BACKGROUND: Survival is lower in ulcer perforation patients than in the general population. This study assesses the causes of death in patients treated for peptic ulcer perforation. METHODS: Cause-specific mortality in a population-based cohort of 817 patients treated for ulcer perforation in western Norway during the period 1962-1990 was compared with cause-specific population death rates. Analyses were based on observed and expected mortality curves for major causes of death and on standardized mortality rates (SMRs). Cox regression models were used to analyse possible differences on the basis of sex, birth cohort, surgical procedure, and ulcer location. RESULTS: Ulcer perforation patients experienced increased mortality from neoplasms (SMR = 1.8; 95% confidence interval (CI) = 1.4-2.1), lung cancer (SMR = 3.6; 95% CI = 2.3-4.9), circulatory diseases (SMR = 1.3; 95% CI = 1.1-1.6), ischaemic heart disease (SMR = 1.3; 95% CI = 1.03-1.6), and respiratory diseases (SMR = 1.9; 95% CI = 1.3-2.6). Postoperative deaths accounted for 38% of all excess deaths. Death from recurrent peptic ulcer was increased also in subjects who survived the 1st year after the perforation (SMR = 5.8; 95% CI = 1.2-10.4) but accounted for only a few deaths. The increase in mortality from lung cancer was higher in subjects born after 1910 than in patients of older generations. Excess mortality from lung cancer and from circulatory diseases was higher in male than in female patients. CONCLUSIONS: Increased mortality in ulcer perforation patients could mainly be attributed to smoking-related diseases. This is indirect evidence that smoking may be an important aetiologic factor for ulcer perforation.