Last verified: April 2026

Di Forti 2019 — The Landmark Multi-Site Study

The most influential study on cannabis and psychosis risk is Marta Di Forti’s 2019 paper published in The Lancet Psychiatry. This was not a single-site study vulnerable to local confounding — it was a multi-center case-control study spanning 11 sites across Europe and Brazil, including London, Amsterdam, Paris, Barcelona, and São Paulo.

The study compared 901 patients presenting with first-episode psychosis (FEP) to 1,237 population-based controls. The central finding: individuals who used cannabis daily and consumed products with ≥10% THC had approximately a five-fold increased odds of developing a psychotic disorder compared to never-users. This dose-response relationship was consistent across sites despite dramatically different cultural contexts, healthcare systems, and cannabis markets.

Daily use of high-potency cannabis was associated with a five-fold increased odds of psychotic disorder compared with never users, across all 11 sites.

Di Forti et al., The Lancet Psychiatry 2019

The study’s cross-cultural consistency is its most important feature. London and Amsterdam — cities with predominantly high-potency cannabis markets — had the highest proportions of first-episode psychosis attributable to daily high-potency use. The estimated population-attributable fraction (PAF) — the proportion of psychosis cases in the population that would not have occurred without cannabis exposure — was 30.3% in London and 50.3% in Amsterdam. If these estimates are correct, eliminating daily high-potency cannabis use from these cities would reduce new psychosis cases by roughly one-third and one-half, respectively.

The Danish Registry — A Rising Fraction

In 2021, Carsten Hjorthøj and colleagues published a study in JAMA Psychiatry using Denmark’s comprehensive national health registries — databases that capture every psychiatric diagnosis for the country’s entire 5.8-million population. The study examined trends in the population-attributable risk fraction for cannabis use disorder as a risk factor for schizophrenia from 1972 through 2016.

The key finding: the proportion of schizophrenia cases attributable to CUD had increased from approximately 2% in 1995 to 6–8% by 2010. This temporal trend coincided with the increasing potency of cannabis products available in European markets, consistent with the dose-response relationship identified by Di Forti. Among young males (the highest-risk demographic for both cannabis use and psychosis), the attributable fraction was even higher.

The Danish data do not prove causation, but they demonstrate a population-level correlation between rising cannabis potency, increasing CUD prevalence, and a growing proportion of schizophrenia diagnoses associated with cannabis use disorder. This is precisely what a causal model would predict, and it is difficult to explain through confounding alone.

Genetic Vulnerability — AKT1 and COMT

The cannabis-psychosis association is not uniform across the population. Genetic studies have identified specific variants that dramatically modify risk, supporting a gene-environment interaction model in which cannabis acts as an environmental trigger in genetically vulnerable individuals.

The best-characterized genetic moderator is AKT1 rs2494732. AKT1 encodes a serine-threonine kinase in the dopamine signaling cascade. In a study published by Di Forti et al. (2012), individuals carrying the C/C genotype at this locus who also used cannabis daily had approximately a seven-fold increased risk of psychosis compared to T/T carriers who used daily. The C/C genotype alone did not elevate risk significantly, nor did daily cannabis use without the C/C genotype produce the same magnitude of risk — the interaction was synergistic.

A second candidate is COMT Val158Met (rs4680). COMT (catechol-O-methyltransferase) is the primary enzyme responsible for dopamine degradation in the prefrontal cortex. The Val allele produces a high-activity enzyme that rapidly clears dopamine, while the Met allele produces a low-activity enzyme that allows dopamine to accumulate. Some (but not all) studies have found that Val/Val carriers who use cannabis in adolescence show greater risk of psychotic symptoms, though the COMT findings have been less consistent than the AKT1 data.

These genetic findings do not currently support population-level pharmacogenomic screening before cannabis use — the predictive value of any single variant is too low, and most carriers never develop psychosis. But they demonstrate that the cannabis-psychosis relationship is not a simple dose-response but a complex interaction between exogenous cannabinoid exposure, endogenous dopaminergic signaling, and individual genetic architecture.

The Component Cause Model — Neither Necessary nor Sufficient

The relationship between cannabis and psychosis is best understood through Rothman’s component cause model (also called the sufficient-component cause framework). In this model, a disease (psychosis/schizophrenia) requires a "sufficient cause" — a complete set of component causes that together inevitably produce the outcome. No single component is sufficient alone, and different individuals can reach the threshold through different combinations of components.

Cannabis is a component cause of psychosis. It is neither necessary (most schizophrenia occurs in people who have never used cannabis) nor sufficient (most cannabis users, even heavy daily users, never develop psychosis). But in the presence of other component causes — genetic vulnerability (AKT1, COMT, and likely dozens of other variants), developmental factors (adolescent brain exposure), stress, trauma, urban environment — cannabis exposure can complete a sufficient cause set and trigger a psychotic episode that would not have occurred without it.

This framework resolves the apparent contradiction between "cannabis causes psychosis" and "most cannabis users never develop psychosis." Both statements are true. Cannabis increases population-level psychosis incidence (as the Danish data show) while leaving most individual users unaffected (because they lack the other component causes needed to complete a sufficient cause set).

CBD as Antipsychotic — McGuire 2018

In a remarkable inversion, the cannabis plant may also contain a treatment for psychosis. In 2018, Philip McGuire and colleagues at King’s College London published a randomized, double-blind, placebo-controlled trial in the American Journal of Psychiatry evaluating cannabidiol (CBD) as an adjunctive treatment for schizophrenia.

The trial randomized 88 patients with schizophrenia (already stabilized on antipsychotic medication) to receive either 1,000 mg/day of CBD or placebo for six weeks. Patients receiving CBD showed statistically significant improvement in positive psychotic symptoms (PANSS positive subscale), overall symptom severity (CGI-S), and were more likely to be rated as "improved" by their treating clinician. Importantly, CBD was well tolerated, with no significant differences in adverse events between the CBD and placebo groups.

CBD was associated with significant improvement in positive psychotic symptoms, and patients receiving CBD were more likely to have been rated as improved by their treating clinician.

McGuire et al., American Journal of Psychiatry 2018

CBD’s antipsychotic mechanism is distinct from conventional antipsychotics (which primarily block dopamine D2 receptors). CBD does not bind D2 with meaningful affinity. Instead, it likely acts through multiple mechanisms: inhibition of anandamide reuptake (increasing endocannabinoid tone), 5-HT1A partial agonism (serotonergic modulation), GPR55 antagonism, and TRPV1 activation. The clinical significance is that CBD could offer antipsychotic efficacy without the metabolic syndrome, extrapyramidal symptoms, and hyperprolactinemia that limit conventional antipsychotics.

The McGuire trial was a Phase II study — it demonstrated proof of concept, not definitive treatment efficacy. Larger Phase III trials are needed. But the finding that a compound from the same plant that can trigger psychosis might also treat it captures the extraordinary pharmacological complexity of Cannabis sativa.

The Honest Summary

Cannabis does not cause psychosis in the way that Mycobacterium tuberculosis causes tuberculosis — deterministically, in every exposed individual. But it is a genuine causal contributor to psychosis in the epidemiological sense: it increases risk in a dose-dependent manner, the association is consistent across studies and populations, the temporal sequence is correct, biological mechanisms are plausible and partially characterized, and the population-attributable fraction is rising as cannabis potency increases.

The groups at highest risk are those who use daily, start young, consume high-potency products, and carry vulnerability genotypes — particularly at AKT1 rs2494732. For a 30-year-old occasional user of moderate-potency cannabis with no family history of psychotic illness, the absolute risk is very low. For a 16-year-old daily dabber with a first-degree relative with schizophrenia, the risk calculation changes dramatically.