OBJECTIVE: To compare the 15-year mortality of people with a history of opioid dependence that had achieved stable abstinence, with the mortality associated with continued drug use. Another objective was to study the influence of hospitalization with comorbid psychosis on the 15-year mortality. METHOD: In 1984, 188 persons (122 men and 66 women) with a history of intravenous narcotics addiction were interviewed about their drug-use pattern. A registry-based follow-up continued through 1999 and mortality was assessed. Three 1984-drug-use categories were formed. In category 1, cohort members had achieved stable abstinence from drug use by 1984. Using Cox multiple regression analysis, we (i) estimated reduced mortality of category 1 drug users, and (ii) studied the influence of hospitalization with comorbid psychosis on mortality. RESULTS: About 32% had died during the 15-year follow-up. The 15-year mortality associated with stable abstinence was reduced by 56% when compared with the perceived worst drug-use pattern. Hospitalization for comorbid psychosis was not independently associated with mortality in this sample. When drug-use categories were compared with mortality expectations for the general population, the standard mortality rates (SMRs) were clearly elevated. Even in the stably abstinent drug-use category (category 1), SMR was significantly elevated by at least seven-fold in both genders. CONCLUSION: People who had achieved stable abstinence from injecting narcotics use were at lower risk of premature death than people with continued drug use. A residual observed excess mortality in people who had apparently achieved stable abstinence from drug use is consistent with the view of drug addiction as a chronic disease.