The 15-year follow-up of mortality and the factors associated with death from various causes were studied in an unselected group of patients surviving deliberate self-poisoning in 1978. The cohort included 152 females and 101 males. By the end of 1993 a total of 37 (24%) of the females and 33 (33%) of the males admitted in 1978 had died. The total follow-up mortality was 4.5 times greater than expected for the female group (95% confidence interval: 3.1-6.1) and 3.6 times greater than expected (2.5-5.1) for the male group. It was highest in the first 5-year period. With regard to specific causes the mortality ratio was highest for deaths from suicide. For females it was 61.1 (30.5-109.4) and for males: 38.8 (20.4-65.4) times the expected ratio. It was also significantly raised for deaths from cardiovascular diseases in females: SMR = 3.7 (2.0-6.4) and from respiratory diseases in males: SMR = 3.3 (1.2-7.1). Significant predictors for death from all causes were age > or = 30 years: RR = 4.4 (2.3-8.5) and male sex: RR = 2.1 (1.2-3.5). Imprisonment was found to be a protective factor: RR = 0.2 (0.1-0.5). Predictors for death from suicide were age > or = 30: RR = 3.1 (1.2-8.1), male sex: RR = 3.3 (1.4-7.9) and a serious suicidal attempt, as evaluated by a psychiatrist: RR = 3.4 (1.4-7.9). It is concluded that patients who survive parasuicide by deliberate self-poisoning are at increased risk of death. The predictors for death are not very specific and are difficult to apply in clinical work with these patients.