Systematic Review and Meta-analysis: Prevalence, Correlates, and Impact of Cannabis Use and Cannabis Use Disorder in Early-Onset Psychosis.

OBJECTIVE

To quantify prevalence of cannabis use (CU) and cannabis use disorder (CUD) in early-onset psychosis (EOP) (age <18) and examine their correlates and clinical and functional impact.

METHOD: This systematic review and meta-analysis searched 6 databases until July 1, 2025, for studies evaluating CU/CUD in EOP. Data were extracted by independent researchers, and quality assessment was conducted using the Newcastle-Ottawa Scale. Heterogeneity, publication bias, subgroup, and meta-regression analyses were performed, followed by a narrative synthesis.

RESULTS

Forty studies (N = 3,473; mean [SD] age = 16.2 [1.6] years; 59% male) were included. In EOP, pooled prevalence was 32.8% (95% CI 18.5-51.2) for current CU, 40.2% (95% CI 31.3-49.7) for lifetime CU, and 36.6% (95% CI 18.2-59.9) for lifetime CUD. In meta-regression analyses, current CU was associated with male sex (β = .012, p = .021) and hospitalization rates (β = .006, p = .021), but with less severe negative symptoms (β = -.191, p = .001) and lower proportion of schizophrenia diagnosis (β = -.026, p = .049). Lifetime CU was associated with cocaine use (β = .101, p = .003) and less severe negative symptoms (β = -.173, p = .030). Narrative synthesis indicated that CU commonly preceded psychosis onset and was linked to negative outcomes (eg, longer duration of untreated psychosis and higher hospitalization rates) in EOP. Study quality was mostly fair (67.5%); Grading of Recommendations Assessment, Development and Evaluation certainty ranged from low to moderate.

CONCLUSION

CU and CUD are common in individuals with EOP and relate to clinical presentation. Screening and integrated substance use care should be embedded in primary care and clinical settings. Further longitudinal studies and randomized clinical trials to evaluate the efficacy of interventions to reduce CU and CUD are needed.