Introduction: Body mass index (BMI) and total blood volume are not always considered as variables that affect serum cotinine concentrations.
Method: We used data from the National Health and Nutrition Examination Survey (NHANES) for the years 1999-2008 and fitted regression models for smokers. In addition to traditionally used covariates like age, race, gender, and average number of cigarettes smoked daily, we used BMI and total blood volume (TBV) as continuous variables to evaluate the impact of these variables on serum cotinine levels.
Results: Adjusted serum cotinine levels increased statistically significantly with increase in age (p<0.001). Serum cotinine levels increased statistically significantly (p<0.001) with average number of cigarettes smoked daily. Levels of adjusted serum concentrations from high to low by race/ethnicity were: non-Hispanic blacks, non-Hispanic whites, other race/ethnicity, and Mexican-Americans; and all differences were statistically significant. A model of serum cotinine including BMI without TBV found BMI to be a significant predictor (p<0.001) and similarly a model including TBV without BMI found TBV to be a significant predictor (p<0.001). When BMI and TBV were both included in the model, the significance of BMI changed markedly (p=0.93) with substantive changes in the BMI coefficient and the significance of TBV changed also (p=0.024) with small change in the TBV coefficient.
Discussion: TBV and BMI are significant predictors of serum cotinine concentrations. TBV or BMI, but not both, should be included in predictive models of serum cotinine concentrations.
Published by Elsevier B.V.