An 18-year addiction career, 1985-2003, for 157 heroin dependent subjects (73% men; 49% human immunodeficiency virus seropositive) admitted for the first time to Stockholm's Methadone Maintenance Treatment program during 1989 to 1991 was analyzed with data from seven official registers and patient records. Regression analyses and incidence rates for various outcomes were calculated for subjects in first methadone maintenance treatment at the end of the observation period, discharged from first methadone maintenance treatment, in second methadone maintenance treatment, and discharged from second methadone maintenance treatment. Being human immunodeficiency virus positive (HR = 3.8), lodging (HR = 1.9) and prison sentence (HR = 1.7) predicted mortality for the 45% deceased. Approximately 70% of living subjects participated in methadone maintenance treatment at some period each year. Subjects in first or second methadone maintenance treatment had less criminality and had spent more time in methadone maintenance treatment (70% to 100%) than those discharged from first or from second methadone maintenance treatment (50%). Efforts and interventions should be intensified to increase time in treatment also for those with high problem severity.