Osteonecrosis in HIV disease: epidemiology, etiologies, and clinical management
".....based on the available studies, there is little evidence to suggest that HAART (or HIV itself) acts as an independent risk factor for the development of osteonecrosis. It should be noted, however, that a number of conditions listed in Table 1 have been linked to HAART, HIV, or both. These disorders include hyperlipidemia, pancreatitis, osteopenia, hyperuricemia, and osteomyelitis.....three recent reports cite annual incidence data ranging from 0.080 to 1.33% in HIV....compared to the annual incidence of symptomatic osteonecrosis in the general population estimated to be between 0.010 and 0.135%...one recent study utilized magnetic resonance imaging (MRI) to screen for osteonecrosis in 339 asymptomatic HIV-infected individuals..found 4% had asymptomatic osteonecrosis...... any HIV-infected patient providing a recent history of periarticular (joint) pain and harboring one or more of the risk factors listed in Table 1 should undergo imaging studies to exclude osteonecrosis...... X-ray and MRI can be used to assess of potential osteonecrotic disease (see discussion of imaging methods below)....preventive strategies (see below regarding corticosteroid use, lipids, alcohol) aimed at reduction of risk should be considered in HIV-infected patients with one or more risk factors for osteonecrosis..because osteonecrosis is an insidious disease doctors should maintain a high suspicion index in HIV-infected patients...."|
Summary: Osteonecrosis has been increasingly associated with HIV disease throughout the 1990s, and the incidence appears to be rising. The hip is most commonly involved and often bilaterally. Although anecodotal reports suggest an association between osteonecrosis and highly active antiretroviral therapy, controlled epidemiologic studies do not support a direct link. Many patients with osteonecrosis have established risk factors, some of which may be related to HIV disease or its therapy, including corticosteroid use and hyperlipidemia. Alcoholism, hypercoagulability, megesterol acetate use,
immune reconstitution, and other factors may also contribute. Plain radiographs and magnetic resonance imaging are the cornerstones of diagnosis. Management is dependent on the stage of bone disease and ranges from observation to total joint arthroplasty. Clinicians may help to prevent HIV-associated osteonecrosis by encouraging patients to limit their
exposure to the established risk factors for the disease. AIDS Jan 3, 2003; 17(1):1-9.
Osteonecrosis, also known as ischemic necrosis, avascular necrosis, aseptic necrosis, or osteochondritis dessicans, has been reported in the medical literature for over 60 years. As the most definitive term for the disease suggests, osteonecrosis is characterized by the in situ death of bone. Although human osteonecrotic disease occurs most frequently in the hip, it may arise at a number of locations throughout the body, including, but not limited to, the shoulder, wrist, knee, and ankle. Osteonecrosis is different than osteopenia and osteoporosis which results in reduced bone mineral density and bone loss and may lead to bone fracture.
Although there have been many reports during the 1990s of osteonecrosis in HIV-infected patients, the true incidence of symptomatic osteonecrosis has not been well defined. Three recent reports cite annual incidence data ranging from 0.080 to 1.33%. The annual incidence of symptomatic osteonecrosis in the general population has been estimated to be between 0.010 and 0.135%. One recent study utilized magnetic resonance imaging (MRI) to screen for osteonecrosis in 339 asymptomatic HIV-infected individuals. The percentage of patients with asymptomatic osteonecrosis in this population was approximately 4%. To our knowledge, however, the prevalence of asymptomatic disease has not been investigated in the population at large.
In several clinical settings, the apparent incidence of osteonecrosis in HIV-infected patients has risen, sometimes dramatically, within the past few years. At the same time, the incidence of osteonecrosis in the general population has remained relatively stable. At least in part, this phenomenon may be secondary to increased detection of osteonecrosis in HIV-infected patients as a result of heightened awareness and vigilance on the part of the clinician. However, because highly active antiretroviral therapy (HAART) has grown to become the mainstay of HIV treatment over the same period of time, it is possible that HAART might predispose HIV-infected patients to the development of osteonecrosis. The available evidence in support of this claim is summarized in the following section.
In the general population, a high degree of heterogeneity exists among the various known pathologies that lead to the development of osteonecrosis. In a final pathway common to all of these etiologies, however, death of bone tissue is invariably caused by vascular (of or related to blood vessels) compromise. This compromise often develops in response to vascular occlusion (vascular occlusion is obstruction of blood vessels, as in thromboembolic disease), but it may also appear following physical damage to vessel walls (as in radiation injury or trauma). Such an interruption of the blood supply results in ischemia, hyperemia, an increase in intraosseous pressure, and, finally, death of osteocytes. Osteonecrosis almost always affects the bone closest to the joint space, as this area is most susceptible to vascular compromise.
In approximately 80% of all osteonecrotic disease in the general population, one or more risk factors can be shown to have predisposed patients to development of the disorder. The remaining 20% is considered to be idiopathic (no known cause). Table 1 summarizes the established risk factors for osteonecrosis, as well as some frequently associated conditions. Several of these risk factors are routinely encountered among HIV-infected patients, including hyperlipidemia and corticosteroid exposure. In HIV-infected and uninfected populations alike, common risk factors include the use of ethanol and corticosteroids.
Table 1. Established risk factors for and conditions associated with osteonecrosis
Use of systemic corticosteroids
Systemic lupus erythematosus
Rheumatoid arthritis (people with HIV appear to have higher risk for arthritis type conditions
Hyperuricemia and gout (elevated uric acid and gout; gout can be associated with ribavirin use for HCV therapy as ribavirin as well as ddI for HIV therapy may lead to accumulation of uric acid as ddI therapy)
Idiopathic (no known cause)
Pancreatitis (related to ddI)
Heavy weight bearing and trauma, leading to fractures and microfactures
osteopenia or oateoporosis, leading to fractures and microfractures
dysbaria (difficulty in walking associated with mental disorders or nervous system disorders)
Osteomyelitis (inflammation of the bone or bone marrow)
Vaculitides (inflammation of the blood or lymph vessel)
tobacco and nicotine use
Although the pathogenesis of steroid- and ethanol-induced osteonecrosis remains obscure, these agents have been known to cause damaging alterations in lipid metabolism. Investigators have shown that these harmful
metabolic changes can cause fatty infiltration of the bone marrow, thereby potentially obstructing blood flow to and from bone. In addition, ethanol- or steroid-induced hyperlipidemia may promote fat embolization within the
Ethanol and corticosteroids
To date, two case-control studies and one small prospective study have systematically examined various known and proposed risk factors for osteonecrosis in HIV infection. All three studies compared groups of
HIV-infected patients with osteonecrosis to HIV-infected patients without osteonecrosis and found a significant increase in the number of recognized risk factors present in the osteonecrosis group. All three of these reports
uncovered a link between prior corticosteroid use and the subsequent diagnosis of osteonecrosis. Two of these studies reported a positive association between alcohol use and the development of osteonecrotic disease. In another pair of these investigations, glucocorticoid use was found to be the most common risk factor associated with osteonecrosis in the HIV-infected population.
Two recent studies have associated the use of megesterol acetate, an anticachectic progesterone derivative, with osteonecrosis in HIV-infected patients. Because the progesterone receptor-ligand complex has been shown to bind to DNA glucocorticoid response elements, it is possible that megesterol may predispose patients to the development of osteonecrosis by acting like a glucocorticoid.
High-dose anabolic steroid use has been linked to osteonecrotic disease in athletes. To our knowledge, however, no association has been drawn between physiologic testosterone replacement therapy and osteonecrotic
disease. In one case-control study, Scribner and colleagues evaluated testosterone levels in HIV-infected patients and found no significant differences between those with osteonecrosis and the controls. However, like progesterones, the testosterones are known to retain some degree of glucocorticoid-like activity. Therefore, testosterone and its derivatives may have the potential to contribute to the development of osteonecrosis.
Vasculitis, anticardiolipin antibodies, and hypercoagulability
Numerous vasculitides have been reported in association with HIV infection, and these inflammatory states have been shown to cause subacute swelling and disruption of the vascular endothelium, insidiously resulting in
luminal occlusion (blockage). This vascular blockage may then lead to necrosis of bone.
Although the pathogenesis of the HIV-linked vasculitides has yet to be elucidated, investigators have suggested a role for vasculitis in HIV-associated osteonecrosis and have speculated on the potential etiologic contribution of anticardiolipin antibodies. The reasons for this focus are threefold. First, anticardiolipin antibodies are estimated to be present in 50-86% of all HIV-infected patients. These numbers are thought to represent much higher rates than are found in the general population. Second, anticardiolipin antibodies have been associated with osteonecrosis in various immune vasculitic syndromes. In systemic lupus erythematosus, for example, the presence of these antibodies has been thought to serve as an independent risk factor for the development of osteonecrosis (although a recent study has seriously questioned this assumption). Third, in addition to their potential for endothelial disruption, anticardiolipin antibodies have been linked to platelet aggregation and vascular thrombosis (fibrous clot formed in a blood vessel or in a chamber of the heart). Therefore, thrombosis represents a second potential mechanism by which these antibodies might contribute to the development of osteonecrotic disease.
In one case series, four HIV-infected patients who were diagnosed with osteonecrosis lacked any classical risk factors for the complication. However, three out of the four had anticardiolipin antibodies in their blood. Because these antibodies are just as prevalent in the HIV population at large, however, no definitive conclusions can be drawn from this report.
Another condition that might play an etiologic role in HIV-associated osteonecrosis is an acquired deficiency of the antithrombotic factor protein S. Although present in only a minority of patients, HIV infection has been associated with both protein S deficiency and with the presence of anti-protein S antibodies.
A 1997 case series of six HIV-infected patients concluded that vascular thrombosis was the most likely mechanism to account for the development of osteonecrosis in HIV disease. The authors postulated that thrombosis
might result either from the actions of anticardiolipin antibodies or from a deficiency of protein S. At least half of the patients in this series, however, had known risk factors for osteonecrosis, including the use of ethanol and
glucocorticoids. In addition, this report did not give either blood levels of protein S or antibody titers.
Highly active antiretroviral therapy and HIV infection
The development of HAART as the standard of care for HIV disease has resulted in both prolonged patient survival and enhanced quality of life. Although it is well known that most of these regimens may cause adverse metabolic effects, both the scope and the mechanisms of these effects are only just beginning to become clear. Because clinicians have increasingly reported osteonecrosis in the era of HAART, a number of groups have proposed that antiviral therapy might serve as a risk factor for osteonecrosis. In contrast, two case-control studies and one recent case series have concluded that HAART is not a consistent risk factor for the development of
osteonecrosis in HIV-infected patients.
Two factors may contribute to a spurious association between osteonecrosis and HAART in uncontrolled studies. First, receipt of HAART is common and thus the likelihood that a given HIV-infected patient with osteonecrosis is
taking a HAART regimen is relatively high. Second, because of the beneficial impact of HAART on survival, survival bias may play a role. As patients live longer with HIV infection as a result of HAART, they have more opportunity to develop osteonecrosis. Nonetheless, HAART may contribute to the development of osteonecrosis by secondary effects on lipids or by the possibility of immune reconstitution. One case-control study reported that
patients with osteonecrosis tended to have longer durations of HAART and greater rises in CD4 counts from nadir levels.
Therefore, based on the available studies, there is little evidence to suggest that HAART (or HIV itself) acts as an independent risk factor for the development of osteonecrosis. It should be noted, however, that a number of conditions listed in Table 1 have been linked to HAART, HIV, or both. These disorders include hyperlipidemia, pancreatitis, osteopenia, hyperuricemia, and osteomyelitis.
In the general population, osteonecrosis primarily affects younger adults between the ages of 20 and 50 years. On average, patients are less than 40 years of age at diagnosis. The disease appears to favor the male sex by a ratio of eight to one and in 85% of those affected, osteonecrosis exclusively involves one or both of the femoral heads. In a large percentage of these (40-60%), involvement is bilateral. The remaining 15% of occurrences may
involve any of the other bones of the human body, either in addition to the femoral head or exclusive of it. Many of these latter cases also involve multiple sites. Of these alternative locations, the humeral heads, the femoral condyles, and the scaphoid and lunate bones are the most commonly affected. The clinical presentation of osteonecrotic disease appears to be identical in all patients, without regard to their HIV status.
Although occasionally acute in presentation, the typical form of osteonecrosis is notoriously insidious in onset and estimates of its progression from asymptomatic disease to symptomatic disease vary widely. The best estimates of disease progression result from investigations of the relationship between both long-term and short-term steroid
exposure and osteonecrosis. These studies separate the time of initial steroid ingestion and the appearance of symptoms by between 2 months and 10 years, with most disease becoming clinically apparent within 3 years of drug exposure. One retrospective study of idiopathic osteonecrosis places the average interval from radiographic signs to clinical symptoms at 5-6 years, with a range of 3-8 years.
History and physical examination
The symptoms of osteonecrosis are usually subtle and non-specific, with most patients complaining of mild to moderate periarticular (surrounding a joint) pain. The pain is usually triggered either by weight bearing or by moving the affected limb toward the extreme outer ranges of its motion. If one of the lower limbs is affected, a slight limp may consequently develop. There may or may not be a clear history of radiation of the pain, either toward the groin or in the direction of an adjacent joint (especially the knee or the elbow).
Most frequently, a focused physical examination will fail to uncover any striking abnormalities. Uncommonly, a focused physical examination may reveal periarticular point tenderness or the examiner might discover a decreased range of motion in the affected joint. Extensive attempts to reproduce the pain should not be attempted, as they are unlikely to alter patient management, and they have the potential to cause considerable harm.
Based on the history and physical examination, both of which are non-specific for osteonecrosis, the disorder may be confused with a number of other clinical conditions including osteoarthritis and rheumatoid arthritis. Clinically, however, any HIV-infected patient providing a recent history of periarticular pain and harboring one or more of the risk factors listed in Table 1 should undergo imaging studies to exclude osteonecrosis.
Clinical imaging has been a subject of great interest in the field of osteonecrotic disease. Plain X-ray films, computed tomography (CT), MRI, and radionuclide bone scans have all been successfully used to diagnose the disease. Because osteonecrosis has a characteristic appearance when imaged, all of these studies share a similarly high specificity for the disease. Their sensitivities, however, differ significantly.
The earliest recorded cases of osteonecrosis were diagnosed using anterior-posterior and lateral X-ray views of the affected joint. The osteonecrotic changes that might be seen on such films include cystic sclerosis, subchondral lucency, bony collapse, and secondary degenerative joint disease. Because all of these changes represent moderate to
advanced osteonecrosis, the sensitivity of plain X-ray films increases with the stage of disease. Overall, this approach has been shown to be about 40% sensitive for osteonecrosis. Despite their relatively low sensitivity, plain X-ray films are by far the least expensive of all studies used in the diagnosis of osteonecrosis. For this reason, when excluding the disease, anterior-posterior and lateral X-ray films of the affected joint remain the tests of first choice.
The range of osteonecrotic changes seen on non-contrast CT scans is nearly identical to the variety seen on plain films. Hence, the two techniques result in similar sensitivities for the disorder. CT scans are best able to detail some
of the more subtle changes of cystic sclerosis and, therefore, they are often required in the staging of osteonecrotic disease. However,, the high cost of CT combined with its relatively low sensitivity precludes its use a primary tool of choice in the evaluation of suspected osteonecrosis.
In the face of a negative X-ray, MRI is the preferred technique for the initial assessment of potential osteonecrotic disease. With its estimated sensitivity well over 90%, MRI has proven to be the most accurate overall imaging study for the detection of osteonecrosis. The value of MRI is particularly significant in the diagnosis of stage I disease.
he sensitivity of radionuclide bone scans for osteonecrosis has proven to be intermediate to that of X-ray/CT and that of MRI. Because a single study is able to sense bony changes anywhere in the body, bone scans are most often used in the detection of occult foci of osteonecrosis. Hence, the bone scan is best utilized as a follow-up measure, especially when used to discover new sites of bony necrosis in patients who have been previously
diagnosed with the disorder.
In the initial assessment of suspected osteonecrosis of the hip joint, a plain film of both the painful side and the contralateral side are indicated. In the primary evaluation of other joints, only an X-ray of the affected side is
suggested. An MRI should then be performed in the location of each negative film. If all of these tests are negative, and clinical suspicion remains high, a bone scan may be ordered, followed by CT of the affected area if necessary. Following a diagnosis of osteonecrosis, a baseline bone scan should be considered so that any involvement of other bones in the body can be excluded.
According to the results of these imaging studies, staging of the disease should next be performed. The most widely accepted method for the staging of osteonecrosis is the quantitative method detailed in Table 2. Proper
staging of the disease will help to guide the clinician toward the most appropriate treatment options.
The dearth of prospective outcome studies in osteonecrosis has resulted in a lack of clearly established treatment guidelines. Current recommendations and suggestions for managing the disease are based on expert opinion and can likely be generalized to HIV-infected patients with osteonecrosis. The published article goes on to a lengthy discussion of the various stages of osteonecrosis and possible treatment strategies.
Despite the high prevalence of risk factors for osteonecrosis in the population with HIV infection, HIV-associated osteonecrotic disease remains relatively uncommon. In our anecdotal experience, HIV- infected patients who do acquire osteonecrosis often initially present with lesions at or above stage IIIA, thereby necessitating aggressive surgical management at the outset of therapy. Therefore, preventive strategies aimed at reduction of risk should be considered in HIV-infected patients with one or more risk factors for osteonecrosis.
Glucocorticoids are prescribed for various conditions associated with HIV. Bartlett currently recommends the use of glucocorticoids only for the treatment of the complications of HIV disease that are listed in Table 3. In addition to the disorders noted in this table, glucocorticoids are often prescribed for serious comorbidities (such as severe, persistent asthma) in HIV-positive patients. Because there are studies to suggest that even the short-term use of glucocorticoids may predispose patients to osteonecrosis, these agents should be used judiciously, in the lowest effective dosages, and for the shortest possible length of time.
All patients taking HAART should be regularly monitored and treated for hyperlipidemia. Excessive ethanol use should be aggressively targeted in all HIV-infected patients.
Any of the other associated or comorbid conditions listed in Table 1 should be diagnosed and treated as appropriate for each individual disorder. Any documented hypercoagulable state should be properly managed with anticoagulant therapy. There is currently no evidence to suggest the general prophylactic use of aspirin or other mild anticoagulants in the prevention of osteonecrosis. Similarly, there is no current indication to treat asymptomatic elevations of anticardiolipin antibodies in HIV-infected patients.
Summary and conclusions
Osteonecrosis is not a new disease, but it has only been recognized in HIV-infected patients in recent years. It now seems clear that HIV-infected patients are at a higher risk for the development of osteonecrotic disease. However, these same patients are also more likely to be exposed to risk factors that have been shown to predispose the general population to the establishment of osteonecrosis.
In both HIV disease and in the general population, osteonecrosis is often insidious in nature. Because the history and the physical examination are both non-specific, the clinician must maintain a high index of suspicion for the disease in all HIV-infected patients.
X-ray and MRI are the studies of choice in all patients being evaluated for suspected osteonecrosis. Proper staging of the disorder, frequent follow-up imaging studies, and stage-specific management should follow the initial patient assessment. Clinicians may help to prevent HIV-associated osteonecrosis by encouraging HIV-infected patients to limit their exposure to the established risk factors for the disease.