Background:  Low bone mineral density (BMD) is common among HIV⁺ individuals, but its clinical significance is unclear. BMD is relatively stable in premenopausal women on established antiretroviral therapy (ART) and fracture data are limited.

Methods:  We compared time to first fracture at any site (fragility and non-fragility) after a median follow-up of 5.4 years in 2391 (1728 HIV⁺, 663 HIV^‑) women enrolled in the Women’s Interagency HIV Study (WIHS), and determined risk factors for incident fracture. Self-report of fracture was recorded at semiannual visits using a specific questionnaire. Cox Proportional Hazards Models were utilized to evaluate association of traditional risk factors for fracture and HIV disease characteristics measured at baseline with subsequent incident fracture.

Results:  HIV⁺ individuals were older than HIV^‑ individuals (40+/-8 vs 36+/-10 years, P <0.0001) and had lower body mass index (28+/-7 vs 30+/-8 kg/m², P <0.0001); the proportion African American individuals was similar (56% vs 58%, P =0.30). HIV⁺ women were more often post-menopausal by self-report (20% vs 11%, P <0.0001), HCV-infected (25% vs 15%, P <0.0001), and taking vitamin D supplementation (42% vs 28%, P <0.0001), and less likely to be a current smoker (45% vs 51%, P =0.02) or use alcohol>2 drinks/day (2% vs 4%, P <0.0001). Personal history of previous fracture and serum creatinine levels did not differ by HIV status. Among HIV⁺, mean CD4 was 319+/-273 cells/μL; most were taking ART, with 30% on protease inhibitor-based and 30% on non-nucleoside reverse transcriptase inhibitor-based ART. Unadjusted fracture rates were similar between HIV⁺ and HIV^‑ for any fracture (1.7 vs 1.4/100 person-years, P =0.18) and for fractures of the hip, spine or wrist (0.6/100 person-years in both groups, P =0.96). In multivariate analysis, white (vs African American) race, menopause, and higher serum creatinine were associated with increased fracture rate, but HIV status was not. Among HIV⁺ women, traditional risk factors for fracture (white race, previous fracture, and menopause) and history of AIDS defining illness were associated with increased fracture rate, while CD4 and cumulative exposure to ART were not.

Conclusions:  After 5 years of follow up, fracture rates were not significantly increased in HIV⁺ women compared to HIV^‑ women. Traditional risk factors were important predictors, while HIV status was not. Our data provide some reassurance that fracture risk is modest in predominantly premenopausal HIV⁺ women, but further research is necessary to define risk after menopause.