HIV/AIDS • CID 2007:44 (1 March) • 739
H I V / A I D SM A J O R A R T I C L E
The Incidence and Natural History of Osteonecrosis
in HIV-Infected Adults
Caryn G. Morse,2JoAnn M. Mican,1Elizabeth C. Jones,3Galen O. Joe,4Margaret E. Rick,5Elizabeth Formentini,2
and Joseph A. Kovacs2
1Office of Clinical Research, National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH), and
Department and Departments of
2Critical Care Medicine
4Rehabilitation Medicine, and
5Laboratory Medicine, NIH Clinical Center, Bethesda, Maryland
ciency virus (HIV) infection, but the natural history has not been well described. We previously documented a
high prevalence (4.4%) of magnetic resonance imaging (MRI)–documented osteonecrosis of the hip in a cohort
of 339 asymptomatic HIV-infected patients. The present study was designed to determine the incidence of newly
diagnosed osteonecrosis in this cohort and to describe the natural history of osteonecrosis in HIV-infectedpatients.
Asymptomatic HIV-infected patients with a previous hip MRI negative for osteonecrosis underwent
follow-up MRI. Patients with asymptomatic or symptomatic osteonecrosis were enrolled in a natural history study,
which included serial MRIs and a physiotherapy follow-up.
Two hundred thirty-nine patients underwent a second MRI a median of 23 months after the initial
MRI. Osteonecrosis of the femoral head was diagnosed in 3 patients (incidence, 0.65 cases per 100 person-years).
During the period of January 1999 through April 2006, symptomatic hip osteonecrosis developed in 13 clinic
patients (incidence, 0.26 cases per 100 person-years). Among 22 patients enrolled with symptomatic hip osteo-
necrosis, 18 had bilateral involvement of the femoral heads, and 7 had osteonecrosis involving other bones. Two
(11%) of 18 asymptomatic patients and 13 (59%) of 22 symptomatic patients underwent total hip replacement.
The percentage of involvement of the weight-bearing surface of the femoral head and the rate of progression to
total hip replacement was significantly greater () in symptomatic patients than in asymptomatic patients.P ! .001
HIV-infected patients are at ∼100-fold greater risk of developing osteonecrosis than the general
population. Disease progression is slower in asymptomatic patients than in symptomatic patients. Given the high
frequency of total hip replacement in symptomatic patients, studies to assess preventive and treatment strategies
Osteonecrosis is increasingly recognized as a debilitating complication of human immunodefi-
Since 1996, the broad availability of HAART in the
United States has led to a dramatic decrease in the
incidence of opportunistic infections and malignancies
and prolonged survival for HIV-infected patients .
During this period, previously unrecognized compli-
cations of long-standing HIV infection and treatment
have had an increasing impact on the quality of life for
these patients. Osteonecrosis of the hip and of other
Received 16 August 2006; accepted 13 November 2006; electronicallypublished
23 January 2007.
Presented in part: 7th International Workshop on Adverse Drug Reactions and
Lipodystrophy in HIV, Dublin, Ireland, 13–17 November 2005 (abstract 86).
Reprints or correspondence: Dr. Caryn G. Morse, National Institutes of Health,
9000 Rockville Pike, Bldg. 10, Rm. 2C145, MSC 1662, Bethesda, MD 20892-1662
Clinical Infectious Diseases2007;44:739–48
This article is in the public domain, and no copyright is claimed.
bones is one such debilitating complication. First de-
reports and retrospectivecasestudieshavesubsequently
appeared in the medical literature . An increased
incidence of previously recognized risk factors for os-
teonecrosis, such as corticosteroid use, hypercoagulable
state, alcohol abuse, and tobacco use, has been asso-
ciated with osteonecrosis in these patients [3–13]. Ad-
ditional reported risk factors include the use of anti-
retroviral drugs (especially protease inhibitors), the
presence of lipodystrophy syndrome, and use of me-
gestrol acetate or testosterone [10, 12, 14, 15].
The annual incidence of symptomatic osteonecrosis
in the general population has been estimated to be
approximately 0.003–0.006 cases per 100 person-years
[16, 17]. Recent retrospective case studies of HIV-in-
fected patients have reported incidences ranging from
0.03 to 0.37 cases per 100 person-years [7, 8, 12, 18–
by guest on October 28, 2015
740 • CID 2007:44 (1 March) • HIV/AIDS
of the patients in whom the diagnosis of osteonecrosis was made after the development of clinical symptoms (right). The patients indicated in blue
and red were those included in the natural history analysis for the asymptomatic and symptomatic cohorts, respectively.
Flow diagram indicating the outcome of the asymptomatic patients who underwent MRI screening for osteonecrosis (left) and the source
20], although the true incidence of symptomatic osteonecrosis
evaluation of osteonecrosis in HIV-infected patients has been
undertaken. Moreover, the natural history of osteonecrosis—
specifically, the frequency of and time to hip replacement—is
undefined in this population. Given that recent studies have
shown that treatment with bisphosphonates can reduce the
need for total hip replacement (THR) in certain populations,
it is critical to understand the natural history of osteonecro-
sis in both symptomatic and asymptomatic HIV-infected
In 2001, our research group reported an unexpectedly high
prevalence (4.4%) of MRI-documented osteonecrosisofthehip
in a cohort of 339 asymptomatic HIV-infected patients.The
identification of a cohort of HIV-infected adults with negative
results of hip MRIs, as well as a cohort of patients with MRI-
documented osteonecrosis, provided a unique opportunity to
study the incidence and naturalhistoryofosteonecrosisinHIV-
Here, we present a prospective study documenting the in-
cidence of osteonecrosis in asymptomatic HIV-infected pa-
tients. In addition, we report longitudinal follow-up data on a
cohort of prospectively identified HIV-infected patients with
asymptomatic osteonecrosis and a cohort of symptomaticHIV-
infected patients who were enrolled after receiving a diagnosis
The methods of the initial prevalencestudyhavebeendescribed
elsewhere . In brief, 339 HIV-infected adults enrolled in
studies of the treatment or natural history of HIV infection at
the National Institutes of Health Clinical Center (Bethesda,
MD) underwent MRI of both femoral heads. Because the study
by guest on October 28, 2015
HIV/AIDS • CID 2007:44 (1 March) • 741
patient with prospectively identified disease. The initial MRI (from 20
July 1999; A) findings were normal. A follow-up screening MRI on 14
June 2001 (not shown) identified bilateral osteonecrosis. Representative
high-resolution T-1 scans of each hip obtained shortly thereafter are
shown on the left (right hip, 27 June 2001 [B]; left hip, 25 June 2001
[D]) demonstrate the characteristic findings of osteonecrosis (arrows),
with linear areas of abnormal signal in the femoral heads bilaterally
without gross deformity of the femoral heads. No evidence of sclerosis
or marrow edema was noted. The lesions showed mild progression at 1
month, then remained stable until 19 months after the diagnosis (right),
when progression can be seen in both hips in the T-1 scans. The T-2scans
(bottom) demonstrate the increase in bone marrow edema and effusion
in the left hip on 3 February 2003 (E and G), compared with 25 June
2001 (D and F). The patient underwent total hip replacement on the right
hip 23 months after diagnosis and on the left hip 28 months after
Development of osteonecrosis of both femoral heads in a
initially focused on asymptomatic patients, persons with hip
or groin pain were excluded. As part of the initial screening
study, all participants completed a standard questionnaire that
addressed joint symptoms, medical history, medication use,
exercise routine, and substance use.
Participants without evidence of osteonecrosis on the initial
screening MRI underwent a second MRI 17–31 months after
the first study. The medical records of all patients who partic-
ipated in the initial trial were reviewed through April 2006; no
additional patients with osteonecrosis were identified, other
than 1 patient who had already been enrolled in the study.
Patients with MRI-documented osteonecrosis, which was di-
agnosed either as part of this study or outside the context of
this study, were eligible to participate in a natural history study
that included additional laboratory evaluation,serialMRIs,and
physiotherapy. Participants with osteonecrosis underwent op-
tional additional MRI after approximately 3, 6, and 12 months
and annually thereafter. Laboratory and clinical data for all
patients were retrieved from a patient database; data through
April 2006 are included in this report. This protocol was ap-
proved by the National Institute of Allergy and Infectious Dis-
eases institutional review board, and all participants provided
written informed consent.
MRIs were performed using an LX Horizon 1.5-T
MRI system (General Electric Medical Systems) in accordance
with a previously described method . The percentage of
involvement of the weight-bearing surface of the femoral head
was determined using previously reported methods and was
graded as !25%, 25%–50%, and 150% [23, 24]. All MRI find-
ings were interpreted by one of the investigators (E.C.J.).
In our previous evaluation for a
hypercoagulable state, we found that, among a panel of assays,
only anticardiolipin antibody levels were elevated significantly
more in patients with osteonecrosis than in HIV-infected pa-
tients without osteonecrosis. Therefore, from the prior panel,
only anticardiolipin antibody levelsweredeterminedinpatients
with osteonecrosis. Commercially available kits were used
(Quantilite; Innova) .
Participants with MRI evidence of
osteonecrosis underwent a detailed clinical evaluation, which
included a functional history of vocational and avocational
activity and a physical examination of the hips, as previously
described . Functional mobility status was determined us-
ing the Sickness Impact Profile Ambulation Subscale .
culated by dividing the number of cases of osteonecrosis by
the total number of person-years of follow-up for HIV-infected
patients regularly observed at our clinic during the period of
January 1999 through April 2006, or, for the asymptomatic
group, the total number of person-years between the 2 MRIs.
For comparison of categorical variables, the x2test or Fisher’s
by guest on October 28, 2015
742 • CID 2007:44 (1 March) • HIV/AIDS
infected patients with hip osteonecrosis.
Comparison of select clinical and laboratory characteristics of asymptomatic and symptomatic HIV-
(n p 18)
(n p 22)Pb
Age, median years (range)
CD4 cell count, median ?109cells/L (range)
HIV load, median copies/mL (range)
Duration of HIV infection, median years (range)
History of opportunistic infection
Protease inhibitor use
Duration of protease inhibitor use, median years (range)
Lipid lowering agents
Positive ACA IgGd
ACA 123 U
Bilateral hip osteonecrosis
Total hip replacement
1073 (!50 to 123,973)
!50 (!50 to 242,844)
aCharacteristics are at the time of diagnosis, except for bilateral hip osteonecrosis and frequency of progression to total hip
replacement. Baseline characteristics for 15 of the asymptomatic patients have been previously reported  and are included to
allow comparison of the entire cohort of asymptomatic patients with symptomatic patients.
bP values were determined using either the x2test, Fischer’s exact test, or the nonparametric Mann-Whitney U test.
cDiagnosed by patient report and confirmed by physical examination by a physician.
dACA values were available for 18 asymptomatic and 20 symptomatic patients.
Data are no. (%) of patients, unless otherwise indicated. ACA, anti-cardiolipin antibodies; NS, not significant.
exact test was used. For comparison of continuous variables,
the nonparametric Mann-Whitney U test was used. Kaplan-
Meier survival analysis to examine the probability of under-
going THR was performed using JMP, version 6.0, Macintosh
version (SAS Institute).
an initial screening MRI, 15 were found to have osteonecrosis,
as previously reported . Of the 324 participants without
evidence of osteonecrosis, 81 patients either declined to un-
dergo a second MRI or did not return to the clinic during the
follow-up period, and 4 died before undergoing an additional
MRI (figure 1).
Two hundred thirty-nine participants (median age, 43 years;
range, 23–70 years) underwent a second screening MRI during
the period from February 2001 through January2002,amedian
of 23 months (range, 17–31 months) after the initial MRI.
Clinical and laboratory characteristics of the 239 patients are
similar to those previously reported for the entire cohort and
Of 339 asymptomatic participants who underwent
to those of patients who did not undergo a second MRI .
On the basis of the follow-up MRI, 3 patients (1.3%) received
a diagnosis of osteonecrosis of the femoral head (bilateral for
all patients), for an incidence of 0.65 cases per 100 person-
years (95% CI, 0.13–1.89 cases per 100 person-years). An ad-
ditional patient, whose screening and follow-up MRIs yielded
normal findings, received a diagnosis of symptomatic left hip
osteonecrosis 3 years after his second MRI. The patients with
newly diagnosed osteonecrosis did not differ significantly from
the patients without osteonecrosis with regard to sex, age, ex-
posure group, duration of HIV infection, or antiretroviraltreat-
ment history (data not shown).
All MRI-identified lesions had features characteristic of os-
teonecrosis (i.e., diminished signal on T1-weighted imageswith
a corresponding bright signal on fat-suppressed, T2-weighted
images) (figure 2). The initial negative MRI results were re-
viewed after the diagnosis of osteonecrosis, and again, no ab-
normalities were noted. Thus, 4 (1.7%; 95% CI, 0.5%–4.2%)
of the 239 asymptomatic patients with previously negativeMRI
results developed MRI-documented osteonecrosis within 5
by guest on October 28, 2015
HIV/AIDS • CID 2007:44 (1 March) • 743
atic HIV-infected patients with hip osteonecrosis.
Clinical and radiographic outcomes for asymptomatic and symptom-
Symptom or change
Mild (not requiring pain medications)
Moderate-severe (requiring pain medications)
Total hip replacement
with longitudinal clinical data and radiographic changes over time for patients with serial ra-
diographic studies are shown. Hips that required total hip replacement were excluded from
both clinical and radiographic analysis.
aData are for 22 hips in the asymptomatic cohort and 20 hips in the symptomatic cohort.
bData are for 20 hips in the asymptomatic cohort and 12 hips in the symptomatic cohort.
Data are no. of hips. Clinical symptoms at the last clinical follow-up visit for patients
bearing portion of the femoral head at the time of diagnosis of
hip osteonecrosis in asymptomatic and symptomaticHIV-infected
Comparison of the extent of involvement of the weight-
No. of hips
Percentage of femoral
aDetermined using the x2test.
Data are no. (%) of hips.
years of the initial negative result, and a cumulative 5.6% (95%
CI, 3.4%–8.6%; 19 patients) of the initial cohort of 339 asymp-
tomatic patients had MRI-documented osteonecrosis.
During the period from January 1999 through April 2006,
thirteen patients who were enrolled in research studies at the
National Institutes of Health Clinical Center developed hip or
groin pain and received a diagnosis of hip osteonecrosis (in-
cluding the patient with 2 negative screening MRI results, but
excluding patients who were identified by screening MRI find-
ings and who subsequently developed symptoms) and were
prospectively observed. The incidence of symptomatic osteo-
necrosis in the clinic population during this time was ∼0.26
cases per 100 person-years (95% CI, 0.14–0.44 cases per 100
person-years). Clinical characteristics of this group are sum-
An additional 9 patients were referred to
the study after they received a diagnosis of symptomatic os-
teonecrosis, for a total of 40 patients with documented hip
osteonecrosis. Because the 18 patients whose diagnoses were
based on screening MRIs were initially asymptomatic and po-
tentially had a natural history that differed from the 22 who
separately (figure 1).
Among the 18 patients identifiedby
for the 3 patients identified by the second screening MRI were
similar to those previously reported for the other 15 patients
(table 1). By a median duration of follow-up of 5.7 years, most
of the patients remained asymptomatic or reported mild pain
(table 2). Two patients underwent THR bilaterally (23 and 28
months after diagnosis for one patient and 57 and 59 months
for the other). Three patients died (1 died of lymphoma, 1
died of lymphoma and stage IV anal carcinoma, and 1 died of
With regard to the initial abnormal MRI findings, the ma-
jority of the hips had !25% or 25%–50% involvement of the
femoral head (table 3). Serial MRIs were available for 16 of the
18 patients over a median period of 48 months (range, 10–74
months) (figure 2). Most of the patients demonstrated radio-
graphic stability or improvement (table 2). One patientshowed
resolution of a small left hip lesion at 28 monthsafterdiagnosis,
as well as improvement in a larger lesion of the right hip (figure
3). Two patients, both of whom initially had 150%involvement
The findings of physiatric evaluations were available for 14
of these patients. The initial findings have been reported pre-
viously: detailed physical examinations were unable to detect
by guest on October 28, 2015
744 • CID 2007:44 (1 March) • HIV/AIDS
MRI (from 18 July 1999; top) revealed bilateral osteonecrosis of the
femoral heads (arrows). As illustrated by a follow-up MRI (from 25 June
2004; bottom), the lesion in the left hip resolved over time, and the
lesion in the right hip decreased in size (arrow).
Serial MRIs of an asymptomatic patient with bilateral hip
Longitudinal physiatric evaluation findings for 18 asymptomatic subjects with
No. of years
after diagnosis of
range of motion
abnormality PROM Pain
upon awakening Gellinga
1 Year (n p 10)
2 Years (n p 8)
3 Years (n p 7)
6 Years (n p 3)
hip stress maneuvers.
aGelling, stiffness, or limited mobility after sitting for any period of time.
Data are no. of patients with each finding. PROM, pain on passive end range of motion or provocative
clinical signs in the asymptomatic osteonecrosis cohort that
differed significantly from the clinical signs in HIV-infected
patients without MRI evidence of osteonecrosis . Serial
examination findings are available for 15 of the 18 patients for
up to 6 years after the initial diagnosis (table 4). The majority
of patients had some range of motion abnormalityateachtime;
other abnormalities were seen less frequently.
Of the 22 patients with symptomatic
hip osteonecrosis (figure 1), 18 had bilateral involvement of the
femoral heads, and 7 had symptomatic osteonecrosis involving
other bones, including the shoulders (5 patients), knees (3 pa-
tients), and ankles (1 patient) (figure 4). Potential risk factors
for osteonecrosis were also common in this cohort (table 1).
After a median duration of follow-up of 26 months, 13 pa-
tients (59%) underwent THR at a median of 10 months (range,
1–45 months) after diagnosis; 8 of these required THR within
1 year after diagnosis. Three of these patients underwent bi-
lateral THR, one of whom had previously undergone bilateral
core decompression and bilateral shoulder surgery to help re-
lieve symptoms; a fourth had a bone graft in the second hip;
a fifth underwent a shoulder replacement before the THR. The
majority of patients that did not require THR had persistent
pain requiring use of nonsteroidal or narcotic painmedications
(table 2). Four patients were lost to long-term follow-up, and
2 died (one died of lymphoma, and the other died of an un-
Findings of radiographic studies (16 MRIs, 1 CT, and 1 plain
film examination) were available for 18 patients (30 hips) at
approximately the time of diagnosis. Eighty-seven percent of
the hips (26 hips in 18 patients) had 150% involvement of the
femoral head. Serial imaging data were available for 10 patients
(table 2). Two patients developed new lesions in the contra-
lateral hip. No patient showed improvement or resolution.
Eight (89%) of 9 subjects who underwent physiatric evalu-
ation had range of motion loss noted by passive range-of-
motion testing, pain at end range, and with provocative tests,
morning stiffness, gelling, and hip pain. Seven subjects (78%)
described a limitation related to functional mobility.
In a comparison of the baseline characteristics of the symp-
tomatic and asymptomatic patients with osteonecrosis (tables
1 and 3), the most striking difference was the percentage of
involvement of the femoral head, which was significantly
greater in symptomatic patients, (
tory was also markedly different: THR was significantly more
common in symptomatic patients (
survival analysis (figure 5, top) demonstrated a significantly
more rapid progression to THR in symptomatic patients, com-
pared with asymptomatic patients (
A similar analysis, which combined data for all patients, high-
lighted the relationship between the percentage of involvement
of the femoral head and clinical outcome: only patients with
). The natural his-
P ! .0001
P p .003
, by log rank test).
P ! .001
by guest on October 28, 2015
HIV/AIDS • CID 2007:44 (1 March) • 745
(top), bilateral osteonecrosis of the hips (middle), bilateral osteonecrosis of the distal femurs (bottom left), and osteonecrosis of the left proximal tibia
MRIs for a patient with multiple lesions due to osteonecrosis (arrows). T1-weighted images show bilateral osteonecrosis of the shoulders
150% involvement later required THR (
test) (figure 5, bottom).
, by log rankP ! .001
The results of this prospectively studied cohort provide strong
evidence that HIV-infected patients are at substantially in-
creased risk for the development of symptomatic or asymp-
tomatic osteonecrosis. Moreover, there is a high rate of pro-
gressive disease and a need for THR—especially in patients in
whom osteonecrosis affected 150% of the weight-bearing por-
tion of the femoral head, which included the majority of symp-
Our finding of incidence rates of 0.65 cases per 100 patient-
years for asymptomatic patients and 0.26 cases per 100 patient-
years for symptomatic patients is ∼100-fold higher than the
estimated incidence in the general population [16, 17]. The
prevalence of osteonecrosis in the initial cohort of 339 asymp-
tomatic patients who were evaluated by MRI is also extraor-
dinarily high at 5.6% and is similar to prevalences reported
among patients at high risk for osteonecrosis in the context of
a variety of underlying diseases (table 5) . The 5.6% prev-
alence represents a minimum estimate, becauseadditionalcases
involving asymptomatic disease may have developed in this
population since the last series of MRIs.
by guest on October 28, 2015
746 • CID 2007:44 (1 March) • HIV/AIDS
hip replacement. Top, Comparison of the time to total hip replacement
in 18 patients with asymptomatic osteonecrosis (red) and 22 patients
with symptomatic osteonecrosis (blue). For patients with bilateral hip
replacement, the time to the first procedure was used. There was a
significant difference in time to progression to total hip replacement
( , by the log rank test). Bottom, Comparison of the time to totalP ! .001
hip replacement in individual hips in which osteonecrosis involved !25%
(red;), 25%–50% (green;
n p 14n p 12
the weight-bearing surface of the femoral head (
), and 150% (blue;
P ! .001
) ofn p 34
, by the log rank
A high rate of asymptomatic osteonecrosis is not unique to
with a high frequency in MRI surveys of patients with alco-
holism or systemic lupus erythematosus and in renaltransplant
recipients, many of whom were receiving corticosteroids, as
well as a cohort of patients with primary antiphospholipid
syndrome, none of whom had receivedcorticosteroidtreatment
(table 5) [22, 27–31]. Giventhattheprimarydifferencebetween
asymptomatic and symptomatic patients is the extent of in-
volvement of the femoral head, it appears likely that the 2
groups represent a continuum of the same process, with the
size of the lesion being the primary determinant of symptoms.
The incidence of osteonecrosis among HIV-infected patients
increased after the introduction of HAART that included pro-
tease inhibitors ∼10 years ago [20, 32], raising the possibility
that this class of drugs or these combination regimens played
an etiologic role. However, on the basis of current data, it is
difficult to conclude that protease inhibitors or antiretroviral
combinations are independently associated with the develop-
ment of osteonecrosis [5, 6, 9, 12, 19, 33]. It is important to
note, however, that a number of risk factors for the develop-
ment of osteonecrosis are associated with HIV infection,
management of HIV-related complications, or antiretroviral
therapy, including pancreatitis, hyperlipidemia, osteopenia/os-
teoporosis, and use of corticosteroids. Chronic inflammation
in the context of long-standing HIV infection may also con-
tribute to the development of osteonecrosis, as has been pos-
tulated for systemic lupus erythematosus .
Corticosteroid use is noteworthy, because it has been con-
sistently identified as one of the most important risk factors
for osteonecrosis among HIV-infected persons, as well as
among other patient populations [6, 12]. Protease inhibitors
may exacerbate the effects of corticosteroids by altering their
cytochrome p450-mediated metabolism. In a study that was
prompted in part by our observations in the initial screening
MRI study, we havedemonstratedthatlowdosesoftheprotease
inhibitor ritonavir significantly increase systemic exposure to
prednisolone in healthy volunteers . Corticosteroids are
used in the management of a number of HIV-related oppor-
tunistic infections, such as Pneumocystis pneumonia, as well as
for non–HIV-related medical conditions, such as asthma and
allergic reactions; they have also been used to explore the role
of immune activation in the pathogenesis of HIV-related im-
munodeficiency. In one such study, osteonecrosiswasidentified
by MRI in 2 (18%) of 11 asymptomatic patients, highlighting
the potential risks in this population and further emphasizing
how corticosteroids must be used cautiously in HIV-infected
We have documented a rapid progression of diseaseinsymp-
tomatic patients, such that 59% of patients had progression
sufficient to require THR. Although the inclusion of patients
from outside of our clinic may have resulted in a selection bias,
the referred patients hadconditionsthatprogressedtothepoint
that required surgery at a rate similar to those identified pro-
spectively within our clinic population and at rates similar to
another report, in which nearly 50% of symptomatic HIV-
infected patients with osteonecrosis of the hip required THR
. In our cohort, this appears primarily to be related to the
extent of disease involvement at the time of diagnosis, which
was significantly greater in symptomatic patients. Moreover,
THR was performed only for patients who presented with
150% involvement of the femoral head, regardless of the pres-
ence of symptoms at diagnosis. The natural history of osteo-
necrosis in other populations has not been clearly defined but
depends on the stage and size of the lesion; one review of
published studies found that 80% of hips that had been treated
by guest on October 28, 2015
HIV/AIDS • CID 2007:44 (1 March) • 747
Table 5. Incidence and prevalence of osteonecrosis in select at-risk patient populations.
no. of cases per
100 person-years Prevalence, %
General population 
Patients with systemic lupus erythematosus
Not receiving therapy 
Receiving therapy, asymptomatic osteonecrosis 
Receiving therapy, symptomatic osteonecrosis 
Patients with primary antiphospholipid syndrome not receiving
therapy, asymptomatic osteonecrosis 
Patients with various autoimmune disorders, receiving high-
dose corticosteroid treatment, asymptomatic osteone-
Renal transplant recipients receiving high-dose corticosteroid
treatment, asymptomatic osteonecrosis 
Patients with acute lymphoblastic leukemia receiving therapy,
asymptomatic osteonecrosis 
Bone marrow transplant recipients receiving therapy, symptom-
atic osteonecrosis 
aFrom the present report.
conservatively ultimately required THR . Although some
studies suggest that core decompression can mitigate the nat-
ural history , the utility of this procedure remains contro-
versial and may be applicable to only a limited number of
Asymptomatic disease is not benign, because the conditions
of 2 patients progressed to the point that THR was required
during a median follow-up period of 5.8years,andtheduration
of follow-up may not be adequate to determine whether the
remaining patients’ conditions will remain stable. In a pro-
spective study of small, asymptomatic lesions in 40 patients
with symptomatic osteonecrosis in the contralateral hip, col-
lapse of the asymptomatic hip occurred in 29 patients a mean
of ∼7.5 years after diagnosis (range, 6–11.5 years), and all 29
patients required surgical intervention (20 required THR) .
Given the high frequency of disease progression thatrequires
THR in HIV-infected patients with symptomatic osteonecrosis
(59%), interventions to prevent progression must be critically
evaluated. Conservative management does not appear to affect
the natural history, and the role of core decompressionremains
controversial. However, recently published studies, including
one randomized trial, have suggested that bisphosphonatescan
modulate thenaturalhistoryof osteonecrosisofthefemoralhead
[21, 40, 41]. In light of these data, additional investigations of
bisphosphonates for the treatment of HIV-associated osteone-
crosis are warranted; reports of bisphosphonate-associated os-
teonecrosis of the jaw should, however, temper routine use of
these drugs for these causes until safetyhasbeenestablished.
We thank the patients for their willingness to participate in the study,
Henry Masur for his helpful suggestions and critical review of the man-
uscript, Dean Follmann for statistical assistance, the medical staff of the
National Institute of Allergy and Infectious Diseases and Critical Care
Medicine Department for the care of the patients described in this report,
and the technicians in the Radiology Department whoperformedthescans.
The Intramural Research Program of the National
Institute of AllergyandInfectiousDiseasesandtheClinicalCenter,National
Institutes of Health.
Potential conflicts of interest.
All authors: no conflicts.
1. Kaplan JE, Hanson D, Dworkin MS, et al. Epidemiology of human
immunodeficiency virus-associated opportunistic infections in the
United States in the era of highly active antiretroviral therapy. Clin
Infect Dis 2000;30(Suppl 1):S5–14.
2. Goorney BP, Lacey H, Thurairajasingam S, Brown JD. Avascular ne-
crosis of the hip in a man with HIV infection. GenitourinMed1990;66:
3. Smoking and avascular necrosis. AIDS Patient Care STDS 2003;17:
4. Aguiar De Aragao P, Lemos RB, Aquino MZ, Souza Marques HH.
Hyperlipidaemia a risk factor for femoral head osteonecrosis (Legg-
Calv -Perthes-like disease) in children with AIDS: case report. Braz J
Infect Dis 2002;6:142–5.
5. Allison GT, Bostrom MP, Glesby MJ. Osteonecrosis in HIV disease:
epidemiology, etiologies, and clinical management. AIDS 2003;17:1–9.
6. Glesby MJ, Hoover DR, Vaamonde CM. Osteonecrosis in patients in-
fected with human immunodeficiency virus: a case-control study. J
Infect Dis 2001;184:519–23.
7. Hasse B, Ledergerber B, Egger M, et al. Antiretroviral treatment and
osteonecrosis in patients of the Swiss HIV Cohort Study: a nested case-
control study. AIDS Res Hum Retroviruses 2004;20:909–15.
by guest on October 28, 2015
748 • CID 2007:44 (1 March) • HIV/AIDS
8. Martin K, Lawson-Ayayi S, Miremont-Salame G, et al. Symptomatic
bone disorders in HIV-infected patients: incidence in the Aquitaine
cohort (1999–2002). HIV Med 2004;5:421–6.
9. Miller KD, Masur H, Jones EC, et al. High prevalence of osteonecrosis
of the femoral head in HIV-infected adults. Ann Intern Med 2002;
10. Molia AC, Strady C, Rouger C, Beguinot IM, Berger JL, Trenque TC.
Osteonecrosis in six HIV-infected patients receiving highly active an-
tiretroviral therapy. Ann Pharmacother 2004;38:2050–4.
11. Blacksin MF, Kloser PC, Simon J. Avascular necrosis of bone in human
immunodeficiency virus infected patients. Clin Imaging 1999;23:
12. Scribner AN, Troia-Cancio PV, Cox BA, et al. Osteonecrosis in HIV:
a case-control study. J Acquir Immune Defic Syndr 2000;25:19–25.
13. Lawson-Ayayi S, Bonnet F, Bernardin E, et al. Avascular necrosis in
HIV-infected patients: a case-control study from the AquitaineCohort,
1997–2002, France. Clin Infect Dis 2005;40:1188–93.
14. Valencia ME, Barreiro P, Soriano V, Blanco F, Moreno V, Lahoz JG.
Avascular necrosis in HIV-infected patients receiving antiretroviral
treatment: study of seven cases. HIV Clin Trials 2003;4:132–6.
15. Koller E, Mann M, Malozowski S, Bacsanyi J, GibertC.Asepticnecrosis
in HIV seropositive patients: a possible etiologic role for megestrol
acetate. AIDS Patient Care STDS 2000;14:405–10.
16. Bauer M, Thabault P, Estok D, Christiansen C, Platt R. Low-dose
corticosteroids and avascular necrosis of the hip and knee. Pharma-
coepidemiology and Drug Safety 2000;9:187–91.
17. Mont MA, Hungerford DS. Non-traumatic avascular necrosis of the
femoral head. J Bone Joint Surg Am 1995;77:459–74.
18. Brown P, Crane L. Avascular necrosis of bone in patients with human
immunodeficiency virus infection: report of 6 cases and review of the
literature. Clin Infect Dis 2001;32:1221–6.
19. Calza L, Manfredi R, Mastroianni A, Chiodo F. Osteonecrosis and
highly active antiretroviral therapy during HIV infection: report of a
series and literature review. AIDS Patient Care STDS 2001;15:385–9.
20. Gutierrez F, Padilla S, Ortega E, et al. Avascular necrosis of the bone
in HIV-infected patients: incidence and associated factors. AIDS
21. 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;87:2155–9.
22. Tervonen O, Mueller DM, Matteson EL, Velosa JA, Ginsburg WW,
Ehman RL. Clinically occult avascular necrosis of the hip: prevalence
in an asymptomatic population at risk. Radiology 1992;182:845–7.
23. Khanna AJ, Yoon TR, Mont MA, Hungerford DS, Bluemke DA. Fem-
oral head osteonecrosis: detection and grading by using a rapid MR
imaging protocol. Radiology 2000;217:188–92.
24. LafforgueP, Dahan E,ChagnaudC,SchianoA,KasbarianM,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
25. Joe G, Kovacs J, Miller K, et al. Diagnosis of avascular necrosis of the
hip in asymptomatic HIV-infectedpatients:clinicalcorrelationofphys-
ical examination with magnetic resonance imaging. Journal of Back
and Musculoskeletal Rehabilitation 2002;16:135–9.
26. Bergner M, Bobbitt RA, Pollard WE, Martin DP, Gilson BS. The sick-
ness impact profile: validation of a health status measure. Med Care
27. Assouline-Dayan Y, Chang C, Greenspan A, Shoenfeld Y, Gershwin
ME. Pathogenesis and natural history of osteonecrosis. SeminArthritis
28. Nagasawa K, Tada Y, Koarada S, et al. Very early development of
steroid-associated osteonecrosis of femoral head in systemic lupus er-
ythematosus: prospective study by MRI. Lupus 2005;14:385–90.
29. Sakamoto M, Shimizu K, Iida S, Akita T, Moriya H, Nawata Y. Os-
teonecrosis of the femoral head: a prospective study with MRI. J Bone
Joint Surg Br 1997;79:213–9.
30. Takatori Y, Kokubo T, Ninomiya 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;75:217–21.
31. Tektonidou MG, Malagari K, Vlachoyiannopoulos PG, Kelekis DA,
Moutsopoulos HM. Asymptomatic avascular necrosis in patients with
primary antiphospholipid syndrome in the absence of corticosteroid
use: a prospective study by magnetic resonance imaging. Arthritis
32. Keruly JC, Chaisson RE, Moore RD. Increasing incidence of avascular
necrosis of the hip in HIV-infected patients. J Acquir Immune Defic
33. Mary-Krause M, Billaud E, Poizot-Martin I, et al. Risk factors for
osteonecrosis in HIV-infected patients: impact of treatment with com-
bination antiretroviral therapy. AIDS 2006;20:1627–35.
34. Lane NE. Therapy insight: osteoporosis and osteonecrosis in systemic
lupus erythematosus. Nat Clin Pract Rheumatol 2006;2:562–9.
35. Penzak SR, Formentini E, Alfaro RM, Long M, Natarajan V, Kovacs
J. Prednisolone pharmacokinetics in the presence and absence of ri-
tonavir after oral prednisone administration to healthy volunteers. J
Acquir Immune Defic Syndr 2005;40:573–80.
36. Wallis RS, Kalayjian R, Jacobson JM, et al. A study of the immunology,
virology, and safety of prednisone in HIV-1-infected subjects withCD4
cell counts of 200 to 700 mm3. J Acquir Immune Defic Syndr 2003;
37. Gutierrez F, Padilla S, Masia M, et al. Osteonecrosis in patientsinfected
with HIV: clinical epidemiology and natural history in a large case
series from Spain. J Acquir Immune Defic Syndr 2006;42:286–92.
38. Mont MA, Carbone JJ, Fairbank AC. Core decompression versus non-
operative management for osteonecrosis of the hip. Clin Orthop Relat
39. Hernigou P, Poignard A, Nogier A, Manicom O. Fate of very small
asymptomatic stage-I osteonecrotic lesions of the hip. J Bone Joint
Surg Am 2004;86-A:2589–93.
40. Agarwala S, Jain D, Joshi VR, Sule A. Efficacy of alendronate, a bis-
phosphonate, in the treatment of AVN of the hip: a prospective open-
label study. Rheumatology (Oxford) 2005;44:352–9.
41. Nishii T, Sugano N, Miki H, Hashimoto J, Yoshikawa H. Does alen-
dronate prevent collapse in osteonecrosis of the femoral head? Clin
Orthop Relat Res 2006;443:273–9.
42. Woo SB, Hellstein JW, Kalmar JR. Narrative review: bisphosphonates
and osteonecrosis of the jaws. Ann Intern Med 2006;144:753–61.
43. Ojala AE, Paakko E, Lanning FP, Lanning M. Osteonecrosis duringthe
treatment of childhood acute lymphoblastic leukemia: a prospective
MRI study. Med Pediatr Oncol 1999;32:11–7.
44. Fink JC, Leisenring WM, Sullivan KM, Sherrard DJ, Weiss NS. Avas-
cular necrosis following bone marrow transplantation: a case-control
study. Bone 1998;22:67–71.
Note added in proof.
of the femoral heads bilaterally, a patient with osteonecrosis of the femoral heads bilaterally received a diagnosis of osteonecrosis of
the knees bilaterally, and a patient with bilateral osteonecrosis of the femoral heads (identified initially as part of the asymptomatic
cohort) had progression to left THR. The latter patient had 25%–50% involvement of the left femoral head.
Since this manuscript was submitted for publication, an additional patient receivedadiagnosisofosteonecrosis
by guest on October 28, 2015