Baseline factors associated with incident HIV and STI in four microbicide trials

Analyzing pooled data from 4 recent microbicide trials, we aimed to determine characteristics of participants at higher risk of HIV and sexually transmitted infections (STIs), to inform targeted recruitment, preserved study power, and potentially smaller study sizes in future trials.

We evaluated the relationships between participants' characteristics and the incidence of HIV, STIs, and reproductive tract infections (RTIs). We calculated incidence rates as the number of infection events divided by the person-years of observation. We applied Cox regression models to assess the relationships between baseline demographic, reproductive and behavioral factors and incident HIV, STIs and RTIs.

The pooled incidence rates for HIV, chlamydia, and gonorrhea were 2.1, 6.4 and 9.9 per 100 person-years, respectively. Proportions of participants with trichomoniasis, bacterial vaginosis (BV), and candidiasis were 0.06, 0.40, and 0.40, respectively. In final multivariable models, age and education were significantly (and inversely) associated with incident HIV; baseline chlamydia, baseline trichomoniasis, and younger age were associated with incident Chlamydia; and baseline gonorrhea infection, younger age, less education, nulliparous status, baseline chlamydia, and condom use for contraception were associated with incident gonorrhea. Three factors were associated with trichomoniasis: baseline trichomoniasis infection, baseline chlamydia, and baseline BV.

Only younger age was robustly associated with multiple STI outcomes in our multivariable analyses. Although there was little evidence of associations between baseline STIs and incident HIV, they were strongly associated with incident STIs. We found no evidence that measured baseline sexual behavior factors were associated with incident HIV or STIs.