Agreement between nicotine metabolites in blood and self-reported smoking status: The Netherlands Epidemiology of Obesity study

Self-report and nicotine detection are methods to measure smoking exposure and can both lead to misclassification. It is important to highlight discrepancies between these two methods in the context of epidemiological research. The aim of this cross-sectional study is to assess the agreements between self-reported smoking status and nicotine metabolite detection. Data of 599 participants from the Netherlands Epidemiology of Obesity study were used to compare serum metabolite levels of five nicotine metabolites (cotinine, hydroxy-cotinine, cotinine N-Oxide, norcotinine, 3-hydroxy-cotinine-glucuronide) between self-reported never smokers (n = 245), former smokers (n = 283) and current smokers (n = 71). We assessed whether metabolites were absent or present and used logistic regression to discriminate between current and never smokers based on nicotine metabolite information. A classification tree was derived to classify individuals into current smokers and non/former smokers based on metabolite information. In 94% of the self-reported current smokers, at least one metabolite was present, versus in 19% of the former smokers and in 10% of the never smokers. In none of the never smokers, cotinine-n-oxide, 3-hydroxy-cotinine-n-glucorinide or norcotinine was present, while at least one of these metabolites was detected in 68% of the self-reported current smokers. The classification tree classified 95% of the participants in accordance to their self-reported smoking status. All self-reported smokers who were classified as non-smokers according to the metabolite profile, had reported to be occasional smokers. The agreement between self-reported smoking status and metabolite information was high. This indicates that self-reported smoking status is generally reliable.