Secondary Logo

Journal Logo


Overall benefit of antiretroviral treatment on the risk of fracture in HIV

nested case–control analysis in a health-insured population

Mundy, Linda M.a; Youk, Ada O.b; McComsey, Grace A.c; Bowlin, Steve J.d

Author Information
doi: 10.1097/QAD.0b013e328351997f
  • Free



The use of antiretroviral treatment in HIV infection has clinical and economic benefits that offset potential untoward drug effects [1]. Associated risks for fracture and for reduced bone mineral density (BMD) among antiretroviral-exposed individuals remain understudied and drug-specific risks have been inconsistent to date among persons with HIV infection [2–9]. Among antiretroviral-naive patients, reduced BMD of 2–6% has consistently been observed within the first year of exposure [2,9–12]. Loss of BMD has been reported among participants randomized to continuous versus intermittent antiretroviral therapy for HIV infection [8]. Independent of HIV infection, factors that influence risk prediction for fracture among persons with HIV infection include the traditional risks associated with reduced BMD such as aging, whereas drug-specific antiretroviral exposures within combination regimens have not been fully explored [6,13–16]. Notably, reduced BMD due to osteoporosis is asymptomatic, and the most common adult clinical presentation of osteoporosis is a low-impact or nontraumatic fracture [2]. To date, quantifying the risk of fragility fractures has been best characterized in persons over 50 years of age [17]. Quantifying low-impact and nontraumatic fracture among persons with HIV infection is especially important, given that changes in BMD are not necessarily unidirectional, acquired HIV infection often occurs prior to 50 years of age, and long-term survival with HIV infection necessitates combination antiretroviral treatment. Findings from studies of reduced BMD and antiretroviral exposure have been inconsistent, and hence inconclusive, for drug-specific risks for facture among persons with HIV infection. We conducted a nested case–control study to assess risks for fracture in a health-insured population with HIV infection.


Study cohort

The cohort was included 59 584 persons with HIV infection identified in the Ingenix Impact National Benchmark Database (INBD), an administrative claims database. The INBD comprises over 74 million members from the United States, and was previously called the Integrated Health Care Information System. The adults with HIV infection were continuously enrolled, health-insured members for 12+ months in INBD between 1 January 1997 and 31 March 2008, had pharmacy-claims eligibility, and evidence of one or more ICD-9-CM diagnostic code of 042 or v08 for HIV infection. The INBD database was deidentified by the vendor in compliance with the Health Insurance Portability and Accountability Act (HIPAA); no additional institutional review board approval was required.

Nested case–control: study design and population

During the 11.25-year study period, there were 2477 participants with fractures identified in the cohort. For the nested case–control study, full risk sets were explicitly constructed from the cohort with age as the primary time dimension using the RISKSET program module in Occupational Cohort and Mortality Analysis Program (OCMAP-Plus) [18]. Risk sets were further reduced by matching on sex and on date of birth at age of fracture to control for cohort effects, and then combined into a single set with multiple cases when cases had the same event date. All participants with incident fracture were included as cases. We randomly selected four noncase individuals from the full risk sets matching on sex, exact age, and year of birth (within 5 years). Time-dependent antiretroviral exposure duration variables for cases and controls were computed as of the date of each event time at risk; in noncases this was the date that the noncase reached the age of the corresponding case at time of fracture.

Subset analysis of abacavir versus tenofovir

In a subset analysis of persons exposed to abacavir versus tenofovir, the dataset was censored to the time interval from 1 November 2001 through 31 March 2008 to align with the postapproval, launch, and availability of each of the two drugs. Similar design, study definitions (for case, noncase, covariates, and outcome), and analyses were conducted for this subset analysis.

Study definitions

Date of birth was restricted to year of birth given vendor compliance with HIPAA; hence, participants were each assigned July 15 for month/date to year of birth for the RISKSET program. Follow-up began on the first (index) date linked to an ICD-9-CM code for HIV infection. The enrollment interval prior to this index date was defined as surveillance and the time interval on and after this index date was defined as the study observation period. Severity of HIV infection was defined using criteria established for advanced HIV infection and for AIDS per the criteria established by the Centers for Disease Control and Prevention (CDC) [19].

Fracture, the primary study outcome, was defined by identification of the first ICD-9-CM code for closed, nontraumatic fracture in each participant's claim history during the study observation period; cases with more than one closed fracture code were individually assessed for potential trauma prior to inclusion as a case. Prior fracture was identified for each participant during the surveillance interval of INBD enrollment.

Prescription claims for drug exposure were identified by Uniform System of Classification Level 4 codes. Excess systemic glucocorticosteroid use was a bivariate response for a cumulative exposure to parenteral and oral glucocorticosteroid drugs equivalent to at least 675 mg of prednisone prior to either the date of fracture or the end of follow-up among participants without fracture [20]. Exposure to either vitamin D or calcium supplementation was a combined variable in the univariate analysis and models. Duration of exposure measures for the specific antiretroviral drugs were computed by restructuring the participant-level prescription history data for drug strength, metric quantity, and days supplied in order to quantify antiretroviral drug class, differential antiretroviral regimens, and drug-specific exposures.

The five categories of antiretroviral drug classes were nucleoside reverse transcriptase inhibitors (NRTI), non-NRTI (NNRTI), protease inhibitors, fusion inhibitors, and entry inhibitors. The duration of antiretroviral exposure measures were further categorized qualitatively (not exposed versus exposed) and quantitatively (in days) by antiretroviral drug class and by each antiretroviral drug.

Data analysis

We estimated odds ratios (ORs) and 95% confidence intervals (CIs) using exact conditional logistic regression programs in Stata (Stata, College Station, Texas, USA). The OR for each of the primary demographic, prescription history, and antiretroviral exposure variables were also adjusted for potential confounding factors if warranted. All antiretroviral drug exposure variables were categorized a priori into approximately equal exposure groups on the basis of the distribution of fractures in an attempt to balance the precision of the risk estimates across subgroups. Multivariate models were adjusted for prior fracture, excess alcohol use, low physical activity, low body weight, hepatitis C virus (HCV) infection, excess steroid use, and treatment for osteoporosis with bisphosphonates. We assessed the statistical significance of each main effect (expressed as a global P-value) with a likelihood ratio statistic and conducted tests for linear trend (expressed as a trend P-value) using equally spaced scores.


Study cohort

The study cohort included 59 584 persons with HIV infection enrolled in INBD over the 11.25-year study period. There were 30 405 persons (51%) who had one or more prescription claim for antiretroviral therapy during the study period (Table 1). Antiretroviral treatment was more frequent among men, more common in the interval from 2003 through 2008 (71.8%) than from 1997 through 2002 (28.8%), and more frequent among participants with advanced HIV infection, low body weight, lipodystrophy, hepatitis B virus coinfection, and HCV coinfection.

Table 1
Table 1:
Demographic and clinical characteristics of 59 584 persons with HIV infection stratified by exposure to antiretroviral drug treatment.

Nested case–control study population and univariate risks for fracture

The final population for the nested case–control study consisted of 11 621 participants, represented in 2286 risk sets with 2477 cases and 9144 noncases (Table 2). There were 157 risk sets with two cases, 15 risk sets with three cases, and one risk set with five cases. Risk for fracture was significantly higher, statistically, among participants with prior fracture (OR = 4.14, 95% CI = 3.28–5.23; P < 0.0001), low physical activity (OR = 2.15, 95% CI = 1.74–2.67, P < 0.0001), excess alcohol use (OR = 1.84, 95% CI = 1.47–2.29. P < 0.0001), low body weight (OR = 1.35, 95% CI = 1.17–1.55, P < 0.0001), HCV coinfection (OR = 1.28, 95% CI = 1.09–1.49, P = 0.002), and advanced HIV infection defined as CDC category B (OR = 1.24, 95% CI = 1.09–1.40; P = 0.001) and CDC category C (OR = 1.14, 95% CI = 1.03–1.27; P < 0.0001). Risk for fracture was statistically significantly reduced among participants exposed to antiretroviral drugs (OR = 0.64, 95% CI = 0.58–0.71; P < 0.0001), with a similar exposure–response relationship identified for antiretroviral drug class and for cumulative duration of exposure (Table 2).

Table 2
Table 2:
Univariate analysis of risk for fracture among 11 621 persons with HIV infection in a nested case–control study.

Antiretroviral drug class-specific exposure–response relationships for fracture

Further analysis of antiretroviral drug-specific exposure–response relationships revealed differential within-class risk estimates (Table 3). Reduced risk for fracture was associated with exposures to the NRTI and to the NNRTI drug classes, with a pattern of incremental reduction of risk with increased duration of exposure. Exposure to the protease inhibitor drugs was associated with a null effect that became slightly reduced in the subset of participants with the longest duration of exposure defined as 18 months or longer. Additionally, a null effect was noted for exposure to the fusion inhibitor in a small number of cases (N = 29) with fracture. There were no exposures to an entry inhibitor drug among cases with fracture.

Table 3
Table 3:
Antiretroviral drug class exposures (in months) and risk for fracture among 11 621 persons with HIV infection.

Antiretroviral drug-specific exposure–response relationships for fracture

Risk for fracture was further examined for antiretroviral drug-specific exposures assessed by approximate quartiles of duration of exposure (in months) and subsequent categorization of exposure–response relationships grouped as risks that were increased, decreased, null, or uncertain (Table 4). Increased risk for fracture was noted for darunavir, delavirdine and saquinavir with each quartile of exposure. Reduced risk for fracture was consistently noted for efavirenz, emtricitabine, lamivudine, tenofovir, and zidovudine. An initial increased risk became protective with increased duration for nevirapine. In a similar pattern, abacavir, didanosine, nelfinavir, ritonavir and stavudine were initially associated with a slightly increased point estimate of risk for fracture in the first quartile of exposure (shortest exposure), after which the risk became reduced with increased duration of exposure. Null or uncertain risk for fracture was associated with amprenavir, atazanavir, enfuvirtide, fosamprenavir, indinavir, lopinavir, tipranavir, and zalcitabine (Table 4).

Table 4-a
Table 4-a:
Antiretroviral drug-specific exposures categorized by associations of risk for fracture in a nested case–control study of 11 621 persons with HIV infection.
Table 4-b
Table 4-b:
Antiretroviral drug-specific exposures categorized by associations of risk for fracture in a nested case–control study of 11 621 persons with HIV infection.

Exposure–response relationships specific to abacavir versus tenofovir

To assess exposure–response relationships specific to abacavir and to tenofovir, a subanalysis was restricted to the 8879 cases and noncases enrolled in care on or after 1 November 2001 (Table 5). Reduced risk for fracture was noted in unadjusted and adjusted models for patients with antiretroviral regimens inclusive of abacavir (OR = 0.75, 95% CI = 0.64–0.88), exclusive of abacavir (OR = 0.61, 95% CI = 0.54–0.69), inclusive of tenofovir (OR = 0.63, 95% CI = 0.55–0.72), and exclusive of tenofovir (OR = 0.68, 95% CI = 0.59–0.78). The estimates of risk for abacavir were slightly increased but not statistically significant among participants with less than 6 months exposure (aOR = 1.12, 95% CI = 0.90–1.40) and with greater than 12 months exposure (aOR = 1.17 95% CI = 0.91–1.52), and reduced, but not statistically significant, for participants with 6–12 months of exposure (aOR = 0.87 95% CI = 0.64–1.17). The estimates of risk for tenofovir were reduced, but not statistically significant, for participants with less than 6 months exposure (aOR = 0.92, 95% CI = 0.76–1.10) and with 6–12 months of exposure (aOR = 0.84, 95% CI = 0.66–1.06), with slightly increased but not statistically significant risk for participants with 12 or more months of exposure (aOR = 1.08 95% CI = 0.83–1.40).

Table 5
Table 5:
Risk assessment for fracture in persons with exposure to abacavir, tenofovir, and other antiretroviral drugs for HIV infection enrolled in medical care between November 2001 and March 2008.


Our study identified an overall reduced risk for fracture in persons treated versus not treated with antiretroviral drugs for HIV infection. The study design and analysis highlights the complexity of estimating time-dependent, antiretroviral drug-specific risks for fracture among persons on combination antiretroviral regimens over different intervals of time for the treatment of HIV infection [21]. Given the known dynamic complexity of bone metabolism in aging populations with HIV infection and differential antiretroviral exposure, we emphasize three noteworthy findings from this nested case–control study.

First, exposures to antiretroviral treatment by drug class, duration, and most drug-specific exposures were associated with reduced risk for fracture among persons with HIV infection. Second, beyond antiretroviral drug class estimates of risk, the antiretroviral drug-specific risk estimates for fracture revealed three different risk categories, with increased risk for fracture identified for three drugs (Table 4). For darunavir, although the number of fracture events were low, the estimate of risk was substantial (aOR = 1.93, 95% CI = 1.05–3.56; (global P-value) Pg = 0.043), and further assessment of this protease inhibitor in other study populations seems justified. For delavirdine, the number of fracture events was low, the estimate of risk was moderate (aOR = 1.59, 95% CI = 0.94–2.71; Pg = 0.095), and future comparisons of risk in other populations may not be feasible given low prescribed use of delavirdine in current practice. For saquinivir, there were 115 exposed cases, and although the approximate quartiles of risk varied, the global and trend values were consistent with increased risk. With two of these three drugs in the protease inhibitor class, these data support some findings from two randomized trials that have identified an association of reduced BMD in the spine and protease inhibitor exposures [11,12]. The potential correlations of reduced BMD short-term versus clinical benefits and risks of various protease inhibitor-based regimens will require additional evaluation. Third, our time-censored subset analysis of abacavir versus tenofovir exposures revealed null risk in adjusted models for known risk variables and other antiretroviral-drug exposures (Table 5). Prior comparative analyses of abacavir versus tenofovir have been inconsistent, yet a significantly larger reduction in BMD has been consistently associated with tenofovir versus other NRTI agents [22–24]. Significant decreases in BMD were especially noted during the first year of exposure in patients randomized to abacavir/lamivudine and to tenofovir/emtricitabine, with either ritonavir-boosted atazanavir or efavirenz; reported fractures were all associated with trauma and not different between study arms [9].

Together, these drug-specific exposure–response relationships suggest an overall benefit of antiretroviral treatment relative to estimates of risk for fracture in participants without antiretroviral treatment. Although it is clinically relevant to discern distinctions in studies that assess changes in BMD versus fracture, our findings are consistent with several longitudinal studies of stable BMD in antiretroviral-exposed patients [5,7,24,25]. As BMD decreases with age, the mechanisms of bone modeling are not unidirectional and untreated HIV infection has been associated with uncoupled bone formation and bone resorption that seems attributed to both viral and inflammatory effects [26]. An undefined direct mechanism for bone remodeling is plausible in treated HIV infection and an indirect mechanism for bone remodeling is supported by lower levels of calcitrol (1,25-dihydroxyvitamin D) as reported in patients with advanced HIV infection compared to patients with early HIV infection [27]. Such changes in calcitrol may promote intestinal calcium absorption and regulation of osteoblast function. Although there is theoretical biological plausibility that increased fat-free mass and decreased inflammation evident via changes in C reactive protein, interleukin-6, soluble tumor necrosis factor receptor (sTNFR) I and II are linked to antiretroviral treatment and altered BMD, such measures are beyond the scope of an observational database analysis [11,27,28].

As noted in a recent study by the HIV Outpatient Study investigators, our findings are consistent with several established risk factors for fracture [29–33]. Despite consistencies in established risks among numerous studies, reported exposure–response relationships for antiretroviral treatment and fracture among persons with HIV infection have varied based on study design, execution, and analyses. Such distinctions are notable by population when restricted to either men or women, covariates such as age when assessed by year or by decade, and quantification of antiretroviral drug exposure when assessed as a dichotomous variable, drug class, or cumulative drug-specific exposure. As an example, a recent study among US women and a case–control study from Australia have each reported that tenofovir was not associated with fracture [30,33]. Given the inconsistencies in reported studies to date, additional investigations of exposure to protease inhibitors and to tenofovir remain important and clinically relevant for future investigations.

We acknowledge several limitations associated with the report of these study findings. First, as a retrospective study executed with administrative claims data, ascertainment bias for both information and measurement exists inclusive of the assumption that a prescription claim equated with administration of the antiretroviral drug as prescribed and the population is likely differential from randomized trial populations. We were unable to assess race (not available in INBD) or tobacco exposure (less than 5% identified in prestudy feasibility assessment), did not include a covariate for opiate use, testosterone use, or untreated hypogonadism, and likely underestimated the severity of advanced HIV infection or AIDS based on available CD4 cell counts and ICD-9-CM claims for opportunistic infections and malignancies. We were able to create claims-based variables for low body weight, low physical activity, and advanced HIV infection. Second, by design, we grouped all participants with low-impact and nontraumatic fracture as cases and did not estimate risks specific for the lumbar spine, hip, and other bone. As such, we are not able to associate with findings from studies that have had BMD loss as the outcome of interest. Third, our study population was restricted to fractures in adults and the findings are not able to be generalized to children or to longitudinal changes in BMD [34]. Fourth, almost all participants with antiretroviral exposure in our study were on NRTI-containing regimens, and we are unable to expand upon findings from a recent report of greater changes in BMD for participants randomized to NRTI-containing antiretroviral regimens versus NRTI-sparing regimens [12]. Lastly, we were unable to assess antiretroviral exposure prior to membership in the insured plan that was captured by INBD. It is nevertheless noteworthy that the study design, execution, and analysis combined the restructure of over 1.9 million antiretroviral prescription claims using occupational epidemiology methods to estimate drug-specific exposure–response relationships.


In summary, bone fractures are common and risks for fracture seem multifactorial among persons with HIV infection. We report significantly reduced risk of fracture in a dichotomous analysis of antiretroviral drug exposure, with a similar exposure–response relationship identified for antiretroviral drug class and for cumulative duration of exposure. Differential antiretroviral drug-specific risks for fracture suggest further assessment of individual drug exposures in other populations is warranted. Given the complexity of antiretroviral treatment for long-term survival among aging populations with HIV infection, our findings contribute evidence to and support for future robust analysis of observational cohort data to assess and identify modifiable risk factors associated with aging and drug exposure–response relationships. Such comparative research efforts will be integral to optimizing incremental net health benefits with tailored antiretroviral treatment in the years to come.


We wish to thank Samantha St. Laurent for assistance with the data harvest and data management, Michael Lann for programming assistance with restructure of the prescription histories for the antiretroviral drug claims, and Dr Keith Pappa for assistance with the antiretroviral drug exposure variables and critical study review.

Conflicts of interest

L.M.M. is a consultant to Worldwide Epidemiology, GlaxoSmithKline, Limited Liability Corporation. A.O.Y. received a grant from G.S.K. to perform the methodological and statistical work associated with this publication. G.A.M. has served as a scientific advisor or speaker for Bristol Myers Squibb, GlaxoSmithKline, Abbott, Tibotec, and Gilead Sciences. She has received research grants from Bristol Myers Squibb, GlaxoSmithKline, Abbott, Merck, and Gilead Sciences, and is currently serving as the DSMB Chair for a Pfizer-sponsored study. S.J.B. conducted this work while employed at GSK and owns GSK stock. S.J.B. is currently employed by Medco Health Solutions, Inc., and the opinions, assertions and/or findings contained in this study are those of the authors and are not to be construed as reflecting the views of Medco Health Solutions, Inc. or any of its affiliates.


1. Beck EJ, Harling G, Gerbase S, DeLay P. The cost of treatment and care for people living with HIV infection: implications of published studies, 1999–2008. Curr Opin HIV AIDS 2010; 5:215–224.
2. Mallon PWG. HIV and bone mineral density. Curr Opin Infect Dis 2010; 23:1–8.
3. Carr A, Miller J, Eisman JA, Cooper DA. Osteopenia in HIV-infected men: association with asymptomatic lactic acidemia and lower weight preantiretroviral therapy. AIDS 2001; 15:703–709.
4. Nolan D, Upton R, McKinnon E, John M, James I, Adler B, et al. Stable or increasing bone mineral density in HIV-infected patients treated with nelfinavir or indinavir. AIDS 2001; 15:1275–1280.
5. 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.
6. Brown TT, Qaqish RB. Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review. AIDS 2006; 20:2165–2174.
7. Bolland MJ, Grey AB, Gamble GD, Reid IR. 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.
8. Grund B, Peng G, Gibert CL, Hoy JF, Isaksson RL, Shlay JC, et al. Continuous antiretroviral therapy decreases bone mineral density. AIDS 2009; 23:1519–1529.
9. van Vonderen MG, Lips P, van Agtmael MA, Hassink EA, Brinkman K, Geerlings SE, et al. First line zidovudine/lamivudine/lopinavir/ritonavir leads to greater bone loss compared to nevirapine/lopinavir/ritonavir. AIDS 2009; 23:1367–1376.
10. Gallant JE, Staszewski S, Pozniak AL, DeJesus E, Suleiman JM, Miller MD, et al. Efficacy and safety of tenofovir DF vs. stavudine in combination therapy in antiretroviral-naive patients: a 3-year randomized trial. JAMA 2004; 292:191–201.
11. Brown TT, McComsey GA, King MS, Qaqish RB, Bernstein BM, da Silva BA. Loss of bone mineral density after antiretroviral therapy initiation, independent of antiretroviral regimen. J Acquir Immune Defic Syndr 2009; 51:554–561.
12. Duvivier C, Kolta S, Assoumou L, Ghosn J, Rozenberg S, Murphy LR, 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.
13. Triant VA, Brown TT, Lee H, Grinspoon SK. Fracture prevalence among human immunodeficiency virus (HIV)-infected versus non-HIV-infected patients in a large U.S. healthcare system. J Clin Endocrinol Metab 2009; 93:3499–3504.
14. Arnsten JH, Freeman R, Howard AA, Floris-Morre M, Santoro N, Schoenbaum EE. HIV infection and bone mineral density in middle-aged women. Clin Infect Dis 2006; 42:1014–1020.
15. Arnsten JH, Freeman R, Howard AA, Floris-Morre M, Lo Y, Klein RS. Decreased bone mineral density and increased fracture risk in aging men with or at risk for HIV infection. AIDS 2007; 21:617–623.
16. McComsey GA, Tebas P, Shane E, Yin MT, Overton ET, Huang JS, et al. Bone disease in HIV infection: a practical review and recommendations for HIV care providers. Clin Infect Dis 2010; 51:937–946.
17. Wasnich RD. Epidemiology of Osteoporosis in the United States of America. Osteoporos Int 1999; 7 (Suppl 3):S68–72.
18. Marsh G, Youk A, Stone R, Sefcik S, Alcorn C. OCMAP-PLUS: a new program for the comprehensive analysis of occupational cohort data. J Occup Environ Med 1998; 40:351–362.
19. Centers for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition of AIDS among adolescents and adults.MMWR 1992; 41(No. RR-17):1–19.
20. De Vries F, Bracke M, Leufkens HGM, Lammers JJ, Cooper C, van Staa TP. Fracture risk with intermittent high-dose oral glucocorticoid therapy. Arthritis Rheum 2007; 56:208–214.
21. Dollery CT. The challenge of complexity. Clin Pharmacol Ther 2010; 88:1315.
22. Cassetti I, Madruga JV, Suleiman JM, Etzel A, Zhong L, Cheng AK, et al. The safety and efficacy of tenofovir DR in combination with lamivudine and efavirenz through 6 years in antiretroviral-naïve HIV-1 infected patients. HIV Clinical Trials 2007; 8:164–172.
23. Stellbrink HJ, Orkin C, Arribas JR, Compston J, Gerstoft J, van Wijngaerden E, et al. Comparison of changes in bone density and turnover with abacavir-lamuvidine versus tenofovier-emtricitabine in HIV-infected adults: 48-week results from the ASSERT study. Clin Infect Dis 2010; 51:963–972.
24. McComsey GA, Kitch D, Daar ES, Tierney C, Jahed NC, Tebas P, et al. Bone mineral density and fractures in antiretroviral-naive persons randomized to receive abacavir-lamivudine or tenofovir disoproxil fumarate-emtricitabine along with efavirenz or atazanavir-ritonavir: AIDS Clinical Trials Group A5224, a substudy of ACTG A5202. J Infect Dis 2011; 203:1791–1801.
25. Dube MP, Qian D, Edmondson-Melancon H, Sattler FR, Goodwin D, Martinez C, et al. Prospective, intensive study of metabolic changes associated with 48 weeks of amprenavir-based antiretroviral therapy. Clin Infect Dis 2002; 35:475–481.
26. Yin MT, Lu D, Cremers S, Tien PC, Cohen MH, Shi Q, et al. Short-term bone loss in HIV-infected premenopausal women. J Acquir Immune Defic Syndr 2010; 54:e5.
27. Brown TT, Tassipoulos K, Bosch RJ, Shikuma C, McComsey GA. Association between systemic inflammation and incident diabetes in HIV-infected patients after initiation of antiretroviral therapy. Diab Care 2010; 33:2244–2249.
28. Ross AC, Armentrout R, O’Riordan MA, Storer N, Rizk N, Harrill D, et al. Endothelial activation markers are linked to HIV status and are independent of antiretroviral therapy and lipoatrophy. JAcquir Immune Defic Syndr 2008; 49:499–506.
29. Collin F, Duval X, Le Moing, Piroth L, Al Kaied F, Massip P, et al. Ten-year incidence and risk factors of bone fracture in a cohort of treated HIV1-infected adults. AIDS 2009; 23:1021–1024.
30. Yin MT, Shi Q, Hoover DR, Anastos K, Sharma A, Young M, et al. Fracture incidence in HIV-infected women: results from the Women's Interagency HIV Study. AIDS 2010; 24:2679–2686.
31. Womack JA, Goulet JL, Gibert C, Brandt C, Chou Chang C, Gulanski B, et al. Increased risk of fragility fractures among HIV infected compared to uninfected male veterans. PLoS One 2011; 6:e17217.
32. Young B, Dao CN, Buchacz K, Baker R, Brooks JT. HIV Outpatient Study (HOPS) InvestigatorsIncreased 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.
33. Yong MK, Elliott JH, Woolley IJ, Hoy JF. 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.
34. Mora S, Zamproni I, Beccio S, Bianchi R, Giacome V, Vigano A. Longitudinal changes of bone mineral density and metabolism in antiretroviral-treated HIV-infected children. J Clin Endocrinol Metab 2004; 89:24–28.

antiretroviral drug; bone mineral density; fracture; HIV; risk

© 2012 Lippincott Williams & Wilkins, Inc.