Background:  Half of HIV-1-infected adults in Sub-Saharan Africa have an uninfected partner. National serosurveys and mathematical modeling suggest as many as two-thirds of new HIV-1 infections in Africa occur within stable sexual partnerships. Little is known about sexual behaviors and risk of HIV-1 acquisition for HIV-1-uninfected partners in serodiscordant partnerships.

Methods:  We conducted a prospective study of 3380 HIV-1 serodiscordant partnerships (2283 with HIV-1-uninfected male partners, 1097 with HIV-1-uninfected female partners) from 7 countries in eastern and southern Africa. At quarterly visits for ≥24 months, HIV-1-uninfected partners were asked about sexual behavior during the prior month.

Results:  During follow-up, the proportion of HIV-1-uninfected partners reporting sex with their HIV-1-infected partner in the prior month declined significantly (from 93.5% at enrollment to 73.2% at 24 months, p <0.001) while the proportion reporting sex with an outside partner increased significantly (from 3.1 to 13.9%, p <0.001). Sex with the known HIV-1-infected partner and an outside partner in the same month was uncommon (3 to 5%) and stable throughout follow-up. HIV-1-uninfected men were significantly more likely than HIV-1-uninfected women to report partners other than their known HIV-1-infected partner (p <0.001). Sex unprotected with condoms was reported significantly more frequently with outside partners than with primary HIV-1-infected partners (RR 4.6; 95%CI 4.2 to 5.2). The proportion who acquired HIV-1 during study follow-up was similar for those who ever versus never reported an outside partner (3.1 vs 4.0%, HR 1.1, p = 0.6). However, based on HIV-1 sequencing, 18 of 21 (86%) of those who acquired HIV-1 and who reported outside partners were found to have an HIV-1 isolate that was virologically distinct from that of their primary partner, compared to 23 of 106 (22%) for those who reported no outside partners during follow-up (p <0.001).

Conclusions:  Among HIV-1 serodiscordant couples followed for 24 months, sexual frequency declined within the partnership as outside partnerships increased. Those with outside partners were no more likely to acquire HIV-1. In addition to ART for the HIV-1-infected partner to reduce infectiousness, prevention strategies for HIV-1-serodiscordant couples need to include strategies to reduce acquisition from outside partners, such as oral or topical pre-exposure prophylaxis.