Self-reported tobacco smoke exposure and plasma cotinine levels during pregnancy--a validation study in Northern Japan.

Maternal smoking is a critical public health concern requiring the establishment of its prevalence rate and clinical impact. Maternal self-reported information of tobacco smoke exposure requires validation using accurate biochemical analysis. This study examined the association between self-reported exposure to tobacco smoke and plasma cotinine level in Japanese pregnant women. We collected information about smoking and secondhand smoke (SHS) exposure during pregnancy from 5128 pregnant women in a prospective cohort design, and analyzed biochemically maternal blood samples using the enzyme-linked immunosorbent assay (ELISA) technique. Based on self-reports, the subjects were classified into three groups: 650 smokers, 728 ex-smokers and 3750 non-smokers. Using the receiver operating characteristic (ROC) curve, plasma cotinine cut-off value of 11.48 ng/mL was established for separating smokers from non-smokers, resulting in a smoking prevalence of 14%. A cotinine cut-off value of 0.21 ng/mL for discriminating exposed and unexposed nonsmokers resulted in a 63% prevalence of exposure to tobacco smoke among nonsmokers. Cotinine biomarker analysis proved accurate in validating self-reported smoking information in the subjects. Lower validity of SHS exposure suggests a need to confirm questionnaire information with biochemical analysis.