In a population-based case-control study of dietary risk factors in French Canadians in Montreal, a total of 179 cases and 239 controls were interviewed between 1984 and 1988. It was found that the frequency of diabetes was almost three times higher in cases of pancreatic cancer (16%) than in the controls (6%), with an Odds Ratio (OR) of 2.52 and a Confidence Interval (CI) of 1.04-6.11. Fifty per cent of the pancreatic cancer cases had this disease before the cancer was diagnosed, compared with 71% of the controls at the time of the study. Those aged 50 years and over accounted for 75% of the cases and 71% of the controls (diabetic subjects); 68% of the cases were treated with a combination of diet and oral hypoglycemic agents compared with 86% of the controls. Insulin was used by 32% of the cases and 14% of the controls. These results suggest that diabetes may play a major role in the etiopathogenesis of pancreatic cancer in this urban population. A lot of controversy remains about the precise role of diabetes, and further studies are needed for a better assessment and understanding of the mechanisms of this association.