Daily high-potency cannabis raises the odds of psychosis 4.8-fold, regardless of genetics. The good news: nine months after quitting, the risk returns to baseline.
In Toronto and Vancouver, cannabis is legal. In Tehran, it is available. Neither word means safe. The useful question is which pattern of use, in which brain, carries what cost — and how that cost travels from the mind to the bedroom.
The short answer: yes, but not for everyone and not with every pattern of use. In the largest European case-control study of its kind (EU-GEI, 2019), daily cannabis use raised the odds of a first psychotic episode 3.2-fold, and daily use of high-potency cannabis (THC above 10 percent) raised them 4.8-fold.
The study compared 901 first-episode patients with 1,237 controls across 11 sites in Europe and Brazil. Its most striking estimate: if high-potency cannabis were unavailable, 12.2 percent of first-episode cases across all sites could have been prevented — 30.3 percent in London and 50.3 percent in Amsterdam.
The limitation should be stated plainly. This is a case-control design, those prevention figures assume causality, and reverse causation cannot be fully excluded. The absolute risk of psychosis in the general population is low, and a multiple of a small number is still a small number [Needs verification: the absolute risk for daily high-potency users was not reported in this study]. For someone using a high-potency product every night, that small number stops being ignorable.
One sentence I hear constantly: “Nobody in my family has had psychosis, so I am safe.” A 2024 analysis of the same cohort tested exactly that claim. Genetic risk and cannabis use each raised the odds independently, and regular use of high-potency cannabis still carried an odds ratio of 5.09 after adjusting for genetic risk. No gene-by-cannabis interaction was found. In plain terms: the risk of high-potency cannabis lands on everyone, good genes or bad.
The first question is not “how much?” but “what type, how often, since when?” — Image: AI-generated / Gemini 3.1 Flash via OpenRouter
What does cannabis do to desire and erections?
Here the evidence is weaker. A 2019 meta-analysis pooled five case-control studies covering 3,395 men: 1,035 cannabis users and 2,360 non-users. The prevalence of erectile dysfunction was 69.1 percent among users and 34.7 percent among non-users, an odds ratio of 3.83.
Two caveats soften that number. Heterogeneity between the studies was 90 percent, and the prediction interval crossed 1, meaning the next study could plausibly find no effect at all. No dose-response relationship has been shown. The accurate sentence is: the evidence points toward more erectile dysfunction among users, but it is not settled.
Against that data stands the user’s own experience. “Sex is better with cannabis,” clients tell me. Both things can be true at once. Cannabis can lower performance anxiety in the moment while slowly weakening the physiological basis of an erection; the user feels the first and not the second, until nothing works without it. That is where recreation quietly becomes performance dependence.
If I stop, does the risk go back down?
This is the most hopeful finding. A 2025 analysis of the EU-GEI data showed that the risk of psychosis falls with time since cessation and, after 37 or more weeks, becomes indistinguishable from that of people who never used. The odds relative to never-users: stopped one to four weeks ago, 6.89; five to twelve weeks, 2.70; thirteen to thirty-six weeks, 1.53; thirty-seven weeks and beyond, no difference.
Two clinical points fall out of those numbers. First, the highest risk sits in the first weeks after stopping. When a client in week two is sleepless and suspicious and says “quitting made me worse,” the data answers: this is the peak of the curve, not its destination. Second, the researchers flagged a preliminary signal that frequent users of high-potency products may keep an elevated risk even beyond 181 weeks [Needs verification: reported as preliminary, from a small subgroup]. Both analyses rely on retrospective self-report.
One more layer: a 2025 medRxiv preprint (not yet peer reviewed) followed 1,014 participants, 734 of them at clinical high risk for psychosis. Tobacco and cannabis co-use raised conversion to psychosis 2.53-fold, heavy co-use 3.63-fold. That fits what I wrote earlier about tobacco and cannabis used together.
After roughly nine months, the brain that stopped carries the same psychosis risk as a brain that never used. — Image: AI-generated / Gemini 3.1 Flash via OpenRouter
What this means for a couple in Toronto or Tehran
The pattern I see most often in the therapy room: one partner uses every night “for sleep” or “for anxiety,” and the other complains of a fading — less desire, less presence, and suspicious or jealous readings that sharpen after use. The user hears “you have changed” as an attack, not a report, and the conversation ends there.
In Canada, legality quietly translates into harmlessness, especially for the younger diaspora. In Tehran, nobody knows the potency of what they are using, and that not-knowing is a risk factor in itself. My job is not moral judgment. It is mapping: how often, what type, how many hours before sex, since when. The data says the risk returns to baseline about nine months after stopping. Those nine months work best as a shared plan with weekly goals and no scorekeeping, not a solitary battle under a reproachful eye. My position is one sentence: how much you use is not the main variable — what type (high-potency or not) and in what pattern (daily or occasional) is what builds the risk.
In EU-GEI the large effects belonged to daily use, above all daily high-potency use: 3.2-fold and 4.8-fold odds. For sporadic use the evidence is much weaker. Today’s high-THC products are not what “occasional” meant twenty years ago.
If there is no psychosis in my family, am I safe?
No. The 2024 genetic analysis found that schizophrenia polygenic risk and cannabis use raise the odds independently. Regular high-potency use still carried a 5.09-fold odds ratio after adjusting for genetic risk. A clean family history is not a shield.
Does cannabis cause erectile dysfunction?
A 2019 meta-analysis of 3,395 men reported erectile dysfunction in 69.1 percent of cannabis users versus 34.7 percent of non-users. But heterogeneity of 90 percent and a prediction interval crossing 1 mean the finding is not settled. Probably yes, pending longitudinal studies.
How long after quitting does the risk return to normal?
According to the 2025 analysis of EU-GEI data, from 37 weeks onward the risk of psychosis matches that of people who never used. The hardest stretch is the first one to four weeks, when the risk peaks.
Conclusion
Cannabis is neither a demon nor harmless; it is a variable with a dose, a potency and a clock. The psychosis risk is real, it is tied mostly to high-potency daily use, and it is reversible. A brain that steps away for nine months returns to baseline risk. A couple that walks those nine months together, without judgment, usually finds something it did not have before: a shared language for the thing both were afraid of.
Looking for professional support with your relationship or mental health?
1. Di Forti M, 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. 2019;6(5):427-436. PMID: 30902669 · DOI
2. Bond BW, et al. Cannabis Use Cessation and the Risk of Psychotic Disorders: A Case-Control Analysis from the First Episode Case-Control EU-GEI WP2 Study. Can J Psychiatry. 2025;70(3):182-193. PMID: 39810593 · DOI
3. Austin-Zimmerman I, et al. The impact of schizophrenia genetic load and heavy cannabis use on the risk of psychotic disorder in the EU-GEI case-control and UK Biobank studies. Psychol Med. 2024;54(15):4160-4172. PMID: 39637925 · DOI
4. Pizzol D, et al. Relationship Between Cannabis Use and Erectile Dysfunction: A Systematic Review and Meta-Analysis. Am J Mens Health. 2019;13(6). PMID: 31795801 · DOI
5. Bello D, et al. Cannabis and Tobacco Co-Use Predicts Psychosis in Clinical High Risk Cohorts. medRxiv [preprint, not peer-reviewed]. 2025. PMID: 41001470 · DOI
Image Credits
All images in this article are AI-generated illustrations (Google Gemini 3.1 Flash Image via OpenRouter) created for behckam.com. They depict no real person.
