| |
Osteonecrosis in Patients Infected With HIV: Clinical Epidemiology and Natural History in a Large Case Series From Spain
|
| |
| |
JAIDS Journal of Acquired Immune Deficiency Syndromes: Volume 42(3) July 2006 pp 286-292
Gutierrez, Felix MD, PhD*; Padilla, Sergio MD*; Masia, Mar MD*; Flores, Juan MD ; Boix, Vicente MD ; Merino, Esperanza MD ; Galindo, Josefa MD ; Ortega, Enrique MD ; Lopez-Aldeguer, Jose MD ; Galera, Carlos MD**; HIV-related Osteonecrosis Study Group
From the *Infectious Diseases Unit, Internal Medicine Department, Hospital General Universitario de Elche, Alicante; Infectious Diseases Unit, Hospital Arnau de Vilanova, Valencia; Infectious Diseases Unit, Hospital General Universitario, Alicante; Infectious Diseases Unit, Internal Medicine Department, Hospital Clinico Universitario, Valencia; Infectious Diseases Unit, Hospital General Universitario, Valencia; Infectious Diseases Unit, Hospital Universitario La Fe, Valencia; and **HIV Unit, Hospital Virgen de la Arrixaca, Murcia, Spain.
HIV-related Osteonecrosis Study Group members: Joan Gregori (Hospital de la Vega Baja, Alicante), Concepcion Amador (Hospital de la Marina Baixa, Alicante), Jose M. Cuadrado and Pablo Roig (Hospital Universitario de San Juan, Alicante), Onofre Martinez and Jose A. Garcia (Hospital Virgen del Rosell, Cartagena, Murcia), Clara Escolano (Hospital General Universitario, Elche), Inigo Lopez Azkarreta (Hospital General Universitario, Alicante), and Ildefonso Hernandez (Departamento de Salud Publica, Universidad Miguel Hernandez, Alicante).
Predisposing Factors for Osteonecrosis
Forty-seven (87%) of 54 cases had at least 1 identifiable risk factor previously associated with osteonecrosis in HIV-negative individuals. In 29 cases (53.7%), there was more than one of these traditional predisposing factors. The most common risk factors were dyslipidemia, alcohol abuse, use of corticosteroids or megestrol acetate, antiphospholipid antibodies, and local trauma (Table 3). Corticosteroids were prescribed mainly for opportunistic infections (P. jiroveci pneumonia, 40%; toxoplasmic encephalitis, 20%) and HIV-associated thrombocytopenia (13%). Median time on corticosteroids was 30 days (range, 15 days to 3.2 years). Mean (±SD) CD4 T-cell count at the time of prescribing corticosteroids was 116 ± 108 cells/mL. Almost half (46.6%) of the patients who received corticosteroids were on PI-based HAART regimens, and 27.2% were receiving regimens containing ritonavir. Fasting serum cholesterol and triglyceride levels at diagnosis of osteonecrosis were available in 49 and 47 patients, respectively. Mean cholesterol level was 191 ± 189 mg/dL, and mean triglyceride level was 182 ± 182 mg/dL. Twenty-two patients (44.9%) had cholesterol levels higher than 200 mg/dL, and 47 (46.8%) had triglyceride levels greater than 150 mg/dL. Nine cases had marked elevation of serum lipid levels, defined as a cholesterol level greater than 240 mg/dL or a fasting triglyceride level higher than 400 mg/dL. Only 2 patients were receiving therapy with lipid-lowering agents. Seventeen patients (31.4%) had clinical features consistent with lipodystrophy at diagnosis of osteonecrosis. Seven (13%) of 54 patients did not have any potential risk factor for developing osteonecrosis.
PATIENTS AND METHODS
We analyzed all cases of osteonecrosis diagnosed in HIV-infected patients from 1990 through 2003 in 19 HIV clinics located in 2 provinces in southeastern Spain. These clinics have provided medical care for a population of more than 10,000 HIV-infected patients since the beginning of the AIDS epidemic. Cases of osteonecrosis were identified from the records of HIV clinics by physicians caring for HIV-infected patients based on clinical presentation and radiology. Inclusion as a case of osteonecrosis required evidence of both a clinical and a radiographic diagnosis.
Information was collected at 2 time points: (1) in the first term of 2001, all cases of osteonecrosis diagnosed from 1990 through 2000 were identified and retrospectively reviewed using a standardized data collection sheet,13 and (2) in the last term of 2003, data on follow-up of all patients and information on new cases diagnosed from 2001 through 2003 were obtained. Potential risk factors associated with osteonecrosis evaluated in this study included the following: history of local trauma, previous use of systemic corticosteroids, use of systemic corticosteroids at the time of diagnosis of osteonecrosis and duration of corticosteroids therapy, alcohol abuse, presence of antiphospholipid antibodies, previous diagnosis of deep venous thrombosis or other hypercoagulable state, previous diagnosis of pancreatitis, previous diagnosis of gout, hyperlipemia, pregnancy, and use of megestrol acetate. Information concerning symptoms and functional status of the subjects at the last visit was obtained with the validated Spanish version of the Western Ontario and McMaster Universities Index (WOMAC) questionnaire.16 This questionnaire comprises 24 questions divided into 3 domains: pain (5 questions), stiffness (2 questions), and physical function (17 questions). Scores range from 0 to 25 for pain, 0 to 10 for stiffness, and 0 to 85 for physical function. A score of 0 represents the best possible health.
Patients were classified as having a history of local trauma, pancreatitis, gout, use of megestrol acetate, alcohol abuse, intravenous drug use, cocaine consumption, and lipodystrophy if any of those was recorded in the chart by the patient's clinician. Hyperlipemia was defined as elevated serum triglyceride (>150 mg/dL) or cholesterol levels (>200 mg/dL)17 or use of lipid-lowering agents.
Continuous variables were summarized by computation of the mean and/or median. We used the 2 test or Fisher exact test where appropriate to compare proportions of qualitative variables, and we used the Mann-Whitney U test to compare quantitative variables. The multivariable analysis of factors potentially associated with surgical intervention was performed by stepwise logistic regression, including all clinically important variables. A 2-tailed P value of 0.05 was considered significant. Statistical analyses were performed by means of the SPSS Version 8 (Chicago, Ill).
Abstract
Background: There is a paucity of data on clinical epidemiology of osteonecrosis in HIV-infected patients. We aimed to describe patients' characteristics and natural history of this poorly known condition.
Methods: All cases of symptomatic HIV-related osteonecrosis diagnosed from 1990 through 2003 in 19 Spanish clinics were reviewed. Functional status at the last visit was assessed with the validated Western Ontario and McMaster Universities Index questionnaire.
Results:
Of 54 patients analyzed, 29 (53.7%) had a single bone necrosis, and 25 (46.3%) had 2 or more sites involved. Progression of symptoms happened more often in patients with hip involvement (17/39 vs 0/8 patients; P = 0.019). Twenty patients (37%) required surgical intervention. Male sex and higher CD4 cell count were associated with surgery on multivariable analysis. Overall, at the end of the follow-up period, half of the patients had moderate to severe disability (Western Ontario and McMaster Universities Index score 60). During a follow-up period of 137 person-years, only 2 new episodes of osteonecrosis were observed (rate of recurrences, 1.5/100 person-years; 95% confidence interval, 0.4-5.1).
Conclusions: HIV-related osteonecrosis is associated with significant disability over time. Location of bone necrosis, sex, and CD4 cell count may influence the outcome. The risk for recurrences for patients who have experienced 1 episode is low.
Osteonecrosis, which is also referred to as avascular necrosis of the bone, is a disabling condition that has been increasingly described in the setting of HIV infection since first reported in the early 1990s.1-10 Four studies carried out in the last few years have reported annual incidence rates ranging from 0.08% to 1.33%,11-14 a much higher rate than that expected in the general population.8,11 In a recent cross-sectional study, 4% of 339 HIV-infected individuals were found to have evidence of asymptomatic osteonecrosis of the hip by magnetic resonance imaging,15 raising concerns on the potential magnitude of this condition in the future.
Previous studies on HIV-associated osteonecrosis have consisted of case reports and small series of cases or case control studies with a limited number of patients, and very few have provided data regarding patients' characteristics and follow-up. As a result, there is little information on the natural history of this condition, and clinical outcome remains largely unknown. Questions such as the risk of recurrences, the long-term physical prognosis, and the safety of orthopedic interventions in patients with this condition remain unanswered. The lack of data on outcome is most meaningful because this information is crucial to base recommendations for managing the disease. Large case series are clearly needed to accumulate experience on this rare event.
This article aims to describe the clinical epidemiology, natural history, and outcomes of osteonecrosis in a large case series.
DISCUSSION
This is the largest published series of HIV-related osteonecrosis. The study expands upon the known clinical spectrum of this condition and offers relevant information on its natural history and outcome that should be useful for clinicians managing this problem.
The natural history of HIV-related osteonecrosis is highly heterogeneous. In the present series, more than half of the patients remained clinically stable during the follow-up period and one third worsened. Interestingly, some patients tended to improve spontaneously. With regard to functional status, at the end of the follow-up period, half of the patients had developed moderate to severe physical dysfunction, highlighting the importance of this condition. Hip location seems to be associated with the poorest outcome. Indeed, in this series, all patients with osteonecrosis located in sites other than the hip had a good prognosis, and none of them had a worsening of symptoms during the follow-up period. In contrast, half of the patients with hip involvement required joint replacement an average of 2 years after the episode of osteonecrosis. Although this study did not address the indications for orthopedic interventions, it was reassuring to note that all except 1 of the patients requiring orthopedic surgical procedures had an uncomplicated clinical course after surgery, suggesting that orthopedic interventions can usually be performed safely in these patients.
Male sex and higher CD4 count at diagnosis of osteonecrosis were significantly associated with surgery on multivariable analysis. The reason why patients with osteonecrosis undergoing surgical intervention were apparently in a better immunologic condition than patients managed medically remains unknown. Although this may only reflect a reluctance of clinicians and orthopedic surgeons to recommend surgery in patients with advanced HIV disease, a potential influence of immune status in the clinical course of osteonecrosis should not be definitely ruled out. A few cases have been described in whom osteonecrosis developed after improvement in CD4 lymphocyte counts.7 It is unclear why an improved immune status could lead to osteonecrosis or aggravate its clinical course. An increased production of antiprotein S or antiphospholipid antibodies as a result of enhanced humoral immunity has been proposed,7 but to date, there are no data supporting this hypothesis.
The pathophysiology of HIV-related osteonecrosis remains uncertain. Several contributing factors might be involved. Like other clinical reports,11,15,18,19 our study disclosed a remarkably high frequency of risk factors previously associated with osteonecrosis. Indeed, potential predisposing factors were identified in 86% of the patients, the most common being the presence of dyslipidemia, alcohol abuse, use of corticosteroids or megestrol acetate, and antiphospholipid antibodies. Of note, most cases of osteonecrosis from this series were diagnosed after the introduction of HAART. Although most patients had heavy exposure to antiretroviral drugs, the lack of a control group precludes any conclusion to be drawn on the potential role of antiretroviral therapy as risk factor for osteonecrosis. Most controlled studies have not supported the contention that antiretroviral therapy increases the risk of osteonecrosis,11,15,18-20 but others have found an association14,21 with HAART, and a close relationship between duration of antiretroviral therapy and incidence of osteonecrosis has recently been reported.22 In our cohort more than half of patients were receiving stavudine alone or in combination with didanosine, both drugs being potent inhibitors of mitochondrial gamma polymerase linked to mitochondrial dysfunction and lactic acidosis.23 Whether mitochondrial toxicity might be involved in the development of osteonecrosis in patients receiving antiretroviral therapy, as it was suggested for osteopenia in HIV-infected men,24 remains undetermined.
A possible link between antiretroviral therapy and traditional risk factors for developing osteonecrosis cannot be overlooked. In a recent study, ritonavir significantly increased the systemic exposure of prednisolone in healthy subjects, suggesting that corticosteroid exposure is likely elevated in HIV-infected patients receiving PI.25 An unexpected finding was the high prevalence of hyperlipidemia in our patients, although in most cases, the lipid increases observed were mild. Hyperlipidemia is one of the most common metabolic complications of HAART, and it is a well-recognized predisposing factor for osteonecrosis in the general population. An association between dyslipidemia and osteonecrosis in HIV-infected patients receiving HAART has been found in a recent case control study.21 Whether a better metabolic control of patients receiving HAART might help to prevent HIV-associated osteonecrosis remains unknown.
In this study, osteonecrosis was seen in patients with different levels of immune status. In most patients, CD4 counts at diagnosis of osteonecrosis were above 200 cells/mL, but many of them had been severely immunocompromised in the past, as reflected by the high frequency of previous opportunistic infections and the low CD4 nadirs. A history of severe immunosuppression and low CD4 cell count nadirs have been associated with a higher risk of developing osteonecrosis in previous reports.20,22
The main caveat of this study relates to the retrospective collection of data for cases diagnosed from 1990 through 2000. This limitation, along with the lack of a control group, did not allow the identification of risk factors associated with osteonecrosis. Incomplete or missing information on follow-up in some of the old cases, and the impossibility to assess functional status by the WOMAC questionnaire in patients lost to follow-up, reduced the number of cases in which detailed information was available for subanalysis of outcomes. Those limitations may have precluded us the identification of other potential factors that may influence the outcome, such as the size of the involved bone segment that was not available in many of the cases. In addition, the follow-up period may have not been sufficiently long to avoid underestimation of those cases that eventually would require surgery at a later date. Finally, it should be acknowledged that the WOMAC questionnaire we used to assess the functional status of the subjects was developed for patients with osteoarthritis of the hip or knee and has not been evaluated in patients with osteonecrosis. However, this questionnaire is highly reliable, and it is one of the most commonly used outcome instruments in rheumatology. Because physicians taking part in the study reported all cases of HIV-related osteonecrosis cared for in their clinics during the study period, selection bias was not of great concern.
Despite the limitations, in this study, we have analyzed by far the largest number of patients with HIV-related osteonecrosis until now. Furthermore, this is the first study in which questionnaires assessing symptoms and functional status of patients have been used, and this is the first study to provide data on clinical outcome of HIV-infected patients with osteonecrosis. This information should form an essential adjunct to present knowledge of epidemiology of HIV-related osteonecrosis and should be useful to clinicians.
The study confirms that osteonecrosis is an important complication that may lead to significant disability in adults with long-lasting HIV disease under different antiretroviral regimens. Most patients have other known predisposing factors for this bone disease and have a history of severe immunosuppression. Clinicians need to be alert for early diagnosis of osteonecrosis in those patients. Because this bone disorder may present at multiple sites, the complication should be considered in patients who complain of bone pain in the proximity of any joint. Although many patients remain clinically stable after an episode of osteonecrosis, more than one third of them experience pain worsening and/or progressive limitation of joint movement over time, requiring surgery. Location of bone necrosis, sex, and CD4 cell count may influence the outcome. Hip location is associated with the poorest outcome. The risk of recurrences after having an episode of osteonecrosis is low, and life expectancy is not apparently compromised.
The optimal management of osteonecrosis in HIV-infected patients is unknown. The present study confirms that many cases of hip osteonecrosis will indeed require joint replacement. In contrast, in patients with osteonecrosis located in sites other than the hip, a conservative management may be warranted. The question of modifying antiretroviral therapy or the need of further medical interventions such as antithrombotic therapy remains unanswered. Whether any of those interventions can lead to an improvement of the outcome is unknown. However, the low rate of recurrences observed in our case series without any of those interventions is reassuring.
RESULTS
Patients' Characteristics and Antiretroviral Therapy
From 1990 through 2003, a total of 54 symptomatic cases of osteonecrosis were identified. The estimated overall annual incidence rate was 0.4 cases per 1000 HIV-infected patients. In 45 cases, the clinical diagnosis was confirmed by magnetic resonance imaging, in 5 patients, by conventional radiographs and bone scintigraphy, and in the remaining 4 cases, by conventional radiographs. Demographic and clinical characteristics are shown in Table 1. Forty-nine (90.7%) of the 54 cases were diagnosed during the highly active antiretroviral therapy (HAART) era (1997-2003). Symptomatic osteonecrosis usually presented late in the course of HIV disease. Median time from HIV diagnosis to osteonecrosis was 6.2 years, and 81.5% of cases had a previous diagnosis of AIDS. Pneumocystis jiroveci pneumonia and tuberculosis were the most common previous opportunistic infections.
Information on antiretroviral therapy was obtained in 52 cases. Fifty (96%) had previous exposure to antiretroviral drugs (Table 2). Median time on antiretroviral therapy was 3.8 years (range, 5 months to 12.1 years). At the time of diagnosis, 47 patients (90.4%) were receiving nucleoside analog reverse transcriptase inhibitors, 28 (53.8%) were on treatment with protease inhibitors, and 16 (30.8%) were receiving nonnucleoside reverse transcriptase inhibitors (NNRTIs) as part of HAART regimens. The most commonly used nucleoside analog reverse-transcriptase inhibitor were lamivudine in 31 cases (59.6%), stavudine in 24 cases (51.1%), and didanosine in 12 cases (25.5%). Six patients (11.5%) were receiving stavudine plus didanosine, 4 of them in combination with protease inhibitor (PI). Five patients (9.6%) were not receiving antiretroviral therapy at the time of diagnosis of osteonecrosis. Only 3 patients had their antiretroviral regimen modified after the diagnosis of osteonecrosis was established.
Osteonecrosis Characteristics
Twenty-nine (53.7%) of the 54 patients had a single bone necrosis, and 25 (46.3%) had 2 or more sites involved, with a total of 82 sites affected (Table 4). The hip was involved in 46 patients (85.2%), accounting for 68 (82.9%) of the 82 sites affected. Other affected sites included the knees (4 cases), shoulders (4 cases), and ankles (3 cases). Three patients (5.5%) had multiple site involvement.
No significant differences were noted between patients with single bone necrosis and those who had 2 or more sites involved, in demographic characteristics, HIV transmission category, clinical stage of HIV disease, median time since diagnosis of HIV infection, median time of follow-up since diagnosis of osteonecrosis, type and duration of antiretroviral therapy, or HIV viral load at the time of diagnosis of osteonecrosis. Patients who had 2 or more sites involved tended to have a lower nadir CD4 lymphocyte count (median [interquartile range, or IQR], 33 [51] vs 76 [247] cells/mL; P = 0.48) and a higher CD4 lymphocyte count at the time of osteonecrosis (median [IQR], 345 [435] vs 227 [433] cells/mL; P = 0.47), and they were more likely to have more than 1 potential risk factor for osteonecrosis (32% vs 6.8%; P = 0.014). No significant differences between both groups were observed in the prevalence of any individual risk factor.
Natural History, Treatment, and Outcome
Patients were followed-up for a median of 14.4 months (range, 3 months to 14.41 years), with a total of 137 person-years. Four patients died during the follow-up (2 with disseminated cytomegalovirus disease, 1 with invasive aspergillosis, and 1 with liver failure). There were only 2 new episodes of osteonecrosis throughout the follow-up period, giving a rate of recurrences of 1.5 per 100 person-years (95% confidence interval, 0.4-5.1). One of the recurrences happened in the right knee 2 years after the patient had osteonecrosis in the left knee, and in the other case, it occurred in the shoulder 2 months after being diagnosed of bilateral osteonecrosis of the hip.
There was detailed information on the clinical evolution of osteonecrosis in 47 patients. Twenty-six (53.2%) of the 47 patients remained clinically stable, and 17 (36.2%) experienced pain worsening and/or limitation of joint movement during the follow-up period. In 4 cases (8.5%), there was symptomatic improvement over time. Patients with osteonecrosis involving the hip were more likely to have worsening of symptoms than those with osteonecrosis located in sites other than the hip. Although progression of symptoms was observed in 17 (43.6%) of 39 patients with hip involvement that were medically controlled during a median of 13.3 months (range, 4.8 months to 14.4 years), no instances of progression was found among the 8 patients with osteonecrosis without hip involvement during a median follow-up period of 24.5 months (range, 3 months to 4 years) (P = 0.019).
Twenty (37%) of the 54 patients with osteonecrosis underwent surgery; 19 of them had a hip replacement. The median time from diagnosis of osteonecrosis to surgical intervention was 12.7 months (range, 1 month to 2.5 years). Surgical infection leading to removal of the prosthesis occurred in 1 case. The other 19 patients had an uncomplicated clinical course after surgery. The median follow-up period after surgery was 10.3 months (range, 1 month to 6.5 years).
Patients who underwent surgical intervention tended to be more often males with higher CD4 lymphocyte count at the time of osteonecrosis, and they were less likely to have a previous diagnosis of AIDS (Table 5). On multivariable analysis, male sex (odds ratio, 9.26; 95% confidence interval, 1.08-79.70; P = 0.030) and higher CD4 count at diagnosis of osteonecrosis (a 3% increase in the likelihood of surgery per 10 cells/mL more; P = 0.038) were significantly associated with surgery.
Information concerning the outcome of the osteonecrosis at the end of the follow-up period, as determined by the WOMAC questionnaire obtained at the last visit, was available in 36 patients (21 patients managed medically and 15 patients undergoing surgery). The mean (±SD) WOMAC scores for pain, rigidity, and function for the whole population were 10.92 ± 4.56, 4.19 ± 2.29, and 42.11 ± 16.69, respectively. Sixteen patients (44.4%) had an overall WOMAC score of at least 60, and in 2 cases (5.5%), it was at least 90. No significant differences in the mean WOMAC scores were observed between those who underwent surgery and those who did not (pain, 10.87 vs 10.95; P = 0.96; rigidity, 4.20 vs 4.19; P = 0.99; function, 46.67 vs 38.86; P = 0.17).
|
|
| |
| |
|
|
|
