OBJECTIVE: Our purpose was to analyze (1) the effects of prevalent lower reproductive tract infections and (2) the effect of systematic diagnosis and treatment to reduce risks of early pregnancy loss (<22 weeks), preterm premature rupture of membrances, and overall preterm birth. STUDY DESIGN: A prospective, controlled treatment trial was conducted on 1260 women. During the first 7 months of the program (observation, phase I), women were examined at initiation of prenatal care for a panel of lower genital tract microorganisms and bacterial vaginosis. Women were followed up with reexaminations at 22 to 29 weeks and after 32 weeks' gestation. The recommended treatments of the Centers for Disease Control (i.e., 300 mg of clindamycin orally twice daily for 7 days for bacterial vaginosis) were used for infected women during the second 8 months of the study (treatment, phase II). Data were analyzed according to intent to treat by means of univariate and multivariate methods. RESULTS: Overall, presence of bacterial vaginosis (32.5%) at enrollment was associated with pregnancy loss at <22 weeks' gestation (relative risk 3.1, 95% confidence interval 1.4 to 6.9). Among women in the observation phase bacterial vaginosis was associated with increased risk of both preterm birth (relative risk 1.9, 95% confidence interval 1.2 to 3.0) and preterm premature rupture of membranes (relative risk 3.5, 95% confidence interval 1.4 to 8.9). Within this population (phase 1) 21.9% of preterm birth overall (43.8% premature rupture of membranes) is estimated as attributable to bacterial vaginosis. Among women with bacterial vaginosis phase II (treatment) was associated with reduced preterm birth (relative risk 0.5, 95% confidence interval 0.3 to 0.9); there was a similar reduction for women with preterm premature rupture of membranes (relative risk 0.5, 95% confidence interval 0.2 to 1.4). Women with both bacterial vaginosis and trichomoniasis were at highest risk of preterm birth (28%); treatment of both conditions (phase II) reduced preterm birth (17%) but did not eliminate this risk. Earlier patient enrollment and oral antibiotic treatment were associated with reduced preterm birth. CONCLUSIONS: This prospective, controlled trial confirms that the presence of bacterial vaginosis is associated with increased risks of pregnancy loss at < 22 weeks, preterm premature rupture of membranes, and preterm birth. Orally administered clindamycin treatment is associated with a 50% reduction of bacterial vaginosis-linked preterm birth and preterm premature rupture of membranes. Women at risk for preterm birth or preterm premature rupture of membranes because of bacterial vaginosis or common genital tract infections should be screened, treated, reevaluated for cure, and re-treated if necessary,
