No patients were lost to fol-
lowup. Minimum followup was 24 months (mean,
36 months; range, 24–50 months). After obtaining institu-
tional review board approval, a prospective database was
used to collect relevant surgical, clinical, and radiographic
data.
We identified the following risk factors and associated
conditions with osteonecrosis of the femoral head: corti-
costeroid usage (defined as a dose greater than 2 g
prednisone or its equivalent per month for 3 months min-
imum [42]) in 9 patients (12 hips), alcohol abuse (defined
as alcohol consumption of more than 400 mL per week
[25]) in 8 patients (eight hips), systemic lupus erythemat-
osus in 6 patients (seven hips), tobacco abuse (defined as
20 cigarettes or more per day [25]) in 3 patients (four hips),
hepatitis C in 2 patients (three hips), and HIV infection in 2
patients (two hips). Of the remaining 4 patients (four hips),
one each had an underlying diagnosis of ulcerative colitis,
sickle cell disease, high levels of plasminogen activator
inhibitor with hypofibrinolysis, and chronic obstructive
pulmonary disease. Three patients (four hips) had no
apparent associated risk factors and were deemed idio-
pathic osteonecrosis. Some patients had more than one
associated risk factor.
We (TMS, SDU) assessed patients using the Harris hip
rating system [14]. We defined failure as patients who
underwent total hip arthroplasty surgery.
Anteroposterior and lateral radiographs were made
preoperatively and postoperatively at 6 weeks, 3 months,
6 months, 1 year, and annually thereafter. We determined
Ficat and Arlet stage [10], combined Kerboul angle [12],
presence or absence of new bone formation, location of the
Volume 466, Number 5, May 2008
lesion, and disease progression. Preoperative radiographs
were evaluated by two of us (TMS, SDU) to determine the
staging according to the system by Ficat and Arlet. The
size of the lesions was measured using the combined
necrotic angle technique described by Kerboul et al. [19].
The combined angle is derived from evaluating antero-
posterior and lateral radiographs by adding the sums of the
angle of the lesions delineated on each view. In vague
cases in which the lesion was not clearly demarcated on
plain radiographs, MRI and computed tomography evalu-
ations were used to assist in the evaluation of lesion size.
Using this method, the extent of the necrosis was stratified
into three categories: (1) large lesions, when the combined
necrotic angle was 200° or greater; (b) medium lesions,
when the angle was between 150° and 200°; and (c) small
lesions, when the angle was 150° or less. The location of
the lesion was defined using a system initially described by
Ohzono et al. [41]. Lesions were classified as type A, B,
C1, or C2. A Type A lesion was one that occupied the
medial third or less of the weight-bearing portion. A Type
B lesion occupied the medial two-thirds or less of the
weight-bearing portion. A Type C lesion occupied more
than the medial two-thirds of the weight-bearing portion.
The subtypes C1 and C2 were used to further stratify the C
type lesions with a Type C2 lesion extending laterally to
the acetabular edge, whereas a Type C1 lesion did not.
Because of the possible introduction of error assessing
radiographic measurements, an evaluation of interobserver
and intraobserver error in radiographic assessment was
performed by two of us (TMS, SDU) before reviewing
study-related radiographs. The intraobserver agreement
was 100% in the 10 pilot cases and the interobserver
agreement was an exact match in 90% of the pilot cases. To
guarantee objectivity and avoid the problem of intraob-
server and interobserver variability in assessing the various
radiographic parameters, two of the authors (TMS, SDU)
independently evaluated the radiographs 2 weeks apart. If
there was a disagreement, the senior author (MAM) inter-
preted the films until a unanimous decision could be made
regarding the best estimate at staging, size, or extent of
collapse of lesion evaluation. The various radiographic
variables (Ficat and Arlet stage and Kerboul angle) were
assessed to see whether they had any prognostic value. Of
the 39 hips, 22 hips were classified as Ficat and Arlet Stage
II and 17 hips were classified as Ficat and Arlet Stage III
preoperatively. The assessment of lesions size using the
Kerboul technique revealed seven small lesions, 23 med-
ium lesions, and nine large lesions.
All procedures were performed by the senior author
(MAM) using a trapdoor made at the femoral head-neck
junction (Fig. 1). Large lesions were not considered a
contraindication for the present patient cohort. The tech-
nique [48, 49] was performed using the anterolateral
Nonvascularized Bone Grafting in Hip ON 1127
approach (Watson-Jones [62]) with the patient lying in the
lateral decubitus position. The skin incision was started just
distal to the anterosuperior iliac spine and carried out to a
point just posterior to the greater trochanter. The incision
was then angled at about 110° anteriorly and extended
distally to parallel the femoral shaft for 8 to 10 cm. In the
next step, the interval between the tensor fascia latae
muscle and the gluteus medius muscle was identified by
dividing the gluteus maximus fascia and the fascia latae.
The dissection was carefully extended proximally to
expose but protect the superior gluteal nerve. The fascia
latae was then split in the direction of the skin incision and
the anterior 40% of the gluteus medius was detached and
retracted posteriorly. The gluteus minimus muscle was
detached fully revealing the capsule with the head of the
rectus femoris muscle attached to the upper part of ace-
tabular rim. The capsule was then excised with the labrum
left intact and the capsule peeled anteriorly to preserve the
medial circumflex artery and its branches posteriorly. This
approach allowed for preservation of the blood supply to
the femoral head. We inspected the femoral head cartilage
in situ by rotating and distracting the femur without dis-
locating the femoral head. The femoral head cartilage was
then inspected to ascertain whether there were any full-
thickness defects or areas of delaminated cartilage. We
considered a defect of 1 cm or greater, cartilage delami-
nation, or erosive areas as contraindications for performing
this bone grafting procedure. This occurred in five cases
during the time period of the study and these patients
received a total hip arthroplasty. An approximate 2-cm 9
2-cm window was then made at the femur head-neck
junction (trapdoor) using a microoscillating saw and os-
teotomes. The window segment was preserved and stored
in normal saline-wrapped gauze for replacement at the end
of the procedure. A 6-mm mushroom-tipped burr was used
to débride necrotic bone in the femoral head using the
trapdoor as an entrance point. If 70% or more of the
femoral head was involved with the disease (necrotic
bone), the procedure was abandoned and a hip arthroplasty
was performed. Accidental head penetration with the burr
was avoided. The cavity was filled with cancellous bone
chips and bone marrow. In addition, recombinant human
bone morphogenetic protein 7 was added to promote new
bone formation. Each sterile unit of implant contained
3.5 mg OP-1 (purchased from Stryker Biotech, Hopkinton,
MA) mixed with 1 g Type I bovine bone-derived collagen.
The material was tightly packed into the cavity with a
layered approach and the saved bony window segment was
put back and fixed with three, 2-mm poly-p-dioxanone
resorbable pins (Orthosorb1, Johnson and Johnson, New
Brunswick, NJ). Finally, the hip was relocated, and the
gluteus minimus muscle was reattached to bone and the
gluteus medius muscle and fascia latae were repaired with
123
1128 Seyler et al. Clinical Orthopaedics and Related Research

interrupted sutures. The procedure had a mean operative Table 1. Correlation between lesion size, location, Ficat and Arlet
time of 62 minutes (range, 37–102 minutes). stage, and incidence of collapse
All patients were maintained at toe-touch weightbearing Number Number Incidence
with two crutches or a walker for 5 to 6 weeks. For the next of hips collapsed of collapse
5 to 6 weeks, patients were advanced to approximately
Lesion Size
50% weightbearing using a cane or crutch in the opposite
Small 7 1 13%
hand. Patients were then advised to start full weightbearing
Medium 23 5 17%
as tolerated at 10 weeks postoperatively. Participation in
Large 9 7 78%
sports and higher impact loading activities such as running
Location of Lesion
were not recommended for the first 10 months
A 8 1 13%
postoperatively.
B 12 5 42%
Survival was defined by whether the patient had sub-
C1 12 2 17%
sequent surgery on the hip.
C2 7 5 71%
To assess how the results of the procedures for our
cohort compared other nonvascularized bone grafting Ficat and Arlet Stage
procedures in similarly aged patients at a similar length of Stage I 0 NA NA
followup, the authors carried out an extensive literature Stage II 22 4 18%
review of the databases of the National Library of Medi- Stage III 17 9 53%
cine, the National Institutes of Health, and EMBASE. We Stage IV 0 NA NA
identified all articles concerning nonvascularized bone NA = not applicable.
grafting for osteonecrosis of the femoral head. The key
words used in the search were ‘‘hip,’’ ‘‘femoral head,’’
‘‘osteonecrosis,’’ ‘‘avascular necrosis,’’ and ‘‘necrosis.’’ The mean preoperative Harris hip scores for all patients
The initial search was refined with the addition of the in this series was 50 points (range, 28–76 points). The
keywords ‘‘core decompression,’’ ‘‘bone grafting,’’ ‘‘non- preoperative scores for the hips that subsequently failed
vascularized,’’ ‘‘trapdoor,’’ and ‘‘lightbulb’’ [49]. All (mean, 47 points; range, 28–72 points) were similar
articles identified in this manner were then subject to a (p = 0.175) to those of the survival group (mean, 52 points;
review by two of us (TMS, DRM, MAM, or TF). The range, 28–76 points). At a mean followup of 35 months, the
search revealed 26 published studies. A similar review was mean postoperative score for the entire series improved to
conducted to identify reports of nonoperative treatment for 75 points (range, 27–100 points) (p \ 0.001). There were
osteonecrosis of the head. This search revealed 11 reports. no perioperative complications documented.
For both the nonoperative and nonvascularized bone Medically, one patient had a urinary tract infection
grafting literature reviews we collected the following data: which resolved without any sequelae. There were no other
failure rates (in terms of later receiving a total hip arthro- medical complications.
plasty), surgical technique, bone grafting procedure, and The overall early clinical success (defined as not later
demographic variables. undergoing total hip arthroplasty) rate of 67% (26 of 39
hips) for this procedure as well as the 80% (24 of 30 hips)
success rate for small and medium sized lesions compared
Results similarly to other nonvascularized procedures performed at
similar mean followup (range, 28–144 months) (Table 3).
Overall, 26 of the hips survived out of the 39 hips treated We have also provided results of nonoperative studies for
(67%). At most recent followup, 24 of the 30 hips (80%) comparison (Table 4).
with small- or medium-sized lesions had avoided further
surgery. Patients with large lesions fared poorly with only
two of nine hips avoiding further surgery. When stratified Discussion
by Ficat and Arlet stage, 18 of the 22 hips with Stage II
disease did not undergo further surgery. Stage III hips were Nonvascularized bone grafting techniques for the treatment
less successful with only eight of 17 hips surviving. There of osteonecrosis of the femoral head were popularized in
were similar results when analyzing location of lesion, with the 1950s and 1960s [3, 4, 43]. The literature reports a wide
more lateral lesions faring more poorly than centrally range of success rates with these techniques and this may
located lesions (Table 1). Failures (n = 13) had a mean be a result of the various surgical techniques and/or reflect
time to femoral head collapse of 13 months (range, the problem of choosing the appropriate treatment modality
2–34 months) (Table 2). for the various disease stages. We evaluated our recent

123
Volume 466, Number 5, May 2008 Nonvascularized Bone Grafting in Hip ON 1129

Table 2. Characteristics of clinical failures

t Age Gender Risk factors Ficat and Lesion size Location Time to failure
Arlet stage (Kerboul) of lesion (months)

1 36 Male hyperlipidemia, smoking II Medium B 16

2 22 Male alcohol III Large B 31
3 51 Female SLE, corticosteroids III Large C1 12
4 27 Female SLE, corticosteroids II Medium B 8
5 37 Male HIV, hepatitis C III Large C2 6
6 30 Female SLE, corticosteroids II Small B 34
7 36 Male alcohol, smoking III Large C2 2
8 44 Female SLE, corticosteroids III Large B 16
9 52 Male HIV, hepatitis C III Large C2 8
10 31 Female SLE, corticosteroids II Medium A 24
11 55 Female SLE, corticosteroids II Medium C2 10
12 29 Female alcohol, smoking III Large C1 7
13 41 Male corticosteroids II Medium C2 9
SLE = systemic lupus erythematosus.

experience with nonvascularized bone grafting. The pri-
mary questions were whether this technique effectively
deferred further surgical treatment when compared to those
reported in studies using nonoperative treatment. In addi-
tion, we questioned whether the outcomes in this study
were comparable to other studies of nonvascularized bone
grafts.
Our study has several shortcomings including the small
number of patients and the short-term followup. Never-
theless, the early results encourage the continued use and
further study of this procedure. A larger series with longer
followup will further help assess positive and negative
predictors of outcome.
Several authors have described results comparable to
ours using variations of these nonvascularized bone graft-
ing procedures. Saito et al. [51] reported various treatment
modalities for idiopathic necrosis of the femoral head.
Their series included 18 hips with Ficat and Arlet Stage II
osteonecrosis treated with a similar technique of nonvas-
cularized bone grafting using cancellous bone obtained
from the ipsilateral iliac crest. At a minimum followup of
24 months (mean, 48 months; range, 24–144 months), the
clinical evaluation revealed Merle D’Aubigné [27] scores
of 15 or more points in 13 of 18 hips. However, radio-
graphic results demonstrated less favorable results, with
seven of the 18 hips showing progressive femoral head
collapse. We included both Ficat and Arlet Stage II and III
hips, which may have contributed to the slightly lower
chance of having a Harris hip score above 70.
The percentage of hips in our cohort of patients with
nonvascularized bone grafting patients whom we consid-
ered had success treatment (67%) was similar to that in
other reports in the literature (Table 3). The clinical
success of the lightbulb technique ranged from 68% to 87%
compared to a range of 36% to 90% reports for the
Phemister technique. Similarly, the clinical success of the
trapdoor technique ranged from 71% to 89%.
The proportion of nonvascularized bone grafting
patients in our cohort who underwent total hip replacement
(67%) was lower than six of the 11 studies that reported the
outcomes of patients who were treated nonoperatively. The
success (defined as not having total hip replacement by
final followup) in studies from 1986 to 2007 of nonoper-
ative treatment ranged from 9% to 86% (Table 4).
Other authors combined this technique with intertro-
chanteric osteotomy, use of growth factors, or gluteus
medius muscle pedicle bone graft [31, 48, 52]. Scher and
Jakim [52] prospectively studied 45 hips with Ficat and
Arlet Stage III osteonecrosis treated with intertrochanteric
osteotomy and nonvascularized bone grafting through a
window in the femoral neck. The 5-year survival rate was
87%. This encouraging survival rate, however, should be
critically evaluated because of the stringent inclusion cri-
teria that were employed. The study included only patients
younger than 45 years of age, with Ficat and Arlet Stage III
of the anterosuperior part of the femoral head, with no
underlying metabolic bone disease or systemic condition
treated with chemotherapy or corticosteroids, and with no
extensive involvement of the posterior part of the femoral
head. Rosenwasser et al. [48] reported the long-term results
of their series using the lightbulb technique. At a minimum
followup of 120 months (mean, 144 months; range, 120–
180 months), the survival rate was 87% with minimal
disease progression. In three patients, the authors used a
gluteus medius muscle pedicle graft to augment blood
supply to the femoral neck. Mont et al. [31] reported on a

123
1130 Seyler et al. Clinical Orthopaedics and Related Research

Table 3. Literature review of nonvascularized bone grafting techniques

Study Year Hips Followup (months) Clinical Radiographic
success (%) success (%)

Trapdoor technique
Meyers et al. [29] 1983 21 NA 71 NA
Meyers et al. [29] 1991 9 35 (12–107) 89 NA
Ko et al.* [21] 1995 14 53 (24–108) 85 70
Mont et al. [33] 1998 30 56 (30–60) 73 73
Phemister technique
Bonfiglio and Voke [4] 1968 116 67 (24–204) 78 NA
Boettcher et al. [2] 1970 38 72 (24–204) 79 76
Marcus et al. [24] 1973 11 N/A (24–48) 90 91
Dunn and Grow [9] 1977 23 40 (27–98) 74 30
McBeath and Oeljen [26] 1977 6 NA 83 0
Smith et al. [55] 1980 56 144 (24–332) 57 NA
Steinberg et al. [57] 1984 19 [6 82 36
Buckley et al. [6] 1991 20 96 (24–228) 90 90
Nelson and Clark [39] 1993 52 NA (24–144) 77 13
Steinberg et al. [58] 2001 312 63 (23–146) 64 61
Mont et al. [31] 2003 21 48 (36–55) 86 76
Plakseychuk et al. [45] 2003 50 60 (36–96) 36 28
Rijnen et al. [46] 2003 28 50 (24–119) 71 57
Lieberman et al. [23] 2004 17 53 (26–94) 82 82
Kim et al. [20] 2005 30 50 (36–67) 78 80
Israelite et al. [16] 2005 276 N/A (24–145) 62 NA
Wang et al. [61] 2005 28 26 (24–39) 68 64
Keizer et al. [18] 2006 80 84 (36–NA) 46 43
Lightbulb technique
Saito et al. [51] 1988 18 48 (24–168) 72 61
Scher and Jakim* [53] 1993 45 65 (36–126) 87 71
Rosenwasser et al. [48] 1994 15 138 (108–180) 86 86
Mont et al. [31] 2003 21 48 (36–55) 86 NA
Our study 2007 47 28 (12–50) 68 64
*
NA = data not available; combined with osteotomy.

series of 19 patients (21 hips) treated with bone morpho-

Table 4. Literature review of nonoperative treatment outcomes genetic protein-enriched allograft to avoid donor site
Musso et al./1986 [38] 50 30 32 morbidity. At a minimum followup of 36 months (mean,
48 months; range, 36–55 months), three hips had failed the
Steinberg et al./1989 [56] 55 21 (6–120) 16 bone grafting procedure. Interestingly, all failures occurred
Churchill and Spencer/1991 [8] 18 60 50 in hips with large-sized lesions, suggesting lesion size was
Stulberg et al./1991 [59] 22 27 9 associated with failure.
Robinson and Springer/1992 [47] 16 39 (24–61) 56 Despite the limitations of the study, we are encouraged
Bradway and Morrey/1993 [5] 15 23 (3–66) 13 by these early results using cancellous bone chips, bone
Jergesen and Khan/1997 [17] 19 111 (51–81) 42 marrow, and bone morphogenetic protein-7 as a nonvas-
Lai et al./2005 [22] 25 24 32 cularized bone grafting technique for the treatment of Stage
Hernigou et al./2006 [15] 121 168 (120–240) 25 II and III osteonecrosis of the femoral head. The decreased
Neumayr et al./2006 [40] 21 36 86 progression of symptoms at a mean of 36 months suggests
Morse et al./2007 [37] 67 23 (17–31) 70 the natural progression of the disease and subsequent hip
*Defined as not requiring conversion to total hip arthroplasty by final arthroplasty surgery has been delayed. This technique is
followup. straightforward, has low donor site morbidity, and

123
Volume 466, Number 5, May 2008
demonstrates a high degree of efficacy for Stage II and
small to medium sized lesions.
Acknowledgments We thank Colleen Kazmarek for her assistance
in the preparation of this manuscript.
References
1. Australian Orthopaedic Association Web site. The National
Joint Replacement Registry: Annual Report. Available at: http://
www.dmac.adelaide.edu.au/aoanjrr/documents/aoanjrr_2006_
supplementary.pdf. Accessed March 15, 2007.
2. Boettcher WG, Bonfiglio M, Smith K. Non-traumatic necrosis of
the femoral head. II. Experiences in treatment. J Bone Joint Surg
Am. 1970;52:322–329.
3. Bonfiglio M, Bardenstein MB. Treatment by bone-grafting of
aseptic necrosis of the femoral head and non-union of the femoral
neck (Phemister technique). J Bone Joint Surg Am. 1958;40:
1329–1346.
4. Bonfiglio M, Voke EM. Aseptic necrosis of the femoral head and
non-union of the femoral neck. Effect of treatment by drilling and
bone-grafting (Phemister technique). J Bone Joint Surg Am.
1968;50:48–66.
5. Bradway JK, Morrey BF. The natural history of the silent hip in
bilateral atraumatic osteonecrosis. J Arthroplasty. 1993;8:383–387.
6. Buckley PD, Gearen PF, Petty RW. Structural bone-grafting for
early atraumatic avascular necrosis of the femoral head. J Bone
Joint Surg Am. 1991;73:1357–1364.
7. Canadian Institute for Health Information Web site. Canadian
Joint Replacement Registry: Annual Report. Available at:
http://secure.cihi.ca/cihiweb/dispPage.jsp?cw_page = services_
cjrr_e#publications. Accessed April 1, 2007.
8. Churchill MA, Spencer JD. End-stage avascular necrosis of bone
in renal transplant patients. The natural history. J Bone Joint Surg
Br. 1991;73:618–620.
9. Dunn AW, Grow T. Aseptic necrosis of the femoral head.
Treatment with bone grafts of doubtful value. Clin Orthop Relat
Res. 1977;249–254.
10. Ficat RP. Idiopathic bone necrosis of the femoral head. Early
diagnosis and treatment. J Bone Joint Surg Br. 1985;67:3–9.
11. Ganz R, Buchler U. Overview of attempts to revitalize the dead
head in aseptic necrosis of the femoral head–osteotomy and
revascularization. Hip. 1983;296–305.
12. Gibson A. Posterior exposure of the hip joint. J Bone Joint Surg
Br. 1950;32:183–186.
13. Hardinge K. The direct lateral approach to the hip. J Bone Joint
Surg Br. 1982;64:17–19.
14. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
15. Hernigou P, Habibi A, Bachir D, Galacteros F. The natural history
of asymptomatic osteonecrosis of the femoral head in adults with
sickle cell disease. J Bone Joint Surg Am. 2006;88:2565–2572.
16. Israelite C, Nelson CL, Ziarani CF, Abboud JA, Landa J,
Steinberg ME. Bilateral core decompression for osteonecrosis of
the femoral head. Clin Orthop Relat Res. 2005;441:285–290.
17. Jergesen HE, Khan AS. The natural history of untreated asymp-
tomatic hips in patients who have non-traumatic osteonecrosis.
J Bone Joint Surg Am. 1997;79:359–363.
18. Keizer SB, Kock NB, Dijkstra PD, Taminiau AH, Nelissen RG.
Treatment of avascular necrosis of the hip by a non-vascularised
cortical graft. J Bone Joint Surg Br. 2006;88:460–466.
Nonvascularized Bone Grafting in Hip ON 1131
19. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. The
conservative surgical treatment of idiopathic aseptic necrosis of
the femoral head. J Bone Joint Surg Br. 1974;56:291–296.
20. Kim SY, Kim YG, Kim PT, Ihn JC, Cho BC, Koo KH. Vascu-
larized compared with nonvascularized fibular grafts for large
osteonecrotic lesions of the femoral head. J Bone Joint Surg Am.
2005;87:2012–2018.
21. Ko JY, Meyers MH, Wenger DR. ‘‘Trapdoor’’ procedure for
osteonecrosis with segmental collapse of the femoral head in
teenagers. J Pediatr Orthop. 1995;15:7–15.
22. Lai KA, Shen WJ, Yang CY, Shao CJ, Hsu JT, Lin RM. The use
of alendronate to prevent early collapse of the femoral head in
patients with nontraumatic osteonecrosis. A randomized clinical
study. J Bone Joint Surg Am. 2005;87:2155–2159.
23. Lieberman JR, Conduah A, Urist MR. Treatment of osteone-
crosis of the femoral head with core decompression and human
bone morphogenetic protein. Clin Orthop Relat Res. 2004;429:
139–145.
24. Marcus ND, Enneking WF, Massam RA. The silent hip in idio-
pathic aseptic necrosis. Treatment by bone-grafting. J Bone Joint
Surg Am. 1973;55:1351–1366.
25. Matsuo K, Hirohata T, Sugioka Y, Ikeda M, Fukuda A. Influence
of alcohol intake, cigarette smoking, and occupational status on
idiopathic osteonecrosis of the femoral head. Clin Orthop Relat
Res. 1988;234:115–123.
26. McBeath AA, Oeljen CG. Phemister bone graft for osteonecrosis
post renal transplant. Clin Orthop Relat Res. 1977:164–168.
27. Merle D’Aubigne R. [Numerical classification of the function of
the hip. 1970]. Rev Chir Orthop Reparatrice Appar Mot.
1990;76:371–374.
28. Merle D’Aubigne R, Postel M, Mazabraud A, Massias P,
Gueguen J, France P. Idiopathic necrosis of the femoral head in
adults. Journal of Bone and Joint Surgery Br. 1965;47:612–633.
29. Meyers MH, Jones RE, Bucholz RW, Wenger DR. Fresh
autogenous grafts and osteochondral allografts for the treatment
of segmental collapse in osteonecrosis of the hip. Clin Orthop
Relat Res. 1983;174:107–112.
30. Mont MA, Einhorn TA, Sponseller PD, Hungerford DS. The
trapdoor procedure using autogenous cortical and cancellous
bone grafts for osteonecrosis of the femoral head. J Bone Joint
Surg Br. 1998;80:56–62.
31. Mont MA, Etienne G, Ragland PS. Outcome of nonvascularized
bone grafting for osteonecrosis of the femoral head. Clin Orthop
Relat Res. 2003;417:84–92.
32. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of
the femoral head. J Bone Joint Surg Am. 1995;77:459–474.
33. Mont MA, Jones LC, Einhorn TA, Hungerford DS, Reddi AH.
Osteonecrosis of the femoral head. Potential treatment with
growth and differentiation factors. Clin Orthop Relat Res.
1998;355(Suppl):S314–335.
34. Mont MA, Jones LC, Elias JJ, Inoue N, Yoon TR, Chao EY,
Hungerford DS. Strut-autografting with and without osteogenic
protein-1 : a preliminary study of a canine femoral head defect
model. J Bone Joint Surg Am. 2001;83:1013–1022.
35. Mont MA, Jones LC, Hungerford DS. Nontraumatic osteone-
crosis of the femoral head: ten years later. J Bone Joint Surg Am.
2006;88:1117–1132.
36. Mont MA, Ragland PS, Biggins B, Friedlaender G, Patel T, Cook
S, Etienne G, Shimmin A, Kildey R, Rueger DC, Einhorn TA.
Use of bone morphogenetic proteins for musculoskeletal appli-
cations. An overview. J Bone Joint Surg Am. 2004;86(Suppl
2):41–55.
37. Morse CG, Mican JM, Jones EC, Joe GO, Rick ME, Formentini
E, Kovacs JA. The incidence and natural history of osteonecrosis
in HIV-infected adults. Clin Infect Dis. 2007;44:739–748.
123
1132 Seyler et al.
38. Musso ES, Mitchell SN, Schink-Ascani M, Bassett CA. Results of
conservative management of osteonecrosis of the femoral head. A
retrospective review. Clin Orthop Relat Res. 1986;207:209–215.
39. Nelson LM, Clark CR. Efficacy of phemister bone grafting in
nontraumatic aseptic necrosis of the femoral head. J Arthroplasty.
1993;8:253–258.
40. Neumayr LD, Aguilar C, Earles AN, Jergesen HE, Haberkern
CM, Kammen BF, Nancarrow PA, Padua E, Milet M, Stulberg
BN, Williams RA, Orringer EP, Graber N, Robertson SM,
Vichinsky EP. Physical therapy alone compared with core
decompression and physical therapy for femoral head osteone-
crosis in sickle cell disease. Results of a multicenter study at a
mean of three years after treatment. J Bone Joint Surg Am.
2006;88:2573–2582.
41. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
42. Ono K, Tohjima T, Komazawa T. Risk factors of avascular
necrosis of the femoral head in patients with systemic lupus
erythematosus under high-dose corticosteroid therapy. Clin
Orthop Relat Res. 1992;277:89–97.
43. Phemister DB. Treatment of the necrotic head of the femur in
adults. J Bone Joint Surg Am. 1949;31:55–66.
44. Phemister DB. Repair of bone in the presence of aseptic necrosis
resulting from fractures, transplantations, and vascular obstruc-
tion. 1930. J Bone Joint Surg Am. 2005;87:672.
45. Plakseychuk AY, Kim SY, Park BC, Varitimidis SE, Rubash HE,
Sotereanos DG. Vascularized compared with nonvascularized
fibular grafting for the treatment of osteonecrosis of the femoral
head. J Bone Joint Surg Am. 2003;85:589–596.
46. Rijnen WH, Gardeniers JW, Buma P, Yamano K, Slooff TJ,
Schreurs BW. Treatment of femoral head osteonecrosis using
bone impaction grafting. Clin Orthop Relat Res. 2003;417:74–83.
47. Robinson Jr. HJ, Springer JA. Success of core decompression in
the management of early stages of avascular necrosis: A four year
prospective study. Orthop Trans. 1993;16:707.
48. Rosenwasser MP, Garino JP, Kiernan HA, Michelsen CB. Long
term followup of thorough debridement and cancellous bone
grafting of the femoral head for avascular necrosis. Clin Orthop
Relat Res. 1994;306:17–27.
49. Rosenwasser MP, Michelsen CB, Kiernan HA. Treatment of
avascular necrosis of the femoral head with curettage, iliac
cacellous bone grafting, and revascularization with gluteus
muscle pedicle flap. Orthop Trans. 1983;7:396.
123
Clinical Orthopaedics and Related Research
50. Rue JP, Inoue N, Mont MA. Current overview of neurovascular
structures in hip arthroplasty: anatomy, preoperative evaluation,
approaches, and operative techniques to avoid complications.
Orthopedics. 2004;27:73–81; quiz 82–73.
51. Saito S, Ohzono K, Ono K. Joint-preserving operations for idi-
opathic avascular necrosis of the femoral head. Results of core
decompression, grafting and osteotomy. J Bone Joint Surg Br.
1988;70:78–84.
52. Scher MA, Jakim I. Intertrochanteric osteotomy and autogenous
bone-grafting for avascular necrosis of the femoral head. J Bone
Joint Surg Am. 1993;75:1119–1133.
53. Scher MA, Jakim I. Late follow-up of femoral head avascular
necrosis managed by intertrochanteric osteotomy & bone graft-
ing. Acta Orthop Belg. 1999;65(Suppl 1):73–77.
54. Smith-Petersen MN. Evolution of mould arthroplasty of the hip
joint. J Bone Joint Surg Br. 1948;30:59–75.
55. Smith KR, Bonfiglio M, Montgomery WJ. Non-traumatic
necrosis of the femoral head treated with tibial bone-grafting. A
follow-up note. J Bone Joint Surg Am. 1980;62:845–847.
56. Steinberg ME, Brighton CT, Corces A, Hayken GD, Steinberg
DR, Strafford B, Tooze SE, Fallon M. Osteonecrosis of the
femoral head. Results of core decompression and grafting with
and without electrical stimulation. Clin Orthop Relat Res.
1989;249:199–208.
57. Steinberg ME, Brighton CT, Steinberg DR, Tooze SE, Hayken
GD. Treatment of avascular necrosis of the femoral head by a
combination of bone grafting, decompression, and electrical
stimulation. Clin Orthop Relat Res. 1984:137–153.
58. Steinberg ME, Larcom PG, Strafford B, Hosick WB, Corces A,
Bands RE, Hartman KE. Core decompression with bone grafting
for osteonecrosis of the femoral head. Clin Orthop Relat Res.
2001;386:71–78.
59. Stulberg BN, Davis AW, Bauer TW, Levine M, Easley K.
Osteonecrosis of the femoral head. A prospective randomized
treatment protocol. Clin Orthop Relat Res. 1991;268:140–151.
60. Swedish Orthopaedic Association Web site.The Swedish Hip
Arthroplasty Register: Annual Report. Available at: http://
www.jru.orthop.gu.se/. Accessed March 15, 2007.
61. Wang CJ, Wang FS, Huang CC, Yang KD, Weng LH, Huang
HY. Treatment for osteonecrosis of the femoral head: comparison
of extracorporeal shock waves with core decompression and
bone-grafting. J Bone Joint Surg Am. 2005;87:2380–2387.
62. Watson-Jones R. Fractures of the neck of the femur. Br J Surg.
1935;23:787–808.
Clin Orthop Relat Res (2008) 466:1133–1140
DOI 10.1007/s11999-008-0204-9
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Long-term Followup of Vascularized Fibular Grafting
for Femoral Head Necrosis
Myung-Chul Yoo MD, PhD, Kang-Il Kim MD, PhD,
Chung-Soo Hahn MD, PhD, Javad Parvizi MD
Published online: 11 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Vascularized fibular grafting has been used for
treatment of osteonecrosis of the femoral head and although
some reports demonstrate successful short- to mid-term
outcomes, long-term results are still unknown. We retro-
spectively reviewed 135 patients (151 hips) who underwent
vascularized fibular grafting for osteonecrosis of the femoral
head. One-hundred and ten patients (124 hips) were followed
for a minimum 10 years (mean, 13.9 years; range, 10–
23.7 years). The mean Harris hip score improved from 72 to
88. At the latest followup, we found improved or unchanged
radiographs in 37 of 59 hips initially Stage II hips and 39 of
65 Stage III hips. Thirteen hips (13 patients) (10.5%) failed
treatment and underwent total hip arthroplasty. The location
and size of the necrotic lesion and the patient’s age influ-
enced long-term survival of the graft. Postoperative
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution either has waived or
does not require approval for the human protocol for this investigation
and that all investigations were conducted in conformity with ethical
principles of research.
M.-C. Yoo (&), K.-I. Kim
Department of Orthopaedic Surgery, Center for Joint Diseases,
East-West Neo Medical Center, Kyung Hee University, 149
Sangil-dong, Gangdong-gu, Seoul 134-727, Korea
e-mail: hipsurgeon@hanmail.net
C.-S. Hahn
Department of Orthopaedic Surgery, Kyung Hee Medical
Center, Kyung Hee University, 1 Haegi-dong, Dongdaeman-gu,
Seoul 130-702, Korea
J. Parvizi
Rothman Institute at Thomas Jefferson University Hospital,
Philadelphia, PA, USA
complications included clawing of the big toe in 17 patients,
partial peroneal nerve palsy in two, and superficial infection
in two. Subtrochanteric fracture occurred in two hips. The
data suggest free vascularized fibular grafting was successful
in maintaining joint function and delaying the need for joint
replacement procedure. Graft survival was associated with
the patient’s age and size and location of the lesion but not
etiology and stages of the disease.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis of the femoral head (ONFH) typically
affects younger patients [23, 43]. Treatment options
include joint preserving procedures such as electrical
stimulation, drilling, core decompression, vascularized or
non-vascularized fibular grafting, and osteotomy; and joint
replacement procedures such as resurfacing and hemi- or
total hip arthroplasty (THA) [1, 2, 3, 11, 18, 28, 32, 34–36,
38, 41–44]. Early diagnosis and appropriate surgery may
reduce the risk of progression and improve the outcome [6,
12, 18, 19, 21, 24, 29]. Among the joint-preserving surgical
procedures, free vascularized fibular grafting (VFG)
reportedly has a survival of 61%–96% at mid-term (4–
7 years) followup [4, 16, 17, 20, 28, 31, 43, 44]. Although
VFG appears successful in the short- to mid-term [4, 16,
17, 19, 20, 28, 31, 43–45], its long-term benefits are not
known. While most authors report their results in relation
to preoperative etiology or collapse stage, the influence of
other variables (e.g., the radiographic extent or location of
the necrotic lesion) on long-term graft survival are also
unknown.
123
1134 Yoo et al.
We addressed the following questions: (1) Does this
procedure provide a long-term (10 years or more)
improvement in function (as measured by Harris hip
score)?; (2) Does the procedure avert the need for THA
and, if so, for how long?; (3) Do factors such as age, size
and location of the lesion, or etiology of the disease
influence long-term survivorship?; and (4) Does radio-
graphic appearance change over time?
Materials and Methods
We retrospectively reviewed 135 patients (151 hips) who
underwent free VFG in for Ficat and Arlet stage II and III
ONFH between August 1979 and December 1995 in a single
institution. Important clinical and imaging data including
patient age (two groups: younger than and older than
35 years at the time of operation), etiology, Harris hip score,
and various radiographic parameters (collapse stage, extent
of involvement, and location of the necrotic lesion) were
collected for all the patients. The outcome and survivorship
of VFG was evaluated using conversion to THA as the
endpoint. Seventeen patients (19 hips) died and eight patients
(eight hips) were lost to followup. This left 110 patients
(81%) with 124 hips for review. The minimum followup was
10 years (mean, 13.9 years; range, 10–23.7 years). There
were 94 men and 16 women with a mean age at surgery of
35.5 years (range, 13–63 years). The diagnosis of osteone-
crosis was confirmed in all cases by a histologic examination
of the subchondral bone that was obtained from a core biopsy
of the femoral head during the operation. We divided hips
into four groups based on etiology of the disease: idiopathic
(n = 59), alcoholic (n = 31), posttraumatic (n = 21), and
steroid-induced (n = 13) (Table 1).
We (KKI, PSW) examined the preoperative radiographs
of the patients to determine the location and size of the
necrosis as well as the presence or absence of collapse. For
the presence of radiographic collapse, we used the Ficat
Table 1. Demographic data
Variables n
Average patient age (years) 35.5 (range, 13–63)
Number of hips B 35 years old 68
Number of hips [ 36 years old 56
Male:female 94:16
Mean patient weight (kg) 63.5 (range, 45–97)
Etiology (number of hips)
Idiopathic 59
Alcoholic 31
Posttraumatic 21
Steroid-induced 13
123
Clinical Orthopaedics and Related Research
and Arlet system [9] and for quantifying the lesion, the
Steinberg classification [37] was used. Size A lesions
involved less than 15% of the femoral head, Size B
involved 15% to 30%, and Size C involved more than 30%
of the femoral head. Also, for evaluating the location of the
lesion, the method by Sugano et al. [40] was used, which
was adopted as a classification by Ohzono et al. [27]. In
Type A, the lesion involved the medial third of the weight-
bearing dome (sourcil) of the acetabulum. Type B lesions
involved the middle third (medial two-thirds or less) of the
weight-bearing dome, and Type C lesions involved the
lateral third or more of the dome. Preoperatively, 59 hips
were classified as Ficat Stage II and 65 hips as Stage III.
Based on the Steinberg classification, 27 hips were Size A,
38 hips were Size B, and 59 hips were Size C. Based on
location of the lesion, nine hips were classified as Type A,
35 hips were Type B, and 80 hips were Type C (Table 2).
The operative technique has been previously described
[44] and was originally designed by the senior author
(YMC) [44]. One surgical team by the senior author
operates on the femoral side. Briefly, this involves expo-
sure of the femur using a Watson-Jones approach.
Following release of the gluteus maximus insertion and
release of the vastus lateralis from linea aspera the first or
the second perforating branch of the profunda femoris
artery was dissected carefully. In general, the second per-
forating branch was preferable for the recipient vessels
owing to enough length and diameter of the vessels. After
complete dissection of the recipient vessels, a large hole 2
cm in diameter in the lateral cortex just beneath the flare of
the trochanter was made. At this point we performed a
biopsy of the subchondral bone using an 8-mm trephine
directed towards the necrotic lesion. A tunnel was then
created in the femoral neck to admit the fibula and its
vessels without compressing them. The tunnel was directed
towards the lesion and as much of the subchondral bone as
possible was removed. At this point we brought the fibular
Table 2. Classifications of osteonecrosis of the femoral head
Variables Number of hips (%)
Ficat stage [9]
II 59 (47.6)
III 65 (52.4)
Steinberg classification [37]
A 27 (21.8)
B 38 (30.6)
C 59 (47.6)
Ohzono classification [27]
A 9 (7.3)
B 35 (28.2)
C 80 (64.5)
Volume 466, Number 5, May 2008
graft harvested by another team (HCS) to the field and
performed the vascular anastomosis. The fibula was har-
vested using a curvilinear incision over the fibula of the
contralateral leg. The dissection then proceeded between
the peroneus longus and the soleus muscles. With careful
dissection of the flexor hallucis longus muscle, we exposed
the entire course of the peroneal artery. Then the antero-
lateral musculature attached to the fibula was released and
fibular freed from the interosseous membrane on the
medial side. We used the middle third of the fibula as the
graft. The peroneal vessels supplying this part of the fibula
are usually cut in a sufficient length after checking the
vascularity of the fibula with deflation of the tourniquet.
The average length of the harvested fibula usually ranged
from 8 to 10 cm. Our technique differs from others in some
respects. First, we have used the first or the second per-
forating branch of the profunda femoris artery rather than a
branch of lateral femoral circumflex artery. We think it is
easier and has less morbidity to the hip joint. We do not
routinely perform angiography to assess patency of the
anastomosed vessels. Instead we raise a small area of the
skin overlying fibula during harvest that allows monitoring
the vascular patency of the grafted fibula. We also perform
autografting of subchondral region using cancellous bone
chips obtained from the greater trochanter region.
We gave all patients antibiotic prophylaxis. Postopera-
tive thromboembolic prophylaxis included intravenous
infusion of dextran for 3 days after surgery and application
of compression stockings. Passive range-of-motion exer-
cises were encouraged after removal of suction drains on
Day 3. Postoperative rehabilitation included complete
nonweightbearing (wheelchair-bound) for 1 week, minimal
weightbearing and ambulation with crutches for 10 weeks,
followed by partial weightbearing for a total of 6 months.
Patients were encouraged to bear full weight after this
period. Because of the potential injury to the flexor hallucis
longus and the risk of clawing of the big toe from the fibular
harvest side, patients used a short leg splint including the
big toe for 3 weeks and we encouraged the patient to do
physiotherapy (active and passive dorsiflexion and plan-
tarflexion of the toe) for 6 months to prevent toe clawing.
Clinical evaluation was performed by the senior surgeon
(YMC) whenever the patient visited his outpatient clinic.
Clinical results were recorded preoperatively and postop-
eratively using the Harris hip score [13] (HHS).
Radiographic evaluation was performed by two (KKI,
PSW) individuals who were blinded to the functional
results. We categorized final radiographs (AP, lateral, and
frog leg view) in one of three classes: (1) Improved—
Those cases in which the osteonecrosis had healed or was
being replaced with new bone formation. For the Stage II
lesion, the crescent had disappeared or the density of cystic
lesion had increased with trabecular formation of the tip of
Free Vascularized Fibular Grafting 1135
the vascularized fibula. For the Stage III lesion, the col-
lapsed lesion healed or became more rounded with
trabecular formation of the tip of the vascularized fibula;
(2) No change—compared with the preoperative status;
and (3) Progressed—Those cases with progression
observed based on stage or those with more than 3-mm of
collapse. We confirmed any definite change on any of three
radiographs. For the exact measurement of a collapsed
lesion, we used Mose’s template of concentric circles. To
evaluate interobserver validity we compared the radio-
graphic results that were made by the two different
observers. Of 124 total cases, there was agreement between
observers in five, 61, and 48 cases with improved,
unchanged and progressed respectively. The level of
agreement was tested by Kappa statistics (k = 0.85,
p \ 0.0001) which we considered highly in agreement.
We recorded the number of cases converted to THA
during followup resulting from progression of osteone-
crosis or degenerative change. We then calculated survival
with THA as an endpoint for each radiographic group,
etiology, and patient’s age using the Kaplan-Meier method.
The changes in HHS were evaluated with the Wilcoxon
signed rank test. We used Cox proportional hazard model
to assess the independent effects of location and size of
lesion on survivorship; for the independent variables with a
categorical characteristics, we created ‘dummy variables’
if the variables had more than 2 groups. We performed all
analyses using SPSS (version 12.0; SPSS Inc, Chicago, IL).
Results
Preoperative HHS improved from 72 (range, 52–81) to 88
(range, 62–100) at the latest followup (p \ 0.001). There
were 85 hips with HHS of more than 90 points, 13 with
HHS between 80 and 90, 13 with HHS between 70 and 79,
and 13 with HHS of less than 69. HHS over 80 was
observed in 48 of 59 Stage II hips (81%) and in 50 of 65
Stage III hips (77%).
Thirteen patients with 13 hips (10%) had undergone
THA resulting in a survivorship of 93% at ten years and
83% at 20 years. Moreover, the rate of graft survival at ten
years of the patients without preoperative femoral collapse
was 93% and 92% in patients with collapse. The time
interval between fibular grafting and THA averaged
8.4 years (range, 1.3–18.8 years). Conversion to THA was
in 7 hips with Ficat Stage II and 6 hips with stage III at the
time of fibular grafting. The conversion rate to THA was
not statistically different between Stage II or Stage III hips
at 12% and 9% respectively.
We observed a higher (p = 0.019) survival rate in
patients younger than 35 years of age compared with those
older than 35 years. The location of lesion (p = 0.032), as
123
1136 Yoo et al. Clinical Orthopaedics and Related Research

well as the extent of involvement (p = 0.015) indepen- Improved or unchanged results were found in 22 of 27
dently influence survivorship. Survivorship was not (82%) hips with osteonecrosis less than 15%, 22 of 38
influenced by Ficat stage of the hip (p = 0.574) or the (58%) hips with osteonecrosis less than 30%, and 32 of 59
etiology (p = 0.204). (54%) hips with osteonecrosis more than 30%. According
Radiographically seven hips (6%) improved, 69 hips to etiology, improved or unchanged results were observed
(56%) were unchanged, and 48 hips (39%) progressed in 31 of 59 (53%) hips (53 patients) with idiopathic
(Fig. 1A–D). We observed improved or unchanged radio- necrosis, 18 of 31 (58%) hips (26 patients) with alcoholic
graphs in 37 of 59 (63%) hips at Ficat-Arlet Stage II and 39 necrosis, 17 of 21 hips (21 patients) with traumatic
of 65 (60%) hips at Stage III (Table 3). According to the necrosis, and 10 of 13 hips (10 patients) with steroid-
location of femoral head necrosis, improved or unchanged induced osteonecrosis.
results were seen in seven of nine hips with osteonecrosis Clawing of the big toe developed in 17 cases and most
in the medial region, 22 of 35 (63%) hips in the central of the patients were treated nonoperatively including
region, and 47 of 80 (59%) hips in the lateral region. physiotherapy except three cases having surgical release of
Fig. 1A–D (A) Preoperative radio-
graph shows a 19-year-old man
with collapse of the femoral head
secondary to osteonecrosis. Vas-
cularized fibular grafting was
performed resulting in an excel-
lent outcome. Anteroposterior
radiograph of the same hip is
shown at (B) 3 months, (C)
7 years, and (D) 15 years.

123
Volume 466, Number 5, May 2008 Free Vascularized Fibular Grafting 1137

Table 3. Summary of the outcomes previous studies report the outcome based on etiology and
Variables ‘‘Improved’’ or HHS over 80, Conversion to
radiographic stage of the disease [4, 7, 14, 16, 20, 31, 43,
‘‘Unchanged,’’ number of THA, number 44]. However, size and location of necrotic involvement
number of hips (%) of hips (%) also influence outcomes [25–27, 39, 40]. Most of these
hips (%) reports reflect short- or mid-term followup and rather than
Ficat stage [9] the long-term followup. We therefore addressed the fol-
II 37 (62.7) 48 (81.3) 7 (11.8) lowing questions: (1) Does this procedure provide a long-
III 39 (60.0) 50 (76.9) 6 (9.2) term (10 years or more) improvement in function (as
p-value* NS NS NS measured by Harris hip score)?; (2) Does the procedure
Ohzono classification [27]
avert the need for THA and, if so, for how long?; (3) Do
A 7 (77.8) 9 (100) 0 (0)
factors such as age, size and location of the lesion, or eti-
ology of the disease influence long-term survivorship?; and
B 22 (62.8) 31 (88.6) 1 (2.8)
(4) Does radiographic appearance change over time?
C 47 (58.7) 58 (72.5) 12 (15.0)
We are aware of some limitations of our study. First, we
p-value* NS 0.011 0.039
had no control group treated with alternative joint-pre-
Steinberg classification [37]
serving procedures. Second, because the criteria for
A 22 (81.5) 27 (100) 0 (0)
determining both clinical and radiographic assessment is
B 22 (57.9) 30 (78.9) 3 (78.9)
different in each article, direct comparison of our findings
C 32 (54.2) 41 (69.5) 10 (16.9)
to those reported by others is difficult. We also recognize
p-value* 0.026 0.004 0.012
that even for appropriately experienced surgeons VFG is a
Etiology
complex operation often with long operation time, the
Idiopathic 31 (52.5) 47 (79.6) 8 (13.5)
necessity of two surgical teams to reduce the time, donor
Alcoholic 18 (58.0) 22 (70.9) 4 (12.9)
site morbidity, and substantial postoperative complications.
Posttraumatic 17 (80.9) 18 (85.7) 0 (0) Marcus et al. [21] proposed the most desirable time for
Steroid-induced 10 (76.9) 11 (84.6) 1 (7.7) joint-sparing surgery for ONFH is before collapse of the
p-value* 0.039 NS NS femoral head. Springfield and Enneking [32], on the other
Age hand, suggested the possibility for regeneration of the
B 35 years 50 (73.5) 57 (83.8) 3 (4.4) femoral head still exists even in some cases with collapse
[ 36 years 26 (46.4) 41 (73.2) 10 (17.8) and subchondral fracture. Others have also reiterated that
p-value* 0.009 NS 0.014 collapse does not necessarily imply a poor prognosis and
* The chi square test was used to test significance among the groups. cessation of collapse can be expected in a certain per-
centage of hips [25]. Marciniak et al. [20] found no
correlation between the initial radiographic stage and
the flexor hallucis longus tendon. This complication was clinical outcomes or the overall rate of graft survival. Their
caused by the extensive dissection and injury of the flexor 5-year results were even more encouraging for the Marcus-
hallucis longus muscle. Partial peroneal nerve palsy Enneking Stage 3 and 4 hips than for the hips with Stage 2.
developed in two patients; their symptoms resolved within Moreover, in the large series of mid-term followup, Scully
1.5 years. Superficial infections developed in two patients et al. [30] also concluded VFG could delay or prevent
and were successfully resolved with adequate antibiotics collapse in hips that have Ficat Stage II or III. Our long-
therapy. Subtrochanteric fractures occurred in two patients; term data also suggests similar survival in Ficat Stage II
these were successfully treated with open reduction and and III, findings similar to those of Judet and Gilbert [14].
internal fixation (Fig. 2A–B). We obtained a rate of 92.1% of graft survival at 10 years in
65 patients with femoral head collapse. These findings
confirm VFG can be used successfully even in the period of
Discussion early collapse such as Ficat Stage III.
Previous reports have suggested a direct correlation
Although free VFG in the treatment of ONFH is reported to between the size of necrotic lesion and the outcome of
have encouraging short to mid-term results, long-term VFG [33]. While others have refuted the latter being
outcome of this procedure is largely unknown [4, 6, 7, 10, unable to identify an association between lesion size and
12, 14, 15, 17–21, 31, 43, 44]. Furthermore, some studies survivorship of VFG at a mean followup of 4.3 years [4].
suggest the success rate after VFG decreases with time [3, The classification system they used was different from the
20], various results have been reported both clinically and widely used Steinberg classification because they classified
radiographically [5, 7, 19, 31, 43, 44] (Table 4). Most femoral involvement as less than 25%, 25% to 50%, or

123
1138 Yoo et al. Clinical Orthopaedics and Related Research

Fig. 2A–B (A) Preoperative radio-

graph of the right hip of a 46-year-
old man who slipped and fell
1 month after VFG shows subtro-
chanteric fracture. (B) 16-year
followup radiograph shows excel-
lent bony union and well-
maintained fibular graft without
substantial collapse of the femoral
head.

Table 4. Comparison of our results with those in the literature

Study Number of hips Followup (years) HHS [ 80 (%) Radiographic Survival (%)
progression (%)

Malizos et al. [19] 40 2.7 87.5 12.5 92.5

Louie et al. [17] 59 4.2 NA 48 73
Soucacos et al. [31] 184 4.7 NA 37.5 92.4
Plakseychuk et al. [28] 50 5 70 24 86
Yoo et al. [44] 81 5 91 11 96
Marciniak et al. [20] 101 5 NA 57.4 61
Berend et al. [4] 121 5.7 63 NA 64.5
Urbaniak et al. [43] 103 7 81 NA 70
Brunelli and Brunelli [5] 18 7.7 78 46 NA
Judet and Gilbert [14] 68 18 52* NA 73.5
Current study 124 13.9 79 38.7 89.6
* Merle d’Aubigne score [ 15.

more than 50%. A direct comparison between the two
results is thus impossible. Our data showed a relationship
between the size of the lesion and survivorship. Our data
support the quantitative analysis of lesion morphology by
Nishii et al. [26], demonstrated lesion volume correlated
with progression.
Currently, the importance of lesion location and size is
well accepted [22, 26]. A laterally located lesion on the
weight-bearing surface of the acetabulum predicts poor
outcomes [25–27, 39, 40]. In such types, regardless of size,
head collapse is likely to occur soon after the onset of the
disease [27]. Thus, joint-preserving surgery may have a
limited role for treatment of lateral lesions, particularly
when they are large. Our data showed a relationship
between the size of the lesion and outcome. Twelve of 13
failures in our series occurred in patients with lateral lesion
(Type C). On the contrary, we had a 100% rate of graft
survival with nine cases of Type A lesions, but such lesions
are believed less predictive than Type B or C [25]. Thus,
we no longer perform this procedure in patients with Type
A lesions unless a simpler procedure of core decompres-
sion fails to alleviate consistent pain.

123
Volume 466, Number 5, May 2008
A few studies considered patient age at the time of
operation as a possible factor predicting survival [8, 14, 20,
43]. Several authors [7, 20, 43] suggest age does not affect
the results in mid-term followup, whereas Judet and Gilbert
[14] reported, in 68 cases of VFG with an average followup
of 18 years, that better results were obtained in patients
younger than 40 years of age. Another report [8] suggests a
trend toward a lower rate of failure in younger patients.
Our long-term results suggest a better outcome in patients
younger than 35 years of age. Thus, the results of Judet and
Gilbert [14] and our long-term results support the view that
the patient’s age can be one of the key factors in long-term
survival. Because this procedure involves a microvascular
repair, it is generally agreed younger patients undergoing
microvascular surgery have a higher success rate than older
patients.
Berend et al. [4] and others [43, 44] reported etiology
was not a factor in the success of VFG in their group.
Urbaniak et al. [43] also found no difference in survival
rate according to etiology. Our long-term study confirms
those findings.
We conclude this method as a joint-preserving treatment
for osteonecrosis is a reasonable option. Preoperative
evaluation should include not only the stage, but also the
extent and location of the necrosis to predict long-term
graft survival. This modality appears especially effective in
young patients.
Acknowledgment We thank Dr. Sung-Woo Park for the radio-
graphic assessment.
References
1. Aaron RK. Treatment of osteonecrosis of the femoral head with
electrical stimulation. Instr Course Lect. 1994;43:495–498.
2. Arai K, Toh S, Tsubo K, Nishikawa S, Narita S, Miura H.
Complications of vascularized fibula graft for reconstruction of
long bones. Plast Reconstr Surg. 2002;1097:2301–2306.
3. Berend KR, Gunneson E, Urbaniak JR, Vail TP. Hip arthroplasty
after failed free vascularized fibular grafting for osteonecrosis in
young patients. J Arthroplasty. 2003;184:411–419.
4. Berend KR, Gunneson EE, Urbaniak JR. Free vascularized fibular
grafting for the treatment of postcollapse osteonecrosis of the
femoral head. J Bone Joint Surg Am. 2003;85:987–993.
5. Brunelli G, Brunelli G. Free microvascular fibular transfer for
idiopathic femoral head necrosis: long-term follow-up. J Recon-
str Microsurg. 1991;7:285–295.
6. Chillag KJ. Survival analysis of hips treated with core decom-
pression or vascularized fibular grafting because of avascular
necrosis. J Bone Joint Surg Am. 2000;82:289–290.
7. Cho BC, Kim SY, Lee JH, Ramasastry SS, Weinzweig N, Baik
BS. Treatment of osteonecrosis of the femoral head with free
vascularized fibular transfer. Ann Plast Surg. 1998;40:586–593.
8. Dean GS, Kime RC, Fitch RD, Gunneson E, Urbaniak JR.
Treatment of osteonecrosis in the hip of pediatric patients by free
vascularized fibular graft. Clin Orthop Relat Res. 2001;386:106–
113.
Free Vascularized Fibular Grafting 1139
9. Ficat P, Arlet J, Vidal R, Ricci A, Fournial JC. Therapeutic
results of drill biopsy in primary osteonecrosis of the femoral
head (100 cases) [in French]. Rev Rhum Mal Osteoartic.
1971;38:269–276.
10. Garberina MJ, Berend KR, Gunneson EE, Urbaniak JR. Results
of free vascularized fibular grafting for femoral head osteone-
crosis in patients with systemic lupus erythematosus. Orthop Clin
North Am. 2004;353:353–357.
11. Gonzalez Della Valle A, Bates J, Di Carlo E, Salvati EA. Failure
of free vascularized fibular graft for osteonecrosis of the femoral
head: a histopathologic study of 6 cases. J Arthroplasty.
2005;203:331–336.
12. Goodman SB. Survival analysis of hips treated with core
decompression or vascularized fibular grafting because of avas-
cular necrosis. J Bone Joint Surg Am. 2000;82:289.
13. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
14. Judet H, Gilbert A. Long-term results of free vascularized fibular
grafting for femoral head necrosis. Clin Orthop Relat Res.
2001;386:114–119.
15. Kane SM, Ward WA, Jordan LC, Guilford WB, Hanley EN Jr.
Vascularized fibular grafting compared with core decompression
in the treatment of femoral head osteonecrosis. Orthopedics.
1996;19:869–872.
16. Kim SY, Kim YG, Kim PT, Ihn JC, Cho BC, Koo KH. Vascu-
larized compared with nonvascularized fibular grafts for large
osteonecrotic lesions of the femoral head. J Bone Joint Surg Am.
2005;879:2012–2018.
17. Louie BE, McKee MD, Richards RR, Mahoney JL, Waddell JP,
Beaton DE, Schemitsch EH, Yoo DJ. Treatment of osteonecrosis
of the femoral head by free vascularized fibular grafting: an
analysis of surgical outcome and patient health status. Can J
Surg. 1999;42:274–283.
18. Malizos KN, Quarles LD, Dailiana ZH, Rizk WS, Seaber AV,
Urbaniak JR. Analysis of failures after vascularized fibular
grafting in femoral head necrosis. Orthop Clin North Am.
2004;353:305–314.
19. Malizos KN, Soucacos PN, Beris AE. Osteonecrosis of the
femoral head. Hip salvaging with implantation of a vascularized
fibular graft. Clin Orthop Relat Res. 1995;314:67–75.
20. Marciniak D, Furey C, Shaffer JW. Osteonecrosis of the femoral
head. A study of 101 hips treated with vascularized fibular
grafting. J Bone Joint Surg Am. 2005;874:742–747.
21. Marcus ND, Enneking WF, Massam RA. The silent hip in idio-
pathic aseptic necrosis. Treatment by bone-grafting. J Bone Joint
Surg Am. 1973;55:1351–1366.
22. Mont MA, Jones LC, Einhorn TA, Hungerford DS, Reddi AH.
Osteonecrosis of the femoral head. Potential treatment with
growth and differentiation factors. Clin Orthop Relat Res.
1998;355(Suppl):S314–S335.
23. Mont MA, Jones LC, Hungerford DS. Survival analysis of hips
treated with core decompression or vascularized fibular grafting
because of avascular necrosis. J Bone Joint Surg Am.
2000;82:290–291.
24. Mulliken BD. Osteonecrosis of the femoral head: current con-
cepts and controversies. Iowa Orthop J. 1993;13:160–166.
25. Nishii T, Sugano N, Ohzono K, Sakai T, Haraguchi K, Yoshik-
awa H. Progression and cessation of collapse in osteonecrosis of
the femoral head. Clin Orthop Relat Res. 2002;400:149–157.
26. Nishii T, Sugano N, Ohzono K, Sakai T, Sato Y, Yoshikawa H.
Significance of lesion size and location in the prediction of
collapse of osteonecrosis of the femoral head: a new three-
dimensional quantification using magnetic resonance imaging.
J Orthop Res. 2002;201:130–136.
123
1140 Yoo et al.
27. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular
necrosis of the femoral head. J Bone Joint Surg Br. 1991;73:
68–72.
28. Plakseychuk AY, Kim SY, Park BC, Varitimidis SE, Rubash HE,
Sotereanos DG. Vascularized compared with nonvascularized
fibular grafting for the treatment of osteonecrosis of the femoral
head. J Bone Joint Surg Am. 2003;85:589–596.
29. Plancher KD, Razi A. Management of osteonecrosis of the
femoral head. Orthop Clin North Am. 1997;283:461–477.
30. Scully SP, Aaron RK, Urbaniak JR. Survival analysis of hips
treated with core decompression or vascularized fibular grafting
because of avascular necrosis. J Bone Joint Surg Am.
1998;80:1270–1275.
31. Soucacos PN, Beris AE, Malizos K, Koropilias A, Zalavras H,
Dailiana Z. Treatment of avascular necrosis of the femoral head
with vascularized fibular transplant. Clin Orthop Relat Res.
2001;386:120–130.
32. Springfield DS, Enneking WJ. Surgery for aseptic necrosis of the
femoral head. Clin Orthop Relat Res. 1978;130:175–185.
33. Steinberg ME, Bands RE, Parry S, Hoffman E, Chan T, Hartman
KM. Does lesion size affect the outcome in avascular necrosis?
Clin Orthop Relat Res. 1999;367:262–271.
34. Steinberg ME, Brighton CT, Corces A, Hayken GD, Steinberg
DR, Strafford B, Tooze SE, Fallon M. Osteonecrosis of the
femoral head. Results of core decompression and grafting with
and without electrical stimulation. Clin Orthop Relat Res.
1989;249:199–208.
35. Steinberg ME, Brighton CT, Hayken GD, Tooze SE, Steinberg
DR. Electrical stimulation in the treatment of osteonecrosis of the
femoral head—a 1-year follow-up. Orthop Clin North Am.
1985;16:747–756.
36. Steinberg ME, Brighton CT, Steinberg DR, Tooze SE, Hayken
GD. Treatment of avascular necrosis of the femoral head by a
123
Clinical Orthopaedics and Related Research
combination of bone grafting, decompression, and electrical
stimulation. Clin Orthop Relat Res. 1984;186:137–153.
37. Steinberg ME, Hayken GD, Steinberg DR. A quantitative system for
staging avascular necrosis. J Bone Joint Surg Br. 1995;77:34–41.
38. Steinberg ME, Larcom PG, Strafford B, Hosick WB, Corces A,
Bands RE, Hartman KE. Core decompression with bone grafting
for osteonecrosis of the femoral head. Clin Orthop Relat Res.
2001;386:71–78.
39. Sugano N, Atsumi T, Ohzono K, Kubo T, Hotokebuchi T,
Takaoka K. The 2001 revised criteria for diagnosis, classification,
and staging of idiopathic osteonecrosis of the femoral head.
J Orthop Sci. 2002;75:601–605.
40. Sugano N, Takaoka K, Ohzono K, Matsui M, Masuhara K, Ono
K. Prognostication of nontraumatic avascular necrosis of the
femoral head. Significance of location and size of the necrotic
lesion. Clin Orthop Relat Res. 1994;303:155–164.
41. Sugioka Y, Hotokebuchi T, Tsutsui H. Transtrochanteric anterior
rotational osteotomy for idiopathic and steroid-induced necrosis
of the femoral head. Indications and long-term results. Clin
Orthop Relat Res. 1992;277:111–120.
42. Trancik T, Lunceford E, Strum D. The effect of electrical stim-
ulation on osteonecrosis of the femoral head. Clin Orthop Relat
Res. 1990;256:120–124.
43. Urbaniak JR, Coogan PG, Gunneson EB, Nunley JA. Treatment
of osteonecrosis of the femoral head with free vascularized fibular
grafting. A long-term follow-up study of one hundred and three
hips. J Bone Joint Surg Am. 1995;77:681–694.
44. Yoo MC, Chung DW, Hahn CS. Free vascularized fibula grafting
for the treatment of osteonecrosis of the femoral head. Clin
Orthop Relat Res. 1992;277:128–138.
45. Zhang C, Zeng B, Xu Z, Sui S, Song W, Jin D, Shi H, Wang K.
Treatment of osteonecrosis of femoral head with free vascular-
ized fibula grafting [in Chinese]. Zhongguo Xiu Fu Chong Jian
Wai Ke Za Zhi. 2004;18:367–369.
Clin Orthop Relat Res (2008) 466:1141–1147
DOI 10.1007/s11999-008-0202-y
SYMPOSIUM: MOLECUAR AND SURGICAL ADVANCES IN OSTEONECROSIS
THA Using an Anatomic Stem in Patients With Femoral Head
Osteonecrosis
Yong-Chan Ha MD, Hee Joong Kim MD,
Shin-Yoon Kim MD, Tae-Young Kim MD,
Kyung-Hoi Koo MD
Published online: 8 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Treating young patients with femoral head
osteonecrosis (ON) remains challenging. Anatomic stems
were introduced in the 1980s and 1990s to improve the
proximal canal fit in an attempt to enhance long-term
implant survival, an important aspect of treating young
patients. We began using one design in 1993 and asked
three questions to confirm whether the design criteria
improved outcomes in patients with ON: (1) What is the
long term survivorship of these implants?; (2) What is the
amount and rate of wear?; and (3) What is the incidence
of osteolysis? We retrospectively reviewed 56 patients
(69 hips) who underwent THA for femoral head ON with
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
Y.-C. Ha, H. J. Kim, K.-H. Koo
Department of Orthopaedic Surgery, Seoul National University
College of Medicine, Seoul, South Korea
S.-Y. Kim
Department of Orthopaedic Surgery, Kyungpook National
University College of Medicine, Daegu, South Korea
T.-Y. Kim
Department of Orthopaedic Surgery, Korea University College
of Medicine, Seoul, South Korea
K.-H. Koo (&)
Department of Orthopaedic Surgery, Seoul National University
Bundang Hospital, 300 Gumi-dong, Bundang-gu, Seongnam
463-707, South Korea
e-mail: khkoo@snu.ac.kr
a cementless anatomic stem proximally coated with
hydroxyapatite. Four patients (four hips) were lost to fol-
lowup and 16 patients (19 hips) died. In the remaining 36
patients (46 hips) the minimum followup was 10 years
(mean, 11.2 years; range, 10–13 years). The mean age at
operation was 48.6 years. The average Harris hip score at
last followup was 87 points. Worst-case survivorship was
58.1% at 13 years and best-case was 93.3%. The average
linear wear of the polyethylene liner was 2.02 mm and the
average annual wear was 0.18 mm per year. Thirty-seven
hips (80%) had femoral osteolysis and 14 (30%) had ace-
tabular osteolysis. One patient who had extensive femoral
osteolysis and stem loosening was revised at 11.2 years
postoperatively. The high rates of polyethylene wear and
osteolysis are of concern.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Femoral head osteonecrosis (ON) occurs in young patients
with a mean age of younger than 50 years [19, 29, 31] and
treating these young patients remains a major therapeutic
challenge [5, 18, 22, 34].
Cementless THA was developed to obtain biologic fix-
ation and increase the longevity of the implant. Short-term
results of various types of cementless THA were encour-
aging [3, 4, 7]. However, THA using the first-generation
straight femoral stems was associated with high rates of
failure as a result of thigh pain, subsidence of the femoral
stem, aseptic loosening, proximal loss of bone attributable
to stress shielding, and polyethylene particle-induced
osteolysis [1, 2, 8, 13, 21, 24, 28, 32, 33, 35].
123
1142 Ha et al. Clinical Orthopaedics and Related Research

Anatomic stems were designed to maximize the proxi- Table 1. Demographic data
mal fit in both the coronal and the sagittal planes and Variable Data
hopefully enhance implant survival compared to that of the
first-generation stems. Implant survival would, of course, Number of patients 36
be particularly important for the younger patients with ON. Number of hips 46
However, the effectiveness of the anatomic stem remains Gender (male/female) 30 (39 hips) / 6 (7 hips)
controversial and it is not known whether the design cri- Age (years) mean, 48.6; range, 22–65
teria solved the problems with the first-generation straight \ 50 years (%) 22 (61.1%)
stems [3, 20, 28]. C 50 years (%) 14 (38.9%)
We therefore asked three questions: (1) What is the long Body weight (kg) mean, 64.8; range, 49–89
term survivorship of these implants?; (2) What is the Height (cm) mean, 167.2; range, 151–183
amount and rate of wear?; and (3) What is the incidence of Body mass index mean, 23.9; range, 22.1–25.8
osteolysis? Preoperative Harris hip score mean, 44.2; range, 20–62
Duration of followup (years) mean, 11.2; range, 10–13
Risk factors for osteonecrosis
Materials and Methods Idiopathic 17 patients (23 hips)
Alcohol abuse 13 patients (16 hips)
We retrospectively reviewed 56 patients (69 hips) with Steroid 4 patients (5 hips)
femoral head ON treated with an anatomic cementless stem Femoral neck fracture 2 patients (2 hips)
design between July 1993 and December 1995. Of the 56 Ficat stage
patients (69 hips), 16 patients (19 hips) died because of Stage IIB 6 hips
problems unrelated to the operation and four patients (four Stage III 29 hips
hips) were lost before a minimum followup of 10 years. Stage IV 11 hips
These 20 patients (23 hips) were followed for an average of Occupational activity
4.7 years (range, 1–9 years). The causes of death in the 16 Sedentary workers 9 patients (11 hips)
deceased patients were chronic liver disease in seven, Moderate workers 8 patients (10 hips)
chronic renal failure in two, heart disease in two, malignant
Intermediate workers 16 patients (20 hips)
neoplasm in two, traffic accident in two, and unknown in
Intensive workers 3 patients (5 hips)
one. None of the 16 deceased patients (19 hips) and four
lost patients (four hips) had a revision or a reoperation at
last followup. Their mean Harris hip score was 91.4 (range, The stem design was a cementless anatomic metaphyseal
82–98) at the last followup. None of these patients used fitting titanium stem. The stem had a rough surface with an
support and one patient (one hip) had a mild limp; one average roughness of 1.2 micron and the proximal 50% of
patient had mild thigh pain. None of the 20 hips had the stem was coated with hydroxyapatite in a thickness of
loosening at last followup but femoral osteolysis was seen 100 micron (Profile, DePuy, Leeds, UK) (Fig. 1).
in one hip in Gruen zones 1 and 7 and acetabular osteolysis All operations were performed by one surgeon (K-HK)
was seen in one hip in DeLee and Charnley zone I. Of the using the posterolateral approach. All of the acetabular and
36 surviving patients there were 30 men (39 hips) and six femoral components were inserted in a press-fit manner. The
women (seven hips). The mean age of these 36 patients was surgeon tried to position the acetabular cup at the abduction
48.6 years (range, 22–65 years) at the time of the index of 40° ± 10° as suggested by Lewinnek et al. [26]. A por-
operation. Twenty-two (61%, 27 hips) of 36 patients were ous-coated hemispheric cup (Duraloc 1200; DePuy,
less than 50 years old. On preoperative radiographic clas- Warsaw, IN), an ultrahigh-molecular-weight polyethylene
sification with the system of Ficat [11], six hips were in liner (Enduron; DePuy), and a 28-mm cobalt-chromium
stage II-B, 29 in stage III and 11 in stage IV. Nine patients head (DePuy) were used in all hips.
(11 hips) were sedentary workers, eight patients (10 hips) Patients were instructed to walk with partial weight-
moderate workers, 16 patients (20 hips) intermediate bearing with the aid of two crutches for 4 weeks after
workers, and three patients (five hips) intensive workers surgery.
[12] (Table 1). In the remaining 36 patients (46 hips), the We performed followup evaluations at 6 weeks; at 3, 6, 9,
minimum followup was 10 years (mean, 11.2 years; range, and 12 months; and every 6 months thereafter. Patients who
10–13 years). The study was approved by our Institutional had not returned for regularly scheduled visits were con-
Review Board; all patients were informed his or her tacted by telephone. Two nurses and one doctor located and
medical data could be used in a scientific study and pro- visited nonresponders. These three individuals performed a
vided consent preoperatively. clinical evaluation using the Harris hip scoring system [15].

123
Volume 466, Number 5, May 2008 Cementless THA Using an Anatomic Stem 1143

Fig. 1 The cementless anatomic

Profile stem, shown from three
angles, has a large proximal seg-
ment to obtain a maximized fit
into the endosteal cavity of the
proximal part of the femur. The
proximal part of the stem is
coated with hydroxyapatite.

We considered the 6-week anteroposterior and cross- at the final followup. No patient used support and three
table lateral radiographs the baseline study for radiographic patients (three hips [7%]) had a mild limp. Thigh pain was
comparison. Two of us (Y-CH and S-YK) measured the present in three patients (three hips [7%]).
abduction and anteversion angles of the acetabular com- Worst-case survivorship was 58.1% (95% confidence
ponent using the method of Widmer and Zurfluh [36] and interval, 41.8%–71.3%) and best-case survivorship was
the offset of femur before and after the total hip arthro- 93.3% (95% confidence interval, 80.7%–100%) at 13 years
plasty using the method of Jolles et al. [16]. On the final (Fig. 2).
radiographs we assessed fixation of the femoral and ace- The average linear wear of the polyethylene liner was
tabular components, polyethylene liner wear, osteolysis, 2.02 mm (range, 0.3–4.7 mm), and the average amount of
and heterotopic ossification. The fixation of the femoral volumetric wear was 1243.43 mm3 (range, 184.6–2892.6
component was classified with use of the method of Engh mm3). The average linear wear rate was 0.18 mm per year
et al. [10] and the fixation of the acetabular component (range, 0.03–0.48 mm per year), and the average annual rate of
with use of the method of Latimer and Lachiewicz [25]. volumetric wear was 110.87 mm3 (range, 18.5–295.4 mm3).
The wear of the polyethylene liner was calculated Femoral osteolysis was seen in 37 hips (80%); 19 hips
according to the method developed by Livermore et al. (41%) had osteolysis in Gruen zone 1 and 17 hips (37%) in
[27]. Volumetric wear was calculated with the equation zones 1 and 7. These osteolytic lesions were small (less
V = [pi]r2w, where V = volumetric wear, r = the radius than 1 cm2) and were not progressive on serial radiographs.
of the femoral head, and w = measured linear wear [23, Forty-five stems (98%) had bone on growth stability
27]. Osteolytic lesions were defined according to the cri- (Fig. 3). However, one stem (2%) had diffuse extensive
teria of Engh et al. [9]. The lesions were recorded
according to the three zones described by DeLee and
Charnley [6] on the acetabular side and the seven zones
described by Gruen et al. [14] on the femoral side.
We performed Kaplan-Meier survival analysis for all
hips with a minimum ten-year followup with revision of
either component as an endpoint [17]. We did two analy-
ses: a best-case scenario (in which all 23 hips with less than
10 year followup were considered to have had no revision
through the 13 year followup) and worst-case scenario (in
which all 23 hips were considered to have required revision
before the 13 year followup). For all analyses we used
SPSS version 11.0 (Chicago, IL).

Fig. 2 The Kaplan-Meier survival curve including 95% confidence

Results
In the 36 patients (46 hips) who had a minimum followup of
10 years, the mean Harris hip score was 87 (range, 56–95)
interval estimates with implant revision as the endpoint. Excluding
patients lost to followup (best-case scenario) there was only one
failure of the femoral stem at 11.5 years (straight line). The dotted
line shows the worst-case scenario presuming all 23 hips lost to
followup had failed.

123
1144 Ha et al. Clinical Orthopaedics and Related Research

Fig. 3A–B A 28-year-old man

underwent THA resulting from
ON of the femoral head with use
of a Profile stem. (A) An antero-
posterior radiograph obtained
6 weeks postoperatively shows
radiolucent line in DeLee and
Charnley zones II and III as well
as screw sites. No remodeling
changes are seen around the stem.
(B) On anteroposterior radiograph
obtained 13 years postopera-
tively, the line disappeared in
DeLee and Charnley zone III.
Focal osteolysis (less than
1 cm2) is seen in Gruen zone 1
(arrows) around the stem and a
shelf pedestal at the tip of the
stem.

osteolysis and loosening, which was revised 11 years
2 months after the index operation (Fig. 4) (Table 2).
Acetabular osteolysis was identified in 14 hips (30%); two
hips (4%) in DeLee and Charnley zone I, eight hips (17%)
in zone II, three hips (7%) in zones I to II, and one hip (2%)
in Zones II to III (Fig. 4) (Table 2).
Reactive radiolucent line formation was found around
the uncoated distal tip of the stem in nine hips (20%) at
nine to 24 months after the operation. The line was seen in
only lateral radiograph (Gruen zones 10, 11 and 12) in
eight hips and in both anteroposterior and lateral radio-
graphs (Gruen zones 3, 4, 5, 10, 11 and 12) in one hip. The
line was parallel to the distal tip and the thickness was less
than two millimeters. It was not associated with the for-
mation of a halo pedestal and was not progressive. Thus, it
was not thought to be osteolysis. Three hips (7%) had a
shelf pedestal at the tip of the stem. Cortical hypertrophy
was observed in the distal zone (Gruen zones 3, 4 and 5) in
nine hips (20%). All hips had cortical thinning and canc-
ellization of the cortex in the calcar femorale (Fig. 3).
No acetabular component had loosening. We observed
radiolucent lines around the acetabular cup in 29 hips
(63%) at 6-week radiographs, which was presumed to be
polar gap or peripheral gap. No line had progressed on
serial radiographs. The line disappeared at six to 12 months
after the operation in 20 hips and persisted until the latest
evaluation in nine hips (20%) (Fig. 3).
There was an intraoperative linear fracture of the calcar
femorale in one hip, which was treated with a cerclage wire
and healed completely. One hip dislocated at postoperative
day one, which was treated successfully with closed
reduction and an abduction brace for 3 months. One patient
sustained a periprosthetic fracture of the proximal femur
because of a traffic accident 57 months postoperatively. The
fracture was treated with open reduction and internal fixa-
tion using cerclage wires. The fracture united uneventfully
and the stem was still well-fixed 7 years later (Fig. 5).
Discussion
Treating patients with femoral head ON is a challenge
owing to their relative young age and anticipated many
decade survival. Owing to disappointing survival rates of

123
Volume 466, Number 5, May 2008 Cementless THA Using an Anatomic Stem 1145

Fig. 4A–C A 53-year-old man underwent THA because of ON of the radiograph obtained 11 years postoperatively show extensive femoral
femoral head with use of a Profile stem. (A) An anteroposterior osteolysis (arrows). Cavitary osteolysis (arrowheads) is also seen
radiograph obtained 6 weeks postoperatively shows no remodeling around the acetabular cup in DeLee and Charnley zones I and II.
changes. (B) An anteroposterior radiograph and a (C) lateral

Table 2. Periprosthetic bone change along the implants might have influenced the long-term results. The unex-
Bone change type Percentage
pectedly high death rate in our patients seemed related to
causes of ON including alcohol abuse and steroid use,
Radiolucent line which might be potentially confounding comorbidities.
Femoral 9 (19.6%) The effectiveness of the anatomic stem remains con-
Acetabular 9 (19.6%) troversial and the long-term followup results have not yet
Osteolysis been determined. One previous study [3] reporting 2-year
Femoral 37 (80.4%) results of 50 cementless THAs with the use of Porous-
Acetabular 14 (30.4%) Coated Anatomic (PCA) stems (Howmedica, Rutherford,
Shelf pedestal 3 (6.5%) NJ), the clinical results were encouraging. However, pro-
Cortical hypertrophy 9 (19.6%) gressive radiodense femoral lines developed in 41%,
Calcar atrophy 46 (100%) progressive loosening of beads from the stem in 24%, and
thigh pain in 16%. In another study [28], 52 THAs with the
use of the PCA stem were evaluated and a high failure rate
some straight stems, anatomic stems were developed in the was reported. At a mean of 2.4 years followup, four (8%)
1980s and 1990s to improve the proximal canal fit and femoral components were revised because of subsidence
presumably long-term survival. We began using one design and persistent thigh pain was reported in 33%. Kim et al.
in 1993 and asked three questions to confirm whether the [20] reviewed 108 patients (116 hips) who were followed
design criteria improved outcomes in patients with ON: (1) for 10 to 12 years after THA using a PCA hip prosthesis.
What is the long term survivorship of these implants?; (2) The rate of revision of the femoral component was 11%,
What is the amount and rate of wear?; and (3) What is the the rate of revision of the acetabular component was 15%,
incidence of osteolysis? and the incidence of thigh pain was 28%. Moskal et al. [30]
We note several limitations to our study. Among the 56 evaluated 107 PCA arthroplasties at an average of
patients (69 hips), 16 patients (19 hips) died and four 12.4 years. The failure rate was 13% in the acetabular
patients (four hips) were lost before the 10 year minimum components and 4% in the stems. The design of the Profile
followup; additionally, the followup rate was 64%, which stem used in our study has a similar design to a PCA stem.

123
1146 Ha et al. Clinical Orthopaedics and Related Research

Fig. 5A–B A 59-year-old man

underwent THA because of ON
of the femoral head with use of a
Profile stem. The patient sustained
injury resulting from a traffic acci-
dent 57 months postoperatively.
(A) An anteroposterior radiograph
shows a periprosthetic fracture of
the proximal femur. The fracture
was treated with open reduction
and internal fixation using cerclage
wires and the fracture united
uneventfully. (B) On an anteropos-
terior radiograph obtained 7 years
after the fracture operation, the
stem is still well fixed.

Both stems are cementless anatomic metaphyseal fitting
stems with a circumferential coating on the proximal one-
third. However, the proximal portion of the Profile stem is
larger than that of the PCA stem and has a closer match to
the endosteal cavity of the proximal part of the femur and a
better surface coating.
We observed only one patient with a loose stem (2%) and
three with thigh pain (7%) at a mean of 11.2 years followup.
Our findings agree with those of Kim et al. [23] who eval-
uated 118 THAs with the use of the Profile stem. DePuy
supplied two versions of the Profile stem. The two versions
are identical with regard to geometry and stem material. The
two versions differed only with regard to the surface treat-
ment of the proximal portion; one was porous-coated and the
other was hydroxyapatite-coated. Although we used
hydroxyapatite-coated Profile stems, Kim et al. [23] used
porous-coated Profile stems. At a mean of 9.8 years fol-
lowup they had no aseptic loosening and the incidence of
transitory thigh pain was 10%. We believe the improved
design features of the Profile stem might lead to the better
results with the Profile stem compared to the PCA stem.
We observed a femoral osteolysis rate of 80% and an
acetabular osteolysis rate of 30%. These rates were much
higher than rates in the study of Kim et al. [23] who
reported 12% femoral osteolysis and 9% acetabular oste-
olysis. Their low incidence of osteolysis around the
acetabular and femoral components might be related to the
relatively low rate of volumetric wear of polyethylene
because they used a 22-mm head. The different surface
treatments, hydroxyapatite coating and porous coating,
might have influenced the rate of osteolysis.
Our data suggest the anatomic fit cementless Profile stem
had durable fixation in young, active patients who had fem-
oral head ON. However, a high wear rate of the polyethylene
liner and a high incidence of osteolysis are concerning.
References
1. Bourne RB, Rorabeck CH, Ghazal ME, Lee MH. Pain in the thigh
following total hip replacement with a porous-coated anatomic
prosthesis for osteoarthrosis. A five-year follow-up study. J Bone
Joint Surg Am. 1994;76:1464–1470.
2. Callaghan JJ. Results of primary total hip arthroplasty in young
patients. Instr Course Lect. 1994;43:315–321.
3. Callaghan JJ, Dysart SH, Savory CG. The uncemented porous-
coated anatomic total hip prosthesis. Two-year results of a pro-
spective consecutive series. J Bone Joint Surg Am. 1988;70:337–
346.
4. Cameron HU. The results of early clinical trials with a micro-
porous coated metal hip prosthesis. Clin Orthop Relat Res.
1982;165:188–190.
5. Cornell CN, Salvati EA, Pellicci PM. Long-term follow-up of
total hip replacement in patients with osteonecrosis. Orthop Clin
North Am. 1985;16:757–769.
6. DeLee JG, Charnley J. Radiological demarcation of cemented
sockets in total hip replacement. Clin Orthop Relat Res.
1976;121:20–32.

123
Volume 466, Number 5, May 2008
7. Engh CA. Hip arthroplasty with a Moore prosthesis with porous
coating. A five-year study. Clin Orthop Relat Res. 1983;176:52–66.
8. Engh CA, Culpepper WJ II, Engh CA. Long-term results of use
of the anatomic medullary locking prosthesis in total hip
arthroplasty. J Bone Joint Surg Am. 1997;79:177–184.
9. Engh CA, Hooten JP Jr, Zettl-Schaffer KF, Ghaffarpour M,
McGovern TF, Bobyn JD. Porous-coated total hip replacement.
Clin Orthop Relat Res. 1994;298:89–96.
10. Engh CA, Massin P, Suthers KE. Roentgenographic assessment
of the biologic fixation of porous-surfaced femoral components.
Clin Orthop Relat Res. 1990;257:107–128. Erratum in: Clin
Orthop Relat Res. 1992;284:310–312.
11. Ficat RP. Idiopathic bone necrosis of the femoral head: Early
diagnosis and treatment. J Bone Joint Surg Br. 1985;67:3–9.
12. Flugsrud GB, Nordsletten L, Espehaug B, Havelin LI, Meyer HE.
Risk factors for total hip replacement due to primary osteoar-
thritis: A cohort study in 50,034 persons. Arthritis Rheum. 2002;
46:675–682.
13. Galante JO, Jacobs J. Clinical performances of ingrowth surfaces.
Clin Orthop Relat Res. 1992;276:41–49.
14. Gruen TA, McNeice GM, Amstutz HC. ‘Modes of failure’ of
cemented stem-type femoral components: a radiographic analysis
of loosening. Clin Orthop Relat Res. 1979;141:17–27.
15. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
16. Jolles BM, Zangger P, Leyvraz PF. Factors predisposing to dis-
location after primary total hip arthroplasty: a multivariate
analysis. J Arthroplasty. 2002;17:282–288.
17. Kaplan EL, Meier P. Nonparametric estimation from incomplete
observations. J Am Stat Assoc. 1958;53:457–481.
18. Katz RL, Bourne RB, Rorabeck CH, McGee H. Total hip
arthroplasty in patients with avascular necrosis of the hip: follow-
up observation on cementless and cemented operations. Clin
Orthop Relat Res. 1992;281:145–151.
19. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R. The
conservative surgical treatment of idiopathic aseptic necrosis of
the femoral head. J Bone Joint Surg Br. 1974;56:291–296.
20. Kim YH, Kim JS, Cho SH. Primary total hip arthroplasty with a
cementless porous-coated anatomic total hip prosthesis: 10- to
12-year results of prospective and consecutive series. J Arthro-
plasty. 1999;14:538–548.
21. Kim YH, Kim VE. Results of the Harris-Galante cementless hip
prosthesis. J Bone Joint Surg Br. 1992;74:83–87.
Cementless THA Using an Anatomic Stem 1147
22. Kim YH, Oh JH, Oh SH. Cementless total hip arthroplasty in
patients with osteonecrosis of the femoral head. Clin Orthop
Relat Res. 1995;320:73–84.
23. Kim YH, Oh SH, Kim JS. Primary total hip arthroplasty with a
second-generation cementless total hip prosthesis in patients
younger than fifty years of age. J Bone Joint Surg Am. 2003;85:
109–114.
24. Lachiewicz PE, Anspach WE 3rd, DeMasi R. A prospective study
of 100 consecutive Harris-Galante porous total hip arthroplasties.
2–5 year results. J Arthroplasty. 1992;7:519–526.
25. Latimer HA, Lachiewicz PF. Porous coated acetabular compo-
nents with screw fixation. Five to ten-year results. J Bone Joint
Surg Am. 1996;78:975–981.
26. Lewinnek GE, Lewis JL, Tarr R, Compere CL, Zimmerman JR.
Dislocation after total hip-replacement arthroplasties. J bone
Joint Surg Am. 1978;60:217–220.
27. Livermore J, Ilstrup D, Morrey B. Effect of femoral head size on
wear of the polyethylene acetabular component. J Bone Joint
Surg Am. 1990;72:518–528.
28. Maric Z, Karpman RR. Early failure of noncemented porous
coated anatomic total hip arthroplasty. Clin Orthop Relat Res.
1992;278:116–120.
29. Merle d’Aubigne R, Postel M, Mazabraud A, Massias P, Gueguen
J, France P. Idiopathic necrosis of the femoral head in adults.
J Bone Joint Surg Br. 1965;47:612–633.
30. Moskal JT, Jordan L, Brown TE. The porous-coated anatomic total
hip prosthesis: 11- to 13-year results. J Arthroplasty. 2004;19:
837–844.
31. Ohzono K, Saito M, Takaoka K, Ono K, Saito S, Nishina T,
Kadowaki T. Natural history of nontraumatic avascular necrosis
of the femoral head. J Bone Joint Surg Br. 1991;73:68–72.
32. Owen TD, Moran CG, Smith SR, Pinder IM. Results of unce-
mented porous-coated anatomic total hip replacement. J Bone
Joint Surg Br. 1994;76:258–262.
33. Piston RW, Engh CA, De Carvalho PI, Suthers K. Osteonecrosis
of the femoral head treated with total hip arthroplasty without
cement. J Bone Joint Surg Am. 1994;76:202–214.
34. Saito M, Nishina T, Ohzono K, Ono K. Long-term results of total
hip arthroplasty for osteonecrosis of the femoral head: a compari-
son with osteoarthritis. Clin Orthop Relat Res. 1989;244:198–207.
35. Smith SE, Garvin KL, Jardon OM, Kaplan PA. Uncemented total
hip arthroplasty. Prospective analysis of the tri-lock femoral
component. Clin Orthop Relat Res. 1991;269:43–50.
36. Widmer KH, Zurfluh B. Compliant positioning of total hip compo-
nents for optimal range of motion. J Orthop Res. 2004;22:815–821.
123
Clin Orthop Relat Res (2008) 466:1148–1153
DOI 10.1007/s11999-008-0180-0
SYMPOSIUM: MOLECULAR AND SURGICAL ADVANCES IN OSTEONECROSIS
Metal-on-Metal Hip Arthroplasty Does Equally Well
in Osteonecrosis and Osteoarthritis
Manish R. Dastane MD, William T. Long MD,
Zhinian Wan MD, Lisa Chao BS, Lawrence D. Dorr MD
Published online: 19 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Many previous reports suggest total hip
arthroplasty performs suboptimally in young patients with
osteonecrosis. We retrospectively compared the perfor-
mance of metal-on-metal articulation in a select group
of 107 patients with 112 hips (98 uncemented and 14
cemented stems) 60 years of age or younger with either
osteonecrosis (27 patients, 30 hips) or primary osteoar-
thritis (80 patients, 82 hips). We evaluated all patients with
patient-generated Harris hip score forms and serial radio-
graphs. Five mechanical complications were caused by
impingement, two with pain, two dislocations, and one
liner dissociation. At a minimum followup of 2.2 years
(mean, 5.5 years; range, 2.2–11.7 years), we observed no
osteolysis or aseptic loosening in the osteonecrosis group,
whereas one osteoarthritic hip had cup revision for loos-
ening (none showed evidence of osteolysis). None of the
stems were loose. Patients with osteonecrosis or primary
osteoarthritis were similar in clinical and radiographic
performance. The patients with metal-on-metal hip
arthroplasty for osteonecrosis had no revisions for aseptic
loosening, but did have one liner change in a cup for
painful impingement.
One or more of the authors (LDD) have received benefits from
Zimmer Inc for conducting this study.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
M. R. Dastane, W. T. Long, Z. Wan, L. Chao, L. D. Dorr (&)
Arthritis Institute, 501 E Hardy Street, 3rd Floor, Inglewood,
CA 90301, USA
e-mail: Patriciajpaul@yahoo.com
123
Level of Evidence: Level III, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Osteonecrosis (ON) is predominantly a disease of the
young, causing secondary arthritis [22]. Few reports sug-
gest longevity of THA for ON is inferior when compared
with THA for osteoarthritis (OA) [24, 28, 32]. Young,
more active patients have higher revision rates with THA
performed for any disease, but failure rates with ON are
reportedly even higher for reasons not clearly understood
[23, 24].
Metal-on-metal articulations in younger active patients
have a low rate of wear and osteolysis [30, 31]. Low wear
has been the most important factor in long-term perfor-
mance of metal-on-metal articulations [35].
We asked whether metal-on-metal articulation, com-
bined with noncemented fixation of the stem and cup,
would provide equivalent survival and clinical scores in
patients aged 60 years or younger with ON compared with
primary OA.
Materials and Methods
We retrospectively compared clinical outcomes from
clinical examination, medical records, and radiographs in
129 selected patients (135 hips) who were 60 years
or younger with ON and OA and who had Metasul1
(Zimmer, Inc., Warsaw, IN) metal-on-metal articulation.
The group is selected, in part, because of inconsistent
availability of the implant until 1999. In addition, from
Volume 466, Number 5, May 2008 Metal-on-Metal THA for Osteonecrosis 1149

1999 to 2000, there was contamination of the Sulzer
InterOp metal shell (Sulzer Orthopedics, Austin, TX) used
by us, which required revision in 14 patients; these patients
were eliminated from this study group. Nine of the 129
patients (nine hips) had surgery between 1991 and 1993; 33
patients (35 hips) had surgery between 1993 and 1998 in an
investigational device exemption for Metasul1; and 87
patients (91 hips) had surgery 1999 to 2003 after the
Metasul1 articulation was approved by the US Food and
Drug Administration. The total number of hips operated on
by us from 1999 to 2003 was 1016, so the 91 metal-on-
metal hips in that same period represent 9% of our volume.
Of the 129 patients potentially available, three (three hips)
died, five (five hips) were lost to followup, and 14 patients
(15 hips) were eliminated from the study because they
had revision surgery from a failed recalled InterOp cup
(Sulzer Orthopedics, Austin, TX) [20]. The final study
group therefore had 107 patients (112 hips), of which 80
patients (82 hips, two bilateral) had primary OA and 27 (30
hips, three bilateral) had ON. The patients with OA were
7 years older on average (p \ 0.0005) than those with
ON with no other differences between the two groups
(Table 1). All patients were contacted for recall for clinical
and radiographic examination with five patients lost to
followup. We obtained prior Institutional Review Board
consent for review of the records.
The causes of ON were idiopathic (48%), posttraumatic
(26%), alcohol abuse (15%), corticosteroids (7%), and
sickle cell disease (4%). There were six hips in Stage 3 of
ON with an obvious sequestrum and segmental collapse of
the femoral head and 24 hips in Stage 4 with severe fem-
oral head deformity and secondary OA using a staging
system designed by Ficat [14]. Only one patient with
bilateral ON had received surgical treatment in the past in
the form of core decompression.
Patients were operated on with a traditional posterior
approach as previously described [6]. All procedures were
performed by the senior surgeons (LDD, WTL). Epidural
anesthesia was augmented with general anesthesia to keep
the average arterial pressure between 60 and 80 mmHg.
Between 1991 and 1998, we used 12 cemented Weber cups
(Sulzer Medica, Winterthur, Switzerland) and 32 cement-
less APR cups (Anatomic Porous Replacement; Zimmer)
(Table 2). The APR cup was a 3.5-mm wall thickness
titanium shell with a cancellous structured titanium porous
coating. These 44 cups were coupled with 44 APR stems
(11 cemented and 33 cementless). Between 1999 and 2003,
we used 28 cementless InterOp cups (Sulzer) and 40
cementless Converge cups (Zimmer) with 67 cementless
stems and one cemented APR stem. The InterOp was
changed to the Converge in 2001 because of the recall of
the InterOp in 2000 [20]. The InterOp metal shell and the
Converge metal shell did not differ from the APR metal
shell in thickness or porous surface. The only difference in
the InterOp and Converge cups was the locking mechanism
for the acetabular insert. The Metasul1 acetabular insert
for all cups had an articulation surface of cobalt-chromium
metal, which was inlaid into a polyethylene hemisphere
having the locking mechanism for the metal shell. This
metal insert was manufactured separately and then
Table 2. Various implant combinations used during the study period
Implant Osteonecrosis Primary osteoarthritis
(n = 30) (n = 82)
Cup design
Weber 4* 8*
APR 8 24
Interop 10 18
Converge 8 32
Stem type
APR 30 (2*) 82 (10*)
* Cemented fixation; APR = anatomic porous replacement.

Table 1. Patient demographic and preoperative clinical data according to group

Variable Osteonecrosis Primary osteoarthritis p Value

Number of patients (hips) 27 (30) 80 (82)

Mean age, years (SD) 44.7 (±6.9) 51.67 (±6.7) \ 0.0005
Male patients (hips) 22 (25) 56 (57)
Mean followup, years (range) 5.5 (2.25–11.7) 5.35 (1.1–13.2) 0.2
Female patients (hips) 5 (5) 24 (25)
Weight, kg (SD) 80.6 (± 12.43) 87.95 (± 21.8) 0.181
Height, m (SD) 1.71 (± 0.10) 1.76 (± 0.10) 0.108
Body mass index (SD) 27.53 (± 4.96) 28.21 (± 6.18) 0.675
Dorr bone type A = 14, B = 16, C = 0 A = 31, B = 50, C = 1 0.56
Preoperative Harris hip score 45.92 (± 10.26) 46.15 ( ± 15.25) 0.96
SD = standard deviation.

123
1150 Dastane et al.
embedded into the polyethylene [35]. This insert was the
same for all metal shells used with the exception of the
locking mechanism of the polyethylene into the metal
shell. Metasul1 (Zimmer) was designed to have better
clearance of surfaces to avoid excess friction, promote
lubrication, and allow clearance of debris [29, 35]. The
cobalt-chromium metal used in the acetabular surface of
the Metasul1 liner and the femoral head was Protasul-21
WF (Zimmer) cobalt-chromium alloy, a high carbon
wrought forged alloy. The diameter of the femoral head
used in this study was 28 mm. The clearance between the
femoral and acetabular articulation surfaces was a mean
120 lm (range, 70–170 lm). The Metasul1 insert had an
elevated metal edge where it was inlaid. This protruding
edge could cause impingement of the metal neck and cup
(Fig. 1).
We obtained data on pain and functional outcome pre-
operatively and at the final followup with a patient self-
assessment form (patient-generated Harris hip score [16];
Orthographics, Salt Lake City, UT). We determined
activity separately by asking the patients their activity
level; activity was graded as unlimited community ambu-
lation (more than eight blocks), active community
ambulation (can walk up to eight blocks), limited com-
munity ambulation (can walk two blocks), and household
ambulation (limited to household activities) [12]. Patients
completed their forms either by mail or during followup in
the office. Ninety-six of the 107 patients (101 hips) were
seen at final followup, and 11 (11 hips) were graded by
forms returned by mail. We reviewed the medical records
for revisions, complications, and clinical scores.
Fig. 1 Retrieved Metasul1 liner with a protruded edge of the inlay
has damage to the polyethylene caused by impingement.
123
Clinical Orthopaedics and Related Research
We obtained anteroposterior pelvic radiographs and lat-
eral radiographs (iliac oblique views) preoperatively and at
each followup visit. The preoperative radiograph was used
to determine the diagnosis. We used the 6-week postoper-
ative radiographs as the baseline for comparison with final
followup radiographs for fixation and osteolysis. Wear
could not be measured from the radiographs because it was
not possible to distinguish between the edge of the femoral
head and the metal articulation surface of the acetabular
liner. One of us (ZW) measured inclination of the cup using
the technique of Callaghan et al. [4] and anteversion using
the modified technique of Ackland et al. [1, 11]. Femoral
radiolucent lines and osteolysis were recorded in each of
seven Gruen zones [15] on the anteroposterior and lateral
radiographs [17]. Calcar resorption was a focal radiolucent
area that was seen immediately under the collar of the stem
and was located between the cortex of the calcar and the
medial edge of the stem. Fixation and osteolysis of the cup
were measured by the zones of DeLee and Charnley [8].
Cup loosening was diagnosed when there was: (1) a cir-
cumferential radiolucent line of 1 mm or wider; (2)
appearance of a new radiolucent line in any zone; (3) pro-
gression of a radiolucent line; or (4) migration of the cup by
more than 2 mm of vertical or horizontal shift or a change in
inclination of more than 5° [33].
We used the Student’s t-test to compare demographic
parameters such as age, height, weight, and body mass
index and for pre- and postoperative Harris hip scores, pain
and function scores, and number of years of followup in
both groups. Chi square test was used to compare the
categorical data (gender, bone type [9], patient self-
assessment score, and postoperative level of activity) in the
two groups. A probability value of p B 0.05 was consid-
ered significant. All data were analyzed using SPSS
software (SPSS Inc, Chicago, IL).
Results
At a mean followup of 5.5 years, the postoperative score
was similar for patients with ON and those with OA
(93.73 ± 8.2 versus 93.12 ± 8.5, respectively). Mean pain
scores at the final followup were also similar (40.48 ± 5.78
for ON versus 40.74 ± 5.76 for OA) and only three
patients with ON (11.11%) and six patients with OA
(7.5%) had self-described mild to moderate pain. There
was no difference in the mean functional scores for either
group (44.64 ± 2.78 for ON versus 43.93 ± 4.71 for OA)
with only two patients in each group requiring an assistive
device for walking. Seventeen patients (20 hips) out of 27
(30 hips) with ON reported their outcome as excellent in
the ‘‘patient self-assessment form’’ compared with 60
patients (62 hips) out of 80 (82 hips) patients with OA.
Volume 466, Number 5, May 2008 Metal-on-Metal THA for Osteonecrosis 1151

Table 3. Clinical outcome in patients with osteonecrosis and

primary osteoarthritis at final followup
Outcome Patients in the Patients in the p Value
(clinical grade) osteonecrosis primary osteoarthritis
group (hips) group (hips)

Excellent 17 (20) 60 (62) 0.58

Good 8 (8) 15 (15)
Fair 2 (2) 5 (5)

Table 4. Functional activity in patients with osteonecrosis and

primary osteoarthritis at final followup
Outcome Patients in the Patients in the p Value
(functional osteonecrosis primary osteoarthritis
activity) group (hips) group (hips)
Unlimited 25 (28) 68 (70)
ambulator
Active community 2 (2) 10 (10)
ambulator
Limited community 0 2 (2)
ambulator
Overall clinical outcome and functional activity were
similar for all measured parameters (Tables 3, 4).
Radiographic results were also similar between the OA
and ON hips. The mean cup inclination for the ON group
was 39° ± 5.9° compared with 38° ± 5.7° for OA; the
mean cup anteversion for ON hips was 18.2° ± 5.4° and
17.9° ± 5.1° for OA hips. No pelvic osteolysis was
observed in either of the groups. Femoral osteolysis was
seen in Zones 3 and 4 in one of 82 hips (1.22% in the OA
group and none of the ON hips) and calcar lysis of 3 mm x
3 mm or less was observed in two hips with ON and in
three with OA.
There were no loose stems in either group and one loose
cup in the OA group. Femoral radiolucent lines, confined to
Zones 3, 4, and 5, were seen in two hips with ON and eight
with OA. Radiolucent lines were more prevalent around the
cup with seven cups in the ON group and 20 in the OA
group affected. These incomplete radiolucent lines were in
three zones in one hip with ON and in three hips with OA;
in two zones in three hips with ON and eight hips with OA;
and in one zone in three hips with ON and nine hips with
OA. Other than the revised cup, no cup had a complete
radiolucent line in all three zones or progressive lines, and
no cup showed migration.
Overall, there were six acetabular revisions with four
metal shell replacements and two liner changes. There
were no stem revisions. We revised one cup in the OA
group for aseptic loosening. This cup had bone graft for
protrusio acetabuli and had three zone circumferential
radiolucent lines on the postoperative radiograph. We
Fig. 2 Anteroposterior pelvic radiograph shows a disassociated
0.32 Metasul1 liner in the right THA at 7.7 years postoperatively.
revised two cups in the OA group for dislocation. There
were two liner changes, one in each group, for painful
impingement at 2 years and 8.5 years postoperatively. One
patient with OA had a liner dissociation, which occurred at
7.7 years followup (Fig. 2). In this hip, the acetabular
component was malpositioned with 30° of inclination and
2° of anteversion, which resulted in impingement and
failure of the locking mechanism. This hip had a revision
of the cup to a position that prevented impingement. The
only other reoperation was in one patient with ON in whom
a periprosthetic fracture occurred distal to the femoral
component as a result of a traffic accident; it was suc-
cessfully fixed with open reduction and internal fixation.
Discussion
Traditionally, ON of the femoral head as the reason for
THA has had outcomes inferior to the hips with OA as the
cause of the operation [24, 28]. Specifically, the ON hips
have had more loosening of components [24, 26, 28]. We
asked whether the outcome of THA was different in
patients 60 years old or younger with ON compared with
OA when using noncemented implant fixation and metal-
on-metal articulation (Metasul1).
Our study was limited in that it was not randomized and
has nonconsecutive patient selection with a smaller sample
size. These patients were operated on at various time
periods, depending on availability of Metasul1 implants.
The study did, however, include all of the patients 60 years
or younger operated on by us who had ON and OA and a
Metasul1 insert. The study included both cementless and
cemented cups, which have different propensities for
osteolysis. However, we had no case of osteolysis around
the cup in either group so this did not influence results.

123
1152 Dastane et al.
Finally, wear cannot be measured in metal-on-metal
articulation so we could not compare it with series using
polyethylene articulations.
We found no difference in the clinical or radiographic
outcome for the length of our followup. Our mean followup
was 5.35 years for OA and 5.5 years for ON with the
longest followup being 13.2 years and 11.7 years, respec-
tively. There was no osteolysis of the pelvis in either group
and only one OA hip showed osteolysis in the femur.
Fixation was secure in all cemented or noncemented hips
and failures in both groups were caused by impingement
except for one OA hip with a technical error, which was
revised.
The importance of our data is reflected in the similarity
of outcomes in ON and OA hips. The results of cemented
THAs with metal-on-polyethylene articulation in ON have
not generally been satisfactory in studies reported with a
similar age group and followup as our mean 5.5 years
[5, 7, 10, 24, 26, 28]. One study [7] reported an overall
failure rate of 37% at a mean followup of 7.6 years in 28
cemented arthroplasties; a second study [28] reported
unsatisfactory results in 14 of 29 ON hips (48%) compared
with 16 of 63 (25%) in patients with OA, and a third study
[26] had a poor outcome in 11 of 12 cemented ON hips. At
a long-term followup of 15 years, Ortiguera et al. [24] had
nine of 18 failed hips in 35 patients as compared with four
of 17 in patients with OA.
In contrast, cementless arthroplasties with metal-on-
polyethylene articulation have had better survival than
cemented arthroplasties for ON. One study [23] reported no
difference in fixation at 3 years among 52 matched hips; a
second [25] reported loosening of 3% of acetabular com-
ponents in 35 ON hips at 7.5 years; and a third [19] had
98% survival of 100% cementless hips at 10 years.
Failure with cemented components may be related to
changes in cancellous bone structure and remodeling with
ON. Defective cancellous bone might not support the
interdigitation of cement and the increased load placed on
it. The framework of cancellous bone in ON is apparently
weak. Arlot et al. [2] studied the histomorphometry of iliac
bone in 77 patients with ON and normal kidney function
and found osteomalacia in nine patients with the remaining
68 patients having reduced trabecular bone volume, a
reduced calcification rate, and a thin osteoid seam indi-
cating defective osteoblastic apposition and healing. Calder
et al. [3] described extensive osteocyte death and an
abnormal remodeling capacity in the proximal femur in
ON, and proposed premature loosening of implants in
patients with ON may be related to this presence of
abnormal cancellous bone at the implant-bone and cement-
bone interfaces.
Metasul1 metal articulation also protects against bone
failure by a low volume of particles, which prevents
123
Clinical Orthopaedics and Related Research
Fig. 3 The current Metasul1 liner has a buried metal inlay and a
chamfered polyethylene rim to reduce impingement.
changes in osteolysis. In retrieval analysis of 118 second-
generation Metasul1 articulations for up to 8 years, there
was volumetric wear of 0.3 mm3 per year, which is 60
times less than metal-on-polyethylene articulation [31].
Osteolysis with Metasul1 metal-on-metal articulation has
been rarely observed in studies with no osteolysis in 39
cementless arthroplasties (20 ON, 19 OA hips) in one study
[21]; in a second study [13], none at 5 to 12 years followup
of 56 cemented and cementless arthroplasties; in a third
[27], none in 106 cementless arthroplasties at a mean
6.4 years; and in the fourth [18], one of 68 hips with pelvic
osteolysis needing revision and another hip with two small
areas of focal femoral lesions around a stable stem.
We observed several mechanical complications from
impingement. Two dislocations, two liner changes, and one
liner disassociation had evidence of impingement at revi-
sion with indentations on the femoral neck and acetabular
liner. The Metasul1 liners used in our patients had an inlay
with a prominent metal rim. In addition, we used a 28-mm
head on a 12 to 14 taper femoral neck giving a head-neck
ratio of 2.0, which is a risk for impingement [34].
Metasul1 liners have been redesigned to have a buried
metal inlay in a chamfered polyethylene rim (Fig. 3), and
we use a 32-mm or larger femoral head (36-mm+ Durom;
Zimmer) when we implant Metasul1 articulations.
THA for arthritis secondary to ON of the femoral head
did not result in higher failure rates than OA with modern
cementless implants. We continue to use cementless fixa-
tion and this metal-on-metal articulation for patients
60 years of age or younger with either ON or OA as a
cause of their arthritic hip.
Volume 466, Number 5, May 2008
References
1. Ackland MK, Bourne WB, Uhthoff HK. Anteversion of the
acetabular cup. Measurement of angle after total hip replacement.
J Bone Joint Surg Br. 1986;68:409–413.
2. Arlot ME, Bonjean M, Chavassieux PM, Meunier PJ. Bone his-
tology in adults with aseptic necrosis. Histomorphometric
evaluation of iliac biopsies in seventy-seven patients. J Bone
Joint Surg Am. 1983;65:1319–1327.
3. Calder JD, Pearse MF, Revell PA. The extent of osteocyte death
in the proximal femur of patients with osteonecrosis of the
femoral head. J Bone Joint Surg Br. 2001;83:419–422.
4. Callaghan JJ, Salvati EA, Pellicci PM, Wilson PD Jr, Ranawat
CS. Results of revision for mechanical failure after cemented
total hip replacement, 1979 to 1982. A two to five-year follow-up.
J Bone Joint Surg Am. 1985;67:1074–1085.
5. Chandler HP, Reineck FT, Wixson RL, McCarthy JC. Total hip
replacement in patients younger than thirty years old. A five-year
follow-up study. J Bone Joint Surg Am. 1981;63:1426–1434.
6. Cohen J, Bindelglass DF, Dorr LD. Total hip replacement using
the APR II system. Techniques Orthop. 1991;6:40–58.
7. Cornell CN, Salvati EA, Pellicci PM. Long-term follow-up of
total hip replacement in patients with osteonecrosis. Orthop Clin
North Am. 1985;16:757–769.
8. DeLee JG, Charnley J. Radiological demarcation of cemented
sockets in total hip replacement. Clin Orthop Relat Res.
1976;121:20–32.
9. Dorr LD, Faugere MC, Mackel AM, Gruen TA, Bognar B,
Malluche HH. Structural and cellular assessment of bone quality
of proximal femur. Bone. 1993;14:231–242.
10. Dorr LD, Takei GK, Conaty JP. Total hip arthroplasties in
patients less than forty-five years old. J Bone Joint Surg Am.
1983;65:474–479.
11. Dorr LD, Wan Z. Ten years of experience with porous acetabular
components for revision surgery. Clin Orthop Relat Res.
1995;319:191–200.
12. Dorr LD, Wan Z, Gruen T. Functional results in total hip
replacement in patients 65 years and older. Clin Orthop Relat
Res. 1997;336:143–151.
13. Dorr LD, Wan Z, Longjohn DB, Dubois B, Murken R. Total hip
arthroplasty with use of the Metasul metal-on-metal articulation.
Four to seven-year results. J Bone Joint Surg Am. 2000;82:789–
798.
14. Ficat RP. Idiopathic bone necrosis of the femoral head. Early
diagnosis and treatment. J Bone Joint Surg Br. 1985;67:3–9.
15. Gruen TA, McNeice GM, Amstutz HC. ‘Modes of failure’ of
cemented stem-type femoral components: a radiographic analysis
of loosening. Clin Orthop Relat Res. 1979;141:17–27.
16. Harris WH. Traumatic arthritis of the hip after dislocation and
acetabular fractures: treatment by mold arthroplasty. An end-
result study using a new method of result evaluation. J Bone Joint
Surg Am. 1969;51:737–755.
17. Johnston RC, Fitzgerald RH Jr, Harris WH, Poss R, Muller ME,
Sledge CB. Clinical and radiographic evaluation of total hip
replacement. A standard system of terminology for reporting
results. J Bone Joint Surg Am. 1990;72:161–168.
18. Kim SY, Kyung HS, Ihn JC, Cho MR, Koo KH, Kim CY.
Cementless Metasul metal-on-metal total hip arthroplasty in
patients less than fifty years old. J Bone Joint Surg Am.
2004;86:2475–2481.
Metal-on-Metal THA for Osteonecrosis 1153
19. Kim YH, Oh SH, Kim JS, Koo KH. Contemporary total hip
arthroplasty with and without cement in patients with osteone-
crosis of the femoral head. J Bone Joint Surg Am. 2003;85:675–
681.
20. Manning DW, Ponce BA, Chiang PP, Harris WH, Burke DW.
Isolated acetabular revision through the posterior approach: short-
term results after revision of a recalled acetabular component.
J Arthroplasty. 2005;20:723–729.
21. Migaud H, Jobin A, Chantelot C, Giraud F, Laffargue P,
Duquennoy A. Cementless metal-on-metal hip arthroplasty in
patients less than 50 years of age: comparison with a matched
control group using ceramic-on-polyethylene after a minimum
5-year follow-up. J Arthroplasty. 2004;19(Suppl 3):23–28.
22. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of
the femoral head. J Bone Joint Surg Am. 1995;77:459–474.
23. Mont MA, Seyler TM, Plate JF, Delanois RE, Parvizi J. Unce-
mented total hip arthroplasty in young adults with osteonecrosis
of the femoral head: a comparative study. J Bone Joint Surg Am.
2006;88S:104–109.
24. Ortiguera CJ, Pulliam IT, Cabanela ME. Total hip arthroplasty
for osteonecrosis: matched-pair analysis of 188 hips with long-
term follow-up. J Arthroplasty. 1999;14:21–28.
25. Piston RW, Engh CA, De Carvalho PI, Suthers K. Osteonecrosis
of the femoral head treated with total hip arthroplasty without
cement. J Bone Joint Surg Am. 1994;76:202–214.
26. Ranawat CS, Atkinson RE, Salvati EA, Wilson PD Jr. Conven-
tional total hip arthroplasty for degenerative joint disease in
patients between the ages of forty and sixty years. J Bone Joint
Surg Am. 1984;66:745–752.
27. Saito S, Ryu J, Watanabe M, Ishii T, Saigo K. Midterm results of
Metasul metal-on-metal total hip arthroplasty. J Arthroplasty.
2006;21:1105–1110.
28. Saito S, Saito M, Nishina T, Ohzono K, Ono K. Long-term results
of total hip arthroplasty for osteonecrosis of the femoral head. A
comparison with osteoarthritis. Clin Orthop Relat Res.
1989;244:198–207.
29. Santavirta S, Bohler M, Harris WH, Konttinen YT, Lappalainen
R, Muratoglu O, Rieker C, Salzer M. Alternative materials to
improve total hip replacement tribology. Acta Orthop Scand.
2003;74:380–388.
30. Schmalzried TP, Peters PC, Maurer BT, Bragdon CR, Harris WH.
Long-duration metal-on-metal total hip arthroplasties with low
wear of the articulating surfaces. J Arthroplasty. 1996;11:322–
331.
31. Sieber HP, Rieker CB, Kottig P. Analysis of 118 second-gener-
ation metal-on-metal retrieved hip implants. J Bone Joint Surg
Br. 1999;81:46–50.
32. Stulberg BN, Singer R, Goldner J, Stulberg J. Uncemented total
hip arthroplasty in osteonecrosis: a 2- to 10-year evaluation. Clin
Orthop Relat Res. 1997;334:116–123.
33. Udomkiat P, Wan Z, Dorr LD. Comparison of preoperative
radiographs and intraoperative findings of fixation of hemispheric
porous-coated sockets. J Bone Joint Surg Am. 2001;83:1865–
1870.
34. Usrey MM, Noble PC, Rudner LJ, Conditt MA, Birman MV,
Santore RF, Mathis KB. Does neck/liner impingement inc-
rease wear of ultrahigh-molecular-weight polyethylene liners?
J Arthroplasty. 2006;21:65–71.
35. Weber BG. Experience with the Metasul total hip bearing system.
Clin Orthop Relat Res. 1996;329S:69–77.
123
Clin Orthop Relat Res (2008) 466:1154–1161
DOI 10.1007/s11999-008-0194-7
ORIGINAL ARTICLE
Do Patient Expectations of Spinal Surgery Relate to Functional
Outcome?
Albert Yee MD, Nana Adjei BSc, Jennifer Do,
Michael Ford MD, Joel Finkelstein MD
Received: 27 June 2007 / Accepted: 15 February 2008 / Published online: 18 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract It is important for surgeons to understand
patients’ expectations for surgery. We asked whether
patient factors and preoperative functional outcome scores
reflect the degree of expectations patients have for pos-
terior spinal surgery. Second, we asked whether patients’
expectations for surgery predict improvements in func-
tional outcome scores after surgery. We prospectively
enrolled 155 consecutive surgical patients with greater than
90% followup. Patients’ expectations were evaluated pre-
operatively along with SF-36 and Oswestry disability
questionnaires. Postoperatively (6 months for decompres-
sion; 1 year for fusions), we quantified patient-derived
satisfaction regarding whether expectations were met and
by patient-derived functional outcome scores. In patients
undergoing decompression, gender, SF-36 general health
domain, and SF-36 physical component score predicted
patients with high expectations for surgery. Patients with
high expectations also reported greater postoperative
improvements in SF-36 role physical domain scores after
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
A. Yee (&), N. Adjei, J. Do, M. Ford, J. Finkelstein
The Spine Program, Division of Orthopaedic Surgery,
Sunnybrook Health Sciences Centre, 2075 Bayview Avenue,
Room MG 371-B, Toronto, ON, Canada M4N 3M5
e-mail: Albert.Yee@sunnybrook.ca
A. Yee, M. Ford, J. Finkelstein
Department of Surgery, University of Toronto, Toronto, ON,
Canada
123
surgery. Expectations for surgery were met in 81% of
patients. In a subset of patients (21 of 143), expectations
were not met. These patients reported lower mean preop-
erative SF-36 general health, vitality, and mean mental
component scores.
Level of Evidence: Level I, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Surgery for degenerative conditions of the lumbar spine is
a frequent indication for spinal surgery. The efficacy of
spinal surgical procedures and its potential impact on the
natural history of the degenerative process is under
increasing scrutiny [5, 6]. The rates of spinal surgery for
degenerative conditions have increased with time and our
general population continues to age [3]. Various factors can
influence patient outcome after lumbar spinal surgery that
may be important in guiding patient selection for surgery
[5, 6, 15].
Physical variables such as preoperative functional status
and medical comorbidity have influenced surgical outcome
[12, 15]. Psychologic variables reportedly correlate with
patient satisfaction after spine surgery [13]. A patient’s
preoperative rating of his or her health may be an important
predictor of symptom severity, walking capacity, and sat-
isfaction after lumbar decompression [13]. In one study of
patients undergoing total joint arthroplasty, patients with
high mental distress, such as anxiety and depression, before
surgery were more likely to have worse reported physical
outcomes than those with minimal or no mental distress [1].
What patients expect from spine surgery also appears to
influence outcome [4, 10, 18]. Patients’ satisfaction after
Volume 466, Number 5, May 2008
lumbar surgery may not always correlate with postoperative
physical functioning [22]. Patient gender and type of
surgery performed also may influence surgical outcome
[19, 20]. The association between patients’ expectation for
surgery and how that may relate to postoperative outcomes
has not been clearly delineated. Patients with high expec-
tations for surgery may have a greater result from surgery if
their expectations reflect their motivation. Alternately,
patients with high expectations for surgery may have
unrealistic expectations in what surgery can accomplish
and experience a poorer result from surgery.
The primary question was whether patient factors and
baseline preoperative functional outcome scores reflected
the degree of expectations that patients have for surgery.
Second, we asked whether patients’ expectations for sur-
gery predict patient-reported improvements in functional
outcome after surgery.
Materials and Methods
We prospectively followed 155 consecutive patients
undergoing posterior lumbar spinal surgery for degenera-
tive conditions of the lumbar spine between 1998 and 2002.
Indications for surgery included back, buttock, and/or
lower extremity pain of spondylogenic origin. We included
patients undergoing decompression and/or spondylodesis
(spinal fusion) regardless whether they had prior lumbar
surgery. Patients with spinal stenosis underwent decom-
pression and patients with spondylolisthesis underwent
decompression and fusion. We excluded 26 patients
because they were not capable of completing the ques-
tionnaires (eg, cognitive or language limitations) (n = 17)
or declined participation (n = 9). After the exclusions, 143
remaining patients were assessed with questionnaires
(SF-36, Oswestry Disability Index) preoperatively and
6 months (decompressions) and 1 year (fusions) after sur-
gery. The mean age of the 143 patients was 52 years
(range, 18–84 years; male:female ratio, 1:1) (Fig. 1). We
calculated preoperative patient comorbidity using the
Charlson Comorbidity Index [2]. Institutional review board
approval was obtained for the study.
Baseline patient and surgical demographics of patients
in the study included an average age at surgery of 52 years
(range, 18–84 years) and a male:female ratio of 1:1. Forty-
three percent of surgeries were performed for disc herniation,
9% for isthmic spondylolisthesis, 10% for degenerative
spondylolisthesis, 30% for stenosis, 6% for spondylosis,
1% for pseudarthrosis, and 1% for adjacent segment dis-
ease. Among the patients in whom decompression was
performed (n = 94), single-level decompressions were
performed in 70%, two-level decompressions in 21%, and
multilevel in 9%. Thirty-four percent involved spinal
Patient Expectations of Spinal Surgery 1155
fusion of which 75% were instrumented (Fig. 1). Among
the patients in whom fusions were performed (n = 49),
49% had single-level fusions and 51% had multilevel
fusions. Forty-two of 49 patients who had fusions also
underwent spinal decompression. Revision cases and
workers’ compensation or litigation cases accounted for
11% and 9%, respectively.
The mean Mental Component and Physical Component
scores preoperatively were 42.1 ± 1.2 and 22.3 ± 1.1,
respectively. These values are approximately one and three
standard deviations below age- and gender-matched
national norms. The mean preoperative Oswestry score was
48.7 ± 1.7%.
Surgery was performed by one of two spine fellowship-
trained surgeons (AY, JF). Our standard approach for a disc
herniation was a discectomy with partial medial facetec-
tomy and lateral recess decompression either unilaterally or
bilaterally according to patient symptoms. For fusion pro-
cedures, a posterior intertransverse process lumbar fusion
was performed using autogenous iliac crest bone. Instru-
mented procedures used a pedicle screw/rod system (USS,
Synthes, Inc, (Monument, CO) from 1998–2000 and Xia,
Stryker1, (Kalamazoo, MI) from 2000–2002).
Patients were permitted activity ad lib after surgery. No
lumbar orthoses were used postoperatively. A 6-week
course of active and passive physiotherapy was recom-
mended at the 6-week postoperative followup.
Patients were reviewed clinically and radiographically
(fusion procedures) at 6 weeks, 3 months, 6 months, and
1 year (fusion procedures) after surgery. Study personnel
not involved in care of the patients and blinded to outcome
status sent the patients questionnaires before surgery and
again 6 months and 1 year after surgery. In the cover letter,
patients were instructed to contact the study coordinator for
clarification needed for any of the survey questions.
Patients were requested to complete the preoperative
questionnaire during the week before surgery. Preopera-
tively, patients completed a generic health status measure
(SF-36) and a disease-specific questionnaire (Oswestry
Disability Index). In addition, patients completed an
expectations questionnaire asking them to rate their
expectations for surgery regarding relief of back and leg
pain, their ability to sleep, recreational and daily activities,
and return to work (Table 1). Postoperatively, we evalu-
ated patients’ expectations and compared them with
patient-derived functional outcome measures at 6 months
for decompressions and 1 year for procedures involving
spinal fusion. The postoperative expectation questionnaire
investigated the patients’ attitudes toward the outcome of
surgery as it related to meeting their initial expectations
(Table 2). Surgeons were blinded to the patients’ expec-
tations and functional outcome questionnaires until
completion of the study.
123
1156 Yee et al. Clinical Orthopaedics and Related Research

Fig. 1 This consort flow chart

depicts patient recruitment and
followup during the study period.

Table 1. Preoperative expectations questionnaire

Parameter Not at all Slightly Somewhat Very Extremely Not
likely likely likely likely likely applicable

Relief from back pain 1 2 3 4 5 6

Relief from leg pain 1 2 3 4 5 6
Relief from numbness, weakness, instability 1 2 3 4 5 6
To do more everyday household or yard activities 1 2 3 4 5 6
To sleep more comfortably 1 2 3 4 5 6
To go back to my usual job and normal activities 1 2 3 4 5 6
To exercise and do recreational activities 1 2 3 4 5 6

Patients’ responses to the expectation questionnaires
were used to define a priori the patients who had either high
or low preoperative expectations for surgery and to deter-
mine if surgery met preoperative expectations. We
assumed a high preoperative expectation for surgery if the
average of applicable items (Table 1) was greater than 3.5
(ie, majority of replies to items in Table 1 indicating either
a very likely or extremely likely response). Excluding
items in which patients indicated the particular question
stem was not applicable (ie, last column, Table 2), we
summed and averaged patient-specific raw scores for the
items that were applicable to the patient. We assumed a
score of 2.5 or less would constitute overall expectations
for surgery being met (ie, majority of items that were
applicable to the patient would be recorded as either defi-
nitely yes or probably yes as depicted in Table 2). We
considered 94 four patients to have lower preoperative
expectations and 51 to have higher preoperative expectations.

123
Volume 466, Number 5, May 2008 Patient Expectations of Spinal Surgery 1157

Table 2. Postoperative expectations questionnaire

Parameter Definitely yes Probably yes Not sure Probably not Definitely not Not applicable

Relief from back pain 1 2 3 4 5 6

Relief from leg pain 1 2 3 4 5 6
Relief from numbness, weakness, instability 1 2 3 4 5 6
To do more everyday household or yard activities 1 2 3 4 5 6
To sleep more comfortably 1 2 3 4 5 6
To go back to my usual job and normal activities 1 2 3 4 5 6
To exercise and do recreational activities 1 2 3 4 5 6
* Are you satisfied with your surgical result?; Are the results of your treatment what you expected? Yes No.

The distribution of patients with high versus low preoper- Results

ative expectations for surgery was similar between the two
surgical groups (decompression versus fusion). Several patient demographic and preoperative functional
We created two prediction models: (1) Patient age, gen- outcome variables reflected patients with high preoperative
der, comorbidity, and preoperative functional outcome expectations for surgery. In the subgroup analysis of
scores were evaluated to determine if these variables patients with decompressions (n = 94), gender, SF-36
reflected patients who had either high or low expectations for general health domain, and SF-36 Physical Component
surgery; and (2) Degree of improvement from baseline scores reflected patients who had high preoperative
generic and disease-specific functional outcome scores after expectations for surgery (Table 3). Specifically, male
surgery was evaluated to determine if they could predict gender, better preoperative SF-36 general health domain,
patients who possessed greater expectations for surgery. We and poorer preoperative SF-36 Physical Component score
performed a post hoc analysis by the operating surgeon to reflected patients who possessed high expectations for
ensure surgery by one of the two surgeons did not influence surgery (Table 3). Remaining SF-36 domain and compo-
patients’ outcomes and expectations in our study. nent scores were not statistically significant. In the
Improvement in patient-derived functional outcome subgroup analysis of patients undergoing fusion (n = 49),
measures after surgery was analyzed by paired analysis patient age, gender, and preoperative functional outcome
comparing patient-specific preoperative with postoperative scores did not reflect higher patient expectations. We
SF-36 and Oswestry scores. After dichotomizing patients observed no difference in patient age (p = 0.18), comorbidity
into those with high or low preoperative expectations,
logistic regression analysis was performed. Because of the Table 3. Patient parameters and baseline functional outcome scores
differences in length of followup according to surgery type as predictors
(decompression versus fusion), subgroup analysis was Parameter Odds 95% Confidence p Value*
performed separately for these two procedures. In the first ratio interval
prediction model on preoperative data, factors including
Age 1.08 0.98–1.19 0.11
patient age, gender, comorbidity, and preoperative func-
Gender 0.18 0.03–1.00 0.05*
tional scores (SF-36 domain scores, SF-36 Mental
Comorbidity 1.01 0.97–1.05 0.75
Component score, SF-36 Physical Component score,
SF-36 Physical functioning 1.03 0.98–1.09 0.20
Oswestry Disability Index) were evaluated to determine if domain
these variables reflected the likelihood of patients having SF-36 Role physical domain 1.01 0.99–1.04 0.34
high preoperative expectations for surgery. The second
SF-36 Bodily pain 1.04 0.99–1.10 0.12
prediction model focused on postoperative analysis. We
SF-36 General health 1.04 1.01–1.08 0.02*
calculated the improvements each patient had after surgery
SF-36 Vitality 1.04 1.00–1.09 0.06
for the eight domains of the SF-36, mental and physical
SF-36 Social functioning 1.01 0.97–1.06 0.58
components, and Oswestry Disability Index scores. We
SF-36 Role emotional 1.01 0.97–1.04 0.74
then determined if improvements to these variables pre-
SF-36 Mental health 1.05 0.97–1.13 0.24
dicted the likelihood of patients having had high
SF-36 Mental Component score 0.81 0.63–1.04 0.10
expectations for surgery. Analysis was performed with and
SF-36 Physical Component score 0.75 0.59–0.96 0.02*
without inclusion of patients who are known to have poor
Oswestry Disability Index 1.00 0.96–1.04 0.98
prognostic factors for outcome (workers’ compensation/
litigation, revision surgical cases). * p \ 0.05.

123
1158 Yee et al. Clinical Orthopaedics and Related Research

Table 4. Results of SF-36 and Oswestry Disability Index scores

Parameter Preoperative Postoperative
± SD ± SD

SF-36 Physical functioning domain 30.0 ± 2.1 61.6 ± 2.4

SF-36 Role physical domain 10.8 ± 2.0 50.7 ± 3.5
SF-36 Bodily pain domain 23.8 ± 1.5 57.9 ± 2.2
SF-36 General health domain 58.4 ± 2.3 67.9 ± 2.1
SF-36 Vitality domain 38.3 ± 1.8 56.3 ± 1.8
SF-36 Social functioning domain 40.7 ± 2.3 73.3 ± 2.4
SF-36 Role emotional domain 45.2 ± 3.8 69.6 ± 3.4
SF-36 Mental health domain 61.9 ± 1.9 72.8 ± 1.5
SF-36 Mental Component score 42.1 ± 1.2 48.1 ± 1.1
SF-36 Physical Component score 22.3 ± 1.1 38.6 ± 1.2
Oswestry Disability Index 48.7 ± 1.7 23.1 ± 1.9
(percent disability)
SD = standard deviation.

Fig. 2 Patients generally had higher expectations for relief of leg

(p = 0.5), distribution by revision (p = 0.6), or workers’
versus back pain and for improvements in sleep and return to
compensation/litigation cases (p = 0.4) comparing de- household and recreational activities. There appeared to be a lower
compressive versus fusion groups. There were more expectation for return to work-related activities.
(p = 0.02) males in the group having decompressions and
patients undergoing decompression recorded higher SF-36
physical function (p = 0.04) and general health domain
(p = 0.03) scores when compared with patients undergo-
ing fusion preoperatively. In both groups, significant
improvements in functional scores were observed after
surgery (Table 4).
In our second prediction model, higher preoperative
expectations predicted greater improvement in some but
not all functional outcome measures after surgery. Patients
with higher preoperative expectations had greater
improvement in mean SF-36 role physical domain scores
(OR, 3.7; 95% CI, 1.59–8.58; p = 0.002). In the majority
of patients expectations for surgery were met (Figs. 2, 3).
Overall patient satisfaction with surgery as quantified in
our postoperative questionnaire was 81% (116 of 143). In
general, patients had a higher expectation for the relief of
leg versus back pain and a higher expectation for
improvements in sleep and return to household and recre- Fig. 3 The majority of patients believed their expectations for
surgery were met in the seven categories relating to symptoms and
ational activities (Fig. 2). Expectations for return to work- physical function.
related activities were lower than expectations for relief of
leg pain (Fig. 2). In 19% (27 of 143 patients) of patients,
surgery did not meet overall expectations. Of these 27 Component scores (p = 0.02, 0.01, and 0.04, respectively).
patients, there were two cases of pseudarthroses, one case Patients in whom expectations were not met also reported
of pedicle screw misplacement, and three additional cases less (p \ 0.05) improvement in SF-36 and Oswestry scores
in which medical comorbidities were believed to be con- when compared with patients in whom expectations were
tributing factors. met. Patients were likely to be less satisfied if they had
For the 21 of 27 patients whose expectations for surgery prior lumbar surgery (p = 0.02) and were involved with
were not met and no other poor prognostic factor could be workers’ compensation or litigation (p \ 0.001). They also
identified, we observed lower reported preoperative SF-36 were more likely (p = 0.004) to undergo additional spinal
general health, vitality domain scores, and mean Mental surgery.

123
Volume 466, Number 5, May 2008
Discussion
What patients expect from spine surgery is important as it
relates to patient satisfaction. Patients’ satisfaction after
lumbar surgery may not always correlate with postopera-
tive physical functioning [22]. We therefore asked whether
patient factors and baseline preoperative functional out-
come scores reflected the degree of expectations that
patients have for surgery. We then asked whether patient
expectations for surgery predict patient-reported improve-
ments in functional outcome after surgery. By determining
if patients’ expectations for surgery are met and whether
this translates into satisfaction after surgery is important to
our role as clinicians.
We note several limitations. Varying diagnostic degen-
erative conditions of the lumbar spine were evaluated. We
did perform a subgroup analysis in the current study
evaluating decompressive and fusion procedures sepa-
rately. Although the SF-36 includes a mental component to
the assessment, a formal psychometric evaluation was not
performed on the study cohort. This study evaluated the
short-term satisfaction and outcomes of patients undergo-
ing posterior lumbar spinal surgery for degenerative
conditions. At the last followup, despite improvement after
surgery, the mean SF-36 Physical Component score of our
patients was still on average one standard deviation below
age- and gender-matched national norms [8]. Longer-term
followup may be needed to determine any potential addi-
tional improvement after surgery and how this may
influence patients’ expectations [12, 14].
We observed several patient demographic and preop-
erative functional outcome variables reflecting patients
who reported higher preoperative expectations for sur-
gery. Male gender, greater perceived SF-36 general
health, and lower reported SF-36 Physical Component
scores were predictors of patients possessing greater
expectations for surgery in the decompressive group. In a
study of older patients than those in our study cohort,
Shabat et al. evaluated patients older than 65 years
undergoing decompression for lumbar spinal stenosis
[19]. They observed the satisfaction rate after surgery at
a minimum of 1-year followup as determined by tele-
phone interview was worse in women, although both
genders reported improvements in activities of daily
living and reduction in pain perception. Hakkinen et al.
also reported gender differences existed in baseline
Oswestry Disability Index scores in patients undergoing
surgery for lumbar disc herniation [7]. They observed
greater baseline disability in walking, sex life, social life,
and traveling items of the Oswestry Disability Index in
women. In our study, a lower baseline SF-36 Physical
Component score was predictive of greater expectations
for surgery. When considering an outcome measure that
Patient Expectations of Spinal Surgery 1159
gauges a patient’s perception of their overall function
from a physical perspective, the issue of chronicity of
clinical course may be an important variable in addition
to the absolute magnitude of that function that a patient
reports before surgery. Because current indications for
surgery are primarily for pain and function, patients
reporting lower function physically may have more to
gain by surgical intervention and therefore have greater
expectations for their surgical results.
Patients with high expectations also reported greater
improvements in SF-36 role physical domain after surgery,
but not greater improvements in other SF-36 domains,
component scores, and Oswestry Disability Index. Longer
patient followup and/or a larger sample size may be nec-
essary to more accurately define parameters in which
expectations impact the magnitude of improvements in
functional outcomes. We did observe that patients’
expectations for surgery were met in the majority of
patients (81%). We also observed that when expectations
for surgery were met, greater patient satisfaction and
improvements in functional outcome were reported. Spinal
surgery for degenerative conditions of the lumbar spine are
more likely to be successful in meeting patients’ expecta-
tions for relief of radicular leg symptoms when compared
with low back pain [13]. Revision lumbar surgery, workers’
compensation, and litigation predicted a poorer functional
outcome; however, these cases accounted for a small pro-
portion of our study cohort. Katz et al. studied the
prognostic importance of patients’ assessment of their
health and comorbidity in outcome after surgery in
degenerative lumbar spinal stenosis [15]. They observed
that patients who perceived their general health to be poor
were less likely to show substantial improvement after
surgery when compared with patients who perceived their
general health to be good. In our subgroup analysis of
patients undergoing decompression, we also observed that
patients who reported a higher SF-36 general health
domain score also tended to have greater expectations for
surgery. Gepstein et al. [4] reported the expectations rela-
ted to satisfaction and preoperative expectations could aid
in predicting postoperative satisfaction in elderly patients
with lumbar spinal stenosis. However, in their study, pre-
operative expectation was assessed by one question
regarding either low or high expectations for successful
surgical treatment when compared with our study evalu-
ating several parameters of pain, physical, and daily
function relating to expectations [4]. A patient’s motivation
appears important in that patients with high expectations
for surgery may more likely report improved function and
satisfaction after surgery [10, 17].
The notion of meeting patients’ expectations for treat-
ment is important in their perception of satisfaction
[10, 11, 16]. We observed concordance in that patients who
123
1160 Yee et al.
believed their expectations for surgery were not met also
reported less improvement as quantified by functional
outcome measures. Although surgery type (fusion versus
decompression) was not associated with the probability of
meeting expectations after surgery in our study, the study
by Toyone et al. of 98 patients with decompression sug-
gested patients who were undergoing surgery for stenosis
were more likely to have unrealistic expectations when
compared with patients undergoing surgery for disc her-
niation [20]. Overall satisfaction with spinal surgery was
71% to 86% in their series, which is consistent with the
results of our study.
An important consideration in patient satisfaction and
expectations for surgery is the surgeon-patient discussion
regarding the role of surgery for the patient’s condition.
Although there is the potential for surgeon bias in edu-
cating what patients may expect from surgery, we used a
standard discussion between the operating surgeon and the
patient regarding the role and expected benefits of surgery.
Preoperative information was provided by our two
surgeons, including a standardized patient-oriented infor-
mation package for either decompressive or fusion surgery.
The provision of information whereby leg pain is managed
more satisfactorily than back pain may prebias patients’
expectations for outcome [13]. What patients retain from
preoperative instructions and surgical counseling in
informed consent is generally poor [9, 21]. Patients’
expectations for surgery is a multifaceted issue, and patient
characteristics and surgical counseling regarding what
patients expect from surgery influence outcome. Although
we have dichotomized patients into those with high versus
low expectations for surgery based on their responses to
aspects of physical function and symptoms, what consti-
tutes unrealistic expectations patients may have for surgery
and conversely surgical counseling that is appropriate to
guide realistic expectations for surgery requires additional
study.
Posterior lumbar spinal surgery for degenerative condi-
tions can assist in improving patients’ symptoms and
functioning. The expectations for surgery are met in the
majority of patients. Patients’ preoperative perception of
general health, vitality, and mental health was worse in
patients in whom surgery did not meet expectations. In
patients undergoing decompressive surgery, gender and
some baseline functional measures including the SF-36
physical component score could reflect patients who were
more likely to possess a higher degree of expectation for
their surgical result. A higher degree of expectation for
surgery modestly predicted greater improvement in some
but not all functional outcome measures after surgery.
Functional outcome measures may be of value in guiding
patient selection and education regarding potential surgical
outcomes.
123
Clinical Orthopaedics and Related Research
Acknowledgments We appreciate the contributions of Christine Le,
Emi Do, Dawn Barbieri, Dr. Mubarak Al-gahtany, and Lynn Antwi-
Boasiako to this study, and we thank Dr. Marjan Vidmar for assis-
tance with the statistical analysis.
References
1. Ayers DC, Franklin PD, Trief PM, Ploutz-Snyder R, Freund D.
Psychological attributes of preoperative total joint replacement
patients: implications for optimal physical outcome. J Arthro-
plasty. 2004;19(7 suppl 2):125–130.
2. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of
a combined comorbidity index. J Clin Epidemiol. 1994;47:
1245–1251.
3. Ciol MA, Deyo RA, Howell E, Kreif S. An assessment of surgery
for spinal stenosis: time trends, geographic variations, compli-
cations, and reoperations. J Am Geriatr Soc. 1996;44:285–290.
4. Gepstein R, Arinzon Z, Adunsky A, Folman Y. Decompression
surgery for lumbar spinal stenosis in the elderly: preoperative
expectations and postoperative satisfaction. Spinal Cord.
2006;44:427–431.
5. Gibson JN, Grant IC, Waddell G. Surgery for lumbar disc pro-
lapse. Cochrane Database Syst Rev. 2000(3):CD001350.
6. Gibson JN, Waddell G, Grant IC. Surgery for degenerative
lumbar spondylosis. Cochrane Database Syst Rev. 2000(3):
CD001352.
7. Häkkinen A, Kautiainen H, Järvenpää S, Arkela-Kautiainen M,
Ylinen J. Changes in the total Oswestry Index and its ten items in
females and males pre- and post-surgery for lumbar disc herni-
ation: a 1-year follow-up. Eur Spine J. 2007;16:347–352.
8. Hopman WM, Towheed T, Anastassiades T, Tenenhouse A,
Poliquin S, Berger C, Joseph L, Brown JP, Murray TM, Adachi
JD, Hanley DA, Papadimitroupolos E. Canadian normative data
for the SF-36 health survey. Canadian Multicentre Osteoporosis
Study Research Group. CMAJ. 2000;163:265–271.
9. Hutson MM, Blaha JD. Patients’ recall of preoperative instruction
for informed consent for an operation. J Bone Joint Surg Am.
1991;73:160–162.
10. Iversen MD, Daltroy LH, Fossel AH, Katz JN. The prognostic
importance of patient pre-operative expectations of surgery for
lumbar spinal stenosis. Patient Educ Couns. 1998;34:169–178.
11. Jackson JL, Kroenke K. The effect of unmet expectations among
adults presenting with physical symptoms. Ann Intern Med.
2001;134(9 pt 2):889–897.
12. Jönsson B, Annertz M, Sjöberg C, Strömqvist B. A prospective,
consecutive study of surgically treated lumbar spinal stenosis.
Part II: Five-year follow-up by an independent observer. Spine.
1997;22:2938–2944.
13. Katz JN, Lipson SJ, Brick GW, Grobler LJ, Weinstein JN, Fossel
AH, Lew RA, Liang MH. Clinical correlates of patient satisfac-
tion after laminectomy for degenerative lumbar spinal stenosis.
Spine. 1995;20:1155–1160.
14. Katz JN, Lipson SJ, Chang LC, Levine SA, Fossel AH, Liang
MH. Seven- to 10-year outcome of decompressive surgery for
degenerative lumbar spinal stenosis. Spine. 1996;21:92–98.
15. Katz JN, Stucki G, Lipson SJ, Fossel AH, Grobler LJ, Weinstein
JN. Predictors of surgical outcome in degenerative lumbar spinal
stenosis. Spine. 1999;24:2229–2233.
16. Lin CT, Albertson GA, Schilling LM, Cyran EM, Anderson SN,
Ware L, Anderson RJ. Is patients’ perception of time spent with
the physician a determinant of ambulatory patient satisfaction?
Arch Intern Med. 2001;161:1437–1442.
17. Mahomed NN, Liang MH, Cook EF, Daltroy LH, Fortin PR,
Fossel AH, Katz JN. The importance of patient expectations in
Volume 466, Number 5, May 2008
predicting functional outcomes after total joint arthroplasty.
J Rheumatol. 2002;29:1273–1279.
18. Rönnberg K, Lind B, Zoëga B, Halldin K, Gellerstedt M, Brisby
H. Patients’ satisfaction with provided care/information and
expectations on clinical outcome after lumbar disc herniation
surgery. Spine. 2007;32:256–261.
19. Shabat S, Folman Y, Arinzon Z, Adunsky A, Catz A, Gepstein R.
Gender differences as an influence on patients’ satisfaction rates in
spinal surgery of elderly patients. Eur Spine J. 2005;14:1027–1032.
Patient Expectations of Spinal Surgery 1161
20. Toyone T, Tanaka T, Kato D, Kaneyama R, Otsuka M. Patients’
expectations and satisfaction in lumbar spine surgery. Spine.
2005;30:2689–2694.
21. Turner P, Williams C. Informed consent: patients listen and read,
but what information do they retain? N Z Med J. 2002;115:U218.
22. Yamashita K, Hayashi J, Ohzono K, Hiroshima K. Correlation of
patient satisfaction with symptom severity and walking ability
after surgical treatment for degenerative lumbar spinal stenosis.
Spine. 2003;28:2477–2481.
123
Clin Orthop Relat Res (2008) 466:1162–1168
DOI 10.1007/s11999-008-0129-3
ORIGINAL ARTICLE
Is the Long-term Outcome of Cemented THA Jeopardized
by Patients Being Overweight?
Daniël Haverkamp MD, PhD, F. Harald R. de Man MD,
Pieter T. de Jong MD, Renée A. van Stralen MSc,
René K. Marti MD, PhD
Received: 26 June 2007 / Accepted: 9 January 2008 / Published online: 21 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Although the effect of being overweight on the
long- and short-term outcome of THA remains unclear, the
majority of orthopaedic surgeons believe being overweight
negatively influences the longevity of a hip implant. We
asked whether complications and long-term survival of
cemented THA differed in overweight patients (body mass
index [BMI] [ 25 kg/m2) and obese patients (BMI [ 30
kg/m2) compared with normal-weight patients (BMI \ 25
kg/m2). We retrospectively analyzed 411 consecutive
patients (489 THAs) treated with cemented THA between
1974 and 1993. Except for cardiovascular comorbidity, we
observed no differences in demographics among these
weight groups. We found no differences in the number of
intraoperative or postoperative complications. The survival
rates for the three BMI groups were similar. The 10-year
survival for any revision was 94.9% (95% confidence
interval, 91.6%–98.2%), 90.4% (95% confidence interval,
85.6%–95.2%), and 91% (95% confidence interval, 81.2%–
100%) for normal-weight, overweight, and obese patients,
respectively. Cox regression analysis showed BMI and
weight had no major influence on survival rates. The
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution either has waived or
does not require approval for the human protocol for this investigation
and that all investigations were conducted in conformity with ethical
principles of research.
D. Haverkamp (&), F. H. R. de Man, P. T. de Jong,
R. A. van Stralen, R. K. Marti
Department of Orthopedic Surgery G4-No, Academic Medical
Centre Amsterdam, Orthopaedic Research Centre Amsterdam,
PO Box 22660, 1100DD Amsterdam, The Netherlands
e-mail: D.Haverkamp@osteotomie.nl
123
differences in mean Harris hip score at final followup were
4.8 between normal-weight and overweight patients and
7.1 between normal-weight and obese patients. Being
overweight and obesity had no influence on perioperative
complication rates in this cohort and did not negatively
influence the long-term survival of cemented THA.
Level of Evidence: Level III, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Whether being overweight influences the fate of a THA is
still debated. One study suggests obese patients are more
likely to undergo THA for osteoarthritis (OA) of the hip
than control patients with lower body mass index (BMI)
[7]. Therefore, it is important for the orthopaedic surgeon
who is planning the joint arthroplasty to know the effect of
obesity on the fate of THA [7, 11, 22]. Although being
overweight or obese have a negative influence on health
and mobility, it is not certain whether they have a negative
influence on the short- and long-term results after THA as
well [4, 6, 19, 21].
The assumption that being overweight or obese neg-
atively influences the long-term survival of THA could
preclude some obese patients from having joint arthro-
plasty. Recently, the Wall Street Journal mentioned more
orthopaedic surgeons refuse to perform THA in obese
patients because of the fear of complications [15]. A
large international survey of orthopaedic surgeons per-
formed in 12 European countries revealed 80.9% believe
the long-term outcome of THA is impaired by being
overweight [20]. Several short-term outcome studies,
summarized in two reviews [4, 19], however, failed to
Volume 466, Number 5, May 2008 Long-term Effect of Obesity on Cemented THA 1163

show a negative influence of obesity on the short-term
results of THA.
We asked whether obesity influences the long-term
survival, clinical outcomes scores, and perioperative com-
plication rates. We also asked whether BMI and body
weight were risk factors for revision.
Materials and Methods
We retrospectively reviewed the medical records of 411
consecutive patients (489 hips) who underwent primary
THA between 1974 and 1993. We divided our patients into
three groups based on body mass index (BMI) at the time
of surgery: (1) patients with a normal body weight
(BMI \ 25 kg/m2); (2) patients who were overweight
(BMI [ 25 kg/m2); and (3) patients who were morbidly
obese (BMI [ 30 kg/m2). One hundred sixty-three patients
(201 hips [41%]) had a normal body weight. One hundred
forty-two patients (172 hips [35%]) had a BMI greater than
25 kg/m2 and 35 (42 hips [9%]) of these patients had a
BMI greater than 30 kg/m2. For 106 patients (116 hips
[24%]), no BMI (weight and/or height) was documented
preoperatively. To avoid selection bias, these patients were
included in the overall (survival) analysis. During fol-
lowup, 164 patients (184 hips) died after a minimum
followup of 1 year (mean, 11.6 years; range, 1–29.3 years)
and an additional 37 patients (50 hips) were lost to fol-
lowup after a minimum followup of 0.1 year (mean,
6.8 years; range, 0.1–15.6 years). These patients are
included in the survival analysis and radiographic analysis
until their last outpatient clinic contact. Of these patients
lost to followup, two had a BMI greater than 30 kg/m2,
eight had a BMI greater than 25 kg/m2, and 12 had a
normal BMI; for 16 patients, no BMI was documented.
Sample size power analysis was performed assuming a
10-year survival rate of 95% in normal-weight individuals.
We assumed a difference of 10% survival rate in over-
weight patients was of clinical importance. When using a
power of 0.8 and an alpha of 0.05, a sample size of 159
hips is needed per group. Our number of patients with a
BMI greater or less than 25 kg/m2 therefore seems
sufficient.
For maximum followup, 210 patients (255 hips) were
available. The minimum followup in these 210 patients was
10 years (mean, 14.9 years; range, 10–28.1 years). We
then compared long-term survivorship, functional outcome,
and perioperative complication rate. The average age at the
time of surgery was 67 years (range, 22–88 years). One
hundred seventeen (24%) of these patients were male
(Table 1). The indication for THA was idiopathic OA in
235 hips (48%), acetabular dysplasia in 165 hips (34%),
rheumatoid arthritis in eight (2%), avascular necrosis in 30
(6%), posttraumatic in 23 (5%), and other causes in 28
(4%). Apart from cardiologic comorbidity, which occurred
more often in overweight and obese patients (Fisher’s exact
test, p = 0.028 for BMI [ 30 kg/m2 versus BMI \ 30 kg/m2
and p = 0.044 for BMI [ 25 kg/m2 versus BMI \ 25
kg/m2), we observed no differences between the patients
who were obese or overweight and the normal-weight
patients (Table 1). The average BMI of all patients was
25.3 kg/m2 (range, 17.9–41.1 kg/m2).
The same prosthetic implant and surgical procedure
were used in all patients. All patients were placed in a
supine position and all had an anterolateral approach and a
cemented Weber Rotation THA System (Allopro, Baar,
Switzerland) implanted [5]. This system consists of a
wrought CoNiCrMo alloy stem (Protasul1 10; Sulzer AG,
Winterthur, Switzerland) with a cylindrical neck (the
trunnion) made of a cast CoCrMo alloy (Protasul1 2)
composite welded to the stem, which is grit-blasted with
glass particles. The 32-mm head was made from Protasul1
2 or Al2O3 ceramic (Biolox1; Feldmühle, Plochingen,
Germany) and placed on a Protasul1 2 cylinder. The stem

Table 1. Demographic data per BMI group shown in number and percentage
Demographics BMI \ 25 kg/m2 BMI [ 25 kg/m2 p Value* BMI [ 30 kg/m2 p Value*
and comorbidity (n = 201 hips) (n = 172 hips) (n = 42 hips)

Age (years)à 65.0 (21–83) 65.7 (22–87) 0.50 64.0 (49–79) 0.56
Percent idiopathic osteoarthritis 90 (44.8%) 82 (42.7%) 0.46 23 (54.8%) 0.22
Female 152 (75.6%) 134 (69.8%) 0.63 30 (71.4%) 0.84
Comorbidity
Central nervous system 14 (7.0%) 17 (8.9%) 0.35 5 (11.9%) 0.33
Respiratory 11 (5.5%) 10 (5.2%) 1.0 5 (11.9%) 0.16
Cardiovascular§ 43 (21.4%) 51 (26.6%) 0.07 16 (38.1%) 0.03
Diabetes 8 (4.0%) 9 (4.7%) 0.62 4 (9.5%) 0.12
* p values show comparison with the group with a BMI of less than 25 kg/m2; àage is given as an average, with range in parentheses; age was
compared using a t test; §cardiovascular comorbidity is higher (p \ 0.05) in the group with a BMI of greater than 30 kg/m2; for all the other
demographic data, no differences were found using a Fisher’s exact test; BMI = body mass index.

123
1164 Haverkamp et al.
and the nonhighly crosslinked polyethylene Weber socket
were cemented using low-viscosity Sulfix1 (Sulzer AG)
cement. Until the 1980s, we used two types of cups, a flat
type and a hemispheric type. Because of the inferior results
of the flat type, their use was discontinued. In this study,
112 flat type and 377 hemispheric type sockets were used.
The percentages of flat cups used were not different among
the weight groups.
We (DH, RKM, FHRdM) obtained Harris hip scores
(HHS) for patients whose THA was not revised at final
followup.
We (DH, FHRdM) performed a radiographic analysis
using the weightbearing pelvic and lateral radiographs
taken at the latest followup. Loosening of the stem was
ranked according to Harris et al. [8] and loosening of the
cup according to Hodgkinson et al. [9]. For both compo-
nents, loosening was scored as definitive, probable,
possible, or no loosening. Loosening was scored by com-
paring the radiographs at last followup with previous
radiographs.
Complications were retrieved from the clinical charts.
We noted the presence of hematoma when patients under-
went exploratory surgery for suspected hematoma. Early
infection was defined as requiring antibiotic treatment and/
or débridement within 3 months after the operation.
A survival analysis was performed using the Life Table
Method using revision for aseptic loosening, revision for
any reason, and radiographic loosening (definitive loosen-
ing) as end points. We performed survivorship analysis for
the acetabular and femoral component separately and for
both components combined. Because all patients were seen
annually or biannually, all could be included in the survival
analysis until their last followup. Equality of the survival
curves for the normal-weight, overweight, and obese
patients were compared using a log rank test. Differences
in HHS among the three study groups were evaluated using
analysis of variance. A difference greater than 4 points was
considered clinically important [10]. We also compared
BMI as a continuous variable with the HHS at maximum
followup by means of Pearson correlation analysis to
explore the overall influence of BMI on outcome. Differ-
ences in loosening between the normal-weight, overweight,
and obese patients were evaluated using Fisher’s exact test.
Differences in perioperative and postoperative complica-
tions were compared using Fisher’s exact test. Cox
regression analysis was performed for survival of the
implant (any revision) with weight and BMI as risk factors.
Results
We observed no differences between the survival rates for
normal-weight patients and overweight patients and
123
Clinical Orthopaedics and Related Research
morbidly obese and normal-weight patients for all end
points using a log rank test (Table 2; Fig. 1). Fifty four
patients (64 hips) underwent revision surgery, of which five
hips were revised for septic loosening, 54 for aseptic
loosening of at least one of the components, and five for
other reasons (periprosthetic fractures and heterotopic
ossifications). The rate of infection causing septic loosening
was similar in patients with a BMI of between 25 kg/m2 and
30 kg/m2 (n = 4) and with a normal body weight (n = 1)
(p = 0.13).
Patients with a BMI greater than 30 kg/m2 had lower
(p = 0.02) HHS than patients with a BMI less than 25 kg/m2
and patients with a BMI greater than 25 kg/m2 had
lower (p = 0.02) HHS than patients with a BMI less than
25 kg/m2 (Table 3). The differences in average HHS
between the three groups were greater than 4 points,
indicating these differences were clinically relevant. Body
mass index showed a poor correlation (rho = -0.17;
p = 0.024) with HHS.
Several local and systemic complications occurred,
which were similarly distributed among the normal-weight,
overweight, and obese patients (Table 4). We observed no
differences in the rates of radiographic loosening among
the normal-weight versus overweight patients (p = 0.30)
and normal-weight versus obese patients (p = 0.47)
(Table 5).
Body mass index and body weight were not risk factors
for revision (Exp[B] = 1.00 [95% confidence interval,
0.93–1.08] and Exp[B] = 1.01 [95% confidence interval,
0.99–1.03], respectively).
Discussion
The influence of being overweight on the long- and short-
term outcome of THA is controversial in the literature but
the majority of orthopaedic surgeons believe being over-
weight negatively influences the longevity of a hip implant
[20]. Because the issue is controversial, we asked whether
obesity influences the long-term survival, clinical outcomes
scores, and perioperative complication rates, and whether
BMI and body weight were risk factors for revision.
We note several limitations of our study. First, we did
not study wear. It could be hypothesized that more body
weight causes more wear. Although it can be expected that
excessive wear may influence the rate of revision, we did
not see a difference in revision rates between the weight
groups [2]. Second, we studied only patients with cemented
THA. Our analysis may not be valid for uncemented THA.
One study of 300 patients with the cementless PM pros-
thesis suggested obesity negatively influenced medium-
term survival, showing a twofold increase in loosening/
revision rate in obese patients [6]. Another recent study
Volume 466, Number 5, May 2008 Long-term Effect of Obesity on Cemented THA 1165

Table 2. Survival rates

Number at risk and revisions All BMI \ 25 kg/m2 BMI [ 25 kg/m2 BMI [ 30 kg/m2

Number at risk
At start 489 201 172 42
At 10 years 336 161 122 30
At 15 years 181 92 69 14
At 20 years 49 29 17 4
Any revision
At 10 years 92.4 (89.8–95.0) 94.9 (91.6–98.2) 90.4 (85.6–95.2) 91.0 (81.2–100)
At 15 years 83.7 (79.4–88.0) 85.9 (80.0–91.8) 83.1 (76.2–90.0) 79.5 (61.5–97.4)
At 20 years 72.6 (64.5–96.4) 75.6 (65.5–85.6) 68.3 (53.0–83.6) 79.5 (61.5–97.4)
Aseptic stem loosening
At 10 years 95.1 (92.9–97.2) 96.6 (93.9–99.3) 94.2 (90.3–98.1) 91.0 (81.2–100)
At 15 years 89.3 (85.7–92.9) 91.4 (86.6–96.2) 87.5 (81.1–93.9) 79.5 (61.5–97.4)
At 20 years 84.1 (78.1–90.0) 85.2 (76.4–94.0) 82.7 (73.8–91.6) 79.5 (61.5–97.4)
Aseptic cup loosening
At 10 years 96.9 (95.1–98.6) 97.7 (94.4–100) 97.2 (94.5–99.9) 97.1 (91.4–100)
At 15 years 90.0 (86.4–93.5) 89.6 (84.3–94.9) 91.5 (85.9–97.1) 84.9 (67.5–100)
At 20 years 79.9 (72.3–98.5) 79.4 (69.7–89.0) 80.0 (66.4–93.6) 84.9 (67.5–100)
Aseptic loosening, both components
At 10 years 94.0 (91.6–96.4) 96.0 (93.1–98.9) 92.8 (88.5–97.1) 91.0 (81.2–100)
At 15 years 85.9 (81.8–90.0) 86.7 (80.8–92.6) 86.1 (79.4–93.0) 79.5 (61.5–97.4)
At 20 years 74.5 (66.3–82.7) 85.2 (76.4–94.0) 70.8 (55.1–86.5) 79.5 (61.5–97.4)
Radiographic stem loosening
At 10 years 94.9 (92.7–97.1) 96.5 (93.7–99.3) 94.0 (89.9–98.1) 91.0 (81.2–100)
At 15 years 88.9 (85.1–92.7) 91.1 (86.1–96.1) 86.7 (79.8–93.6) 78.7 (59.8–97.6)
At 20 years 78.1 (70.7–85.5) 78.9 (68.3–98.5) 77.5 (65.8–98.5) 63.0 (31.5–98.5)
Radiographic cup loosening
At 10 years 96.8 (95.0–98.6) 97.6 (95.2–99.9) 97.1 (94.3–99.9) 97.1 (91.4–100)
At 15 years 89.3 (85.5–93.1) 84.9 (79.4–90.4) 90.8 (84.7–96.9) 84.2 (65.6–100)
At 20 years 76.6 (68.5–84.7) 76.2 (65.8–86.6) 76.1 (61.6–98.5) 67.3 (34.2–100)
Radiographic loosening, both components
At 10 years 93.4 (91.0–95.8) 95.9 (93.5–98.3) 92.6 (88.2–97.4) 91.0 (81.2–100)
At 15 years 85.1 (80.8–89.4) 85.6 (81.3–89.9) 85.4 (78.4–92.4) 78.7 (59.8–97.6)
At 20 years 67.4 (58.4–76.4) 69.6 (60.6–78.6) 63.5 (46.9–80.1) 50.0 (16.1–83.8)
Values are expressed as percentages, with 95% confidence intervals in parentheses; BMI = body mass index.

suggested no difference in the outcome of uncemented
THA in obese versus normal-weight patients, although a
high revision rate for the acetabular component was present
[13].
Our data suggest BMI and weight do not influence the
long-term survival of cemented THA. We also found no
differences in the incidence of THA-related complications
for the overweight patients undergoing THA. Cardiovas-
cular comorbidity was more common in the obese patients;
however, we observed no differences in perioperative
cardiac complications.
The percentage of overweight and obese individuals in
our study is lower than those reported in American studies.
In a study including 1071 American patients undergoing
THA, 36% of the patients had a BMI greater than 30 kg/m2
[16]. In The Netherlands, the annual incidence of obesity
(BMI [ 30 kg/m2) gradually inclined from 5% in 1981 to
7% in 1993 and 10% in 2005 [3]. In our study, 9% had a
BMI greater than 30 kg/m2. For the overweight patients
(BMI [ 25 kg/m2), these percentages were 33% in 1981
and 37% in 1993 and 35% in our study. Because OA is
more common in overweight patients, we believe these
percentages indicate our patient group is comparable to the
average Dutch population [7]. This also indicates absence
of a selection bias. All patients were operated on in our
hospital regardless of their weight. Another major

123
1166 Haverkamp et al. Clinical Orthopaedics and Related Research

Fig. 1A–C Survival rates are shown for patients with (A) a BMI less than 25 kg/m2, (B) a BMI greater than 25 kg/m2, and (C) a BMI greater
than 30 kg/m2. The x-axis shows years and the y-axis shows survival rates. The solid line represents survival rate and the dotted lines represent
the 95% confidence intervals.

Table 3. Average Harris hip score per BMI group
BMI \ 25 kg/m 2
BMI [ 25 kg/m 2
BMI [ 30 kg/m
91.6 (89.3–93.9) 86.8 (83.5–90.1)* 83.7 (74.5–92.3)*
Values are expressed as averages, with 95% confidence intervals
in parentheses; * difference with group with a BMI of less than
25 kg/m2 (p = 0.02); BMI = body mass index.
difference between our Dutch population and the American
population is extreme obesity (BMI [ 40 kg/m2) was low
in our country before 1993. We had only two patients who
had a BMI greater than 40 kg/m2 (neither had revision and
had HHS of 87 and 90). This low number of patients with a
BMI greater than 40 kg/m2 means our study does not
supply an answer for the long-term fate of THAs in these
extremes.
Several publications report on the short-term results of
THA in the obese in which the HHS after surgery are
compared between obese and normal-weight patients. The
literature contains controversial data suggesting either
similar or worse outcomes for obese patients undergoing
2 THA. Two large studies reported lower HHS in obese
patients after short-term followup [1, 14]. Both showed
lower HHS with an average difference of 5 points, but
neither compared the preoperative HHS among the differ-
ent groups. The clinical relevance of these small
differences in the postoperative HHS without a comparison
of the preoperative HHS is debatable, especially because
other studies showed no differences in postoperative HHS
between the several weight groups [18]. Another study
suggested the level of activity is lower, which continues to
be so after THA [12]. The same problem occurs in our
study because no preoperative HHS was available for
analysis. If patients who are more obese have initial lower
HHS and similar improvement as normal-weight patients
after the arthroplasty, the same difference remains.
Although our data suggest differences between the average
HHS in the weight groups, the differences between the
mean HHS were small (4.8 and 7.1). However; the only
study on the responsiveness and discriminative ability of

123
Volume 466, Number 5, May 2008 Long-term Effect of Obesity on Cemented THA 1167

Table 4. Complications per BMI group

Complication All (n = 489 hips) BMI \ 25 kg/m2 BMI [ 25 kg/m2 p Value* BMI [ 30 kg/m2 p Value*
(n = 201 hips) (n = 172 hips) (n = 42 hips)

Venous thromboembolism 2 (0.4%) 1 (0.5%) 0 0

Cardiac 6 (1.2%) 1 (0.5%) 4 (2.3%) 1.0 2 (4.8%) 1.0
Respiratory 3 (0.6%) 1 (0.5%) 1 (0.6%) 1.0 1 (2.4%) 0.31
Abdominal 4 (0.8%) 2 (1.0%) 1 (0.6%) 1.0 0
Other systemic (including urinary 18 (3.7%) 8 (4.0%) 9 (5.2%) 0.62 1 (2.4%) 1.0
tract infection)
Hematoma 10 (2.0%) 6 (3.0%) 2 (1.2%) 0.30 2 (4.8%) 0.63
Early infection 5 (0.8%) 2 (1.0%) 3 (1.7%) 0.67 0
Intraoperative complication 24 (4.9%) 11 (5.5%) 11 (6.4%) 0.67 2 (4.8%) 1.0
Any complication 68 (13.9%) 30 (14.9%) 28 (16.3%) 0.78 6 (14.3%) 1.0
* p values are given for the comparison with the group with a BMI of less than 25 kg/m2 (Fisher exact test); BMI = body mass index.

Table 5. Radiographic analysis of the unrevised hips

Component Definitive loosening Probable loosening Possible loosening
Number of hips Time until Number of hips Time until Number of hips Time until
loosening (years)* loosening (years)* loosening (years)*

Acetabular  2 18.3 (18.2–18.3) 2 16.2 (13.9–18.4) 15 15.9 (9.0–23.0)

à
Femoral 6 18.9 (15.9–22.6) 1 23.2 11 17.9 (14.0–22.8)
  à
* Values are expressed as averages, with ranges in parentheses; according to the criteria of Hodgkinson et al. [9]; according to the criteria of
Harris et al. [8].

the HHS showed a difference of 4 points is enough to be THA be necessary in an overweight or obese patient, the
clinically relevant, indicating our measured differences are arguments that survival is shorter in obese patients and that
clinically relevant [10]. However, the correlation of HHS obese patients have a higher risk of perioperative compli-
with BMI as a continuous variable was poor (rho = cations do not seem valid.
-0.17), but the content validity of the HHS is poor, eg, a
large ceiling effect is visible, which could influence the
correlation coefficient measured (Fig. 2).
In a review of patient characteristics affecting the
outcome of THA, a body weight greater than 70 kg was
mentioned as a factor that negatively influences the
outcome of THA [23]. They suggest weight alone is a
much stronger predictor for the outcome than BMI
because height has no influence on the prosthesis. In
our series, neither body weight nor BMI influenced
outcome.
One study stated patients who underwent bariatric sur-
gery before having THA had an excellent outcome,
although the average postoperative BMI of 29 kg/m2 still
indicated overweight. The main question we would ask is
whether the outcome would have been worse if no bariatric
surgery was performed [17].
We do not intend to suggest being overweight has no
risks. We believe it is important to motivate overweight
patients to lose weight. Being overweight could increase
the rate of OA and has an increased risk for several non- Fig. 2 A scatterplot shows HHS versus BMI. The ceiling effect of
orthopaedic morbidities [7]. However, should a (cemented) the HHS can be seen.

123
1168 Haverkamp et al.
References
1. Aderinto J, Brenkel IJ, Chan P. Weight change following total hip
replacement: a comparison of obese and non-obese patients.
Surgeon. 2005;3:269–272, 305.
2. Bordini B, Stea S, De Clerico M, Strazzari S, Sasdelli A, Toni A. Factors
affecting aseptic loosening of 4750 total hip arthroplasties: multivariate
survival analysis. BMC Musculoskelet Disord. 2007;8:69.
3. Central Bureau of Statistics, The Netherlands. Available at:
http://www.cbs.nl. Accessed January 2007.
4. Crawford RW, Murray DW. Total hip replacement: indications
for surgery and risk factors for failure. Ann Rheum Dis.
1997;56:455–457.
5. de Jong PT, van der Vis HM, de Man FH, Marti RK. Weber
rotation total hip replacement: a prospective 5- to 20-year fol-
lowup study. Clin Orthop Relat Res. 2004;419:107–114.
6. Dickob M, Martini T. The cementless PM hip arthroplasty:
four-to-seven-year results. J Bone Joint Surg Br. 1996;78:
195–199.
7. Flugsrud GB, Nordsletten L, Espehaug B, Havelin LI, Engeland
A, Meyer HE. The impact of body mass index on later total hip
arthroplasty for primary osteoarthritis: a cohort study in
1.2 million persons. Arthritis Rheum. 2006;54:802–807.
8. Harris WH, McCarthy JC Jr, O’Neill DA. Femoral component
loosening using contemporary techniques of femoral cement
fixation. J Bone Joint Surg Am. 1982;64:1063–1067.
9. Hodgkinson JP, Shelley P, Wroblewski BM. The correlation
between the roentgenographic appearance and operative find-
ings at the bone-cement junction of the socket in Charnley low
friction arthroplasties. Clin Orthop Relat Res. 1988;228:
105–109.
10. Hoeksma HL, van den Ende CH, Ronday HK, Heering A,
Breedveld FC. Comparison of the responsiveness of the Harris
Hip Score with generic measures for hip function in osteoarthritis
of the hip. Ann Rheum Dis. 2003;62:935–938.
11. Karlson EW, Mandl LA, Aweh GN, Sangha O, Liang MH,
Grodstein F. Total hip replacement due to osteoarthritis: the
importance of age, obesity, and other modifiable risk factors. Am
J Med. 2003;114:93–98.
123
Clinical Orthopaedics and Related Research
12. McClung CD, Zahiri CA, Higa JK, Amstutz HC, Schmalzried TP.
Relationship between body mass index and activity in hip or knee
arthroplasty patients. J Orthop Res. 2000;18:35–39.
13. McLaughlin JR, Lee KR. The outcome of total hip replacement in
obese and non-obese patients at 10- to 18-years. J Bone Joint
Surg Br. 2006;88:1286–1292.
14. Moran M, Walmsley P, Gray A, Brenkel IJ. Does body mass
index affect the early outcome of primary total hip arthroplasty?
J Arthroplasty. 2005;20:866–869.
15. Naik G. Surgeons’ weighty dilemma; wary of extra risk, work,
some doctors won’t replace knees, hips of obese patients. Wall
Street Journal. February 28, 2006:B1, B8.
16. Namba RS, Paxton L, Fithian DC, Stone ML. Obesity and peri-
operative morbidity in total hip and total knee arthroplasty
patients. J Arthroplasty. 2005;20(suppl 3):46–50.
17. Parvizi J, Trousdale RT, Sarr MG. Total joint arthroplasty in
patients surgically treated for morbid obesity. J Arthroplasty.
2000;15:1003–1008.
18. Stickles B, Phillips L, Brox WT, Owens B, Lanzer WL. Defining
the relationship between obesity and total joint arthroplasty. Obes
Res. 2001;9:219–223.
19. Stukenborg-Colsman C, Ostermeier S, Windhagen H. [What
effect does of obesity have on the outcome of total hip and knee
arthroplasty: review of the literature] [in German]. Orthopade.
2005;34:664–667.
20. Stürmer T, Günther KP, Brenner H. Obesity, overweight and
patterns of osteoarthritis: the Ulm Osteoarthritis Study. J Clin
Epidemiol. 2000;53:307–313.
21. Sutherland CJ, Wilde AH, Borden LS, Marks KE. A ten-year
follow-up of one hundred consecutive Muller curved-stem total
hip-replacement arthroplasties. J Bone Joint Surg Am.
1982;64:970–982.
22. Wendelboe AM, Hegmann KT, Biggs JJ, Cox CM, Portmann AJ,
Gildea JH, Gren LH, Lyon JL. Relationships between body mass
indices and surgical replacements of knee and hip joints. Am J
Prev Med. 2003;25:290–295.
23. Young NL, Cheah D, Waddell JP, Wright JG. Patient charac-
teristics that affect the outcome of total hip arthroplasty: a review.
Can J Surg. 1998;41:188–195.
Clin Orthop Relat Res (2008) 466:1169–1176
DOI 10.1007/s11999-008-0133-7
ORIGINAL ARTICLE
Reliable Angle Assessment During Periacetabular Osteotomy with
a Novel Device
Anders Troelsen MD, Brian Elmengaard MD, PhD,
Lone Rømer MD, Kjeld Søballe MD, DMSc
Received: 9 July 2007 / Accepted: 15 January 2008 / Published online: 9 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract We developed and assessed a measuring device
for intraoperative assessment of the acetabular index and
center edge angle during acetabular reorientation in peri-
acetabular osteotomy. We asked whether reliable
assessment of angles could be made using the device; to be
reliable we presumed the variability of angle measurements
should not exceed that of inherent variability when
assessing angles on radiographs (± 5°). The device was
mounted bilaterally on the pelvis, and using fluoroscopy,
angle measurements were obtained with adjustable mea-
suring discs. We conducted a cadaver study to assess
intraobserver and interobserver variability of the device
and to assess if pelvic positioning influenced variation of
measurements. Intraoperative measurements of 35 con-
secutive patients were compared with measurements on
postoperative radiographs. Intraoperatively obtained angle
measurements differed less than ± 5° from measurements
on postoperative radiographs and the intraobserver and
interobserver variability of the device were confined within
± 5°. Positioning did not influence the variation of angle
One or more of the authors (AT) have received funding from a grant
from the Danish Rheumatism Association during conduction of this
study.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
A. Troelsen (&), B. Elmengaard, K. Søballe
Orthopaedic Research Unit, University Hospital of Aarhus,
Tage-Hansens Gade 2, Building 7B, DK-8000 Aarhus C,
Denmark
e-mail: a_troelsen@hotmail.com
L. Rømer
Department of Radiology, University Hospital of Aarhus, Tage-
Hansens Gade 2, Building 2B, DK-8000 Aarhus C, Denmark
measurements beyond intraobserver variability of the
device when applying arcs of tilt and rotation of ± 12.5°.
We believe the device is a potentially helpful tool in the
periacetabular osteotomy. It is simple to use and facili-
tates repeated reliable angle measurements during
acetabular reorientation, making intraoperative radiographs
unnecessary.
Introduction
Developmental dysplasia of the hip (DDH) is characterized
by an excessively oblique and shallow acetabulum with
insufficient coverage of the femoral head laterally and
anteriorly [1, 10, 19]. The periacetabular osteotomy is a
well-established treatment for DDH in young adults [6, 20,
21, 24, 32, 33]. In this procedure, the acetabulum is
reoriented to enhance coverage of the femoral head and the
aim is to achieve congruity, to stabilize the joint, to med-
ialize the joint center, and to reduce contact pressures [6, 7,
13, 31]. This will relieve pain and improve function and is
likely to prevent additional overload of the labrum, carti-
lage, and soft tissues, thereby delaying or preventing the
development of osteoarthritis [20, 21, 24, 32, 33]. Under-
correction or overcorrection of the acetabulum can cause
symptoms such as the feeling of instability and impinge-
ment, respectively [9, 18, 25]. One study reported a
postoperative acetabular index (AI) outside the interval of
0° to 10° negatively influenced the outcome after peri-
acetabular osteotomy [24], and another study found the
postoperative center edge (CE) angle averaged 29° for hips
that had an increase in the apparent joint space compared
with an average 21° in hips that did not [33]. Although
not thoroughly investigated, it seems warranted to consider
123
1170 Troelsen et al.
acetabular reorientation an important step in surgical
decision making when performing periacetabular
osteotomy.
The AI and the CE angle are commonly used to char-
acterize the dysplastic anatomy of the hip [1, 10, 19, 26, 29,
34]. In the dysplastic hip undergoing a periacetabular
osteotomy, these angles must be assessed intraoperatively
to ascertain whether the intended reorientation of the ace-
tabulum has been achieved [5, 6, 17, 20, 25, 31]. One way
to assess the AI and the CE angle intraoperatively is by
taking anteroposterior pelvic radiographs during reorien-
tation of the acetabulum. Several radiographs may be
necessary and the surgeon must wait for the radiographs to
be developed and the angles to be measured. Alternatively,
assessment can be made by eye using fluoroscopy, but this
approach is qualitative and might only be reliable for an
experienced surgeon. Image-guided techniques seemingly
facilitate surgery but without improving acetabular cor-
rection, and are relatively expensive [8, 14, 15].
The senior author (KS) developed a measuring device
that allows intraoperative assessment of the AI and the CE
angle during acetabular reorientation when performing
periacetabular osteotomy. The device is mounted bilater-
ally at the anterior superior iliac spines and angle
measurements are performed with adjustable measuring
discs using fluoroscopy.
The primary research question was whether the mea-
suring device could assist the surgeon in obtaining good
correction of the acetabulum in terms of reliable intraop-
erative determination of the AI and the CE angle. We also
determined (1) the variability between angle measurements
obtained with the device and angle measurements on
postoperative radiographs of the same patients, (2) whether
this variability was increased compared with the intraob-
server variability of angle measurements obtained on
postoperative radiographs, (3) the intraobserver and inter-
observer variability when using the device for angle
measurements, and (4) whether patient positioning in terms
of differing pelvic tilt and rotation influenced variation of
angle measurements.
Materials and Methods
Prospective intraoperative assessment of the AI and the CE
angle was made in 35 patients undergoing unilateral peri-
acetabular osteotomies performed by the senior author
(KS) from May 2006 to March 2007. Intraoperative angle
measurements were performed by the senior author (KS)
using the measuring device. Indications for surgery were
symptomatic DDH defined by persistent hip pain, a CE
angle less than 25°, a congruent joint, Tönnis osteoarthritis
Grade 0 to 1, hip flexion greater than 110°, and internal
123
Clinical Orthopaedics and Related Research
rotation greater than 15°. The study group consisted of 31
females and four males with a median age of 30 years
(range, 14–57 years). Double measurements of the AI and
the CE angle were performed on the patients’ postoperative
anteroposterior pelvic radiographs by one author (AT). We
presumed the device satisfactory from clinical and radio-
graphic points of view if the measurements differed less
than ± 5° from measurements on postoperative radio-
graphs and if intraobserver and interobserver variability of
the device were within ± 5°. We selected this level as it
corresponds approximately to the inherent variability of AI
and CE angle measurements on plain radiographs [2, 11,
22, 30]. In a cadaver study, measurements of the AI and the
CE angle were made by two authors (AT, KS) in a setup
mimicking differing tilt and rotation of the pelvis.
The landmarks for the CE angle [34] are the center of
the femoral head and the most lateral point of the sclerotic
acetabular roof. For the AI [29], the landmarks are the most
medial and lateral points of the sclerotic acetabular roof.
Intraoperative and postoperative measurements on antero-
posterior pelvic radiographs demand alignment of the
pelvis using a line of reference (Fig. 1).
The measuring device is used under fluoroscopy in the
anteroposterior plane. It is mounted bilaterally at the
anterior-superior iliac spines and the position is secured by
inserting small spikes. To secure alignment of the pelvis
for angle measurement, a rod connects the spikes (Fig. 2).
Adjustable angle measuring discs can be mounted on the
alignment rod. We use two different angle measuring discs:
(1) the disc for AI measurement is positioned and adjusted
Fig. 1 Drawing a line of reference precedes construction of radio-
graphic angles. It is constructed by drawing a line going through the
most caudal points on the inferior ramus bilaterally. The acetabular
index (AI) is constructed by drawing a line from the most lateral to
the most medial limit of the sclerotic acetabular roof and from there
another line parallel to the line of reference. The center edge (CE)
angle is constructed by drawing a line from the most lateral limit of
the sclerotic acetabular roof to the center of the femoral head and
another line perpendicular to the line of reference and through the
center of the femoral head.
Volume 466, Number 5, May 2008
Fig. 2A–B The measuring device is used under fluoroscopy in the
anteroposterior plane. It is mounted bilaterally at the anterior-superior
iliac spines. The position is secured by inserting small spikes. To
secure alignment of the pelvis for angle measurement, a rod connects
the spikes. (A) The measuring device is mounted on a sawbones
model with the disc for CE angle measurement. (B) The measuring
device is mounted during surgery with the disc for CE angle
measurement.
until it is placed correctly in relation to the most medial and
lateral points of the sclerotic acetabular roof (Fig. 3A); and
(2) the disc for CE angle measurement is positioned and
adjusted until it is placed correctly in relation to the center
of the femoral head and the most lateral point of the
sclerotic acetabular roof (Fig. 3B). If fine-tuning of the
reorientation is necessary, the measuring discs are easily
adjusted for new assessment of the angle measurements.
Version of the acetabulum cannot be measured using the
device but is equally important and must be addressed to
achieve an appropriate anteversion [20, 30]. This is
accomplished by identifying the anterior and posterior
acetabular rim (Fig. 3A). In appropriate anteversion, it is
observed (1) the posterior rim is lateral to the anterior rim
and the center of the femoral head and (2) the anterior rim
is medial to the center of the femoral head [16, 20, 31].
Crossing of the anterior and posterior rim (crossover sign)
Intraoperatively Obtained AI and CE Angle 1171
is a sign of inappropriate anteversion or retroversion [12,
16]. Acetabular version, that is, the appearance of the
acetabular rims, is dependent on the pelvic tilt [12, 23, 28,
30]. We attempt to minimize intraoperative misinterpreta-
tion of the acetabular version by ensuring the pelvic tilt
intraoperatively corresponds to that of the anteroposterior
pelvic radiograph used for diagnosis and preoperative
planning. Pelvic tilt can be assessed in the anteroposterior
plane measuring the distance from the symphysis to the
sacrococcygeal joint [23, 27, 28]. We use fluoroscopy to
obtain an approximate measure intraoperatively. The aim
of the reorientation is to achieve an AI between 0° and 10°
and CE angle between 30° and 40°. To achieve appropriate
coverage, we approximate the AI to 0°. We avoid a neg-
ative value of the AI and a CE angle greater than 40° when
approximating the AI to 0°. A negative AI, a CE angle
greater than 40°, or the crossover sign are indicators for
overcorrection and it is recognized impingement and
reduced range of motion can result from a periacetabular
osteotomy [9, 18, 25]. We routinely assess range of motion
intraoperatively after reorientation. If the CE angle
becomes excessive, it is adjusted to 40° or less and the AI
is adjusted keeping it between 0° and 10°, which in our
experience is possible in almost every case. Also, in some
patients, the hip is too dysplastic to achieve a CE angle
greater than 30° when approximating the AI to 0°. Using
the above guidelines, the senior author (KS) decided when
optimal reorientation was achieved during surgery; the AI
(range, 0°–10°) and CE angle (range, 23°–36°) measure-
ments then were registered. After becoming familiar with
the measuring device, mounting and adjusting the discs for
measurement of the AI and the CE angle take approxi-
mately 2 to 3 minutes.
Anteroposterior pelvic radiographs were taken on the
first or second postoperative day and were stored digitally.
Using a workstation allowing digital management of
radiographs, the AI (range, -1°–9°) and the CE angle
(range, 24°–36°) were measured by one blinded observer
(AT). To assess intraobserver variability, a second blinded
measurement of the AI and the CE angle was performed by
the same observer 4 weeks later. Angle measurements
obtained with the device were compared with angle mea-
surements on postoperative radiographs to assess the
variability. We also compared intraobserver variability
with the variability when comparing angle measurements
obtained intraoperatively with those obtained on postop-
erative radiographs.
To assess intraobserver and interobserver variability and
evaluate the effect of pelvic tilt or rotation on the AI and
CE angle measurements, we did a cadaver study. The
female cadaver specimen was partial and consisted of the
pelvis, hips, thighs, and knees with intact skin and soft
tissues. Initially, the spikes and alignment rod of the
123
1172 Troelsen et al.
Fig. 3A–B (A) The disc for AI measurement is positioned and
adjusted until it is placed correctly in relation to the most medial
(Arrow 1) and lateral (Arrow 2) points of the sclerotic acetabular roof.
The AI in this case is 1° (Arrow 3). The anterior (row of white arrows)
and posterior (row of black arrows) acetabular rims are seen in the
anteroposterior fluoroscopic view. The fluoroscopic view is rotated
slightly clockwise, but alignment is secured by the measuring device.
(B) The disc for CE angle measurement is positioned and adjusted
until it is placed correctly in relation to the center of the femoral head
(Arrow 1) and the most lateral point of the sclerotic acetabular roof
(Arrow 2). The CE angle in this case is 35° (Arrow 3). 0° on the
measuring disc is marked. The fluoroscopic view is rotated slightly
clockwise, but alignment is secured by the measuring device.
Table 1. Angle measurements in the cadaver study
Angle Observer Tilting of the C-arm
*
Acetabular index 1 Right, 12.5°; left, 12.5°
Cranial, 12.5°; caudal, 12.5°
2 Right, 12.5°; left, 12.5°
Cranial, 12.5°; caudal, 12.5°
Center edge 1* Right, 12.5°; left, 12.5°
Cranial, 12.5°; caudal, 12.5°
2 Right, 12.5°; left, 12.5°
Cranial, 12.5°; caudal, 12.5°
*
Observer 1 did double measurements to determine intraobserver variability.
123
Clinical Orthopaedics and Related Research
measuring device were mounted as described previously.
The position of the pelvis then was adjusted until it was in
a neutral position. This was confirmed by achieving a
foramen obturator index of 1.0 and a distance between the
pubic symphysis and the sacrococcygeal joint of 4 cm on
an anteroposterior pelvic radiograph (the tube was oriented
perpendicular to the table and the tube to film distance was
110 cm, resulting in a magnification of 15%).
Measurement of the AI and the CE angle using the
measurement device was done as described previously.
Two observers (Observer 1: KS, experienced with the
device; Observer 2: AT, inexperienced with the device)
performed the angle measurements on the right hip of the
cadaver. The C-arm of the fluoroscope was tilted in 2.5°
increments to the left and to the right of the cadaver in an
arc totaling 25°. In this way, 11 measurements of both
angles were done from an angle of 12.5° to the left and to
12.5° to the right of the cadaver (Table 1). An angle of 0°
indicates the anteroposterior plane. Another 11 measure-
ments of both angles were performed similarly tilting the
C-arm of the fluoroscope in 2.5° increments in the cranial
and caudal directions in an arc totaling 25° (Table 1). For
each measurement, the femoral head was focused in the
middle of the image. The observers were blinded to each
other’s measurements and to the actual position of the C-
arm. The order of measurements was random with respect
to the position of the C-arm. We assessed intraobserver and
interobserver variability of angle measurements obtained
using the device.
We determined the variability between intraoperatively
and postoperatively obtained AI and CE angle measure-
ments and present data as mean of the difference with
standard deviation (SD) and 95% limits of agreement
presented in a Bland-Altman plot [3, 4]. The intraobserver
variability of angle measurements on postoperative radio-
graphs and intraobserver and interobserver variability of
angle measurements in the cadaver study were assessed
and presented the same way. We compared the intraob-
server variability of angle measurements on postoperative
Number of measurements Range Mean Standard deviation
11 1°–4° 2.73° ± 1.01°
11 1°–5° 3.00° ± 1.18°
11 2°–3° 2.36° ± 0.50°
11 0°–3° 1.64° ± 0.92°
11 34°–37° 35.73° ± 0.90°
11 32°–37° 34.27° ± 1.90°
11 35°–38° 36.27° ± 1.10°
11 35°–38° 36.09° ± 1.14°
Volume 466, Number 5, May 2008 Intraoperatively Obtained AI and CE Angle 1173

Table 2. Comparison of angle measurements

Compared measurements Angle Number of Mean of the Standard 95% Limits of Pitman’s variance
measurements difference deviation agreement ratio test

Perioperative versus postoperative AI 35 0.06° ± 2.22° -4.30°–4.42° AI: p = 0.14; CE: p = 0.24
CE 35 -0.03° ± 2.43° -4.80°–4.74°
Postoperative versus postoperative AI 35 -0.09° ± 1.72° -3.46°–3.28°
CE 35 -2.09° ± 2.02° -6.05°–1.87°
Cadaver study Observer 1 versus Observer 2 AI 22 0.86° ± 1.21° -1.51°–3.23° AI: p = 0.39; CE: p = 0.13
CE 22 -1.18° ± 1.84° -4.79°–2.43°
Cadaver study Observer 1 versus Observer 1 AI 22 -0.77° ± 1.45° -3.60°–2.06°
CE 22 0.55° ± 1.30° -2.00°–3.10°
AI = acetabular index; CE = center edge.

radiographs with (1) the variability between angle mea-

surements obtained intraoperatively and on postoperative
radiographs and (2) the intraobserver and interobserver
variability of angle measurements in the cadaver study; for
this comparison we used Pitman’s variance ratio test.
Angle measures of the different series in the cadaver study
are presented as range, mean, and SD. The variability of
these series of measurements is compared with the intra-
observer variability of the device using Pitman’s variance
ratio test. Analyses were performed using the Stata1
software package (Intercooled Stata version 9.2; StataCorp
LP, College Station, TX).

Results

We found the measuring device assisted the surgeon in

obtaining good correction of the acetabulum as the intra-
operatively obtained angle measurements differed less than
± 5° from measurements on postoperative radiographs and
the intraobserver and interobserver variability of the device
were confined within ± 5° (Table 2; Fig. 4).
We found 95% of differences between angle measure-
ments obtained with the device and angle measurements on
postoperative radiographs were between -4.3° and +4.4°
for the AI and between -4.8° and +4.7° for the CE angle
(Table 2; Fig. 4).
The level of variability (device versus postoperative) did
not differ (AI, p = 0.14; CE angle, p = 0.24) from the
intraobserver variability assessed on double measurement Fig. 4A–B Bland-Altman plots of difference against average for
of angles on postoperative radiographs (Table 2). With measurements of (A) AI and (B) CE angle on postoperative
radiographs and with the device intraoperatively are shown. The
intraobserver variability assessment on postoperative
mean of differences (solid line), the SD of differences (dotted line),
radiographs, we found 95% of differences between two and the 95% limits of agreement (dashed line) are presented. The 95%
repeated measurements would be expected to lie between limits of agreement show we can expect 95% of differences between
-3.5° and +3.3° for the AI and between -6.1° and +1.9° the two measurement methods to lie (A) between -4.3° and +4.4° for
the AI and (B) between -4.8° and +4.7° for the CE angle. Because
for the CE angle (Table 2).
inherent variability of radiographic angle assessment is ± 5°, the
For intraobserver and interobserver variability we found reported levels are satisfactory for reliable intraoperative angle
95% limits of agreement for both angle measures were assessment.

123
1174 Troelsen et al.
confined between -4.8° and +3.3° (Table 2). We also
found intraobserver and interobserver variability when
using the measuring device did not differ from each other
(AI, p = 0.39; CE angle, p = 0.13) (Table 2) or from the
intraobserver variability assessed on double measurements
on postoperative radiographs (AI, p = 0.45–0.96; CE
angle, p = 0.06–0.92).
Patient positioning (differing pelvic tilt and rotation) did
not influence the variation of angle measurements beyond
intraobserver variability of the device, within the arcs (±
12.5°) of tilt and rotation applied in this study. The dif-
ferent series of angle measurements performed by both
observers in the cadaver study showed variations (expres-
sed as ± SD) for AI measurement between ± 0.5° and ±
1.2° and for CE angle measurement between ± 0.9° and ±
1.9° (Table 1). These variations either did not differ from
intraobserver variability (AI, p = 0.65–0.81; CE angle,
p = 0.17–0.52), or in one series of AI measurement by
Observer 2 variation actually was less (p = 0.03) than the
intraobserver variability.
Discussion
Periacetabular osteotomy with reorientation of the acetab-
ulum is a well-established treatment to relieve pain,
increase function, and delay or prevent development of
osteoarthritis in young adults with DDH [20, 21, 24, 32,
33]. The AI and the CE angle are commonly used to
describe the morphologic characteristics of DDH [1, 10,
19, 26, 29, 34] and are assessed intraoperatively to ascer-
tain whether the intended reorientation has been achieved
[5, 6, 17, 20, 25, 31]. The senior author developed a
measuring device for intraoperative assessment of the AI
and the CE angle using fluoroscopy. We asked whether the
device could assist the surgeon in obtaining good correc-
tion of the acetabulum in terms of reliable intraoperative
determination of the AI and the CE angle.
Perhaps the major limitation of our study is that we had
no gold standard against which to compare angles using the
measuring device. Measurements of the AI and the CE
angle using three-dimensional imaging techniques could be
considered a gold standard, although for example, com-
puted tomography scanning is unlikely to be conducted
postoperatively. We therefore judged our measurements
against the variability of the measures, presuming the
correct (accurate) measure was within those ranges. It
seems meaningful to compare the angle measurements
obtained with the device with those on postoperative
radiographs as this is the commonly used imaging tech-
nique to evaluate the result of the acetabular reorientation.
We found the measuring device was able to assist the
surgeon in obtaining good correction of the acetabulum as
123
Clinical Orthopaedics and Related Research
the intraoperatively obtained AI and CE angle measure-
ments differed less than ± 5° from measurements on
postoperative radiographs and the intraobserver and inter-
observer variability of the device were confined well within
± 5°. The limit for a satisfactory result of the variability
assessment of ± 5° was chosen from clinical and radio-
graphic points of view. Using a two-dimensional imaging
technique for assessment of the AI and the CE angle, there
are some inherent problems adding to variability: (1)
identification of the center of the femoral head; (2) iden-
tification of the medial and lateral limits of the sclerotic
acetabular roof; and (3) construction of the line of refer-
ence. Previously, variability of AI and CE angle
measurements were confined to ± 5° [2, 11, 22, 30]. The
device is built to mimic construction of the AI [29] and the
CE angle [34] when they are drawn on an anteroposterior
pelvic radiograph. Therefore, the expectation was not to
find decreased variability when using the device, but to
make sure it did not exceed existing variability of angle
assessment using two-dimensional imaging techniques.
The intraobserver and interobserver variability of the
device did not differ from each other, suggesting the device
is equally reliable in the hands of an inexperienced
(Observer 2) and experienced (Observer 1) user. This is
likely attributable to the simple construction and straight-
forward use of the device.
One way to assess angles intraoperatively is by record-
ing anteroposterior pelvic radiographs. This is time-
consuming and several radiographs might be needed before
the intended correction has been achieved. An advantage is
that assessment of angles can be made constructing a line
of reference (Fig. 1) securing pelvic alignment. Using the
device, this is achieved by a rod connecting the bilaterally
and symmetrically inserted spikes. Also, the device is used
under fluoroscopy, which allows fast and easy repeated
measurements during fine-tuning of acetabular correction
without exposing patients to heavy radiation of conven-
tional radiographs. A potential disadvantage when using
the device is the insertion of spikes in the anterior-superior
iliac spine. In our experience, it does not convey specific
site-related postoperative pain or skin problems.
Another way is intraoperative assessment of the AI and
the CE angle by eye using fluoroscopy. In our opinion, this
can be done only by someone very experienced, if at all.
The surgeon using this approach for angle assessment can
evaluate postoperative radiographs to see if surgery resul-
ted in the intended reorientation, but to our knowledge,
there is no study evaluating whether this is a reliable
approach for AI and CE angle assessment, either for the
experienced or less experienced surgeon.
Computer-assisted and image-guided techniques using
preoperative computed tomography scans have been
introduced as potentially helpful and eliminate or reduce
Volume 466, Number 5, May 2008
the need for radiographs in acetabular reorientation surgery
[8, 14, 15]. However, these techniques are relatively
expensive and have not been shown to improve the
achieved acetabular correction compared with a conven-
tional approach [8]. Our measuring device is relatively
inexpensive and, in our opinion, simple and fast to use. In
our hands, mounting and adjusting the discs for both angle
measurements takes approximately 2 to 3 minutes. Also,
the device can assist the surgeon in making a reliable in-
traoperative determination of the AI and the CE angle, with
variability at the same level as angle measurements made
on conventional radiographs.
The appearance of acetabular version on two-dimen-
sional radiographic imaging depends on pelvic tilt [12, 23,
28, 30]. Radiographic dysplastic parameters, such as the AI
and the CE angle, generally are only affected beyond
inherent measuring error if the image is severely distorted
[2, 11, 22, 34]. We wanted to make sure the measuring
device did not add variability to angle measurements when
tilting and rotating the pelvis. This could introduce a
potential risk of misinterpreting the achieved angle mea-
surements. We found patient positioning, in terms of
differing pelvic tilt and rotation, did not influence variation
of angle measurements beyond intraobserver variability of
the device, within the arcs (± 12.5°) of tilt and rotation
applied in this study.
The measuring device is a potentially helpful tool during
acetabular reorientation. Using the device, one can obtain
reliable measures of the AI and the CE angle with a vari-
ability confined well within that of angle measurements on
conventional radiographs. Compared with existing tech-
niques for intraoperative AI and CE angle assessment, it
has the advantages of (1) being simple to use, (2) being
relatively inexpensive, (3) making conventional intraoper-
ative radiographs unnecessary, (4) facilitating fast repeated
angle measurements during fine-tuning of correction, and
(5) being equally reliable in the hands of inexperienced and
experienced users.
Acknowledgments We thank associate professor Peter Holm-
Nielsen from the Institute of Anatomy, Faculty of Health Sciences,
Aarhus University, Denmark, for providing access to the cadaver
specimen.
References
1. Anda S, Terjesen T, Kvistad KA, Svenningsen S. Acetabular
angles and femoral anteversion in dysplastic hips in adults: CT
investigation. J Comput Assist Tomogr. 1991;15:115–120.
2. Ball F, Kommenda K. Sources of error in the roentgen evaluation
of the hip in infancy. Ann Radiol (Paris). 1968;11:298–303.
3. Bland JM, Altman DG. Statistical methods for assessing agree-
ment between two methods of clinical measurement. Lancet.
1986;1:307–310.
Intraoperatively Obtained AI and CE Angle 1175
4. Bland JM, Altman DG. Applying the right statistics: analyses of
measurement studies. Ultrasound Obstet Gynecol. 2003;22:85–
93.
5. Clohisy JC, Barrett SE, Gordon JE, Delgado ED, Schoenecker
PL. Periacetabular osteotomy in the treatment of severe acetab-
ular dysplasia. J Bone Joint Surg Am. 2006;88(suppl 1, pt 1):65–
83.
6. Ganz R, Klaue K, Vinh TS, Mast JW. A new periacetabular
osteotomy for the treatment of hip dysplasias: technique and
preliminary results. Clin Orthop Relat Res. 1988;232:26–36.
7. Hipp JA, Sugano N, Millis MB, Murphy SB. Planning acetabular
redirection osteotomies based on joint contact pressures. Clin
Orthop Relat Res. 1999;364:134–143.
8. Hsieh PH, Chang YH, Shih CH. Image-guided periacetabular
osteotomy: computer-assisted navigation compared with the
conventional technique: a randomized study of 36 patients fol-
lowed for 2 years. Acta Orthop. 2006;77:591–597.
9. Hussell JG, Rodriguez JA, Ganz R. Technical complications of
the Bernese periacetabular osteotomy. Clin Orthop Relat Res.
1999;363:81–92.
10. Jacobsen S, Rømer L, Søballe K. The other hip in unilateral hip
dysplasia. Clin Orthop Relat Res. 2006;446:239–246.
11. Jacobsen S, Sonne-Holm S, Lund B, Søballe K, Kiær T, Rovsing
H, Monrad H. Pelvic orientation and assessment of hip dysplasia
in adults. Acta Orthop Scand. 2004;75:721–729.
12. Jamali AA, Mladenov K, Meyer DC, Martinez A, Beck M, Ganz
R, Leunig M. Anterior pelvic radiographs to assess acetabular
retroversion: high validity of the ‘‘cross-over-sign.’’ J Orthop
Res. 2007;25:758–765.
13. Kralj M, Mavcic B, Antolic V, Iglic A, Kralj-Iglic V. The Ber-
nese periacetabular osteotomy: clinical, radiographic and
mechanical 7–15-year follow-up of 26 hips. Acta Orthop.
2005;76:833–840.
14. Langlotz F, Bächler R, Berlemann U, Nolte LP, Ganz R. Com-
puter assistance for pelvic osteotomies. Clin Orthop Relat Res.
1998;354:92–102.
15. Langlotz F, Stucki M, Bächler R, Scheer C, Ganz R, Berlemann
U, Nolte LP. The first twelve cases of computer assisted peri-
acetabular osteotomy. Comput Aided Surg. 1997;2:317–326.
16. Mast JW, Brunner RL, Zebrack J. Recognizing acetabular version
in the radiographic presentation of hip dysplasia. Clin Orthop
Relat Res. 2004;418:48–53.
17. Mechlenburg I, Nyengaard J, Rømer L, Søballe K. Changes in
load-bearing area after Ganz periacetabular osteotomy evaluated
by multislice CT scanning and stereology. Acta Orthop Scand.
2004;75:147–153.
18. Myers SR, Eijer H, Ganz R. Anterior femoroacetabular
impingement after periacetabular osteotomy. Clin Orthop Relat
Res. 1999;363:93–99.
19. Nakamura S, Yorikawa J, Otsuka K, Takeshita K, Harasawa A,
Matsushita T. Evaluation of acetabular dysplasia using a top view
of the hip on three-dimensional CT. J Orthop Sci. 2000;5:533–
539.
20. Peters CL, Erickson JA, Hines JL. Early results of the Bernese
periacetabular osteotomy: the learning curve at an academic
medical center. J Bone Joint Surg Am. 2006;88:1920–1926.
21. Pogliacomi F, Stark A, Wallensten R. Periacetabular osteotomy:
good pain relief in symptomatic hip dysplasia, 32 patients fol-
lowed for 4 years. Acta Orthop. 2005;76:67–74.
22. Portinaro NM, Murray DW, Bhullar TP, Benson MK. Errors in
measurement of acetabular index. J Pediatr Orthop.
1995;15:780–784.
23. Siebenrock KA, Kalbermatten DF, Ganz R. Effect of pelvic tilt
on acetabular retroversion: a study of pelves from cadavers. Clin
Orthop Relat Res. 2003;407:241–248.
123
1176 Troelsen et al.
24. Siebenrock KA, Schöll E, Lottenbach M, Ganz R. Bernese
periacetabular osteotomy. Clin Orthop Relat Res. 1999;363:9–20.
25. Søballe K. Pelvic osteotomy for acetabular dysplasia. Acta Ort-
hop Scand. 2003;74:117–118.
26. Tallroth K, Lepistö J. Computed tomography measurement of
acetabular dimensions: normal values for correction of dysplasia.
Acta Orthop. 2006;77:598–602.
27. Tannast M, Murphy SB, Langlotz F, Anderson SE, Siebenrock
KA. Estimation of pelvic tilt on anteroposterior x-rays: a com-
parison of six parameters. Skeletal Radiol. 2006;35:149–155.
28. Tannast M, Zheng G, Anderegg C, Burckhardt K, Langlotz F,
Ganz R, Siebenrock KA. Tilt and rotation correction of acetab-
ular version on pelvic radiographs. Clin Orthop Relat Res.
2005;438:182–190.
29. Tönnis D. Congenital Dysplasia and Dislocation of the Hip in
Children and Adults. Berlin, Germany: Springer; 1987.
123
Clinical Orthopaedics and Related Research
30. Troelsen A, Jacobsen S, Rømer L, Søballe K. Weightbearing
anteroposterior pelvic radiographs are recommended. Clin Ort-
hop Relat Res. 2008; DOI: 10.1007/s11999-008-0156-0.
31. Trousdale RT, Cabanela ME. Lessons learned after more than
250 periacetabular osteotomies. Acta Orthop Scand. 2003;
74:119–126.
32. Trousdale RT, Ekkernkamp A, Ganz R, Wallrichs SL. Periace-
tabular and intertrochanteric osteotomy for the treatment of
osteoarthrosis in dysplastic hips. J Bone Joint Surg Am. 1995;
77:73–85.
33. Trumble SJ, Mayo KA, Mast JW. The periacetabular osteotomy:
minimum 2 year followup in more than 100 hips. Clin Orthop
Relat Res. 1999;363:54–63.
34. Wiberg G. Studies on dysplastic acetabula and congenital sub-
luxation of the hip: with special reference to the complication of
osteoarthritis. Acta Chir Scand. 1939;83(suppl 58):5–135.
Clin Orthop Relat Res (2008) 466:1177–1185
DOI 10.1007/s11999-008-0165-z
ORIGINAL ARTICLE
Curved-stem Hip Resurfacing
Minimum 20-year Followup
James W. Pritchett MD
Received: 5 July 2007 / Accepted: 25 January 2008 / Published online: 13 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Hip resurfacing is an attractive concept because
it preserves rather than removes the femoral head and neck.
Most early designs had high failure rates, but one unique
design had a femoral stem. Because that particular device
appeared to have better implant survival, this study asses-
sed the clinical outcome and long-term survivorship of a
hip resurfacing prosthesis. Four hundred forty-five patients
(561 hips) were retrospectively reviewed after a minimum
of 20 years’ followup or until death; 23 additional patients
were lost to followup. Patients received a metal femoral
prosthesis with a small curved stem. Three types of
acetabular reconstructions were used: (1) cemented poly-
urethane; (2) metal-on-metal; and (3) polyethylene secured
with cement or used as the liner of a two-piece porous-
coated implant. Long-term results were favorable with the
metal-on-metal combination only. The mean overall Harris
hip score was 92 at 2 years of followup. None of the 121
patients (133 hips) who received metal-on-metal articula-
tion experienced failure. The failure rate with polyurethane
was 100%, and the failure rate with cemented polyethylene
was 41%. Hip resurfacing with a curved-stem femoral
component had a durable clinical outcome when a metal-
on-metal articulation was used.
The author certifies that he has no commercial associations (eg.
consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc.) that might pose a conflict of interest in connection
with the submitted article.
The author certifies that his institution has approved the human
protocol for this investigation, that all investigations were conducted
in conformity with ethical principles of research, and that informed
consent for participation in the study was obtained.
J. W. Pritchett (&)
168 Lake WA Boulevard E, Seattle, WA 98112, USA
e-mail: bonerecon@aol.com
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Hip resurfacing offers several functional benefits over
THA: the size of the femoral head and neck remains close
to normal, the resurfaced hip is stable [1, 2, 14, 40] and
capable of an excellent range of motion [2, 3, 14, 30], and
the joint retains a greater degree of normal biomechanical
function [25, 30, 33, 42, 43]. It also offers several
procedural benefits: it is more bone conserving than con-
ventional hip arthroplasty because it does not involve
decapitation of the femur, and it results in less blood loss
and rehabilitates more easily [25, 42, 43]. The disadvan-
tages of this procedure include the risk of femoral neck
fracture (0%–7%) [2, 3, 14, 29, 40] and collapse of the
femoral head resulting from osteonecrosis (0%–4%) [8,
26]. Additionally, it is a demanding procedure that requires
anterior and posterior dislocation of the joint [2, 14, 40].
The first total hip resurfacing arthroplasty was devel-
oped by Charnley [10] using a polytetrafluorethylene-on-
polytetrafluorethylene (Teflon1 or Fluon1) bearing. The
procedure failed because of osteonecrosis of the femoral
head. Townley and Walker [39] introduced a device (total
articular replacement arthroplasty, or TARA) with a small
and short curved femoral stem; none of the other sub-
sequent designs, including that of Charnley, used a femoral
stem. In the 1970s, hip resurfacing was popular in several
centers in Europe, Japan, England, and the United States.
However, initial promising results [15, 16, 17, 44] gave
way to unacceptable failure rates (22%–76%) [21, 22, 41],
owing primarily to acetabular loosening from polyethylene
123
1178 Pritchett
wear (10%–46%) [4, 28, 36]. Less commonly, femoral
neck fracture, osteonecrosis, or loosening of the femoral
component occurred [4, 21, 22, 36, 41, 44].
Resurfacing was largely abandoned again until the
1990s when it was resurrected for the same reasons that
made it attractive initially: many patients want an active
lifestyle that would include participation in sports or rec-
reation [33, 42, 43], they want to keep their bone, and some
patients have the perception that revisions, if necessary, are
easier than with a conventional stem-supported intramed-
ullary hip prosthesis [2, 5, 6]. Modifications were made to
the original design based on the presumed cause of failure,
one of which was the requirement for a very thin and
flexible polyethylene component when retaining a femoral
implant the size of the normal femoral head. However, as
noted earlier, none of the original devices that had high
failure rates had a femoral stem. The one device that did
have a femoral stem appeared to have better midterm
outcomes than the nonstemmed devices [31, 38].
Therefore the following questions were raised: what are
the functional results from curved-stem total hip resurfac-
ing and what is the survivorship of the prosthesis over
long-term followup?
Materials and Methods
Four hundred forty-five patients (561 hips) who underwent
total hip resurfacing (TARA) procedures from 1960 to
1987 were retrospectively reviewed. Conventional THAs
also were performed and it is estimated that approximately
20% of patients treated received a resurfacing prosthesis.
Patients were generally selected for resurfacing procedures
if they were younger than 60 years. Patients older than
60 years were offered the resurfacing procedure if they
were active and if they had excellent bone quality on
their radiographs. None of the patients had a prior implant
arthroplasty, although 18 had previous surgery for a dis-
located hip or fracture. The underlying diagnosis was
osteoarthritis in 334 patients (75%), osteonecrosis in 44
(10%), posttraumatic arthritis in 31 (7%), inflammatory
arthritis in 18 (4%), and developmental dysplasia in 18
(4%). The patient population consisted of 218 women and
227 men with a mean body weight of 71 and 82 kg,
respectively (range, 50–107 kg). The mean age was
52 years (range, 30–74 years) with 97 patients aged 30 to
40 years, 118 aged 40 to 50 years, 109 aged 50 to 60 years,
100 aged 60 to 70 years, and 21 aged 70 to 74 years.
All patients were followed up until death or a minimum
of 20 years; 374 (84%) of the 445 patients had died by the
time of final followup. Twenty-three additional patients
underwent hip resurfacing but were lost to followup and
are not otherwise included in the results. The mean age at
123
Clinical Orthopaedics and Related Research
Table 1. Survivorship among original 445 patients treated with hip
resurfacing
Years since surgery Number (%) Mean age at death
or follow-up
(years; range)
Patients who had died 374 (84) 80 (58–99)
Less than 5 years 19 (5)
5–9 years 24 (6)
10–19 years 54 (14)
20–30 years 166 (45)
Longer than 30 years 111 (30)
Patients alive at followup 71 (16) 75 (53–94)
Survival periods
20–30 years 51 (72)
30–40 years 18 (25)
40 years 2 (3)
Patients lost to followup 23
the time of death was 80 years (range, 58–99 years) and
the mean survival time from surgery until the time of death
was 22 years. The remaining 71 patients (16%) had been
followed a minimum of 20 years (average, 27 years; range,
20–41 years) (Table 1). Prior Institutional Review Board
approval was obtained for this study.
Surgery was performed by one of two surgeons (JWP,
COT). Each surgical procedure was done through an
anterolateral approach without trochanteric osteotomy. The
hip was dislocated anteriorly and the femur prepared. The
femoral head was downsized when possible, trying not to
notch the femoral neck. The zenith of the femoral head was
removed at an approximate 140° angle to the femur
(measured by a goniometer), and all cystic or structurally
damaged at-risk bone was removed. The guide stem then is
placed into the femoral canal. Fitting this curved stem into
the femur creates slight valgus relative to the medial tra-
becular system of the femur. Cylinder and chamfer cutters
were made to complete the preparation of the femoral
head [38]. Prostheses were placed using an interference
fit, cemented, or porous-coated technique. The surgeon
attempted to place the femoral component in valgus.
The type of prosthesis varied with the time at which the
procedure was performed. In the earliest procedures from
1960 to 1962, the acetabular surface used was polyurethane
(24 patients). This polymer was prepared by mixing the
prepolymer with resin and the catalyst at the time of surgery
and shaping it in situ or on the back table to the femoral
prosthesis. Polyurethane therefore served as the anchoring
cement for the femoral side and as the articular replacement
and cement for the acetabulum. Although it is a plastic, it had
a fairly rough finish. The length of the femoral stem varied
from 127 to 165 mm, with longer stems used more com-
monly in the earlier cases. Metal-on-metal implants became
Volume 466, Number 5, May 2008
Fig. 1 A photograph shows the curved-stem metal-on-metal hip
resurfacing prosthesis. The acetabular component has a small fin and
screw holes for adjunctive fixation.
available in 1962 and were used in 121 patients through the
mid 1970s; these were made of cobalt chromium (DePuy
Orthopaedics, Inc, Warsaw, IN; Howmedica, Rutherford,
NJ; Zimmer, Inc, Warsaw, IN) (Fig. 1). They were implan-
ted without cement on the acetabular side and with or without
cement on the femoral side. Polyethylene components
(DePuy), which became available in the 1970s (222 pati-
ents), had an initial thickness of 4.5 mm. These components
were later increased to 6.0 mm and cemented in place using
polymethylmethacrylate (Simplex1; Howmedica, NJ). The
two-piece metal-polyethylene component (78 patients) was
porous-coated with a coxcomb fin for adjunctive fixation
(Fig. 2). Fifteen patients received a two-piece cementless
acetabular prosthesis in one hip and a cemented polyethylene
prosthesis in the other (Fig. 2).
Patients were followed prospectively and asked to return
at 1 year, 2 years, 5 years, and every 5 years thereafter.
Curved-stem Hip Resurfacing 1179
Fig. 2 An anteroposterior radiograph shows a pelvis with a cemented
polyethylene cup on the left side and a cementless two-piece
acetabular prosthesis on the right side. The radiograph was taken
21 years after insertion of the prosthesis on the right and 29 years
after insertion of the prosthesis on the left.
The Harris hip score was used (COT, JWP) to evaluate the
surgical results [20]. The hip score at 2 years was used to
grade the functional result. When this was not possible,
patients were sent a written questionnaire (Appendix 1) or
contacted by telephone and interviewed using the same
questionnaire. Clinical examinations to final followup were
available for 226 (51%) patients. Written questionnaires
were available for 88 (20%) patients and telephone ques-
tionnaires were available for 131 (29%) patients. Patients
were queried specifically about the need for additional
surgery on their hip. If surgery had been performed, the
patient was asked to provide information about that pro-
cedure. The date of death was obtained by direct
communication with the family. Information regarding the
patient’s hip function was obtained from the family for
patients who had died. Twenty seven patients (6%)
underwent a resurfacing procedure on one side and a
conventional THA on the other. They were asked which
was the better hip based on their perception of a more
natural feel and superior strength or function.
Immediate postoperative radiographs were assessed and
the abduction angle of the acetabular component and the
stem shaft angle of the femoral prosthesis were measured
[2, 7]. The femoral component was considered malposi-
tioned if it was 5° more horizontal (varus) than the medial
trabecular system of the proximal femur [12, 35]. The
acetabular component was considered malpositioned if
the abduction angle was greater than 65° or less than 30°.
The observers (JWP, COT) were not blinded to the results.
Survivorship was computed using Kaplan-Meier survi-
vorship estimates [24]. and the end points consisted of
123
1180 Pritchett
Fig. 3 A Kaplan-Meier survivorship curve for the cemented poly-
ethylene acetabular component is shown. The survivorship is 59% at
20 years after prosthesis insertion. Dashed lines indicate 95%
confidence intervals.
revision or removal (or recommendation for revision or
removal) of either component for any reason. Patients were
censored at death or at revision. A 95% confidence interval
was calculated. Survivorship analyses were calculated for
each type of acetabular reconstruction used (Figs. 3, 4).
Failure was defined by removal or revision of the pros-
thesis or consideration for revision based on reduction in
function of the hip with radiographic evidence of loosening
of the components, such as change in position of either the
femoral or acetabular component or extensive radiolucent
lines around the acetabular component and resorption of
bone [1].
Results
The mean peak Harris hip score improved from 57 (range,
8–79) to 92 (range, 63–100) at 2 years. Flexion improved
from a mean of 83° (range, 5°–118°) to a mean of 110°
(range, 65°–140°) between preoperative and postoperative
evaluations. Most patients experienced no pain and only
four (less than 1%) experienced severe pain. Of the 445
.
patients assessed for postsurgical activity, 1 3 participated
in strenuous athletics or work and only 22 (5%) did not
work or participate in activities. Ninety percent were not
limited in their activities (Table 2). Of the 27 patients who
123
Clinical Orthopaedics and Related Research
Fig. 4 A Kaplan-Meier survivorship curve for the curved-stem
femoral component is shown. The survivorship rate is 84% at
20 years. Dashed lines indicate 95% confidence intervals.
had a resurfacing procedure on one side and a conventional
THA on the other, all indicated the hip that had resurfacing
was the better hip.
Among living patients and those who died with their
implant in place, the survivorship for the femoral prosthesis
(including patients with all three acetabular implants) was
84% (Fig. 4). However, the metal-on-metal prostheses had
100% survivorship. Failure rates for the remaining ace-
tabular prostheses ranged from 34% to 100% (Table 3).
The highest failure rate (100%) was seen with polyure-
thane. This bearing surface disappeared radiographically
with time (Fig. 5); thereafter, this prosthesis seemed to
function as a hemiarthroplasty. Of the two patients with
polyurethane prostheses undergoing revision, one had
metal-on-metal resurfacing with a good outcome, and the
other underwent THA because of a femoral neck fracture.
The cemented polyethylene acetabular prosthesis (Fig. 2)
also resulted in high failure rates. The 15 patients who
received a two-piece cementless acetabular prosthesis in
one hip and a cemented polyethylene prosthesis in the other
also experienced high failure rates (Table 3). All but two
of the 141 revisions were in patients with metal-on-
polyethylene articulation and two involved a metal-on-
polyurethane prosthesis. None of the metal-on-metal
prostheses underwent revision (Table 3). We removed both
components and inserted an entirely new resurfacing
prosthesis in two patients. The acetabular prosthesis alone
Volume 466, Number 5, May 2008 Curved-stem Hip Resurfacing 1181

Table 2. Complications of hip resurfacing procedures

Outcome Number of patients (%) Comments

Complications
Deep infection 11 (2) Over lifetime of prosthesis
Dislocation 5 (\ 1)
Periprosthetic fracture (hips) 6 (\ 1) Intertrochanteric and subtrochanteric
Femoral neck fractures 10 (1.7)
Intraoperative femoral neck fracture 1 Converted to THA
Femoral nerve palsy 2 Both patients recovered
Sciatic palsy 5 (\ 1) Recovery: 2 full; 2 partial; 1 limited
as a result of peroneal and tibial involvement

Table 3. Functional results of hip resurfacing

Pain Number (%) Assessed 2 years after procedure

No pain 459 (82)

Slight pain 86 (15)
Moderate pain 12 (2)
Severe pain 4 (\ 1)
Function: postsurgical activity Assessed 2 years after procedure in 445 patients
Highly active 147 (33) Strenuous sports or job
Active and no limitations necessary 254 (57)
Moderately active 22 (5)
Inactive 22 (5)
Patient satisfaction
Satisfied with outcome 427 (96)
Dissatisfied with outcome 18 (4) Nine patients were dissatisfied because of a limp
or weakness Nine patients were dissatisfied because of pain

was revised in 22 hips. The femoral component was secure Discussion

in these cases. Revision surgery in the remaining patients
(117 hips) was conventional THA. Thirty-two of 44
patients (73%) with osteonecrosis experienced prosthesis
failure (mean time to failure, 7 years; range, 3–12 years).
Postoperative radiographs revealed technical errors in
approximately 13% of patients, most commonly a malpo-
sitioned femoral component (28 hips or 5%), and smaller
numbers of malpositioned acetabular components, malpo-
sitioned femoral and acetabular components in the same
patient, with notched femoral necks, and incompletely
seated femoral components in 36 instances (Table 4).
Twenty-one of these 64 patients (33%) had a body mass
index greater than 35. The most common complications
seen at any time during the followup included deep
infection, dislocation, and periprosthetic fracture. The
periprosthetic fractures occurred sporadically any time
after the surgical procedure from 6 months to 36 years
later. Less frequently, intraoperative fracture and nerve
palsy occurred (Table 5). Medical complications of various
types occurred in approximately 5% of patients.
The curved-stem hip resurfacing prosthesis was the second
attempt (1960) at total hip resurfacing [15]. John Charnley
made the first attempt (1951) before his work on low-
friction arthroplasty [11]. The innovator (COT) continued
to use the curved-stem prosthesis for over 40 years and
long-term followup of the patients is available.
There are some limitations of this study. The investi-
gation is retrospective, but the primary outcome,
prosthesis, and survival are known on all but a few patients
who were lost to followup. Second, this study investigates
a prosthesis that was in evolution as it was being used.
Three different materials were used for the acetabular
resurfacing, although most were metal-on-metal and metal-
on-polyethylene a few were metal-on-polyurethane. Also
the femoral component was secured without or with
cement and varied in stem length adding variations for
which statistical analysis could not be done. Pathologic
specimens of failed cases or autopsy retrievals also are not
available to show the reasons for success or failure. The

123
1182 Pritchett
Fig. 5A–C Anteroposterior radiographs show a metal-on-polyurethane curved-stem resurfacing prosthesis. (A) There is 9 mm of polyurethane
immediately after insertion. (B) Two years later, approximately ½ of the polyurethane has worn away. (C) All the polyurethane has worn away
6 years after prosthesis insertion.
Table 4. Radiographic findings after hip resurfacing
Radiographic finding Number of hips (%)
Femoral component malpositioned 28 (5)
Acetabular component malpositioned 17 (4)
Acetabular and femoral components malpositioned 6 (1)
Notched femoral neck 11 (2)
Femoral component incompletely seated 2 (\ 1)
two surgeons involved performed all of the clinical and
radiographic analyses. Complete followup data (particu-
larly radiographs) were not available on many patients so
questionnaires were relied on for some of the information.
There are no long-term functional outcomes reported in
this study and validated instruments were not available at
the time of surgery for the earlier patients. Finally, because
this is a single patient series, there are no patients or groups
available with other stemmed femoral devices for direct
comparison.
The data suggest a survivorship of 59% at 20 years
when using a polyethylene acetabular component. Mesko
et al. [31] reported a 75% survivorship with the curved-
stem TARA prosthesis at 10 years, but failure rates of 57%
and 76% were reported in two other studies [21, 41]. The
THARIESTM total resurfacing prosthesis had a failure rate
of 50% at 10 years and 80% at 15 years [4, 28]. The
123
Clinical Orthopaedics and Related Research
Comments
Greater than 5° more varus postoperatively
measured versus medial trabecular system
Includes 11 with hip resurfacing failure
Includes three with hip resurfacing failure
Includes three with a femoral neck fracture
Includes one with hip resurfacing failure
Indiana conservative hip had a failure rate of 66% at
9 years and the Wagner had a failure rate of 60% at 8 years
[22, 36].
Metal-on-metal hip resurfacing had a survival rate of
100% in this study and as a generic type is the most
commonly used resurfacing today. Success rates of 94% or
better are reported with as much as 9 (range, 2–9 years)
years of followup [2, 3, 14, 40].
Femoral neck fracture is a rare complication after hip
resurfacing, occurring at reported rates of 0% to 7% [2, 3,
29, 37]. The rate of femoral fracture and femoral compo-
nent failure was low in this series. This was despite the
effort made to downsize the femoral head that resulted in
femoral neck notching in some cases. Placing the femoral
component in valgus reduces the stresses in the superior
aspect of the femoral head and neck [23, 27, 45]. Femoral
components placed in 5° valgus have a factor of 6.1
Volume 466, Number 5, May 2008
Table 5. Revisions of hip resurfacing prostheses
Item Type of prosthesis
Metal-on-polyurethane Metal-on-
metal
Revision needed 2 0
Patients/hips 24/26 121/133
Mean followup (years; range) 24 (20–31) 26 (20–41) 25 (20–31)
Alive at followup 0 0
Lost to followup 0 2
Prosthesis failure rate 100% 0%
Reason for failure
More than one reason Polyurethane wear (26) N/A
in some patients Femoral neck fracture (1)
N/A = not applicable.
reduction in the relative risk of an adverse outcome [7].
Obese patients have a higher incidence of varus positioning
and femoral neck fracture.
Most early resurfacing implants involved hemispheric
preparation of the femoral head followed by placement of a
hemispheric femoral implant; unfortunately, shear often
resulted in loosening of these implants. These implants also
did not have a femoral stem [4, 16, 22, 36]. The presence of
a stem reduces the shear force on the prosthetic femoral
head-native femoral neck junction by 34% [13, 19]. A
curved femoral stem reduces the shear force on the femoral
neck by 17% compared with a straight stem [13, 19]. This
may be an additional explanation for the lower rate of
femoral neck fracture and femoral component loosening
observed in the current study with other implants [4, 22, 28,
36]. The preparation required to implant a curved stem
promotes valgus positioning of the femoral component [18,
38]. Although I report what appears to be the first attempt
to use a metal-on-metal resurfacing prosthesis, there were
other early innovators. Gerard [17] used a metal-on-metal
prosthesis but did not fix the acetabular component to the
pelvis; Mueller [32] also performed metal-on-metal resur-
facing procedures. In this series, a prosthesis originally
known as cup-stem arthroplasty was used [39], in which
the hemisphere was replaced by a flat-topped cylinder. The
technique used to place this implant excised at-risk bone in
the femoral head and this may have contributed to the low
failure rate. The head design provides compressive resis-
tance stability; and a short, curved stem on the prosthesis
adds stability without stress relieving the proximal femur
[13, 38]. Current designs use a straight femoral stem [3, 14,
40]. The results in this report and biomechanical
Curved-stem Hip Resurfacing 1183
Metal-on-cemented Metal on two-piece cementless
polyethylene with polyethylene
105 34
222/282 78/120
21 (20–22)
41 30
15 6
41% 34%
Loosening of acetabulum (76) Polyethylene wear (27)
Polyethylene wear (30) Component loosening with
Loosening of femoral prosthesis (5) migration (11)
Femoral neck fracture (6) Femoral neck fracture (3)
considerations suggest a curved femoral stem may be the
superior design for a resurfacing prosthesis.
The difficulties with hip resurfacing in this series were
primarily on the acetabular side. Well-performed femoral
resurfacing rarely fails with time; this was true when an
interference press-fit technique was used when neither
cement nor porous coating was yet available. Early pro-
cedures involved the use of materials that did not provide
an appropriate acetabular surface. Charnley [10, 11] used
polytetrafluorethylene in the first hip resurfacing procedure
and it failed. In this series, polyurethane failed every time.
However, polyurethane does not cause an osteolytic reac-
tion; as a result, patients functioned generally well as it
wore away. They had some pain and radiographs of the hip
looked as though hemiarthroplasty had been performed
(Fig. 5). The crude polyurethane used in the early days has
now been reformulated. Thus far, the wear characteristics
of the new formulation seem favorable [9]. Another con-
tributor to resurfacing arthroplasty failure in this series
(and in others) was the use of cemented polyethylene
acetabular components that loosened and wore through,
often resulting in osteolysis [1, 4, 16, 21, 22, 36]. Metal-
backed cemented polyethylene sockets were not used in
this series, but others have reported prosthesis failure when
they were used in such procedures [34, 41].
Exposing and positioning the acetabular component
with the femoral head in the way is technically difficult [2,
3, 37]. Exacting preparation of the femoral head is neces-
sary. Hip resurfacing using the curved femoral stem
performed well on the femoral side for more than 20 years.
Although polyethylene acetabular components failed reg-
ularly, metal acetabular components performed well. Hip
123
1184 Pritchett
resurfacing may be an attractive option for a young patient
fearing a potentially difficult future total hip replacement
revision.
Acknowledgments I thank Charles O. Townley, MD, for contrib-
uting his cases and knowledge. He died on December 22, 2006.
References
1. Amstutz H. Surface replacement arthroplasty. In: Amstutz HC,
ed. Hip Arthroplasty. Edinburgh, UK: Churchill Livingstone;
1991:295–332.
2. Amstutz HC, Ball ST, Le Duff MJ, Dorey FJ. Resurfacing THA
for patients younger than 50 years. Clin Orthop Relat Res.
2007;460:159–164.
3. Amstutz HC, Beaule PE, Dorey FJ, Le Duff MJ, Campbell PA,
Gruen TA. Metal-on-metal hybrid surface arthroplasty: two to
six-year follow-up study. J Bone Joint Surg Am. 2004;86:28–39.
4. Amstutz HC, Dorey F, O’Carroll PF. THARIES resurfacing
arthroplasty: evolution and long-term results. Clin Orthop Relat
Res. 1986:213:92–114.
5. Alberton GM, High WA, Morrey BF. Dislocation after total hip
arthroplasty: an analysis of risk factors and treatment options.
J Bone Joint Surg Am. 2002:84:1788–1797.
6. Ball ST, Le Duff MJ, Amstutz HC. Early results of conversion of
failed femoral component in hip resurfacing arthroplasty. J Bone
Joint Surg Am. 2007:89:735–741.
7. Beaule PE, Lee JL, Le Duff MJ, Amstutz HC, Ebramzadeh E.
Orientation of the femoral component in surface arthroplasty of
the hip. J Bone Joint Surg Am. 2004:86:2015–2021.
8. Campbell P, Beaule PE, Ebramzadeh E, Le Duff M, DeSmet K,
Lu Z, Amstutz HC. The John Charnley Award: a study of implant
failure in metal-on-metal surface arthroplasties. Clin Orthop
Relat Res. 2006;453:35–46.
9. Carbone A, Howie DW, McGee M, Field J, Pearcy M, Smith N,
Jones E. Aging performance of a compliant bearing acetabular
prosthesis in an ovine hip arthroplasty model. J Arthroplasty.
2006;21:899–906.
10. Charnley J. Arthroplasty of the hip: a new operation. Lancet.
1961;1:1129–1132.
11. Charnley J. Low Friction Arthroplasty of the Hip: Theory,
Practice. New York, NY: Springer; 1979.
12. Clark JM, Freeman MA, Witham D. The relationship of neck
orientation to the shape of the proximal femur. J Arthroplasty.
1987;2:99–109.
13. Collier J, Kennedy F, Mayor M, Townley CO. The importance of
stem geometry, porous coating and collar angle of femoral hip
prosthesis on the strain distribution in the normal femur. Trans
Soc Biomater. 1983;9:96.
14. Daniel J, Pynsent PB, McMinn DJ. Metal-on-metal resurfacing
arthroplasty of the hip in patients under 55 years with osteoar-
thritis. J Bone Joint Surg Br. 2004;86:177–184.
15. Freeman MA. Total surface replacement hip arthroplasty. Clin
Orthop Relat Res. 1978;134:2–4.
16. Freeman MA, Cameron HU, Brown GC. Cemented double cup
arthroplasty of the hip: a 5-year experience with the ICLH
prosthesis. Clin Orthop Relat Res. 1978;134:45–52.
17. Gerard Y. Hip arthroplasty by matching cups. Clin Orthop Relat
Res. 1978;134:25–35.
18. Grecula MJ, Thomas JA, Kreuzer SW. Impact of implant design
on femoral head hemiarthroplasty. Clin Orthop Relat Res.
2004;418:41–47.
123
Clinical Orthopaedics and Related Research
19. Haboush EJ. A new operation for arthroplasty of the hip based on
biomechanics, photoeleasticity, fast setting dental acrylic, and
other considerations. Bull Hosp Joint Dis. 1953;14:242–276.
20. Harris WH. Traumatic arthritis of the hip after dislocation
and acetabular fractures: treatment by mold arthroplasty: an
end-result study using a new method of result evaluation. J Bone
Joint Surg Am. 1969;51:737–755.
21. Head WC. Total articular resurfacing arthroplasty: analysis of
failure in 67 hips. J Bone Joint Surg Am. 1984;66:28–34.
22. Howie DW, Campbell D, McGee M, Cornish BL. Wagner
resurfacing hip arthroplasty: the results of one hundred consec-
utive arthroplasties after eight to ten years. J Bone Joint Surg Am.
1990;72:708–714.
23. Huiskes R, Strens PH, van Heck J, Slooff TJ. Interface stresses in
the resurfaced hip: finite element analysis of load transmission in
the femoral head. Acta Orthop Scand. 1985;56:474–478.
24. Kaplan E, Meier P. Nonparametric estimation from incomplete
observations. J Am Stat Assoc. 1958;53:457–481.
25. Lilikakis AK, Arora A, Villar RN. Early rehabilitation comparing
hip resurfacing and total hip replacement. Hip Int. 2005;15:189–
194.
26. Little CP, Ruiz AL, Harding IJ, McLardy-Smith P, Gundle R,
Murray DW, Athanasou NA. Osteonecrosis in retrieved femoral
heads after failed resurfacing arthroplasty of the hip. J Bone Joint
Surg Br. 2005;87:320–323.
27. Long JP, Bartel DL. Surgical variables affect mechanics of a
hip resurfacing system. Clin Orthop Relat Res. 2006;453:115–
122.
28. Mai MT, Schmalzried TP, Dorey FJ, Campbell PA, Amstutz HC.
The contribution of frictional torque to loosening at the cement
bone interface in THARIES hip replacements. J Bone Joint Surg
Am. 1996;78:505–511.
29. Marker DR, Seyler TM, Jinnah RH, Delanois RE, Ulrich SD,
Mont MA. Femoral neck fractures after metal-on-metal total hip
resurfacing. J Arthroplasty. 2007;22(7 suppl 3):66–71.
30. Mont MA, Seyler TM, Ragland PS, Starr R, Erhart J, Bhave A.
Gait Analysis of patients with resurfacing hip arthroplasty com-
pared with hip osteoarthritis and standard total hip arthroplasty.
J Arthroplasty. 2007;22:100–108.
31. Mesko JW, Goodman FG, Stanescu S. Total articular replace-
ment arthroplasty: a three-to ten-year case-controlled study. Clin
Orthop Relat Res. 1994;300:168–177.
32. Mueller M. The benefits of metal-on-metal total hip replacement.
Clin Orthop Relat Res. 1995;311:54–59.
33. Naal FD, Maffiuletti NA, Munzinger U, Hersche O. Sports after
hip resurfacing. Am J Sports Med. 2007;35:705–711.
34. Pritchett JW. Success rates of the TARA hip. Am J Orthop.
1998;27:658.
35. Pritchett JW, Perdue KD, Dona GA. The neck shaft-plate shaft
angle in slipped capital femoral epiphysis. Orthop Rev.
1989;28:1187–1192.
36. Ritter MA, Lutgring JD, Berend ME, Pierson JL. Failure mech-
anisms of total hip resurfacing. Clin Orthop Relat Res.
2006;453:110–114.
37. Shimmin AJ, Back D. Femoral neck fractures following
Birmingham hip resurfacing: a national review of 50 cases.
J Bone Joint Surg Br. 2005;87:463–464.
38. Townley CO. Hemi and total articular replacement arthroplasty
of the hip with a fixed femoral cup. Orthop Clin North Am.
1982;13:869–894.
39. Townley CO, Walker S. Intramedullary cup-stem arthroplasty of
the hip. J Bone Joint Surg Am. 1961;43:602.
40. Treacy RB, McBryde CW, Pynsent PB. Birmingham hip resur-
facing: a minimum follow-up of five years. J Bone Joint Surg Br.
2005;87:167–170.
Volume 466, Number 5, May 2008
41. Treuting RJ, Waldman D, Hooten J, Schmalzried TP, Barrack
RL. Prohibitive failure rate of the total articular replacement
arthroplasty at five to ten years. Am J Orthop. 1997;26:114–118.
42. Vail TP, Mina CA, Yergler JD, Pietrobon R. Metal-on-metal hip
resurfacing compares favorably with THA at 2 years followup.
Clin Orthop Relat Res. 2006;453:123–131.
43. Vendittoli PA, Lavigne M, Roy AG, Lusignan D. A prospective
randomized clinical trial comparing metal-on-metal total hip
Curved-stem Hip Resurfacing 1185
replacement and metal-on-metal total hip resurfacing in patients
less than 65 years old. Hip Int. 2006;16:73–81.
44. Wagner H. Surface replacement arthroplasty of the hip. Clin
Orthop Relat Res. 1978;134:102–130.
45. Watanabe Y, Shiba N, Matsuo S, Higuchi F, Tagawa Y, Inoue A.
Biomechanical study of resurfacing arthroplasty: finite element
analysis of the femoral component. J Arthroplasty. 2000;15:505–
511.
123
Clin Orthop Relat Res (2008) 466:1186–1192
DOI 10.1007/s11999-008-0168-9
ORIGINAL ARTICLE
Alumina Inlay Failure in Cemented Polyethylene-backed Total
Hip Arthroplasty
Kentaro Iwakiri MD, Hiroyoshi Iwaki MD, PhD,
Yukihide Minoda MD, PhD, Hirotsugu Ohashi MD, PhD,
Kunio Takaoka MD, PhD
Received: 21 July 2007 / Accepted: 30 January 2008 / Published online: 21 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Alumina-on-alumina bearings for THA have
markedly improved in mechanical properties through
advances in technology; however, alumina fracture is still a
concern. We retrospectively reviewed 77 patients (82 hips)
with cemented alumina-on-alumina THAs to identify fac-
tors relating to alumina failure. The mean age of the
patients at surgery was 63 years. The prostheses had a
cemented polyethylene-backed acetabular component with
an alumina inlay and a 28-mm alumina head. Revision
surgery was performed because of alumina inlay failure in
four hips (three fractures and one dissociation; 5.6%), deep
infection in two, and recurrent dislocation in one. The
8-year survival rate was 90.7% with revision for any reason
and 94.4% with revision for alumina failure as the end
point. There were no differences in age, body mass index,
gender, mobility, function, abduction angle, or size of
component among the four hips with alumina failure and
the remaining 68 hips without it; however, radiolucent lines
in the sockets were more apparent in four cases with
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
K. Iwakiri (&), H. Iwaki, Y. Minoda, K. Takaoka
Department of Orthopaedic Surgery, Osaka City University
Graduate School of Medicine, 1-4-3 Asahi-machi, Abeno-ku,
Osaka 545-8585, Japan
e-mail: kentucky@msic.med.osaka-cu.ac.jp
H. Ohashi
Department of Orthopaedic Surgery, Saiseikai Nakatsu Hospital,
Osaka, Japan
123
alumina inlay failure. This alumina-on-alumina THA thus
yielded unsatisfactory medium-term results because we
observed a high rate of catastrophic alumina inlay failure.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
The alumina-on-alumina articulation in THA was intro-
duced in the 1970s to reduce wear and its consequences
[19, 26]. Alumina particles induced less macrophage
reaction and cytokine secretion than polyethylene particles
[8], and THA using alumina-on-alumina articulation
induced little periprosthetic osteolysis [13, 25, 31]. How-
ever, early alumina prostheses, eg, with conically threaded
monoblocks or spherical press-fit acetabular components,
generally were found to have insufficient fixation of the
acetabular component and the risk of fracture of the
alumina component seemed to be a problem [3, 13]. During
the last decade, the quality of materials has improved
considerably and the risk of fracture has decreased [13, 27].
Subsequently, good clinical results have been obtained
with a cementless, press-fit, metal-backed acetabular
component with an alumina insert [3, 31]. However, con-
cerns still remain regarding problems with the alumina
component.
A unique polyethylene-backed acetabular component
with an alumina insert intended for fixation with bone
cement was developed in 1998. Recently, there have been
reports regarding problems with the alumina inlay in this
design [1, 12, 14, 18, 22, 24, 25, 28, 30, 31], but what
factors might relate to the socket or insert failure are
unclear as there have been no reports regarding the clinical
Volume 466, Number 5, May 2008 Polyethylene-backed Cemented Alumina THA 1187

results of this cemented polyethylene-backed alumina-on-

alumina THA.
We asked whether age, body mass index (BMI), gender,
mobility, function, abduction angle, size of component, the
existence of radiolucent lines, or component type with or
without a flange were associated with alumina failure.

Materials and Methods

We retrospectively reviewed 77 consecutive patients (82

hips) who underwent THA between February 1998 and
July 2000. In all 82 hips, cemented THA was performed
using a polyethylene-backed acetabular component with an
alumina inlay (ABS Cup; Kyocera, Kyoto, Japan) (Fig. 1),
KC stem (Kyocera), and 28-mm alumina femoral head.
Four patients (four hips) died within 27 months after the
operation without complications from surgery. Two
patients (two hips) were excluded because of recurrent
dislocation and one patient (one hip) was excluded because
of deep infection; these patients underwent revision sur-
gery within 24 months. Three patients (three hips) were
lost to followup. The remaining 67 patients (72 hips) were
followed for a minimum of 5 years. There were 60 women
and seven men. The mean age at the index operation was
63 years (range, 41–87 years), mean weight 53.9 kg (range,
38–85 kg), mean height 152 cm (range, 138–172 cm), and
mean BMI 23.2 kg/m2 (range, 16.2–34.7 kg/m2). The
primary diagnosis was osteoarthritis in 61 hips, osteone-
crosis in five, rheumatoid arthritis in four, and sequelae of
pyogenic infection in two. The minimum duration of fol-
lowup was 5 years (mean, 6.7 years; range, 5–8.3 years)
(Table 1). The study was approved by the Institutional
Review Board of the hospital, and all patients provided
informed consent.
The acetabular components consisted of a spherical
cemented polyethylene socket with an alumina inlay (ABS;
Kyocera) without a flange (Fig. 1A) in 20 hips and the
same socket with a flange (Fig. 1B) in 52 hips. The outer
diameter of the acetabular component was 42 mm in 14
hips, 44 mm in 42, 46 mm in eight, 48 mm in seven, and
50 mm in one. The femoral component was a tapered
collarless titanium stem (KC stem; Kyocera). The stem was

Fig. 1A–C A spherical cemented polyethylene-alumina composite c

cup, ABS cup (Kyocera, Kyoto, Japan), was developed to obtain
stability between the alumina cup and bone cement. There were two
designs for this cup: (A) one is the ABS cup without a flange and (B)
the other is the ABS cup with a flange. (C) A cross section of the ABS
cup is shown. The thickness of the alumina inlay is fixed at 4 mm in
any size of acetabular component.

123
1188 Iwakiri et al.
Table 1. Demographics of the 67 patients
Variable Value
Age at surgery (years)* 63 (41–87)
Gender (male/female) 7/60
Body mass index (kg/m2)*
Male patients 22.5 (16.4–30.5)
Female patients 23.3 (16.2–34.7)
Preoperative diagnosis (number of hips)
Osteoarthritis resulting from 61 (91%)
developmental dislocation of the hip
Osteonecrosis 5 (7%)
Rheumatoid arthritis 4 (6%)
Sequelae of pyogenic infection 2 (3%)
*Values expressed as mean, with range in parentheses.
tapered only in the anteroposterior projection. The proxi-
mal end of the stem was coated with macrotexture. The
neck-shaft angle of the stem was 130°. The standard offset
was 40 mm in 42 hips, and the offset of the narrow stem
was 35 mm in 30 hips. The femoral component was fixed
with Simplex P1 cement (Stryker Howmedica Osteonics,
Allendale, NJ). A 28-mm alumina femoral head (Kyocera)
was used. A short neck-head component was used in 18
hips, a medium neck in 50, and a long neck in four. All
operations were performed by one surgeon (HO) through
an anterolateral approach in which an attempt was made to
place the acetabular component in an anatomic position.
All patients received intravenous antibiotic prophylaxis
preoperatively and for 3 days after surgery. The patients
received mechanical prophylaxis for thromboembolism by
intermittent pneumatic compression with the A-V Impulse
System1 (Novamedix, Andover, UK) for 2 days, but no
pharmacologic prophylaxis using warfarin, heparin, or
aspirin was administered. Patients were encouraged to walk
with full weightbearing as tolerated without the aid of
crutches 4 weeks after surgery. Routine followups were
scheduled for 3, 6, 9, and 12 months and yearly thereafter.
We (HI) performed clinical evaluation using the Merle
d’Aubigné and Postel score [16] that allocates up to 6
points for each category of pain, mobility, and function
with a total of 18 points given to a normal hip. Patients
were routinely asked at followup whether they had expe-
rienced audible hip noise because we were concerned about
separation of alumina-on-alumina bearings and alumina
fractures [18, 28].
The radiographic evaluation was performed by one
observer (KI) who did not participate in the index opera-
tions. The 6-month anteroposterior and lateral radiographs
were used for assessment of the abduction angle of the
acetabular socket [17]. On the radiographs at the final
examination, radiolucency and osteolysis were evaluated
123
Clinical Orthopaedics and Related Research
around the acetabular component using the zone classifi-
cation of DeLee and Charnley [9] and around the femoral
component using the criteria of Hodgkinson et al. [15] and
Gruen et al. [11]. Migration of the acetabular and femoral
component center was evaluated by comparing the hori-
zontal and vertical distance from the inferior points of
the teardrops and the center of the lesser trochanter,
respectively, on the immediate postoperative and final
radiographs [20]. Loosening of each component was con-
sidered to have occurred when migration of the component
was greater than 2 mm. Alumina component failure was
checked. Heterotopic ossification was defined according to
Brooker et al. [7].
The mean Merle d’Aubigné and Postel hip score of the
remaining 68 hips improved from 10.1 (pain 2.4, mobility
4.2, function 3.5) before the operation to 16.3 (pain 5.8,
mobility 5.6, function 4.9) at final followup. No patient
used any type of walking support. On 6-month anteropos-
terior radiographs of the 76 hips, the mean abduction angle
of the acetabular component was 37.6° (range, 20°–50°).
On the final radiographs of the remaining 68 hips, we found
radiolucencies around the acetabular component in Zone 1
in 20 hips, in Zone 2 in five, and in Zone 3 in four.
However, there was no osteolysis in any zone around the
acetabular component and no evidence of migration of the
acetabular component. On the femoral side, there was no
radiolucency or osteolysis in any of the hips. No patients
had heterotopic ossification. During the followup, there
were no cases of nerve palsy, deep vein thrombosis, or
pulmonary embolism.
Cumulative survival rates were calculated using the
Kaplan-Meier method with failure defined as the end point
of revision for alumina failure or for any reason. To
compare groups with and without alumina failures in age,
BMI, gender, mobility, function, abduction angle, size of
component, or the existence of radiolucent lines, we used
the nonparametric Mann-Whitney U-test. Fisher’s exact
probability test was used to compare the alumina failure
rate between polyethylene acetabular components with and
without flanges. All analyses were performed with SAS1
software (Version 9.1; SAS Institute Inc, Cary, NC).
Results
We detected four (5.6%) ceramic failures at a mean of
5.6 years (range, 3.5–6.8 years) after the index operation.
Three alumina inlays had fractured and one had dissociated
from its polyethylene-backed acetabular component
(Fig. 2); three had been revised, whereas for the remaining
one, revision surgery was intended. None of the failures
occurred during implantation or in association with trau-
matic episodes. Kaplan-Meier survival analysis revealed a
Volume 466, Number 5, May 2008
Fig. 2A–B The radiographs show alumina inlay failure: (A) fracture
of the alumina inlay occurred in three hips without trauma, and (B)
dissociation of the alumina inlay from the polyethylene shell occurred
in one hip without trauma.
survival rate of 94.4% (95% confidence interval) at
6.8 years with failure defined as revision for alumina
failure and 90.7% (95% confidence interval) at 6.8 years
with revision for any reason as the end point (Fig. 3).
There were no differences in age, BMI, gender, mobil-
ity, function, abduction angle, or size of the component
between the four hips with alumina failure and the
remaining 68 hips without it (p = 0.090, 0.278, 0.512,
0.246, 0.239, 0.607, and 0.912, respectively). However, the
four cases of alumina inlay failure had radiolucent lines
(especially in Zones 1 and 2) to a greater (p = 0.046)
Polyethylene-backed Cemented Alumina THA 1189
Fig. 3 Kaplan-Meier survivorship curves of 72 consecutive polyeth-
ylene-backed alumina-on-alumina THAs are shown. The 8-year
survival rate was 90.7% with revision for any reason and 94.4% with
revision for alumina failure as the end point. Dotted lines indicate the
95% confidence intervals.
extent than the cases without alumina inlay failure. We
observed fretting on the stem neck in one of three fracture
cases or a dissociation case (Fig. 4). There was no differ-
ence (p = 0.307) in alumina failure rate between
polyethylene acetabular components with or without a
flange.
Discussion
Although evolution in technology has improved the
quality of alumina considerably and the risk of alumina
fracture has decreased, concerns still remain regarding
problems related to alumina. Numerous factors might be
related to alumina failure. We explored whether age,
BMI, gender, mobility, function, abduction angle, size of
component, the existence of radiolucent lines, or com-
ponent type with or without a flange were associated with
alumina failure.
Our study was limited by few patients with failure and
therefore low power: we observed four (5.6%) alumina
failures at a mean of 5.6 years (range, 3.5–6.8 years) after
the index operation.
The first alumina-on-alumina THA was performed in
April 1970 by Boutin [6]. During the last two decades, the
mechanical strength of alumina has substantially improved.
The third generation of alumina, in which the ABS cup is
classified, is hot isostatically pressed, laser-marked, and
proof-tested [2]. Compared with the first-generation alu-
mina ceramics, grain size has decreased from 4.2 to 1.8 lm
and burst strength has improved from 46 to 65 kN [2].
These advantages of alumina materials are related to its
distinctive tribologic properties resulting from high scratch
resistance and wettability of the material, both of which
reduce third-body and adhesive wear. There are three other
advantages of alumina: lower linear wear rate than metal-
on-polyethylene articulation [10]; lower concentration of
123
1190 Iwakiri et al.
Fig. 4A–C Fretting by the alumina inlay on the stem neck was observed
on the radiographs of (A) a patient with dissociation and (B) a patient
with a fracture. (C) Macroscopic fretting on the stem neck is shown.
123
Clinical Orthopaedics and Related Research
wear particles in the periprosthetic tissue around the
bearing than metal-on-polyethylene articulation [5]; and
less release of TNF-a, which is one of the main factors
inducing osteolysis, than by polyethylene particles [23].
These advantages are related to avoidance of the acetabular
osteolysis observed with alumina-on-alumina THA [13].
Osteolysis has seldom been reported after alumina-on-
alumina THA, and the few cases reported of this usually
were associated with use of a Mittelmeier total hip system
[29, 32], which is made of first-generation alumina with
large grain size, low density, and high porosity. Each of
these characteristics of first-generation alumina could have
led to the production of a large amount of debris. In
addition, this prosthesis had a poor acetabular component
design that was responsible for a rate of failure as much as
27% by the 26-month followup [21]. Meanwhile, fracture
of the alumina component is one of the disadvantages, and
it remains despite the advance in technology [1, 12, 14, 18,
22, 24, 25, 28, 30, 31].
One alumina-on-alumina THA with a cementless, press-
fit, metal-backed acetabular component yielded a survival
rate of 93.7% at 9 years with revision for any reason as the
end point without any alumina fracture [3]. However,
results observed with alumina-on-alumina THA with
cemented acetabular components rarely have been repor-
ted. The survival rate of cemented alumina-on-alumina
THA with a monoblock alumina acetabular component
for 20 years was reported as 61.2% with failure defined as
revision [13]. The main reason for revision was aseptic
loosening of the acetabular component. Loosening of the
cemented acetabular components was always an acute
event related to debonding of the alumina acetabular
component from the bone cement caused by a mechanical
phenomenon resulting from mismatch of stiffness
between the alumina component and either the bone or
cement [13].
The unique design of the polyethylene-backed acetab-
ular component with an alumina inlay (ABS cup; Kyocera)
was developed in 1998 to ensure sufficient fixation of the
alumina surface with bone cement despite mismatch in
stiffness [4, 14, 22] by using the stable fixation among
bone, cement, and polyethylene; this was used in Japan.
Although there were a few reports regarding cementless
metal-backed polyethylene-alumina composite liner, none
was available for the cemented polyethylene-backed socket
while we used this device. Therefore, we continued to
implant these components until July 2000.We observed
four alumina inlay failures (5.6%), including three frac-
tures and one dissociation, but no alumina head fractures,
and the survival rate was 94.4% when failure was revision
for alumina failure 7 years postoperatively. Cases with
alumina inlay failure had considerably more radiolucent
Volume 466, Number 5, May 2008
lines (especially in Zones 1 and 2) than cases without
failure.
There have been some reports of failures of the poly-
ethylene-alumina composite liner within a cementless
titanium alloy shell [1, 12, 14, 22, 24, 30] in which alumina
inlay dissociation and fracture were caused by impinge-
ment, microseparation, and squeeze force [14, 22, 30]. We
thus speculated the following three disadvantageous fea-
tures in this study led to alumina inlay failure in the unique
acetabular component design modification: (1) use of a
much thinner, 4-mm alumina inlay despite improved
quality of material; (2) the narrow clearance of the alumina
inlay and alumina head (5–35 lm), which generated strong
squeeze force leading to separation of the alumina inlay
from the polyethylene shell; and (3) a relatively narrow
oscillation angle, 120°, which readily produced contact
force leading to dissociation of the alumina inlay insert
from its polyethylene shell or increased the chance of
peripheral chip fracture and subsequent crack propagation
resulting from the brittle alumina material under conditions
of impingement. The radiolucent lines, observed to a
considerable extent in the four cases of alumina inlay
failure, might have been induced by large numbers of
polyethylene particles generated from the interface
between the dissociated or fractured alumina inlay and the
polyethylene shell or by high shear stress in the bone-
cement interface resulting from the strong squeeze force
generated on the alumina-on-alumina surface. In July 2000,
we discontinued use of this type of THA to avoid alumina
failure. We suggest all patients with this type of acetabular
component be followed carefully.
Cemented polyethylene-backed alumina-on-alumina
THA with a composite of alumina inlay had a relatively
high rate of catastrophic alumina inlay failure (5.6%)
during a mean of 6.7 years’ followup. We believe the
socket fixation still needs to be improved.
Acknowledgments We thank A. Kobayashi, MD, PhD, M. Ikebuchi,
MD, Y. Ohta, MD, and R. Sugama, MD, PhD, for assistance and advice
concerning polyethylene-backed, cemented, alumina-on alumina THA.
References
1. Akagi M, Nonaka T, Nishisaka F, Mori S, Fukuda K, Hamanishi
C. Late dissociation of an alumina-on-alumina bearing modular
acetabular component. J Arthroplasty. 2004;19:647–651.
2. Bierbaum BE, Nairus J, Kuesis D, Morrison JC, Ward D. Cera-
mic-on-ceramic bearings in total hip arthroplasty. Clin Orthop
Relat Res. 2002;405:158–163.
3. Bizot P, Banallec L, Sedel L, Nizard R. Alumina-on-alumina
total hip prostheses in patients 40 years of age or younger. J Bone
Joint Surg Br. 2004;86:190–194.
4. Boehler M, Knahr K, Plenk H Jr, Walter A, Salzer M, Schreiber
V. Long-term results of uncemented alumina acetabular implants.
J Bone Joint Surg Br. 1994;76:53–59.
Polyethylene-backed Cemented Alumina THA 1191
5. Bohler M, Mochida Y, Bauer TW, Plenk H Jr, Salzer M. Wear
debris from two different alumina-on-alumina total hip arthro-
plasties. J Bone Joint Surg Br. 2000;82:901–909.
6. Boutin P. [Total arthroplasty of the hip by fritted aluminum
prosthesis: experimental study and 1st clinical applications]
[in French]. Rev Chir Orthop Reparatrice Appar Mot. 1972;58:
229–246.
7. Brooker AF, Bowerman JW, Robinson RA, Riley LH Jr.
Ectopic ossification following total hip replacement: incidence
and a method of classification. J Bone Joint Surg Am. 1973;55:
1629–1632.
8. Catelas I, Petit A, Marchand R, Zukor DJ, Yahia L, Huk OL.
Cytotoxicity and macrophage cytokine release induced by cera-
mic and polyethylene particles in vitro. J Bone Joint Surg Br.
1999;81:516–521.
9. DeLee J, Charnley J. Radiological demarcation of cemented
sockets in total hip replacement. Clin Orthop Relat Res.
1976;121:20–32.
10. Dorlot JM, Christel P, Meunier A. Wear analysis of retrieved
alumina heads and sockets of hip prostheses. J Biomed Mater
Res. 1989;23(A3 suppl):299–310.
11. Gruen T, McNeice G, Amstutz H. ‘‘Modes of failure’’ of
cemented stem-type femoral components: a radiographic analysis
of loosening. Clin Orthop Relat Res. 1979;141:17–27.
12. Ha YC, Koo KH, Jeong ST, Joon Yoo J, Kim YM, Joong Kim
H. Ceramic liner fracture after cementless alumina-on-alumina
total hip arthroplasty. Clin Orthop Relat Res. 2007;458:106–
110.
13. Hamadouche M, Boutin P, Daussange J, Bolander ME, Sedel L.
Alumina-on-alumina total hip arthroplasty: a minimum 18.5-
years follow-up study. J Bone Joint Surg Am. 2002;84:69–77.
14. Hasegawa M, Sudo A, Hirata H, Uchida A. Ceramic acetabular
liner fracture in total hip arthroplasty with a ceramic sandwich
cup. J Arthroplasty. 2003;18:658–661.
15. Hodgkinson J, Shelley P, Wroblewski B. The correlation
between the roentgenographic appearance and operative find-
ings at the bone-cement junction of the socket in Charnley low
friction arthroplasties. Clin Orthop Relat Res. 1988;228:105–
109.
16. Merle d’Aubigné R, Postel M. Functional results of hip arthro-
plasty with acrylic prosthesis. J Bone Joint Surg Am. 1954;36:
451–475.
17. Mulliken BD, Nayak N, Bourne RB, Rorabeck CH, Bullas R.
Early radiographic results comparing cemented and cementless
total hip arthroplasty. J Arthroplasty. 1996;11:24–33.
18. Nevelos J, Ingham E, Doyle C, Streicher R, Nevelos A, Walter
W, Fisher J. Microseparation of the centers of alumina-alumina
artificial hip joints during simulator testing produces clinically
relevant wear rates and patterns. J Arthroplasty. 2000;15:793–
795.
19. Nizard RS, Sedel L, Christel P, Meunier A, Soudry M, Witvoet J.
Ten-year survivorship of cemented ceramic-ceramic total hip
prosthesis. Clin Orthop Relat Res. 1992;282:53–63.
20. Nunn D, Freeman MA, Hill PF, Evans SJ. The measurement of
migration of the acetabular component of hip prostheses. J Bone
Joint Surg Br. 1989;71:629–631.
21. O’Leary JF, Mallory TH, Kraus TJ, Lombardi AV Jr, Lye CL.
Mittelmeier ceramic total hip arthroplasty: a retrospective study.
J Arthroplasty. 1988;3:87–96.
22. Park YS, Hwang SK, Choy WS, Kim YS, Moon YW, Lim SJ.
Ceramic failure after total hip arthroplasty with an alumina-on-
alumina bearing. J Bone Joint Surg Am. 2006;88:780–787.
23. Petit A, Catelas I, Antoniou J, Zukor DJ, Huk OL. Differential
apoptotic response of J774 macrophages to alumina and ultra-
high-molecular-weight polyethylene particles. J Orthop Res.
2002;20:9–15.
123
1192 Iwakiri et al.
24. Ravasi F, Sansone V. Five-year follow-up with a ceramic sand-
wich cup in total hip replacement. Arch Orthop Trauma Surg.
2002;122:350–353.
25. Sedel L. Evolution of alumina-on-alumina implants: a review.
Clin Orthop Relat Res. 2000;379:48–54.
26. Sedel L, Nizard R, Kerboull L, Witvoet J. Alumina-alumina hip
replacement in patients younger than 50 years old. Clin Orthop
Relat Res. 1994;298:175–183.
27. Skinner HB. Ceramic bearing surfaces. Clin Orthop Relat Res.
1999;369:83–91.
28. Toni A, Traina F, Stea S, Sudanese A, Visentin M, Bordini B,
Squarzoni S. Early diagnosis of ceramic liner fracture: guidelines
based on a twelve-year clinical experience. J Bone Joint Surg Am.
2006;88(suppl 4):55–63.
123
Clinical Orthopaedics and Related Research
29. Wirganowicz PZ, Thomas BJ. Massive osteolysis after ceramic
on ceramic total hip arthroplasty: a case report. Clin Orthop Relat
Res. 1997;338:100–104.
30. Yamamoto K, Shishido T, Tateiwa T, Katori Y, Masaoka T,
Imakiire A, Clarke IC. Failure of ceramic THR with liner
dislocation: a case report. Acta Orthop Scand. 2004;75:
500–502.
31. Yoo JJ, Kim YM, Yoon KS, Koo KH, Song WS, Kim HJ.
Alumina-on-alumina total hip arthroplasty: a five-year mini-
mum follow-up study. J Bone Joint Surg Am. 2005;87:
530–535.
32. Yoon TR, Rowe SM, Jung ST, Seon KJ, Maloney WJ. Osteolysis
in association with a total hip arthroplasty with ceramic bearing
surfaces. J Bone Joint Surg Am. 1998;80:1459–1468.
Clin Orthop Relat Res (2008) 466:1193–1197
DOI 10.1007/s11999-008-0176-9
ORIGINAL ARTICLE
Late Hardware-induced Sciatic Nerve Lesions After Acetabular
Revision
Martti Vastamäki MD, PhD, Pekka Ylinen MD, PhD,
Asko Puusa MD, Timo Paavilainen MD
Received: 3 September 2007 / Accepted: 4 February 2008 / Published online: 26 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract We encountered late hardware-induced sciatic
nerve lesions after acetabular revision in six patients. There
were five female patients and one male patient. The mean
age of the patients at the time of index acetabular revision
was 59.3 years (range, 42–76 years). The interval from the
index acetabular revision to the onset of sciatic nerve
symptoms averaged 9.4 months (range, 4–16 months) and
that from the onset of symptoms to nerve release was
11.3 months (range, 8–13 months), except in two patients
with intermittent symptoms in which it was 9 and
10.5 years, respectively. Sciatic nerve release was suc-
cessful in two patients, but in four patients, the nerve had
been partly or entirely cut by the metallic hardware
resulting in a permanent deficit. The minimum followup
was 2 years (mean, 4 years; range, 2–7 years). Mechanical
irritation should be suspected in the case of any late sign of
peroneal neuropathy after acetabular revision with a mac-
rocup or antiprotrusion device, and plate fixation of the
posterior column. We recommend exploration and nerve
release before a permanent lesion of the nerve has
developed.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, and that all investigations were
conducted in conformity with ethical principles of research.
M. Vastamäki, P. Ylinen, A. Puusa, T. Paavilainen
ORTON Orthopaedic Hospital, Invalid Foundation, Helsinki,
Finland
M. Vastamäki (&)
Elotie 1 A 2, Kaarina 20780, Finland
e-mail: martti.vastamaki@invalidisaatio.fi
Level of Evidence: Level II, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Nerve lesions after hip arthroplasty are disabling compli-
cations occurring in 0.06% to 2.2% of arthroplasties [4, 7,
10–12, 16]. Most of them are surgery-related and occur
during the procedure or are delayed by a few days [9],
especially when caused by a hematoma [2, 6]. Late nerve
lesions are uncommon and are reported only sporadically
[3, 5]. Late hardware-induced sciatic nerve complications
are rarely reported [1, 8, 14]. In three such reports, a
migrated fragment of a Kirschner wire violated the sciatic
nerve 6 years after THA [1], the posterior flange of a pro-
trusion ring cut the sciatic nerve 3 months after THA [8],
and a loose screw after pelvic reconstruction with a plate
caused sciatic nerve entrapment 6 months after THA [14].
We report six cases in which metallic hardware used for
acetabular revision caused severe late problems to the
sciatic nerve.
Materials and Methods
From 1991 to 1998, we performed 1570 hip revision ar-
throplasties in our hospital. Acetabular revision was
included in approximately 90% of the patients. According
to our practice, the loosened acetabular cup was replaced
by a larger porous-coated press-fit cup whenever possible,
combined with autogeneic or allogeneic bone transplanta-
tion for the cavitary defects. In cases with more severe
segmental defects, we used a reinforcement device with
123
1194 Vastamäki et al. Clinical Orthopaedics and Related Research

allogeneic cancellous bone graft. The Burch-Schneider ring

(Sulzer Medica AG, Winterthur, Switzerland) was most
popular during that period. Plate fixation of the posterior
column also was performed in the patients with a posterior
column fracture or pelvic discontinuity. We encountered
late hardware-induced sciatic nerve lesions in six patients
(0.4%). There were five female patients and one male
patient. The mean age of the patients at the index acetab-
ular revision was 59.3 years (range, 42–76 years). The
interval from the index acetabular revision to the onset
of symptoms averaged 9.4 months (range, 4–16 months)
and that from the onset of symptoms to nerve release was
11.3 months (range, 8–13 months). In two patients,
symptoms were intermittent, and surgery was scheduled 9
and 10.5 years, respectively, after the onset of symptoms.
All patients reported pain and numbness in the peroneal Fig. 1 A radiograph of Patient 1 taken 12 years after a macrocup
revision arthroplasty was interpreted as normal. The arrow shows the
area of the ipsilateral lower limb, and peroneal weakness.
anatomic location prone to nerve lesion.
Some kind of local pain at the hip level also was present.
The minimum followup was 2 years (mean, 4 years; range,
2–7 years).

Patient 1

In 1984, a 40-year-old woman had an uneventful THA

resulting from congenital hip dysplasia-induced arthritis.
She did not have previous hip surgery. She did well for
5 years postoperatively, but in 1991, revision was per-
formed because of loosening of an uncemented threaded
cup. A macrocup (68-mm Universal1; Biomet, Inc,
Warsaw, IN) was implanted. One year after the revision
surgery, the patient reported slight radicular pain in her leg
on the surgically treated side, and in 1994, electroneur-
omyography (ENMG) showed slight neurogenic findings in
Fig. 2 A photograph shows an impingement of the sciatic nerve
the muscles innervated by the peroneal division of the (CHIA) caused by the edge of the acetabular macrocup and scar tissue
sciatic nerve at the level of the hip but no clear injury of the in Patient 1.
sciatic nerve. She did quite well and could walk 2 to 3 km
without difficulties but still reported slight sporadic sciatic connective tissue flap was developed between the ring and
pain. In 2001, the patient experienced temporary severe the nerve (Fig. 3). No resection of the corner of the mac-
numbness and fatigue in her leg. Electroneuromyography rocup was performed to avoid excessive metallic debris
showed slight to moderate neurogenic findings in the formation during the procedure. Sciatic pain resolved, and
muscles innervated by the peroneal division of the sciatic 2 months after surgery, the patient was free of symptoms.
nerve. The injury was localized at the level of the hip. The
plain hip radiographs were interpreted as normal. Magnetic
resonance imaging of the lumbar spine showed only pre- Patient 2
sacral protrusion. Sciatic nerve exploration was postponed
because the symptoms resolved spontaneously. In 2003, A 59-year-old woman with rheumatoid disease since 1974
12 years after the cup revision, disabling sciatic nerve had acetabular revision performed in 1996 with a Burch-
symptoms recurred. Radiographs were interpreted as nor- Schneider ring. Primary THA was performed in 1989.
mal (Fig. 1). However, sciatic nerve injury was apparent During the next 2 years after the revision, a slight peroneal
on ENMG, and surgery was scheduled. The sciatic nerve paresis developed. In December 2005, she experienced
was tightly adherent to the corner of the macrocup and severe radicular pain in her ipsilateral lower limb. Lumbar
severely entrapped (Fig. 2). The nerve was released, and a MRI was normal. On plain radiographs, the distal part of

123
Volume 466, Number 5, May 2008
Fig. 3 A connective tissue flap (arrows) was developed between the
acetabular shell and the nerve shown in Figure 2.
Fig. 4 A radiograph of Patient 2 taken 9 years after acetabular
revision with a Burch-Schneider ring shows some loosening of the
distal plate. The arrow shows the level of the nerve injury.
the plate seemed loose (Fig. 4). Electroneuromyography
suggested a sciatic nerve lesion at the level of the hip; the
peroneal division was severely injured and the tibial divi-
sion moderately injured. In surgery, the distal half of the
plate was loose, the sciatic nerve was adherent to the plate
at the joint level, and there were metallosis and severe
damage of the nerve induced by the corner of the Burch-
Schneider ring. The nerve was released without any
attempt at reconstruction (Fig. 5). The Burch-Schneider
ring was refixed, and the posterior fracture was recon-
structed with a titanium plate (LCP1 Reconstruction Plate,
Synthes AG, Solothurn, Switzerland) and a titanium mesh,
Hardware-induced Sciatic Nerve Lesions 1195
Fig. 5 The sciatic nerve was severely damaged at the joint level by
the Burch-Schneider ring (arrow) in Patient 2. The picture was taken
after reconstruction of the posterior column fracture and before
creation of a protective connective tissue flap.
also using bone transplants. After that, the nerve was
protected with a substantial amount of subcutaneous fat
tissue between the reconstruction plate and the nerve. The
pain resolved, but the peroneal paresis did not recover.
Patient 3
In 1996, a 65-year-old man underwent acetabular revision
with a Burch-Schneider ring. Total hip arthroplasty
resulting from primary arthrosis had been performed
10 years earlier. The patient was free of symptoms for
1 year, but the plate was broken. After 2 years, he was still
free of symptoms. No cup migration was detected, and
bone healing was good. In 2000, the patient reported some
pain and numbness in the peroneal area of the ipsilateral
lower limb. Peroneal weakness was detected. Electro-
neuromyography showed severe nerve injury in the
muscles and nerves innervated by the peroneal division of
the sciatic nerve. The nerve lesion was localized at the
level of the hip. At surgery in 2000, the broken end of the
plate had violated the peroneal division of the sciatic nerve;
.
approximately 1 3 of the nerve was lacerated. The broken
plate was removed, and the nerve was released but not
reconstructed because of the chronic lesion, the age of the
patient, and the long distance to the muscles. The pain
remitted, but the patient had marked peroneal palsy at the
1-year clinical followup, which remained when the patient
was interviewed by telephone 7 years later.
123
1196 Vastamäki et al.
Patient 4
In 1991, a 67-year-old woman had acetabular revision
using a 62-mm Universal1 cup. The primary THA was
performed in 1984 with a Lord prosthesis (Benoist Girard,
Bagneoux, France) as a result of osteoarthrosis. Nine
months after the acetabular revision, hip pain developed,
and during the next 3 months, peroneal palsy appeared.
Electroneuromyography showed a severe lesion of the
peroneal division of the sciatic nerve at the level of the hip
and moderate changes in the tibial division. Surgery was
performed within a few months in 1993. The sharp corner
of the metallic acetabular shell had cut the sciatic nerve
almost entirely. The nerve was only released. After 4 years,
her neuromotor symptoms and signs were unchanged.
Patient 5
In 1996, a 76-year-old woman underwent acetabular revi-
sion with a Burch-Schneider ring 14 years after her
primary THA. Four months after the revision surgery,
sciatic pain developed. There was no muscular weakness.
In May 1997, ENMG showed moderate neurogenic find-
ings, but it was impossible to differentiate the findings
between L5 nerve root lesion and injury of the peroneal
division of the sciatic nerve. Spinal causes of the nerve
lesion were excluded by computed tomography, and the
pain resolved to some extent. However, during the next few
months, peroneal paresis developed, and 1 year after the
index surgery, ENMG showed progression of the nerve
injury in the peroneal division of the sciatic nerve. There
were also slight findings in the muscles innervated by the
tibial division. The nerve lesion could now be localized at
the hip level. At surgery 1 year after acetabular revision,
the sharp corner of the Burch-Schneider plate had cut the
peroneal division of the sciatic nerve. The nerve was
released, and neurorraphy of the peroneal division was
performed by a younger colleague. The pain resolved, but
no recovery was detected during the next 4 years before
her death resulting from unrelated causes.
Patient 6
In 1998, a 42-year-old woman had acetabular revision
with a Burch-Schneider ring. Her primary THA was
performed in 1982 as a result of Perthes disease, and
successive acetabular revisions then were performed in
1989, 1994, and 1995 as a result of recurrent dislocations.
She also had iatrogenic injury of the superior gluteal
nerve diminishing the abduction strength of the hip. The
patient began to report some pain from 2002 to 2004 in
123
Clinical Orthopaedics and Related Research
her ipsilateral lower extremity, and acetabular revision
was again scheduled in 2004 as a result of loosening of
the cup. During surgery, the broken distal plate of the
Burch-Schneider ring was removed, and the sciatic nerve
was released. The sciatic nerve was severely tethered to
the broken plate and compressed by the plate and scar
tissue. Postoperatively, Grade 3/5 peroneal palsy was
detected, but it recovered substantially during the next
3 years. In all likelihood, the patient was treated before
severe damage of the sciatic nerve had developed from
the hardware.
Discussion
A clinical neurologic deficit is not uncommon after THA,
occurring in 0.06% to 2.2% of arthroplasties [1, 4, 7, 10–
12, 16]; many are related to hematomas or technical errors.
In the series of 4339 THAs performed at our institute, we
have observed only 27 sciatic nerve lesions (0.6 %) [11]. In
another series of 3126 THAs, 42 sciatic nerve lesions
occurred (1.3 %) [12]. Few articles describe late neuropa-
thies, especially those induced by hardware. Masses from
particulate debris related to acetabular loosening [5] and
migration of a trochanteric wire after THA [1] can cause
late sciatic neuropathy. An acetabular reinforcement ring
has been reported to cause impingement of the sciatic nerve
between the dorsal aspect of the acetabular reinforcement
ring and scar tissue [13]. We detected similar findings in
two patients (Patients 1 and 4). In one reported case [8], an
antiprotrusion ring caused a sciatic nerve lesion 3 months
after THA when the patient fell on his involved hip. The
.
sciatic nerve was cut through by approximately 1 3 appar-
ently by the prominent edge of the nonanatomic
antiprotrusion ring [8]. The nerve was released approxi-
mately 10 cm proximally and distally from the hip and
protected using a posterior third of the gluteus medius [8].
Our Patients 2, 3, and 5 showed similar lesions of the
sciatic nerve. In another case [14], a loose screw of the
reconstruction plate after THA caused sciatic nerve prob-
lems 6 months after surgery. Screw removal and nerve
release 6 months after the onset of symptoms enabled
substantial improvement in clinical symptoms [14]. We
observed similar findings in one of our patients (Patient 6).
It is important to recognize the cause of symptoms even
years after THA and especially after acetabular revision.
Our first patient was fortunate enough to be treated before a
permanent lesion had developed. Also, Patient 6 had sur-
gery early enough as a result of the need for revision
arthroplasty. At an early stage, it may be difficult to dis-
tinguish between spine- and hip-related causes of
symptoms, and therefore lumbar imaging modalities might
be necessary. We did not find in our material any alerting
Volume 466, Number 5, May 2008
radiographic signs, except breakage or loosening of a
reconstruction plate.
Mechanical irritation should be suspected in cases of
any late sign of peroneal neuropathy after acetabular
revision with a macrocup or antiprotrusion device and after
plate fixation of the posterior column. In posterior column
fractures, we have used a locking compression titanium
reconstruction plate with combined holes. A straight plate
is bent to meet the anatomy of the posterior wall of the
acetabulum. A reconstruction plate seems to cause a bigger
risk to bring about nerve problems than a macrocup. Ten to
15 years ago, we used antiprotrusion ring, macrocup, or
plate fixation only in approximately 10% of all acetabular
revision cases. If the posterior acetabular wall is so
defective that a macrocup extends over the bone rim, the
risk seems greater. The most susceptible place for nerve
injury is the posterior lateral corner of the macrocup or
reconstruction plate in the proximal part of the posterior
acetabular wall.
The sciatic nerve should always be identified during
THA, preferably by palpation [15]. We now routinely
expose the sciatic nerve during revision surgery and
develop a protecting connective tissue flap between the
nerve and hardware if needed. Ten years ago, that was not
routine. We have not observed any adverse effects with
nerve exposure, although harmful irritation by hardware
would be possible, especially without connective tissue flap
protection.
One should always keep in mind the potential risk for
nerve entrapment or lesion when using antiprotrusion
devices or when a posterior column has been reconstructed,
and moreover, if a macrocup has been used in acetabular
reconstruction. In the case of sciatic nerve symptoms,
careful repetitive ENMG examinations are mostly decisive.
Exploration and nerve release should be performed before
a permanent lesion of the nerve has developed. We rec-
ommend using a connective tissue flap between the sciatic
nerve and metallic hardware to minimize the risk of late
nerve lesions.
Hardware-induced Sciatic Nerve Lesions 1197
References
1. Asnis SE, Hanley S, Shelton PD. Sciatic neuropathy secondary to
migration of trochanteric wire following total hip arthroplasty.
Clin Orthop Relat Res. 1985;196:226–228.
2. Cohen B, Bhamra M, Ferris BD. Delayed sciatic nerve palsy fol-
lowing total hip arthroplasty. Br J Clin Pract. 1991;45:292–293.
3. Edwards MS, Barbaro NM, Asher SW, Murray WR. Delayed
sciatic palsy after total hip replacement: case report. Neurosur-
gery. 1981;9:61–63.
4. Eftekhar NS, Stinchfield FE. Experience with low-friction
arthroplasty: a statistical review of early results and complica-
tions. Clin Orthop Relat Res. 1973;95:60–68.
5. Fischer SR, Christ DJ, Roehr BA. Sciatic neuropathy secondary
to total hip arthroplasty wear debris. J Arthroplasty. 1999;14:
771–774.
6. Fleming RE Jr, Michelsen CB, Stinchfield FE. Sciatic paralysis: a
complication of bleeding following hip surgery. J Bone Joint
Surg Am. 1979;61:37–39.
7. Johanson NA, Pellicci PM, Tsairis P, Salvati EA. Nerve injury in
total hip arthroplasty. Clin Orthop Relat Res. 1983;179:214–222.
8. McLean M. Total hip replacement and sciatic nerve trauma.
Orthopedics. 1986;9:1121–1127.
9. Navarro RA, Schmalzried TP, Amstutz HC, Dorey FJ. Surgical
approach and nerve palsy in total hip arthroplasty. J Arthroplasty.
1995;10:1–5.
10. Oldenburg M, Muller RT. The frequency, prognosis and signifi-
cance of nerve injuries in total hip arthroplasty. Int Orthop.
1997;21:1–3.
11. Pekkarinen J, Alho A, Puusa A, Paavilainen T. Recovery of
sciatic nerve injuries in association with total hip arthroplasty in
27 patients. J Arthroplasty. 1999;14:305–311.
12. Schmalzried TP, Amstutz HC, Dorey FJ. Nerve palsy associated
with total hip replacement: risk factors and prognosis. J Bone
Joint Surg Am. 1991;73:1074–1080.
13. Schuh A, Riedel F, Cralovan B, Zeiler G. Verzögerte Läsion des
Nervus ischiadicus nach Totalendoprosthesenimplantaion des
Huftgelenkes mit Stutzringosteosynthese. Zentralbl Chir. 2003;
128:871–873.
14. Stiehl JB, Stewart WA. Late sciatic nerve entrapment following
pelvic plate reconstruction in total hip arthroplasty. J Arthro-
plasty. 1998;13:586–588.
15. Vastamäki M, Paavilainen T. Preventing neurologic complica-
tions of total hip arthroplasty. Complications in Orthopaedics.
1991;6:147–151.
16. Weber ER, Daube JR, Coventry MB. Peripheral neuropathies
associated with total hip arthroplasty. J Bone Joint Surg Am.
1976;58:66–69.
123
Clin Orthop Relat Res (2008) 466:1198–1203
DOI 10.1007/s11999-008-0130-x
ORIGINAL ARTICLE
The Width:thickness Ratio of the Patella
An Aid in Knee Arthroplasty
Farhad Iranpour MD, Azhar M. Merican MS (Orth),
Andrew A. Amis DSc (Eng), Justin P. Cobb MCh, FRCS
Received: 4 September 2007 / Accepted: 10 January 2008 / Published online: 11 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Establishing the appropriate size of the patellar
implant-bone composite is one of the important steps
ensuring functional success in arthroplasty. Convention-
ally, the patella is measured intraoperatively and its
thickness is used to guide the depth of resection. However,
in a diseased joint, this may not reflect the native patellar
thickness. We studied the relationship between the patellar
thickness and various patellar dimensions on three-
dimensional reconstructed computed tomographic scans
from 37 normal adult knees. Patellar width correlated with
thickness. The average patellar width:thickness ratio was
2.0 (standard deviation, 0.106; 95% confidence interval,
1.96–2.03). The cartilage thickness was on average 2.5 mm
(standard deviation, 1.0). The width:thickness ratio was
similar in 79 digital radiographs taken before TKA of
knees without patellofemoral disease (mean, 2.1; standard
deviation, 0.28). When compared with the two other
methods for calculating patellar resection described in the
literature, the width:thickness ratio was more reliable. The
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
F. Iranpour (&), A. M. Merican, J. P. Cobb
Division of Surgery, Oncology, Reproductive Biology
and Anaesthetics, Imperial College London, 7th Floor, Charing
Cross Hospital, Fulham Palace Road, London W6 8RF, UK
e-mail: f.iranpour@imperial.ac
A. A. Amis
Biomechanics Division, Mechanical Engineering Department,
Imperial College London, London, UK
123
width:thickness ratio appears anatomically constant and
may be a useful guide for estimating premorbid patellar
thickness.
Introduction
For surgeons who choose to resurface the patella,
establishing appropriate thickness of the patellar implant-
bone composite during knee arthroplasty is important to
optimize the patellofemoral joint kinematics and to balance
its soft tissues [2]. Knee range of motion can be decreased
by an increased thickness of the patellar prosthesis-bone
composite after knee arthroplasty [5]. Other detrimental
effects of overstuffing the patellofemoral joint, such as
lateral patellar subluxation, increased patellofemoral con-
tact pressure on the lateral condyle, and increased
patellofemoral compression forces, have been reported in
laboratory studies [8, 10, 21].
Surgeons tend to avoid overstuffing the patellofemoral
joint by resecting the amount of bone that corresponds to
the thickness of the patellar implant. In other words, the
final patellar bone-prosthesis composite thickness is
intended to match the original patellar thickness before
surgery. However, the thickness is difficult to estimate
when it has been reduced substantially by the wearing
process. In advanced cases, the patella may be excavated
and the median ridge altered. In the most severe cases, the
patella will be quite thin and will not reflect the original
thickness.
The preresection thickness of the patella typically is
measured as the anteroposterior dimension from the anterior
surface of the patella to the deepest part of the median ridge
of the patella [18]. To address patellae with marked articular
surface wear, two principal methods of reconstructing the
Volume 466, Number 5, May 2008
patella without reference to the articular surface have been
described: (1) the lateral facet subchondral bone thickness
method [9, 18], in which all bone that is farther from the
anterior surface than the shallowest part of the patella –
typically on the lateral facet - is resected, and (2) the tendon-
capsule method, in which all bone is resected deep to the
posterior limit of the quadriceps tendon and the patellar
tendon attachment [12, 13, 15] or the capsular attachment
onto the patella [4].
We first asked whether a reliable relationship could be
found between the thickness of the patella and any of its
other dimensions that might allow the native thickness of a
worn patella to be predicted from the nonarticular parts.
Secondarily we asked whether any predictive method we
explored would be superior to the two described in the
literature.
Materials and Methods
We obtained 37 computed tomographic (CT) scans of
knees from 21 female and 16 male patients older than
55 years (range, 55–70 years) without patellofemoral dis-
ease from the contralateral knees of active people whose
other knee was part of a study of unilateral medial com-
partmental arthritis. Computed tomographic scans were
obtained using an established protocol that reduced the
total radiation exposure to 0.7 mSev, the same radiation
dose as for a long-leg standing film [7]. Three-dimensional
images were reconstructed using computer software; the
surface models enabled manipulation of the images and
measurements.
Fig. 1A–H Three-dimensional images were reconstructed from CT
scans of the knees in extension; software generates all figures
automatically. (A) Posterior reconstruction with the black arrow
shows the length of the medial ridge; (B) the sagittal view with the
white arrow shows the length of the patella; (C) axial and (D) coronal
The Width:thickness Ratio of the Patella 1199
A robust method for aligning the patella was necessary
to produce reliable and reproducible measurements. The
anterior surface of the patella was established by fitting a
plane to multiple points on this surface. We aligned the
patella with this plane vertically and the most posterior
points in seven axial images from superior to inferior were
used to define the deepest points on the median ridge. The
patellar median ridge was simplified by fitting a line to
these points. After aligning the patella with the anterior
plane horizontal and the median ridge line in screen, the
patellar thickness was measured from the anterior surface
of the patella to the median ridge at the proximodistal
center of the median ridge. At this level, the patellar width
was measured. We also made measurements of the patellar
length, lateral:medial facet ratio, and length of the median
ridge (Fig. 1).
To establish precision of the measurements, we
determined the agreement between two different observers
for patellar width and thickness in 20 measurements. We
assessed interobserver agreement by Bland-Altman graphs
and intraclass correlation coefficient [11, 17] for 27 knees.
The intraclass correlation coefficient was 0.95 and 0.97
for the thickness and width, respectively. We found no
systematic biases between observers and the difference
between the readings from both observers was 0.04
(standard deviation [SD], 1.07) and 0.2 mm (SD, 0.78) for
width and thickness, respectively.
The patella can be resected at the level of the deep limit
of the quadriceps tendon attachment and nearly posterior to
the attachment of the patellar tendon (tendon method) [15]
or at the level of the subchondral bone of the lateral facet
(subchondral method) [9, 18]. However, neither of these
views with two-headed arrows show the widths of the medial and
lateral facets; (E) sagittal reconstruction is shown; (F) an axial view
with a black line shows patellar width; (G) a coronal view shows...;
and (H) this sagittal view with white line shows patellar thickness.
123
1200 Iranpour et al.
Fig. 2A–D These images illustrate the method of resecting the
patella at the level of the quadriceps tendon attachments (tendon-
capsule method). (A) A sagittal reconstruction is shown; (B) an axial
image with the two-headed black arrow shows the amount to be
resected; (C) a coronal view at the level of the patellar tendon is
shown; and (D) a sagittal image with a two-headed black arrow shows
the amount to be resected; the attachment of the patella tendon also
can be seen.
methods is precise in relation to the depth of the patella. To
virtually reproduce the tendon method, we used a sagittal
image at the median ridge to locate the posterior (deep)
limit of the quadriceps tendon by noting when the
Hounsfield units just outside the bone changed from fat to
fascia-tendon (Fig. 2). To virtually reproduce the sub-
chondral bone method, after aligning the patella with the
anterior surface horizontally, the axial image at the center
of the median ridge was used to define a plane parallel to
the anterior surface but just down to the subchondral bone
of the lateral facet (Fig. 3). We then measured the thick-
ness of the patellar bone to be resected using these two
methods as references for the depth of patellar resection.
We determined the cartilage thickness at the point from
which the patellar bony thickness was measured. This was
possible on a CT scan because, in the extended relaxed
knee, the articular surface of the patella rests on the
supracondylar area with intervening fat between it and the
underlying femur. Thus, the fat-cartilage junction was
located where the Hounsfield unit changes from negative
(fat) to positive (cartilage).
In addition, we studied the preoperative axial
radiographs of 79 patients undergoing TKA. These patients
had minimal radiographic changes of their patellofemoral
joint. We measured the bony width and thickness of the
patella on a digital axial radiograph using Centricity1
123
Clinical Orthopaedics and Related Research
Fig. 3A–D These images show the method to simulate resection of
the patella at the level of the subchondral bone of the lateral facet. (A)
A sagittal reconstruction is shown; (B) an axial image with a two-
headed black arrow shows the amount to be resected; (C) a coronal
view shows the subchondral bone of the lateral facet; and (D) a
sagittal image with a two-headed black arrow shows the amount to be
resected.
Picture Archiving and Communication Systems (PACS)
3.0 software (GE Healthcare, Chalfont St Giles, UK).
We used Spearman’s rho correlation to ascertain any
relationships between patellar thickness and the other
measurements (patellar width, patellar length, patellar
ridge length). We used the Statistical Package for Social
Sciences (SPSS) Version 13 (SPSS Inc, Chicago, IL).
Results
Patellar thickness correlated (r = 0.89, p \ 0.001) with
its width (Tables 1, 2; Fig. 4). The ratio of the width of
the patella to the thickness was 2.0 ± 0.106 (mean ± SD)
(95% confidence interval, 1.96–2.03). At the point from
which the thickness of the patella was measured, the
mean cartilage thickness was 2.5 ± 1.0 mm (95% confi-
dence interval, 1.8–3.70 mm). However, we found no
correlation between the length and width of the patella or
the length and thickness of the patella (Table 2). The
average ratio of the lateral facet to medial facet width was
1.3 (range, 0.8–1.6).
The width:thickness ratio was the most reliable of the
three ways of restoring native thickness in normal knees.
The alternative methods based on the lateral facet or the
tendon attachments were substantially less reliable in
Volume 466, Number 5, May 2008 The Width:thickness Ratio of the Patella 1201

Table 1. Patellar dimensions of 37 knees

Dimensions Mean Standard deviation Range 95% Confidence intervals

Width (mm) 44.8 4.8 36.8–53 43.2–46.4

Thickness (mm) 22.4 2.3 18.4–27.3 21.7–23.2
Length (mm) 34.3 3.8 24.3–39.2 33.0–35.5
Ridge length (mm) 23.7 2.3 18.7–28.6 22.9–24.5

Table 2. Correlation coefficients for patellar dimensions of 37 knees Discussion

Dimensions Ridge length Length Thickness Width
The disadvantages of overstuffing the patellofemoral joint
Width 0.68* 0.40 0.89* 1 are well recognized [2, 5, 8, 10, 21]. In addition, a thin
Thickness 0.52* 0.22 1 patella resulting from over-resection has had poor strain
Length 0.65* 1 characteristics, which may contribute to early failure [14,
Ridge length 1 19]. Our primary aim was to investigate whether a reliable
* Highly significant correlation. relationship could be found between the thickness of the
patella and any of its other dimensions. Our secondary aim
was to determine whether our predictive method was more
28.00 accurate in reproducing patellar thickness than the two
main methods previously described.
26.00 The major shortcoming of this study is the fact that these
scans were obtained from patients whose other knee had
medial compartment osteoarthritis. This potentially will
Thickness

24.00
22.00
20.00
18.00
36.00 39.00 42.00 45.00 48.00 51.00 54.00
Width
Fig. 4 The regression line shows the relationship when patellar width
was 2.0 times patellar thickness. There was a strong relationship
between the patellar thickness and its width (r = 0.89, p \ 0.001;
thickness = 0.44; width + 2.8).
restoring normal patella thickness when a single thickness
of patella implant was used (Fig. 6). When comparing the
three methods, the width:thickness ratio allowed the native
thickness to be restored with an average of 0.1 mm and a
standard deviation of 1 mm. On average, the tendon
method removed 7.5 mm of bone and the subchondral
method removed 9.4 mm of bone (Table 3). The influence
of these methods on the final thickness of the bone-
prosthesis composite depends on the range of components
available. The ratio of patellar width:thickness measured
on axial radiographs was on average 2.1. The correlation
between width and thickness of the patella was good
(r = 0.63, p \ 0.001) (Fig. 5). The variability of this ratio
in the study group was small (SD, 0.28; 95% confidence
interval, 2.07–2.19).
bias our findings because medial osteoarthritis is a common
variant that is usually symmetric. However, these knees
had not yet succumbed to obvious osteoarthritis by their
sixth decade, although some early changes may be devel-
oping. We measured only patellae whose articular cartilage
remained of normal thickness in active people older than
55 years, so we believe the patellae were essentially nor-
mal. The study is essentially preclinical: we were not
measuring these distances in the operating theater. We used
CT scans rather than radiographs for clinical measure-
ments. However, the method, based on reliably oriented
patellae in three dimensions, is likely to be as accurate as
any method using calipers because there is user variability
in the use of manual measurement devices and intervening
soft tissue can overestimate dimensions. Measurements
obtained in this study, based on three-dimensional images
and the use of Hounsfield units to correctly identify bony
limits, further improve the repeatability of this observation.
This is reflected in the agreement between observers.
The width of the patella appears to be a reliable
indicator for predicting normal patellar thickness. We
found, for normal patellae unaffected by erosive disease
changes, the thickness was ½ of the maximum width. This
simple ratio is independent of damage to the articular
surface and may help the surgeon when deciding on the
thickness of the patella-prosthesis composite during
arthroplasty. The relationship between patellar thickness
and width has not been described, and the low variability in
this measurement is surprising. Two other recent studies

123
1202 Iranpour et al.
Table 3. Thickness of the resected part of the patella
Method of resection Mean (mm)
Subchondral bone of the lateral facet 9.4
Quadriceps tendon 7.5
35.0
30.0
Thickness
25.0
20.0
15.0
10.0
35.0 40.0 45.0 50.0 55.0 60.0
Width
Fig. 5 Patellar thickness and width were measured from digital axial
radiographs of 79 knees. There was good correlation between the
patellar thickness and its width (r = 0.63, p \ 0.001; thick-
ness = 0.40; width + 4.0).
that measured dimensions of the patella during surgery
reported width and thickness [1, 20]. They did not report a
ratio, but if one were to calculate a ratio based on their
average measurements, it would be comparable to ours.
Moreover, they did not emphasize the reliable and constant
relationship nor did they highlight its usefulness.
The width:thickness ratio was more reliable than the
two published conventional methods, especially when
the median ridge is considerably worn away by disease, the
feature most commonly used by surgeons to gauge the
preresection thickness. The alternative methods based on
the lateral facet or the tendon attachments were substan-
tially less reliable in restoring normal patella thickness
when a single thickness of patella implant was used (Fig. 6).
The simple 2:1 ratio for estimating the patellar thickness
from its width does not take into account the thickness of
the patellar cartilage. The patellar cartilage thickness is
approximately 4 mm [3, 6], although there is progressive
thinning after the age of 50 years [16], presumably a nor-
mal aging process. The cartilage thickness in patients in
our CT-based study was on average 2.5 mm and one pos-
sible reason may be all of these patients were older than
60 years. This is more reflective of the population who
undergo knee arthroplasty. Therefore, strictly speaking, to
restore normal patellar thickness, 2.5 mm should be added
to ½ the width of the patella. However, if one considers the
123
Clinical Orthopaedics and Related Research
Standard deviation Range 95% Confidence intervals
1.4 13–6.3 8.9–9.8
1.4 5.3–10.3 7.0–7.9
Patellar Resection Using Three Methods
Overstuff (mm)
2.5
0.0
Understuff (mm)
-2.5
65.0
Quadriceps Tendon Lateral Facet Method Using
Method Method Width:thickness Ratio
Fig. 6 This box plot shows the postoperative thickness of the patella
using each method and a 9-mm patella prosthesis.
ideal thickness of the cement mantle and the fact that
cartilage is more compressible than the prosthesis, it may
be more practical to use the original ratio [6, 7, 9, 12, 21].
The width:thickness ratio of the patella is not the only
variable in reconstructing the patella during arthroplasty; it
clearly will be impacted by different designs of femoral
trochlea and patellar button, which more or less reproduce
the natural morphologic features. In grossly abnormal
patellofemoral joints, such as those with primary patel-
lofemoral arthritis secondary to trochlea dysplasia, we do
not yet know whether changing an abnormal patella into
one with a normal width:thickness ratio is desirable or
appropriate. In these difficult cases, there is a real risk of
overstuffing the joint, but this ratio gives the surgeon some
numeric ground rules to start from. However, the majority
of patellae resurfaced in the course of total condylar knee
arthroplasty will not have substantial morphologic disor-
ders of the patellofemoral joint. For these cases, a
width:thickness ratio of 2 may be a starting point on which
to base decisions in reconstructive surgery of the patella.
Acknowledgments We thank Dr. Robin Richards for technical
support and designing the three-dimensional image analysis software
that was used in this study.
Volume 466, Number 5, May 2008
References
1. Baldwin JL, House CK. Anatomic dimensions of the patella
measured during total knee arthroplasty. J Arthroplasty.
2005;20:250–257.
2. Bengs BC, Scott RD. The effect of patellar thickness on intra-
operative knee flexion and patellar tracking in total knee
arthroplasty. J Arthroplasty. 2006;21:650–655.
3. Cohen ZA, Mow VC, Henry JH, Levine WN, Ateshian GA.
Templates of the cartilage layers of the patellofemoral joint and
their use in the assessment of osteoarthritic cartilage damage.
Osteoarthritis Cartilage. 2003;11:569–579.
4. Fitzpatrick C, FitzPatrick D, Lee J, Auger D. Statistical analysis
of patellar resurfacing in Caucasian and Japanese subjects.
J Biomech. 2006;39(suppl 1):519.
5. Greenfield MA, Insall JN, Case GC, Kelly MA. Instrumentation
of the patellar osteotomy in total knee arthroplasty: the rela-
tionship of patellar thickness and lateral retinacular release. Am J
Knee Surg. 1996;9:129–131; discussion 131–132.
6. Hardy PA, Nammalwar P, Kuo S. Measuring the thickness of
articular cartilage from MR images. J Magn Reson Imaging.
2001;13:120–126.
7. Henckel J, Richards R, Lozhkin K, Harris S, Baena FM, Barrett
AR, Cobb JP. Very low-dose computed tomography for planning
and outcome measurement in knee replacement: the Imperial
knee protocol. J Bone Joint Surg Br. 2006;88:1513–1518.
8. Hsu HC, Luo ZP, Rand JA, An KN. Influence of patellar thickness
on patellar tracking and patellofemoral contact characteristics
after total knee arthroplasty. J Arthroplasty. 1996;11:69–80.
9. Insall JN. Surgical techniques, instrumentation in total knee
arthroplasty. In: Insall JN, ed. Surgery of the Knee. New York,
NY: Churchill Livingstone; 1993:767–768.
10. Jiang CC, Chen CH, Huang LT, Liu YJ, Chang SL, Wen CY, Liu
TK. Effect of patellar thickness on kinematics of the knee joint.
J Formos Med Assoc. 1993;92:373–378.
The Width:thickness Ratio of the Patella 1203
11. Lamraski G, Monsaert A, De Maeseneer M, Haentjens P. Reli-
ability and validity of plain radiographs to assess angulation of
small finger metacarpal neck fractures: human cadaveric study.
J Orthop Res. 2006;24:37–45.
12. Ledger M, Shakespeare D, Scaddan M. Accuracy of patellar
resection in total knee replacement: a study using the medial
pivot knee. Knee. 2005;12:13–19.
13. Lee TQ, Kim WC. Anatomically based patellar resection cri-
teria for total knee arthroplasty. Am J Knee Surg. 1998;11:
161–165.
14. Lie DT, Gloria N, Amis AA, Lee BP, Yeo SJ, Chou SM. Patellar
resection during total knee arthroplasty: effect on bone strain and
fracture risk. Knee Surg Sports Traumatol Arthrosc.
2005;13:203–208.
15. Lombardi AV Jr, Mallory TH, Maitino PD, Herrington SM,
Kefauver CA. Freehand resection of the patella in total knee
arthroplasty referencing the attachments of the quadriceps tendon
and patellar tendon. J Arthroplasty. 1998;13:788–792.
16. Meachim G, Bentley G, Baker R. Effect of age on thickness of
adult patellar articular cartilage. Ann Rheum Dis. 1977;36:563–
568.
17. Pynsent PB. Choosing an outcome measure. J Bone Joint Surg
Br. 2001;83:792–794.
18. Rand JA. Total knee arthroplasty: techniques. In: Morrey BF, ed.
Joint Replacement Arthroplasty. New York, NY: Churchill
Livingstone; 1991:1002–1003.
19. Reuben JD, McDonald CL, Woodard PL, Hennington LJ. Effect
of patella thickness on patella strain following total knee
arthroplasty. J Arthroplasty. 1991;6:251–258.
20. Rooney N, Fitzpatrick DP, Beverland DE. Intraoperative knee
anthropometrics: correlation with cartilage wear. Proc Inst Mech
Eng H. 2006;220:671–675.
21. Star MJ, Kaufman KR, Irby SE, Colwell CW Jr. The effects of
patellar thickness on patellofemoral forces after resurfacing. Clin
Orthop Relat Res. 1996;322:279–284.
123
Clin Orthop Relat Res (2008) 466:1204–1209
DOI 10.1007/s11999-008-0195-6
ORIGINAL ARTICLE
Hybrid Total Knee Arthroplasty
13-year Survivorship of AGC Total Knee Systems with Average 7 Years Followup
Philip M. Faris MD, E. Michael Keating MD,
Alex Farris, John B. Meding MD, Merrill A. Ritter MD
Received: 7 June 2007 / Accepted: 15 February 2008 / Published online: 7 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract A retrospective study of 201 anatomic gradu-
ated component total knee arthroplasties implanted with
hybrid fixation at the authors’ center was performed in
response to conflicting data in the literature concerning the
benefits of a hybrid method. Selection for hybrid fixation
was nonrandomized and based on femoral component fit.
Survivorship analysis was performed, and rates of radio-
lucent lines surrounding the femoral component and
occurrence of osteolysis were noted. At 7 and 13 years,
survivorship with tibial or femoral revision as the end point
was 0.9926 and 0.9732, respectively. Radiolucencies were
found adjacent to 15 femoral components at final followup
(seven in Zone 1, three in Zone 2, five in Zone 3, one in
Zone 4, two in Zone 5, zero in Zone 6). Osteolysis was
observed in one knee after secondary evaluation. Hybrid
fixation in a selected patient population can result in
excellent results in middle to long-term followup.
Level of Evidence: Level IV, prognostic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc.) that might pose a conflict of interest with the
submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent for participation in the study was obtained.
P. M. Faris (&), E. M. Keating, A. Farris,
J. B. Meding, M. A. Ritter
The Center for Hip and Knee Surgery, St Francis Hospital,
1199 Hadley Road, Mooresville, IN 46158, USA
e-mail: pfaris2111@aol.com
123
Introduction
Despite small, but prevalent, incidences of osteolysis [1, 9, 16,
26, 31, 36, 37, 47, 50, 56, 57] (ranging from 4.1% to 34%),
polyethylene wear [10, 14, 15, 32, 38, 41, 53, 55], and lack of
adequate fixation [3, 7, 33, 52, 54], TKA is a highly successful
operation with success rates greater than 90% [4, 12, 17–19, 24,
36, 39, 40, 43, 45, 48, 51]. To further minimize the aforemen-
tioned concerns that lead to TKA failure, various fixation
methods have been proposed [2, 5, 21] and used. Cementless
fixation with porous coating has largely met with mixed results
[8, 11, 12, 13, 17, 19, 20, 27, 39, 40], including a 15-year
survival of 72% [13] and greater loosening among uncemented
tibial components compared with cemented components [7]. In
contrast, short-term studies of hybrid fixation showed promise
for this second alternative [23, 28, 30, 49, 58]. However, one
intermediate-term report [6] of 65 press-fit condylar (PFC)
arthroplasties had unacceptable implant survivorship (89%
after 5 years, 85% after 8 years) and problems with the femoral
component. These problems included six of nine revisions for a
loose femoral prosthesis, two of nine for a fractured femoral
prosthesis, and one for osteolysis, which led the authors of the
study to recommend abandonment of hybrid TKA.
Because of this conflict of data, we examined the results of
a nonrandomized study of the survivorship of hybrid TKA. In
our long-term study performed over a minimum of 2.0 years
(average, 7.9 years; range, 2.0–17.4 years), we examined
the clinical and survivorship results of 201 hybrid TKAs
using one design.
Materials and Methods
Of 403 TKAs performed at our institution between August
24, 1988, and May 17, 1989, 201 (49.9%) were hybrid
Volume 466, Number 5, May 2008
fixation TKAs, all of which used anatomic graduated
components (AGC; Biomet, Warsaw, IN). The femoral
component consisted of a cobalt-chrome articular surface
with a plasma-sprayed titanium undersurface, including
anterior, posterior, distal, and chamfer surfaces. The pegs
were poly grit-blasted. The patellae were all polyethylene,
single-pegged, and the tibial component was a monoblock
titanium-backed component. No other hybrid procedures
were performed after the study period.
The study group comprised 80 women (61.1%) and 51
men (38.9%). Seventy patients (34.8%) received bilateral
hybrid arthroplasties; five others received bilateral
arthroplasties with one knee implanted with hybrid fixa-
tion. The average age of the patients at the time of surgery
was 70.3 ± 8.1 years (mean ± standard deviation; range,
44–87 years). The average body mass index at the time of
surgery was 27.0 ± 4.3 kg/m2 (range, 17.4–39.6 kg/m2).
Diagnoses in this group were osteoarthritis in 181 knees
(90.0%), rheumatoid arthritis in 12 (6.0%), osteonecrosis in
seven (3.5%), and Paget’s disease in one (0.5%). The
preoperative Knee Society score was 51.6 ± 13.7 (range,
7–81); preoperative function score was 49.9 ± 15.8 (range,
10–90); and preoperative pain score was 29.5 ± 11.0
(range, 0–50). Of patients who did not receive hybrid
components during the study period, 83 were women
(69%) and 37 were men (31%). Their average age at the
time of surgery was 73 ± 8.0 years (range, 50–90 years),
and their average body mass index was 29.2 ± 5.6 kg/m2
(range, 19.4–48.9 kg/m2).
During the study period, we performed 202 all cemented
AGC TKAs. Selection for hybrid TKA was based on sur-
geon preference taking into account the coaptation of the
implant; once the trial femoral component was implanted,
the patient received a hybrid prosthesis if the performing
surgeon determined that it fit snugly. Cement was used as
the primary method of fixation in all tibial and patellar
components, and the posterior cruciate ligament (PCL) was
retained in all patients. We used screws in conjunction with
cement in six tibial components to help fill large defects
[42, 44, 46], and the keel was cemented.
We evaluated all patients preoperatively, at 6 months,
and at 1, 3, 5, 7, 10, 12, 15, and 17 years after surgery,
when available. Patient evaluation spanned an average of
7.9 ± 4.9 years (range, 2.0–17.4 years). Radiographs were
taken 2 months postoperatively and at every subsequent
followup. Radiographs were evaluated at followup by the
performing surgeon, who was not blinded, for varus-valgus
alignment and occurrence of radiolucent lines. Primary
evaluation of osteolysis, using the definition given by
Peters et al. [37], was performed after the study period by
two of the authors (PMF, ATF) involved in the study who
were not blinded to patient identification; a secondary
detailed evaluation of possible osteolysis was performed by
Hybrid TKA 1205
another author (MAR), who also was not blinded to patient
identification. Failure was defined as revision of the femoral
or tibial component for reasons other than infection.
We performed statistical analysis using SAS statistical
software (SAS Institute, Inc, Cary, NC). Survivorship rates
were obtained with Kaplan-Meier survival analysis [25]
with failure defined as revision of the tibial or femoral
component.
Results
Kaplan-Meier survival analysis, with failure defined as
revision of either the femoral or tibial component, found
that at 5 years, the followup survival rate was 1.0000; at 7,
10, and 12 years followup, the survival rate was 0.9926;
and at 13 years followup, the survival rate was 0.9732
(Fig. 1). Preoperative knee scores improved from an
average score of 50 to 95 at 1 year followup with similar
improvements in functional scores (Fig. 2).
During the followup period, there were five revisions of
the femoral or tibial component (2.5%), two of which were
not related to an infection (two of 201 [1.0%]). Both aseptic
revisions resulted from medial tibial collapse secondary to
loosening, and both knee systems had a well-fixed femoral
component at revision. The aseptic revisions included one
in a 70-year-old man and one in a 65-year-old woman; both
were treated primarily for osteoarthritis. The two tibial
components were in 90° coronal alignment; the overall
anatomic alignment for one knee was in 5° valgus, whereas
the other was 0°. Both patients had a pain score of 50.
There were three knees (two tibial components, dis-
cussed previously, and one femur) that were considered
radiographically loose. The loose femoral component, in a
62-year-old man treated for osteoarthritis, did not require
revision; the overall alignment was 4° valgus and 87°
coronal tibial alignment. The patient had a pain score of 50.
Although there were radiographic signs of loosening, there
was no change in alignment or migration of the prostheses.
At final followup, we found radiolucencies adjacent to
the femoral component in 15 of 194 knees (7.7%). Seven
radiolucencies were found in Zone 1, three in Zone 2, five
in Zone 3, one in Zone 4, two in Zone 5, and zero in
Zone 6. One hundred twenty-five of 176 knees had no
radiolucency at 1 year followup (71.0%), 117 of 136 had
no radiolucency at 3 years (86.0%), 78 of 87 had
no radiolucency at 5 years (89.7%), 54 of 59 had no
radiolucency at 10 years (91.5%), and 11 of 12 had no
radiolucency at 15 years (91.7%). Radiolucencies greater
than 2 mm at earlier followup, typically found around the
anterior or posterior femoral flanges, did not increase in
size or compromise the implant; all such lucencies did not
progress into adjacent zones.
123
1206 Faris et al. Clinical Orthopaedics and Related Research

Fig. 1 Kaplan-Meier survivor-

ship analysis shows 97.3%
survival at 13 years.

Fig. 2 Average knee society Average Knee Society Scores for Hybrid Fixation (n=201 knees)
scores for hybrid fixation 100
(n = 201 knees) are shown.
90

80

70

60

50

40

30

20

10

0
Preoperative 6 months 1 year 3 years 5 years 7 years 10 years
Knee Score 52 83 93 95 96 89 87
Function Score 50 80 88 86 78 81 79
Pain Score 30 46 48 48 49 47 49
Walk Score 20 43 47 46 47 45 44
Stairs Score 29 38 41 41 35 39 39

On primary radiographic evaluation, we found nine
osteolytic lesions in eight of 201 knees (4.0%), but only
one of these nine lesions (in 0.5% of 201 knees) was
determined to possibly be wear osteolysis after secondary
evaluation; the other eight were small, medial, and non-
progressive. Two lesions were found surrounding a femoral
component, whereas the other seven were found in the
tibia. One femoral lesion was located adjacent to the
anterior flange with the other adjacent to the posterior
flange; these were determined to be related to poor
coaptation of the femoral flanges, and not polyethylene
wear, on secondary evaluation. All seven tibial lesions
were found in the medial tibial plateau. Six of these
appeared within 5 mm of the edge of the plateau; they first
appeared at 6 months followup and were nonprogressive;
thus, they were deemed nonosteolytic at secondary evalu-
ation. The area of concern for osteolysis in the one knee
with possible osteolysis after secondary evaluation was
noted at 3 years followup; the area became more promi-
nent, but it did not expand during the 15-year followup

123
Volume 466, Number 5, May 2008
period. One knee had two lesions that were visualized, one
in the medial tibia and one in the anterior femur but both
were deemed nonosteolytic. All patients with apparent
osteolysis determined at primary evaluation, with the
exception of those who had revision surgery, had a pain
score of 50.
Three supracondylar femoral fractures occurred in the
study group. These fractures, which appeared near the
distal third of the right femur, were treated with Rush rods;
all healed sufficiently. We performed three revision
arthroplasties secondary to infection in two patients within
2 years after the index surgery; these were revised with
modular Insall-Burstein constrained prostheses (Zimmer,
Warsaw, IN). Seven manipulations and three lateral
releases were performed. Other operations included seven
manipulations, three lateral releases, and one patellar
button excision.
Discussion
Cementless fixation in TKA was developed to negate the
problems associated with cemented TKA; however, in light
of the mixed results associated with cementless TKA [11,
12, 13, 17, 19, 20, 27, 39, 40], hybrid fixation was put forth
as an alternative. In initial studies [28, 30, 49, 58], this
method showed success rates comparable with those of the
cemented femoral Miller-Galante and PFC components;
however, the studies relied on short followup and stressed
further study was warranted. Campbell et al. [6] cited 10
revisions among 65 PFC implants and 84.6% survivorship
at 8 years; these findings led them to conclude hybrid
fixation should be abandoned. On seeing these results and
reports [22, 24, 29, 34] suggesting increased success rates
could result from patient selection, we decided to examine
data concerning our center’s experience with hybrid TKAs.
One limitation of this study was the reliance on standard
radiographs to observe osteolysis [23, 35]. However, the
same two technicians took all the radiographs and their
techniques are the same. Moreover, the scarcity of revi-
sions in this study group suggests possible unobserved
osteolysis did not complicate the integrity of the implants.
Previous studies [15, 20, 23, 26, 37] have found an
increased incidence of osteolysis in cementless TKAs
surrounding tibial and femoral components. Specifically,
one study [37] found a 16% occurrence of osteolysis sur-
rounding tibial screws in cementless TKA, and another
[20] found tibial osteolysis in 24 of 113 knees (21%) and
femoral osteolysis in 20 knees (18%). Osteolysis was found
in fewer knees in the current study, in only one of 201
hybrid TKAs after secondary evaluation. Two knees in the
group of 201 (1.0%) showed potential osteolysis around the
cementless femoral component; however, they were
Hybrid TKA 1207
deemed to be a result of failed coaptation of the femoral
flanges because Zones 2, 3, 4, and 6 showed femoral
ingrowth. We believe the overall decreased occurrence of
osteolysis is influenced by the monoblock tibial component
with compression-molded polyethylene, which has been
shown to have decreased wear characteristics [43].
Numerous evaluations of implant designs and fixation
methods [1, 7, 12, 51, 52] have noted the constant but
variable prevalence of radiolucent lines surrounding TKA
components. Lucencies have been found in at least one
zone adjacent to 18 of 55 (32.7%) [12] and five of 58
(8.6%) [51] cementless femoral components in cementless
TKA in groups followed for an average 10 and 11 years,
respectively. Two of the 55 were considered radiographi-
cally loose, whereas none of the 58 was loose at followup.
In our study, we found no radiolucencies at final followup
in 179 of 194 knees (92.3%), with incidence of radiolu-
cency decreasing with each subsequent followup. Only one
femoral component, which did not require revision, was
deemed to be loose at clinical and radiographic followups.
Comparisons among fixation techniques in primary
TKAs have been produced in studies using databases with
greater than 5700 knee arthroplasties [17, 19, 39, 40]. In
two of these reports [19, 40], such comparisons included
the survival rates of cemented, cementless, and hybrid
TKAs. Both found the survivorship of cemented compo-
nents to be superior to the latter two methods, with
cemented implants having success rates of 99% and 92%,
respectively (Table 1). These findings give credence to
cemented fixation’s distinction as the gold standard against
which all other fixation methods are compared. However,
hybrid fixation has consistently shown more success than
cementless techniques, as seen by the 97.3% 13-year suc-
cess rate found in the current study. This is partly
attributable to the improved surgical procedure and
continued, but tempered, evolution of components.
Hybrid TKA has shown good short-term and intermediate-
term results. Although cemented fixation has consistently
had excellent survival rates, careful selection of patients
based on component fit may increase survivorship seen in
hybrid fixation. Further study, especially analysis of greater
than 15-year survival, is warranted and would add consid-
erably to the debate concerning the optimal method of
fixation in primary TKA.
Table 1. Comparison of hybrid TKA results
Study Hybrid
Survival (%) Number
Rand et al., 2003 [40] (10 years) 84 172
Gioe et al., 2004 [19] (11 years) 93.2 837
Faris et al., 2008 (current study) (13 years) 97.3 201
123
1208 Faris et al.
Acknowledgments We thank Matthew Brunsman for statistical
analysis and graphics design.
References
1. Arora J, Ogden AC. Osteolysis in a surface-cemented, primary,
modular Freeman-Samuelson total knee replacement. J Bone
Joint Surg Br. 2005;87:1502–1506.
2. Bauer GC. What price progress? Failed innovations of the knee
prosthesis. Acta Orthop Scand. 1992;63:245–246.
3. Berry DJ, Wold LE, Rand JA. Extensive osteolysis around an
aseptic, stable, uncemented total knee replacement. Clin Orthop
Relat Res. 1993;293:204–207.
4. Bozic KJ, Kinder J, Menegini M, Zurakowski D, Rosenberg AG,
Galante JO. Implant survivorship and complication rates after
total knee arthroplasty with a third-generation cemented system:
5 to 8 years followup. Clin Orthop Relat Res. 2005;430:117–124.
5. Bulstrode CJ, Murray DW, Carr AJ, Pynsent PB, Carter SR.
Designer hips: don’t let your patient become a fashion victim.
BMJ. 1993;306:732–733.
6. Campbell MD, Duffy GP, Trousdale RT. Femoral component
failure in hybrid total knee arthroplasty. Clin Orthop Relat Res.
1998;356:58–65.
7. Carlsson Å, Björkman A, Besjakov J, Önsten I. Cemented tibial
component fixation performs better than cementless fixation: a
randomized radiostereometric study comparing porous-coated,
hydroxyapatite-coated and cemented tibial components over
5 years. Acta Orthop. 2005;76:362–369.
8. Collier JP, Mayor MB, Surprenant VA, Dauphinais LA,
Surprenant HP, Jensen RE, Chae JC, Spector M, Shortkroff S,
Sledge CB, et al. Advances in implant-bone interface. Instr
Course Lect. 1991;40:89–113.
9. Collier MB, Engh CA Jr, McAuley JP, Ginn SD, Engh GA.
Osteolysis after total knee arthroplasty: influence of tibial base-
plate surface finish and sterilization of polyethylene insert. J Bone
Joint Surg Am. 2005;87:2702–2708.
10. Conditt MA, Thompson MT, Usrey MM, Ismaily SK, Noble PC.
Backside wear of polyethylene tibial inserts: mechanism and
magnitude of material loss. J Bone Joint Surg Am. 2005;87:
326–331.
11. Cook SD, Thomas KA, Haddad RJ Jr. Histologic analysis of
retrieved human porous-coated total joint components. Clin
Orthop Relat Res. 1988;234:90–101.
12. Duffy GP, Berry DJ, Rand JA. Cement versus cementless fixation
in total knee arthroplasty. Clin Orthop Relat Res. 1998;356:
66–72.
13. Duffy GP, Murray BE, Trousdale RR. Hybrid total knee arthro-
plasty: analysis of component failures at an average of 15 years. J
Arthroplasty. 2007;22:1112–1115.
14. Engh GA, Dwyer KA, Hanes CK. Polyethylene wear of metal-
backed tibial components in total and unicompartmental knee
prostheses. J Bone Joint Surg Br. 1992;74:9–17.
15. Engh GA, Parks NL, Ammeen DJ. Tibial osteolysis in cementless
total knee arthroplasty: a review of 25 cases treated with and
without tibial component revision. Clin Orthop Relat Res.
1994;309:33–43.
16. Ezzet KA, Garcia R, Barrack RL. Effect of component fixation
method on osteolysis in total knee arthroplasty. Clin Orthop Relat
Res. 1995;321:86–91.
17. Furnes O, Espehaug B, Lie SA, Furnes O, Havelin LI. Early
failures among 7,174 primary total knee replacements: a follow-
up study from the Norwegian Arthroplasty Register 1994–2000.
Acta Orthop Scand. 2002;73:117–129.
123
Clinical Orthopaedics and Related Research
18. Gioe TJ, Bowman KR. A randomized comparison of all-poly-
ethylene and metal-backed tibial components. Clin Orthop Relat
Res. 2000;380:108–115.
19. Gioe TJ, Killeen KK, Grimm K, Mehle S, Scheltema K. Why are
total knee replacements revised?: analysis of early revision in a
community knee implant registry. Clin Orthop Relat Res.
2004;428:100–106.
20. Goldberg VM, Kraay M. The outcome of the cementless tibial
component: a minimum 14-year clinical evaluation. Clin Orthop
Relat Res. 2004;428:214–220.
21. Goodfellow J. Editorials. Knee prostheses—one step forward,
two steps back. J Bone Joint Surg Br. 1992;74:1–2.
22. Hsu RW, Tsai YH, Huang TJ, Chang JC. Hybrid total knee
arthroplasty: a 3- to 6-year outcome analysis. J Formos Med
Assoc. 1998;97:410–415.
23. Huang CH, Ma HM, Liau JJ, Ho FY, Cheng CK. Osteolysis in
failed total knee arthroplasty: a comparison of mobile-bearing
and fixed-bearing knees. J Bone Joint Surg Am. 2002;84:
2224–2229.
24. Illgen R, Tueting J, Enright T, Schreibman K, McBeath A, Heiner
J. Hybrid total knee arthroplasty: a retrospective analysis of
clinical and radiographic outcomes at average 10 years follow-
up. J Arthroplasty. 2004;19(7 suppl 2):95–100.
25. Kaplan EL, Meier P. Nonparametric estimation from incomplete
observations. J Am Stat Assoc. 1958;53:457–481.
26. Kilgus DJ, Funahashi TT, Campbell PA. Massive femoral oste-
olysis and early disintegration of a polyethylene-bearing surface
of a total knee replacement: a case report. J Bone Joint Surg Am.
1992;74:770–774.
27. Kim YH, Oh JH, Oh SH. Osteolysis around cementless porous-
coated anatomic knee prostheses. J Bone Joint Surg Br.
1995;77:236–241.
28. Kobs JK, Lachiewicz PF. Hybrid total knee arthroplasty: two- to
five-year results using the Miller-Galante prosthesis. Clin Orthop
Relat Res. 1993;286:78–87.
29. König A, Kirschner S, Walther M, Eisert M, Eulert J. Hybrid total
knee arthroplasty. Arch Orthop Trauma Surg. 1998;118:66–69.
30. Kraay MJ, Meyers SA, Goldberg VM, Figgie HE II, Conroy PA.
‘‘Hybrid’’ total knee arthroplasty with the Miller-Galante pros-
thesis: a prospective clinical and roentgenographic evaluation.
Clin Orthop Relat Res. 1991;273:32–41.
31. Lachiewicz PF, Soileau ES. The rates of osteolysis and loosening
associated with a modular posterior stabilized knee replacement:
results at 5–14 years. J Bone Joint Surg Am. 2004;86:525–530.
32. Lewis P, Rorabeck CH, Bourne RB, Devane P. Posteromedial
tibial polyethylene failure in total knee replacements. Clin Orthop
Relat Res. 1994;299:11–17.
33. Li MG, Nilsson KG. The effect of the preoperative bone quality
on the fixation of the tibial component in total knee arthroplasty.
J Arthroplasty. 2000;15:744–753.
34. Mitsui H. Hybrid total knee arthroplasties in rheumatoid arthritis.
Bull Hosp Jt Dis. 1993;53:19–20.
35. Miura H, Matsuda S, Okazaki K, Kawano T, Kawamura H,
Iwamoto Y. Validity of an oblique posterior condylar radio-
graphic view for revision total knee arthroplasty. J Bone Joint
Surg Br. 2005;87:1643–1646.
36. O’Rourke MR, Callaghan JJ, Goetz DD, Sullivan PM, Johnston RC.
Osteolysis associated with a cemented modular posterior-cruciate-
substituting total knee design: five to eight-year follow-up. J Bone
Joint Surg Am. 2002;84:1362–1371.
37. Peters PC Jr, Engh GA, Dwyer KA, Vinh TN. Osteolysis after
total knee arthroplasty without cement. J Bone Joint Surg Am.
1992;74:864–876.
38. Puloske SK, McCalden RW, MacDonald SJ, Rorabeck CH,
Bourne RB. Tibial post wear in posterior stabilized total knee
Volume 466, Number 5, May 2008
arthroplasty: an unrecognized source of polyethylene debris.
J Bone Joint Surg Am. 2001;83:390–397.
39. Rand JA, Ilstrup DM. Survivorship analysis of total knee
arthroplasty: cumulative rates of survival of 9200 total knee
arthroplasties. J Bone Joint Surg Am. 1991;73:397–409.
40. Rand JA, Trousdale RT, Ilstrup DM, Harmsen WS. Factors
affecting the durability of primary total knee prostheses. J Bone
Joint Surg Am. 2003;85:259–265.
41. Rao AR, Engh GA, Collier MB, Lounici S. Tibial interface wear
in retrieved total knee components and correlations with modular
insert motion. J Bone Joint Surg Am. 2002;84:1849–1855.
42. Ritter MA. Screw and cement fixation of large defects in total
knee arthroplasty. J Arthroplasty. 1986;1:125–129.
43. Ritter MA, Berend ME, Meding JB, Keating EM, Faris PM,
Crites BM. Long-term follow-up of Anatomic Graduated Com-
ponents posterior cruciate-retaining total knee replacement. Clin
Orthop Relat Res. 2001;388:51–57.
44. Ritter MA, Harty LD. Medial screws and cement: a possible
mechanical augmentation in total knee arthroplasty. J Arthro-
plasty. 2004;19:587–589.
45. Ritter MA, Herbst SA, Keating EM, Faris PM, Meding JB. Long-
term survival analysis of a posterior-cruciate-retaining total
condylar total knee arthroplasty. Clin Orthop Relat Res.
1994;309:136–145.
46. Ritter MA, Keating EM, Faris PM. Screw and cement fixation of
large defects in total knee arthroplasty a sequel. J Arthroplasty.
1993;8:63–65.
47. Robinson EJ, Mulliken BD, Bourne RB, Rorabeck CH, Alvarez C.
Catastrophic osteolysis in total knee replacement: a report of 17
cases. Clin Orthop Relat Res. 1995;321:98–105.
48. Rodriguez JA, Baez N, Rasquinha V, Ranawat CS. Metal-backed
and all-polyethylene tibial components in total knee replacement.
Clin Orthop Relat Res. 2001;392:174–183.
49. Rorabeck CH, Bourne RB, Lewis PL, Nott L. The Miller-Galante
knee prosthesis for the treatment of osteoarthrosis: a comparison
Hybrid TKA 1209
of the results of partial fixation with cement and fixation without
any cement. J Bone Joint Surg Am. 1993;75:402–408.
50. Schmalzried TP, Callaghan JJ. Current concepts review: wear in
total hip and knee replacements. J Bone Joint Surg Am.
1999;81:115–136.
51. Schrøder HM, Berthelsen A, Hassani G, Hansen EB, Solgaard S.
Cementless porous-coated total knee arthroplasty: 10-year
results in a consecutive series. J Arthroplasty. 2001;16:
559–567.
52. Smith S, Naima VS, Freeman MA. The natural history of tibial
radiolucent lines in a proximally cemented stemmed total knee
arthroplasty. J Arthroplasty. 1998;13:3–8.
53. Tanner MG, Whiteside LA, White SE. Effect of polyethylene
quality on wear in total knee arthroplasty. Clin Orthop Relat Res.
1995;317:83–88.
54. Vigorita VJ, Minkowitz B, Dichiara JF, Higham PA. A his-
tomorphometric and histologic analysis of the implant interface
in five successful, autopsy-retrieved, noncemented porous-
coated knee arthroplasties. Clin Orthop Relat Res. 1993;293:
211–218.
55. Wasielewski RC, Galante JO, Leighty RM, Natarajan RN,
Rosenberg AG. Wear patterns on retrieved polyethylene tibial
inserts and their relationship to technical considerations during
total knee arthroplasty. Clin Orthop Relat Res. 1994;299:31–43.
56. Wasielewski RC, Parks N, Williams SI, Surprenant H, Collier JP,
Engh G. Tibia insert undersurface as a contributing source of
polyethylene wear debris. Clin Orthop Relat Res. 1997;345:53–
59.
57. Whiteside LA. Effect of porous-coating configuration on tibial
osteolysis after total knee arthroplasty. Clin Orthop Relat Res.
1995;321:92–97.
58. Wright RJ, Lima J, Scott RD, Thornhill TS. Two- to four-year
results of posterior cruciate-sparing condylar total knee arthro-
plasty with an uncemented femoral component. Clin Orthop
Relat Res. 1990;260:80–86.
123
Clin Orthop Relat Res (2008) 466:1210–1216
DOI 10.1007/s11999-008-0197-4
MULTIMEDIA ARTICLE
Vascular Management in Rotationplasty
Craig R. Mahoney MD, Curtis W. Hartman MD,
Pamela J. Simon RN, B. Timothy Baxter MD,
James R. Neff MD
Received: 22 February 2007 / Accepted: 18 February 2008 / Published online: 18 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract The Van Nes rotationplasty is a useful limb-
preserving procedure for skeletally immature patients with
distal femoral or proximal tibial malignancy. The vascular
supply to the lower limb either must be maintained and
rotated or transected and reanastomosed. We asked whe-
ther there would be any difference in the ankle brachial
index or complication rate for the two methods of vascular
management. Vessels were resected with the tumor in
seven patients and preserved and rotated in nine patients.
One amputation occurred in the group in which the vessels
were preserved. Four patients died secondary to metastatic
disease diagnosed preoperatively. The most recent ankle
brachial indices were 0.96 and 0.82 for the posterior tibial
and dorsalis pedis arteries, respectively, in the
James R. Neff—Deceased.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation, that all investigations were
conducted in conformity with ethical principles of research, and that
informed consent was obtained.
Electronic supplementary material The online version of this
article (doi:10.1007/s11999-008-0197-4) contains supplementary
material, which is available to authorized users.
C. R. Mahoney, C. W. Hartman, P. J. Simon,
B. T. Baxter, J. R. Neff
The Department of Orthopaedic Surgery and Rehabilitation,
University of Nebraska Medical Center, Omaha, NE, USA
C. R. Mahoney (&)
Iowa Orthopaedic Center, 411 Laurel Street, Suite 3300,
Des Moines, IA 50314, USA
e-mail: iowamahoneymd@aol.com
123
reconstructed group. The ankle brachial indices were 0.98
and 0.96 for the posterior tibial and dorsalis pedis arteries,
respectively, in the rotated group. Outcomes appear similar
using both methods of vascular management and one
should not hesitate to perform an en bloc resection when
there is a question of vascular involvement.
Level of Evidence: Level III, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
Borggreve [4] first described rotationplasty in 1930 as a
treatment for limb shortening and knee ankylosis second-
ary to tuberculosis. Van Nes, in the 1950s, popularized this
procedure for correcting c the location of the tumor and
necessary bone resection. From the late 1970s to the
present, the procedure has been popularized as a limb-
preserving procedure for skeletally immature patients with
distal femoral or proximal tibial sarcomas [18, 29]. There
are numerous physical benefits to rotationplasty when
compared with above-knee amputation. Cammisa et al. [6]
compared the energy cost of ambulation in above-knee
amputees with that of patients with a rotationplasty. They
found ambulation by patients after rotationplasty required
38% of maximum aerobic capacity whereas above-knee
amputees required 46% [6]. Rotationplasty maintains a
functioning joint at the level of the knee. This results in a
more efficient gait pattern and decreased oxygen con-
sumption [2, 6, 29].
However, numerous complications have been reported
with this procedure, including nonunion, malunion, local
recurrence, compartment syndrome, superficial infection,
and deep infection [2, 6, 9, 13, 19]. One of the most serious
Volume 466, Number 5, May 2008
potential complications involves vascular compromise of
the reconstructed limb [11, 17, 20–22, 27, 32, 33].
Currently, two methods of vascular management are
reported. Patients may be treated by primary transection
and reanastomosis, sparing only the sciatic nerve, when a
tumor is suspected to involve the major vessels. This
approach eliminates loops of redundant artery and vein and
allows potentially improved resection margins. One report
suggests resection of the vessels is obligatory when per-
forming rotationplasty for tumors of the distal femur [35].
Alternatively, major vascular structures can be retained as
dictated by tumor involvement, response to chemotherapy,
and the ability to provide an adequate vascular bed distally.
The vascular structures when mobilized are externally
rotated 180° with the lower leg and looped with the sciatic
nerve. Reports of complications in this group include
pseudoaneurysm and prolonged distal ischemia eventually
requiring vascular reconstruction [3]. Reported complica-
tions in the anastomosis group include thrombosis at the
anastomosis, requiring vascular revision [9, 11, 17, 20, 21,
27, 32]. Currently, there is no consensus regarding whether
one of the approaches results in a higher rate of
complications.
We asked whether the ankle brachial index (ABI) or
complication rate between the two groups would differ
with the two methods of vascular management for
rotationplasty.
Materials and Methods
We retrospectively reviewed 15 patients (10 boys, five
girls) ranging in age from 7 to 13 years, with an average
age of 11.5 years, who underwent rotationplasty between
April 1991 and May 1998. We resected the distal superfi-
cial femoral and popliteal vessels with the tumor in seven
patients (Group A). In the remaining eight patients, the
vessels were mobilized for rotation without transection
(Group B). All seven patients in Group A had distal fem-
oral tumors managed with a Type A1 rotationplasty as
described by Winkelmann [35]. Five patients in Group B
had proximal tibial tumors managed with a Type A2
rotationplasty [35]. Three patients in Group B had distal
femoral tumors and were managed with a modified Type
A1 rotationplasty in that the vessels were not transected
[35]. Fourteen patients had osteosarcoma and one patient
had synovial sarcoma. The patient with synovial sarcoma
underwent rotationplasty as a salvage procedure to avoid
above-knee amputation after an invasive Aspergillus
infection developed after her initial tumor resection. All
tumors were Stage IIB [10] at the time of initial evaluation.
At the time of surgery, three patients had evidence of
metastatic disease. Four patients died, all secondary to their
Vascular Management in Rotationplasty 1211
sarcomas, at an average of 46 months (range, 28–
72 months) after their surgery. No living patient was lost to
followup. The 15 patients were seen at a minimum fol-
lowup of 28 months (mean, 95 months; range, 28–
165 months).
Preoperative evaluation consisted of a history and
physical examination, routine hematologic studies, routine
radiographs and MRI of the involved extremity, and
computerized axial tomography of the chest. Vascular
studies included a limited arteriogram and a duplex scan of
the deep and superficial venous systems. These studies
were used to assess the extent of the tumor involvement
and to plan the surgical procedure and vascular recon-
struction. A biopsy-proven diagnosis was established in all
patients preoperatively. We reviewed all radiographic data
with a musculoskeletal radiologist (CWW), a vascular
surgeon (BTB), and an orthopaedic oncologist (JRN).
All patients with osteosarcoma had completed a mini-
mum of three courses of neoadjuvant chemotherapy before
the anticipated surgical procedure. Patients typically com-
pleted chemotherapy 2 to 4 weeks before surgery.
Hematologic parameters were followed closely by the
pediatric oncology team and were normal at the time of
surgery.
Patients with tumors showing maintenance of an other-
wise normal fat plane between the sciatic nerve and the
vessels were judged candidates to maintain continuity of
the vessels. Patients having deviation of the vessels but
maintenance of a normal fat plane also were presumed
candidates to maintain vascular continuity. Patients with
tumors showing the absence of a normal circumferential fat
plane or encasement were considered candidates for vas-
cular resection in continuity with the resection specimen to
provide a minimum of a wide margin [10]. In all cases, the
sciatic or tibial and peroneal nerves had to be uninvolved
for a patient to remain a candidate for rotationplasty. The
surgical approach was based on the approach described by
Winkelmann [35] for Types A1 and A2 rotationplasties.
The initial portion of the operation concerned the devel-
opment of appropriate flaps and planes of dissection to
adequately expose the peroneal nerve and sciatic nerve
(Fig. 1). When appropriate, the vessels were mobilized,
preserving the normal fat plane and keeping them isolated
from the tumor. When vascular involvement was antici-
pated, the plane of dissection remained outside the vessels,
thereby maintaining a wide margin of resection (Fig. 2).
The average time to complete the surgery was 10.0 hours
in Group A and 8.1 hours in Group B. During the case and
postoperatively, a pulse oximeter was placed on the great
toe and monitored closely (Fig. 3). Dextran (0.25 mL/kg/
hour) was started at the end of the procedure and continued
for 2 days to prevent thrombus of the venous anastomosis.
To minimize swelling that can result from venous
123
1212 Mahoney et al.
Fig. 1 The initial portion of the operation concerns dissection and
mobilization of the sciatic nerve. The nerve is localized by isolating
the common peroneal nerve and dissecting proximally.
Fig. 2 Vascular shunts are used to maintain perfusion of the distal
extremity during resection of the mass. After the resection is
completed, the respective artery and vein are repaired by end-to-
end anastomoses.
Fig. 3 A pulse oximeter is used on the distal portion of the ipsilateral
extremity throughout the entire procedure with maintenance of
oxygen saturation greater than 90%.
123
Clinical Orthopaedics and Related Research
hypertension, the extremity was carefully dressed with a
circumferential dressing, keeping the toes exposed for
monitoring, and it was continuously elevated. Aspirin,
70 mg once daily, was initiated after the dextran was dis-
continued and continued until discharge from the hospital
for 6 months.
Chemotherapy typically was resumed 2 to 4 weeks
postoperatively when hematologic parameters had returned
to normal. All patients had been familiarized with the
expected cosmetic and functional outcome either by per-
sonal interaction with other patients or using videotapes of
interviews with other patients.
Patients were seen in followup by the operating surgeon
(JRN) at 2 weeks and then monthly for 3 months. Fol-
lowup was more frequent for patients with wound
complications. An ABI was obtained when the wounds
were healed. Patients typically were seen every 3 months
for the first year, every 4 months for the second and third
years, every 6 months for the fourth and fifth years, and on
a yearly basis thereafter. Routine radiographs of the
involved extremity were obtained at each followup to
evaluate the osteotomy site. Metastatic surveillance inclu-
ded a computed tomographic scan of the chest at each
followup and an annual radionuclide bone scan.
The differences in the ABI between the two groups were
determined using Student’s t test with the level of signifi-
cance set at p \ 0.05. Survival was determined by
calculating the percent of patients living at 5 years
followup.
Results
Four patients died, all secondary to their sarcomas, at an
average of 46 months after their surgery. Three deceased
patients were in Group A and one was in Group B. Two of
these patients were known to have systemic disease before
their operation. Of the remaining patients, 10 have no
evidence of residual neoplastic disease. One patient has
evidence of metastatic disease 39 months after surgery.
There have been no local recurrences. The average 5-year
survival rate for the entire cohort was 81%. The 5-year
survival rate for Group A was 71%, and the 5-year survival
rate for Group B was 89%.
The mean ABIs for both groups for the posterior tibial
and dorsalis pedis arteries were similar preoperatively and
just over 2 months postoperatively (Table 1). The most
recent ABIs for Group A at an average of 23 months
postoperatively and Group B at an average of nearly
8 months also were similar (Table 2). The mean change
from the preoperative to postoperative ABIs for both
arteries in both groups was similar (Table 3), as was mean
change from preoperative to final values (Table 4).
Volume 466, Number 5, May 2008 Vascular Management in Rotationplasty 1213

Table 1. Mean change in ankle brachial index group, required vascular exploration within 24 hours of
Group/Artery Preoperative Postoperative p Value
the operation secondary to venous congestion in the first
and arterial thrombosis in the second. Both of these
Group A patients eventually required split-thickness skin grafting
Posterior tibial 0.96 (0.89–1.00) 0.80 (0.64–1.05)* 0.28 for necrotic areas around their respective wounds. Two of
Dorsalis pedis 0.98 (0.93–1.00) 0.57 (0.00–1.13)* 0.27 the patients in Group A had infections requiring irrigation
Group B and débridement. One patient initially was treated for
Posterior tibial 1.03 (0.93–1.10) 1.00 (0.78–1.31)  0.09 synovial sarcoma with soft tissue resection and a medial
Dorsalis pedis 0.97 (0.91–1.00) 0.97 (0.81–1.22)  0.09 gastrocnemius flap. Postoperatively, the wound became
*  
2.4 months postoperatively; 7.2 months postoperatively. infected with Aspergillus, and despite numerous débrid-
ements, it remained infected. The limb was salvaged with
a rotationplasty. The second patient had a postoperative
Table 2. Mean change in ankle brachial index infection and twice underwent irrigation and débridement
Group/Artery Preoperative Last p Value to aid wound healing. This patient ultimately required
hardware removal to achieve wound healing. The
Group A remaining additional operations were for soft tissue cov-
Posterior tibial 0.96 (0.89–1.00) 0.96 (0.64–1.35)* 0.17 erage secondary to wound-healing complications. These
Dorsalis pedis 0.98 (0.93–1.00) 0.82 (0.00–1.13)* 0.27 were equally divided between the two groups.
Group B
Posterior tibial 1.03 (0.93–1.10) 0.98 (0.93–1.31)  0.17
Dorsalis pedis 0.97 (0.91–1.00) 0.96 (0.81–1.22)  0.16 Discussion
*  
23 months postoperatively; 7.7 months postoperatively.
Two methods are available for management of the vascular
structures during a rotationplasty. These involve transec-
Table 3. Mean change in preoperative to postoperative ankle bra- tion and reanastomosis of the vessels or retaining and
chial index rotating the vessels with the leg. It is not clear if one
Artery Group A Group B p Value
method results in superior outcomes. We asked whether the
two methods of vascular management differ regarding ABI
Posterior tibial 0.11 0.19 0.48 and rate of complications.
Dorsalis pedis 0.07 0.22 0.10 Two limitations of this study include the retrospective
nature and the small sample size. A retrospective study
design cannot be randomized and can be subject to bias.
Table 4. Mean change in preoperative to final ankle brachial index With a small variable cohort, one should exercise caution
when generalizing the results.
Artery Group A Group B p Value
The ABI is a well-described tool to indirectly measure
Posterior tibial 0.26 0.10 0.24 the adequacy of arterial circulation of the lower extremity
Dorsalis pedis 0.43 0.12 0.27 [24]. We used the ABI to evaluate the health of the vas-
cular system with two methods of vascular management.
Although the mean preoperative to postoperative changes
Duplex scanning was performed in two of the patients in ABI were similar in Group A, there was a trend toward a
who had reconstruction during their routine followup. lower ABI in the postoperative period, which partially
Notably, there was no evidence of intimal hyperplasia of recovered with time. This likely was attributable to a
the anastomosis. considerable decline in the ABI of two patients during the
There were 12 complications (five in Group A and early postoperative period. We cannot be certain of the
seven in Group B) requiring an additional 36 operations reasons for these findings, but we presume they were
(18 in both groups) in eight patients. Nineteen of the related to swelling around the anastomosis. Only one
operations were performed on two patients. One patient in patient with a noticeable change in the ABI had a com-
Group B experienced vascular insufficiency after dis- plication. All other complications were in patients with
charge from the hospital. On return for the first normal ABIs. A review of the literature finds other reports
postoperative visit, examination in the clinic revealed a of this method of vascular management. Sidiropoulos et al.
largely necrotic, nonunited limb. Further inspection in the [31] reported finding no vascular complications in a series
operating room confirmed the findings and the decision of 34 patients managed with vascular resection and rean-
was made to amputate. Two patients, one from each astomosis. They recommend this method of vascular

123
1214 Mahoney et al. Clinical Orthopaedics and Related Research

management because of the improved resection margins,
surgical time, and low rate of complications. We found no
difference in complication rate between the two methods
but did find the vascular reconstruction took slightly more
time to perform.
Duplex ultrasound was performed in two patients to
evaluate the anastomosis for intimal hyperplasia after a
substantial interval decrease in the ABI. Surprisingly, we
did not observe a single case of hyperplasia. We postulate
the perioperative chemotherapy may have inhibited the
cellular proliferation required for hyperplasia. This
hypothesis is supported by the experience of using drug-
eluting stents to prevent restenosis after percutaneous
intervention for coronary artery disease [7, 26, 30].
We did experience some wound-healing problems at the
site of the rotation. The rate and type of complications we
found were similar to those described in other reports of
limb-sparing surgery [12, 14]. It could be theorized the
wound complications were secondary to a decrease in the
peripheral blood flow to the skin resulting from our
vascular ligation and reanastomosis; however, patients had
wound-healing problems in the group without shunting
procedures as well. Furthermore, the problems encountered
occurred in earlier cases, indicating a possible learning
curve associated with the procedure, an idea both senior
authors (BTB, JRN) support.
Several authors have reported using rotationplasty as a
successful treatment option for a failed limb salvage pro-
cedure, albeit with higher complication rates than primary
rotationplasty [5, 15]. The only patient in our cohort who
did not have osteosarcoma underwent rotationplasty after
failure of the primary resection for synovial sarcoma. After
the rotationplasty, the patient had no additional
complications.
There are some drawbacks to the rotationplasty. There is
a long and intensive period of physical therapy that should
be completed to ensure the limb is functional. This requires
commitment from not only the patient but also the patient’s
family because nearly all of these patients are juveniles.
Much has been written about the self-image of these

Fig. 4A–C A patient is shown

57 months after rotationplasty (A)
wearing his prosthesis, (B) sitting
and (C) standing without his
prosthesis. He is clinically and
radiographically free of disease.
His wound healed with no skin
loss and his prosthesis is well
fitting and highly functional. He is
extremely active and enjoys inline
skating and bicycle riding (Vid-
eos 1 and 2, Supplemental
Website Materials; supplemental
materials are available with the
online version of CORR).

123
Volume 466, Number 5, May 2008
patients after they have their surgeries [25]. It is no doubt
hard to undergo major surgery for an illness, especially a
serious illness like cancer, only to have the residual be an
extremity that is deformed and initially not functional
(Fig. 4). However, published reports suggest patients are
well adjusted (in terms of SF-36 mental health scores) to
the final outcome [1, 16]. One concern more recently has
been the long postoperative hospital stay compared with
amputation. In the age of managed care, prolonged hospital
stays may become harder to justify; however, it is our
opinion that a patient’s well-being should never be com-
promised secondary to cost. Finally, the surgeon must
consider, if the rotationplasty fails, the patient will be left
with a considerably shorter residual limb than if transfe-
moral amputation had been the primary procedure.
Reconstruction with an expandable endoprosthesis could
be considered an alternative option for many of these
patients. Patients with intraarticular extension of the tumor
and those with failed resection of a synovial sarcoma were
not candidates for an expandable prosthesis. In addition,
numerous orthopaedic oncologists, including the senior
author (JRN), have expressed concern regarding the
reported long-term durability and revision rates of the
expandable prosthetic reconstructions [8, 28, 29].
We evaluated two methods of vascular management
with the Van Nes rotationplasty and were unable to show a
difference in the ABI or rate of complications between the
two methods of vascular management. From this, we
conclude resection of the vessels with reanastomosis is a
surgically sound alternative to coiling of the vessels when
the tumor dictates en bloc vascular resection.
Acknowledgments We thank Dr. Craig Walker for assistance in
interpretation of the MRIs, Kerby Selmer for assistance with the
patient database, and Dr. Kaleb Michaud for assistance with the
statistical analysis.
References
1. Akahane T, Shimizu T, Isobe K, Yoshimura Y, Fujioka F, Kato
H. Evaluation of postoperative general quality of life for patients
with osteosarcoma around the knee joint. J Pediatr Orthop B.
2007;16:269–272.
2. Alman BA, Krajbich JI, Hubbard S. Proximal femoral focal
deficiency: results of rotationplasty and Syme amputation. J Bone
Joint Surg Am. 1995;77:1876–1882.
3. Blewitt N, Pooley J, Berridge D, Chamberlain JC. Pseudoaneu-
rysm of the superficial femoral artery: an unusual complication
following rotationplasty. Clin Orthop Relat Res. 1994;306:171–
174.
4. Borggreve J. Kniegelenkseratz durch das in der Beinlangsachse
um 180 Degree gedrehte Fussgelenk. Arch Orthop Unfall Chir.
1930;28:175.
5. Brigman BE, Kumagai SG, McGuire MH. Rotationplasty after
failed limb-sparing tumor surgery: a report of two cases. Clin
Orthop Relat Res. 2003;415:254–260.
Vascular Management in Rotationplasty 1215
6. Cammisa FP Jr, Glasser DB, Otis JC, Kroll MA, Lane JM, Healey
JH. The Van Nes tibial rotationplasty: a functionally viable
reconstructive procedure in children who have a tumor of the distal
end of the femur. J Bone Joint Surg Am. 1990;72:1541–1547.
7. Degertekin M, Regar E, Tanabe K, Smits PC, van der Giessen
WJ, Carlier SG, de Feyter P, Vos J, Foley DP, Ligthart JM,
Popma JJ, Serruys PW. Sirolimus-eluting stent for treatment of
complex in-stent restenosis: the first clinical experience. J Am
Coll Cardiol. 2003;41:184–189.
8. Eckardt JJ, Safran MR, Eilber FR, Rosen G, Kabo JM. Expand-
able endoprosthetic reconstruction of the skeletally immature
after malignant bone tumor resection. Clin Orthop Relat Res.
1993;297:188–202.
9. Eilber FR, Morton DL, Eckardt J, Grant T, Weisenburger T.
Limb salvage for skeletal soft tissue sarcomas: multidisciplinary
preoperative therapy. Cancer. 1984;53:2579–2584.
10. Enneking WF. A system of staging musculoskeletal neoplasms.
Clin Orthop Relat Res. 1986;204:9–24.
11. Fortner JG, Kim DK, Shiu MH. Limb-preserving vascular sur-
gery for malignant tumors of the lower extremity. Arch Surg.
1977;112:391–394.
12. Futani H, Minamizaki T, Nishimoto Y, Abe S, Yabe H, Ueda T.
Long-term follow-up after limb salvage in skeletally immature
children with a primary malignant tumor of the distal end of the
femur. J Bone Joint Surg Am. 2006;88:595–603.
13. Gottsauner-Wolf F, Kotz R, Knahr K, Kristen H, Ritschl P, Salzer
M. Rotationplasty for limb salvage in the treatment of malignant
tumors at the knee: a follow-up study of seventy patients. J Bone
Joint Surg Am. 1991;73:1365–1375.
14. Hanlon M, Krajbich JI. Rotationplasty in skeletally immature
patients: long-term followup results. Clin Orthop Relat Res.
1999;358:75–82.
15. Hillmann A, Gosheger G, Hoffmann C, Ozaki T, Winkelmann W.
Rotationplasty: surgical treatment modality after failed limb
salvage procedure. Arch Orthop Trauma Surg. 2000;120:555–
558.
16. Hopyan S, Tan JW, Graham HK, Torode IP. Function and upright
time following limb salvage, amputation, and rotationplasty for
pediatric sarcoma of bone. J Pediatr Orthop. 2006;26:405–408.
17. Imparato AM, Roses DF, Frances KC, Lewis MM. Major vas-
cular reconstruction for limb salvage in patients with soft tissue
and skeletal sarcomas of the extremities. Surg Gynecol Obstet.
1978;147:891–896.
18. Jacobs PA. Limb salvage and rotationplasty for osteosarcoma in
children. Clin Orthop Relat Res. 1984;188:217–222.
19. Karakousis CP, Emrich LJ, Rao U, Krishnamsetty RM. Feasi-
bility of limb salvage and survival in soft tissue sarcomas.
Cancer. 1986;57:484–491.
20. Kawai A, Hashizume H, Inoue H, Uchida H, Sano S. Vascular
reconstruction in limb salvage operation for soft tissue tumors of
the extremities. Clin Orthop Relat Res. 1996;332:215–222.
21. Koperna T, Teleky B, Vogl S, Windhager R, Kainberger F,
Schatz KD, Kotz R, Polterauer P. Vascular reconstruction for
limb salvage in sarcoma of the lower extremity. Arch Surg.
1996;131:1103–1107; discussion 1108.
22. Krajbich JI. Modified Van Nes rotationplasty in the treatment of
malignant neoplasms in the lower extremities of children. Clin
Orthop Relat Res. 1991;262:74–77.
23. Krajbich JI, Carrol NC. Van Nes rotationplasty with segmental
limb resection. Clin Orthop Relat Res. 1990;256:7–13.
24. McGee SR, Boyko EJ. Physical examination and chronic lower-
extremity ischemia: a critical review. Arch Intern Med.
1998;158:1357–1364.
25. Merkel KD, Gebhardt M, Springfield DS. Rotationplasty as a
reconstructive operation after tumor resection. Clin Orthop Relat
Res. 1991;270:231–236.
123
1216 Mahoney et al.
26. Morice MC, Serruys PW, Sousa JE, Fajadet J, Ban Hayashi E,
Perin M, Colombo A, Schuler G, Barragan P, Guagliumi G,
Molnar F, Falotico R, RAVEL Study Group. RAndomized study
with the sirolimus-coated Bx VElocity balloon-expandable stent
in the treatment of patients with de novo native coronary artery
lesions. A randomized comparison of a sirolimus-eluting stent
with a standard stent for coronary revascularization. N Engl J
Med. 2002;346:1773–1780.
27. Nambisan RN, Karakousis CP. Vascular reconstruction for limb
salvage in soft tissue sarcomas. Surgery. 1987;101:668–677.
28. Neel MD, Wilkins RM, Rao BN, Kelly CM. Early multicenter
experience with a noninvasive expandable prosthesis. Clin Ort-
hop Relat Res. 2003;415:72–81.
29. Rougraff BT, Simon MA, Kneisl JS, Greenberg DB, Mankin HJ.
Limb salvage compared with amputation for osteosarcoma of the
distal end of the femur: a long-term oncological, functional, and
quality-of-life study. J Bone Joint Surg Am. 1994;76:649–656.
30. Serruys PW, Degertekin M, Tanabe K, Abizaid A, Sousa JE,
Colombo A, Guagliumi G, Wijns W, Lindeboom WK, Ligthart J,
de Feyter PJ, Morice MC, RAVEL Study Group. Intravascular
ultrasound findings in the multicenter, randomized, double-blind
123
Clinical Orthopaedics and Related Research
RAVEL (RAndomized study with the sirolimus-eluting VElocity
balloon-expandable stent in the treatment of patients with de
novo native coronary artery lesions) trial. Circulation.
2002;106:798–803.
31. Sidiropoulos A, Tjan TD, Soeparwata R, Wuismann P,
Winkelmann W, Scheld HH. Reconstructive vascular surgery in
rotationplasty for malignant tumors of the femur. Int Angiol.
1994;13:327–330.
32. Steed DL, Peitzman AB, Webster MW Jr, Ramasastry SS,
Goodman MA. Limb sparing operations for sarcomas of the
extremities involving critical arterial circulation. Surg Gynecol
Obstet. 1987;164:493–498.
33. Taminiau AH, Arndt JW, van Bockel JH, Steenhoff JR, Pauwels
EK. Evaluation of the lymph flow with lymphoscintigraphy after
rotationplasty for the treatment of bone tumors. J Bone Joint Surg
Am. 1992;74:101–105.
34. Torode IP, Gillespie R. Rotationplasty of the lower limb for
congenital defects of the femur. J Bone Joint Surg Br.
1983;65:569–573.
35. Winkelmann WW. Rotationplasty. Orthop Clin North Am.
1996;27:503–523.
Clin Orthop Relat Res (2008) 466:1217–1224
DOI 10.1007/s11999-008-0134-6
ORIGINAL ARTICLE
Late Results of Absorbable Pin Fixation in the Treatment
of Radial Head Fractures
Panagiotis K. Givissis MD, PhD, Panagiotis D. Symeonidis MD,
Konstantinos T. Ditsios MD, PhD, Panagiotis S. Dionellis MD,
Anastasios G. Christodoulou MD, PhD
Received: 27 February 2007 / Accepted: 15 January 2008 / Published online: 4 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract The use of bioabsorbable pins with prolonged
degradation periods for fracture fixation has raised con-
cerns about adverse soft tissue reactions, including
seromas, discharging sinuses, or osteolytic changes. We
asked whether bioabsorbable pins of self-reinforced poly-
lactic acid polymer used in radial head fractures resulted in
such reactions. We retrospectively reviewed 21 patients
followed a minimum of 36 months (mean, 81 months;
range, 36–136 months). There were nine Mason II, 10
Mason III, and two Mason IV fractures, which were eval-
uated clinically and radiographically. All fractures healed
well with no radiographic signs of osteolysis. The mean
Mayo Elbow Performance score was 93.8 (range, 20–100),
which is comparable to the outcome of historical groups
with radial head arthroplasty. The mean range of flexion of
the elbow was 9° to 132°, with 79° pronation and 77°
supination. The grip strength of the operated arm was not
affected in comparison to the contralateral arm (mean
range, 38.6 versus 40.9 kg). No material-related adverse
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution either has waived or
does not require approval for the human protocol for this investigation
and that all investigations were conducted in conformity with ethical
principles of research.
P. K. Givissis, P. D. Symeonidis, K. T. Ditsios,
P. S. Dionellis, A. G. Christodoulou
First Orthopaedic Department, Aristotle University of
Thessaloniki ‘‘G Papanikolaou’’ Hospital, Thessaloniki, Greece
P. K. Givissis (&)
9, Papanikolaou Street, Panorama, Mail Box 215, 55210
Thessaloniki, Greece
e-mail: givissis@otenet.gr; givissis@med.auth.gr
effects were observed during and beyond the degradation
period. Our data suggest concerns about soft tissue or bony
reactions from these materials in radial head fractures are
not justified.
Level of Evidence: Level IV, therapeutic study. See the
Guidelines for Authors for a complete description of levels
of evidence.
Introduction
The best method for treating displaced radial head fractures
in adults remains controversial [6, 7, 12]. The decision
regarding whether the fractured head is salvageable by
means of internal fixation is largely based on subjective
criteria, including the surgeon’s experience and preference
and the equipment available. Whereas radial head
replacement is gaining popularity, there is little doubt a
salvageable and properly fixed radial head provides the
best option for a long-term functional outcome [2, 7, 15].
Various metallic implants are being used for operative
fixation of radial head fractures, including minifragment
screws, Kirschner wires, Herbert screws, and miniplates
[10, 13, 19]. However, during the last two decades, bio-
absorbable pins have been introduced to treat radial head
fractures [9, 21, 26]. Their widespread acceptance has been
hampered partly by reports regarding soft tissue reactions
during the fast degradation process and suboptimal
mechanical properties of these implants [4, 24]. Second-
generation implants composed of self-reinforced (SR)
poly-L-lactic acid (PLLA) with improved strength and a
slower degradation rate were intended to eliminate these
drawbacks and apparently did so [22, 24, 25].
Nevertheless, the prolonged degradation period of the
new implants has raised continuing concerns regarding
123
1218 Givissis et al.
whether the adverse soft tissue reactions of these materials
are in fact not diminished but rather take longer to appear
[3]. The strength of SR-PLLA rods decreases to the level of
cancellous bone in 36 weeks, but total loss of implant mass
takes much longer [11, 30]. Remnants of SR-PLLA plates
have been detected as long as 5 years after fixation of
mandibular osteotomies in sheep [28]. This necessitates
clinical studies with a long followup [4, 23].
Furthermore, such studies need to focus on the clinical
outcome of comminuted radial head fractures in particular.
Recently published series underline the substantial con-
troversy regarding the role of open reduction internal
fixation (ORIF) versus arthroplasty for Mason Type III and
Type IV fractures [15, 27]. Although some authors con-
sider severe comminution an absolute indication for
replacement of the radial head, advocates of the ORIF
claim favorable clinical outcomes regardless of the frac-
ture’s complexity [6, 29].
We asked whether adverse reactions, such as seroma
formation, discharging sinuses, or osteolytic reactions,
would occur during and beyond the material’s degradation
period. We then asked whether patients would achieve
reasonable functional scores and whether function scores
related to fracture type (Mason classification [16]), age, or
length of followup.
Materials and Methods
From a group of 68 patients admitted with closed radial
head fractures, we retrospectively reviewed 25 selected
patients treated with ORIF with absorbable pins from
September 1995 until February 2004. We included only
Grades II to IV fractures according to the McKee and
Jupiter modification [17] of the Mason classification [16]
and surgery in the first 48 hours after the traumatic event.
We excluded 23 patients for whom we judged salvage of
the radial head by internal fixation was impractical.
Another 20 patients were treated by other means of ORIF
(screws, Kirschner wires, miniplates) and also were
excluded from the study. Two of the 25 patients were lost
to followup, leaving 23 for evaluation. There were 12 men
and nine women with an average age at the time of surgery
of 37.8 years (range, 18–78 years; Table 1). All fractures
were secured with the use of SR-PLLA pins (SmartPin1;
formerly Bionx Implants Oy, now ConMed Livantec
Biomaterials Ltd, Tampere, Finland). The clinical outcome
was assessed and compared with two historical groups
using radial head arthroplasty [2, 18]. The minimum
followup was 36 months (median, 97 months; range,
36–136 months).
All operations were performed by the senior author
(PKG). With the patient supine, under axillary block or
123
Clinical Orthopaedics and Related Research
general anesthesia, and the use of a pneumatic tourniquet, a
straight lateral 5- to 6-cm incision was made over the radial
head and through a modified Kocher approach and the
fracture was identified. We reduced the major fragments
with the aid of the convex surface of a McDonald instru-
ment and temporarily secured them with miniclamps.
Smaller fragments were reduced and compressed in posi-
tion manually. We used pronation/supination maneuvers
during the reduction process. The fractures were fixed by
two to five predrilled pins (Fig. 1). When the fracture
included the head only, we inserted the pins parallel to the
joint line. When the fracture included the subcapital region,
additional pins were inserted obliquely to the axis of the
joint line. In one case, we removed the comminuted head
from the surgical site, fixed it ex situ with bioabsorbable
pins, and reinserted and attached it to the radial neck with
long (40-mm length) pins in an oblique direction. Any
protruding part of the pins was cut flat at the bone level
with a Number 15 blade. We fixed the major fragments
with 1.5- to 2.0-mm pins of 20- to 30-mm length. Smaller
fragments were either fixed with 1.1-mm pins or packed
manually. Stability and range of motion were tested in-
traoperatively. When present, concomitant lateral ligament
injuries were repaired primarily with bone anchors. The
wound was closed in layers over a drain and we obtained a
postoperative radiograph.
Postoperatively, an above-elbow posterior splint was
applied for up to 2 weeks followed by early mobilization.
A strengthening program was initiated 4 weeks postoper-
atively, including active-assisted and active free weights
and wall pulley exercises.
We performed complete clinical and radiographic
evaluations in 21 of the 23 patients; two patients were
interviewed over the telephone. Patients were interviewed
and examined by one of two independent observers (PDS,
PSD) who did not participate in any of the operations. All
possible adverse effects and complications were recorded.
We evaluated the functional outcome according to the
Mayo Elbow Performance score [20]. This continuous
score combines four criteria: pain, motion, stability, and
function of the affected elbow. The ranges of motion of
the affected and contralateral elbow also were recorded
separately. Grip strengths of the affected and the contra-
lateral arm were measured with a hydraulic hand
dynamometer (Jamar1; Nex Gen Ergonomics Inc,
Quebec, Canada).
We (PKG, AGC) retrospectively reclassified the inju-
ries; in case of discrepancy, final assessment was based on
concurrence. In four of the 25 cases, some of the radio-
graphs were of poor quality or missing from the patients’
files. The radiographic evaluation included plain radio-
graphs in anteroposterior and lateral views in all patients
(Fig. 2).
Volume 466, Number 5, May 2008 Late Results of Absorbable Pin Fixation 1219

Table 1. Demographic data, radiographic classification, measurements, and recorded complications

Patient number Gender Age Followup Mason Mayo Flexion Pronation/ Grip strength Complications
(years) (months) type score (degrees) supination (affected/normal)
(degrees) (kg)

1 Male 19 70 II 95 5–115 80/50 45/50 Supination restriction

2 Male 47 53 II 100 0–140 85/80 54/53 Superficial infection
3 Male 32 63 III 100 0–145 45/90 35/57 Pronation restriction
4 Female 21 46 III 85 30–125 80/60 22/25
5 Female 65 106 IV 100 10–125 65/70 25/30
6 Female 65 39 III 85 10–135 90/80 20/19
7 Male 23 136 II 100 0–120 80/70 54/57
8 Female 18 79 IV 85 20–130 80/65 50/48
9 Male 40 97 III 100 0–118 90/90 51/55
10 Female 55 50 III 100 10–135 85/80 45/40
11 Male 22 100 III 100 0–140 80/80 45/46
12 Female 19 118 III 100 0–150 90/90 23/20
13 Male 25 66 III 100 0–135 80/80 55/58
14 Male 61 120 III 100 0–140 80/80 36/29
15 Female 40 61 II 100 0–145 80/80 35/40
16 Male 30 119 II 100 0–135 80/80 44/25
17 Female 45 97 II 100 0–130 80/80 24/30
18 Male 38 36 II 100 0–140 80/80 55/55
19 Male 28 36 II 20 95–100 70/65 40/55 Medial heterotopic
ossification, stiffness
20 Male 24 118 II 100 0–142 90/90 50/52
21 Female 78 104 III 100 15–125 70/75 18/20
Mean 37.86 81.62 93.81 9.29–131.90 79.05–76.90 38.57/40.90
Standard deviation 17.99 32.10 17.74 21.29–11.96 10.20–10.43 12.98/14.16
Median 38 97 100 0–125 80–80 44/45
Minimum, maximum 18, 78 36–136 20, 100 0, 95–100, 150 45, 90–50, 90 18, 55/19, 58

Categorical variables were presented with frequencies
and percentages. Continuous variables were summarized
with means and 95% confidence intervals and with median
and interquartile ranges whenever the data were not nor-
mally distributed. The grip strengths of the affected and the
normal elbow were compared with the Wilcoxon signed-
rank test for paired samples. Data of two published series
of patients with radial head fractures treated with arthro-
plasty (titanium prosthesis and metal radial heads) were
extracted and compared with data from our sample of
patients [2, 18].
We evaluated the association of categorical variables
(gender distribution and Mason type between groups) with
the chi square test or Fisher’s exact test when the expected
marginal cell count was less than five. Differences in
clinical scores and in variables on range of motion between
groups (Mason II and Mason III to IV fracture types) were
assessed with either Student’s t test for unpaired samples or
the Mann-Whitney U test whenever the variables deviated
from normality; tests were two-tailed. The age and
followup times of the two groups were similar. Analyses
were conducted using SPSS 14 (SPSS Inc, Chicago, IL).
Results
We observed no early or late occurrences of clinically
evident seroma formation, discharging sinus over the
fracture site, or osteolytic changes of the radial head in any
patient. For 15 of the 21 patients, the followup exceeded
5 years, which is the maximum degradation period of the
specific implant.
The median Mayo Elbow Performance score was 93.8
(range, 20–100; Table 1). Three patients with a score of 85
had mild pain on activity. The mean range of flexion of the
elbow was 9° to 132°, with 79° pronation and 77° supi-
nation. The grip strength of the surgically treated arm was
not affected in comparison to the contralateral arm (mean
range, 38.6 versus 40.9 kg). One patient with a poor result
had a fixed flexion deformity resulting from a complete

123
1220 Givissis et al.
Fig. 1A–B (A) This intraoperative figure shows drilling of the radial
head after reduction of the fracture. (B) The SR-PLLA SmartPin1 is
being inserted in the radial head.
bony bridge of heterotopic ossification at the medial side of
the elbow despite the full pronation/supination range of
motion. He was offered and accepted soft tissue release and
removal of the heterotopic ossification with improvement
of the range of motion from 30° to 120°. No other patient
underwent a second operation. The two patients inter-
viewed over the telephone reported no pain or alteration of
their activities resulting from the fracture.
Fracture severity did not correlate with clinical outcome
(Table 1). The median Mayo score was 100 for Mason II
and Mason III to IV fracture type groups. Range of motion
123
Clinical Orthopaedics and Related Research
was similar for the Mason II and Mason III-IV groups
(mean flexion 130° versus 134°, respectively). Pronation
and supination were also similar (Table 2).
We identified no major complications. One patient had a
superficial postoperative infection resulting from Staphy-
lococcus epidermidis, which was treated successfully with
oral antibiotics. One patient had residual moderate insta-
bility. Two cases of radiographically apparent heterotopic
ossification were seen, one over the annular ligament
without any functional impairment and another over the
medial aspect of the elbow with a fixed flexion contracture
(mentioned above). The latter patient had not adhered to
the routine rehabilitation protocol, having prolonged the
immobilization period arbitrarily followed by aggressive
physiotherapy. On the postoperative films of four patients,
we observed a step-off between the fragments of as much
as 1.5 mm although the patients appeared to have no
clinical sequelae.
Discussion
Qualities of an ideal fixation material include adequate
strength and rigidity, lack of adverse reactions, lack of
interference with bone healing, lack of visibility and pal-
pability, and avoidance of an implant removal operation.
Early reports of adverse effects in fracture fixation from the
use of bioabsorbable materials mainly composed of poly-
glycolide acid (PGA) have led to the development of new
implants with a longer degradation period, such as the SR-
PLLA. However, this has raised additional concerns
regarding the possible late occurrence of such reactions.
We therefore asked whether adverse reactions, such as
seroma formation, discharging sinuses, or osteolytic reac-
tions, would occur during and beyond the material’s
degradation period.
We did not have a control group with which we could
compare functional outcomes of alternative approaches in
the long term. Rather, we compared our results with pub-
lished results of historical groups in which the same type of
fractures was treated and the same score was used for
assessment of elbow function. Another limitation is the
inclusion of Mason Type II fractures in our study group.
According to some studies, these injuries may yield a good
clinical outcome by nonoperative treatment [1, 31] or
delayed head excision in selected patients [5]. However,
we agree with Morrey [19] that because today’s patients
demand a greater degree of function, internal fixation of
radial head fractures is being widely used for Type II
fractures. Moreover, a mismatch between radiographic and
intraoperative evaluations of these fractures, the latter
always being more severe, was noted in some of our cases.
Regardless, we found fracture severity did not correlate
Volume 466, Number 5, May 2008 Late Results of Absorbable Pin Fixation 1221

Fig. 2A–D (A) An anteroposterior radiograph shows a Mason Type reconstruction of the radial head. (D) A postoperative lateral
III fracture of the radial head. (B) The lateral radiograph of the same radiograph shows fracture union and satisfactory alignment of the
patient shows the fracture’s comminution. (C) A postoperative radiohumeral joint.
radiograph obtained at the final followup shows successful

Table 2. Comparison of Mason II and Mason III to IV groups

Parameter Mason II (n = 9) Mason III–IV (n = 12) p Value

Age (years)* 32.67 (24.87–40.47) 41.75 (27.91–55.59) 0.221à

Followup (months) *
80.67 (51.59–109.74) 82.33 (64.02–100.65) 0.910à
Flexion (degrees) *
129.67 (118.08–141.25) 133.58 (127.63–139.53) 0.472à
Loss of extension (degrees)  0 (0–2.5) 5 (0–13.75) 0.255§
 
Pronation (degrees) 80.0 (80.0–82.50) 80.0 (72.50–88.75) 0.972§
Supination (degrees) *
75.00 (65.99–84.01) 78.33 (72.23–84.44) 0.483à
 
Mayo score (points) 100 (97.50–100) 100 (88.75–100) 0.883§
* Values expressed as means, with 95% confidence intervals in parentheses;  values expressed as medians, with interquartile ranges in
parentheses; àStudent’s t test; §Mann-Whitney U test.

123
1222 Givissis et al. Clinical Orthopaedics and Related Research

with outcome in our series and this agrees with the study by
Koslowsky et al. [15].
Previous studies suggest, among the various clinical
applications of bioabsorbable implant for fracture fixation,
their use in radial head fractures seems to have the most
benign clinical course. Böstman and Pihlajamäki [4]
reviewed 2528 patients in whom such implants were used
in more than 20 applications. They report patients with
radial head fractures had the lowest adverse reaction in
their series (0.3%), which is in conformity with our series.
This has been partly attributed to an inverse association
between vascularity of the local tissues and the implant’s
reaction rate. Recently, a prospective, randomized study
was published [8] in which bioabsorbable polylactide
implants were compared with standard metal fixation for
treatment of displaced radial head fractures. The authors
report equally satisfactory results for both methods; how-
ever, their followup is limited to 2 years, whereas material-
related foreign body reactions can emerge as late as 4 or
5 years after the original surgery for fracture fixation [3, 4].
Our data confirm these favorable results until and beyond
the complete absorption of this implant. Another bioab-
sorbable material that has been used is PGA [9, 21]. This
material lacks a methyl group, which makes it hydrophilic
and thus more susceptible to hydrolysis and faster degra-
dation than polylactide, usually within 6 months after
insertion [24]. It has been reported to have a high incidence
of soft tissue reactions varying from 2.0% to 46.7%, which
created skepticism regarding its clinical application [4].
We are unaware of reports of delayed adverse effects
from using SR-PLLA pins. Polylactic acid is a derivative of
cyclic diesters of lactic acid from which it has been pro-
duced by ring-opening polymerization resulting in a poly-
alpha-hydroxy derivative of the original acid. The methyl
group makes polylactic acid hydrophobic and is thus
resistant to hydrolysis [22]. Its mechanical in vitro degra-
dation time is 9 months, and animal studies show complete
resorption of SR-PLLA implants within 5 years [8, 14].

Table 3. Comparison of absorbable pin fixation and titanium prosthesis (historical controls)
Parameter Absorbable pin fixation (n = 10) Titanium prosthesis (n = 16) p Value

Gender 1.000à
Males 5 (50.0%) 8 (50.0%)
Females 5 (50.0%) 8 (50.0%)
Age (years)* 41.80 (26.50–57.10) 44.94 (37.23–52.65) 0.658§
Followup (months)* 80.30 (58.28–102.32) 33.69 (27.42–39.95) 0.001§
*
Flexion (degrees) 134.80 (127.79–141.81) 108.94 (86.42–131.45) 0.031§
 
Mayo score (points) 100 (96.25–100) 87.50 (80.00–98.75) 0.012k
*
Values expressed as means, with 95% confidence intervals in parentheses;  values expressed as medians, with interquartile ranges in
parentheses; àchi square test; §Student’s t test; k Mann-Whitney U test.

Table 4. Comparison of absorbable pin fixation and metal radial head (historical controls)
Parameter Absorbable pin fixation (n = 12) Metal radial heads (n = 9) p Value

Gender 1.000à
Males 7 (58.3%) 5 (55.6%)
Females 5 (41.7%) 4 (44.4%)
Mason type 0.331à
III 10 (83.3%) 5 (55.6%)
IV 2 (16.7%) 4 (44.4%)
Age (years)* 41.75 (27.91–55.59) 56.11 (49.37–62.85) 0.056§
Followup (months)* 82.33 (64.01–100.65) 38.78 (32.36–45.19) \ 0.001§
Flexion (degrees)* 133.58 (127.63–139.53) 140.00 (131.16–148.84) 0.175§
Loss of extension (degrees)  5.00 (0–13.75) 8.00 (2.00–15.50) 0.312k
 
Pronation (degrees) 80.00 (72.50–88.75) 84.00 (73.00–87.50) 0.829k
Supination (degrees)* 78.33 (72.23–84.44) 71.67 (64.23–79.10) 0.135§
Mayo score (points)  100 (88.75–100) 83.00 (64.50–85.00) 0.001k
*
Values expressed as means, with 95% confidence intervals in parentheses;  values expressed as medians, with interquartile ranges in
parentheses; àFisher’s exact test; §Student’s t-test; k Mann-Whitney U test.

123
Volume 466, Number 5, May 2008
The SR technique strengthens the polymer and gives it
metal-like mechanical properties without influencing the
degradation process.
Our data using an absorbable pin fixation group suggest
favorable outcomes compared with use of a titanium
prosthesis in Mason Type III radial head fractures
(Tables 3, 4) [2, 18]. The followup was longer in our
absorbable pin fixation group than in the reported arthro-
plasty groups. Compared with the study by Ashwood et al.
[2], flexion and Mayo score were greater in the absorbable
pin fixation group than in the titanium prosthesis group.
Compared with the study by Moro et al. [18], there were no
differences between the two groups with respect to flexion,
loss of extension, supination, or pronation. However, the
Mayo score was greater in our absorbable pin fixation
group than in the titanium prosthesis group.
We consider the midterm outcome of internal fixation of
displaced radial head and neck fractures with PLLA pins
satisfactory in our series. This was documented by a
combination of subjective and objective clinical criteria,
such as pain, range of motion, alignment, stability, and grip
strength of the affected arm.
Our results were superior to those of published series in
which the comminuted fractures were treated with metallic
replacement [2, 18]. Although a recent comparative study
favors the bipolar radial head prosthesis versus ORIF for
treatment of Mason Type III fractures, the followup is
rather short for definite conclusions [27]. The preferential
use of either ORIF or replacement for comminuted radial
head fractures continues to cause debate and controversy.
Whereas the exact clinical and radiographic criteria that
dictate the decision for a head sacrifice have not yet been
established [29], some authors believe, even in commi-
nuted fractures, an initial attempt for internal fixation
should be made; we agree with this concept [6, 12, 15]. In
any case, the surgeon should be prepared for any operation,
including radial head sacrifice or replacement, and the final
decision should be made intraoperatively.
The use of bioabsorbable pins in the treatment of radial
head fractures has yielded satisfactory midterm clinical and
radiographic outcomes. Concerns about soft tissue reac-
tions from the use of these materials in radial head fractures
are not justified according to our results, having overcome
the maximum degradation period.
Acknowledgments We thank Bettina Haidich for assistance with
statistical analysis of the data.
References
1. Akesson T, Herbertsson P, Josefsson PO, Hasserius R, Besjakov
J, Karlsson MK. Primary nonoperative treatment of moderately
displaced two-part fractures of the radial head. J Bone Joint Surg
Am. 2006;88:1909–1914.
Late Results of Absorbable Pin Fixation 1223
2. Ashwood N, Bain GI, Unni R. Management of Mason type-III
radial head fractures with a titanium prosthesis, ligament repair,
and early mobilization. J Bone Joint Surg Am. 2004;86:274–280.
3. Böstman O, Pihlajamäki H. Clinical biocompatibility of biode-
gradable orthopaedic implants for internal fixation: a review.
Biomaterials. 2000;21:2615–2621.
4. Böstman OM, Pihlajamäki HK. Adverse tissue reactions to bio-
absorbable fixation devices. Clin Orthop Relat Res. 2000;371:
216–227.
5. Broberg MA, Morrey BF. Results of delayed excision of the
radial head after fracture. J Bone Joint Surg Am. 1986;68:669–
674.
6. Esser RD, Davis S, Taavao T. Fractures of the radial head treated
by internal fixation: late results in 26 cases. J Orthop Trauma.
1995;9:318–323.
7. Furry KL, Clinkscales CM. Comminuted fractures of the radial
head: arthroplasty versus internal fixation. Clin Orthop Relat Res.
1998;353:40–52.
8. Helling HJ, Prokop A, Schmid HU, Nagel M, Lilienthal J, Rehm
KE. Biodegradable implants versus standard metal fixation for
displaced radial head fractures: a prospective, randomized, mul-
ticenter study. J Shoulder Elbow Surg. 2006;15:479–485.
9. Hirvensalo E, Böstman O, Rokkanen P. Absorbable polyglyco-
lide pins in fixation of displaced fractures of the radial head. Arch
Orthop Trauma Surg. 1990;109:258–261.
10. Hotchkiss RN. Displaced fractures of the radial head: internal
fixation or excision? J Am Acad Orthop Surg. 1997;5:1–10.
11. Hughes TB. Bioabsorbable implants in the treatment of hand
fractures: an update. Clin Orthop Relat Res. 2006;445:169–174.
12. Ikeda M, Sugiyama K, Kang C, Takagaki T, Oka Y. Comminuted
fractures of the radial head: comparison of resection and internal
fixation. Surgical technique. J Bone Joint Surg Am.
2006;88(suppl 1 pt 1):11–23.
13. Josten C, Korner J, Lill H. Radiusköpfchenfrakturen In: Josten C,
Lill H, eds. Ellenbogen Verletzungen: Biomechanik, Diagnose,
Therapie. Heidelberg, Germany: Steinkopf Verlag; 1997:137–149.
14. Jukkala-Partio K, Pohjonen T, Laitinen O, Partio EK, Vasenius J,
Toivonen T, Kinnunen J, Törmälä P, Rokkanen P. Biodegrada-
tion and strength retention of poly-L-lactide screws in vivo: an
experimental long-term study in sheep. Ann Chir Gynaecol.
2001;90:219–224.
15. Koslowsky TC, Mader K, Gausepohl T, Pennig D. Reconstruc-
tion of Mason type-III and type-IV radial head fractures with a
new fixation device: 23 patients followed 1–4 years. Acta Orthop.
2007;78:151–156.
16. Mason ML. Some observations on fractures of the head of the
radius with a review of one hundred cases. Br J Surg.
1954;42:123–132.
17. McKee MD, Jupiter JB. A contemporary approach to the man-
agement of complex fractures of the distal humerus and their
sequelae. Hand Clin. 1994;10:479–494.
18. Moro JK, Werier J, MacDermid JC, Patterson SD, King GJ.
Arthroplasty with a metal radial head for unreconstructible
fractures of the radial head. J Bone Joint Surg Am. 2001;83:
1201–1211.
19. Morrey BF. Radial head fracture. In: Morrey BF, ed. The Elbow and
its Disorders. 3rd ed. Philadelphia, PA: Saunders; 2000:341–364.
20. Morrey BF, Adams RA. Semiconstrained arthroplasty for the
treatment of rheumatoid arthritis of the elbow. J Bone Joint Surg
Am. 1992;74:479–490.
21. Pelto K, Hirvensalo E, Böstman O, Rokkanen P. Treatment of
radial head fractures with absorbable polyglycolide pins: a study
on the security of the fixation in 38 cases. J Orthop Trauma.
1994;8:94–98.
22. Peltoniemi HH, Hallikainen D, Toivonen T, Helevirta P, Waris T.
SR-PLLA and SR-PGA miniscrews: biodegradation and tissue
123
1224 Givissis et al.
reactions in the calvarium and dura mater. J Craniomaxillofac
Surg. 1999;27:42–50.
23. Prokop A, Jubel A, Hahn U, Dietershagen M, Bleidistel M, Peters
C, Höfl A, Rehm KE. A comparative radiological assessment of
polylactide pins over 3 years in vivo. Biomaterials. 2005;26:
4129–4138.
24. Prokop A, Jubel A, Helling HJ, Eibach T, Peters C, Baldus SE,
Rehm KE. Soft tissue reactions of different biodegradable poly-
lactide implants. Biomaterials. 2004;25:259–267.
25. Prokop A, Jubel A, Helling HJ, Udomkaewkanjana C, Brochha-
gen HG, Rehm KE. [New biodegradable polylactide implants
(Polypin-C) in therapy for radial head fractures] [in German].
Chirurg. 2002;73:997–1004.
26. Rehm KE, Helling HJ, Gatzka C. [New developments in the
application of resorbable implants] [in German]. Orthopade.
1997;26:489–497.
123
Clinical Orthopaedics and Related Research
27. Ruan HJ, Fan CY, Liu JJ, Zeng BF. A comparative study of
internal fixation and prosthesis replacement for radial head
fractures of Mason type III. Int Orthop. 2007 Oct 16 [Epub ahead
of print].
28. Suuronen R, Pohjonen T, Hietanen J, Lindqvist C. A 5-year in
vitro and in vivo study of the biodegradation of polylactide
plates. J Oral Maxillofac Surg. 1998;56:604–614; discussion
614–605.
29. Tejwani NC, Mehta H. Fractures of the radial head and neck:
current concepts in management. J Am Acad Orthop Surg.
2007;15:380–387.
30. Waris E, Ashammakhi N, Kaarela O, Raatikainen T, Vasenius J.
Use of bioabsorbable osteofixation devices in the hand. J Hand
Surg Br. 2004;29:590–598.
31. Weseley MS, Barenfeld PA, Eisenstein AL. Closed treatment of
isolated radial head fractures. J Trauma. 1983;23:36–39.
Clin Orthop Relat Res (2008) 466:1225–1231
DOI 10.1007/s11999-008-0169-8
ORIGINAL ARTICLE
Are Inflatable Nails an Alternative to Interlocked Nails
in Tibial Fractures?
Jesper Blomquist MD, Odd J. Lundberg Ing,
Nils R. Gjerdet DDS, PhD, Anders Mølster MD, PhD
Received: 17 September 2007 / Accepted: 30 January 2008 / Published online: 26 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Recently developed inflatable nails avoid
reaming and interlocking screws in tibial fractures and
reflect a new principle for stabilization of long bone frac-
tures. We asked if the bending stiffness, rotational rigidity,
or play (looseness of rotation) differed between an inflat-
able versus large-diameter reamed interlocked nails, and
whether the maximal torque to failure of the two bone-
implant constructs differed. In a cadaveric model, we
compared the biomechanical properties with those of an
interlocked nail in eight pairs of fractured tibial bones.
Bending stiffness, rotational rigidity, play (looseness in
rotation), and torsional strength within 20° rotation were
investigated using a biaxial servohydraulic testing system.
For all biomechanical variables, we found a large
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved the
human protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
J. Blomquist (&)
Department of Surgery, Haraldsplass Deaconess Hospital,
Ulriksdal 8, 5009 Bergen, Norway
e-mail: Jesper.Blomquist@kir.uib.no
J. Blomquist, A. Mølster
Department of Surgical Sciences, University of Bergen, Bergen,
Norway
O. J. Lundberg, N. R. Gjerdet
Department of Oral Sciences–Dental Biomaterials,
University of Bergen, Bergen, Norway
A. Mølster
Department of Orthopaedics, Haukeland University Hospital,
Bergen, Norway
interindividual variance between the pairs attributable to
bone quality (osteoporosis) for both fixation methods. The
inflatable nail had a higher bending stiffness, with a mean
difference of 58 N/mm, and a lower torsional strength,
with a mean difference of 13.5 Nm, compared with the
locked nail. During torsional testing we noted slippage
between the inflatable nail and bone. We observed no
differences in play or rotational rigidity. Given the lower
torsional strength we recommend caution with weight-
bearing until there are signs of fracture consolidation.
Introduction
Tibial fractures are the most common of the long bone
fractures with an annual incidence of two tibial shaft
fractures per 1000 individuals [1]. The choice of treatment
depends on fracture location, displacement, comminution,
soft tissue condition, and local tradition, but for displaced
shaft fractures in the adult, some studies recommend
reamed intramedullary nailing as the preferred method [1,
12, 15].
The reamed intramedullary nail provides biomechanical
advantages through its central placement, large diameter,
and locking screws that secure rotation and length. A small
incision for nail entry away from the fracture reduces the
risk of infection, and autotransplantation of bone through
reaming promotes bone healing [8]. The procedure has a
union rate greater than 95% for closed fractures in
numerous clinical studies [7]. The technique, however, is
not without disadvantages and complications. At many
centers, the procedure requires 60 to 100 minutes of sur-
gery [10, 20]. Fluoroscopy use during distal locking with a
free-hand technique puts surgeons at risk of exposing their
hands in the radiation field [20]. Complications are
123
1226 Blomquist et al.
common, with complication rates as much as 58% [16] and
reoperation rates as much as 35% [3, 11].
Inflatable nails are a relatively new device for intra-
medullary fixation of long bone fractures without reaming
or locking screws [6]. The implant allows fast, minimal
invasive and stable internal fixation for long bone fractures.
The interface between bone and implant differs radically
from traditional nails and is expected to have a different
effect on fracture stability and biology. Good clinical
results have been reported for this device for humeral and
tibial fractures [3, 6, 9, 13, 25], with only one case report of
a complication directly related to the implant [23]. One
humerus cadaveric study compared an inflatable with a
conventional intramedullary nail and the investigators
observed higher bending stiffness but lower rotational
stiffness related to the funnel-shaped intramedullary canal
of the humerus [5]. For the proximal femur, the inflatable
implant consists of an inflatable nail and an inflatable peg
for the femoral neck. When tested mechanically in a
cadaver model, the implant had similar results for pull-out
and a superior torsional stability, compared with a standard
hip lag screw [24].
We asked whether the bending stiffness, rotational
rigidity, or play (looseness of rotation) differed between
inflatable versus large-diameter reamed interlocked nails,
and whether the maximal torque to failure of the two bone-
implant constructs differed.
Materials and Methods
In a cadaveric fracture model, we tested the bone-implant
construct for the inflatable nail compared with a traditional
interlocked nail. Eight pairs of tibia were used, with one
type of implant placed randomly at each side, and paired
comparisons were made.
The inflatable implant is based on an expandable,
stainless-steel tube with four reinforcement bars, a conical
distal end, and a one-way valve in the proximal part
(Fixion1 IM Nail; Disc-O-Tech Medical Technologies
Ltd, Herzliya, Israel). The nail is introduced in an unex-
panded state into the medullary canal without reaming and
without guide wire. By means of a hydraulic pump with a
screw piston, saline is delivered into the nail to inflate it
until a pressure of 50 to 70 bar is reached (Fig. 1A). This
will increase its diameter until the reinforcement bars
connect with the walls of the medullary cavity and lock to
the bone. The 8.5-mm nail can expand to a maximum of
13.5 mm and the 10-mm nail to 16 mm.
The interlocked tibia nail is made of stainless steel (Grosse
& Kempf1; Stryker Osteosynthesis, Kiel, Germany). It
is an unslotted nail with holes for two proximal and two
distal static interlocking screws of 4.6-mm thread diameter
123
Clinical Orthopaedics and Related Research
Fig. 1A–B (A) The Grosse & Kempf1 interlocked nail system and
(B) the Fixion1 inflatable nail system are shown.
and 3.8-mm core diameter (Fig. 1B). The most commonly
used nail diameter in the tibia is 10 to 11 mm. This design
of reamed interlocked nails is well documented in long-
term clinical [7] and in biomechanical studies [21].
Eight pairs of human tibiae were obtained from fresh
cadavers and frozen immediately after excision. The
donors (five female, three male) ranged in age from 55 to
92 years (mean, 74.8 years). Patients with a history of bone
disease or tibial fractures were excluded. Each pair was
fixed by one inflatable and one interlocked nail, respec-
tively, evenly distributed between the left and right sides.
Before use, they were thawed at room temperature for
approximately 12 hours and kept moist in saline until use.
A 1-cm transverse saw cut from the front was placed at the
anatomic longitudinal midpoint of the bone, and then
bending was applied by manual force in flexion in the
sagittal plane until fracture, causing a short oblique fracture
line in all cases.
The bone was placed in a clamp in reduced position.
When using the locked nails, the bone was opened at the
midpoint between the tibial tuberosity and the tibial pla-
teau, and a guide was introduced. Reaming was performed
up to 11 mm. This provided at least 4 cm of bone contact
on either side of the fracture. A 10-mm locked nail was
used in all preparations. The locking screws (two in each
end) were placed through both bone cortices with a jig. The
inflatable nails were inserted through a similar opening;
8.5-mm nails were used. The unexpanded nail could be
fully inserted without reaming in all instances. The pump
then was connected to the valve, and the pump handle was
rotated until the pressure reached 60 bar. By manual testing,
the osteosynthesis appeared stable, which confirmed pro-
per expansion of the nail. Radiographs were taken to
confirm the nails were properly placed in the medullary
Volume 466, Number 5, May 2008 Inflatable Versus Interlocked Nail in Tibia 1227

Indianapolis, IN) at a speed of 0.2 mm per second until

3 mm of deformation was reached and then returned to
0 mm. The stiffness was calculated from the slope of the
curve. The deformation was small and within the elastic
region, so that we observed no visual structural changes.
Thereafter, both ends of the bones were embedded in
methylmethacrylate cement, but not including the locking
screws. The entry hole in the proximal tibia for the nail was
filled with a soft silicone material before embedding to
prevent direct contact between the implant and the cement.
The embedded ends of the specimens were mounted in the
testing machine. Axial traction of 20 N was applied
(Fig. 4). A moment then was imposed, starting at zero until
reaching 1 Nm. This is less than 10% of the failure load
and represents a load that does not cause gross structural
changes. Then the rotational direction was changed until a
moment of -1 Nm was reached. The rotational speed was
0.1 Nm per second. Moment and rotational angle were
recorded continuously by the software. From the resulting

Fig. 2A–D (A, B) Anteroposterior and (C, D) lateral radiographs

show the prepared specimens with the inserted nails (A, C:
interlocked nail; B, D: inflatable nail) before mechanical testing.

Fig. 3 The setup for four-point testing of bending stiffness is shown.

cavity and the interlocking screws were within the holes of

the nail (Fig. 2).
Bending stiffness was first assessed by mounting the
specimens in a mediolateral position in a four-point testing
rig with 220 mm between the outer rollers and 35 mm
between the inner rollers. The two inner rollers were
connected with an axis centered over the fracture to ensure
equal loading on both sides of the fracture (Fig. 3). An
Fig. 4 A specimen loaded in the servohydraulic biaxial testing
axial compression force was applied by a biaxial servo- machine for measurement of play, torsional rigidity, and maximal
hydraulic testing machine (859 Minibionix1 II; MTS Corp, rotational strength is shown.

123
1228 Blomquist et al.
Fig. 5 A typical curve for calculation of play and rotational rigidity
is shown.
Fig. 6 A typical curve for calculation of maximal rotational strength,
defined as maximal moment before 20° deformation, is shown.
GK = interlocked nail; FX = inflatable nail.
curves, the torsional rigidity was calculated as the slope
between 1 Nm and -1 Nm (Fig. 5). The play (the
unconstrained rotation) was defined as the offset in angle
between the curves obtained in the opposite directions. The
offset angle was calculated from the points corresponding
to 0.25 Nm and -0.25 Nm to avoid uncertainty in the
region of zero moment (Fig. 5). In two specimens, the
measurements of play and initial torsional rigidity were
missed. Thereafter, a rotational force was applied at a
speed of 0.5° per second for 60° or until the construct
failed. The highest moment before reaching 20° rotation
was defined as the overall torsional strength of the con-
struct (Fig. 6). During the latter testing procedure, a
compressive axial load of 150 N was applied, assumed to
simulate the force from muscle tone and soft tissue resis-
tance in a nonweightbearing situation.
Mean differences for each pair were calculated for all
mechanical variables and paired comparisons for the two
123
Clinical Orthopaedics and Related Research
nails were analyzed by the one-sample t test of differences,
using SPSS statistical software (SPSS Inc, Chicago, IL).
We calculated statistical power post hoc for the different
variables.
Results
We observed a large interindividual variation for the
mechanical results. The inflatable nail had a higher
(p = 0.02) bending stiffness than the locked nail, with a
mean of 280.8 N/mm and 222.7 N/mm, respectively. We
found no differences in initial torsional rigidity and play
between the two types of nails. We observed a large scatter
for those two parameters in the group of interlocked nails
(Fig. 7). For torsional rigidity, power analysis revealed an
effect size of 4.51 Nm per degree would be needed to
obtain a difference at 80% power, whereas the observed
difference was only 0.08 Nm per degree. The corre-
sponding values for play were a needed effect size of 2.22°
at 80% power and an observed difference of 0.77
(Table 1).
During testing of torsional strength, all interlocked nail-
bone constructs failed by fracture of the bone through the
distal locking screw holes. The hydraulic nail-bone con-
structs failed initially by slippage between the nail and
bone. At higher deformations, the initial fracture site
developed new fracture lines in most of the specimens
(Fig. 8). The rotational strength for the inflatable nail was
lower (p = 0.05) than that for the interlocked nail, with
12.6 Nm and 26.1 Nm peak moments, respectively, before
reaching 20° rotation.
Discussion
Recently developed inflatable nails avoid reaming and
interlocking screws in tibial fractures and reflect a new
principle for stabilization of long bone fractures. These
nails have been used clinically in the humerus and tibia
with some success [3, 6, 9, 13, 25]. We asked if the
bending stiffness, rotational rigidity, or play (looseness of
rotation) differed between an inflatable versus large-
diameter reamed interlocked nails, and whether the maxi-
mal torque to failure of the two bone-implant constructs
differed.
It is difficult to fully mimic the mechanical forces in the
clinical situation with the combined effects of canal pres-
sure from the osteosynthesis, muscular axial compression
forces, friction in the fracture, and soft tissue resistance.
We opposed the tendency of elongation/separation of
fragments by the obliquity of the fracture during rotation in
the unlocked hydraulic nail construct by application of
Volume 466, Number 5, May 2008 Inflatable Versus Interlocked Nail in Tibia 1229

Fig. 7A–D Results of biome-

chanical testing show interindivi-
dual variance in terms of (A)
bending stiffness, (B) rotational
rigidity, (C) play, and (D) maxi-
mal rotational strength. GK =
interlocked nail; FX = inflatable
nail. Arrows indicate mean in all
figures.

Table 1. Biomechanical testing results of an interlocking intramedullary nail and an inflatable nail
Biomechanical test Interlocking Expandable Paired difference Median paired
nail (GK) nail (FX) (GK - FX) (p value) ratio (GK/FX)

Bending stiffness (N/mm)(n = 8) 222.7 (169.5–275.9) 280.8 (224.7–336.8) -58.1 (-103.5–12.6) (p = 0.02) 1.2
Rotational rigidity (Nm/degree)(n = 6) 2.72 (0.07–5.37) 2.80 (1.62–3.97) -0.08 (-3.22–3.07) (p = 0.95) 1.7
Play (degrees)(n = 6) 1.15 (-0.24–2.54) 0.38 (0.12–0.64) 0.77 (-0.77–2.32) (p = 0.26) 1.8
Maximum rotational strength (Nm)(n = 8) 26.1 (14.3–37.9) 12.6 (8.1–17.1) 13.49 (-0.08–27.05) (p = 0.05) 0.4
Values are expressed as means, with confidence intervals in parentheses.

axial load. The amount of resistance in vivo is not known,
but we have estimated the value in a nonweightbearing
situation to 150 N. In vivo, the axial forces will increase
and tend to separate fragments axially, with increased
rotational force. We used an oblique fracture, which gives
increased resistance to rotation by the obliqueness of the
fracture line, and the more subtle irregularities of a frac-
ture, which results in a great increase of friction compared
with osteotomies [22]. This model is considered more
similar to the most frequent types of low-energy clinical
tibial shaft fractures. A more comminuted fracture with low
intrinsic axial and rotational stability would be, in our
opinion, less suitable for stabilization by a nail without
interlocking screws. The data may also be influenced by
bone quality, which we did not determine. However, these
influences would be limited by the paired study design.
Bending stiffness, rotational rigidity, and play influence
the healing environment for a diaphyseal fracture. We
found a 25% higher mean bending stiffness in the
inflatable nail. We presume this relates to better contact
between the implant and the walls of the medullary canal
and a stiffer implant. We have not performed tests of the
implant alone and therefore cannot differentiate between
these factors.
We assume the rotational play indicates an instability
that can occur during normal clinical loading, even without
weightbearing. Play will induce shearing movements in the
initial healing phase that can prevent or delay fracture

123
1230 Blomquist et al.
Fig. 8A–B Specimens after rotational testing to failure for (A) the
interlocked nail and (B) the inflatable nail are shown.
consolidation as reported in animal models [2, 18]. We
found no differences in play, but the scatter of the results
was higher with the interlocked nail (Fig. 6). We observed
the same pattern for rotational rigidity and maximal rota-
tional strength. Owing to what we considered a high effect
size, we found a difference between the implants for
maximal rotational strength. The scatter suggests the
interlocked nail system may be more technique and oper-
ator sensitive, and more influenced by bone quality. The
high scatter is to some degree compensated for by the use
of paired comparison. However, a higher number or a more
homogenous selection of specimens could have revealed
differences not detected in this study.
The maximum torque before fracture for intact tibia
cadaver bones is in the order of 100 Nm [17]. Our findings
of torsional strength of 13 Nm for the inflatable nail and
26 Nm for the interlocked nail are, in relation to intact
tibiae, quite low. In this study, we used the 8.5-mm
inflatable nail with expanding capacity up to 13.5 mm. The
larger-diameter nail (10 mm with expansion to 16 mm)
might have resulted in improved mechanical properties.
However, the manufacturer recommends using the large
diameter only for an isthmus diameter greater than 12 mm,
and reaming of the opposite side resulted in good cortical
contact at lower diameters in all specimens.
123
Clinical Orthopaedics and Related Research
The theoretical basis for the inflatable implant is that
expansion of the nail results in high pressure against the
inner wall of the medullary cavity and an extended area of
contact, both contributing to high frictional forces between
bone and nail. At the same time, the larger diameter of the
nail increases the area and polar moment of inertia,
increasing the torsional and bending resistance of the nail by
a factor of 4. The absence of locking screws may imply
immediate full dynamization of the fracture postoperatively.
Mechanical testing by Blum et al. [5] of the inflatable
nail in the humerus revealed a lower torsional stiffness for
humeral bones with a funnel-shaped medullary cavity than
for bones with a well-defined isthmus. A transverse oste-
otomy fracture model was used, with a gap of 3 mm. Thus,
there is no friction at the fracture site, and with a funnel-
shaped medullary cavity and without soft tissue, there is
virtually no resistance to rotation with the inflatable nail.
We found the same relatively low rotational strength for
the inflatable nail in our tibial model, with a well defined
isthmus and a fracture type with a high inherent rotational
stability and good contact between the fragments.
This new concept of nail construction using a hydraulic
inflatable nail, inserted in an unexpanded condition, has
provided good results in certain diaphyseal fractures of
long bones [3, 9, 13, 25]. The clinical data, however, are
somewhat limited, with only one controlled study pub-
lished [3]. The inflatable nail had a shorter duration of
surgery and fewer reoperations and rehospitalizations
compared with an interlocking intramedullary nail
(Mathys, Bettlach, Switzerland). The control nail in that
study was an unreamed nail, which would affect the out-
come negatively compared with a large-diameter reamed
nail, according to several studies [4, 7, 8]. Intuitively, and
as suggested by our data and that of Blum et al. [5], the
weak link of an inflatable nail without locking screws
would be control of rotation, whereas the ability to fill out
the intramedullary canal diameter would give good control
of shear and bending forces. None of the clinical studies
report malrotation with inflatable nails used in the tibia
[3, 9, 13, 25]. This is in accordance with most studies of tibia
fractures treated with conventional intramedullary nailing.
However, in two studies where malrotation was specifically
examined, malrotation greater than 10° was reported in more
than 20% of the patients treated with a conventional intra-
medullary nail, which suggests malrotation often is
underreported and difficult to assess clinically [14, 19].
The values for torsional strength of approximately
15 Nm for the hydraulic nailing still means withstanding
up to 60 N at the forefoot, which we consider sufficient for
controlled weightbearing. The tested nail-bone consisted of
isolated tibiae, whereas in clinical situations soft tissue and
the presence of the fibula will influence, and probably
increase, the stability, especially in rotation. Our
Volume 466, Number 5, May 2008
experimental design using axial compression force will add
stability but we suspect does not fully compensate for the
lack of the surrounding tissue.
The hydraulic nail, in contrast to the locked nail,
depends on intrinsic stability in the fracture. Based on the
current results, we recommend restricting its use to
diaphyseal fractures with greater than 50% contact between
the main fragments, and weightbearing should be recom-
mended with caution until there are signs of fracture
consolidation. However, as the procedure is fast and pro-
vides good results in clinical studies, clinical use in
selected cases may be warranted.
Acknowledgments We thank Stein Atle Lie, PhD, Department of
Health, University Research Bergen, for assistance in the statistical
analysis.
References
1. Alho A, Benterud JG, Høgevold HE, Ekeland A, Stromsøe K.
Comparison of functional bracing and locked intramedullary
nailing in the treatment of displaced tibial shaft fractures. Clin
Orthop Relat Res. 1992;277:243–250.
2. Augat P, Burger J, Schorlemmer S, Henke T, Peraus M, Claes L.
Shear movement at the fracture site delays healing in a diaphyseal
fracture model. J Orthop Res. 2003;21:1011–1017.
3. Ben-Galim P, Rosenblatt Y, Parnes N, Dekel S, Steinberg EL.
Intramedullary fixation of tibial shaft fractures using an
expandable nail. Clin Orthop Relat Res. 2007;455:234–240.
4. Blachut PA, O’Brien PJ, Meek RN, Broekhuyse HM. Interlocking
intramedullary nailing with and without reaming for the treatment
of closed fractures of the tibial shaft: a prospective, randomized
study. J Bone Joint Surg Am. 1997;79:640–646.
5. Blum J, Karagul G, Sternstein W, Rommens PM. Bending and
torsional stiffness in cadaver humeri fixed with a self-locking
expandable or interlocking nail system: a mechanical study.
J Orthop Trauma. 2005;19:535–542.
6. Capelli RM, Galmarini V, Molinari GP, De Amicis A. The Fixion
expansion nail in the surgical treatment of diaphyseal fractures of
the humerus and tibia: our experience. Chir Organi Mov.
2003;88:57–64.
7. Court-Brown CM. Reamed intramedullary tibial nailing: an over-
view and analysis of 1106 cases. J Orthop Trauma. 2004;18:96–101.
8. Finkemeier CG, Schmidt AH, Kyle RF, Templeman DC, Varecka
TF. A prospective, randomized study of intramedullary nails
inserted with and without reaming for the treatment of open and
closed fractures of the tibial shaft. J Orthop Trauma.
2000;14:187–193.
Inflatable Versus Interlocked Nail in Tibia 1231
9. Franck WM, Olivieri M, Jannasch O, Hennig FF. An expandable
nailing system for the management of pathological humerus
fractures. Arch Orthop Trauma Surg. 2002;122:400–405.
10. Gugala Z, Nana A, Lindsey RW. Tibial intramedullary nail distal
interlocking screw placement: comparison of the free-hand versus
distally-based targeting device techniques. Injury. 2001;32(suppl
4):SD21–SD25.
11. Harris I, Lyons M. Reoperation rate in diaphyseal tibia fractures.
ANZ J Surg. 2005;75:1041–1044.
12. Hooper GJ, Keddell RG, Penny ID. Conservative management or
closed nailing for tibial shaft fractures: a randomised prospective
trial. J Bone Joint Surg Br. 1991;73:83–85.
13. Jovanovic A, Pirpiris M, Semirli H, Doig SG. Fixion nails for
humeral fractures. Injury. 2004;35:1140–1142.
14. Kahn KM, Beals RK. Malrotation after locked intramedullary
tibial nailing: three case reports and review of the literature.
J Trauma. 2002;53:549–552.
15. Karladani AH, Granhed H, Edshage B, Jerre R, Styf J. Displaced
tibial shaft fractures: a prospective randomized study of closed
intramedullary nailing versus cast treatment in 53 patients. Acta
Orthop Scand. 2000;71:160–167.
16. Koval KJ, Clapper MF, Brumback RJ, Ellison PS Jr, Poka A,
Bathon GH, Burgess AR. Complications of reamed intramedul-
lary nailing of the tibia. J Orthop Trauma. 1991;5:184–189.
17. Martens M, van Audekercke R, de Meester P, Mulier JC. The
mechanical characteristics of the long bones of the lower
extremity in torsional loading. J Biomech. 1980;13:667–676.
18. Mølster AO. Effects of rotational instability on healing of
femoral osteotomies in the rat. Acta Orthop Scand. 1984;55:
632–636.
19. Puloski S, Romano C, Buckley R, Powell J. Rotational mala-
lignment of the tibia following reamed intramedullary nail
fixation. J Orthop Trauma. 2004;18:397–402.
20. Sanders R, Koval KJ, DiPasquale T, Schmelling G, Stenzler S,
Ross E. Exposure of the orthopaedic surgeon to radiation. J Bone
Joint Surg Am. 1993;75:326–330.
21. Schandelmaier P, Krettek C, Tscherne H. Biomechanical study of
nine different tibia locking nails. J Orthop Trauma. 1996;10:37–
44.
22. Seligson D, Byrt W, Hogan M, Pope M. The mechanical basis for
the combined use of internal and external fixation by direct
measurement of fracture gap motion (abstract). 2nd International
Symposium on Internal Fixation of Fractures. Lyon, France:
Faculte Medecine Alexis Carrel; 1982:166–167.
23. Smith MG, Canty SJ, Khan SA. Fixion: an inflatable or deflatable
nail? Injury. 2004;35:329–331.
24. Steinberg EL, Blumberg N, Dekel S. The Fixion proximal femur
nailing system: biomechanical properties of the nail and a
cadaveric study. J Biomech. 2005;38:63–68.
25. Steinberg EL, Geller DS, Yacoubian SV, Shasha N, Dekel S,
Lorich DG. Intramedullary fixation of tibial shaft fractures using
an expandable nail: early results of 54 acute tibial shaft fractures.
J Orthop Trauma. 2006;20:303–309.
123
Clin Orthop Relat Res (2008) 466:1232–1238
DOI 10.1007/s11999-008-0155-1
ORIGINAL ARTICLE
Pause Insertions During Cyclic In Vivo Loading Affect Bone
Healing
Michael J. Gardner MD, Benjamin F. Ricciardi BS,
Timothy M. Wright PhD, Mathias P. Bostrom MD,
Marjolein C. H. van der Meulen PhD
Received: 2 June 2007 / Accepted: 23 January 2008 / Published online: 14 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Fracture repair is influenced by the mechanical
environment, particularly when cyclic loads are applied
across the fracture site. However, the specific mechanical
loading parameters that accelerate fracture healing are
unknown. Intact bone adaptation studies show enhanced
bone formation with pauses inserted between loading
cycles. We hypothesized pause-inserted noninvasive
external loading to mouse tibial fractures would lead to
accelerated healing. Eighty mice underwent tibial osteoto-
mies with intramedullary stabilization and were divided
into four loading protocol groups: (1) repetitive loading
(100 cycles, 1 Hz); (2) pause/time-equivalent (10 cycles,
0.1 Hz); (3) pause/cycle-equivalent (100 cycles, 0.1 Hz);
and (4) no load control. Loading was applied daily for
One or more of the authors have received funding from the
Orthopaedic Trauma Association (MJG), the Orthopaedic Research
and Education Foundation (MJG), and NIH Musculoskeletal Core
Center P30AR046121 (TMW, MCHvdM).
Each author certifies that his or her institution has approved the
animal protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
M. J. Gardner (&)
Department of Orthopaedic Surgery, Harborview Medical
Center, 325 9th Ave, Box 359798, Seattle, WA 98104, USA
e-mail: michaelgardnermd@gmail.com
B. F. Ricciardi, T. M. Wright
Laboratory for Biomedical Mechanics and Materials, Hospital
for Special Surgery, New York, NY, USA
M. P. Bostrom
Department of Orthopaedic Surgery, Hospital for Special
Surgery, New York, NY, USA
M. C. H. van der Meulen
Department of Mechanical and Aerospace Engineering, Cornell
University, Ithaca, NY, USA
123
2 weeks. Healing was assessed using histology, biome-
chanical bending tests, and microcomputed tomography.
The pause-inserted, cycle-equivalent group had a greater
percentage of osteoid present in the callus cross-sectional
area compared with no-load controls, indicating more
advanced early healing. The pause-inserted, cycle-equiva-
lent group had a failure moment and stiffness that were 37%
and 31% higher than the controls, respectively. All three
loaded groups had smaller overall mineralized callus vol-
umes than the control group, also indicating more advanced
healing. At an early stage of fracture healing, pause-inserted
loading led to more histologically advanced healing.
Introduction
Numerous characteristics of the mechanical environment
directly affect the subsequent course of fracture healing.
The magnitude of the applied load across the fracture site,
the size of the fracture gap, the relative fracture motion and
resulting strain, and the timing of mechanical intervention
all impact healing [1, 7, 11, 12, 18]. Cyclic compression
applied at an appropriate time, rate, and magnitude across
the fracture site potentiates healing, leading to earlier and
more advanced callus formation [6, 7, 11, 12, 18, 20, 22,
30]. Early dynamization of external fixation also leads to
accelerated healing compared with rigid fixation [6–8, 20].
Conversely, the beneficial healing effects of intermittent
loading across a fracture are negated when superimposed
on static compression [2, 3, 13, 24, 35].
Innovative loading parameter variations cause dramatic
increases in bone formation in intact bone [26, 28, 31].
Cyclic compressive loads applied with a pause between
each cycle are more osteogenic than repetitive, continuous
cyclic loads [23, 26, 31]. An increased duration of
Volume 466, Number 5, May 2008
repetitive cyclic loading does not lead to proportional
increases in bone mass, presumably because cellular sen-
sitivity to the mechanical stimulus diminishes quickly once
an initial threshold for a response has been exceeded [10,
26, 29, 33]. For example, in a rat jumping model, the
femoral compressive force in rats that jumped five times a
day led to considerable bone hypertrophy compared with
controls, but when jumps were increased to 100 per day, no
additional increase occurred [33]. The cellular sensitivity
to load can be recovered, however, as seen in studies in
which loading sessions were separated into discrete tem-
poral bouts [5, 16].
Accelerating fracture repair through mechanical inter-
vention is a potentially noninvasive treatment supplement.
Although used clinically in one trial [22], this approach has
not become a standard treatment method, partly because
the ideal loading parameters have not been fully defined.
Previously, we showed that repetitive cyclic noninvasive
loading without pause insertions in the loading regimens
led to improved fracture healing after 2 weeks in a mouse
tibial fracture model [18].
We hypothesized insertion of pauses in the loading
waveform would lead to improvements in fracture healing
compared with healing of nonloaded and repetitively loa-
ded fractures, specifically in the early phases of healing.
Materials and Methods
To address our hypothesis, we divided 80 C57Bl/J6 12-
week-old male mice (Jackson Laboratories, Bar Harbor,
ME; body weight 29 ± 2 g) into four groups (n = 20 per
group). After tibial osteotomy creation, stabilization, and
loading, the animals were euthanized by carbon dioxide
inhalation. The experimental hind limbs were disarticulated
at the hip, the tibiae were dissected free of soft tissue, and
the intramedullary pin was removed. All specimens except
those designated for histologic sectioning were frozen in
moist gauze at -20°C in preparation for mechanical anal-
ysis and microcomputed tomographic imaging.
All mice underwent surgical osteotomy and intramed-
ullary nailing of the left tibia as described previously [18].
In three of the groups, cyclic compression was applied to
the ends of the tibia with a noninvasive external loading
device (Fig. 1) [17]. Daily loading was begun after a 4-day
fracture consolidation period to allow for stabilization of
the hematoma and early angiogenesis [19, 34] and con-
tinued for 5 days per week for 2 weeks. One group had 100
load cycles applied daily at a rate of 1 Hz. A second group
had a 9-second pause inserted between each load such that
each load-pause cycle was 10 seconds. This loading con-
tinued for 10 cycles and was the time-equivalent
(100 seconds) of the first group but with one-tenth the
Pause-inserted Fracture Loading 1233
Fig. 1 The external loading device consists of titanium plates that
cup the calcaneus and femoral condyles. The actuator is driven by a
function generator, an amplifier, and an electromagnetic coil to apply
loads across the stabilized fracture.
number of cycles. A third group also had a 9-second pause
inserted between loads, but continued for 100 cycles at
0.1 Hz (1000 seconds), which was the cycle equivalent of
the repetitive loading group (Fig. 2). All loading was
applied with the mice under general anesthesia with a tri-
angle waveform with a peak load amplitude of 1 N and a
preload of 0.5 N. In a fourth (control) group, mice under-
went surgical osteotomy and intramedullary stabilization
and were placed in the loading device daily but without any
load application. A pre-hoc power analysis determined that
10 mice per group, plus an additional five specimens for
histologic analysis, would provide greater than 80% power
to detect a difference in biomechanical and microcomputed
tomography parameters of 1.2 times the pooled standard
deviation using a one-factor analysis of variance design
(alpha, 0.05). Our protocol had prior approval of our
Institutional Animal Care and Use Committee.
With the animal under general anesthesia, an incision
was made over the dorsal aspect of the knee, and a 25-
gauge needle was used to bore a hole in the proximal tibia.
An oscillating saw was used to create a transverse osteot-
omy 5 mm distal to the patellar tendon insertion. The
fragments were held in alignment, and a 27-gauge needle
was inserted across the osteotomy into the distal fragment
and cut to the appropriate length. All fibulas were verified
as intact postoperatively using radiography. Buprenorphine
was administered ad libitum for pain relief; no nonsteroidal
antiinflammatory medications were used to avoid phar-
macologic interference with fracture healing. All mice
were allowed full, unrestricted cage activity after surgery.
The animals did not favor the nonoperated limb starting on
the first postoperative day.
The formation of mineralized tissue at the fracture
site was assessed by quantitative microcomputed tomog-
raphy (micro-CT) (MS-8 Small Specimen Scanner; GE
123
1234 Gardner et al.
Fig. 2 A plot of the waveforms
for the loaded groups is shown.
The two pause-inserted loading
groups had a 9-second pause
between each load application
with cyclic loading continuing
for either 100 or 1000 seconds.
Healthcare, formerly EVS Corp, Ontario, Canada). The
tibiae were placed in a saline-filled scan tube and scanned
at 80 V and 80 lA. Each scan included a phantom con-
taining air, saline, and an SB-3 bone analog (1.18 g/cc) for
calibration of image Hounsfield units to tissue mineral
density in milligrams per centimeter. Reconstruction of the
individual projections to CT volume data was accom-
plished with a modified Parker algorithm [15, 25] with an
isotropic voxel resolution of 24 lm.
To distinguish mineralized tissue from marrow, water,
and unmineralized callus, microCT images were threshol-
ded using 25% of the mineral attenuation value of the
cortical bone determined for each specimen. Regional
analyses of the thresholded scans were performed using the
system software (MicroView; GE Healthcare Technologies,
Waukesha, WI). To characterize calluses globally and
locally, two volumes of interest were selected. First, a best-
fit volume around the entire callus was used to determine
the total mineralized callus volume, a slightly different
method than has been used previously [18]. An elliptic
cylinder, 1 mm in height, was centered at the osteotomy site
and used to determine callus mineral density and content.
Care was taken to ensure this section did not contain a
cortical overlap that would affect density measurements.
Bone tissue from five specimens from each group
underwent histologic analysis. Immediately after harvesting
and dissection, the bones were preserved in 10% phosphate-
buffered formalin for 2 days at 4° C. After formalin fixa-
tion, samples were washed overnight in running water,
dehydrated through alcohol at 70%, cleared in xylene, and
embedded in methyl methacrylate [14]. Calcified tissues
were then sectioned longitudinally at 5- to 7-lm thickness
along the extramedullary portion of the callus. Sectioning
was performed with a tungsten carbide blade on a micro-
tome (Reichert-Jung Ultracut E; Leica Microsystems,
Wetzlar, Germany). Nine sections were subsequently
stained with von Kossa to evaluate the mineralized tissue
and with Goldner’s trichrome to evaluate the osteoid
123
Clinical Orthopaedics and Related Research
deposition. Measurements were taken at x40 magnification.
The percentage of mineralized tissue and osteoid at the
osteotomy site was quantified by a blinded investigator
(BFR) using a BioQuant morphometric analysis system
(Bioquant, Nashville, TN). Because of the early time
studied, no attempt was made to quantify trabecular mor-
phology or other indicators of advanced bone formation.
Rather, we considered advanced histologic healing to be the
presence of more osteoid and total mineralized tissue.
All remaining specimens from each group (n = 7–9)
were tested to failure using four-point bending on an
electromagnetic-based load frame (EnduraTEC ELF 3200;
Bose Corp, Minnetonka, MN). Tibiae were placed with the
flat anteromedial surface on the lower supports, which were
set 8.4 mm apart; the upper load points were set 3.5 mm
apart. Displacement was applied at 0.033 mm per second
until failure occurred. The bending moment to failure
(Nmm) and stiffness (Nmm2) were calculated from the
load-deflection curves and the dimensions between the
loading points. Contralateral intact bones were not tested.
We calculated the mean and standard deviation of the
percent osteoid, percent mineralized tissue, bone mineral
density, bone mineral content, bone volume, callus volume,
failure moment, and stiffness for each group. One-way
analysis of variance tests, followed by Bonferroni post hoc
tests, were used to compare differences among the means
of each variable tested for each group using a p value \
0.05 (PASS 6.0; NCSS, Kaysville, UT). The number of
mice per group was based on a power analysis performed
before the experiment.
Results
Histologically, the cycle-equivalent pause-inserted group
(Fig. 3A) had a greater (p = 0.015) percentage of osteoid
present in the callus compared with controls (Fig. 3B),
indicating a more advanced early healing process. The
Volume 466, Number 5, May 2008
Fig. 3A–B (A) Examples of a histologic specimen after trichrome
staining and microcomputed tomography scans from the cycle-
equivalent pause-inserted group are shown. The area staining positive
for osteoid in the lower portion of the callus is highlighted in red
(Magnification, 912.5). (B) Examples of a histologic specimen after
percentage of mineralized callus tissue was similar
between the groups (Fig. 4).
All three loaded groups had smaller mineralized callus
volumes than the control group (Fig. 5). Additional mea-
surements of callus characteristics such as bone mineral
density, bone mineral content, and total bone volume were
similar between the groups.
Bending strength and stiffness were similar across the
treatment groups. (Fig. 6; Table 1).
Fig. 4 The percentage of osteoid relative to total callus tissue area in
the pause-inserted cycle-equivalent group was greater (p = 0.015)
than the control group. No considerable differences were found in
callus mineralization among groups. Means and standard deviations
are shown.
Pause-inserted Fracture Loading 1235
trichrome staining and microcomputed tomography scans from the
control group are shown. The area staining positive for osteoid in the
lower portion of the callus is highlighted in red (Magnification,
912.5). The histology showed more osteoid in the pause-inserted
group, and calluses appeared radiographically more advanced.
Discussion
Directed mechanical stimulation of fracture healing may be
a potent treatment supplement in difficult fractures. To
become potentially clinically useful, however, the specific
loading parameters which optimize osteogenesis must be
further defined. The insertion of pauses into mechanical
loading regimens is particularly osteogenic in intact bone,
and the purpose of this study was to determine the effect of
pause-inserted loading on fracture healing. After 2 weeks
of healing, we found histologic evidence that pause-
Fig. 5 All three loaded groups had smaller callus volumes than the
control group as measured by microcomputed tomography. Means
and standard deviations are shown.
123
1236 Gardner et al. Clinical Orthopaedics and Related Research

chosen for the specific purpose of analyzing early healing

Fig. 6 The biomechanical parameters tested, failure bending moment
and stiffness, were similar between the groups. Means and standard
deviations are shown.
inserted loading accelerates the fracture healing process
compared with controls.
This study had limitations. Most notably, we studied
only one healing time. Although this time period was
differences resulting from loading, this choice was none-
theless somewhat arbitrary, and different results may occur
at later stages of healing. Our microCT and mechanical
testing data showed larger variability than observed pre-
viously [18], contributing to the lack of considerable
effects despite an appropriate power calculation as part of
the planning for this experiment. Reasons for this are
unclear, but may have been partly attributable to technical
error with testing, and the biomechanical results should be
interpreted with caution. Another limitation was the exact
strains at the osteotomy site were unknown. The curvature
of the mouse tibia induces bending moments that produce
compressive and tensile strains at the fracture site. Strains
induced at the middiaphysis with loading have been char-
acterized for intact tibia of mice using this loading device
[9, 17], but the strains that occur with loading across a
stabilized osteotomy in this model are unknown.
The specimens with 100 cycles of daily pause-inserted
loading had histologic evidence of accelerated fracture
healing compared with controls. This group, together with

Table 1. Groupwise comparisons of bone healing

Variable Group
Control Repetitive Pause-inserted, Pause-inserted,
loading time-equivalent cycle-equivalent

Percent osteoid 13.0 19.4 17.4 24.8*

SD 5.0 5.8 5.3 1.9
Number 5 5 5 5
Percent mineralized tissue 25.0 26.8 32.5 31.8
SD 1.7 7.0 9.1 7.8
Number 5 5 5 5
BMD (mg/cm3) 473.80 491.84 488.26 519.24
SD 41.87 64.59 55.96 75.74
Number 16 14 15 16
BMC (mg) 2.38 2.07 2.29 2.23
SD 0.37 0.51 0.73 0.53
Number 16 14 15 16
Bone volume (cm3) 4.40 3.72 4.16 3.87
SD 0.81 0.93 1.42 0.86
Number 16 14 15 16
Callus volume (cm3) 0.035 0.024* 0.026* 0.026*
SD 0.010 0.007 0.010 0.008
Number 16 14 15 16
Failure moment (N-mm) 16.7 18.4 20.4 22.8
SD 6.9 5.2 7.4 9.5
Number 8 7 9 9
Stiffness (N-mm2) 741.4 719.4 776.1 974.7
SD 312.6 223.8 311.3 463.5
Number 8 7 9 9
* p \ 0.05 versus control; SD = standard deviation; BMD = bone mineral density; BMC = bone mineral content.

123
Volume 466, Number 5, May 2008
the other loaded groups, also had a smaller mineralized
callus volume. In general, fracture callus increases in size
before complete calcification and remodeling, which then
leads to a decrease in size. Thus, the smaller size coupled
with increased osteoid in our study may support the
advanced stage of healing afforded by loading.
Substantial data suggest specific loading regimens in
intact bone induce increased bone formation [23, 26–28,
31]. Although fracture repair mechanisms are clearly more
complex than intact bone adaptation, the mechanisms
involved in periosteal osteoblast proliferation in intact bone
are important in fracture healing as well, and activation of a
mechanically induced response may similarly result in
increased bone formation. In a noninvasive mouse tibia
cantilever bending model, a low-magnitude force applied
for 100 cycles daily was minimally osteogenic, whereas the
same load applied for 10 cycles daily with 10-second
pauses between cycles was a potent stimulant of bone
formation [31]. In a similar study loading the rat ulna, a 14-
second pause between individual load cycles led to a
greater response than insertion of shorter pauses [26].
The overall effect from pause-inserted loading during
fracture healing was modest, but present nonetheless. We
found no mechanical differences among the pause-inserted
and repetitively loaded groups despite observing consid-
erably greater strength using the same repetitive loading
protocol previously [18]. Perhaps the beneficial effects of
pause insertions over repetitive loading results from satu-
ration of the continuous cyclic protocol after a set number
of cycles [31, 32]. Based on our data, the 100-cycle/1-Hz
repetitive loading protocol may not saturate the mechano-
sensitivity of the healing tissue. Superimposing externally
applied load on routine weightbearing, as was the case in
this study, also may lessen the effect of load augmentation
[36].
We chose the 2-week loading period to determine the
differences in the early healing phase. Concordantly,
healing appeared at an accelerated stage in the cycle-
equivalent pause-inserted group as indicated by increased
osteoid deposition. This time in the healing process may
have been before substantial mineralization [4, 21],
accounting for the lack of differences in von Kossa staining
or microCT-based bone mineral content and density.
Considerably smaller mineralized callus size with applied
loading was measured by microCT compared with con-
trols. The smaller size of calluses in the loaded groups may
reflect early organization of the healing callus tissue.
We found axial loading of the fractured mouse tibia with
pauses inserted may lead to acceleration of bone healing
compared with nonloaded controls in the early phases. We
suspect the increased osteoid deposition after pause-inser-
ted loading would likely lead to increased mineral
deposition and earlier fracture consolidation in the later
Pause-inserted Fracture Loading 1237
phases. Additional studies will be required to evaluate the
effects of loading and pause insertions at later times.
Acknowledgments We thank Jonathan Zelken, Alexia Hernandez,
and Xu Yang for technical assistance.
References
1. Aro HT, Chao EY. Bone-healing patterns affected by loading,
fracture fragment stability, fracture type, and fracture site com-
pression. Clin Orthop Relat Res. 1993;293:8–17.
2. Carter DR, Blenman PR, Beaupre GS. Correlations between
mechanical stress history and tissue differentiation in initial
fracture healing. J Orthop Res. 1988;6:736–748.
3. Chao EYS, Inoue N, Elias JJ, Aro H. Enhancement of fracture
healing by mechanical and surgical intervention. Clin Orthop
Relat Res. 1998;355(suppl):S163–178.
4. Cheung KM, Kaluarachi K, Andrew G, Lu W, Chan D,
Cheah KS. An externally fixed femoral fracture model for mice.
J Orthop Res. 2003;21:685–690.
5. Chow JW, Jagger CJ, Chambers TJ Characterization of osteo-
genic response to mechanical stimulation in cancellous bone of
rat caudal vertebrae. Am J Physiol. 1993;265:E340–347.
6. Claes LE, Heigele CA. Magnitudes of local stress and strain
along bony surfaces predict the course and type of fracture
healing. J Biomech. 1999;32:255–266.
7. Claes LE, Heigele CA, Neidlinger-Wilke C, Kaspar D, Seidl W,
Margevicius KJ, Augat P. Effects of mechanical factors on
the fracture healing process. Clin Orthop Relat Res. 1998;
355(suppl):S132–147.
8. Claes LE, Wilke HJ, Augat P, Rubenacker S, Margevicius KJ.
Effect of dynamization on gap healing of diaphyseal fractures
under external fixation. Clin Biomech (Bristol, Avon).
1995;10:227–234.
9. Connelly JT, Fritton JC, van der Meulen MCH. Simulation of
in vivo loading in the tibia of the C57BL/6 mouse. Trans Orthop
Res Soc. 2003;28:0409.
10. Cullen DM, Smith RT, Akhter MP. Bone-loading response varies
with strain magnitude and cycle number. J Appl Physiol.
2001;91:1971–1976.
11. Duda GN, Sollmann M, Sporrer S, Hoffmann JE, Kassi JP,
Khodadadyan C, Raschle M. Interfragmentary motion in tibial
osteotomies stabilized with ring fixators. Clin Orthop Relat Res.
2002;396:163–172.
12. Egger EL, Gottsauner-Wolf F, Palmer J, Aro HT, Chao EY.
Effects of axial dynamization on bone healing. J Trauma.
1993;34:185–192.
13. Ehrlich PJ, Lanyon LE. Mechanical strain and bone cell function:
a review. Osteoporos Int. 2002;13:688–700.
14. Erben RG. Embedding of bone samples in methylmethacrylate:
an improved method suitable for bone histomorphometry, histo-
chemistry, and immunohistochemistry. J Histochem Cytochem.
1997;45:307–313.
15. Feldkamp LA, Goldstein SA, Parfitt AM, Jesion G, Kleerekoper M.
The direct examination of three-dimensional bone architecture
in vitro by computed tomography. J Bone Miner Res. 1989;4:3–11.
16. Forwood MR, Turner CH. The response of rat tibiae to incre-
mental bouts of mechanical loading: a quantum concept for bone
formation. Bone. 1994;15:603–609.
17. Fritton JC, Myers ER, Wright TM, van der Meulen MC. Loading
induces site-specific increases in mineral content assessed by
microcomputed tomography of the mouse tibia. Bone. 2005;36:
1030–1038.
123
1238 Gardner et al.
18. Gardner MJ, van der Meulen MC, Demetrakopoulos D, Wright
TM, Myers ER, Bostrom MP. In vivo cyclic axial compression
affects bone healing in the mouse tibia. J Orthop Res.
2006;24:1679–1686.
19. Goodship AE, Cunningham JL, Kenwright J. Strain rate and
timing of stimulation in mechanical modulation of fracture
healing. Clin Orthop Relat Res. 1998;355(suppl):S105–115.
20. Goodship AE, Kenwright J. The influence of induced micro-
movement upon the healing of experimental tibial fractures.
J Bone Joint Surg Br. 1985;67:650–655.
21. Hiltunen A, Aro HT, Vuorio E. Regulation of extracellular matrix
genes during fracture healing in mice. Clin Orthop Relat Res.
1993;297:23–27.
22. Kenwright J, Richardson JB, Cunningham JL, White SH,
Goodship AE, Adams MA, Magnussen PA, Newman JH. Axial
movement and tibial fractures: a controlled randomised trial of
treatment. J Bone Joint Surg Br. 1991;73:654–659.
23. LaMothe JM, Zernicke RF. Rest insertion combined with high-
frequency loading enhances osteogenesis. J Appl Physiol.
2004;96:1788–1793.
24. Lanyon LE, Rubin CT. Static vs dynamic loads as an influence on
bone remodelling. J Biomech. 1984;17:897–905.
25. Parker DL. Optimal short scan convolution reconstruction for
fanbeam CT. Med Phys. 1982;9:254–257.
26. Robling AG, Burr DB, Turner CH. Recovery periods restore
mechanosensitivity to dynamically loaded bone. J Exp Biol.
2001;204:3389–3399.
27. Robling AG, Hinant FM, Burr DB, Turner CH. Improved bone
structure and strength after long-term mechanical loading is
123
Clinical Orthopaedics and Related Research
greatest if loading is separated into short bouts. J Bone Miner
Res. 2002;17:1545–1554.
28. Robling AG, Hinant FM, Burr DB, Turner CH. Shorter, more
frequent mechanical loading sessions enhance bone mass. Med
Sci Sports Exerc. 2002;34:196–202.
29. Rubin CT, Lanyon LE. Regulation of bone formation by applied
dynamic loads. J Bone Joint Surg Am. 1984;66:397–402.
30. Sarmiento A, Schaeffer JF, Beckerman L, Latta LL, Enis JE.
Fracture healing in rat femora as affected by functional weight-
bearing. J Bone Joint Surg Am. 1977;59:369–375.
31. Srinivasan S, Weimer DA, Agans SC, Bain SD, Gross TS. Low-
magnitude mechanical loading becomes osteogenic when rest is
inserted between each load cycle. J Bone Miner Res. 2002;17:
1613–1620.
32. Turner CH. Three rules for bone adaptation to mechanical stim-
uli. Bone. 1998;23:399–407.
33. Umemura Y, Ishiko T, Yamauchi T, Kurono M, Mashiko S. Five
jumps per day increase bone mass and breaking force in rats.
J Bone Miner Res. 1997;12:1480–1485.
34. Wallace AL, Draper ER, Strachan RK, McCarthy ID, Hughes SP.
The vascular response to fracture micromovement. Clin Orthop
Relat Res. 1994;301:281–290.
35. White AA 3rd, Panjabi MM, Southwick WO. Effects of com-
pression and cyclical loading on fracture healing: a quantitative
biomechanical study. J Biomech. 1977;10:233–239.
36. Wolf S, Janousek A, Pfeil J, Veith W, Haas F, Duda G, Claes L.
The effects of external mechanical stimulation on the healing of
diaphyseal osteotomies fixed by flexible external fixation. Clin
Biomech (Bristol, Avon). 1998;13:359–364.
Clin Orthop Relat Res (2008) 466:1239–1244
DOI 10.1007/s11999-008-0181-z
ORIGINAL ARTICLE
Four Weeks of Mobility After 8 Weeks of Immobility Fails
to Restore Normal Motion
A Preliminary Study
Guy Trudel MD, Jian Zhou BSc, Hans K. Uhthoff MD,
Odette Laneuville PhD
Received: 25 July 2007 / Accepted: 5 February 2008 / Published online: 26 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Prolonged immobilization reduces passive range
of motion of joints creating joint contractures. Whether and
to what extent these iatrogenic contractures can be reduced is
unknown. We raised three questions using an animal model:
What degree of contracture remains at the end of a defined
remobilization period? Do contractures in sham-operated
and immobilized joints differ? What is the contribution of
the posterior knee capsule in limiting knee extension? We
immobilized one knee of 11 adult male rats in flexion to
induce a joint contracture; 10 control animals underwent a
sham operation. After 8 weeks, the internal fixation device
was removed, and the animals were allowed to resume
unrestricted activity for 4 weeks at the end of which the knee
range of motion was measured with standardized torques.
The mean flexion contracture was higher in the immobilized
group (51.9° ± 2.8°) than in the sham-operated group
(18.9° ± 2.1°). Eighty-eight percent of the contractures
One or more of the authors have received funding from the Canadian
Institutes of Health Research (GT, HKU, OL) and the Canadian
Institute of Musculoskeletal Health and Arthritis (JZ).
Each author certifies that his or her institution has approved the
animal protocol for this investigation and that all investigations were
conducted in conformity with ethical principles of research.
G. Trudel, J. Zhou, H. K. Uhthoff
Bone and Joint Laboratory, University of Ottawa, Ottawa, ON,
Canada
G. Trudel
Division of Physical Medicine and Rehabilitation, University of
Ottawa, Ottawa, ON, Canada
O. Laneuville (&)
Department of Biochemistry, Microbiology and Immunology,
Faculty of Medicine, University of Ottawa, 451 Smyth Road,
Ottawa, ON, Canada K1H 8M5
e-mail: olaneuvi@uottawa.ca
remained in the immobilized group after dividing skin and
muscle, suggesting an important contribution of the posterior
knee capsule in limiting knee mobility. Based on our pre-
liminary study the range of motion of rat knees immobilized
for 8 weeks remained substantially reduced after a 4-week
period of unassisted remobilization.
Introduction
Motion of synovial joints is essential to maintain mobility in
their full anatomic range of motion (ROM). Immobility can
reduce passive ROM of joints [1, 6, 32], creating a joint
contracture [2, 22, 28]. Contractures diminish joint function
and lead to a reduction in patient autonomy [17, 20]. The
important role and frequent use of immobilizing procedures
to treat musculoskeletal disorders put joints at risk for
having contractures develop [1, 37]. Patients with limited
mobility and those confined to bed have multiple immo-
bility-induced joint contractures develop [7, 9, 15, 25].
A few experimental studies on immobility-induced joint
contractures suggest the timing of development of con-
tractures. In several studies an internal fixation device to
immobilize one hind limb of rats duplicated joint con-
tractures in humans, with substantial reductions in the
passive ROM [8, 12]. More recently, a time course study
over 32 weeks in rat knees immobilized at 135° flexion
caused a gradual decrease of the ROM in extension that
reached 69° after the first 8 weeks and then leveled off
[32]. The posterior capsular structures contributed to the
restriction [31, 32]. The process by which immobility
changes the mechanical properties of various articular
structures and limits joint movement is not fully under-
stood, but shortening, atrophy, and fibrosis of the posterior
capsule have been reported [1, 18, 27, 30, 33].
123
1240 Trudel et al.
Reversibility is clinically and functionally important
for patients. Michelsson [19] used a plastic splint to
immobilize 11 rabbit knees in extension and reported a
reduction in mobility after 5 weeks that disappeared
spontaneously in most cases after the splint was removed.
But, should the immobilization last 6 to 7 weeks or
longer, there was gradual but incomplete recovery after
8 weeks of remobilization, with the contracture being
severe and permanent. Akeson et al. [4] investigated the
reversibility of immobility-induced joint contractures
using a model of rabbit knees fixed in a flexed position
using transarticular wires. Knees immobilized for 9 weeks
recovered completely 3 weeks after the fixation was
removed. We immobilized canine shoulders for 12 weeks
with a cast to create a contracture. Four weeks of remo-
bilization did not reverse the limited motion but the
limitation began to reverse after 8 weeks of remobiliza-
tion and returned to normal after 12 weeks [24]. Van
Harreveld et al. [35] enforced 8 weeks of exercise after
7 weeks of immobilization of the horse metacarpopha-
langeal joint but could not restore joint motion. Finally,
immobilization of the rat ankle for 2 weeks revealed
that treadmill running improved ROM compared with no
running [23]. The literature on reversibility of joint
contractures in animals therefore seems contradictory. The
differences likely stem in part from the variety of joints,
times and methods of immobilization, and animal species.
The development of joint contractures secondary to
immobility is gradual, and reversibility is not well
documented.
We therefore raised three questions using an animal
model: (1) What degree of contracture remains at the end
of a 4-week remobilization period? (2) Do contractures in
sham-operated and immobilized joints differ? (3) What is
the contribution of the posterior knee capsule in limiting
knee extension?
Materials and Methods
We immobilized the knees of 11 adult Sprague Dawley rats
(300–350 g) in 135° flexion for 8 weeks then remobilized
the knees for 4 weeks. Ten similar rats with sham surgery
but no immobilization served as controls. All rats were
obtained from one source (Charles River Laboratories, St
Constant, Quebec, Canada) and housed individually in
standard cages. At the end of 8 weeks we removed the
fixation devices (experimental animals) and sham screws
(control animals) and allowed free activity (unassisted
remobilization) for 4 weeks. We then sacrificed the ani-
mals and measured knee motion. The contribution of
the posterior knee capsule in limiting joint motion was
determined by sectioning all skin and posterior muscles.
123
Clinical Orthopaedics and Related Research
We immobilized the knees of the 11 experimental rats in
135° flexion using an internal system consisting of a Del-
rin1 plate (DuPont Engineering Polymers, Wilmington,
DE) secured with one screw to the proximal femur and one
screw to the distal tibia. The course of the plate was such
that it passed laterally between the peronei and lateral
gastrocnemius muscles at the leg and medial to the tensor
fasciae lata at the thigh without violating the knee. Details
of the surgical protocol and a description of the immobi-
lization system were published previously [32]. The knees
of 10 rats (control group) underwent similar surgery with
the insertion of two screws but not the plate.
Postoperatively the rats were allowed free activity in
individual cages for 8 weeks. At the end of the 8-week
period, the immobilized and control groups underwent
limited surgery to remove the plate and/or screws and then
were returned to their cages for a 4-week period of unre-
stricted activity in their cages. We defined remobilization
as this 4-week period of spontaneous active mobilization
after discontinuation of immobilization.
At the end of the 4-week remobilization period, the
animals were euthanized with carbon dioxide exposure
using a protocol approved by our animal ethics committee.
Joint ROM was measured immediately with a graded
spring-loaded goniometer (arthrometer) adapted to the rat
knee [29]. The arthrometer measured the joint angles
reached at two torque values: 667 g/cm (Torque 1) and
1060 g/cm (Torque 2). The motion in extension was
measured on the operated knee and the contralateral knee.
The angular displacement of the operated leg, immobilized
or sham, was reported as the mean lack of ROM in
extension (ie, an flexion contracture). The mean joint
contracture was calculated by subtracting the ROM miss-
ing to reach full extension of the operated leg from the
measurement taken from the contralateral leg (Fig. 1). To
elucidate the contribution of the posterior knee capsule in
limiting knee ROM, all skin and muscles crossing the knee
were divided with a scalpel and the ROM measurement
repeated. The contribution of the posterior knee capsule in
limiting knee extension was calculated by the lack of ROM
after division of skin and transarticular muscles.
In humans, the knee extends to 0°. A knee flexion
contracture is defined by the number of degrees missing to
full extension. In normal Sprague Dawley rats, the knee
does not reach full extension, possibly owing to the
habitual knee flexion posture. Therefore, to report knee
flexion contracture, we subtracted the lack of extension of
the contralateral knee from the lack of extension of the
surgically treated knee for the immobilized and the control
animals.
All data were expressed as mean ± standard error of the
mean. Lack of extension was compared between the
immobilized and sham-operated groups using an unpaired
Volume 466, Number 5, May 2008 Partial Reversibility of Joint Contractures 1241

Torque 1 Torque 2

Knee Flexion Contracture (degree)

Knee Flexion Contracture (degree)

40 40
Immobilized Sham Immobilized Sham
35 35
30 30
25 25
20 20
15 15
10 * 10 * *
5 * 5
0 0
Muscle On Muscle Off Muscle On Muscle Off
A B

Fig. 1A–B Knee flexion contractures caused by 8 weeks of immo-
bilization were not reversed by 4 weeks of remobilization.
Contractures are calculated as the ROM lacking in extension of the
surgically treated knee (sham or immobilized group) minus the ROM
lacking in extension of the contralateral knee (sham or immobilized
group) (data presented in Table 1) at two torque values: (A) Torque
1 = 667 g/cm and (B) Torque 2 = 1060 g/cm. Muscle on refers to
results calculated from data on the intact knee; muscle off refers to
results calculated from data on the knee after the skin and
transarticular muscles were removed. Error bars represent standard
error of the mean. * = p B 0.05 between immobilized and sham-
operated groups.

t test. In both groups, the ROM was compared with that of (p \ 0.05) in the immobilized legs than in their contralat-
the contralateral knee using paired t tests. For contracture, eral legs (21.1° ± 2.9°) at both torques (Table 1).
an unpaired t test was used to compare the means of the Immobilized legs also lacked more (p \ 0.05) ROM in
immobilized and control animals. We used a post hoc extension than contralateral legs at both torques (Torque 1:
Bonferroni correction for multiple comparisons. The data 21.1° ± 2.9°; Torque 2: 10.3° ± 2.3°). Range of motion
were stored and analyzed using SPSS 15.0 (SPSS Inc, measurements of the surgically treated and contralateral
Chicago, IL). legs in the sham-operated group were similar (Table 1).
Most (88% of the limitation at Torque 1 and 83% of the
limitation at Torque 2) of the contracture remained after
Results dividing the skin and transarticular muscles (Table 1).
Similarly, the knee flexion contracture remained in the
The mean flexion contracture after remobilization was immobilized group after division of skin and muscle at
51.9° ± 2.8° in the immobilized knees and 18.9° ± 2.1° in both torque values (Fig. 1).
the sham-operated knees at Torque 1 (Table 1, muscle on).
The immobilized group lacked 41.9° ± 3.0° extension
while the sham-operated group lacked 9.8° ± 2.4° at Discussion
Torque 2 (Table 1, muscle on).
The contracture was greater (p \ 0.05) in the immobi- Prolonged immobilization reduces passive ROM of joints
lized group (51.9° ± 2.8°) than in the sham-operated group creating joint contractures. Whether and to what extent
(18.9° ± 2.1°) (Table 1). The contracture also was greater these iatrogenic contractures can be reduced is not clearly

Table 1. Range of motion in extension of rat knees*

Group Number of joints Mean lack of knee extension ± SEM (°)
Torque 1 Torque 2
 
Muscle on Muscle off Muscle on Muscle off

Experimental
Immobilized 11 51.9 ± 2.8 45.6 ± 2.7 41.9 ± 3.0 34.7 ± 2.9
Contralateral 11 21.1 ± 2.9à 15.2 ± 2.5à 10.3 ± 2.3à 6.5 ± 2.3à
Control
Sham-operated 8 18.9 ± 2.1à 15.4 ± 1.4à 9.8 ± 2.4à 6.3 ± 2.0à
à à à
Contralateral 8 17.6 ± 2.7 11.5 ± 2.4 6.9 ± 2.0 2.8 ± 2.5à
* Range of motion measurements correspond to the angular displacement missing to reach complete extension (0°) at the knee; Torque
1 = 667 g/cm; Torque 2 = 1060 g/cm;  the muscle off procedure isolated the posterior joint capsule as the tissue limiting knee ROM;
à
p \ 0.05; SEM = standard error of the mean.

123
1242 Trudel et al. Clinical Orthopaedics and Related Research

Yes: with exercise

known from the few published studies. We therefore raised

No: without exercise

Yes: 6, 9 weeks
No: 1, 2, 3 weeks
three questions using an animal model: (1) What degree of

Yes: 12 weeks
No: 4, 8 weeks
contracture remains at the end of a defined remobilization

Reversible
period? (2) Do contractures in sham-operated and immo-
bilized joints differ? (3) What is the contribution of the

No

No
No
No

No
posterior knee capsule in limiting knee extension?
Given the single time point and number of animals, we

Statistics
considered this a preliminary study and therefore did not
perform a power analysis. Additional studies would be

Yes
Yes
Yes
No
No

No

No

No
required to fully characterize the response to remobiliza-
tion, and we cannot say whether the contractures are

Number
permanent. We first tested animals for combined arthro-

5
11
25

22
28
53
39

10
genic and myogenic restriction (muscle on) at both torque
values and then for arthrogenic restriction only (muscle

(measured at the end

of remobilization)
off) at both torque values. In the first series of testing, the

Range of motion
posterior capsule and other arthrogenic structures may
have been damaged beyond their elastic properties. This
limitation would underestimate the restriction imposed by

Yes

Yes
Yes
Yes

Yes
the capsular component of the knee.

No

No

No
Using the model of knee immobility, we previously

Not assisted 8, 16, 32 weeks

Not assisted 1,2,3,6,9 weeks
documented a contracture after 8 weeks of immobility

Not assisted 4, 8, 12 weeks

Not assisted 6, 8 months
with a ROM of 69° before sectioning skin and transartic-
ular muscles [32]. The current data show 4 weeks of

Not assisted 8 weeks

Stretching 4 weeks
remobilization did not restore normal joint mobility and

Exercise 6 weeks

Exercise 2 weeks
Remobilization

there was a residual 42° extension contracture at Torque 1.

Most of the limitation in knee ROM remained after divi-
sion of skin and muscles. Sectioning of the posterior knee
capsule eliminated all resistance to knee extension [31].
Therefore we concluded the posterior capsule of the knee Flexion cast, 2 weeks 6 weeks
caused the major limitation to knee extension after
Extension plaster, 5 weeks

immobilization and remobilization.

Flexion internal, 9 weeks

Flexion internal, 40 days

Immobilization system

In previous reports on the reversibility of immobility-

Table 2. Literature on reversibility of immobility-induced joint contractures

Flexion cast, 2 weeks

induced joint contractures (Table 2), five studies used a
Flexion 8 weeks

protocol of remobilization not complemented with exercise Cast, 12 weeks

Cast, 7 weeks
or stretching and resulted in a lack of complete revers-
ibility [10, 14, 19, 34, 35]. Of those five studies of
unassisted remobilization, only one study reported both
measurements of angular displacement and statistical
analysis [14]. Similar to ours, that study reported incom-
Horse metacarpo-
Canine shoulder

plete reversibility of joint contractures secondary to

phalangeal
Animal joint

Rabbit knee
Rabbit knee

Rabbit knee
Rabbit knee

immobility [14]. However, the duration of the remobili-

Rat ankle
Rat knee

zation period influenced the reversibility according to two

studies [4, 24]: joint contractures were reversible after long
periods of remobilization but irreversible after short peri-
Finsterbush and Friedman [10]

ods although ROM measurements [24] and statistical

analysis [4, 24] were not reported in those studies. Our data
Van Harreveld et al. [35]

on remobilization after 8 weeks of immobility provide

Schollmeier et al. [24]
Hildebrand et al. [14]

Sakakima et al. [23]

evidence for the lack of reversibility at 4 weeks.

Akeson et al. [4]

Usuba et al. [34]

Many anatomic structures potentially limit mobility of

Michelsson [19]

the knee. They include skin, bone, muscle, tendons, liga-

ments, and capsule. Published studies on the response of
Study

those structures to immobility is available except for the

capsule. Although muscle fibers have plasticity for

123
Volume 466, Number 5, May 2008
adaptive lengthening [26, 37], intrafascicular endomysium
and intramuscular septa and perimysium, composed of
dense connective fascia, may be more resistant to rees-
tablishment of their previous length [1]. This also applies
to tendons and ligaments [3, 5, 11, 13, 16, 21, 36, 38, 39].
Our data suggest a possible role of the capsule and intra-
articular ligaments of the knee in the contracture. The
capsular changes, synovial villi adhesion, synovial short-
ening, and atrophy were not amenable to spontaneous
recovery after 4 weeks of remobilization [27, 30]. A tem-
poral study to evaluate the duration of the immobilization
period leading to irreversible joint contractures would
provide the rationale to develop a protocol for timely
clinical intervention. The incomplete reversibility of joint
contractures underscores the need for active preventive
action on immobile joints.
Range of motion of rat knees immobilized for 8 weeks
remains substantially reduced after a 4-week period of
unassisted remobilization. The posterior knee capsule was
the primary structure restricting ROM.
Acknowledgments We thank Julie Courchesne for assistance with
these experiments.
References
1. Akeson WH, Amiel D, Abel MF, Garfin SR, Woo SL. Effects of
immobilization on joints. Clin Orthop Relat Res. 1987;219:28–
37.
2. Akeson WH, Amiel D, Woo SL. Immobility effects on synovial
joints: the pathomechanics of joint contracture. Biorheology.
1980;17:95–110.
3. Akeson WH, Woo SL, Amiel D, Coutts RD, Daniel D. The
connective tissue response to immobility: biochemical changes in
periarticular connective tissue of the immobilized rabbit knee.
Clin Orthop Relat Res. 1973;93:356–362.
4. Akeson WH, Woo SL, Amiel D, Doty DH. Rapid recovery from
contractures in rabbit hindlimbs: a correlative biomechanical and
biochemical study. Clin Orthop Relat Res. 1977;122:359–365.
5. Amiel D, Woo SL, Harwood FL, Akeson WH. The effect of
immobilization on collagen turnover in connective tissue: a bio-
chemical-biomechanical correlation. Acta Orthop Scand. 1982;
53:325–332.
6. Brandt KD. Response of joint structures to inactivity and to
reloading after immobilization. Arthritis Rheum. 2003;49:267–
271.
7. Clavet H, Hebert P, Fergusson D, Doucette S, Trudel G. Joint
contractures following prolonged stays in the intensive care unit.
CMAJ. 2008, DOI: 10.1503/cmaj.071054.
8. Evans EB, Eggers GW, Butler JK, Blumel J. Experimental
immobilization and remobilization of rat knee joints. J Bone Joint
Surg Am.1960;42:737–758.
9. Fergusson D, Hutton B, Drodge A. The epidemiology of major
joint contractures: a systematic review of the literature. Clin
Orthop Relat Res. 2007;456:22–29.
10. Finsterbush A, Friedman B. Reversibility of joint changes pro-
duced by immobilization in rabbits. Clin Orthop Relat Res.
1975;111:290–298.
Partial Reversibility of Joint Contractures 1243
11. Frank C, MacFarlane B, Edwards P, Rangayyan R, Liu ZQ,
Walsh S, Bray R. A quantitative analysis of matrix alignment in
ligament scars: a comparison of movement versus immobilization
in an immature rabbit model. J Orthop Res. 1991;9:219–227.
12. Furlow LT Jr, Peacock EE Jr. Effect of beta-aminopropionitrile
on joint stiffness in the rat. Ann Surg. 1967;165:442–447.
13. Harper J, Amiel D, Harper E. Collagenases from periarticular
ligaments and tendons: enzyme levels during the development of
joint contractures. Matrix. 1989;9:200–205.
14. Hildebrand KA, Sutherland C, Zhang M. Rabbit knee model of
post-traumatic joint contractures: the long-term natural history
of motion loss and myofibroblasts. J Orthop Res. 2004;22:
313–320.
15. Kagaya H, Sharma M, Kobetic R, Marsolais EB. Ankle, knee,
and hip moments during standing with and without joint con-
tractures: simulation study for functional electrical stimulation.
Am J Phys Med Rehabil. 1998;77:49–54.
16. Karpakka J, Vaananen K, Virtanen P, Savolainen J, Orava S,
Takala TE. The effects of remobilization and exercise on colla-
gen biosynthesis in rat tendon. Acta Physiol Scand. 1990;139:
139–145.
17. Leblebici B, Adam M, Bagis S, Tarim AM, Noyan T, Akman
MN, Haberal MA. Quality of life after burn injury: the impact of
joint contracture. J Burn Care Res. 2006;27:864–868.
18. Matsumoto F, Trudel G, Uhthoff H. High collagen type I and low
collagen type III levels in knee joint contracture: an immuno-
histochemical study with histological correlate. Acta Orthop
Scand. 2002;73:335–343.
19. Michelsson JE. Thickness of the rabbit knee during and after
immobilization. IRCS Med Sci Connect Tissue Skin Bone Pathol
Surg Transplant. 1979;7:36.
20. Morris PE. Moving our critically ill patients: mobility barriers
and benefits. Crit Care Clin. 2007;23:1–20.
21. Noyes FR. Functional properties of knee ligaments and altera-
tions induced by immobilization. Clin Orthop Relat Res.
1977;123:210–242.
22. Perry J. Contractures: a historical perspective. Clin Orthop Relat
Res. 1987;219:8–14.
23. Sakakima H, Yoshida Y, Sakae K, Morimoto N. Different fre-
quency treadmill running in immobilization-induced muscle
atrophy and joint contracture of rats. Scand J Med Sci Sports.
2004;14:186–192.
24. Schollmeier G, Sarker K, Fukuhara K, Uhthoff HK. Structural
and functional changes in the canine shoulder after cessation of
immobilization. Clin Orthop Relat Res. 1996;323:310–315.
25. Souren LE, Franssen EH, Reisberg B. Contractures and loss of
function in patients with Alzheimer’s disease. J Am Geriatr Soc.
1995;43:650–655.
26. Tabary JC, Tabary C, Tardieu C, Tardieu G, Goldspink G.
Physiological and structural changes in the cat’s soleus muscle
due to immobilization at different lengths by plaster casts.
J Physiol. 1972;224:231–244.
27. Trudel G, Jabi M, Uhthoff H. Localized and adaptive synoviocyte
proliferation characteristics in rat knee joint contractures sec-
ondary to immobility. Arch Phys Med Rehabil. 2003;84:1350–
1356.
28. Trudel G, Laneuville O, Uhthoff HK. Joint contractures: editorial
comment. Clin Orthop Relat Res. 2007;456:2.
29. Trudel G, O’Neill PA, Goudreau LA. A mechanical arthrometer
to measure knee joint contracture in rats. IEEE Trans Rehab Eng.
2000;8:149–155.
30. Trudel G, Seki M, Uhthoff HK. Synovial adhesions are more
important than pannus proliferation in the pathogenesis of knee
joint contracture after immobilization: an experimental investi-
gation in the rat. J Rheumatol. 2000;27:351–357.
123
1244 Trudel et al.
31. Trudel G, Uhthoff HK. Contractures secondary to immobility: is
the restriction articular or muscular? An experimental longitu-
dinal study in the rat knee. Arch Phys Med Rehabil. 2000;81:
6–13.
32. Trudel G, Uhthoff HK, Brown M. Extent and direction of joint
motion limitation after prolonged immobility: an experimental
study in the rat. Arch Phys Med Rehabil. 1999;80:1542–1547.
33. Trudel G, Uhthoff HK, Laneuville O. Prothrombin gene expres-
sion in articular cartilage with a putative role in cartilage
degeneration secondary to joint immobility. J Rheumatol. 2005;
32:1547–1555.
34. Usuba M, Akai M, Shirasaki BS, Miyakawa S. Experimental joint
contracture correction with low torque: long duration repeated
stretching. Clin Orthop Relat Res. 2007;456:70–78.
35. Van Harreveld PD, Lillich JD, Kawcak CE, Gaughan EM,
Mclaughlin RM, Debowes RM. Clinical evaluation of the effects
123
Clinical Orthopaedics and Related Research
of immobilization followed by remobilization and exercise on the
metacarpophalangeal joint in horse. Am J Vet Res. 2002;63:282–
288.
36. Walsh S, Frank C, Hart D. Immobilization alters cell metabolism
in an immature ligament. Clin Orthop Relat Res. 1992;277:277–
288.
37. Winkelman C. Inactivity and inflammation in the critically ill
patients. Crit Care Clin. 2007;23:21–34.
38. Woo SL, Gomez MA, Inoue M, Akeson WH. New experimental
procedures to evaluate the biomechanical properties of healing
canine medial collateral ligaments. J Orthop Res. 1987;5:425–
432.
39. Woo SL, Gomez MA, Sites TJ, Newton PO, Orlando CA, Akeson
WH. The biomechanical and morphological changes in the
medial collateral ligament of rabbit after immobilization and
remobilization. J Bone Joint Surg Am. 1987;69:1200–1211.
Clin Orthop Relat Res (2008) 466:1245–1250
DOI 10.1007/s11999-008-0193-8
ORIGINAL ARTICLE
Readability of Online Patient Education Materials
From the AAOS Web Site
Sanjeev Sabharwal MD, Sameer Badarudeen MBBS,
Shebna Unes Kunju MBBS
Received: 6 February 2008 / Accepted: 15 February 2008 / Published online: 7 March 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract One of the goals of the American Academy of
Orthopaedic Surgeons (AAOS) is to disseminate patient
education materials that suit the readability skills of the
patient population. According to standard guidelines from
healthcare organizations, the readability of patient educa-
tion materials should be no higher than the sixth-grade
level. We hypothesized the readability level of patient
education materials available on the AAOS Web site would
be higher than the recommended grade level, regardless
when the material was available online. Readability scores
of all articles from the AAOS Internet-based patient
information Web site, ‘‘Your Orthopaedic Connection,’’
were determined using the Flesch-Kincaid grade formula.
The mean Flesch-Kincaid grade level of the 426 unique
articles was 10.43. Only 10 (2%) of the articles had the
recommended readability level of sixth grade or lower. The
readability of the articles did not change with time. Our
findings suggest the majority of the patient education
materials available on the AAOS Web site had readability
scores that may be too difficult for comprehension by a
substantial portion of the patient population.
Each author certifies that he or she has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
S. Sabharwal (&), S. Badarudeen, S. Unes Kunju
Department of Orthopedics, University of Medicine and
Dentistry of New Jersey–New Jersey Medical School,
Doctor’s Office Center, 90 Bergen Street, Suite 7300, Newark,
NJ 07103, USA
e-mail: sabharsa@umdnj.edu
Introduction
In 2000, the American Academy of Orthopaedic Surgeons
(AAOS) introduced the Internet-based patient education
database, ‘‘Your Orthopaedic Connection [3].’’ Every day
more than 35,000 people visit this Web site to access
education materials on diverse orthopaedic conditions [2]
(oral communication, Jim Ogale, AAOS Web site staff,
June 6, 2007). The goal of this Web site is to enhance
patient-physician communication by providing validated
and up-to-date information about various orthopaedic
conditions in a way that is ‘‘sensitive to diversity and
readability and to strengthen the bond between physicians
and patients [24].’’
Readability of a text is the reading comprehension level
a person must have to understand the written material and
is an important determinant of a person’s ability to com-
prehend health information [1, 14, 38]. The Flesch-Kincaid
(FK) grade formula is one of the most commonly used
tools to assess the readability of written materials in terms
of the academic grade [1, 14]. A higher FK grade level of a
text indicates a greater level of difficulty to read and
comprehend the material and thus requires more advanced
reading skills than would be required of a text with a lower
FK grade level. Organizations like the National Institutes
of Health, the National Work Group on Cancer and Health,
and the American Medical Association have recommended
the readability of patient education materials should be no
higher than the sixth-grade level [15, 34, 35]. The average
readability of the US adult population is at the eight-grade
level [18]. However, several studies suggest the readability
of patient education materials in most medical specialties is
beyond these recommended levels [1, 14, 18, 34]. We
recently assessed the readability scores of patient education
materials that were available at the AAOS and Pediatric
123
1246 Sabharwal et al.
Orthopaedic Society of North America (POSNA) Web
sites, pertaining exclusively to pediatric orthopaedic con-
ditions [5]. Only 2% of the articles available in 2007 had a
FK readability score of the sixth grade level or less. It is
unknown whether there is a trend toward greater read-
ability on these two sites, but if the readability of articles
remains high over a period of time, it would suggest lack of
awareness of the importance of readability in the ortho-
paedic community.
Based on the cited reports of relatively high reading
levels of patient education material, we hypothesized the
readability level of the majority of patient education
materials available on the AAOS Web site would be higher
than the recommended FK grade level of sixth grade. We
then asked whether there was a trend toward lower read-
ability scores of articles available on the Web site with
time. Finally, we asked whether the variability of the
readability level of articles would differ between subject
categories in the patient education Web site.
Materials and Methods
We searched the patient education database, ‘‘Your
Orthopaedic Connection,’’ of the AAOS Web site [3]
during April 2007 and downloaded all 663 patient edu-
cation articles. Materials had been created between
September 1999 and July 2006. We excluded articles
written in a language other than English. We noted the
subject category of each article and the date of the last
update and saved the text as a separate Microsoft1
Word1 (Microsoft Corp, Redmond, WA) document. The
text was copied as plain text format to avoid HTML tags
[30]. Any information related to Web page navigation,
copyright notices, postal addresses, phone numbers, uni-
form resource locaters (URLs), disclaimers, date stamps,
author information, citations, references, feedback ques-
tionnaires, or hyperlinks were omitted. Followup editing
was performed as recommended by Flesch and others
[19, 21] by removing decimal points from numbers and
colons and semicolons in sentences. The FK grade for-
mula is best suited for text arranged in a paragraph as
opposed to list format [32]. A running passage containing
a minimum of 35 words was defined as a paragraph. To
get the most representative sample, similar to methods
adopted by other researchers, the three longest paragraphs
of each patient education article were selected [1, 17].
Our Institutional Review Board granted a waiver for the
study.
There were 663 articles grouped under 21 subject cat-
egories in the AAOS patient information Web site, ‘‘Your
Orthopaedic Connection [3].’’ Forty of these articles were
in Spanish and thus excluded from the study. Of the
123
Clinical Orthopaedics and Related Research
remaining 623 articles, 573 articles in 19 subject categories
met the study’s inclusion criteria. Of the 573 articles, 147
were listed in more than one subject category, resulting in
426 unique articles.
The FK grade level of readability was assessed using
Microsoft1 Office Word1 software (2003 Service Pack 2;
Microsoft Corp). The built-in tool to measure readability is
disabled by default, and the user has to enable it by
sequentially selecting the commands ‘‘Tools,’’ ‘‘Options,’’
‘‘Spelling and Grammar,’’ and then enabling the option
‘‘Show readability statistics’’ followed by clicking on the
icon for ‘‘Spelling and Grammar’’ from the tool bar. The
underlying formula for determining the FK grade level is as
follows [21, 27]: (0.39 9 average number of words per
sentence) + (11.8 9 average number of syllables per
word) – 15.59.
The same individual (SUK) calculated all FK grade
levels. Interobserver reliability was assessed by calculating
the intraclass correlation coefficient using 30 randomly
selected articles that were graded by another individual
(SB). An intraclass correlation coefficient of 0 to 0.24
reflects poor correlation; 0.25 to 0.49, low; 0.50 to 0.69,
fair; 0.7 to 0.89, good; and 0.9 to 1.0, excellent [33]. The
intraclass correlation coefficient for assessing FK grade
level was 0.96, indicating excellent interobserver
reliability.
The mean and 95% confidence interval values of the FK
grade level were calculated. Using descriptive statistics and
analysis of variance, the readability grade scores of articles
grouped under different subject categories were analyzed.
A two-sample t test was used to compare the FK grade for
articles in each subject category against the rest of the
articles. The Pearson correlation coefficient (r) was calcu-
lated to study the relationship of the FK grade of articles
with the date when the latest version of the article was
available online. Statistical analysis was performed using
the SAS1 software (Version 9.1; SAS Institute Inc, Cary,
NC).
Results
The majority (98%) of the 426 unique articles on the
patient education Web site had readability scores that were
higher than the sixth-grade level. Only 10 articles (2%) had
the recommended readability level of the sixth-grade level
or less [15, 34, 35]. The mean FK grade level of the 426
articles was 10.4 (95% confidence interval, 10.2–10.6).
Moreover, 85% of the articles had readability above the
eighth-grade level (Fig. 1).
The readability level did not change with time
(r = 0.0003) (Fig. 2). Thus the readability of articles
remained high throughout the entire period.
Volume 466, Number 5, May 2008 Readability of Patient Education Materials 1247

Table 1. Flesch-Kincaid grade level of AAOS patient education

articles (n = 573) in 19 subject categories
Subject categories Number Flesch-Kincaid grade level p Value
of articles
Mean 95% confidence
interval

Arm 13 9.8 9.0–10.7 0.346

Arthritis 22 10.8 10.2–11.3 0.146
Children 48 10.1 9.7–10.5 0.191
Foot* 41 9.9 9.4–10.3 0.028
General 61 10.9 10.5–11.2 0.006
Fig. 1 The distribution of FK readability grades of 426 unique
information**
articles that were available at the AAOS patient education Web site,
‘‘Your Orthopaedic Connection,’’ [3] is illustrated. Hand* 32 9.7 9.3–10.1 0.001
Hip 26 10.4 9.7–11.0 0.99
Injury prevention 64 10.0 9.3–10.7 0.23
Joint replacement 28 10.8 10.0–11.7 0.212
Knee 34 10.9 10.4–11.5 0.101
Neck 7 9.9 8.9–10.9 0.549
Osteoporosis 11 10.9 9.2–12.6 0.373
Patient stories** 33 11.9 11.4–12.4 \ 0.001
Patient-centered 15 11.3 9.8–12.7 0.086
care
Shoulder 21 10.2 9.5–11.0 0.78
Sports/exercise 68 9.8 9.0–10.5 0.083
Spine 22 10.0 9.1–10.9 0.421
Tumors 12 9.6 8.7–10.4 0.172
Women’s health 15 10.9 9.8–12.0 0.307
p value refers to the t test comparing each subject category with the
Fig. 2 A scatterplot displays the lack of relationship of the FK grade
remaining articles; *mean FK grade level was significantly lower than
of the online articles with the date when the latest version of the
the remaining articles; **mean FK grade significantly higher than the
article was available online. Pearson correlation coefficient
remaining articles.
r = 0.0003 (n = 426).

Although none of the 19 subcategories of articles
achieved a mean FK grade level of sixth or lower
(Table 1), we found substantial variation in the readability
scores among the patient education articles in different
subcategories. The FK grades of articles in the Hand and
Foot sections were lower (p = 0.001 and p = 0.028,
respectively), indicating easier readability, whereas the FK
grades of articles in the General Information and Patient
Stories categories were higher (p = 0.006 and p \ 0.001,
respectively) than the remaining articles.
Discussion
There is increasing concern regarding the disparity
between the readability levels of patient education mate-
rials available online and the reading skills of the target
population [1, 14, 18]. Based on our review of the literature
and our recent report dealing exclusively with the read-
ability scores of online pediatric orthopaedic materials [5],
the readability of medical information is higher than
generally recommended. Based on this literature we
hypothesized the readability level of the majority of patient
education materials available on the AAOS Web site,
‘‘Your Orthopaedic Connection [3]’’ would be higher than
the recommended FK grade level of sixth grade. We then
asked whether there was a trend toward lower readability
scores of articles available on the Web site with time.
Finally, we asked whether the variability of the readability
level of articles would differ between subject categories in
the patient education Web site.
We note several limitations in our study. First, there is
an inherent weakness in the assessment of readability of
health-related text using the FK grade level because this
tool relies solely on the number of syllables in a word and
the number of words in a sentence [19]; the number of
syllables may not accurately reflect reading level. In the
field of medicine, the unfamiliarity of nonprofessionals to
medical terms, even if they are short, such as ‘‘lupus,’’
‘‘physis,’’ and ‘‘colon,’’ can lead to underestimation of the
reading skills required to fully comprehend medical liter-
ature with the use of the FK grading system [23]. Although

123
1248 Sabharwal et al.
comprehension of a given material may be enhanced by the
addition of illustrative figures, improved layout, and
appropriate use of font size and color [34], the FK grading
tool does not assess these features. Other instruments such
as the Suitability Assessment of Materials (SAM), a rela-
tively new tool, can assess these factors to measure the
comprehensibility of patient education materials [7, 18, 25,
37]. However, SAM is not as validated as the FK grade
score, is more time-consuming [18], and is ‘‘inherently
subjective [37].’’ Furthermore, the scoring system by SAM
is not based on grades and thus is incompatible with the
recommendations by healthcare organizations [18]. Sec-
ond, we did not directly assess the reading skills of our
patient population. There are many validated tools avail-
able to assess the reading skills of a given population.
These include the Wide Range Achievement Test, Rapid
Estimate of Adult Literacy in Medicine, and the Test of
Functional Health Literacy in Adults [10]. However, in the
context of the current study, because the patient education
materials available through the AAOS Web site are avail-
able in the public domain and do not have a definite target
population, trying to match the reading level of these
materials to the reading skills of our patient population may
be irrelevant. Third, we excluded articles that were avail-
able in a language other than English. Given that 41% of
the people who are visiting the AAOS Web site are from
countries where English is not the primary language, the
need to make patient education articles easier to read is
even more imperative [2]. Fourth, we did not assess the
entire text of the articles on the Web site, but only the
longest three paragraphs. We presume these would reflect
the entire article. Finally, we limited our assessment of
readability of orthopaedic patient education materials to
one Web site. Patients and their caretakers may access
more than one Web site to gain additional insight into their
orthopaedic ailments. Nevertheless, we believe our study is
relevant because orthopaedic surgeons often direct their
patients to the AAOS Web site to find accurate, peer-
reviewed, and up-to-date information. The Web site reach
as measured by number of daily unique visitors of more
than 35,000 adds further credibility to our sample selection
[2] (oral communication, Jim Ogale, AAOS Web site staff,
June 6, 2007).
Numerous authors report patients seeking orthopaedic
care extensively use the Internet as a resource for patient
education [8, 9, 13, 28]. Access to the Internet and its use to
obtain health information are increasing globally at a rapid
pace. One study reported three of every four patients
attending an orthopaedic clinic have access to the Internet
[13]. The Pew Research Center report found, in 2006 alone,
more than 100 million people in the United States searched
the Internet to find health information about diseases from
which they or their relatives and friends suffer [20].
123
Clinical Orthopaedics and Related Research
However, patient education materials available on the
Internet may not have the appropriate reading level for the
average person. A survey conducted in a pediatric ortho-
.
paedic outpatient clinic reported 2 3 of the respondents
believed the health information available on the Internet
was ‘‘too technical [4].’’ However, the authors did not
assess the issue of readability. Our findings highlight the
need for orthopaedic surgeons and educators to recognize
the concept of readability while preparing and reviewing
health education materials for their patients. Although only
2% of the English articles on the AAOS Web site had
readability at the sixth-grade level or lower, creating
musculoskeletal education articles with easier readability
for patients seems attainable.
A recent national adult literacy survey found 40 million
people in the adult US population have literacy skills
equivalent to less than the fifth-grade level and another
50 million have reading skills between the sixth- and
eighth-grade levels [29, 36]. Reading grade level is distinct
from the last academic grade achieved in schools and
colleges. Patients read approximately five grades lower
than their highest attained academic grade [26, 31]. The
reason for this difference between academic grade and
reading level may be multifactorial and possibly related to
a flaw in the education system and the fact that a sub-
stantial portion of the patient population belong to the
lower socioeconomic status [18]. Health literacy is the best
predictor of an individual’s health status [38]. It is defined
as the ‘‘degree to which individuals have the capacity to
obtain, process and understand health information and
services needed to make appropriate health decisions [29,
36].’’ Low health literacy is associated with poor health
status, increased hospitalization rate, poor compliance to
treatment, missed appointments, and increased healthcare
expenditure [6, 10, 18, 34, 36].
The improved access to the Internet is making the online
population increasingly similar to the general population
[20]. The fact that 33% of the ‘‘online health seekers’’ had
only a high school diploma or less education [20] adds
further proof to this assumption. Recently, a national sur-
vey found 20% of adults with ‘‘below basic health
literacy’’’ are getting their health information from the
Internet [29]. Some studies also suggest even people with
good literacy skills prefer materials written in simpler
format and low grade level [16, 37]. Furthermore, surgeons
often are using patient education handouts printed from
Web pages for their patients, and thus the readability level
of Internet-based health information materials should serve
the needs of all segments of society [23]. Although the data
regarding readability skills of Internet users are lacking,
most researchers use readability standards of the general
patient population to assess the readability of online
materials [11, 12, 16, 22, 23, 30, 37].
Volume 466, Number 5, May 2008
Readability of patient education materials can be
assessed readily using the FK grade formula, with a rela-
tively simple set of keystroke instructions and software
available on most personal computers. The FK grade for-
mula originally was developed for the US military in the
early 1970s [27]. Since then, this instrument has been
extensively validated and researched [14, 21, 31]. The
advent of computers and software capable of automating
the calculation has made the formula relatively simple,
quick, and intuitive [14, 21, 31] and within the reach of
almost every healthcare worker. As established in our
study, the FK grade formula has very high interobserver
reliability. Our findings support the hypothesis that the
readability of patient education materials in the AAOS
patient education database would be higher than the rec-
ommended level. In addition, we found the readability
scores of the text did not improve with time. These findings
may reflect the lack of awareness regarding the concept of
readability in the orthopaedic community. Moreover, in a
recent study [5], despite using a smaller and distinct set of
online patient education materials dealing exclusively with
pediatric orthopaedic conditions, we arrived at a similar
conclusion with only 2% of the articles having readability
scores of the sixth grade level or less. The method used in
that study [5] also was different, in that the entire text was
subject to the FK formula as opposed to assessing the
readability score based on the longest three paragraphs, as
in the current study.
In addition to the FK grade level available on the
Microsoft1 Office software, there are other software
packages, such as Corel1 WordPerfect1 Office X3
(Corel Corp, Ottawa, Ontario, Canada), Readability Cal-
culations (Micro Power and Light Co, Dallas, TX),
Readability Studio 1.1 (Oleander Solutions, Vandalia,
OH), and InText (Social Science Consulting, Rudolstadt,
Germany) [35], that have readability assessment tools.
However, we believe the readability score of a given text
should not be the sole criterion to develop patient edu-
cation materials. The ‘‘Living word vocabulary’’ contains
approximately 43,000 words arranged in various grade
levels of complexity and can be used to substitute diffi-
cult words [15, 31]. To allow improved comprehension
by a larger segment of the society, simpler words can
replace complex medical jargon. A list of lay terms that
can be used instead of medical terms can be found at the
following Web sites: http://uuhsc.utah.edu/pated/authors/
substitute2.html and http://plainlanguage.gov/howto/word
suggestions/index.cfm [15]. Detailed instructions on
making more comprehensible patient education literature
are available through the American Medical Association
[38].
Ensuring patients receive education materials they can
understand is the responsibility of physicians, professional
Readability of Patient Education Materials 1249
organizations, and healthcare institutions [15]. Our find-
ings suggest a substantial portion of the patient education
articles available at the patient education library of the
AAOS Web site have readability scores that are higher
than the sixth-grade level. These findings suggest the
online material presented may be too difficult for com-
prehension by a substantial portion of the patient
population. To enhance the patient-physician dialogue,
orthopaedic educators should attempt to keep the read-
ability level of the patient education materials at a FK
grade level of sixth grade or lower. Such measures will
help accomplish the goal of the AAOS Web site of
‘‘improving the communication between orthopaedic sur-
geons and their patients’’ [24] and positively influence the
health outcome of our patients.
Acknowledgments We acknowledge the contributions of Caixia
Zhao, MD, and Emily McClemens, PA-C, in preparation of this
manuscript.
References
1. Albright J, de Guzman C, Acebo P, Paiva D, Faulkner M,
Swanson J. Readability of patient education materials: implica-
tions for clinical practice. Appl Nurs Res. 1996;9:139–143.
2. Alexa1 the Web Information Company. Available at: http://
www.alexa.com/data/details/traffic_details?q=orthoinfo.org&url=
orthoinfo.org/ Accessed June 4, 2007.
3. American Academy of Orthopaedic Surgeons. Your orthopaedic
connection. Available at: http://orthoinfo.aaos.org/all.cfm.
Accessed February 6, 2007.
4. Aslam N, Bowyer D, Wainwright A, Theologis T, Benson M.
Evaluation of Internet use by paediatric orthopaedic outpatients
and the quality of information available. J Pediatr Orthop B.
2005;14:129–133.
5. Badarudeen S, Sabharwal S. Readability of patient education
materials from the American Academy of Orthopaedic Surgeons
and Pediatric Orthopaedic Society of North America Web sites.
J Bone Joint Surg Am. 2008;90:199–204.
6. Baker DW, Parker RM, Williams MV, Clark WS, Nurss J. The
relationship of patient reading ability to self-reported health and
use of health services. Am J Public Health. 1997;87:1027–1030.
7. Bass L. Health literacy: implications for teaching the adult
patient. J Infus Nurs. 2005;28:15–22.
8. Beall MS 3rd, Beall MS Jr, Greenfield ML, Biermann JS. Patient
Internet use in a community outpatient orthopaedic practice. Iowa
Orthop J. 2002;22:103–107.
9. Beall MS 3rd, Golladay GJ, Greenfield ML, Hensinger RN,
Biermann JS. Use of the Internet by pediatric orthopaedic out-
patients. J Pediatr Orthop. 2002;22:261–264.
10. Berkman ND, DeWalt DA, Pignone MP, Sheridan SL, Lohr KN,
Lux L, Sutton SF, Swinson T, Bonito AJ. Literacy, Health Out-
comes: Evidence Report. Rockville, MD: Agency for Healthcare
Research and Quality/US Department of Health and Human
Services; 2004.
11. Berland GK, Elliott MN, Morales LS, Algazy JI, Kravitz RL,
Broder MS, Kanouse DE, Munoz JA, Puyol JA, Lara M, Watkins
KE, Yang H, McGlynn EA. Health information on the Internet:
accessibility, quality, and readability in English and Spanish.
JAMA. 2001;285:2612–2621.
123
1250 Sabharwal et al.
12. Boulos MN. British internet-derived patient information on dia-
betes mellitus: is it readable? Diabetes Technol Ther.
2005;7:528–535.
13. Brooks BA. Using the Internet for patient education. Orthop
Nurs. 2001;20:69–77.
14. Cooley ME, Moriarty H, Berger MS, Selm-Orr D, Coyle B, Short
T. Patient literacy and the readability of written cancer educa-
tional materials. Oncol Nurs Forum. 1995;22:1345–1351.
15. Cotugna N, Vickery CE, Carpenter-Haefele KM. Evaluation of
literacy level of patient education pages in health-related jour-
nals. J Community Health. 2005;30:213–219.
16. D’Alessandro DM, Kingsley P, Johnson-West J. The readability
of pediatric patient education materials on the World Wide Web.
Arch Pediatr Adolesc Med. 2001;155:807–812.
17. Davis TC, Mayeaux EJ, Fredrickson D, Bocchini JA Jr, Jackson
RH, Murphy PW. Reading ability of parents compared with
reading level of pediatric patient education materials. Pediatrics.
1994;93:460–468.
18. Doak CC, Doak LG, Root JH. Teaching Patients With Low
Literacy Skills. 2nd ed. Philadelphia, PA: JB Lippincott; 1996.
19. Flesch RF. How to Write Plain English: A Book for Lawyers,
Consumers. 1st ed. New York, NY: Barnes and Noble; 1981.
20. Fox S. Online Health Search 2006: Report. Washington, DC:
Pew Internet and American Life Project; 2006.
21. Friedman DB, Hoffman-Goetz L. A systematic review of read-
ability and comprehension instruments used for print and web-
based cancer information. Health Educ Behav. 2006;33:352–373.
22. Friedman DB, Hoffman-Goetz L, Arocha JF. Readability of
cancer information on the Internet. J Cancer Educ. 2004;19:117–
122.
23. Graber MA, Roller CM, Kaeble B. Readability levels of patient
education material on the World Wide Web. J Fam Pract.
1999;48:58–61.
24. Hirsch SA. Academy introduces Web-based medical education
for fellows and patients. J Bone Joint Surg Am. 2000;82:1665–
1667.
25. Hoffmann T, McKenna K. Analysis of stroke patients’ and ca-
rers’ reading ability and the content and design of written
materials: recommendations for improving written stroke infor-
mation. Patient Educ Couns. 2006;60:286–293.
123
Clinical Orthopaedics and Related Research
26. Jackson RH, Davis TC, Bairnsfather LE, George RB, Crouch
MA, Gault H. Patient reading ability: an overlooked problem in
health care. South Med J. 1991;84:1172–1175.
27. Kincaid JP, Fishburne RP, Rogers RL, Chissom BS. Derivation of
New Readability Formulas (Automated Readability Index, Fog
Count, and Flesch Reading Ease Formula) for Navy Enlisted
Personnel: Report. Millington, TN: Research Branch, Naval
Technical Training Command; 1975.
28. Krempec J, Hall J, Biermann JS. Internet use by patients in
orthopaedic surgery. Iowa Orthop J. 2003;23:80–82.
29. Kutner M, Greenburg E, Jin Y, Paulsen C. The Health Literacy of
America’s Adults: Results from the 2003 National Assessment of
Adult Literacy. Washington, DC: National Center for Education
Statistics; 2006.
30. Lewiecki EM, Rudolph LA, Kiebzak GM, Chavez JR, Thorpe
BM. Assessment of osteoporosis-website quality. Osteoporosis
Int. 2006;17:741–752.
31. Ley P, Florio T. The use of readability formulas in health care.
Psychol Health Med. 1996;1:7–28.
32. Mosenthal PB, Kirsch IS. A new measure of assessing document
complexity: the PMOSE/IKIRSCH Document Readability For-
mula. J Adolesc Adult Lit. 1998;41:638–657.
33. Munro BH. Correlation. Statistical Methods for Health Care
Research. 5th ed. Philadelphia, PA: Lippincott-Raven; 2005:239–
258.
34. Murero M, D’Ancona G, Karamanoukian H. Use of the Internet
by patients before and after cardiac surgery: telephone survey.
J Med Internet Res. 2001;3:E27.
35. National Institutes of Health. MedlinePlus: How to write easy to
read health materials. Available at: http://www.nlm.nih.gov/
medlineplus/etr.html. Accessed May 1, 2007.
36. Nielsen-Bohlman L, Panzer AM, Kindig DA. Health Literacy: A
Prescription to End Confusion. Washington, DC: Institute of
Medicine, National Academies Press; 2004.
37. Wallace LS, Turner LW, Ballard JE, Keenum AJ, Weiss BD.
Evaluation of Web-based osteoporosis educational materials.
J Womens Health. 2005;14:936–945.
38. Weiss BD. Health Literacy: A Manual for Clinicians. Chicago,
IL: American Medical Association, American Medical Founda-
tion; 2003.
Clin Orthop Relat Res (2008) 466:1251–1256
DOI 10.1007/s11999-008-0125-7
CASE REPORT
Case Report
Parachordoma of Soft Tissues of the Arm
Jonathan Clabeaux MD, Leonard Hojnowski MD,
Alfredo Valente MD, Timothy A. Damron MD
Received: 7 August 2007 / Accepted: 8 January 2008 / Published online: 25 January 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract Parachordoma, or myoepithelioma, is a very
rare tumor histologically resembling chordoma but occur-
ring in the nonaxial soft tissues. It typically has an indolent
nature, with occasional late recurrence and even rare
metastases. Review of existing literature reveals a male
predilection, with the tumor typically occurring in the
fourth decade of life in the lower extremity. It typically is
managed with wide resection. We report the case of a 60-
year-old woman with a right distal upper arm parachord-
oma treated with wide resection of the tumor.
Introduction
Parachordoma, also known as myoepithelioma, is an
extremely rare soft tissue tumor that histologically resem-
bles chordoma of the axial skeleton in a nonaxial location.
Each author certifies that he has no commercial associations
(eg, consultancies, stock ownership, equity interest, patent/licensing,
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his or her institution has approved or waived
approval for the reporting of this case and that all investigations were
conducted in conformity with ethical principles of research.
J. Clabeaux, T. A. Damron (&)
Department of Orthopedic Surgery, Upstate Medical University,
Suite #130, 550 Harrison Street, Syracuse, NY 13202, USA
e-mail: damront@upstate.edu
L. Hojnowski
Department of Radiology, Upstate Medical University, Syracuse,
NY, USA
A. Valente
Department of Pathology, Upstate Medical University, Syracuse,
NY, USA
Fewer than 50 cases are reported in the English literature
[14]. The first description is credited to Laskowski in 1951
[4], who initially labeled it chordoma periphericum.
Dabska [4] later renamed it parachordoma and published
the largest initial case series of 10 cases in 1977. Para-
chordoma is a soft tissue tumor characterized by an
indolent nature, slow growth, occasional late recurrence,
rare metastases, and at least two reported fatalities. Histo-
logically, it resembles chordoma with a wider variation in
appearance. We report the current case to increase aware-
ness of this unusual soft tissue tumor.
Case Report
A 60-year-old right-hand–dominant woman presented with
a soft tissue mass in the posteromedial aspect of her right
distal arm of 3 months’ duration. She denied any trauma to
that area but stated the mass had slowly enlarged. She
denied any pain, numbness, or tingling in the distal
extremity. She also denied fevers, chills, weight loss, loss
of energy, and loss of appetite but did have some occa-
sional night sweats. She had a history of localized thyroid
cancer in 1989 treated with thyroidectomy. She denied any
lumps or bumps elsewhere.
Physical examination revealed a healthy-appearing
woman with a palpable, firm, deep, slightly warm mass
over the posteromedial aspect of her distal right arm that
measured approximately 7 cm proximodistal by 4 cm
mediolateral. Tinel’s test was negative, and there was no
bruit, thrill, or tenderness to palpation over the mass.
The initial radiographic studies included plain radio-
graphs and MRI of the extremity. The radiographs showed
a slight soft tissue density at the posteromedial aspect of
the arm without any evidence of periosteal reaction or
123
1252 Clabeaux et al.
calcification in the mass (Fig. 1). Magnetic resonance
imaging revealed a complex, heterogeneous, deeply situ-
ated intramuscular soft tissue mass adjacent and posterior
to the neurovascular bundle on the medial aspect of the arm
in the triceps muscle belly. The mass was predominantly
dark with heterogeneity on the T1-weighted images
(Fig. 2A) and bright with some heterogeneity on
T2-weighted images (Fig. 2B). There was minimal soft
tissue edema surrounding the mass. There was no bony
involvement and no periosteal reaction, but it did appear to
abut the humerus medially.
The nonspecific MRI features of the deep, large heter-
ogeneous soft tissue mass yielded a broad differential
diagnosis, including benign and malignant entities such as
malignant fibrous histiocytoma, fibrosarcoma, and even
metastatic disease. Tissue for pathologic examination was
obtained through a Tru-Cut core needle (Travenol Labo-
ratories, Inc, Deerfield, IL) biopsy in the office. Staging
studies, which included computed tomographic scans of the
chest, abdomen, and pelvis, were negative for the presence
of metastases. Laboratory values were all within normal
Fig. 1 A plain radiograph of the elbow and distal arm shows only a
nonmineralized soft tissue shadow.
123
Clinical Orthopaedics and Related Research
limits, including a leukocyte count of 6000, erythrocyte
sedimentation rate of 15 mm/hour, and C-reactive protein
less than 0.5.
The Tru-Cut needle biopsy specimen revealed a tumor
composed of nests and cords of cells with eosinophilic and
vacuolated cytoplasm and relatively uniform round to oval
nuclei. The cells were embedded in a myxoid and hyaline
stroma and separated by broad fibrous septa (Fig. 3). Foci
of recent and remote hemorrhage were present, but necrosis
was not identified. In addition, there was a low mitotic rate
not exceeding one mitosis per 20 high-power fields
(Fig. 4).
Differential diagnosis for these histologic findings
included parachordoma, extraskeletal myxoid chondrosar-
coma, and chordoma. Immunohistochemistry revealed
expression of the cytokeratin CAM5.2 and S-100 protein
by tumor cells with negative staining for cytokeratins AE1/
AE3, cytokeratin 19, epithelial membrane antigen (EMA),
muscle-specific actin, desmin, and carcinoembryonic anti-
gen (CEA). An alcian blue stain was strongly positive in
the stroma, with marked reduction of the staining after
hyaluronidase digestion. The diffuse CAM5.2 and S-100
expression is an immunophenotype not seen in extraskel-
etal myxoid chondrosarcoma, although negative
cytokeratin 19 and CEA with hyaluronidase digestion of
the matrix favored parachordoma rather than chordoma.
Because extraskeletal myxoid chondrosarcomas are char-
acterized by the balanced chromosomal translocation
t(9;22)(q22;q12), with the breakpoint involving the EWS
gene on chromosome 22q12 and the CHN gene on 9q22,
fluorescence in situ hybridization studies for the detection
of the EWSR1 gene break-apart rearrangement were per-
formed, and the results were negative.
The patient underwent wide resection of the tumor
2 weeks later. A curvilinear incision was made on the
posteromedial aspect of the arm, which included excision
of the biopsy tract. The ulnar nerve was identified and
protected during the entire procedure. The mass was
removed en bloc with negative intraoperative frozen sec-
tion margins. She experienced some mild numbness along
her small finger and the ulnar border of the ring finger but
full abduction strength. The final pathologic results also
were consistent with parachordoma. There has been no
recurrence to date during the past 3 months of followup.
Discussion
Parachordoma (or myoepithelioma) is a rare soft tissue
tumor first described by Laskowski and then renamed by
Dabska [4] in 1977. The differential diagnosis often
includes extraskeletal myxoid chondrosarcoma, and this
was considered in the pathologic differential diagnosis in
Volume 466, Number 5, May 2008
Fig. 2A–B Sagittal MRI of the distal upper arm soft tissue mass shows a (A) heterogeneous dark signal on the T1-weighted image and a (B)
heterogeneous bright signal on the T2-weighted image.
Fig. 3 A low-power photomicrograph of the tumor shows multino-
dular masses separated by broad collagen bands (Stain, hematoxylin
and eosin; original magnification, 920).
the current case [8]. Parachordoma originally was believed
to be a chordoma occurring in nonaxial sites, but more
recent work suggests it has a distinct immunohistochemical
profile [6]. Presently, parachordoma is considered a unique
entity [14].
Light microscopic features of parachordoma include
cells in clusters, nodules, whorls, and a pseudoglandular
formation of rounded cells in a focally myxoid stroma
separated by fibrous tissue [6]. In general, nuclei are bland,
mitotic figures are rare, and there is no necrosis or vascular
invasion. The tumor is circumscribed but not encapsulated.
Parachordoma of Soft Tissues of the Arm 1253
Fig. 4 A high-power magnification with tumor cells shows vacuo-
lated cytoplasm and lack of considerable nuclear atypia (Stain,
hematoxylin and eosin; original magnification, 9400).
Parachordoma is composed of three cell types, spe-
cifically epithelioid cells, smaller glomoid cells, and
spindle cells [8]. All lesions have a population of cells
with vacuolated cytoplasm resembling the physaliferous
cells found in chordomas. Parachordoma and chordoma
differ further in their expression of immunophenotypes.
Both tumors show immunohistochemical evidence of
positivity with CAM5.2 (which recognizes cytokeratins
8/18), EMA, vimentin, and S-100. However, reactivity to
Type IV collagen is much stronger in parachordoma,
whereas chordoma expresses CK1/10 and CK19 most of
123
1254 Clabeaux et al. Clinical Orthopaedics and Related Research

Table 1. Clinical findings of parachordoma case reports

Authors (Year) Age (years) Gender Site Recurrence Metastasis Fatal

Dabska [4] (1977) 47 Female Ankle 7 years No No

52 Male Knee No No No
18 Male Thigh No No No
12 Male Arm No No No
29 Female Back 12 years No No
50 Female Groin No No No
22 Female Forearm No No No
25 Male Femur Unknown Unknown No
28 Female Finger 2 years No No
62 Male Ilium No No No
Povysil and Matejovsky [19] (1985) Unknown Unknown Tibia Unknown Unknown No
Enzinger and Weiss [5] (1988) 53 Male Back Unknown Unknown No
Miettinen et al. [17] (1992) 67 Female Popliteal No Metastasis Fatal
Ishida et al. [12] (1993) 19 Female Calf 1 year No No
Shin et al. [22] (1994) 4 Male Unknown Unknown Unknown No
8 Male Unknown Unknown Unknown No
12 Male Unknown Unknown Unknown No
13 Male Knee Unknown Unknown No
15 Male Unknown Unknown Unknown No
20 Female Calf Unknown Unknown No
Hirokawa et al. [10] (1994) 43 Male Buttock No No No
Sangueza and White [20] (1994) 25 Male Finger Unknown Unknown No
Carstens [3] (1995) 42 Male Buttock 6 months Metastasis Fatal
Niezabitowski et al. [18] (1995) 45 Female Palm 3 months No No
Carlen [2] (1996) 45 Female Leg Unknown Unknown No
Karabela et al. [13] (1996) 86 Female Pelvis No No No
Fisher and Miettinen [7] (1997) 23 Male Thigh Unknown Unknown No
25 Female Triceps No No No
53 Male Back No No No
14 Male Wrist Unknown Unknown No
Limon et al. [16] (1998) 52 Female Chest 2 years Metastasis Fatal
Imlay et al. [11] (1998) 13 Male Thigh No No No
25 Female Forearm Yes No No
Tihy et al. [23] (1998) 7 Female Nares No No No
Folpe et al. [8] (1999) 42 Female Forearm No No No
7 Female Buttock Unknown Unknown No
45 Male Chest No No No
38 Male Deltoid No No No
30 Female Thigh No No No
62 Male Thigh No No No
Koh et al. [15] (2000) 64 Male Tibia Recurrence No No
Separovic et al. [21] (2001) 20 Female Hand No No No
Hemalatha et al. [9] (2003) 24 Male Knee No No No
Abe et al. [1] (2003) 68 Male Calf No Metastasis Fatal
Kinoshita and Yasoshima [14] (2007) 60 Male Calf Unknown Metastasis Fatal

the time. In contrast to chordoma, the matrix of para- EWSR1 gene break-apart rearrangement is positive in
chordoma is abolished by hyaluronidase predigestion. extraskeletal myxoid chondrosarcoma but not in
Additionally, fluorescence in situ hybridization for the parachordoma.

123
Volume 466, Number 5, May 2008
The existing English literature reveals 45 cases of
parachordomas [1–23]. Two of these previously reported
cases have somewhat ambiguous pathology that may be
more consistent with chordoma, rather than parachordoma
[3, 17]. A review of the studies reveals parachordoma has a
slight male predilection, with 25 of 44 (56.8%) docu-
mented cases occurring in males and 19 of 44 (43.2%)
cases occurring in females (Table 1). The average age of
the patients is 34.4 years (range, 4–86 years). Of the case
reports that mentioned location, the lower extremity was
the most common location, with 21 of 41 (51.2%) cases
occurring there. Eleven cases occurred in the upper
extremity (26.8%). Eight (19.5%) cases occurred in the
thorax, trunk, or pelvis. One case occurred in the nares
(2.5%).
Parachordoma is considered a slow-growing tumor of an
indolent, less aggressive nature than chordoma. There has
been reported late recurrence in nine of 45 cases anywhere
from 3 months to 12 years later [7, 14, 15]. However, it is
difficult to extrapolate a recurrence rate from these num-
bers because many patients in the case reports were not
followed for a sufficiently long time. Furthermore, most
studies do not mention if adequate margins were obtained.
In our case, intraoperative frozen section and final patho-
logic review showed negative margins. The possibility of a
late recurrence of disease is something that should be
considered when following a patient with parachordoma.
There have been five reported cases of metastases from
parachordoma. The first report by Miettinen et al. [17]
describes a ‘‘chordoma-like’’ sarcoma in a 67-year-old
woman with metastatic disease to the lung who died of the
disease 11 months after presentation. However, the path-
ologic features were more consistent with chordoma than
parachordoma. Carstens [3] described a metastatic case
arising in the buttocks with widespread disease, with death
resulting 14 months later after local recurrence. However,
again, the pathology was more like chordoma than para-
chordoma. These two studies were included previously in
the parachordoma literature, but, given the inconsistent
nature of the pathology, we believe they should be
discounted.
Limon et al. [16] illustrated a case of lymph node
metastases from a chest wall parachordoma. The metastatic
cells were described as being more cellular and pleomor-
phic. The fourth metastatic case by Abe et al. [1] reported
multiple metastases in a 68-year-old man with a confirmed
parachordoma originating in the calf. The patient eventu-
ally died 32 months after surgery from disease despite
below-knee amputation, radiotherapy, and chemotherapy.
The most recent report of metastatic disease is from
Kinoshita and Yasoshima [14]. The patient was a 60-year-
old man with primary parachordoma in his left calf who
died 4 months after surgery from lung and brain metastases.
Parachordoma of Soft Tissues of the Arm 1255
The case presented here, a 60-year-old woman with a
parachordoma of the right humerus, adds to the literature
involving this rare tumor. Parachordoma generally is con-
sidered a benign tumor with few metastases. However,
review of the literature shows there have been at least two
confirmed deaths from metastatic parachordoma. Possibly,
parachordoma should be thought of as a potentially
aggressive low-grade sarcoma. Given this, we performed
wide excision of the mass and will continue to follow the
patient with routine imaging studies as we would for a
patient with a soft tissue sarcoma. To clarify the behavior
of this rare tumor, long-term followup studies are needed.
Acknowledgments We thank Julie Davila for assistance with
preparation of this manuscript.
References
1. Abe S, Imamura T, Harasawa A, Ishida T, Unno K, Tateishi A,
Tokizaki T, Yorikawa J, Matsushita T. Parachordoma with
multiple metastases. J Comput Assist Tomogr. 2003;27:634–638.
2. Carlen B. Diagnostic value of electron microscopy in rare
malignant musculoskeletal tumors. Ups J Med Sci. 1996;101:
69–85.
3. Carstens PH. Chordoid tumor: a light, electron microscopic,
and immunohistochemical study. Ultrastruct Pathol. 1995;19:
291–295.
4. Dabska M. Parachordoma: a new clinicopathologic entity.
Cancer. 1977;40:1586–1592.
5. Enzinger FM, Weiss SW. Extraskeletal myxoid chondrosarcoma.
In: Gay SM, ed. Soft Tissue Tumors. St Louis, MO: CV Mosby;
1988:868.
6. Fisher C. Parachordoma exists—but what is it? Adv Anat Pathol.
2000;7:141–148.
7. Fisher C, Miettinen M. Parachordoma: a clinicopathologic and
immunohistochemical study of four cases of an unusual soft tis-
sue neoplasm. Ann Diagn Pathol. 1997;1:3–10.
8. Folpe AL, Agoff SN, Willis J, Weiss SW. Parachordoma is
immunohistochemically and cytogenetically distinct from axial
chordoma and extraskeletal myxoid chondrosarcoma. Am J Surg
Pathol. 1999;23:1059–1067.
9. Hemalatha AL, Srinivasa MR, Parshwanath HA. Parachordoma
of tibia: a case report. Indian J Pathol Microbiol. 2003;46:
454–455.
10. Hirokawa M, Manabe T, Sugihara K. Parachordoma of the but-
tock: an immunohistochemical case study and review. Jpn J Clin
Oncol. 1994;24:336–339.
11. Imlay SP, Argenyi ZB, Stone MS, McCollough ML, Henghold
WB. Cutaneous parachordoma: a light microscopic and immu-
nohistochemical report of two cases and review of the literature.
J Cutan Pathol. 1998;25:279–284.
12. Ishida T, Oda H, Oka T, Imamura T, Machinami R. Parachord-
oma: an ultrastructural and immunohistochemical study.
Virchows Arch A Pathol Anat Histopathol. 1993;422:239–245.
13. Karabela-Bouropoulou V, Skourtas C, Liapi-Avgeri G, Mahaira
H. Parachordoma: a case report of a very rare soft tissue tumor.
Pathol Res Pract. 1996;192:972–978; discussion 979–981.
14. Kinoshita G, Yasoshima H. Case report: fatal parachordoma.
J Orthop Sci. 2007;12:101–106.
15. Koh JS, Chung JH, Lee SY, Cho KJ. Parachordoma of the tibia:
report of a rare case. Pathol Res Pract. 2000;196:269–273.
123
1256 Clabeaux et al.
16. Limon J, Babinska M, Denis A, Rys J, Niezabitowski A. Para-
chordoma: a rare sarcoma with clonal chromosomal changes.
Cancer Genet Cytogenet. 1998;102:78–80.
17. Miettinen M, Gannon FH, Lackman R. Chordomalike soft tissue
sarcoma in the leg: a light and electron microscopic and immu-
nohistochemical study. Ultrastruct Pathol. 1992;16:577–586.
18. Niezabitowski A, Limon J, Wasilewska A. Parachordoma: a
clinicopathologic, immunohistochemical, electron microscopic,
flow cytometric, and cytogenetic study. Gen Diagn Pathol.
1995;141:49–55.
19. Povysil C, Matejovsky Z. A comparative ultrastructural study of
chondrosarcoma, chordoid sarcoma, chordoma, and chordoma
periphericum. Pathol Res Pract. 1985;179:546–559.
123
Clinical Orthopaedics and Related Research
20. Sangueza OP, White CR. Parachordoma. Am J Dermatopathol.
1994;16:185–188.
21. Separović R, Glumbić I, Pigac B, Separović V, Kruslin B.
Parachordoma: a case report. Tumori. 2001;87:207–210.
22. Shin HJ, Mackay B, Ichinose H, Ayala AG, Romsdahl MM.
Parachordoma. Ultrastruct Pathol. 1994;18:249–256.
23. Tihy F, Scott P, Russo P, Champagne M, Tabet JC, Lemieux N.
Cytogenetic analysis of a parachordoma. Cancer Genet Cytoge-
net. 1998;105:14–19.
Clin Orthop Relat Res (2008) 466:1257–1261
DOI 10.1007/s11999-008-0143-5
CASE REPORT
Case Report
Nonoperative Treatment of an Unstable Jefferson Fracture Using a Cervical Collar
Brian M. Haus MD, Mitchel B. Harris MD
Received: 22 June 2007 / Accepted: 17 January 2008 / Published online: 8 February 2008
Ó The Association of Bone and Joint Surgeons 2008
Abstract The treatment of unstable burst fractures of the
atlas (Jefferson fractures) is controversial. Unstable Jef-
ferson fractures have been managed successfully with
either immobilization, typically halo traction or halo vest,
or surgery. We report a patient with an unstable Jefferson
fracture treated nonoperatively with a cervical collar, fre-
quent clinical examinations, and flexion-extension
radiographs. Twelve months after treatment, the patient
achieved painless union of his fracture. The successful
treatment confirms prior studies reporting unstable Jeffer-
son fractures have been treated nonoperatively. The
outcome challenges the clinical relevance of treatment
algorithms that rely on the ‘‘rules of Spence’’ to guide
treatment of unstable Jefferson fractures and illustrates
instability may not necessarily be present in patients with
considerable lateral mass widening. Additionally, it
emphasizes a more reliable way of assessing C1–C2 sta-
bility in unstable Jefferson fractures is by measuring the
presence and extent of anterior subluxation on lateral
flexion and extension views.
Each author certifies that he has no commercial associations (eg,
consultancies, stock ownership, equity interest, patent/licensing
arrangements, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his institution has approved the reporting of
this case report, that all investigations were conducted in conformity
with ethical principles of research, and that a waiver of informed
consent was obtained.
B. M. Haus (&), M. B. Harris
Department of Orthopaedic Surgery, Brigham and Women’s
Hospital, 75 Francis Street, Boston, MA 02115, USA
e-mail: bhaus@partners.org
Introduction
Burst fractures of the atlas (Jefferson fractures) occur when
an axial force is transmitted across the occipital-cervical
junction, causing the atlas to be compressed between the
angulated articular surface of the axis and the occipital
condyles. The impact forces cause an outward spread of the
lateral masses of C1. The resulting four-part atlas fracture,
with two in the posterior arch and two in the anterior arch,
is classically referred to as the Jefferson fracture [20]. Most
are relatively stable and not associated with neurologic
deficits and can be treated by external immobilization with
satisfactory outcomes. Unstable Jefferson fractures reflect a
more severe injury of the atlas that occur when the trans-
verse ligament is also ruptured secondary to the extent of
spread of the C1 arch. These fractures are more difficult to
treat because of the atlantoaxial instability. Many surgeons
recommend operative stabilization of these unstable Jef-
ferson fractures.
We present a patient with an unstable Jefferson fracture
who was successfully treated nonoperatively with a cervi-
cal collar. At 12 months’ followup, he had achieved
painless healing of C1 with ankylosis of C1–C2 with
resumption of full premorbid activities.
Case Report
A 62-year-old man slipped and struck the back of his neck.
He denied loss of consciousness and reported only neck
pain. The day after the injury, he noted a change in his
voice, transient chest tingling, and the inability to lift his
head off his chest.
He delayed medical evaluation until 5 days later when
he presented to his chiropractor. The patient’s chiropractor
123
1258 Haus and Harris Clinical Orthopaedics and Related Research

obtained radiographs of the patient’s neck (Fig. 1A) that
were interpreted as normal. The chiropractor then per-
formed a manipulation on his cervical spine.
Postmanipulation, the patient reported nausea and his wife
described his color as ‘‘ashen’’.
He returned to his chiropractor the next day. Before
performing another manipulation, the chiropractor obtained
another open-mouth view of the patient’s cervical spine
(Fig. 1B). The chiropractor identified the C1 fracture, and
the patient ultimately was transferred to our hospital for
evaluation.
On examination, he was hesitant to flex and extend his
neck. His neck was nontender and he was neurologically
intact. Plain films showed a 14-mm overlap of the C1-C2
lateral masses, confirming the diagnosis of an unstable
Jefferson fracture (Fig. 1B). Coronal views confirmed
separation of the lateral masses (Fig. 2A) and axial views
showed a bony avulsion suggesting transverse ligament
rupture (Fig. 2B). Because he had been walking without a
cervical orthosis before his presentation, active flexion and
extension films were attempted. The films showed no
motion between C1 and C2 with an atlantodens interval of
3 mm. There was no movement identified between the
occiput and the posterior arch of C1.
The radiographs and computed tomographic scans
were reviewed with the patient. The unstable nature of
his fracture was described and he was offered immobili-
zation in a halo vest or surgical stabilization as treatment
options. The patient was opposed to both options. There-
fore, he was given a Miami J Cervical CollarTM (Jerome
Medical, Moorestown, NJ) and observed with monthly
clinical examinations, quarterly computed tomographic
scans, and flexion-extension views at 6 months and 1 year
postinjury.

Fig. 1A–B (A) An inadequate

open-mouth anteroposterior view
was taken 5 days after the injury.
(B) A repeat open-mouth anter-
oposterior view taken 6 days
after the injury shows 14 mm of
bilateral offset of the lateral
masses, indicating an unstable
Jefferson fracture.

Fig. 2A–B (A) A computed

tomographic scan in the coronal
plane reveals 12-mm lateral mass
separation, confirming the diag-
nosis of an unstable Jefferson
fracture. (B) A computed tomo-
graphic scan in the plane of C1
shows the four fractures of the
ring, two anterior and two pos-
terior. Avulsion of the transverse
ligament is evident on the right.

123
Volume 466, Number 5, May 2008 Management of Unstable Jefferson Fractures 1259

At 3 months’ followup, the patient was encouraged to
mobilize his neck as tolerated and was gradually weaned
from wearing the collar. Examination revealed he was
able to flex his neck within three fingerbreadths of his
chest with extension to neutral only. He laterally rotated
his neck 20° bilaterally. His neck was stiff. Motor testing
remained normal. Computed tomographic scans showed
evidence of healing of the posterior arch on the left
with 11-mm overlap of the C1-C2 lateral masses.
Flexion-extension films showed no motion across the
occipitocervical junction and no increase in the atlanto-
dens interval.
At his 1-year followup, the patient had discontinued
wearing the collar and continued to be pain-free. Exami-
nation revealed he had extension to neutral only, 30° lateral
neck rotation bilaterally, and flexion to within two finger-
breadths of his chest. A lateral flexion-extension
radiograph and computed tomographic scan showed com-
plete healing of C1, ankylosis at C1-C2, and no evidence of
C1-C2 subluxation (Figs. 3, 4).

Fig. 3A–B Computed tomo-

graphic scans in the (A) axial
and (B) coronal planes show
progression of healing of the C1
Jefferson fracture with partial
bony fusion along the left ante-
rior, right anterior, and left
posterior rings. There is a mild
shift of C1 to the left and mild
deformation of the atlas ring.
There is evidence of transverse
ligament calcification.

Fig. 4A–B Lateral (A) flexion

and (B) extension views of the
cervical spine taken at the 1-year
followup reveal a healed poster-
ior arch of the C1 fracture. The
atlantodens interval is preserved,
confirming stability at C1-C2 on
flexion and extension. A minimal
degree of flexion-extension is
evident by the interval between
the skull and the posterior arch of
C1.

123
1260 Haus and Harris
Discussion
Stability at C1-C2 is maintained primarily by the transverse
ligament. With considerable axial loading, separation of
the lateral masses can occur, and the transverse ligament
can be torn. Based on the clinical and radiographic
assessments of Spence et al. [19] in cadavers, transverse
ligament rupture occurs with lateral mass separation less
than 6.9 mm. Jefferson fractures with torn transverse lig-
aments are inherently unstable and referred to as unstable
Jefferson fractures. Instability at the atlantoaxial joint and
subsequent subluxation can occur with unprotected flexion
and extension movements. The degree of subluxation is
observed on flexion-extension views by a change in the
atlantodens interval; an interval greater than 2 to 3 mm in
adults is deemed abnormal.
Although there is agreement regarding treatment of
stable Jefferson fractures [6, 10, 11, 21], management of
unstable Jefferson fractures remains controversial. Suc-
cessful treatment of unstable Jefferson fractures should
preserve rotation at the atlantoaxial joint and maintain
alignment to allow healing of the atlas. Unstable Jefferson
fractures have been treated successfully with immobiliza-
tion with halo traction or vest and surgery. Treatment with
a cervical collar is not the standard of treatment because it
is presumed a collar would not sufficiently limit motion to
prevent subluxation or allow healing.
Many authors argue most unstable Jefferson fractures
heal satisfactorily with immobilization in halo traction or
vest without surgery [5, 6, 9, 12, 13, 21]. Advocates of
treatment with a halo vest argue it provides traction to
align the splayed lateral masses through ligamentotaxis
and reduces any impaction below C1-C2, thereby pre-
venting subluxation and promoting healing. Critics of
nonoperative immobilization cite high rates of nonunion
and persistent posttraumatic pain as reasons to instead
perform surgical stabilization [4, 20]. Several authors
recommend initial immobilization for 8 weeks in a halo
vest followed by C1-C2 fusion to stabilize the joint once
the residual instability is documented through flexion-
extension studies [16, 18]. Hein et al. [7] advocate
immediate surgical stabilization, warning that repair of
the dislocation after long-term immobilization can cause
irreversible incongruence of the atlantoaxial joint fol-
lowed by arthrosis and increasing neck pain. Kesterson
et al. [8] recommend primary occipitocervical stabiliza-
tion of Jefferson fractures; however, others recommend
transfacet screw fixation at C1-C2 to maintain motion
between the occiput and C1. Ruf et al. [17] preserved C1-
C2 rotation, obtained anatomic reconstruction of the atlas,
and achieved bony fusion using transoral reduction and
osteosynthesis of C1. McGuire and Harkey [15]
123
Clinical Orthopaedics and Related Research
recommend primary stabilization using the Magerl trans-
facet screw technique [14].
Our case report underscores many important points
regarding the treatment of unstable Jefferson fractures.
First, the rules of Spence [19] may not correlate directly to
C1-C2 stability. Historic treatment algorithms have relied
on the radiographic distinction to establish stability at the
atlantoaxial joint and to guide treatment. However, rupture
of the transverse ligament does not definitively imply C1-
C2 instability exists. As Dickman [1] discussed, the most
reliable way of assessing C1-C2 stability in unstable Jef-
ferson fractures is by measuring the extent of anterior
subluxation on lateral flexion and extension views. The
patient in our case report had considerable lateral mass
widening (14 mm) with evidence of an avulsion of the
transverse ligament. However, on serial flexion-extension
radiographs, he showed only 3 mm of atlantoaxial sepa-
ration (Figs. 1, 2). In this patient, it is likely the residual
stability is provided by intact alar ligaments and/or portions
of facet capsules, and scarring of the avulsed transverse
ligament. Protective immobilization with a cervical collar
produced adequate long-term stability to enable bony
fusion of the atlas fracture.
This case report also supports prior studies that
emphasized the importance of examining the type of
transverse ligament fracture when making the decision to
pursue surgical intervention. Dickman et al. [2, 3] classi-
fied transverse ligament injuries into two main types, which
have two distinctly different treatment outcomes. Type I
injuries are intrasubstance ruptures of the transverse liga-
ment that are incapable of healing and because of
instability are believed by most to require surgery. Type II
injuries are avulsions of the transverse ligament on the C1
lateral mass that can heal if the fracture is immobilized
nonoperatively. Although the study was limited to isolated
injuries of the transverse ligament, its principles can be
applied to treatment of Jefferson fractures with transverse
ligament disruption. Our case report suggests avulsion
fractures of the transverse ligament can be sufficiently
stabilized by a cervical collar to allow union of the atlas
fracture and healing of the avulsion (Figs. 3, 4).
Although our single case obviously cannot justify
treatment of all unstable Jefferson fractures with a cervical
collar, it does show instability may not necessarily be
present in patients with considerable lateral mass widening.
In our patient, either the cervical collar limited his motion
sufficiently in flexion-extension to allow scarring of the
ligament or the spread of the lateral masses of C1 did not
alter the integrity of the transverse ligament. In patients
with unstable Jefferson fractures who are opposed to sur-
gery, we believe it is important to observe clinical
instability before recommending surgical stabilization.
Volume 466, Number 5, May 2008
References
1. Dickman CA. Letters. Spine. 2004;29:2196.
2. Dickman CA, Greene KA, Sonntag VK. Injuries involving the
transverse atlantal ligament: classification and treatment guide-
lines based upon experience with 39 injuries. Neurosurgery.
1996;38:44–50.
3. Dickman CA, Hadley MN, Browner C, Sonntag VK. Neurosur-
gical management of acute atlas-axis fractures: a review of 25
cases. J Neurosurg. 1989;70:45–49.
4. Guiot B, Fessler RG. Complex atlantoaxial fractures. J Neuro-
surg. 1999;91:139–143.
5. Hadley MN, Dickman CA, Browner CM, Sonntag VK. Acute
traumatic atlas fractures: management and long-term outcome.
Neurosurgery. 1988;23:31–35.
6. Han SY, Witten DM, Mussleman JP. Jefferson fractures of the
atlas: report of six cases. J Neurosurg. 1976;44:368–371.
7. Hein C, Richter HP, Rath SA. Atlantoaxial screw fixation for the
treatment of isolated and combined unstable Jefferson fractures:
experiences with 8 patients. Acta Neurochir (Wein). 2002;144:
1187–1192.
8. Kesterson L, Benzel E, Orrison W, Coleman J. Evaluation and
treatment of atlas burst fractures (Jefferson fractures). J Neuro-
surg. 1991;75:213–220.
9. Koch RA, Nickel VL. The halo vest: an evaluation of motion and
forces across the neck. Spine. 1978;3:103–107.
10. Landellis CD, Van Peteghem PK. Fractures of the atlas: classi-
fication, treatment and morbidity. Spine. 1988;13:450–452.
11. Lee TT, Green BA, Petrin DR. Treatment of stable burst fracture
of the atlas (Jefferson fracture) with rigid cervical collar. Spine.
1998;23:1963–1967.
Management of Unstable Jefferson Fractures 1261
12. Levine AM, Edwards CC. Treatment of injuries in the C1-C2
complex. Orthop Clin North Am. 1986;17:31–44.
13. Levine AM, Edwards CC. Fractures of the atlas. J Bone Joint
Surg Am. 1991;73:680–691.
14. Magerl F, Seemann PS. Stable posterior fusion of the atlas and
axis by transarticular screw fixation. In: Weidner PA, ed. Cer-
vical Spine. New York, NY: Springer-Verlag; 1987:322–327.
15. McGuire RA Jr, Harkey HL. Primary treatment of unstable Jef-
ferson’s fractures. J Spinal Disord. 1995;8:233–236.
16. O’Brien JJ, Butterfield WL, Gossling HR. Jefferson fracture with
disruption of the transverse ligament: a case report. Clin Orthop
Relat Res. 1977;126:135–138.
17. Ruf M, Melcher R, Harms J. Transoral reduction and osteosyn-
thesis C1 as a function-preserving option in the treatment of
unstable Jefferson fractures. Spine. 2004;29:823–827.
18. Schlicke LH, Callahan RA. A rational approach to burst fractures
of the atlas. Clin Orthop Relat Res. 1981;154:18–21.
19. Spence KF Jr, Decker S, Sell KW. Bursting atlantal fracture
associated with rupture of the transverse ligament. J Bone Joint
Surg Am. 1970;52:543–549.
20. Vieweg U, Meyer B, Schramm J. Differential treatment in acute
upper cervical spine injuries: a critical review of a single-insti-
tution series. Surg Neurol. 2000;54:203–210; discussion 201–
211.
21. Zimmerman E, Grant J, Vise WM, Yashon D, Hunt WE. Treat-
ment of Jefferson fracture with a halo apparatus: report of two
cases. J Neurosurg. 1976;44:372–375.
123
Clin Orthop Relat Res (2008) 466:1262–1266
DOI 10.1007/s11999-008-0135-5
ORTHOPAEDIC
RADIOLOGY
PATHOLOGY CONFERENCE
Knee Lesion in a 62-year-old Woman
Aditya V. Maheshwari MD, Carlos A. Muro-Cacho MD,
H. Thomas Temple MD
Published online: 8 February 2008
Ó The Association of Bone and Joint Surgeons 2008
History and Physical Examination
A 62-year-old woman presented for evaluation of a knee
contusion sustained approximately 2 weeks previously.
Her pain had decreased considerably since the injury, and
at the time of presentation, she had neither pain nor
swelling. Several years earlier, she had sustained a trau-
matic fracture of the right pubic and ischial rami, which
resulted in intermittent radicular pain to her right leg. She
denied any constitutional symptoms and her medical his-
tory was noncontributory.
The general physical examination was unremarkable.
Point tenderness with no discrete palpable mass was
noticed on the posterior aspect of her knee, adjacent to the
Each author certifies that he has no commercial associations (eg,
consultancies, stock ownership, equity interest, patent/licensing
arrangement, etc) that might pose a conflict of interest in connection
with the submitted article.
Each author certifies that his institution has approved the reporting of
this case report, that all investigations were conducted in conformity
with ethical principles of research, and that informed consent was
obtained.
A. V. Maheshwari, H. T. Temple
Division of Musculoskeletal Oncology, Department of
Orthopedics, University of Miami Miller School of Medicine,
Miami, FL, USA
A. V. Maheshwari (&)
Cedars Medical Center, 1400 NW 12th Avenue, Suite 4036,
4th floor, East Building, Miami, FL 33136, USA
e-mail: adi_maheshwari@rediffmail.com;
adityavikramm@gmail.com
C. A. Muro-Cacho
Division of Musculoskeletal Oncology, Department of
Pathology, University of Miami Miller School of Medicine,
Miami, FL, USA
123
medial femoral epicondyle. Neither warmth nor erythema
was detected, and there was no palpable inguinal or pop-
liteal lymphadenopathy. Examination of the knee showed
no effusion. There was full painless range of motion with
no ligamentous instability or joint line tenderness. The rest
of her musculoskeletal examination was normal. The lab-
oratory workup was unremarkable.
Because of the concerning radiographic features (Fig. 1),
computed tomography (CT) scans (Fig. 2) and magnetic
resonance (MR) images (Figs. 3, 4) were obtained.
Based on clinical history, physical examination, labo-
ratory tests, and imaging studies, what is the differential
diagnosis?
Imaging Interpretation
Anteroposterior (Fig. 1A) and lateral (Fig. 1B) radiographs
showed an area of cortical irregularity at the posterior
aspect of the medial femoral condyle, associated with a
partially mineralized soft tissue mass. On CT (Fig. 2), the
soft tissue mass measured 3.0 9 3.0 9 2.5 cm and was
contiguous with the posteromedial cortex at the origin of
the medial gastrocnemius head. It had an attenuation sim-
ilar to that of adjacent skeletal muscle, with scattered areas
of mineralization. On MRI, the lesion had hyperintense
signal on gradient echo (Fig. 3A) and intermediate signal
on proton density-weighted (Fig. 3B) sagittal images. On
axial images, the somewhat lobular mass was directly
contiguous with the femur but did not invade it, with
normal underlying marrow signal (Fig. 3C). The mass
demonstrated heterogeneous enhancement on postgadolin-
ium images (Fig. 4). Degenerative changes in the medial
meniscus, a popliteal cyst, and a joint effusion were also
noted.
Volume 466, Number 5, May 2008
Fig. 1A–B (A) Anteroposterior and (B) lateral radiographs show an
area of cortical irregularity at the posterior aspect of medial femoral
condyle, associated with a partially mineralized soft tissue mass
(arrows).
Fig. 2 A CT scan shows a 3.0- 9 3.0- 9 2.5-cm soft tissue mass
(arrow) that is contiguous with the posteromedial cortex at the origin
of the medial head of the gastrocnemius. The attenuation coefficient
of the lesion is similar to adjacent skeletal muscle, with scattered
areas of mineralization and no organization (zonation).
Differential Diagnosis
Parosteal/periosteal (surface) osteosarcoma
Other surface tumors, including juxtacortical chondro-
sarcoma, osteochondroma, and periosteal chondroma
Soft tissue sarcoma
Myositis ossificans
Distal femoral cortical irregularity
Florid reactive periostitis and bizarre parosteal osteo-
chondromatous proliferation (Nora’s lesion)
The patient underwent a CT-guided needle biopsy,
which showed a cytologically bland fibrous proliferation.
Knee Lesion in a 62-year-old Woman 1263
To mitigate potential sampling error, an open biopsy was
performed and the histology of the lesion was studied
(Fig. 5).
Based on the clinical history, physical examination, lab-
oratory studies, radiographic images, and histologic picture,
what is the diagnosis and how should this lesion be treated?
Histology Interpretation
The material consisted of multiple, tan-grey, soft tissue
fragments measuring 8.0 9 6.0 9 3.0 mm in aggregate.
Microscopically, there was a hypocellular, fibroblastic
proliferation associated with fragments of tendon, fibro-
cartilage, and focal myxoid areas of dissecting collagen
fibers. There was no cellular atypia or mitoses (Fig. 5).
Diagnosis
Distal femoral cortical irregularity.
Discussion and Treatment
The diagnosis of distal femoral cortical irregularity was
based on its location (distal medial femur) and supporting
histologic features, despite the worrisome atypical radio-
graphic features. The diagnosis was further supported by
the lack of additional progression after 7 years’ followup.
Usually, distal femoral cortical irregularity differs from
malignant tumors by the absence of a soft tissue mass, the
preservation of soft tissue planes, the absence of periosteal
reaction, and the lack of associated warmth or local ten-
derness on physical examination [21]. Histologically, the
absence of mitotic activity and pleomorphism will help
distinguish it from sarcomas. Sarcomas of bone exhibit
greater cellularity with plump hyperchromatic nuclei,
nuclear and cellular pleomorphism, and increased mitotic
activity. These differences, together with knowledge of
radiographic variation, should make the diagnosis appar-
ent. Serial radiographs can be very beneficial. Fibrous
cortical defect is characteristically present near the meta-
diaphysis of long bones. It is eccentric in bone, has a
geographic ‘‘soap bubble’’ appearance, and erodes the
cortex from within, whereas the avulsive cortical irregu-
larity erodes the external surface of the bone. A fibrous
cortical defect is generally an isolated osseous lesion
without a soft tissue component. It tends to migrate more
proximally with growth [25]. Histologically, the two enti-
ties are distinct, as the fibrous cortical defect has a whirling
spindled pattern of growth with giant cells and hemosid-
erin-rich macrophages. Osteochondroma has a typical
123
1264 Maheshwari et al.
Fig. 3A–C MRI shows the lesion (arrows) having (A) hyperintense
signal on gradient echo (TR 800, TE 15) and (B) intermediate signal on
proton density-weighted (TR 2442, TE 15) sagittal images. (C) An
axial image (TR 5000, TE 21.7) shows the mass has a lobular
Fig. 4 A postgadolinium contrast axial MR image (TR 550, TE 9)
shows heterogeneous enhancement of the mass (arrow), with no
enhancement of adjoining bone or soft tissue.
cartilage cap with enchondral ossification. The lack of
cortical and medullary continuity between the mass and the
adjacent bone excludes the diagnosis of osteochondroma.
Florid reactive periostitis commonly affects the digits of
the hands and feet of adolescents and young adults. Clin-
ically, there is a history of gradually progressive swelling,
123
Clinical Orthopaedics and Related Research
morphology and is contiguous with the femur but does not appear to
invade it, with normal underlying marrow signal. A blooming artifact
(predominant low signal) can be seen on (A) and (B) (arrowheads) in an
area corresponding to an area of mineralization on radiographs and CT.
erythema, and pain or a painful mass in the affected part
and approximately 50% of patients have a history of
trauma. There is mature and immature osteoid and woven
bone, cartilage, or a mix of chondroid, osteoid, and myxoid
elements with frequent multinucleated giant cells. They
show intense uptake on bone scan and may also show zonal
maturation of the tissues similar to that observed in myo-
sitis ossificans. It is thought to progress to bizarre parosteal
osteochondromatous proliferation, an entity also common
in hands and feet. Bizarre parosteal osteochondromatous
proliferation has a cartilaginous cap covering a bone stalk
with areas of ossification attached to the cortex by a broad
base; cortical erosion or corticomedullary continuity is not
present. Periosteal chondromas typically show external
cortical saucerization and a thick periosteal reaction (but-
tressing) with or without mineralized chondroid matrix.
Since its first description by Kimmelstiel and Rapp [13]
in 1951 as ‘‘periosteal desmoid,’’ this lesion has also been
called cortical desmoid, avulsive cortical irregularity,
subperiosteal dermoid, subperiosteal abrasion, cortical
abrasion, subperiosteal cortical defect, parosteal or juxta-
cortical desmoid, medial distal metaphyseal femoral
irregularity, or fibrous metaphyseal defect [1–4, 8, 9, 11–
15, 20–27]. This is a benign entity that may have an
atypical and aggressive appearance or a focal geographic
Volume 466, Number 5, May 2008
Fig. 5A–B Photomicrographs show (A) hypocellular fibroblastic
proliferation with no cellular atypia or mitosis and (B) focal myxoid
areas of dissecting collagen fibers (Stain, hematoxylin and eosin;
original magnification, 9200).
radiolucency within the posterior cortex of medial femoral
condyle. It typically occurs between the ages of 3 to
17 years, with a peak incidence at 10 to 15 years, and has
been reported in 11.5% of boys and 3.6% of girls [21]. At
epiphyseal closure, the irregularity decreases in size and
often disappears, although it may persist into adulthood [7,
9, 13, 20, 27], as documented in one report in a 57-year-old
man [17]. Suh et al. [23] also found distal femoral cortical
irregularities in 44 of 100 knee MRIs (mostly adults) and
divided them into concave (four cases with cortical con-
cavity), convex (36 cases with cortical convexities), and
divergent (four cases with wide and split cortex). The
lesion is often asymptomatic, producing no palpable mass,
pain, or swelling and has a benign course [8, 9, 16, 19].
Consequently, they are frequently an incidental finding
when radiographs are reviewed for unrelated conditions or
knee trauma as in our patient [7, 9].
The etiology is uncertain and controversial. Some [21]
consider it a developmental anomaly while others consider
Knee Lesion in a 62-year-old Woman 1265
it an earlier phase of the more common fibrocortical defect
[4, 7]. Most lesions can be related to chronic traction or
avulsive injury at the insertion of the extensor aspect of the
adductor magus [2, 5, 9, 14, 20] or to the origin of the
medial head of the gastrocnemius [7, 17, 20, 23, 26] and
plantaris muscles [7]. Some authors, however, have not
found tendinous attachments at this location [4, 21, 27]. In
addition to the strong muscle pull exerted in this area,
intense bone remodeling occurs simultaneously during
periods of rapid skeletal growth. The cortex of the bone is
consequently weakened, and the excessive mechanical
stress that occurs at this site is believed to produce mi-
croavulsions of the cortical bone that elicit a hypervascular
and fibroblastic response, which in turn stimulates osteo-
clastic activity and bone resorption [5]. It is postulated this
lesion is a fibroblastic periosteal response resulting in
periosteal new bone formation on the surface and con-
comitant cortical osteoclastic bone resorption. If bone
formation prevails, soft tissue mineralization is evident
radiographically. If the fibroblastic reaction is predomi-
nant, a cortical defect is seen. In contrast, Young et al. [27]
found the lesion indistinguishable from an osteochon-
droma, with a layer of cartilage and an underlying
fibroblastic layer. Resnick and Greenway [20] classified
them as excavations or proliferative cortical irregularities
and found thickened periosteum with fibrous connective
tissue in the proliferative cortical irregularity. Marek [16]
has described the lesion as a ‘‘cork in the bottle,’’ with part
of the mass within the bone and part outside bone. The
morphologic features are reminiscent of a reactive process
rather than a neoplastic one. The wide variation in the
nomenclature of these lesions reflects differences in hist-
opathologic interpretation [8].
Although most common at the posteromedial aspect of
the medial femoral condyle, similar lesions have been
documented in the humerus, tibia, fibula, radius, metatar-
sal, metacarpal, and even distal phalanx [5, 24], all sites of
strong tendon insertions. A proximal humeral lesion at the
insertion of the pectoralis major has been described as a
‘‘ringman’s shoulder,’’ occurring primarily in gymnasts
[10]. The femoral lesion is best demonstrated radiograph-
ically with the knee in 20° to 45° of external rotation [9].
The involved area is typically 1 to 2 cm in length, occa-
sionally with reactive bone formation extending into the
soft tissue [9]. Examination of both knees is recommended
since the lesion is frequently bilateral (up to 35%) [21, 26].
Additional imaging is sometimes required to differentiate
this benign process from malignant bone and soft tissue
tumors and, thus, CT and MRI are important imaging tools
to ascertain anatomic relationship, bone destruction, and
soft tissue involvement [22, 23, 26]. Technetium bone
scintigraphy generally reveals no uptake in the area of
irregularity, although this negative finding may be masked
123
1266 Maheshwari et al.
by the proximity of a growing epiphysis in children, a
degenerative joint in the elderly [6, 24], or concomitant
pathologic conditions such as osteomyelitis and lymphoma
[6]. There is also a report of a stress fracture that simulated
a distal femoral cortical irregularity radiographically and
was differentiated by a bone scan, MRI, and histopatho-
logic findings [18]. Single-photon emission CT imaging of
the knee will show increased uptake [12].
The distal femur is a common site of developmental
anomalies and primary bone tumors [21]. Whether or not a
particular radiograph is interpreted as a normal variant
depends on the clinical situation (patient age and location),
the degree of cortical irregularity, the radiographic pro-
jection, and the experience of the radiologist and the
orthopaedist interpreting the data [9]. Thus, as one gains
familiarity with this variant, there is less likelihood of
performing an unnecessary biopsy [9]. Six of seven cases
(85.7%) in the study of Craigen et al. [8] were diagnosed
primarily as malignant tumors and five (71.4%) underwent
a biopsy. A biopsy may even result in a false-positive
diagnosis of malignancy [24]. Therefore, recognition of
this benign lesion is important, as it could easily be mis-
taken for a malignant change and unnecessary surgery may
be performed. Amputation has been reported because of
confusion with a malignant process [13].
Although a biopsy is not recommended for this lesion,
the patient described herein underwent biopsy twice.
Advanced age, the location at the posteromedial distal
femur, point tenderness, and an associated soft tissue mass
with mineralization, along with cortical irregularity and
mild periosteal reactive changes, were worrisome for
malignancy. Because of the atypical clinical and radio-
graphic features, concerns were expressed about the
adequacy of the CT-guided biopsy material. This diag-
nostic uncertainty and concern over sampling error led to
an open biopsy. This biopsy confirmed the diagnosis of a
distal femoral cortical irregularity, and observation was
recommended, along with the judicious use of antiinflam-
matory medications. The patient remains asymptomatic at
7 years’ followup, with no evidence of disease progression.
References
1. Allen DH. A variation of diaphyseal development which simu-
lates the roentgen appearance of primary neoplasms of bone. Am
J Roentgenol Radium Ther Nucl Med. 1953;69:940–943.
2. Barnes GR Jr, Gwinn JL. Distal irregularities of the femur sim-
ulating malignancy. Am J Roentgenol Radium Ther Nucl Med.
1974;122:180–185.
3. Bernasek TL, Sim FH, Wold LE, Beabout JW, Shives TC, Dunitz SJ.
Avulsive cortical irregularities. Orthopedics. 1987;10:1423–1425.
123
Clinical Orthopaedics and Related Research
4. Brower AC, Culver JE Jr, Keats TE. Histological nature of the
cortical irregularity of the medial posterior distal femoral
metaphysis in children. Radiology. 1971;99:389–392.
5. Bufkin WJ. The avulsive cortical irregularity. Am J Roentgenol
Radium Ther Nucl Med. 1971;112:487–492.
6. Burrows PE, Greenberg ID, Reed MH. The distal femoral defect:
technetium-99 m pyrophosphate bone scan results. J Can Assoc
Radiol. 1982;33:91–93.
7. Caffey J. On fibrous defects in cortical walls of growing tubular
bones: their radiologic appearance, structure, prevalence, natural
course, and diagnostic significance. Adv Pediatr. 1955;7:13–51.
8. Craigen MA, Bennet GC, MacKenzie JR, Reid R. Symptomatic
cortical irregularities of the distal femur simulating malignancy.
J Bone Joint Surg Br. 1994;76:814–817.
9. Dunham WK, Marcus NW, Enneking WF, Haun C. Develop-
mental defects of the distal femoral metaphysis. J Bone Joint
Surg Am. 1980;62:801–806.
10. Fulton MN, Albright JP, El-Khoury GY. Cortical desmoid-like
lesion of the proximal humerus and its occurrence in gymnasts
(ringman’s shoulder lesion). Am J Sports Med. 1979;7:57–61.
11. Hatcher CH. The pathogenesis of localized fibrous lesions in the
metaphyses of long bones. Ann Surg. 1945;122:1016–1030.
12. Jung C, Choi YY, Cho S, Park KC. Symptomatic cortical desmoids
detected on knee SPECT. Clin Nucl Med. 2002;27:437–438.
13. Kimmelstiel P, Rapp IR. Cortical defect due to periosteal
desmoids. Bull Hosp Joint Dis. 1951;12:286–297.
14. Kirkpatrick JA, Wilkinson RH. Case report 52: post-traumatic
fibrous tissue lesion distal end of femur-parosteal (juxtacortical)
desmoid. Skeletal Radiol. 1978;2:189–190.
15. Kreis WR, Hensinger RN. Irregularity of the distal femoral
metaphysis simulating malignancy: case report. J Bone Joint Surg
Am. 1977;59:38.
16. Marek F. Fibrous cortical defect (periosteal desmoid). Bull Hosp
Joint Dis. 1955;16:77–87.
17. Pennes DR, Braunstein EM, Glazer GM. Computed tomography
of cortical desmoid. Skeletal Radiol. 1984;12:40–42.
18. Pistolesi GF, Caudana R, D’Attoma N, Residori E, Pregarz M.
Case report 686: stress fracture at distal end of femur simulating
‘‘periosteal desmoid’’. Skeletal Radiol. 1991;20:454–457.
19. Prentice AI. Variations on the fibrous cortical defect. Clin Radiol.
1974;25:531–533.
20. Resnick D, Greenway G. Distal femoral cortical defects, irregu-
larities, and excavations. Radiology. 1982;143:345–354.
21. Simon H. Medial distal metaphyseal femoral irregularity in
children. Radiology. 1968;90:258–260.
22. Sklar DH, Phillips JJ, Lachman RS. Case report 683: distal
metaphyseal femoral defect (cortical desmoid; distal femoral
cortical irregularity). Skeletal Radiol. 1991;20:394–396.
23. Suh JS, Cho JH, Shin KH, Won JH, Park SJ, Shin DH, Kim SJ,
Lee HY. MR appearance of distal femoral cortical irregularity
(cortical desmoid). J Comput Assist Tomogr. 1996;20:328–332.
24. Velchik MG, Heyman S, Makler PT Jr, Goldstein HA, Alavi A.
Bone scintigraphy: differentiating benign cortical irregularity of
the distal femur from malignancy. J Nucl Med. 1984;25:72–74.
25. Verdonk PC, Verstraete K, Verdonk R. Distal femoral cortical
irregularity in a 13-year old boy: a case report. Acta Orthop Belg.
2003;69:377–381.
26. Yamazaki T, Maruoka S, Takahashi S, Saito H, Takase K, Na-
kamura M, Sakamoto K. MR findings of avulsive cortical
irregularity of the distal femur. Skeletal Radiol. 1995;24:43–46.
27. Young DW, Nogrady MB, Dunbar JS, Wiglesworth FW. Benign
cortical irregularities in the distal femur of children. J Can Assoc
Radiol. 1972;23:107–115.
Clin Orthop Relat Res (2008) 466:1267–1268 / DOI 10.1007/s11999-008-0218-3
Lumbosacral Fusion: The Mortised Transfacet
Method by Use of the Vibrating Electric Saw
for Circular Bone Blocks
Earl D. McBride MD, Howard B. Shorbe MD
R
eliably achieving spine
fusion has been a problem
since fusion was first descri-
bed by Russell Hibbs of the New
York Orthopaedic Hospital in 1911
[1]. Since that time, many dozens, if
not hundreds of techniques have
been proposed. PubMed lists over
1700 articles on the topic since 1935,
many of which relate to techniques,
and these 1700 articles undoubtedly
reflect the tip of the iceberg. The
reasons for performing a spine fusion
have also varied and been greatly
refined over the last century, and the
success rates have varied depending,
in addition to other factors, on both
the reasons for the fusion and the
technique. The article we highlight
this month on lower lumbar spine
fusion was published by Dr. Earl
McBride and his colleague, Dr.
Howard Shorbe, both of Oklahoma
City [5]. Dr. McBride had been the
founder and first President of the
Association of Bone and Joint Sur-
geons (1947) [2, 3] which organized
and has sponsored this journal since
its inception in 1953.
McBride and Shorbe noted, ‘‘a
successful fusion depends on where
and how the bone graft is implanted,
not upon the amount of bone uti-
lized’’ [5]. One of the more common

The Association of Bone and Joint Surgeons 2008
CORR 1958;12:268–275
reasons for surgery at the time was a
ruptured disc, and they commented,
‘‘Removal of the offending disk
protrusion alone is not likely to
relieve the residual effects of pro-
gressive arthritic erosion and
ligamentous weakness. How can
such facts be ignored so com-
pletely?’’ [5] Their proposal was to
fuse the facet joints, which often
were arthritic in advance cases,
using a dowel grafts cut by trephines
on a vibrating saw and inserting
those grafts into similar size cylin-
drical hole across the fact joints.
They commented, ‘‘The nerve will
not be damaged by the saw if it is
directed properly and if correct
depth adjustment has been made’’
[5]. Although mentioned in but a
single sentence, this same point has
been emphasized the past ten years
or so with pedicle screws placed
across the facet joints to achieve the
same purpose: directing an implant
in the proper direction requires great
skill and knowledge.
McBride and Shorbe [5] were
not uniformly successful in achiev-
ing fusion: their rate of successful
lumbar fusion was 91% at the L5-
S1 level. When they operated on
two levels, their rates of success
were much lower: 65% at both
Published online: 21 March 2008
levels. They noted, however, a
successful fusion seen on a radio-
graph did not necessarily correlate
with improvement of the patient:
‘‘…good and fair clinical healing
was obtained in 97 percent of the
total cases’’ [5]. They did not
have the sophisticated instruments
we do today in assessing success,
and undoubtedly a contemporary
assessment would lead to a lower
clinical success rate. Nonetheless,
their finding that successful bone
fusion did not predict clinical out-
come is consistent with a more
recent report with the same finding
in patients who had multiple back
operations and other sorts of
fusion [4]. Thus, their conclusion,
‘‘Satisfactory clinical results of
facet graft fusions have always
exceeded the number of cases
showing solid bony union’’ [5]
likely holds today.
Richard A. Brand MD
Editor-in-Chief
Clinical Orthopaedics
and Related Research
References
1. Bick EM. Source Book of Orthopae-
dics. New York, NY: Hafner Publish-
ing Company; 1978.
123
1268 McBride and Shorbe
2. Derkash RS. History of the Associ-
ation of Bone and Joint Surgeons.
Clin Orthop Relat Res. 1997;337:
306–309.
3. Earl Duwain McBride, MD: 1891 to
1975. J Bone Joint Surg Am. 1976;
58:287.
4. Finnegan WJ, Fenlin JM, Marvel JP,
Nardini RJ, Rothman RH. Results of
surgical intervention in the symptom-
123
atic multiply-operated back patient.
Analysis of sixty-seven cases fol-
lowed for three to seven years. J
Bone Joint Surg Am. 1979;61:1077–
1082.
5. McBride ED, Shorbe HB. Lumbosa-
cral fusion: the mortised transfacet
method by use of the vibrating electric
saw for circular bone blocks. Clin
Orthop Relat Res. 1958;12:268–275.
Clinical Orthopaedics and Related Research
50 Years Ago in CORR:
Lumbosacral Fusion: The Mortised
Transfacet Method by Use of the
Vibrating Electric Saw for Circular Bone
Blocks
Earl D. McBride MD, Howard B. Shorbe
MD
Richard A. Brand MD