Reduced bone mineral density (BMD) is common in persons living with HIV. It is estimated that HIV-infected adults have over 3 times the risk for osteoporosis and almost 7 times the risk of osteopenia as their uninfected counterparts.1 Traditional risk factors such as tobacco use, age, and low body weight have been implicated in bone loss among HIV-infected individuals, as have direct effects of HIV infection, inflammation, and antiretroviral therapy.2–7 Initiation of highly active antiretroviral therapy (HAART) in treatment-naive patients has been consistently associated with a marked and clinically significant loss of BMD (2%–6%) independent of regimen.8–11 Most longitudinal studies show stable or increasing BMD on established HAART,3,4,12 although the SMART body composition substudy suggested that continuous antiretroviral use may be associated with decline in BMD and greater fracture risk relative to intermittent antiretroviral therapy.13
Some large cohort studies have found higher incidence of fractures in HIV-infected than in HIV-uninfected individuals,14–16 whereas others have not.17–19 In our previous analysis in the Women's Interagency HIV Study (WIHS), we found that fracture rates were similar in predominantly premenopausal HIV-infected and HIV-uninfected women [1.8/100 person-years (py) vs. 1.4/100 py, P = 0.13] with a median age of 40 years for HIV-infected women and 36 years for HIV-uninfected women at index and follow-up of 5.5 years.17 Given the concern for increased rates of fracture in older HIV-infected women, especially as they transition through menopause, we now compare rates and determinants of fracture in the same cohort of HIV-infected and HIV-uninfected women after 5 additional years of observation.
The WIHS, an ongoing multicenter prospective cohort study of the natural and treated history of HIV infection in women, enrolled 3766 women (2791 HIV-infected and 975 HIV-uninfected women) in 1994–1995 (n = 2623) and 2001–2002 (n = 1143) from 6 sites (Bronx/Manhattan, NY; Brooklyn, NY; Chicago, IL; Los Angeles, CA; San Francisco, CA; and Washington, DC). WIHS methods and baseline cohort characteristics have been described previously.20 HIV-uninfected women in the WIHS were recruited from groups known to have a high risk for HIV infection. The HIV-infected and HIV-uninfected recruits were comparable for a wide array of characteristics, including drug use, history of chronic illness, perceived health status, reproductive history, and income.20 At semiannual visits, participants complete physical examinations and provide biological specimens; they also undergo extensive assessment of clinical, behavioral, and demographic characteristics. Informed consent was obtained using procedures approved by committees on human research at each of the collaborating institutions.
In 2003 (visit 18), all WIHS participants were asked about personal history of fracture of the hip, wrist, and spine, both ever and within the prior 6 months. In those with a fracture history, fracture type(s) was determined, that is, fragility (resulting from fall from standing height or less) and nonfragility. In all subsequent semiannual study visits, participants were asked if they had fractures of the hip, wrist, spine, and/or other site since the last visit. Two investigators reviewed and adjudicated all fracture reports and excluded nonfractures. Of the 3766 participants enrolled, 2393 had at least 1 additional core visit after the index visit (visit 17, October 2002–March 2003), 18 participants who seroconverted during follow-up were excluded, and the remaining 1713 HIV-infected, and 662 HIV-uninfected participants were included in the analysis. Visit 17 was designated as the index visit because participant report of incident fractures was not collected before visit 17. We considered fragility fractures, nonfragility fractures, and all fractures (fragility and nonfragility combined) at any body site as study outcomes given recent data that high-trauma fractures are associated with lower BMD and are similar to fragility fractures in predicting future fractures.21 This analysis includes fracture data from semiannual observations from October 2002 (visit 17) to September 2013 (visit 38). Our prior publication included data up to September 2008 (visit 28)17; therefore, this analysis increases the observation time by approximately 5 years.
Demographics, medical history, HIV treatment history, and laboratory data were extracted from the WIHS database at the index visit and subsequent visits. Presence of previously identified prognostic factors for fracture22 at the index visit was quantified from the database: including age, menopause (no menses for greater than 1 year), personal history of fracture, parental history of hip fracture, body mass index (in kilograms per square meter), race/ethnicity, current use of cigarettes, greater than 2 alcoholic drinks per day, glucocorticoid use (ever/never), and history of rheumatoid arthritis or other cause of secondary osteoporosis. In addition, at the index visit, we obtained self-reported history of diabetes; current calcium use or multivitamin supplementation; any use of injection drugs (IDUs), opiates or cocaine, statins, hormonal contraception, or hormone replacement therapy (HRT); and estimated glomerular filtration rate was determined using the Modification of Diet in Renal Disease calculation.23 Chronic hepatitis C virus (HCV) infection status was determined by the presence of HCV seropositivity and detectable HCV RNA level. We also considered the following HIV-specific variables: current and nadir CD4+ lymphocyte counts, history of AIDS-defining illness (ADI), any HAART use at index visit, protease inhibitor (PI)-based HAART or nonnucleoside reverse transcriptase inhibitor (NNRTI)-based HAART at index visit, and cumulative HAART exposure as well as cumulative exposure by each antiretroviral class (total years of exposure to nucleoside reverse transcriptase inhibitors, NNRTIs, PIs, and tenofovir from initiation, whether before or after enrollment into WIHS, to either last visit or fracture event.
Means, medians, standard deviations, interquartile ranges, and proportions were derived to summarize study variables, depending on whether they were continuous or categorical. The primary outcome was time to self-reported fracture event, defined as any fracture (fragility or nonfragility) at any body site. Incidence rates were calculated as the number of participants with a new fracture divided by the total time (in person-years) at risk for new fracture during the study period. Visit 17 was considered the start time; the end of follow-up was either the visit at which the first fracture event was observed or, if no fracture was observed, the last visit before October 2013 for each participant. For women reporting multiple fractures, only the first fracture was included in these analyses.
Proportional hazards models were constructed to determine predictors of incidence of new fracture. Testing for HCV was performed at study entry, and cumulative ART exposure was calculated from time of ART initiation to time of fracture event or last visit; all other covariates were assessed at the index visit. Bivariate analyses evaluating each predictor in association with the outcome were considered for HIV-infected and HIV-uninfected participants together, and for HIV-infected participants separately. All variables were entered into multivariate models and a stepwise selection strategy with a P value of ≤0.05 was used to determine which variables remained in the final model. A P value <0.05 was considered statistically significant. A separate multivariate model, restricted to HIV-infected women, was constructed using similar methodology to evaluate the contribution of measures of HIV disease severity and effects of antiretroviral therapy on fracture. Crude and adjusted proportional hazards curves were also constructed. The Kalbfleisch and Prentice24 approach was used to estimate baseline hazards and was fit twice, with all variables (except HIV) set to their mean values and HIV infection status set to negative and positive.
Participant Characteristics at Index Visit
Table 1 presents participant characteristics at the index visit. HIV-infected women were older, had lower body weight, and more likely to be HCV infected, report postmenopausal status, and use HRT, statins, and calcium and vitamin D supplements. HIV-infected women were also less likely than uninfected women to report heavy alcohol consumption or current smoking.
Among the 1713 HIV-infected women at the index visit, 39% reported a prior ADI and 63% reported taking HAART; 49% of HAART regimens were PI based, 47% were NNRTI based, and 4% were neither PI based nor NNRTI based. Mean (±SD) duration of HAART exposure at the index visit (including self-reported HAART use before enrollment in WIHS) was 4.31 ± 3.4 years. Among the 37% of HIV-infected women not receiving antiretroviral therapy at the index visit, half were antiretroviral naive and half reported prior antiretroviral use.
During the study period, new fractures occurred in 390 of 2375 (16%) participants, with new fragility fractures occurring in 106/2375 (4%). At the time of fracture, the median ages of HIV-infected and HIV-uninfected women were 42 vs. 40 years (P = 0.07) and 57% vs. 47% (P = 0.09) were postmenopausal, respectively. Fractures occurred more commonly at the hip in HIV-infected than in HIV-uninfected women (2% vs. 1%, P = 0.06) but were similar at the spine and wrist. Fractures also occurred more commonly for HIV-infected women at other sites not associated with fragility fractures (14% vs. 11%, P = 0.08), including foot, toe, hand, finger, ankle, and knee (Table 2). Unadjusted incidence rates of fracture at any site were higher in HIV-infected women than in HIV-uninfected women (2.19/100 py vs. 1.54/100 py, P = 0.002) (Fig. 1, Table 2).
Among the 390 fractures, 106 (27%) were self-reported as fragility fractures. The number of fragility fractures did not differ in HIV-infected vs. HIV-uninfected women (5% vs. 4%, P = 0.22) (Table 2). Unadjusted incidence rates of fragility fractures were similar between HIV-infected and HIV-uninfected women (0.56/100 py vs. 0.39/100 py, P = 0.13) as were rates of fracture limited to the hip, spine, or wrist (Table 2).
Correlates of Incident Fracture
In bivariate analyses of all participants, traditional risk factors that were statistically associated with incident fracture were older age, self-report of menopause, history of fracture before index visit, HCV infection, reduced estimated glomerular filtration rate, diabetes mellitus, smoking, history of IDU, and cocaine or opiate use (Table 3). Use of statins, HRT, and calcium supplementation were also associated with new fracture. HIV infection was a significant predictor of incident fracture (HR = 1.45; 95% CI: 1.14 to 1.83). In multivariate analysis, HIV infection remained a statistically significant predictor of incident fracture, as were older age, white race, history of fracture before index visit, and history of IDU and cocaine use (Table 3).
Survival curves depicting incident fractures among HIV-infected and HIV-uninfected women are presented without adjustment (Fig. 1A). The rate of incident fracture among HIV-infected women was significantly higher than that for HIV-uninfected women (P = 0.002). After adjustment for significant covariates (age, race, fracture before index, cocaine use ever, and ID use ever), the HIV-infected group remained at greater risk for new fracture throughout the study period (P = 0.037) (Fig. 1B).
In bivariate analyses restricted to HIV-infected participants, older age, postmenopausal status, prior fracture, HCV infection, smoking, cocaine use, history of IDU, as well as CD4 nadir and history of AIDS were associated with incident fracture (Table 4). In contrast, race, body mass index, and HAART use at index or cumulative use, including tenofovir use, were not associated with incident fracture in bivariate analysis. In multivariate analyses restricted to HIV-infected participants, older age, white race, smoking, and history of fracture before index visit were associated with greater fracture incidence, as was prior ADI (adjusted hazard ratio (aHR) = 1.57; 95% CI: 1.24 to 1.99; P = 0.0002) (Table 4).
In our middle-aged cohort, HIV-infected women had higher incidence of fracture than HIV-uninfected women over a 10-year follow-up period. The difference remained statistically significant in multivariate models, in contrast to our earlier findings, where fracture rates were similar in HIV-infected and HIV-uninfected women.17 We also found that incidence of fragility fracture remained similar between HIV-infected and HIV-uninfected women; however, the proportionate increase in fragility fracture risk is similar to that seen with overall fracture incidence, suggesting that this may reflect a lack of power to detect differences in fragility fracture incidence in this cohort. Traditional risk factors for fracture including older age, white race, and history of fracture before index visit were significant predictors of incidence of new fracture, as was calcium supplementation, and history of IDU and cocaine. Compared with our previous study that only had 5 years of follow-up data, fractures at any site occurred in twice as many women (17.5% vs. 8.6%) and the incidence rate was 22% higher (2.19/100 py vs. 1.79/100 py).17 These differences in number and fracture incidence may reflect the higher age of the women under study, or a greater ability to detect differences between the groups because of increased observation time with 5 years of additional follow-up.
Our findings are similar to a number of other published studies that have found associations between HIV infection and fracture occurrence; however, many of these studies are limited by case–control design,25–27 use of general population data as an HIV-uninfected comparison group,15,19 or have been conducted in predominantly male subjects.16,27 In a population-based nationwide cohort study using Danish health registries, HIV-infected patients had a 50% increase in risk of fracture compared with population controls, and among those HIV-infected, this increased fracture risk was seen only for low-energy (osteoporotic) fracture in HAART-exposed persons.14 A Spanish population-based retrospective cohort study found a significant increase in overall fracture risk in HIV-infected compared with uninfected adults (aHR = 1.75; 95% CI: 1.24 to 2.48) and for hip fracture risk (aHR = 4.72; 95% CI: 2.35 to 9.47); however, when analyses were stratified by age <59 years vs. 60 and older, the risk of fracture associated with HIV status was significantly increased in the older group only, among whom HIV more than doubled fracture risk (HR = 2.11; 95% CI: 1.05 to 4.22).28 Our data suggest that fracture rates will increase over time in HIV-infected women as they age. In particular, rates may increase among nonblack women and in women with a history of substance use. This could be explained by a greater prevalence of reduced BMD in HIV-infected women combined with a differential increase in fall risk with age perhaps related to disorders that are more common in HIV-infected persons, such as multiple comorbidities, peripheral neuropathy, neurocognitive deficits, central nervous system effects of medications or illicit drugs, or functional impairment.29
In our previous report, we found that in addition to white race and renal insufficiency, HCV infection was associated with fracture17; however, in this report, HIV but not HCV remained associated with fracture in a multivariate model, after adjusting for race/ethnicity, and history of cocaine use. Our findings differ from several other cohort studies of HIV-infected individuals in which HCV remains a significant predictor of fracture even after adjustment for traditional risk factors including some measure of drug use.15,30 The finding that HCV is no longer significant after adjustment could mean that the association between HCV and fracture is due to true confounders or that given the multicollinearity in our variables, we cannot ascertain the true association between HCV and fracture. For example in our cohort, HCV-infected women were older and more likely to have used illicit drugs. Because both HIV and HCV infection are often associated with substance use, studies that investigate the relation between HCV and fracture must specifically consider the role of substance abuse, including not only opioid use but additionally cocaine and alcohol use, in risk of both falls and fracture. In addition, factors that influence fall risk such as muscle strength and balance can influence fracture rates independently of bone strength; therefore, models including independent assessments of fall risk and bone strength are necessary to determine the true relationships between HCV infection, and HIV infection and fracture.
We were surprised to find that a history of cocaine use remained strongly associated with fracture in our multivariate model. Although cocaine use has not been shown to have a direct effect on bone cells or bone metabolism, cocaine use may have indirect effects on nutritional/hormonal factors that might impact on bone mass or muscle strength. Additionally, cocaine use may lead to increased fall risk resulting in increased fracture. Interestingly, other variables associated with substance use (ID use, opiate use, alcohol) were not associated with fracture. It is also possible that cocaine use is a true confounder in this model because it is associated with both HCV status and older age. Without cocaine use in the model, age becomes a significant predictor but HCV is still not a significant predictor of fracture.
There are several strengths to our study. WIHS has a comparison group of HIV-uninfected women who were enrolled based on having similar risk behaviors as HIV-infected women, detailed data regarding fractures and fracture risk factors collected semiannually, and excellent participant retention. In addition, fracture estimates from our study may differ from others because of more complete ascertainment resulting from use of a dedicated fracture questionnaire at semiannual visits and the use of the same questionnaire in prospectively enrolled HIV-infected and HIV-uninfected controls. Several other studies have reported lower fracture incidence rates, perhaps in part be due to differences in methods of fracture ascertainment. In a French cohort of antiretroviral treated adults, fracture incidence was 0.33/100 py; however, only serious adverse events that were reported were included in analyses, specifically including only fractures leading to severe or complete limitation of activity and/or hospitalization.19 Similarly, a retrospective analysis of the ACTG A5001 [ACTG Longitudinal-Linked Randomized Trial (ALLRT)] database reported fracture incidence was 0.40/100 py among all participants; yet, because the parent studies included did not specify the reporting of fractures, fractures were likely underreported.31 The fracture incidence rates we observed in the WIHS cohort are comparable to those reported by Hansen et al14 in a population-based nationwide cohort study using Danish registries, with mean age 37.6% and 76% male sex, and an incidence rate of 2.10/100 py in the HIV-infected population and 1.35/100 py in the general Danish population. By linking data from the population-based Danish HIV Cohort Study with the Danish Civil Registration System and Danish National Hospital Registry, all fractures diagnosed at hospital admissions and at outpatient or emergency visits for both HIV-infected persons and matched general population controls were captured, and the authors note a high-positive predictive value of the registry diagnoses, which is as high as 93% for hip fracture. Limitations include collinearity of HCV with substance use variables and use of self-report for fracture ascertainment without radiographic confirmation or medical record verification. We also did not assess BMD, vitamin D levels, muscle strength, and balance, which are other potential determinants of falls and fracture. Because the majority of women in WIHS are African American and overweight, our results may not be generalizable to all HIV-infected women (although our enrolled cohort is representative of the HIV epidemic among US women).
In conclusion, after 10 years of median follow-up in predominantly premenopausal women of color, we found that HIV-infected women had higher fracture rates than HIV-uninfected women. History of cocaine use, but not HCV infection, remained predictive of fracture in multivariable analysis. Our data also suggest that fracture rates have increased in our HIV-infected women over the last 5 years as the cohort has aged. Further research may be warranted to delineate the mechanisms by which cocaine use could predispose women to fracture, and/or whether cocaine affects either bone strength or fall risk, but in the meantime, these data support optimizing musculoskeletal health in older HIV-infected women.
WIHS (Principal Investigators): UAB-MS WIHS (Michael Saag, Mirjam-Colette Kempf, and Deborah Konkle-Parker), U01-AI-103401; Atlanta WIHS (Ighovwerha Ofotokun and Gina Wingood), U01-AI-103408; Bronx WIHS (Kathryn Anastos), U01-AI-035004; Brooklyn WIHS (Howard Minkoff and Deborah Gustafson), U01-AI-031834; Chicago WIHS (Mardge Cohen), U01-AI-034993; Metropolitan Washington WIHS (Mary Young), U01-AI-034994; Miami WIHS (Margaret Fischl and Lisa Metsch), U01-AI-103397; UNC WIHS (Adaora Adimora), U01-AI-103390; Connie Wofsy Women's HIV Study, Northern California (Ruth Greenblatt, Bradley Aouizerat, and Phyllis Tien), U01-AI-034989; WIHS Data Management and Analysis Center (Stephen Gange and Elizabeth Golub), U01-AI-042590; Southern California WIHS (Alexandra Levine and Marek Nowicki), U01-HD-032632 (WIHS I–WIHS IV).
1. Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS. 2006;20:2165–2174.
2. Cazanave C, Dupon M, Lavignolle-Aurillac V, et al.. Reduced bone mineral density in HIV-infected patients: prevalence and associated factors. AIDS. 2008;22:395–402.
3. Dolan SE, Kanter JR, Grinspoon S. Longitudinal analysis of bone density in human immunodeficiency virus-infected women. J Clin Endocrinol Metab. 2006;91:2938–2945.
4. Mondy K, Yarasheski K, Powderly WG, et al.. Longitudinal evolution of bone mineral density and bone markers in human immunodeficiency virus-infected individuals. Clin Infect Dis. 2003;36:482–490.
5. Bolland MJ, Grey AB, Gamble GD, et al.. Clinical Review: low body weight mediates the relationship between HIV infection and low bone mineral density: a meta-analysis. J Clin Endocrinol Metab. 2007;92:4522–4528.
6. Fakruddin JM, Lawerence J. HIV-1 Vpr enhances production of activated NF-kappaB ligand (RANKL) via potentiation of glucocorticoid receptor activity. Arch Virol. 2005;150:67–78.
7. Gibellini D, de Crignis E, Ponti C, et al.. HIV-1 triggers apoptosis in primary osteoblasts and HOBIT cells through TNFalpha activation. J Med Virol. 2008;80:1507–1514.
8. Brown TT, McComsey GA, King MS, et al.. Loss of bone mineral density after antiretroviral therapy initiation, independent of antiretroviral regimen. J Acquir Immune Defic Syndr. 2009;51:554–561.
9. Cassetti I, Madruga JV, Suleiman JM, et al.; Study 903E Team*. The safety and efficacy of tenofovir DF in combination with lamivudine and efavirenz through 6 years in antiretroviral-naïve HIV-1-infected patients. HIV Clin Trials. 2007;8:164–172.
10. Duvivier C, Kolta S, Assoumou L, et al.. Greater decrease in bone mineral density with protease inhibitor regimens compared with nonnucleoside reverse transcriptase inhibitor regimens in HIV-1 infected naive patients. AIDS. 2009;23:817–824.
11. Finkelstein JS, Brockwell SE, Mehta V, et al.. Bone mineral density changes during the menopause transition in a multiethnic cohort of women. J Clin Endocrinol Metab. 2008;93:861–868.
12. Yin MT, Lu D, Cremers S, et al.. Short-term bone loss in HIV-infected premenopausal women. J Acquir Immune Defic Syndr. 2010;53:202–208.
13. Grund B, Peng G, Gibert CL, et al.. Continuous antiretroviral therapy decreases bone mineral density. AIDS. 2009;23:1519–1529.
14. Hansen AB, Gerstoft J, Kronborg G, et al.. Incidence of low and high-energy fractures in persons with and without HIV infection: a Danish population-based cohort study. AIDS. 2012;26:285–293.
15. Young B, Dao CN, Buchacz K, et al.; HIV Outpatient Study (HOPS) Investigators. Increased rates of bone fracture among HIV-infected persons in the HIV Outpatient Study (HOPS) compared with the US general population, 2000-2006. Clin Infect Dis. 2011;52:1061–1068.
16. Womack JA, Goulet JL, Gibert C, et al.; Veterans Aging Cohort Study Project Team. Increased risk of fragility fractures among HIV infected compared to uninfected male veterans. PLoS One. 2011;6:e17217.
17. Yin MT, Shi Q, Hoover DR, et al.. Fracture incidence in HIV-infected women: results from the Women's Interagency HIV Study. AIDS. 2010;24:2679–2686.
18. Arnsten JH, Freeman R, Howard AA, et al.. Decreased bone mineral density and increased fracture risk in aging men with or at risk for HIV infection. AIDS. 2007;21:617–623.
19. Collin F, Duval X, Le Moing V, et al.. Ten-year incidence and risk factors of bone fractures in a cohort of treated HIV1-infected adults. AIDS. 2009;23:1021–1024.
20. Bacon MC, von Wyl V, Alden C, et al.. The Women's Interagency HIV study: an observational cohort brings clinical sciences to the bench. Clin Diagn Lab Immunol. 2005;12:1013–1019.
21. Mackey DC, Lui LY, Cawthon PM, et al.. High-trauma fractures and low bone mineral density in older women and men. JAMA. 2007;298:2381–2388.
22. Kanis JA, McCloskey EV, Johansson H, Oden A, Strom O, Borgstrom F. Development and use of FRAX in osteoporosis. Osteoporosis Int. 2010;21(Suppl 2):S407–S413.
23. Levey AS, Bosch JP, Lewis JB, et al.. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130:461–470.
24. Kalbfleisch JD, Prentice RL. The Statistical Analysis of Failure Time Data. 2nd ed. Hoboken, NJ: Wiley; 2002.
25. Mundy LM, Youk AO, McComsey GA, et al.. Overall benefit of antiretroviral treatment on the risk of fracture in HIV: nested case-control analysis in a health-insured population. AIDS. 2012;26:1073–1082.
26. Prieto-Alhambra D, Güerri-Fernández R, De Vries F, et al.. HIV infection and its association with an excess risk of clinical fractures: a nationwide case-control study. J Acquir Immune Defic Syndr. 2014;66:90–95.
27. Yong MK, Elliott JH, Woolley IJ, et al.. Low CD4 count is associated with an increased risk of fragility fracture
in HIV-infected patients. J Acquir Immune Defic Syndr. 2011;57:205–210.
28. Güerri-Fernandez R, Vestergaard P, Carbonell C, et al.. HIV infection is strongly associated with hip fracture risk, independently of age, gender, and comorbidities: a population-based cohort study. J Bone Miner Res. 2013;28:1259–1263.
29. Erlandson KM, Allshouse AA, Jankowski CM, et al.. Risk factors for falls in HIV-infected persons. J Acquir Immune Defic Syndr. 2012;61:484–489.
30. Maalouf NM, Zhang S, Drechsler H, et al.. Hepatitis C co-infection and severity of liver disease as risk factors for osteoporotic fractures among HIV-infected patients. J Bone Miner Res. 2013;28:2577–2583.
31. Yin MT, Kendall MA, Wu X, et al.. Fractures after antiretroviral initiation. AIDS. 2012;26:2175–2184.