AIM: To measure the association between cannabis use disorder (CUD) and adverse cardiovascular disease (CVD) outcomes.

DESIGN AND SETTING: We conducted a matched, population-based retrospective cohort study involving five linked administrative health databases from Alberta, Canada. We identified participants with CUD diagnosis codes and matched them to participants without CUD codes by gender, year of birth and time of presentation to the health system. We included 29 764 pairs (n = 59 528 individuals in total).

MEASUREMENTS: CVD events were defined by at least one incident diagnostic code within the study period (1 January 2012-31 December 2019). Covariates included comorbidity, socio-economic status, prescription medication use and health service use. Using mortality-censored Poisson regression models, we computed survival analyses for time to incident CVD stratified by CUD status. In addition, we calculated crude and stratified risk ratios (RRs) across various covariates using the Mantel-Haenszel technique.

FINDINGS: The overall prevalence of documented CUD was 0.8%. Approximately 2.4% and 1.5% of participants in the CUD and unexposed groups experienced an incident adverse CVD event (RR = 1.57; 95% confidence interval = 1.40-1.77). CUD was significantly associated with reduced time to incident CVD event. Individuals who appeared to have greater RRs for incident CVD were those without mental health comorbidity, who had not used health-care services in the previous 6 months, who were not on prescription medications and who did not have comorbid conditions.

CONCLUSIONS: Canadian adults with cannabis use disorder appear to have an approximately 60% higher risk of experiencing incident adverse cardiovascular disease events than those without cannabis use disorder.