Exploring Discriminant Capacity of Urinary CEMA as Combustion Marker in Tobacco Users - A Population Pharmacokinetic Approach

      Claussen, A.; Donelli, A.; Copalu, W.; Lama, N.; Weitkunat, R.; Haziza, C.; Lüdicke, F.

      Conference date
      Feb 22, 2019
      Conference name
      Society for Research on Nicotine and Tobacco (SRNT) 2019

      Background: Population pharmacokinetics (PK) analysis was performed to characterize the concentration-time profile of urinary 2-cyanoethylmercapturic acid (CEMA), a marker of exposure to acrylonitrile, which is present in tobacco smoke but not in aerosol from uncombusted tobacco, such as the IQOS® heat-not-burn tobacco product. Methods: Data from four studies, assessing exposure reduction to harmful and potentially harmful constituents (HPHC) following IQOS use compared with HPHC exposure from smoking, were pooled for analysis. Modeling and simulations were conducted using the nonlinear mixed-effect method. Goodness-of-fit diagnostics and posterior predictive checks were used to evaluate the adequacy of the model fit and prediction. The potential of CEMA as a diagnostic marker to distinguish between smoking and non-combustible tobacco use was explored by means of receiver operating characteristics (ROC) analysis. Results: A total of 632 subjects, 322 training and 310 validation datasets, were available for analysis. CEMA was best described by a two-compartment linear dispersion with first-order absorption for product use. In the final PK model the elimination rate increased with increasing baseline CEMA and in subjects outside the U.S. region; the inter-compartmental rate constant was lower in smoking. Following single use, the levels of absorption of CEMA from IQOS were 1.3% of those absorbed from smoking. Model simulations were conducted assuming an average consumption of 13 sticks or cigarettes. The area under the ROC curve was 94% and 97% using simulated and observed CEMA values, respectively. A threshold of 40 [ng/mg creat.] resulted in a discriminant accuracy of 90% (false positive rate 2%; false negative rate 26%). Conclusions: The population PK model characterized the kinetics of CEMA adequately. Exploratory analysis showed an excellent performance for CEMA to discriminate between smoking and IQOS use.