Document Type


Publication Date





Introduction: The extraordinarily high incidence of HIV in sub-Saharan Africa led to the search for cofactor infections that could explain the high rates of transmission in the region. Genital inflammation and lesions caused by sexually transmitted infections (STIs) were a probable mechanism, and numerous observational studies indicated several STI cofactors. Nine out of the ten randomized controlled trials (RCTs), however, failed to demonstrate that treating STIs could lower HIV incidence. We evaluate all 10 trials to determine if their design permits the conclusion, widely believed, that STI treatment is ineffective in reducing HIV incidence.

Discussion: Examination of the trials reveals critical methodological problems sufficient to account for statistically insignificant outcomes in nine of the ten trials. Shortcomings of the trials include weak exposure contrast, confounding, non-differential misclassification, contamination and effect modification, all of which consistently bias the results toward the null. In any future STI-HIV trial, ethical considerations will again require weak exposure contrast. The complexity posed by HIV transmission in the genital microbial environment means that any future STI-HIV trial will face confounding, non-differential misclassification and effect modification. As a result, it is unlikely that additional trials would be able to answer the question of whether STI control reduces HIV incidence.

Conclusions: Shortcomings in published RCTs render invalid the conclusion that treating STIs and other cofactor infections is ineffective in HIV prevention. Meta-analyses of observational studies conclude that STIs can raise HIV transmission efficiency two- to fourfold. Health policy is always implemented under uncertainty. Given the known benefits of STI control, the irreparable harm from not treating STIs and the likely decline in HIV incidence resulting from STI control, it is appropriate to expand STI control programmes and to use funds earmarked for HIV prevention to finance those programmes.


Creative Commons Attribution 3.0 Unported (CC BY 3.0) License

Required Publisher's Statement

Original version is available from the publisher at: