This 5-year follow-up study includes all patients (n= 934; 50% females) treated for self-poisoning in Oslo during 1 year. Seventeen percent were considered suicide attempts upon admission, 25% among the nonabusers and 8% among the abusers. At follow-up, 122 patients were dead (61% males). The mortality rate was highest among the abusers. The mortality rate was similar (13%) among those who were considered to be suicidal on admittance and those who were not. The causes of death were suicide (28%), opiate abuse (16%), heart disease (14%), accidents or wounds (11%), alcoholism (9%) and others (22%). The standard mortality rate was highly increased in all groups (8 times on average), highest among the female opiate abusers, whose rate was 63 times higher than expected. The increased suicide rates (87 times for females, 27 times for males), however, may be a more relevant measure of mental morbidity than the standard mortality rate. Logistic regression analysis demonstrated that male sex, age above 50 years and the lowest social group were factors on admission associated with death in the follow-up period. Age above 50 years and suicidal attempt on admission were associated with subsequent suicide. The study strongly supports the idea of self-destructiveness and slow suicide in substance abuse.