Comparative risk of suicide by specific substance use disorders: A national cohort study

Substance use disorders (SUDs) are important risk factors for suicide, yet little is known about how suicide risks vary by specific SUDs. We examined these risks for the first time in a large general population to facilitate comparisons across SUDs. A national cohort study was conducted of all 6,947,191 adults in Sweden. SUDs (opioid, sedative/hypnotic, hallucinogen, cannabis, amphetamine, cocaine, and alcohol use disorders) were identified using inpatient, outpatient, and crime data, and suicide deaths using nationwide death data with follow-up during 2003-2016. Cox regression was used to compute hazard ratios (HRs) for suicide death while adjusting for sociodemographic factors and psychiatric, SUD, and somatic comorbidities. Co-sibling analyses assessed for confounding by unmeasured shared familial (genetic and/or environmental) factors. In 79.8 million person-years of follow-up, 15,616 (0.2%) suicide deaths were identified. All SUDs were associated with significantly increased risks, with HRs ranging from 12- to 26-fold and 2.5- to 6.4-fold before and after adjusting for covariates, respectively. After adjusting for all covariates, opioid use disorder was the strongest risk factor (HR, 6.39; 95% CI, 5.53-7.38) (P ≤ 0.002 compared with any other SUD), followed by sedative/hypnotic use disorder (4.62; 4.06-5.27) (P ≤ 0.009 compared with any other SUD except opioid or hallucinogen). Most associations persisted after controlling for shared familial factors, consistent with causal effects. In this large national cohort, all SUDs were associated with significantly increased risks of suicide death, especially opioid and sedative/hypnotic use disorders. These findings may improve risk stratification and inform interventions to prevent suicide in the highest-risk subgroups with SUDs.