The relationship between alcohol, tobacco and coffee consumption and pancreatic cancer was investigated using population-based case-control data obtained from eight Canadian provinces. Our findings are based on analyses performed on 583 histologically confirmed pancreatic cancer cases and 4813 controls. Questionnaire data were obtained directly from 76% of the cases. Male subjects with 35 or more cigarette pack-years had an increased risk of developing pancreatic cancer relative to never smokers (OR= 1.46, 95% CI 1.00-2.14). Similarly, women reporting at least 23 cigarette pack-years of smoking had an odds ratio of 1.84 (95% CI 1.25-2.69). For the most part, consumption of total alcohol, wine, liquor and beer was not associated with pancreatic cancer. Coffee drinking was not related to pancreatic cancer. More work is needed to clarify the role of these and other potentially modifiable risk factors as a means to reduce the incidence of this disease for which treatment results remain disappointing.