Relative sensitivity and specificity of salivary and serum cotinine in identifying tobacco-smoking status of self-reported nonsmokers and smokers of tobacco and/or marijuana.

Serum and salivary cotinine levels were measured in 327 smoking and nonsmoking participants in a study of the health effects of marijuana with and without tobacco. These individuals had no reason to misrepresent their current tobacco-smoking status. The sensitivity, specificity, and predictive values positive and negative of the cotinine levels in distinguishing self-reported current tobacco smokers from nonsmokers was high (88-100%) and essentially the same for both fluids. Agreement between self-report and cotinine levels was not influenced by the presence or absence of marijuana smoking. A good correlation was found between serum and salivary cotinine levels in self-reported tobacco smokers (r = 0.84, p less than 0.001). Mean average levels were 279 +/- 144 ( +/- standard deviation) ng/ml for serum and 360 +/- 195 ng/ml for saliva. In a separate group of seven tobacco smokers, cotinine levels in saliva were found to be essentially independent of salivary flow rate. An analogous relationship has been observed by others for various compounds that are filtered to saliva from the blood. This may explain the close relationship observed between serum and salivary cotinine levels, and the observation made by others that the half-life of salivary cotinine is similar to that of serum cotinine.