We analyzed the relation between mutations in codon 12 of the K-ras oncogene and lifetime consumption of tobacco in patients with exocrine pancreatic cancer (EPC).Incident cases of EPC were prospectively identified and interviewed during hospital admission about smoking and other factors. Exact logistic regression was used to compare EPC cases (N = 107) with and without K-ras mutations (case-case study).Mutated cases were nonsignificantly less likely to have been smokers than wild-type cases: the odds ratio adjusted by age and sex was 0.54 (95% confidence interval, 0.10-2.69; P = 0.613). With respect to never smokers, adjusted odds ratios for former and current smokers were 0.79 and 0.36, respectively (P = 0.193). Pack-years smoked, years of smoking, and cigarettes smoked per year also tended to be higher in nonmutated than in mutated cases. Neither age at onset of smoking nor the time between quitting and diagnosis were associated with K-ras.Tobacco does not play a major part in the acquisition of K-ras mutations in the pancreatic epithelium. Although both smoking and K-ras mutations have important roles in the etiopathogenesis of EPC, the 2 processes may act independently.