| |
80% Bone Abnormalities in France Aquitane Cohort, 27% Osteoporosis
|
| |
| |
Reduced bone mineral density in HIV-infected patients: prevalence and associated factors
AIDS:Volume 22(3)30 January 2008p 395-402
Cazanave, Charlesa; Dupon, Michela; Lavignolle-Aurillac, Valeried; Barthe, Nicoleb; Lawson-Ayayi, Sylvied,g; Mehsen, Nadiac; Mercie, Patrickd,e; Morlat, Philliped,f; Thiebaut, Rodolphed; Dabis, Francoisd,g; for the Groupe d'Epidemiologie Clinique du SIDA en Aquitaine
From the aFederation de Maladies Infectieuses et Tropicales, France
bService de Medecine Nucleaire, France
cService de Rhumatologie, Hopital Pellegrin, France
dINSERM U593, ISPED, Universite Victor Segalen, France
eService de Medecine Interne et Maladies Tropicales, France
fService de Medecine Interne et Maladies Infectieuses, Hopital Saint-Andre, France
gCentre d'Information et de Soins de l'Immunodeficience Humaine, Bordeaux, France.
"....Among the 492 patients, representative of the Aquitaine cohort, who were recruited, the percentage of bone abnormalities was 80.5%, with 53.7% osteopenia and 26.8% osteoporosis.... Osteoporosis prevalence was high in spite of the young age....
....Osteoporosis was diagnosed in 132 patients (26.8%), 33.7% among men and 8.3% among women. Osteoporosis was diagnosed in 3.9% and 22.6% of premenopausal and menopausal women..... Among men, Independent factors associated with the diagnosis of osteoporosis were older age, homosexual HIV transmission, low BMI and HIV plasma viral load < 500 copies/ml. Only older age and lower BMI were marginally associated with osteopenia....In women, Older age [odds ratio (OR), 1.69; P = 0.02; and low CD4 cell count nadir (OR, 1.43; P = 0.008) were identified as factors associated with reduced BMD..... In consideration of the high frequency of diagnosed osteoporosis, it would be useful to propose osteodensitometry for HIV-infected patients in a targeted way: that is for patients with both the traditional risk factors of osteoporosis and the specific risk factors of HIV infection highlighted in this study...."
Abstract
Background: There is a high prevalence of bone demineralization among HIV-infected patients but mechanisms of alteration of bone turnover are still unclear and it is thought to be multifactorial.
Methods: A cross-sectional survey of 492 HIV-infected patients within the Aquitaine cohort estimated the prevalence of osteoporosis/osteopenia and investigated associated factors. Bone mineral density of total body, lumbar spine and femoral neck was measured by dual-energy X-ray absorptiometry. Multivariable analyses of the association with HIV disease status, treatment and anthropometric parameters were stratified according to gender.
Results:
Median age was 43 years (interquartile range, 38-50); 73% were male; 19.7% patients had reached AIDS, 93.1% were treated with HAART; and 28.5% had lipodystrophy.
Based on World Health Organization criteria, osteopenia was diagnosed in 54.6% of men [95% confidence interval (CI), 49.4-59.7) and 51.1% of women (95% CI, 42.6-59.6) and osteoporosis in 33.7% of men (95% CI, 28.8-38.6) and 8.3% of women (95% CI, 3.6-13.9).
Using a polytomous logistic regression, older age, homosexual transmission group, low body mass index and low HIV plasma viral load were associated with the diagnosis of bone abnormalities in men, whereas older age and low CD4 lymphocyte count nadir were independently associated with osteoporosis/osteopenia in women. The use of HAART was not related to osteoporosis after adjustment (P = 0.58).
Conclusions: This cohort-based survey showed a high prevalence of osteopenia and osteoporosis of multifactorial origin. Mechanisms and consequences of these bone disorders need to be investigated.
Introduction
Use of HAART has drastically improved the prognosis of patients infected with HIV [1]. Long-term HAART is associated with several metabolic and morphological complications, including lipodystrophy, insulin resistance, diabetes and dyslipidaemia [2]. Accelerated bone mass loss, osteopenia and osteoporosis, have recently been described in HIV-infected subjects, particularly in middle-aged men, with variable prevalence estimates [3,4] mostly a consequence of limited sample sizes and possible selection of patients. A recent meta-analytic review has shown a prevalence of osteoporosis of 15% in HIV-infected individuals, 3.7 times greater than in HIV-uninfected controls [5]. Underlying mechanisms leading to these complications are still unclear but it is thought to be a multifactorial process [3-13]. There have been few studies of factors considered to be possibly associated with bone metabolism in HIV-infected patients, including the long-term use of HAART, and results have been inconclusive [3,12]. Among the potential causes of osteoporosis, a direct effect of HIV upon osteogenic cells, the persistent activation of proinflammatory cytokines and alterations in the metabolism of vitamin D have been most often quoted, but all remain a matter of speculation [14-18].
The present study sought an accurate estimate of the prevalence of reduced bone mineral density (BMD) in a large cohort of French HIV-1 infected patients that had unrestricted enrolment, contained individuals with varying risk characteristics and of both sexes. The study also investigated systematically potential factors associated with such disorders.
Discussion
This cross-sectional survey within the Aquitaine cohort of HIV-infected patients was conducted to estimate the prevalence of BMD disorders and to investigate associated factors. Among the 492 patients, representative of the Aquitaine cohort, who were recruited, the percentage of bone abnormalities was 80.5%, with 53.7% osteopenia and 26.8% osteoporosis. Osteoporosis prevalence was high in spite of the young age of the population and reflected well the early demineralization problem faced by HIV-infected patients. This frequency is higher than that reported in other studies, which had prevalence rates ranging from 0% to 22% [11,13,18,22-31], with narrow boundaries.
Another interesting finding of this osteodensitometry-based survey is the site of bone demineralization. For men, there is preferentially cortical anatomic osteoporosis, whereas for women, there are trabecular abnormalities. Apart from HIV disease, male osteoporosis in normal populations has a cortical predominance, especially in secondary osteoporosis, whereas for women, the postmenopausal osteoporosis is classically trabecular, first involving the spine. The few studies that have described BMD abnormalities according to the anatomic site have been inconclusive [4,17,24].
The main strength of our study is the relatively large number of patients included and the number of pathological fractures (10.2%). No publication provides such an estimate and we could only identified some individual case reports [32]. The potential impact of BMD reduction on fracture risk remains unclear for HIV-infected patients. This may be because the majority are young, have few visual or balance problems and are not prone to falls. As the HIV-infected population gets older, clinicians may see an increase in the fracture rate.
The main limitation in the interpretation of our report is the absence of French references for the T-score for men. This could explain the surprising high male prevalence of BMD disorders detected in our group. There is a French reference of BMD values for women, but an American database had to be used for males in the three anatomic sites. Consequently, we may have overestimated the frequency of male osteoporotic events, because Americans have a diet that is richer in vitamin D than the French and so have a higher bone mineralization peak. By using an American reference, we are perhaps overinterpreting the BMD results from the whole body and the femoral neck. For males, we have compared the T-scores obtained for the spine here based on data in American databases with the T-scores that we would have obtained using French databases, in order to see if the use of a non-French reference database was important, and found no statistical significance between the two sets of T-scores (data not shown). The gender stratification, imposed by the difference in reference frame between men and women, led to a loss in statistical power of the study as it generated two weaker samples statistically. This element partly explains the low number of associated factors highlighted in multivariate analysis.
Among the factors found to be associated with BMD, some were expected, such as older age or lower BMI, but others are described for the first time, such as homosexual HIV transmission group, low HIV plasma RNA and low CD4 cell count nadir. The link with homosexual transmission group could indicate that this is a proxy for the abuse of substances that might be toxic for bone metabolism or coinfection, such as human herpes virus 8 and Kaposi disease. Our finding of higher risk of BMD disorder with low plasma viral load does not agree with some previously published reports [15,18], which hypothesized a potential role of the virus itself. In our study, low plasma viral load was the consequence of successful reduction of viral replication with antiretroviral drug exposure. This indicates, indirectly, a bone harmful effect of treatment itself; however, the investigations of the effect of the antiretroviral treatment, unadjusted for the plasma viral load, did not show a significant difference. Therefore, an effect on bone of antiretroviral drugs does not explain why a plasma viral load < 500 copies/ml arose as a risk factor for the osteoporotic men.
The potential effect of CD4 cell count nadir has been studied previously among HIV-positive women and was not recognized as a factor associated with weakening osteopathies [33]. Such a linkage in our study is an original finding. CD4 cell count nadir is directly linked to the severity and length of immunosuppression. This suppression will lead to immunological disorders and an increase in proinflammatory cytokines, inducing modifications of bone tissue metabolism at the beginning of an early demineralization. Another possible explanation of the effect of the CD4 cell nadir observed is that patients with a low CD4 cell nadir are those who have been treated for a longer period with antiretroviral drugs. This would also indirectly support a negative role for antiretroviral drugs on bone metabolism.
The cumulated exposure to PI drugs was significantly associated with bone abnormalities in univariate analysis, as seen by Nolan et al. [22] and Moore et al. [23], but this factor was not significant in the multivariate model. Finally, neither lipodystrophy nor reduced physical activity was associated with early demineralization.
In consideration of the high frequency of diagnosed osteoporosis, it would be useful to propose osteodensitometry for HIV-infected patients in a targeted way: that is for patients with both the traditional risk factors of osteoporosis and the specific risk factors of HIV infection highlighted in this study. Moreover, our analysis indicates a cortical prevalence of bone demineralization, especially for men; consequently, particularly vigilant is suggested for this risk, in particular for femoral neck fractures.
Ongoing studies will provide better knowledge of the physiopathological mechanisms at the initiation of early demineralization among HIV-infected patients, thus allowing better diagnostic, preventive and therapeutic evaluation to support care for this chronic infection in the coming years.
Results
Characteristics of the study sample
The study comprised 492 patients, who did not differ statistically from the 3182 patients followed actively in the Aquitaine Cohort in 2004-2005 according to sociodemographics and HIV characteristics (gender, median age, HIV transmission group, years since HIV diagnosis, AIDS stage, hepatitis B and C virus serological status, alcohol and tobacco consumption, CD4 cell count, plasma HIV RNA; data not shown). There were 359 men (73.0%); 31 of the 133 women were menopausal (23.3%). Median age was 43 years [interquartile range (IQR), 39-51] for men and 41 years (IQR, 38-46) for women (P = 0.01).
The median follow-up since the date of HIV infection diagnosis was 10.9 years (IQR, 5.9-15.3) for men and 11.9 years (IQR, 6.3-15.2) for women (P = 0.38). Transmission risk among men was predominantly homosexual (men who have sex with men; 57.1%) followed by heterosexual transmission (19.5%), whereas 69.9% of women were classified as heterosexual transmission followed by 18.1% by intravenous drug use. All patients combined, 97 (19.7%) had developed AIDS. Median plasma viral load was < 1.7 log copies/ml (IQR, < 1.7-3.0) (i.e., < 50 copies/ml); plasma viral load was < 500 copies/ml in 354 patients (72.0%). The median CD4 cell count was 459 cells/ƒÊl (IQR, 315-643); 45 patients had < 200 cells/ƒÊl. The median nadir CD4 cell count was 200 cells/ƒÊl (IQR, 100-300). The median plasma calcium concentration was 2.34 mmol/l (IQR, 2.27-2.41) and the median plasma phosphate was 1.06 mmol/l (IQR, 0.94-1.18). No patient had major kidney failure (creatinine clearance was > 70 ml/min for all patients). Chronic active hepatitis C (HCV RNA positive) occurred in 103 patients (21.6%); 36 (7.6%) had chronic hepatitis B (HBV surface antigen positive); and 10 (2.1%) had both. Daily alcohol consumption > 10 g was reported by 117 patients (23.8%), 13 (2.7%) of whom had an excessive consumption (> 30 g per day). Ever smoking tobacco (> 1 cigarette per day) was reported by 343 patients (69.7%) [265 men (73.9%) and 78 women (58.6%); P = 0.005)] of whom 181 (37.0%) smoked more than 15 pack-years. Of 423 patients with data available on calcium consumption, only 39 patients (7.9%) consumed more than 1 g calcium per day in their diet. Eighty-two patients (16.7%) did not report any physical activity (< 30 min/day); these patients were more likely to be unemployed and at the most advanced stages of HIV disease. At the time of the survey, 93.1% of the patients were taking antiretroviral drugs; 80.0% were treated with NRTI-based HAART: 37.0% with tenofovir, 28.7% with NNRTI and 52.0% with PI. The median cumulated duration on treatment was 71.8 months (IQR, 30.0-103.0) for NRTI, 7.4 months (IQR, 0.0-30.1) for NNRTI and 20.0 months (IQR, 0.0-52.0) for PI. Fifty patients (10.2%) presented at least one pathological fracture (i.e., for a low-energy traumatism), acquired before or after the diagnosis of HIV infection.
Data on BMI were available for 482 patients, 353 men and 129 women. BMI was < 19 kg/m2 in 57 of these patients (11.8%) and < 20.6 kg/m2 in 118 patients (24.5%) [comprising 68 of the 353 men (19.3%) and 50 of the 129 women (38.8%); P < 0.0001]. Clinically defined lipodystrophy was diagnosed in 140 patients (28.5%), 54 (11.0%) had lipoatrophy, 32 (6.5%) had lipoaccumulation and 54 (11.0%) had a mixed syndrome.
Statistically significant differences between men and women occurred for age (men were older), BMI (women had lower BMI) and tobacco consumption (men smoked more than women). There were no statistical differences on the other parameters described above.
Bone density
Based on WHO criteria, osteopenia was diagnosed in 264 patients (53.7%), 54.6% among men [95% confidence interval (CI), 49.4-59.7) and 51.1% among women (95% CI, 42.6-59.6). Osteopenia was diagnosed in 50.0% (95% CI, 40.0-60.0) and 54.8% (95% CI, 37.5-72.5) of premenopausal and menopausal women, respectively. Osteoporosis was diagnosed in 132 patients (26.8%), 33.7% among men (95% CI, 28.8-38.6) and 8.3% among women (95% CI, 3.6-13.0). Osteoporosis was diagnosed in 3.9% (95% CI, 0.2-7.8) and 22.6% (95% CI, 7.9-37.3) of premenopausal and menopausal women, respectively. Table 3 shows the distribution of median BMD according to gender, site and patients' diagnostic category. Osteoporosis predominated at the femoral neck for men (median BMD, 0.66 g/cm2) and for women (median BMD, 0.59 g/cm2).
Correlates of low bone density
Among men, the following factors significantly associated with the diagnosis of bone mass loss in the univariable model were included in the multivariable analysis: age, follow-up time since HIV diagnosis, transmission group, AIDS clinical stage, HIV plasma viral load < 500 copies/ml, log-based plasma viral load, tobacco consumption, physical activity, cumulative exposure to antiretroviral drug class, BMI < 20.6 kg/m2 and lipodystrophy.
Independent factors associated with the diagnosis of osteoporosis were older age, homosexual HIV transmission, low BMI and HIV plasma viral load < 500 copies/ml (Table 4). Only older age and lower BMI were marginally associated with osteopenia (Table 4).
In women, all bone disorders were pooled without distinction between osteopenia and osteoporosis because of the lower number of observations in each subgroup. Factors analysed in the multivariable model were menopausal status, age, follow-up time since HIV diagnosis, transmission group, AIDS clinical stage, HIV plasma viral load < 500 copies/ml, log-based plasma viral load zenith, CD4 lymphocyte count nadir, alcohol consumption, calcium intake > 1 g/day, physical activity, cumulative exposure to antiretroviral drug class and lipodystrophy. Older age [odds ratio (OR), 1.69; 95% CI, 1.10-2.60; P = 0.02] and low CD4 cell count nadir (OR, 1.43; 95% CI, 1.10-1.85; P = 0.008) were identified as factors associated with reduced BMD. To explore further whether the association between low HIV plasma RNA and osteoporosis in men was related to antiretroviral exposure, the effect of cumulative exposure to HAART was analysed without adjustment for HIV plasma RNA. Three different multivariable models were developed but in none was the treatment variable effect significant: cumulative exposure to any antiretroviral drug (OR, 1.01; 95% CI, 1.00-1.02; P = 0.07), cumulative exposure to HAART (OR, 1.02; 95% CI, 0.95-1.10; P = 0.58), and naive versus drug experienced status (OR, 0.28; 95% CI, 0.06-1.31; P = 0.11). Furthermore, this analysis was carried out for each drug class, again without being significant: cumulated NRTI (OR, 1.01; 95% CI, 1.00-1.02; P = 0.06), cumulated NNRTI (OR, 1.01; 95% CI, 1.00-1.03, P = 0.43) and cumulated PI (OR, 1.01; 95% CI, 1.00-1.03; P = 0.09).
Methods
Study population
A cross-sectional survey was carried out within the ANRS CO3 Aquitaine cohort. The Aquitaine cohort is an open and dynamic prospective hospital-based cohort of HIV-1-infected patients under routine clinical management in southwestern France [19]. It was initiated in 1987 in the Bordeaux University Hospital and four other public hospitals in this region by the Groupe d'Epidemiologie Clinique du Sida en Aquitaine (GECSA). Inclusion criteria for the cohort were adults attending in- or outpatient services of the participating hospitals, HIV-1 infection confirmed by Western blot, regardless of clinical stage, either at least one follow-up after the first clinic visit or with a known date of death, and having given informed consent.
Patients were included consecutively in the present study between November 2004 and May 2005. Patients were eligible if they were still alive and followed on 1 November 2004 and without chronic kidney failure (creatinine clearance < 70 ml/min), liver failure (prothrombin rate < 70%) or prolonged immobilization (> 30 days). Written informed consent was obtained from all participants for this specific study.
Patient characteristics registered included gender; age; HIV transmission group; date of HIV diagnosis and AIDS stage according to US Centers for Disease Control and Prevention (CDC) classification; hepatitis B and C virus serological status; alcohol and tobacco consumption; medication intake; and type and duration of specific antiretroviral classes used, including nucleoside and nucleotide reverse transcriptase inhibitors (NRTI), nonnucleoside reverse transcriptase inhibitors (NNRTI) and protease inhibitors (PI). All these variables were extracted from the cohort data base and checked with medical records. A specific questionnaire was filled during the first clinic visit during the study period to document body mass index (BMI), calcium consumption and physical activity (physical activity, professional or leisure, > 30 min/day was considered sufficient).
Bone mineral density assessment
Total mean BMD and total mean T-score of total body, lumbar spine and femoral neck were measured by dual-energy X-ray absorptiometry (DEXA; Hologic, Bedford, Connecticut, USA) by a single radiologist. The study equipment was registered in a French centralized quality control programme validated by the Groupe de Recherche et d'Information sur les Osteoporoses. This quality control included a daily phantom scan, allowing follow up of the stability of BMD over time, and the use of Shewhart rules and Cusum tests to monitor changes in scanner performances. The manufacturer checked the scanner regularly to verify precision and ensure examination reproducibility with a calibration error < 0.5%. The coefficient of variation of phantom BMD was evaluated at 0.42% for the study period. The T-score is the SD interpreted by comparison with the maximum value reached by young adults (30 years) of the same sex. Measurement was made on the total body to explore bone mass, fatty mass and lean mass, and on two specific anatomic sites: the femoral neck, which allows a quantitative evaluation of cortical bone tissue, and the lumbar spine L2-L4, which allows a quantitative evaluation of trabecular bone tissue.
The database used to assess T-scores for women was the French database (ISOS, OFELY and GENSET studies), validated by the Groupe de Recherche et d'Information sur les Osteoporoses. For men, reference curves for the total body and the femoral neck were lacking in France, and the American databases (TK curves) were used for the femoral neck, the lumbar spine and the total body [20].
The World Health Organization (WHO) classification was used for diagnosis purposes [21]. Osteopenia was defined as a T-score between -1 and -2.5 SD, and osteoporosis was defined as a T-score less than -2.5 SD.
Laboratory methods
HIV plasma RNA was measured by real-time PCR using the COBAS AmpliPrep/Cobas TaqMan HIV-1 Test (Roche Molecular Systems, Branchburg, New Jersey, USA), and CD4 cell counts were measured using flow cytometry.
Statistical analysis
Prevalence of bone abnormalities at the time of the study was estimated for each diagnostic category, dividing the number of patients fulfilling the above diagnostic criteria by the total number of patients screened. Data were stratified according to gender because the reference curves differed by gender (see above). A multivariable analysis using polytomous logistic regression evaluated factors associated with the presence of osteopenia or osteoporosis. Variables with P < 0.25 in univariable analyses were included in the full models. The final models were selected by using a stepwise descending procedure. Fits of final models were checked by Hosmer and Lemeshow ƒÔ2 test. Analyses were processed with the use of SAS software (SAS Institute, Cary, North Carolina, USA).
|
|
| |
| |
|
|
|
