oa Variability in HIV infectiousness across Sub-Saharan Africa

Background & Objectives: A recent randomized clinical trial, CAPRISA 004, reported levels of HIV incidence rate among general population women in South Africa that are comparable or higher than those observed among high-risk populations such as female sex workers, men who have sex with men, or injecting drug users despite low coital frequency. Understanding the factors leading to such high HIV incidence levels is critical for guiding prevention efforts. To this end, we have assessed the risk of HIV transmission among stable couples across a range of epidemic settings in Sub-Saharan Africa (SSA). Methods: We constructed a mathematical model describing the process of HIV discordancy in a population in the context of an HIV epidemic to estimate the risk of HIV transmission within stable partnerships. The model was parameterized using nationally representative demographic health survey data for 20 countries in SSA. Uncertainty and sensitivity analyses were performed to explore the robustness of our estimates to systematic biases in model parameters. Results: HIV incidence rate among discordant couples hovered around 10 per 100 person-years across SSA. A clustering based on HIV prevalence was observed with a median of 7.8 per 100 person-years in low HIV prevalence countries (<5%) compared to 18.4 per 100 person-years in high prevalence countries (>5%). HIV population prevalence explained 49% of the variation in HIV transmission, with every 1% increase in HIV prevalence implying an increase of 1.2 per 100 person-years in HIV infectiousness. The uncertainty and sensitivity analyses suggested that potential systematic biases are more likely to lead to a slight overestimation in HIV incidence rate particularly in high prevalence countries. Conclusions: Our results suggest that a variability in HIV infectiousness may have contributed to the contrasting HIV epidemic trajectories across Africa. The key drivers of this variability may include male circumcision, different circulating virus subtypes, sexually transmitted co-infections, tropical co-infections, hormonal contraception, and host genetics and immunology. The role of behavioral cofactors such as coital frequency and the uptake of condom use are less evident.