Study register · detailMeta-Analyse · Depression Risk · 2026
The association between major depressive disorder and cannabis use disorder: A meta-analysis and meta-regression analysis.
Pini Alemar et al.·Journal of psychiatric researchImpact 3.0
UnclearGRADEModerate1 citations
Samplek = 55 Studien n = 3.279.774 Pat.
EndpointComorbid prevalence
Blindingunklar
DesignMeta-Analyse
”Key finding
Meta-analysis shows a strong bidirectional association between major depression and cannabis use disorder (MDD prevalence in CUD: 19-22%, CUD prevalence in MDD: 4,6-28,5% depending on setting).
Summary
Meta-analysis on the bidirectional association between Major Depression (MDD) and Cannabis Use Disorder (CUD); k=55 studies, n=3.279.774 participants (454.547 with CUD, 112.328 with MDD). Current MDD prevalence in CUD patients: 19,24% (psychiatric samples) and 21,65% (community samples). Current CUD prevalence in MDD patients: 28,45% (psychiatric) vs. 4,61% (community). Strong bidirectional comorbidity across different settings; no publication bias (Egger's test).
P
PopulationPersons with Cannabis Use Disorder (CUD) or Major Depressive Disorder (MDD), pooled n=3.279.774 (of which 454.547 with CUD, 112.328 with MDD)
I
InterventionCannabis Use Disorder (CUD) as exposure/comorbidity
O
OutcomeCurrent MDD prevalence in CUD: psychiatric samples 19,24%, community samples 21,65%; current CUD prevalence in MDD: psychiatric samples 28,45%, community samples 4,61%
Confidence in the evidence
Very lowLowModerateHigh
Moderate
The third of four GRADE levels, the effect estimate is probably reliable.
Downgraded for
Publication bias
Quality profile
Sample size★★★★★
Blinding—
Effect size—
Citations / year★★★★★
Authors
Pini Alemar J, Pozzolo Pedro MO, Leopoldo K, Mandaji JVG, Gimenes GK, Blaas IK, Torales J, Ventriglio A, Castaldelli-Maia JM
Objective: Major Depressive Disorder (MDD) and Cannabis Use Disorder (CUD) frequently co-occur, yet prevalence estimates vary widely across settings. This meta-analysis updates the evidence on the bidirectional association between MDD and CUD, emphasizing current-diagnosis subgroups, which are the most clinically relevant.
Methods: Following PRISMA and MOOSE guidelines, we systematically searched PubMed, Google Scholar, and SciELO. Random-effects models estimated current-diagnosis prevalence of MDD among individuals with CUD and of CUD among individuals with MDD. Subgroup analyses differentiated psychiatric and community samples. Sensitivity analyses (leave-one-out) and Egger's tests assessed robustness and publication bias. Meta-regressions evaluated demographic, methodological, and geographic moderators.
Results: In total, 55 studies comprising 3,279,774 individuals were included (454,547 and 112,328 living with CUD and MDD, respectively). Current MDD prevalence among individuals with CUD was elevated in both psychiatric samples (19.24%) and community samples (21.65%), indicating consistent comorbidity across settings. Current CUD prevalence among individuals with MDD showed stronger contextual variation, being substantially higher in psychiatric populations (28.45%) compared with community samples (4.61%). Sensitivity analyses demonstrated stable estimates across model specifications, although psychiatric samples exhibited greater variance. Studies with older populations and using ICD-10 (compared to DSM) presented higher MDD prevalence among individuals living with CUD in meta-regression models. Egger's tests revealed no consistent evidence of publication bias.
Conclusion: Current-diagnosis estimates highlight a strong and clinically meaningful bidirectional association between MDD and CUD. Differences between psychiatric and community samples-especially the markedly higher current CUD prevalence in patients with MDD-underscore the need for systematic screening across treatment settings. Future work should improve diagnostic differentiation, particularly regarding the overlap between depressive symptoms and cannabis withdrawal.