Discrepancies between self-reported and validated cigarette smoking in a community survey of New Mexico Hispanics

In a population-based survey of respiratory disease in New Mexico Hispanics, we validated self-reports of cigarette use by 1,317 subjects against salivary cotinine level and end-tidal carbon monoxide concentration. For identifying likely deceivers about cigarette smoking among self-reported never smokers and former smokers, we used cutoff values of 20 ng/ml and 8 parts per million (ppm) for salivary cotinine and carbon monoxide, respectively. Among males and females, age-standardized prevalences of current smokers based upon questionnaire reports were 30.9 and 27.1%, respectively. After adjustment for cotinine alone, these percentages were 36.2 for males and 31.1 for females, and after adjustment for cotinine and carbon monoxide level, the corresponding percentages were 39.1 for males and 33.2 for females. We conclude that self-reports about smoking habits may lead to underestimation of the prevalence of current smokers and that questionnaire responses should be validated with biologic markers of tobacco smoke exposure.