To examine prospectively the relation between alcohol consumption and pancreatic cancer risk, the authors analyzed data from the Netherlands Cohort Study. Participants were 120,852 persons who completed a baseline questionnaire in 1986. After 13.3 years of follow-up, 350 cases of pancreatic cancer (67% microscopically confirmed) were available for analysis. Compared with abstention, the highest category of alcohol consumption (> or =30 g/day of ethanol) was positively associated with pancreatic cancer risk (for all cases, rate ratio = 1.57, 95% confidence interval: 1.03, 2.39; P(trend) = 0.12; for microscopically confirmed cases, rate ratio = 1.54, 95% confidence interval: 0.94, 2.54; P(trend) = 0.22). In a subgroup of stable alcohol users (no change during the 5 years before baseline), a similarly increased risk of pancreatic cancer was found. This increased risk was limited to the first 7 years of follow-up. No associations were observed between consumption of specific alcoholic beverages and risk of pancreatic cancer. The associations were not modified by folate intake or smoking. Overall, these findings suggest an increased pancreatic cancer risk for persons with a high ethanol intake (> or =30 g/day). However, this increased risk was observed only during the first 7 years of follow-up.