Although an association between microbial invasion of amniotic cavity and preterm birth has been extensively demonstrated, there is conflicting evidence regarding the benefits of antibiotic therapy in patients with preterm labor and intact membranes. We attempted to assess the efficacy of amoxicillin and erythromycin on pregnancy outcome in those patients. A randomized, double-blinded, placebo-controlled trial was designed and implemented. A total of 196 patients with singleton pregnancies and preterm labor with intact membranes (22-36 weeks) were randomly allocated to receive either antibiotics or placebo, plus adjunctive parenteral tocolysis, and 173 patients (antibiotics group n = 83 vs. placebo group n = 90) completed the treatment. The overall prevalence of microbial invasion of the amniotic cavity was 5.2% (9/173). No significant difference between both groups was found in maternal outcomes, including duration of randomization-to-delivery interval, frequency of preterm delivery, and frequency of clinical chorioamnionitis and endometritis. Rate of cesarean section was significantly higher in the placebo group (28% vs. 12%). Regarding neonatal outcome, no significant difference was detected between both groups in neonatal death, respiratory distress syndrome, proven sepsis, and birthweight. Suspected sepsis was significantly more frequent in the placebo group (6/90 vs. 0/78). The results of this trial indicate that amoxicillin and erythromycin do not prolong pregnancy in patients with preterm labor and intact membranes. A significant reduction in the rate of cesarean section was observed in patients receiving antibiotics. A significant reduction in the rate of neonatal suspected sepsis was also demonstrated.