Objective: We hypothesized that upper genital tract microbial infection associated with spontaneous preterm birth may precede conception. Our objective was to estimate if antibiotic administration during the interpregnancy interval in nonpregnant women with a previous preterm birth before 34 weeks' gestational age would reduce the rate of spontaneous preterm birth in the subsequent pregnancy. Study design: Women with a spontaneous preterm birth <34 weeks' gestational age were randomized at 4 months' postpartum to receive oral azithromycin 1 g twice (4 days apart) plus sustained-release metronidazole 750 mg daily for 7 days, or identical-appearing placebos. This regimen was repeated every 4 months until the subsequent pregnancy. Results: A total of 241 women were randomized; 124 conceived a subsequent pregnancy and were available for study, including 59 in the antibiotic group and 65 in the placebo group. In the antibiotic versus placebo group, neither subsequent spontaneous preterm birth (<37 weeks: 52% vs 46%, P = .568; <35 weeks: 40% vs 30%, P = .276; <32 weeks: 31% vs 23%, P = .376) nor miscarriage (<15 weeks: 12% vs 14%, P = .742) was significantly different. Although not statistically significant, mean delivery gestational age in the subsequent pregnancy was 2.4 weeks earlier in the antibiotic versus placebo group (32.0 ± 7.9 vs 34.4 ± 6.3 weeks, P = .082), and mean birth weight was lower in the antibiotic group (2046 ± 1209 vs 2464 ± 1067 g, P =.060). Conclusion: Intermittent treatment with metronidazole plus azithromycin of nonpregnant women with a recent early spontaneous preterm birth does not significantly reduce subsequent preterm birth, and may be associated with a lower delivery gestational age and lower birth weight. © 2006 Mosby, Inc. All rights reserved.