In this study, the authors attempted to validate answers to smoking-habit questions contained in the European Community Respiratory Health Survey questionnaire. The respondents were invited to visit the chest clinic at Verona, Italy, and their serum cotinine levels were measured. The authors invited each of 504 subjects to complete a respiratory interview and to give a blood sample for a radioimmunoassay serum cotinine measurement. A total of 375 subjects responded, of whom 129 were smokers (34.4%), 79 were exsmokers (21.1%), and 167 (44.5%) had never smoked. Exposure to environmental tobacco smoke was reported by 216 subjects (57.6% [mean exposure = 3.8 hr/day (+/- 3.4 hr/day standard deviation)]). In smokers, serum cotinine levels were directly related to the number of cigarettes smoked/day. The authors excluded from analysis nonsmokers who had serum cotinine levels that were > or = 14 ng/ml, and the resulting mean values were 1.7 ng/ml (+/- 2.1 ng/ml standard deviation) in nonsmokers unexposed to environmental tobacco smoke and 2.6 ng/ml (+/- 2.6 ng/ml standard deviation) (p < .002) in nonsmokers exposed to environmental tobacco smoke. There was a relationship between serum cotinine levels and hours of exposure to environmental tobacco smoke (R2 = .136, p < .05). Serum cotinine, which is an objective and accepted measure of tobacco exposure, confirmed the validity of the European Community Respiratory Health Survey questionnaire with respect to smoking habits and environmental tobacco smoke exposure.