DESCRIPTION: Bacterial vaginosis (BV) is extremely prevalent among low-income, urban pregnant women. The current standard of medical care does not involve screening pregnant women for BV unless clinical symptoms are reported; however, the majority of pregnant women with BV are asymptomatic. A number of studies have found associations between BV and late pregnancy outcomes; such as, preterm labor, premature rupture of membrane, chorioamnionitis, and low birth weight. The impact of BV on the risk for spontaneous abortion (SAB) is unclear. In the proposed prospective cohort study, all women will be screened for BV early in pregnancy regardless of symptoms. The specific aims of this study are to: 1) characterize the prevalence and predictors of BV in women early in pregnancy and 2) evaluate whether BV during pregnancy is an important, independent predictor of SAB. Women attending their first clinical prenatal care visit at the Hospital of the University of Pennsylvania Obstetric Clinic with a pregnancy of 12 weeks gestation or less as determined by last menstrual period will be recruited. We will screen all women for bacterial vaginosis and follow-up through 22 weeks gestation to identify women experiencing a spontaneous abortion. We will enroll 2200 women over a three year period arid compare SAB rates for the estimated 400 women found to test positive for BY (20 percent of patients) and the 1600 women found to test negative for By. Baseline data collection will be standardized and include a structured in-person interview, a vaginal smear used to detect By, and urine analysis to determine alcohol, cocaine and cotinine. Follow-up telephone interviews will be conducted at 22 weeks gestation to determine the status of pregnancy (SAB vs. non-SAB) and BV diagnosis and treatment. Pregnancy outcome status will also be ascertained through ongoing review of medical records, pathology logs and birth certificates. Initial analyses will be exploratory and descriptive, characterizing the prevalence and predictors of BV and the risk factors for SAB. The primary analysis will be logistic regression, with relative risks and 95 percent confidence intervals, to explore whether BV is an independent predictor of SAB. This study will provide a unique opportunity to evaluate the prevalence of BV among pregnant women and to determine the relationship between BV and incident SAB.