Cannabis and the Incidence of Psychotic Disorder – A Multi-cohort Study

While cannabis (Cannabis sativa, Cannabaceae) prohibition is being discarded by many nations, epidemiological studies and biological evidence continue to report links between its use and psychotic disorder (PD). A meta-analysis reported a dose-response link, with the highest risks of PD among people with heaviest cannabis use. It is unclear whether cannabis use affects the incidence of PD. It is often used by individuals to self-treat psychiatric symptoms, as is also seen with other drugs and alcohol.

The European Network of National Schizophrenia Networks Studying Gene-Environment Interactions (EU-GEI; May 1, 2010 – May 1, 2015) sought to identify genetic, clinical, and environmental factors involved in the development, severity, and outcome of psychosis. As part of the EU-GEI, the authors describe differences in cannabis use across sites, identify measures of use with the strongest impact on PD rates across sites (assuming causality), calculate the population attributable fraction (PAF) for patterns of use associated with highest PD rates, and assess whether varying patterns contribute to varying incidence. Patients presenting with their first episode of psychosis within 16 catchment areas (study sites) who were 18-64 years of age, resided in the area at the time of presentation, and diagnosed as psychotic by ICD-10 criteria were eligible and, through their clinicians, invited to participate. Quota sampling was used to recruit matching controls in each of the 11 areas ultimately included. The use of local demographic data ensured the controls’ representative age, gender, and ethnicity. Controls were excluded if they had ever had a diagnosis of PD. No mention was made of cannabis or its potential role as a risk factor for PD in the recruitment of cases or controls.

Cases and controls completed a sociodemographic questionnaire and an updated Cannabis Experience Questionnaire (CEQ). The CEQ queries use of cannabis and other so-called recreational drugs including alcohol (mean n of drinks/d in avg. week); tobacco (Nicotiana tabacum, Solanaceae; never used vs. smoked <10 cigarettes/d vs. smoked >10/d); stimulants, hallucinogens, ketamine, and novel psychoactive (never vs. ever tried). Opiates (from opium poppy [Papaver somniferum, Papaveraceae]) and opioids are, oddly, omitted. Sociodemographic and CEQ variables associated with case-control status were controlled for in all analyses. Cannabis queries included lifetime use (never vs. ever) and, for those responding “ever,” whether cannabis was then used, age at first use, most consistent use frequency over a lifetime, and amount spent weekly for cannabis during periods of most consistent frequency. Respondents were asked to describe the type of cannabis used, in any language or by any name. Official reports were used to gauge levels of tetrahydrocannabinol (THC), cannabis’s main psychoactive compound, in cannabis available across Europe during the study period. The authors divide cannabis into low-potency (<10% THC), or high-potency, ≥10%, but it is unclear how accurately these designations mirror cannabis used by cases or controls.

Of 1519 potential cases approached, 356 (23%) refused, and 33 (1.9%) did not meet criteria, leaving 1130. Five sites lacked ≥10% sociodemographic or CEQ data and were excluded, leaving 901 cases and 1237 matched controls. Overall, cases were somewhat younger, more likely male, and more likely ethnic minority members than controls (P < 0.0001). Controls were more likely than cases to have higher education (P < 0.0001) and to have been employed a year before enrollment. More cases than controls had used cannabis, smoked ≥10/d tobacco cigarettes and tried other recreational drugs. There was no significant difference between cases and controls in alcohol consumption. Adjusted logistical regression showed that those who had used cannabis had slightly higher odds of PD than those who had never used it. This was not changed by age at first use, money spent, or type of cannabis used. The use of high-potency cannabis modestly increased risks of PD. This was unchanged when daily use was controlled for, but those who started using high-potency cannabis by age 15 had doubled risks without interaction (P = 0.63).

Daily use and high-potency cannabis had the highest odds ratios (ORs) for PD risk and the two measures were combined. The analysis found that the daily use of high-potency cannabis carried a risk increase of 400% compared with never use. There was no interaction between use frequency and type of cannabis (P = 0.25). ORs for daily use and type of cannabis did not vary significantly across sites compared with never use. In three sites with the most use of high-potency cannabis, daily use carried the highest PD risks compared with never use (Paris, 400% greater; London, 500%; Amsterdam, 900%). PAFs were estimated for the entire study sample and each site, and again assuming causality, the proportion of new PD cases in the entire sample attributable to daily use was 20.4%; to use of high-potency cannabis, 12.2%. PAFs among sites varied, with higher proportions of new PD cases attributed to sites with more daily use of high-potency cannabis.

The authors, and Gage in her letter, remark that despite the study’s large sample, when split among site cohorts were small (15-201 cases) and perhaps underpowered. Links between cannabis and PD were driven by daily use of high-potency cannabis; there was no link between less-than-weekly use and PD regardless of potency. Links between cannabis and PD may be bidirectional, as suggested by genetic studies. This report adds to the evidence that, for some, the daily use of high-potency cannabis may raise PD risk. Educational and interventional strategies are needed to allay these risks.


Di Forti M, Quattrone D, Freeman TP, et al. The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study. Lancet Psychiatry. May 2019;6(5):427-436.  doi: 10.1016/S2215-0366(19)30048-3.

Gage SH. Cannabis and psychosis: triangulating the evidence. Lancet Psychiatry.  May 2019;6(5):364-365. doi: 10.1016/S2215-0366(19)30086-0.