RESUMO
BACKGROUND: During pregnancy and postpartum period, the sexual behaviors of women and their partners change in ways that may either increase or reduce HIV risks. Pregnant women are a priority population for reducing both horizontal and vertical HIV transmission. SETTING: Nine sites in 4 South African provinces. METHODS: Women aged 18-30 years were randomized to receive pericoital tenofovir 1% gel or placebo gel and required to use reliable modern contraception. We compared HIV incidence in women before, during, and after pregnancy and used multivariate Cox Proportional hazards models to compare HIV incidence by pregnancy status. RESULTS: Rates of pregnancy were 7.1 per 100 woman-years (95% confidence interval [CI]: 6.3 to 8.1) and highest in those who reported oral contraceptive use (25.1 per 100 woman-years; adjusted hazard ratio 22.97 higher than other women; 95% CI: 5.0 to 105.4) or had 2 children. Birth outcomes were similar between trial arms, with 59.8% having full-term live births. No difference was detected in incident HIV during pregnancy compared with nonpregnant women (2.1 versus 4.3%; hazard ratio = 0.56, 95% CI: 0.14 to 2.26). Sexual activity was low in pregnancy and the early postpartum period, as was consistent condom use. CONCLUSIONS: Pregnancy incidence was high despite trial participation being contingent on contraceptive use. We found no evidence that rates of HIV acquisition were elevated in pregnancy when compared with those in nonpregnant women. Risks from reductions in condom use may be offset by reduced sexual activity. Nevertheless, high HIV incidence in both pregnant and nonpregnant women supports consideration of introducing antiretroviral-containing pre-exposure prophylaxis for pregnant and nonpregnant women in high HIV prevalence settings.