Relationship between gestational cocaine use and pregnancy outcome: a meta-analysis

Despite a growing number of studies that have investigated the reproductive effects of maternal cocaine use, a homogeneous pattern of fetal effects has not been established and there is little consensus on the adverse effects of the drug. We used meta-analysis to evaluate the reproductive risks of cocaine. We reviewed the 45 scientific papers published in the English language dealing with effects of cocaine used during pregnancy on pregnancy outcome in humans, and identified 20 papers eligible for meta-analysis (cocaine use in pregnancy, pregnancy/fetal outcome studies, human studies, original work, cohort or case control studies, control group present, English language). Our analysis revealed that very few adverse reproductive effects could be shown to be significantly associated with cocaine use by polydrug users when compared to control groups of polydrug users not using cocaine [genitourinary malformations; odds ratio of 6.08 (95% CI 1.18-31.3); gestation age: Cohen's d 0.37 (CI 0.2-0.55)]. When the control groups consisted of no drug users, the polydrug users abusing cocaine had a higher risk for spontaneous abortions [odds ration 10.50 (CI 11.74-64.1)]. Similarly, comparison of users of cocaine alone or no drug users revealed a higher risk for in utero death, in addition to genitourinary tract malformations. Analysis of continuous variables (head circumference, gestational age, birth weight and length) revealed that the effect size was dependent upon the nature of the comparison. Comparison of cocaine users to no drug users consistently yielded a medium effect size (Cohen's d) between 0.50 and 0.58, while comparison of polydrug/cocaine users to polydrug/no cocaine users provided effect sizes small to non existent (0.06-0.37). These discrepancies suggest that a variety of adverse reproductive effects commonly quoted to be associated with maternal use of cocaine may be caused by confounding factors clustering in cocaine users.