We have known for a long time that cannabis use is somewhat heritable: twin studies estimate this to be around 45%. However, we don’t yet know what genetic variants predict cannabis use.
Perhaps relatedly, we also know from observational studies that cannabis use is more common in populations with certain behavioural traits, and with mental health problems. But understanding the direction of causality is difficult.
A new paper has used three large datasets to run a genome-wide association study of cannabis use (Pasman et al, 2018).
The authors used data from three pre-existing cohorts (the International Cannabis Consortium, American genetics-testing-company 23andme, and the UK Biobank study), giving them a total of 184,765 participants. Each participant had previously been asked a version of the question ‘have you ever used cannabis, at any time in your life?’, giving a binary measure of ever use of cannabis. The paper doesn’t report the exact number of individuals who had used cannabis, but rates were 42.8% in ICC, 43.2% in 23andme, and 22.3% in UK Biobank.
The GWAS analysis investigated variation between cannabis users and non-users at over 11.7 million genetic variant sites (known as single nucleotide polymorphisms – SNPs) across the genome. Because this is a huge number of statistical tests to run, differences were considered statistically significant where p-values were less than 5×10-8.
The authors then investigated whether SNPs that predicted cannabis use were also linked with any other traits, and used a technique called Mendelian randomisation to investigate whether cannabis might be a causal risk factor for schizophrenia, or vice versa.
The GWAS analysis identified 8 SNPs at genome-wide significance that predict whether or not an individual in the study had ever tried cannabis. The strongest association was shown in three genetic variants that are all part of a gene called CADM2. This gene has also been identified as a predictor of risk-taking, body mass index and reduced feelings of anxiety, as well as alcohol use and processing speed.
The authors also investigated genetic correlation of their GWAS results with that of other studies, and found associations with smoking and alcohol use, as well as ADHD, schizophrenia, and risk-taking behaviour, openness to experience, educational attainment, and a negative correlation with conscientiousness.
The two-sample Mendelian randomisation study found weak evidence in support of cannabis use causally influencing schizophrenia risk, but stronger evidence that schizophrenia predicts likelihood to use cannabis.
The authors conclude that the study provides new insights into the aetiology of cannabis use, and it’s relation with mental health.
Strengths and limitations
The study has a number of strengths: the sheer size of the sample for one makes it the best powered study of the genetics of cannabis use to date. It also uses a number of different methods to explore the results, which highlight some other interesting phenotypes associated with cannabis use.
However, as with any GWAS, the strength of the evidence is dependent on the quality of the phenotype under investigation. A binary ‘have you ever used cannabis’ measure mixes together individuals who have used cannabis every day for the last 20 years with people who have tried it once at a party and never again. So the phenotype the GWAS is investigating is supposed to be ever use of cannabis, but could perhaps equally well represent openness to try new experiences, willingness to break the law (in some countries) or risk-taking behaviour. The authors note this, when pointing out that the genetic variants identified are linked to a variety of other substance use, risk taking behaviour, and state that perhaps ‘liability to start using cannabis is an indication of one’s personality’.
The authors also point out that although the overall GWAS had high power, some of the other analyses undertaken did not, in particular the Mendelian randomisation analysis that found evidence suggesting schizophrenia predicts likelihood to use cannabis was underpowered.
Finally, the study was not able to tease apart different types of cannabis, or the levels of THC, cannabidiol or other cannabinoids in the cannabis that was being used by the participants of the study. There is growing evidence that THC and CBD in particular might have differing effects on the link between cannabis and mental health, and this is an area where further research is needed (although the practicalities of collecting this sort of data accurately are extremely challenging).
Implications for practice
This study further improves our understanding of the genetic basis for cannabis use, and gives an indication of the links between cannabis use and other health and mental health related behaviours. As with many studies, it’s one piece in a complex puzzle. The evidence from Mendelian randomisation studies as to causal links between cannabis and schizophrenia is as yet inconsistent, although this study supports another paper (by me, full disclosure) that found stronger evidence for schizophrenia predicting cannabis use than the other way round (Gage et al, 2017). Both of these papers only investigate ever use of cannabis: we know from observational literature and case control studies that the links between cannabis use and schizophrenia are driven by heavy users, and potentially by use of certain types of cannabis. Another study by different authors, found stronger evidence that cannabis use predicts risk of schizophrenia, but did not look in the opposite direction (Vaucher et al, 2018).
The use of a variety of different analytic methods, each with different limitations, and the triangulation of results, is a way this field can continue to move forward, and unpick the complicated relationships between recreational drug use and mental health.
Conflicts of interest
Pasman JA, Verweij KJH, Gerring Z, et al. (2018) GWAS of lifetime cannabis use reveals new risk loci, genetic overlap with psychiatric traits, and a causal influence of schizophrenia. Nat Neurosci. 2018 Aug 27. doi: 10.1038/s41593-018-0206-1.
Gage SH, Jones HJ, Burgess S, Bowden J, Davey Smith G, Zammit S, Munafò MR. (2017) Assessing causality in associations between cannabis use and schizophrenia risk: a two-sample Mendelian randomization study. Psychol Med. 2017 Apr;47(5):971-980. doi: 10.1017/S0033291716003172. Epub 2016 Dec 8.
Vaucher J, Keating BJ, Lasserre AM, Gan W, Lyall DM, Ward J, Smith DJ, Pell JP, Sattar N, Paré G, Holmes MV. (2018) Cannabis use and risk of schizophrenia: a Mendelian randomization study. Mol Psychiatry. 2018 May;23(5):1287-1292. doi: 10.1038/mp.2016.252. Epub 2017 Jan 24.