Meconium testing for cocaine metabolite: prevalence, perceptions, and pitfalls.

Abstract

We determined the prevalence of prenatal cocaine use in a racially mixed sample of urban and suburban mothers and correlated its use with maternal demographics and newborn measurements. Meconium from 621 consecutive newborns delivered at two university-affiliated urban hospitals were assayed for benzoylecgonine. Maternal and infant characteristics were linked anonymously with the results. Statistical analysis included t test, Fisher's exact test, Duncan's multiple range analysis, and analysis of covariance, with a value of p < 0.05 considered significant. We found that 3.4% of meconium samples had benzoylecgonine levels exceeding 0.1 micrograms/ml. Its presence was statistically correlated with maternal and neonatal characteristics. A nurse's opinion of cocaine use was correct 22% of the time. Prenatal cocaine use was statistically associated with multiparity, multigravidity, late-onset and clinic-based prenatal care, public assistance, nonwhite race, and low academic achievement. A nurse's opinion was a poor predictor of maternal cocaine use. Cocaine-exposed infants were significantly smaller, and this correlated best with nonwhite background.