| Background: Vaginal infections are common, and associated with adverse health outcomes. Periodic presumptive treatment (PPT) is an effective intervention for preventing these infections. The PPT effect could be improved if it were restricted to women most likely to benefit from it.;Methods: We conducted a subgroup analysis of data from a PPT trial in Mombasa, Kenya using Andersen-Gill proportional hazards models with robust variance estimates to assess the effect of the intervention among participants with a baseline vaginal infection versus those without.;Results: Of 302 women eligible for analysis, 137 (45%) had one or more vaginal infections at baseline, while 165 (55%) had none. In the subgroup with a baseline vaginal infection, the intervention reduced the risk of incident BV [hazard ratio (HR)=0.55, 95% confidence interval (CI): 0.41-0.76] and increased vaginal colonization with any Lactobacillus species (HR=1.61, 95% CI: 1.01-2.56) as well as H2O2-producing Lactobacillus species (HR=1.85, 95% CI: 1.00-3.45) compared to placebo. In the subgroup without a baseline vaginal infection, the effect of the intervention to reduce incident BV (HR=0.71, 95% CI: 0.47-1.09) and increase vaginal colonization with any Lactobacillus species (HR=1.32, 95% CI: 0.85-2.04) as well as H2O2-producing Lactobacillus species (HR=1.48, 95% CI: 0.73-3.01), was less strong and not statistically significant.;Conclusion: In this PPT trial, the effect of the intervention was driven primarily by participants with a baseline vaginal infection. Restricting PPT to women who are most likely to benefit from it may improve the effect of the intervention. |
