BACKGROUND: Single-dose nevirapine given to women and infants reduces mother-to-child HIV transmission, but nevirapine resistance develops in a large percentage of women. OBJECTIVE: To determine whether the maternal nevirapine dose could be eliminated in the setting of zidovudine prophylaxis. DESIGN, SETTING, AND PARTICIPANTS: A 2 x 2 factorial, randomized, clinical trial, with a double-blinded peripartum factor designed to assess the equivalence of maternal single-dose nevirapine versus placebo with respect to HIV transmission. A total of 709 HIV-infected pregnant women were randomized from four district hospitals in Botswana, resulting in 694 live first-born infants. HAART was available for women with AIDS. INTERVENTION: All women received a background of zidovudine from 34 weeks' gestation through delivery, and all infants received single-dose nevirapine at birth and zidovudine from birth through 1 month. Women were randomized to receive either single-dose nevirapine or placebo during labor. MAIN OUTCOME MEASURES: The primary endpoint was infant HIV infection by the 1-month visit. RESULTS: Of the 694 infants in this equivalence study, 15 (4.3%) of 345 in the maternal nevirapine arm were HIV infected by 1 month, versus 13 (3.7%) of 349 in the maternal placebo arm (95% confidence interval for difference, -2.4% to 3.8%), meeting pre-determined equivalence criteria. Nevirapine resistance at 1 month postpartum was detected in 45% of a random sample of women who received nevirapine. CONCLUSIONS: In the setting of maternal zidovudine and infant zidovudine plus single-dose nevirapine, infant HIV infection rates were similar whether women received single-dose nevirapine or placebo. This strategy avoids the potential for maternal nevirapine resistance.