Injectable and oral contraceptives and risk of HIV acquisition in women: an analysis of data from the MDP301 trial

STUDY QUESTION Do injectable and oral contraceptives increase the risk of human immunodeficiency virus (HIV) acquisition in women? SUMMARY ANSWER After adjusting for confounders, evidence of a significantly increased risk of HIV remained for women using injectable depo-medroxyprogesterone (DMPA) (hazard ratio = 1.49, 95% confidence interval (1.06-2.08)) but not for injectable norethisterone-enanthate (Net-En) or oral contraceptive pills (OC). WHAT IS KNOWN ALREADY An association between the use of some types of hormonal contraception (HC) methods and an increased risk of HIV, possibly through changes in the genital tract environment and alterations in the immune response, has been previously observed, although not consistently. A recent systematic review of these studies has highlighted the need for more definitive evidence. STUDY DESIGN, SIZE, DURATION A secondary data analysis of the MDP301 phase 3 microbicide trial was conducted to estimate the effects of use of different methods of HC on the risk of HIV acquisition in women. HIV-negative women (n = 8663) with a median age of 28 years were included in the analysis; 382 HIV seroconverted by 52 weeks follow-up; 10% of women-years were lost to follow-up before 52 weeks. PARTICIPANTS/MATERIALS, SETTING, METHODS Contraceptive use was reported at each 4-weekly visit. Cox proportional hazards (PH) models were used to estimate the effects of baseline and current use of injectable DMPA, injectable Net-En and OC compared with no HC, on the risk of HIV, adjusting for baseline and time-updated covariates. Causal effects for 52 weeks of HC use compared with no HC were estimated in a weighted Cox model, censoring women at deviation from baseline HC use (or non-use) or pregnancy. MAIN RESULTS AND THE ROLE OF CHANCE At baseline, 2499 (29%) women were on DMPA, 1180 (14%) on Net-En, and 1410 (16%) on OC; 3574 (40%) not on HC, started HC in follow-up. Adjusted hazard ratios (HR) for baseline HC use, compared with no HC, were 1.38 (95% confidence interval (CI) 1.07-1.78) for DMPA; 1.18 (0.86-1.62) for Net-En and 0.97 (0.68-1.38) for OC. The estimated causal effects of DMPA and Net-En over 52 weeks were: HR = 1.49 (95% CI 1.06-2.08) and HR = 1.31 (95% CI 0.62-1.61), respectively. LIMITATIONS, REASONS FOR CAUTION A main limitation of the study was that it was a secondary analysis of data from a study that was not designed to investigate this question. Despite our best efforts, we cannot exclude residual confounding to explain the effect of DMPA. WIDER IMPLICATIONS OF THE FINDINGS The results of this study should be reviewed by the World Health Organization to determine whether current recommendations on the use of DMPA in settings with high HIV prevalence require modification. STUDY FUNDING/COMPETING INTERESTS MDP is a partnership of African and European academic/government institutions with commercial organizations, which is funded by the UK Government (DFID and MRC), with support from IPM and EDCTP. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: None.