Contraception and HIV-1 prevention among women in Africa: informing choices amid expanding options

Introduction Options for both contraception and HIV-1 prevention are expanding in sub-Saharan Africa, offering important opportunities to provide women with products they need to achieve their prevention goals. Decision-making in the face of many options is complex, and we aimed to generate evidence to inform women’s selection of contraception and HIV-1 prevention with the following specific aims: (1) evaluate the association between the increasingly-available copper intrauterine device (Cu-IUD) and a potential side effect, bacterial vaginosis (BV), and assess whether the association varies over time; (2.1) determine the association between reporting any receptive anal intercourse and HIV-1 risk and efficacy of a dapivirine intravaginal ring for HIV-1 prevention; (2.2) estimate the average per-exposure ring effect among women engaged in both vaginal and anal intercourse; and (3) develop and evaluate the predictive performance of age-specific HIV-1 risk scoring tools to optimize provision of HIV-1 prevention options to women at risk of HIV-1. Methods We used data from two large clinical trials conducted among women in sub-Saharan Africa, the ASPIRE trial of a dapivirine-containing intravaginal ring for HIV-1 prevention and the ECHO trial, which evaluated HIV-1 risk and contraceptive efficacy of three contraceptive methods. Primary statistical methods were proportional hazards models and a microsimulation mathematical model. Results and conclusions Risk of BV among Cu-IUD users: BV risk was 28% (95% CI: 12, 46) higher among Cu-IUD users than among women using no contraception or another non-hormonal method. Elevated BV risk persisted throughout Cu-IUD use and declined to pre-initiation levels within seven months to one year following discontinuation. Women and their providers may wish to consider BV risk when selecting contraception. Association between receptive anal intercourse (RAI) and HIV-1 risk and ring efficacy: RAI was not associated with HIV-1 acquisition (aHR: 0.93, 95% CI: 0.57, 1.54). The ring reduced HIV-1 risk by 25% (95% CI: -3, 55) among all women, by 27% (95% CI: -5, 49) among women reporting only vaginal intercourse, and by 18% (95% CI: -57, 57) among women reporting any RAI (interaction p-value=0.77), suggesting that RAI had a minimal impact on estimates of ring efficacy. Per-exposure ring effect among women engaged in both vaginal and anal intercourse: Among women with high adherence engaged in RAI for 6.3% of their acts (the median proportion among women engaged in both vaginal and anal intercourse), the per-exposure ring effect was 56% (IQR: 40, 63) and declined to 25% (IQR: 11, 37) among women for whom 30% of acts were RAI. The ring provides substantial overall risk reduction to the vast majority of women, supporting ring use for most women. HIV-1 predictive performance of age-specific risk scores: Both age-specific and non-age-specific risk scores had moderate HIV-1 predictive performance among women enrolled in a clinical trial in South Africa. Precision public health approaches for targeted PrEP provision in South Africa may require additional data to improve prediction. Availability of both the Cu-IUD and an intravaginal ring are important additions to help women achieve their contraception and HIV-1 prevention goals, as both have important advantages compared to currently available options. Counseling approaches to help women select the prevention methods that will best suit their values and preferences are needed as options expand, as well as complementary risk reduction strategies to mitigate the disadvantages of each method.