Women in Sub-Saharan Africa are at disproportionate risk of HIV acquisition and have a high prevalence of hormonal contraceptive use including the injectable progestin, depot-medroxyprogesterone acetate (DMPA), and combined oral contraceptives (OCs). Understanding the impact of hormonal contraception on HIV acquisition is a critical unanswered public health question. We propose to conduct an individual participant data (IPD) meta-analysis of the relationship between hormonal contraception (HC) including DMPA and OCs and HIV acquisition. Specifically, we will evaluate whether or not these hormonal contraceptive methods increase the risk of HIV acquisition in women of reproductive ages overall and among important subgroups including young women (<25 years) and HSV-2 negative women. We will use data from eleven high-quality prospective studies comprising over 21,000 African women and 1,000 incident HIV infections drawn from the Vaginal Practices Research Partnership (VPRP). The VPRP has already compiled exposure and outcome data and important covariates in a standard format from each study. We will first do descriptive analyses to gain a detailed understanding of the data received within and between studies. We will use two-stage meta-analysis to assess heterogeneity between studies and to examine the overall association between HC exposures and HIV acquisition in each study individually using a common set of covariates. We will then conduct meta-regression to combine summary measures to provide an overall estimate of the HC effect on HIV acquisition. We will also use a one-stage meta-analysis approach that combines the IPD from all studies to perform a single analysis to compare with the results from the two-stage meta-analysis. The one-stage approach will apply multivariable Cox random effect models stratified by study that includes other prognostic factors to evaluate and summarize heterogeneous HC exposure effects on HIV acquisition. The IPD meta-analysis will allow us to examine the hormonal contraceptive-HIV acquisition relationship consistently across studies while maximizing study power and precision (particularly in subgroups of young and HSV-2 negative women). Additionally, we should be able to better control for potential confounding variables and assess and understand heterogeneity in effects across studies. The proposed meta-analysis will provide important information about the impact of hormonal contraception on HIV acquisition. It offers an important opportunity to fill the gap between scientific knowledge and the urgent need to provide women in high HIV prevalence areas with effective and safe contraception.
We propose to conduct an individual participant data meta-analysis of the relationship between hormonal contraceptive use - the injectable progestin DMPA and oral contraceptives - and HIV acquisition. Using data from eleven high-quality prospective studies in which over 21,000 African women were enrolled, we will use one and two-stage meta-analysis approaches to examine the association between hormonal contraceptive use and HIV acquisition overall and among young (<25 years) and HSV-2 negative women.