Send to:

Choose Destination
See comment in PubMed Commons below
AIDS. 2006 Nov 14;20(17):2165-74.

Antiretroviral therapy and the prevalence of osteopenia and osteoporosis: a meta-analytic review.

Author information

  • 1Johns Hopkins University, 1830 East Monument Street, Baltimore, MD 21287, USA.



Prevalence estimates of osteopenia and osteoporosis (reduced bone mineral density; BMD) in HIV-infected patients and the role of antiretroviral therapy (ART) varies in the literature.


We conducted a meta-analytical review of cross-sectional studies published in English to determine the pooled odds ratios (OR) of reduced BMD and osteoporosis in the following groups: HIV-positive versus HIV-negative; ART-treated versus ART-naive; protease inhibitor (PI)-treated versus PI-untreated. We searched the MEDLINE, PubMed, and EMBASE databases for eligible references between January 1966 and November 2005. Random effects models were used to generate pooled OR estimates and confidence intervals.


Of 37 articles identified, 20 met the inclusion criteria. Of the 884 HIV-infected patients, 67% had reduced BMD, of whom 15% had osteoporosis, yielding a pooled OR of 6.4 and 3.7, respectively, compared with HIV-uninfected controls (n = 654) using 11 studies with available data. Compared with ART-naive patients (n = 202, 10 studies), ART-treated individuals (n = 824) had a 2.5-fold increased odds of prevalent reduced BMD. The risk of prevalent osteoporosis (seven studies) was similarly elevated in ART-treated individuals. Compared with non-PI-treated HIV patients (n = 410, 14 studies), PI-treated patients (n = 791) had increased odds of reduced BMD and osteoporosis (12 studies). Few studies adjusted for important covariates such as HIV disease severity or treatment duration.


The prevalence of osteoporosis in HIV-infected individuals is more than three times greater compared with HIV-uninfected controls. ART-exposed and PI-exposed individuals had a higher prevalence of reduced BMD and osteoporosis compared with their respective controls. The influence of other disease and treatment variables on these estimates could not be determined.

[PubMed - indexed for MEDLINE]
PubMed Commons home

PubMed Commons

How to join PubMed Commons

    Supplemental Content

    Full text links

    Icon for Lippincott Williams & Wilkins Icon for PubMed Health
    Loading ...
    Write to the Help Desk