Cocaine use among pregnant women and reports of its adverse perinatal consequences have increased substantially over the past 10 years. However, most researchers have studied patients registered at drug treatment centers or have relied on voluntary participation by patients, either of which introduces the possibility of selection bias. To determine the frequency and consequences of prenatal cocaine use among an unselected inner-city obstetric population, we collected urine samples from parturient women at a municipal hospital and anonymously tested these specimens for metabolites of cocaine, marijuana, opiates, and methadone. Urine specimens, with linked obstetric data sheets, were available from a study population of 1111 patients, and pediatric data sheets were available for 846 mother-infant pairs. Cocaine metabolites were found in 11.5% of the urine samples collected, whereas metabolites of marijuana, opiates, and methadone, respectively, were present in 1.1%, 1.2%, and 0.3% of the specimens. Cocaine users were more likely than nonusers to have had no prenatal care (51% vs 8.8%; p less than 0.0001), to be American-born rather than Caribbean-born (71% vs 33%; p less than 0.001), and to have a higher parity (1.83 vs 1.14; p less than 0.0001). Infants of cocaine users had a lower mean gestational age (-0.93 weeks; p less than 0.01), a lower mean birth weight (2560 +/- 788 vs 3151 +/- 699 gm; p less than 0.001), and an increased probability of having an Apgar score of less than 7 at 5 minutes (12.5% vs 3.2%; p less than 0.0001). Multiple linear regression analysis that isolated confounding variables such as the presence or absence of prenatal care, maternal age and parity, and the use of cigarettes and alcohol did not substantially affect the differences described above. The effect of cigarette smoking on reducing fetal size was cumulative. In conclusion, cocaine is the most commonly used illicit drug among parturients in this community and is strongly associated with underutilization of prenatal care services. Infants of cocaine users are more likely to be preterm and depressed at birth and to have a low birth weight. Cocaine use, through the above-noted effects, increases the need for prenatal care while simultaneously decreasing the likelihood that it will be obtained.