Predictors of pancreatic cancer mortality among a large cohort of United States adults.

Cigarette smoking is considered an important risk factor for pancreatic cancer, but other purported risk factors are less well established. To learn more about the epidemiology of this important cause of mortality we examined associations with a variety of possible risk factors for death from pancreatic cancer in a large, prospective study of United States adults. We used proportional hazards models to obtain adjusted estimates of relative risks (hazards ratios). During 14 years of follow-up, 3751 persons died of pancreatic cancer in a cohort of 483,109 men and 619,199 women who had no reported history of cancer at enrollment in 1982. Cigarette smoking at baseline was associated with fatal pancreatic cancer among men (multivariate relative risk [RR] = 2.1, 95% confidence interval [CI] 1.9-2.4) and among women (RR = 2.0, 95% CI 1.8-2.3). A trend in risk was observed with increasing number of cigarettes smoked per day among current smokers at baseline. With several variables included in separate models for men and women, we found additional factors to be predictive of pancreatic cancer mortality, including family history of pancreatic cancer, black race, diabetes, and increased body mass index. History of gallstones was predictive of pancreatic cancer among men. An inverse association with vegetable consumption was observed among men, that was not statistically significant. Our findings confirm that cigarette smoking is an important predictor of pancreatic cancer mortality, and identify several other factors that may contribute to increased risk.