Incident HIV Infection Among Young Men Associated With Female Sexual Partner Types Identified Through Latent Class Analysis, Rakai, Uganda.

Abstract

Sexual partner characteristics are important determinants of HIV acquisition, but little is known about partner types of young men in sub-Saharan Africa. Sexually active men aged 15-24 years from 5 rounds (2005-2013) of the Rakai Community Cohort Study in Uganda reported characteristics of up to 4 past-year female partners. Partner types were identified using latent class analysis. HIV incidence rates (IRs) were calculated by partner-type combinations, and individual-level risk adjusted IR ratios (aIRRs) relative to the lowest incidence type were estimated using the Poisson regression with generalized estimating equations. Young men (N = 1771) reported 4539 past-year female sexual partners. Three partner types were identified: type A: noncohabiting, student, medium duration partnerships; type B: cohabiting, nonstudent, longer duration partnerships; and type C: noncohabiting, nonstudent shorter duration partnerships. Type C partners engaged in the most HIV-related risk behaviors. Many men (29%) had more than 1 partner type/round. IR overall was 9.8/1000 person-years [95% confidence interval (CI): 4.7 to 20.6]. IR was 4.0 (95% CI: 1.2 to 12.7) for men with type A partners alone (41% of men). Relative to them, IR for those with type B partners alone (25%) was not significantly different. Men with type C partners alone (5%) had higher risk (aIRR = 3.2; 95% CI: 1.0 to 9.9), as did men with >1 partner type, including men with both type A and type B partners (12%; aIRR = 6.3; 95% CI: 2.5 to 15.9) and men with type C and other partner types (17%; aIRR = 4.3; 95% CI: 1.7 to 10.8). Partner-type combination was strongly associated with HIV incidence; type C partners and having more than 1 partner type were the riskiest patterns.