HIV and STI among women in Uganda, Zimbabwe and Thailand: associations with male circumcision and changes in condom use

The results of most HIV-prevention programs over the 25-year history of the AIDS epidemic have been disappointing. Two interventions have been notable exceptions in the string of prevention failures: in 1983, researchers reported that consistent use of male condoms reduced risk of HIV transmission, and much more recently, in February 2007, large-scale randomized trials determined that circumcision reduced men's risk of HIV acquisition by 40-65%. These separate interventions - male circumcision and condom use - are the focus of these dissertation analyses. We examined whether the circumcision status of women's primary sexual partner was associated with her risk of HIV and three sexually transmitted infections (STIs): Chlamydia trachomatis (Ct), Neisseria gonorrhoeae (GC), and Trichomonas vaginalis (Tv). We used data from a prospective cohort study on hormonal contraception and incident HIV and STI (HC-HIV study) among women from Uganda, Zimbabwe and Thailand (HIV analyses included 4,417 Ugandan and Zimbabwean women; STI analyses included 5,925 women from Uganda, Zimbabwe and Thailand). After adjustment, women with circumcised partners had similar risk to women with uncircumcised partners for HIV (hazard ratio (HR): 1.03, 95% confidence interval (CI): 0.69-1.53), Ct (HR: 1.22, 95% CI: 0.94-1.59), GC (HR: 0.93, 95% CI: 0.70-1.24), and Tv (HR: 1.05, 95% CI: 0.81-1.37). Among women who became HIV-infected during HC-HIV, we also examined whether HIV diagnosis, together with counseling and free condoms, was sufficient to induce changes in women's condom use over both short (2-6 months) and longer time periods (12-16 months). After diagnosis, the number of HIV-infected women reporting any unprotected acts in a typical month declined significantly (short-term: from 72% to 56%; long-term: from 74% to 56%). After adjustment, among women reporting any unprotected acts, HIV-infected women also reduced the number of unprotected acts by 29% (short term) and 38% (long term). When assessing the proportion of acts where male condoms were used, however, women had no reduction over time. Circumcision was not associated with women's risk of HIV, Ct, Tv or GC among most participants. HIV-infected women reduced their overall number of unprotected sex acts, but the proportion of unprotected acts was unchanged from pre-infection behavior