Data on mortality during a 48-month follow-up period in a group of 1410 alcoholics who had received inpatient treatment were evaluated. In 1266 patients known to be either living or deceased the death rate was 7.6%. The percentage of deceased subjects was highest in the group over SO years of age. The mortality rate was higher for men (9.8%) than for women (4.8%); for those with more than one divorce (16.8%); for those who were not fit for work (18.1%) of were retired at the start of the treatment (43.3%); who were employed in the alcohol business (21. 7%); who had reduced their alcohol consumption before treatment (13.4%); who were unemployed 6 months after discharge (12.4%). The mortality rate was higher for those with high scores on a scale assessing calmness in a personality inventory (7.9%) and low scores on a questionnaire assessing motivation (10.9%) and insight into the need of change (12.4%). Alcohol-related illness before the index treatment played an important role: the mortality rate was higher for those who had had Wemicke-Korsakoff syndrome (40%), delirium tremens (15.3%), pancreatitis (13.9%) or cardiomyopathy (14.1%). The mortality rate was higher for treatment dropouls (12.9%) and for those who regularly or occasionally took sleeping pills (28.5%), psychoactive drugs (15.1%) or other drugs (11.5%) during treatment. In the follow-up periods substance use had a great effect on mortality. The mortality rate or those patients who still abstained from alcohol after 6 months (4.4%) was only a third of that for the patients who had relapsed (12.4%). Patients who took drugs during the 18-month follow-up period also had higher mortality rates: sleeping pills, 9.6%; psychoactive drugs, 9.7%, analgesics 10.7%. Logistic regression identified significant variables at admission: not fit for work, retired, job related to alcohol, suffering from Wernicke-Korsakoff syndrome and no insight for change.