In a follow-up study of 1 000 patients subjected to Billroth II resection for duodenal ulcer in the period 1948-1956, the late mortality up to the end of 1977 (522 patients) was analyzed. The observed mortality was compared with the expected mortality calculated by the life-table method and indirectly standardized for age, sex, domicile and time and cause of death. The overall mortality rate was significantly increased, mainly due to suicide, which occurred at five times the expected rate and could be related to excessive alcohol consumption and psychiatric disturbances. Malignant neoplasms, including gastric carcinoma, were not more common than expected as causes of death. The same applied to pulmonary tuberculosis, other respiratory disorders, cardiovascular, gastro-intestinal and urogenital disease and to accidents. It is concluded that the main mortality risk factors were psychiatric disease and alcohol consumption.