OBJECTIVE: The use of antibiotics to prevent preterm birth has achieved mixed results. Our goal in this study was to determine if antibiotics given prenatally and during labor reduce the incidence of preterm birth and histologic chorioamnionitis. STUDY DESIGN: A double-blind randomized placebo-controlled trial of antibiotics to reduce preterm birth was conducted in 4 African sites. Both HIV-infected and uninfected pregnant women were given 2 courses of antibiotics, prenatally at 24 weeks (metronidazole 250 mg and erythromycin 250 mg tid orally for 7 days), and during labor (metronidazole 250 mg and ampicillin 500 mg q 4 hours) or identically appearing placebos. Two thousand ninety-eight HIV-infected and 335 HIV-uninfected women had evaluable end points, including gestational age determined by both obstetric and pediatric criteria and birth weight (BWT). Pre- and post-treatment rates of various sexually transmitted infections (STI) were determined and placentas were evaluated for histologic chorioamnionitis. RESULTS: Comparing antibiotic versus placebo treated HIV-infected and uninfected women, there were few differences in mean gestational age at delivery, the percent of preterm births, the time between randomization and delivery, or BWT. Four weeks after the 24-week antibiotic/placebo course, bacterial vaginosis, and trichomoniasis were reduced by 49% to 61% in the antibiotic groups compared with the placebo groups. However, in both the HIV-infected and uninfected groups, the placentas showed no difference in the rate of histologic chorioamnionitis. There were significant differences between HIV-infected and uninfected women, with the former having less education, a history of more stillbirths, more STIs, and in this pregnancy, a lower BWT (2949 vs 3100 g, P < .0001). CONCLUSION: Despite reducing the rate of vaginal infections, the antibiotic regimen used in this study did not reduce the rate of preterm birth, increase the time to delivery, or increase BWT. Failure of this regimen to reduce the rate of histologic chorioamnionitis may explain the reason the antibiotics failed to reduce preterm birth.