Cocaine use during pregnancy was assessed by interviews and urine assays obtained prenatally and immediately postpartum from 679 urban women enrolled in prenatal care. Of these, 17% were found to have used cocaine at least once during pregnancy. Eight percent had urine assays positive for cocaine metabolites using the enzyme-mediated immunoassay technique with a cut-off of 300 ng/mL of benzoylecgonine. Of the cocaine users, 24% denied use at the time of the interview and were identified solely by urine assay. Cocaine users were significantly (P less than .01) less likely than nonusers to be married, Hispanic, or black born outside of the United States and were less well nourished. Users reported significantly (P less than .01) more sexually transmitted diseases, prior low birth weight infants, spontaneous and elective abortions, and greater use of alcohol, cigarettes, marijuana, opiates, and other illicit drugs during pregnancy. Because cocaine use is correlated with many potential risk factors, large sample sizes and multivariate statistical techniques are needed to determine whether cocaine use during pregnancy poses an independent risk for adverse neonatal outcomes.