Hormonal contraception and HIV acquisition: reanalysis using marginal structural modeling

Abstract

Hormonal contraceptives are used widely worldwide; their effect on HIV acquisition remains unresolved. We reanalyzed data from the Hormonal Contraception and HIV Study using marginal structural modeling to reduce selection bias due to time-dependent confounding. Replicating our original analysis closely, we found that depo-medroxyprogesterone acetate (DMPA) but not combined oral contraceptive (COC) was associated with increased HIV acquisition. Also, young (18-24 years) but not older women who used DMPA and COCs were at increased HIV risk.