Prenatal Cocaine Exposure as a Risk Factor for Later Developmental Outcomes—Reply

In Reply: These correspondents share 2 concerns. The first is disagreement with our selection criteria for articles¹—that the criteria were excessively stringent in restriction to human samples, peer review, masked assessment, and prospective recruitment, and simultaneously inadequately selective with respect to attrition and sample size. Animal studies are, as the thalidomide tragedy showed, unreliable indicators of a teratogenicity of a substance or its safety in humans. Methodologic issues also influence interpretation of animal data—whether cocaine effects are found varies by species, sex, timing of exposure, mode of administration, age at assessment, and whether control animals were pair fed with exposed animals.² Except for the article by Leech et al,³ which was presented in detail in our review, the other articles with human subjects cited by Drs Stanwood and Levitt were in a non-peer–reviewed publication.⁴ Without the exposition of the methods demanded by a peer-reviewed journal, these works-in-progress cannot be interpreted with confidence. We make no apologies for excluding, in spite of large sample sizes, articles that lack prospective recruitment or masked assessment; these criteria were justified in detail in our review.¹