Data Selection Importance in the Study of the Association Between Maternal Smoking During Pregnancy and Schizophrenia: Response to Meier et al.

To the Editor: We appreciate the comments of Meier et al. as they give us an opportunity to further clarify some key design elements of our study. We do not believe that the evidence these authors present is adequate to conclude that maternal smoking during pregnancy is not an independent risk factor for schizophrenia. First, while their sample was also a national cohort and was large, they used maternal self-report data on smoking, which are not considered to be reliable, and such misclassification most likely biased the association toward the null in the main analysis. It is for this reason that we used serum cotinine, a reliable biomarker of maternal smoking, in our study. Based on our studies, nearly 8% of women do not disclose their smoking during pregnancy (1). Second, this misclassification is also likely to have diminished the magnitude of association calculated from sibling comparisons considered to be differentially exposed and unexposed to smoking during pregnancy based on maternal report. Third, our results persisted following adjustment for both maternal and paternal psychopathology as well as other confounding factors that are related to smoking, suggesting that familial or genetic vulnerability did not account for the associations that we observed.