Extending research on the protective effect of flavonoids in cell culture and animal studies, we examined the association between consumption of flavonoids and flavonoid-rich foods and development of exocrine pancreatic cancer within the alpha-Tocopherol, beta-Carotene Cancer Prevention Study cohort. Of the 27,111 healthy male smokers (50-69 years) who completed a self-administered dietary questionnaire at baseline, 306 developed exocrine pancreatic cancer during follow-up (1985-2004; median, 16.1 years). Intakes of total flavonoids, three flavonoid subgroups, seven individual flavonoids, and flavonoid-rich foods were estimated from a validated food frequency questionnaire. Hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards models. Overall, flavonoid intake was not significantly associated with pancreatic cancer. However, in stratified analysis, greater total flavonoid intake was associated with decreased pancreatic cancer risk in participants randomized during the trial to placebo (fourth versus first quartile: hazard ratio, 0.36; 95% confidence interval, 0.17-0.78; P trend = 0.009) and not to supplemental alpha-tocopherol (50 mg/d) and/or beta-carotene (20 mg/d; P interaction = 0.002). Similar patterns and significant interactions were observed for flavonols, flavan-3-ols, kaempferol, quercetin, catechin, and epicatechin. Our data suggest that a flavonoid-rich diet may decrease pancreatic cancer risk in male smokers not consuming supplemental alpha-tocopherol and/or beta-carotene.