Bayesian derivation of an oral cancer slope factor distribution for 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK)

The tobacco-specific nitrosamine (TSNA) 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) is classified by the International Agency for Research on Cancer as a Group 1 carcinogen. Cancer risk assessment in humans exposed to TSNAs largely relies on potency values estimated from animal studies, but available cancer potency values for NNK derived from such studies are conflicting. In this analysis, oral cancer slope factors (CSFo) for NNK were derived according to U.S. Environmental Protection Agency guidelines. An animal study in which rats were exposed to NNK in drinking water was selected as the key study. The multistage-cancer model was fit to the tumor incidence data to determine a point of departure for low dose linear extrapolation, using a benchmark response of 10%. CSFo distributions were then computed using Bayesian methods and Monte Carlo simulation. The resultant CSFo point estimate (BMR/BMDL10) was 19.2 (mg/kg day)−1 based on lung tumor data and 12.2 (mg/kg day)−1 based on pancreatic tumors. The 95th percentiles of the CSFo distributions were 27.3 and 19.3 (mg/kg day)−1 based on lung and pancreatic tumors, respectively. The approach using Bayesian methods better accounts for the uncertainty inherent in the values generated using input assumptions and provides for a more robust probabilistic dose–response assessment.