Predictors of concurrent sexual partnerships and association with recent HIV infection in a large population-based survey in Botswana

Abstract

BACKGROUND : Multiple concurrent sexual partnerships (MCP) may drive new HIV infections. We investigated the association between MCP and recent or incident HIV infection in a cluster-randomized HIV prevention trial that followed a population-based HIV incidence cohort across 30 communities in Botswana. METHODS : We used structured questionnaires to evaluate MCP for prior 12 months, defined as either (1) MCP per UNAIDS definition or (2) concurrent sexual relationship per survey questions. Recent HIV infection was determined using an avidity assay-based algorithm or seroconversion within 2 years, and incident infection was determined through annual HIV testing for up to 3 years. We estimated prevalence ratios (PRs) and 95% confidence intervals (CIs) for MCP predictors using univariate and multivariable Poisson regression with log-link and fixed effects for matched pairs. RESULTS : We included 11,965 (94.9%) of the 12,610 participants in the Botswana Combination Prevention Project with sexual history data. Among 9363 sexually active persons in prior 12 months, 2770 (29.6%) were engaged in MCP. Factors independently associated with MCP included male sex (aPR = 1.57; 95% CI: 1.45 to 1.71), age < 25 years (aPR = 1.25, 95% CI: 1.01 to 1.56), alcohol consumption ≥ 2 times/week (aPR = 1.38, 95% CI: 1.26 to 1.51), transactional sex (aPR = 1.69, 95% CI: 1.49 to 1.92), having a partner with MCP (aPR = 1.82, 95% CI: 1.65 to 2.02), and intergenerational sex (partners 10 years younger: aPR = 1.16, 95% CI: 1.06 to 1.28 or 10 years older: aPR = 1.32, 95% CI: 1.15 to 1.51). Reporting prior MCP was associated with HIV seroconversion during follow-up (aPR= 1.28, 95% CI: 1.05 to 1.57) but not with prevalent or recent HIV infection at baseline. CONCLUSIONS : MCP was common and associated with incident HIV infection. People reporting MCP may benefit from pre-exposure prophylaxis.