Correlates of expired air carbon monoxide: effect of ethnicity and relationship with saliva cotinine and nicotine.

AIMS To assess the accuracy of expired air carbon monoxide (CO) measurement vs. saliva cotinine and nicotine and to estimate the degree of misclassifications of smoking status as a function of ethnicity. DESIGN AND MEASUREMENTS Comparison for accuracy of expired air CO, saliva nicotine and cotinine in simultaneously collected specimens. SETTING Outpatient clinic of a clinical research ward. PARTICIPANTS 228 current African-American and Caucasian cigarette smokers. RESULTS Expired-air CO concentration was significantly associated with saliva cotinine, but not with saliva nicotine. Saliva cotinine but not expired CO or saliva nicotine showed a significant between-ethnic difference when adjusted for number of cigarettes smoked and for time since last cigarette. Agreement between expired air CO and saliva cotinine was substantial at expired CO < or = 8 ppm but only moderate at < or = 10 ppm. False negative rates were twice as high at < or = 10 ppm than at < or = 8 ppm at each saliva cotinine cut-off tested. At saliva cotinine of < or = 15 ng/ml, more African-Americans were classified as false negative. CONCLUSIONS Expired CO is strongly associated with saliva cotinine but not with saliva nicotine. Despite this association, misclassifications for no smoking are frequent if true classification is based on saliva cotinine. False negative results occur more frequently in African-Americans.