A hospital-based case-control study of pancreatic cancer was conducted in Athens in 1991-92. One hundred and eighty-one patients operated on for cancer of the exocrine pancreas in eight teaching hospitals formed the case series, whereas hospital patient controls and hospital visitor controls formed two independent comparison series. Cases and controls were matched by hospital, gender, and age in a 1:1:1 ratio, and every matched triplet was interviewed in person by the same researcher. Results indicate that tobacco smoking increased the risk of pancreatic cancer, whereas neither coffee drinking nor consumption of alcoholic beverages were associated with the disease. Diabetes mellitus, cholelithiasis, and pancreatitis were associated positively with risk of pancreatic cancer, whereas allergic asthma was inversely (but not significantly) related to the disease. There was a suggestion that earlier age at menarche was associated with increased risk of pancreatic cancer and that parous women were at lower risk. No consistent associations were noted with respect to gastrectomy, other medical conditions or operations, birth order, height, weight, broad occupational groups, or other reproductive variables. The two comparison series were remarkably similar with respect to the whole spectrum of the study variables.