Editors' ChoiceSchizophrenia

Cannabis use and schizophrenia: Chicken or egg?

See allHide authors and affiliations

Science Translational Medicine  26 Sep 2018:
Vol. 10, Issue 460, eaav0342
DOI: 10.1126/scitranslmed.aav0342


A genome-wide association study identifies risk genes for cannabis use and examines direction of causality for its association with schizophrenia.

Does using cannabis increase the risk of developing schizophrenia? Or does having schizophrenia make one more likely to use cannabis? Although there is plenty of evidence that the two are linked, causality is intensively debated. Clearly, a randomized controlled trial to test causality in this case would be unethical, but it is possible to address questions like these using the “natural” randomization of Mendelian inheritance [Mendelian randomization (MR)]. This is now being applied in genome-wide association studies (GWAS), but two recent papers looking at the relationship between lifetime use of cannabis and schizophrenia came to differing conclusions.

However, for GWAS, sample size is everything. Pasman et al. combined datasets from the International Cannabis Consortium, UK Biobank, and 23andMe to create the largest GWAS on lifetime cannabis use. Using a variety of analytical approaches, they found significant associations between cannabis use and a number of genetic variants, confirming previous indications that genes such as CADM2 and NCAM1 were strongly associated with cannabis use and identifying new candidates. Interestingly, cannabis use was genetically correlated with a wide range of behaviors and personality traits, such as alcohol use and dependence, increased risk taking, and decreased conscientiousness, as well as a variety of mental health disorders. When they applied MR to address causality, they found only weak evidence for a causal effect of cannabis use on schizophrenia but much stronger evidence for a causal effect of schizophrenia on cannabis use. Put simply, this suggests that individuals who have—or are about to develop—schizophrenia are more likely to try cannabis.

Caveats remain. MR is by no means infallible; the different datasets varied from each other in several respects, weakening their combined power; and the result might well have been different if cannabis use disorder (as opposed to lifetime use) had been examined instead. Nevertheless, these findings improve our understanding of how cannabis use and schizophrenia may interact and could ultimately be helpful in the management of schizophrenia.

Highlighted Article

Stay Connected to Science Translational Medicine

Navigate This Article