Background: We need to better understand how the use of different substances and psychiatric comorbidity influence premature death generally and cause-specific death by overdose, intoxication and somatic disorders in people with substance use disorders.

Method: A cohort of 1405 patients consecutively admitted to a Swedish detoxification unit for substance use disorders in 1970–1995 was followed-up for 42 years. Substances were identified by toxicological analyses. Mortality figures were obtained from a national registry. Causes of death were diagnosed by forensic autopsy in 594 patients deceased by 2012. Predictions were calculated by competing risks analysis.

Results: Forty-two per cent of the cohort died during follow-up; more men than women (46.3% vs 30.4%). The standardised mortality ratio (SMR) was calculated as the ratio of observed deaths in males and females in specific age groups in the cohort versus expected deaths in corresponding groups in the general population. SMR was 5.68 for men (CI 95%; 5.04–6.11) and 4.98 (CI 95%; 4.08–5.88) for women. The crude mortality rate (number of deaths divided by number of person observation years) was 2.28% for men and 1.87% for women.

Opiates predicted increased risk of premature death while amphetamine and cannabis predicted lower risk. Comorbid psychiatric disorders were identified in 378 cases and personality disorders in 763 cases. Primary psychoses or mood/depression and anxiety disorders predicted a higher risk of premature mortality.

Death by overdose was predicted by male gender, younger age at admission to substance treatment, opiate use, and comorbid depression and anxiety syndromes. Cannabis and amphetamine use predicted a lower risk of overdose. Death by intoxication was predicted by male gender, use of sedatives/hypnotics or alcohol/mixed substances, primary psychoses and depression/anxiety syndromes. Premature death by somatic disorder was predicted by male gender and alcohol/mixed abuse.

Conclusion: Psychiatric comorbid disorders were important risk factors for premature drug-related death. Early identification of these factors may be life-saving in the treatment of patients with substance use disorders.