The idea that cannabis is a ‘gateway drug’ to more harmful substances such as opioids is highly controversial, yet has substantially impacted policy, education and how we conceptualize substance use. Given a rise in access to cannabis products and opioid‐related harm, the current study aimed to conduct the first systematic review and meta‐analysis on the likelihood of transitioning from cannabis use to subsequent first‐time opioid use, opioid use disorders (OUD), dependence or abuse.

Following Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines, pubMed/MEDLINE, Scopus, EMBASE, PsychINFO, Cochrane Central Register of Controlled Trials and Informit Health Collection were searched for full‐text articles assessing the likelihood of transitioning from cannabis to subsequent opioid use, and from opioid use to OUD, abuse or dependence given prior cannabis use. Analysis of subpopulations within studies were discussed narratively, and E‐values were calculated to assess the potential influence of unmeasured confounding.

Six studies provided relevant data from the United States, Australia and New Zealand between 1977 and 2017, a total sample of 102 461 participants. Random‐effects analysis of the adjusted pooled effect size indicates that the likelihood of transitioning from cannabis to opioid use, relative to non‐cannabis users, is odds ratio (OR) = 2.76, 95% confidence interval (CI) = 2.26–3.36, whereas the likelihood of transitioning from opioid use to OUD, abuse or dependence given prior cannabis use is OR = 2.52, 95% CI = 1.65–3.84. While the evidence was determined to be of low quality with moderate risk of bias, E‐values suggest that these findings are robust against unmeasured confounding.

A systematic review and meta‐analysis found that while people who use cannabis are disproportionately more likely to initiate opioid use and engage in problematic patterns of use than people who do not use cannabis, the low quality of the evidence must be considered when interpreting these findings.