Objectives: To compare the proportion of HIV seroconversions attributable to other sexually transmitted diseases in the intervention and comparison arms of the Mwanza sexually transmitted diseases (STD) intervention trial.
Design: Case-control study of 96 cases of HIV seroconversion and 974 HIV-negative controls, nested within the Mwanza trial cohort.
Methods: Data on reported STD symptoms during 2 years of follow-up, and serological evidence of recent syphilis, were used to obtain odds ratios (ORs) for HIV seroconversion, adjusted for community, age, marital status, sex partners and travel. Population-attributable fractions (PAF) of HIV seroconversions associated with these STD exposures were calculated separately for the intervention and comparison arms, and for men and women.
Results: In men in the comparison arm, adjusted ORs for ulcers (14.8), discharge (3.3), any symptom (4.1) and any STD (4.0) were highly significant. There were no significant associations between HIV incidence and STD exposures in the intervention arm. The PAF were consistently higher in the comparison arm than the intervention arm. In men, the PAF for any STD was 39.6% [95% confidence interval (CI), 12.4-58.3)] in the comparison arm but only 12.0% (CI, 0.0-35.9) in the intervention arm. The PAF for women were lower than for men.
Conclusions: These are minimal PAF estimates and they do not account for STD effects on HIV infectiousness. Nevertheless, a substantial proportion of new HIV infections in men in the comparison arm were attributable to STD. Lower PAF in the intervention arm than in the comparison arm for men provide further evidence of the role of STD cofactors in HIV transmission, supporting the hypothesis that the Mwanza intervention reduced the duration of symptomatic STD, thus reducing the HIV risk associated with such STD.