BACKGROUND: Long-term impact of antiretroviral therapy (ART) on sexual HIV-transmission risk in Africa is unknown. We assessed sexual behavior changes and estimated HIV transmission from HIV-infected adults on ART in Uganda. METHODS: Between 2003 and 2007, we enrolled and followed ART-naive HIV-infected adults in a home-based AIDS program with annual counseling and testing for cohabitating partners, participant transmission risk-reduction plans, condom distribution and prevention support for cohabitating discordant couples. We assessed participants' HIV plasma viral load and partner-specific sexual behaviors. We defined risky sex as intercourse with inconsistent/no condom use with HIV-negative or unknown serostatus partners in previous 3 months. We compared rates using Poisson regression models, estimated transmission risk using established viral load-specific transmission estimates, and documented sero-conversion rates among HIV-discordant couples. RESULTS: Of 928 participants, 755 (81%) had 36 months data: 94 (10%) died and 79 (9%) missing data. Sexual activity increased from 28% (baseline) to 41% [36 months (P < 0.001)]. Of sexually active participants, 22% reported risky sex at baseline, 8% at 6 months (P < 0.001), and 14% at 36 months (P = 0.018). Median viral load among those reporting risky sex was 122 500 [interquartile range (IQR) 45 100-353 000] copies/ml pre-ART at baseline and undetectable at follow-up. One sero-conversion occurred among 62 cohabitating sero-discordant partners (0.5 sero-conversions/100 person-years). At 36 months, consistent condom use was 74% with discordant partners, 55% with unknown and 46% with concordant partners. Estimated HIV transmission risk reduced 91%, from 47.3 to 4.2/1000 person-years. CONCLUSIONS: Despite increased sexual activity among HIV-infected Ugandans over 3 years on ART, risky sex and estimated risk of HIV transmission remained lower than baseline levels. Integrated prevention programs could reduce HIV transmission in Africa.

This paper examines HIV risk behavior among HIV-uninfected adults living with people taking antiretroviral therapy (ART) in Uganda. A prospective cohort of 455 HIV-uninfected non-spousal household members of ART patients receiving home-based AIDS care was enrolled. Sexual behavior, HIV risk perceptions, AIDS-related anxiety, and the perception that AIDS is curable were assessed at baseline, 6, 12 and 24 months. Generalized linear mixture models were used to model risk behavior over time and to identify behavioral correlates. Overall, risky sex decreased from 29% at baseline to 15% at 24-months. Among women, risky sex decreased from 31% at baseline to 10% at 6 months and 15% at 24 months. Among men, risky sex decreased from 30% at baseline to 8% at 6 months and 13% at 24 months. Perceiving HIV/AIDS as curable and lower AIDS-related anxiety were independently associated with risky sex. No evidence of behavioral disinhibition was observed. Concerns regarding behavioral disinhibition should not slow down efforts to increase ART access in Africa.