Clinical Takeaway
Clinical Takeaway: Current randomized controlled trial evidence does not demonstrate established efficacy or safety of cannabinoids as primary treatment for mental disorders or substance use disorders. Existing research gaps mean cannabinoid use for these conditions should remain within carefully monitored clinical contexts with realistic patient expectations about therapeutic benefit.
#1 The efficacy and safety of cannabinoids for the treatment of mental disorders and substance use disorders: a systematic review and meta-analysis.
Citation: Wilson Jack et al.. The efficacy and safety of cannabinoids for the treatment of mental disorders and substance use disorders: a systematic review and meta-analysis.. The lancet. Psychiatry. 2026. PMID: 41856154.
Design: 6 Journal: 4 N: 0 Recency: 3 Pop: 3 Human: 1 Risk: -2
- Preclinical only
Abstract: BACKGROUND: Mental disorders and substance use disorders (SUDs) are among the leading reasons for which the medical use of cannabinoids has been approved, but their efficacy and safety in treating these conditions is yet to be established. We conducted a systematic review and meta-analysis of randomised controlled trials (RCTs) testing the efficacy and safety of cannabinoids as the primary treatment for mental disorders or SUDs. METHODS: We searched Ovid MEDLINE, PsychINFO, Cochrane Central Register of Controlled Clinical Trials, Cochrane Database of Systematic Reviews, and Embase for peer-reviewed articles published between Jan 1, 1980, and May 13, 2025, evaluating the efficacy of cannabinoids in reducing or treating mental disorders and SUDs as the primary indication. Primary outcomes were remission of disorder or reduction in disorder symptoms. Safety was assessed via synthesis of all-cause and serious adverse events, which was used to calculate the number needed to treat to harm (NNTH). Two independent reviewers screened all studies and performed data extraction. Evidence was synthesised as odds ratios (ORs) for dichotomous measures and standardised mean differences (SMDs) for continuous measures, via random-effects meta-analysis in Review Manager, version 5.4. Risk of bias was assessed using the Cochrane Collaboration Risk of Bias 2.0 tool. We evaluated the quality of the primary outcomes using the GRADE framework. The study was registered with PROSPERO (CRD42023392718). FINDINGS: 54 trials were identified for inclusion (2477 participants; 1713 [69%] males, 764 [31%] females; median age 33·3 years [IQR 28·1-38·05; ethnicity data not available). 24 (44%) of these trials had a high risk of bias, and the certainty of evidence for most outcomes was low. Our meta-analysis revealed that a combination of cannabidiol and delta-9-tetrahydrocannabinol reduced cannabis withdrawal symptoms (SMD -0·29, 95% CI -0·57 to -0·02) and weekly grams of cannabis use (-1·00, -1·69 to
What This Study Teaches Us
A systematic review of 54 randomized trials found that evidence for cannabinoids treating mental disorders and substance use disorders remains low-quality and uncertain. Most trials had significant bias problems, making definitive claims about efficacy and safety premature.
Why This Matters Clinically
Patients and clinicians are increasingly asking about cannabis for depression, anxiety, PTSD, and addiction, but this review shows we don’t yet have reliable evidence to support these uses as primary treatments. This matters because off-label enthusiasm can outpace actual clinical benefit data.
Study Snapshot
| Study Design | Systematic review and meta-analysis of randomized controlled trials |
| Population | 54 trials with 2,477 participants (69% male, 31% female, median age 33 years) with mental disorders or substance use disorders |
| Intervention | Cannabinoids as primary treatment (specific doses and formulations not detailed in abstract) |
| Primary Outcome | Remission of disorder or reduction in disorder symptoms; safety measured via adverse events and number needed to treat to harm |
| Key Result | Abstract truncated, but authors note low certainty of evidence for most outcomes and high risk of bias in 44% of included trials |
Where This Paper Deserves Skepticism
The abstract is incomplete, cutting off before key findings are stated, making it impossible to assess the actual efficacy or safety numbers. The fact that 44% of trials had high risk of bias and GRADE certainty was low for most outcomes severely limits what we can conclude. The median trial participant was male, middle-aged, and ethnicity data were unavailable, raising questions about generalizability to diverse populations and to older or younger patients commonly seen in primary care.
Dr. Caplan’s Take
I appreciate that this meta-analysis took the rigorous step of acknowledging low certainty rather than overselling preliminary data. The reality is that while cannabinoids are increasingly used off-label for mental health and addiction, the RCT evidence base remains thin and methodologically weak. This doesn’t mean there’s no signal worth investigating further, but it does mean I counsel patients that we’re still in early stages and that established treatments (medications, therapy, behavioral interventions) remain the foundation of care. We need better trials, clearer outcome measures, and honest conversation about what we actually know versus what we hope might work.
Clinical Bottom Line
Cannabinoids lack reliable evidence as primary treatment for mental disorders and substance use disorders; clinicians should apply standard evidence-based treatments first and discuss the uncertain evidence base transparently with patients interested in cannabis.
|
