Sweden and many other countries have experienced increases in suicide and accidental overdose deaths. An analysis examined the associations between recency of non-medical opioid misuse and frequency of use of non-medical opioids with death due to either suicide or accidental overdose within a sample of 15,000 Swedish adults who completed an Addiction Severity Index (ASI) assessment for risky substance use or a substance use disorder.

Methods and materials: Suicide (n = 136) and death due to overdose (n = 405) were identified in the official Cause of Death Registry from the Swedish National Board of Health and Welfare. Control variables included demographic characteristics and risks associated with either overdose or suicide. Cox regression analyses controlled for variables statistically significantly at the bivariate level.

Results: At the multivariable level: a) a higher (modified) ASI Composite Score for mental health; b) history of suicide attempt; c) having used non-medical opioids for 1-2 times per week for at least a year; d) history of injection drug use; and, e) early onset of drug use, were all significantly and positively associated with death due to suicide. At the multivariable level: a) a higher the revised ASI Composite Score for mental health; b) recency of use of non-medical opioids; c) frequency of non-medical opioid use; d) being a male; and e) being of ages 18-24 years compared to ages 43-51 years were all positively and significantly associated with death due to accidental overdose.

Conclusion: These findings underscore the need to integrate mental health and substance use disorder treatment and provide suicide and overdose prevention interventions for individuals with an opioid use disorder. Recency and frequency of non-medical opioids were only associated with death due to overdose and not suicide. However, other drug use related variables (using opioids 1-2 times per week for at least a year, early onset of drug use and drug injection) were significantly associated with death due to suicide.