Share this article on:

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

doi: 10.1097/QAD.0b013e328351997f
Clinical Science

Objectives: Fractures are common and associated with multiple risk factors. We assessed the risks of fracture associated with time-dependent, differential antiretroviral drug exposures among a cohort of persons with HIV infection.

Design: Nested case–control study from an HIV cohort of 59 594 medically insured persons with HIV infection enrolled in a medical care between January 1997 and March 2008.

Methods: Cases were participants with a low-impact, nontraumatic fracture identified by ICD-9-CM codes; noncases were 1 : 4 matched and without fracture.

Results: Cases included 2477 persons with HIV infection with fractures, who were risk-set matched to 9144 persons with HIV infection without fractures. Exposure to antiretroviral therapy by drug class and by duration (any drug/class) was associated with reduced risk for fracture. Drug-specific antiretroviral exposures over time identified an increased risk for fracture associated with darunavir, delavirdine and saquinavir, whereas reduced risk was associated with efavirenz, emtricitabine, lamivudine, tenofovir, and zidovudine. An initial null risk became a reduced risk with increased duration for nevirapine. In a similar pattern, abacavir, didanosine, nelfinavir, ritonavir and stavudine were initially associated with increased risk for fracture, after which the risk became null with increased duration of exposure. Null or uncertain risk for fracture was associated with amprenavir, atazanavir, enfuvirtide, fosamprenavir, indinavir, lopinavir, tipranavir, and zalcitabine.

Conclusion: Our findings suggest an overall reduced risk for facture in persons treated versus not treated with antiretroviral drugs for HIV infection. Differential drug-specific exposure–response relationships for fracture will need to be further evaluated in other study populations.

aGlaxoSmithKline, Collegeville

bDepartment of Biostatistics, Center for Occupational Biostatistics and Epidemiology, Graduate School of Public Health, University of Pittsburgh, Pittsburgh, Pennsylvania

cCase Medical Center, Case Western Reserve University and University Hospitals, Cleveland, Ohio

dMedco Research Institute, Limited Liability Corporation, Medco Health Solutions, Inc., Bethesda, Maryland, USA.

Correspondence to Linda M. Mundy, MD, PhD, GlaxoSmithKline, 1250 South Collegeville Road, Collegeville, PA 19426, USA. Tel: +1 610 917 6923; e-mail:

Received 21 September, 2011

Revised 12 January, 2012

Accepted 19 January, 2012

© 2012 Lippincott Williams & Wilkins, Inc.