Identifying risk-thresholds for the association between frequency of cannabis use and development of cannabis use disorder: A systematic review and meta-analysis

Abstract

BackgroundCannabis use disorder (CUD) affects one-in-five cannabis users, presenting a major contributor to cannabis-associated disease burden. Epidemiological data identify the frequency of cannabis use as a risk factor for CUD. This review aimed to determine quantifiable risk-thresholds of the frequency of cannabis use for developing CUD. MethodsSystematic search of Medline, EMBASE, PsycInfo, CINAHL, and Web of Science for cohort/case-control studies that assessed the association between frequency of cannabis use and CUD from 2000 to 2022. Effect estimates were converted to risk ratios (RR). A random-effects multi-level multivariate meta-analytic approach was utilized, and sensitivity analyses conducted. Quality of included studies was assessed with the Newcastle Ottawa Scale. ResultsSix prospective cohort studies were included in this review, drawn from two main source studies. Random-effect modeling showed a significant log-linear dose-response association between the frequency of cannabis use and CUD risk (p < 0.0001). The risk of CUD increased from RR:2.03 (95% CI:1.85–2.22) for ‘yearly’ use, to RR:4.12 (95% CI:3.44–4.95) for ‘monthly” use, RR:8.37 (95% CI:6.37–11.00) for ‘weekly’ use, and RR:16.99 (95% CI:11.80–24.46) for ‘daily’ use. Multi-level modeling showed an absolute risk increase (ARI) from 3.5% (95% CI:2.6–4.7) for ‘yearly’ use, to 8.0% (95% CI:5.3–12.1) for ‘monthly’ use, to 16.8% (95% CI:8.8–32.0) for ‘weekly’ use, and 36% (95% CI:27.047.9) for ‘daily’ use. ConclusionA limited risk of CUD as a potential outcome of cannabis use exists even at infrequent levels of use, but significantly increases as frequency of use increases. Corresponding information should be conveyed to cannabis users as part of targeted prevention messaging to promote safer cannabis use.