One familiar sceptic case says there is no link at all between cannabis and schizophrenia. It shows up often in popular cannabis writing, and it rests on three arguments and a quotation from a Harvard psychiatrist. All four deserve a proper look rather than a dismissal. Three of them fail for reasons we can identify, and one (the gateway point) is broadly right. A study that the same sceptic case cites, Leweke’s 2012 CBD trial, actually works against its own conclusion.
Nothing here is medical advice. If you have a personal or family history of psychosis or serious mental illness and you use cannabis, talk to a doctor about it. In Thailand, cannabis flower is a controlled herb. It is dispensed on prescription by authorised practitioners (e.g. form PT 33).
1. Why Answer the Sceptics Rather Than Ignore Them
The sceptic case is not just incomplete. It argues a position, and it argues it well enough to deserve a reply.
The reply also matters in practice. A personal or family history of psychosis is the strongest single caution in medical cannabis, and prescribing guidance treats it as a standard caution. A patient who reads a confident denial of that risk may skip the conversation with a doctor that would have protected them.
2. Argument One — the Population Argument
The claim. The argument goes like this. If cannabis caused schizophrenia, more than 150 million recreational users worldwide would produce a matching wave of the illness, and no such wave has appeared. Professor David Nutt is quoted from a 2009 lecture saying three things: schizophrenia appears to be disappearing from the general population even though cannabis use rose substantially over 30 years; data from Keele University showed psychosis and schizophrenia trending downward; and no evidence of a relationship between cannabis use and schizophrenia could be found.
Why it seemed strong. It is a real epidemiological test. If exposure to a cause rises sharply and the disease does not follow, that is genuine evidence against a strong causal effect. Nutt is a serious pharmacologist, and the data he cited were real. The Keele work is Frisher and colleagues’ 2009 analysis of UK general-practice records from 1996 to 2005. It found diagnosed schizophrenia and psychoses stable or declining over that period.
Where it fails. There are three problems, listed from least to most serious.
If cannabis acts on a small susceptible group, the population effect is small. Suppose cannabis mainly triggers psychosis in people who already have an underlying vulnerability, and that group is a small share of users. Then even a large rise in overall use changes incidence only slightly, and everything else changing at the same time can easily hide that. The population argument can only rule out a strong, general causal effect, and nobody claims one.
The way incidence is measured changed. Diagnostic criteria, case-finding, early-intervention services and help-seeking all changed a great deal over those decades. A downward trend in recorded incidence across thirty years is not a clean measuring instrument.
Newer evidence has overtaken it. This is the decisive point. The population argument inferred from aggregate trends because direct evidence was thin at the time. Direct evidence now exists, and section 3 covers it.
3. What the Direct Evidence Shows
Di Forti and colleagues, “The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI),” The Lancet Psychiatry, 2019. This was a multicentre case-control study at 11 sites in Europe and Brazil, with 901 patients with first-episode psychosis and 1,237 population controls.
| Exposure | Adjusted odds of psychotic disorder versus never-users |
|---|---|
| Daily cannabis use | 3.2 (95% CI 2.2–4.1) |
| Daily use of high-potency cannabis (THC ≥10%) | 4.8 (95% CI 2.5–6.3) |
What sets this study apart from earlier work is its population-level estimates. Across the 11 sites, the authors estimated that 12.2% of new cases of first-episode psychosis could be prevented if high-potency cannabis were no longer available. The estimate was 30.3% in London and 50.3% in Amsterdam, where high-potency cannabis is widely available.
When the study came out, Di Forti summarised it as showing for the first time how cannabis use affects the incidence of psychotic disorder at a population level. That is exactly the question the sceptics’ argument was trying to answer with cruder tools.
The honest caveat. Population attributable fractions assume the association is causal. Correspondence in The Lancet Psychiatry titled “High-potency cannabis and incident psychosis: correcting the causal assumption” challenged that assumption directly, and the authors published a reply. In a case-control design, reverse causation (early psychotic symptoms leading to heavier cannabis use) cannot be fully ruled out, so this is a live methodological argument. The odds ratios are the firmer finding. The attributable fractions are contested.
Still, look at how the debate has moved. It used to ask “is there a relationship?” It now asks “how much of the population burden does it explain?” Those are very different questions, and only the second is still open.
4. Argument Two — the Gateway Theory
The claim. The sceptic case makes four points. First, medical evidence does not support the gateway hypothesis, the idea that cannabis users move on to heroin and cocaine. Second, researchers have found the concept fits tobacco and alcohol better, because those are usually the first addictive substances people encounter. Third, most cannabis users do not progress to harder drugs, and many do not keep using cannabis regularly. Fourth, it adds that cannabis is less addictive than caffeine, and that modern scientific journals will no longer publish on the gateway topic.
Our assessment: the core of this is broadly right, and it is worth saying so plainly. The strong form of the gateway theory has not held up. The order in which people try drugs is better explained by shared risk factors and by access than by a pharmacological escalator, and the point that tobacco and alcohol come first is a fair one.
Two corrections to the add-on claims. The claim that journals will not publish on the topic is overstated. The question is still studied, usually with more sophisticated designs than the original idea. The claim that cannabis is less addictive than caffeine understates cannabis use disorder. It is a recognised diagnosis with a withdrawal syndrome. A meta-analysis of 47 studies covering 23,518 participants found a pooled prevalence of cannabis withdrawal syndrome of roughly 47% among people with regular or dependent use (Bahji et al., JAMA Network Open, 2020).
Being wrong on psychosis does not make the sceptic case wrong on everything. On the gateway question it is much closer to right.
5. Argument Three — Potency
The claim. The argument says modern varieties may contain more cannabinoids, but cannabis has been shown to be safe whatever its cannabinoid content. If today’s cannabis is stronger, it says, the only consequence is that users get higher and can treat their illness more easily with smaller doses.
This is the argument that has aged worst, and section 3 contradicts it directly. In the EU-GEI data, potency is not a neutral variable. It separates an odds ratio of about 3.2 from one of about 4.8, and it is central to the gap in estimated attributable fractions between Amsterdam and London.
The “smaller doses” reasoning assumes users adjust their dose precisely to the effect they want and then stop. That kind of adjustment works with inhalation, where feedback is fast. It works much less well with oral products, where the delayed onset leads people to take too much. It also has nothing to do with the risk at issue here, which is cumulative exposure in a vulnerable person, not how intense a single session is.
6. Argument Four — the Grinspoon Quotation
The claim. The sceptic case quotes Dr Lester Grinspoon (1928–2020), Associate Professor (Emeritus) of Psychiatry at Harvard Medical School. The quotation makes two points. First, since the brain produces its own cannabinoid-like compounds, it would be illogical for the same kind of compound to harm the human brain. Second, even before endocannabinoids were discovered, experiments did not show cannabis damaging the brain.
The logic does not hold, and the reason is general enough to be worth learning. The body produces cortisol, dopamine, glutamate and insulin, and every one of them causes serious harm in excess, at the wrong time or in the wrong tissue. The fact that the body makes a compound says nothing about whether an outside compound is safe at any dose, on any schedule, in any person.
The second point, that cannabis does not cause gross structural brain damage, is a narrower claim and broadly defensible. It is also not what anyone is arguing. The concern is that cannabis can trigger psychotic illness in susceptible people. That is very different from brain damage, and evidence about brain damage does not address it.
7. Where the Sceptic Case Is Right, and Its Own Best Point
The same sceptic case cites a 2012 study by Leweke and colleagues in acute schizophrenia. It reports that cannabidiol enhances anandamide signalling and reduces psychotic symptoms, with efficacy comparable to amisulpride, an established antipsychotic, and with fewer side effects.
That finding is real, and it is the most important thing in the sceptic case. But it undercuts the sceptics’ own argument rather than supporting it.
The later trial evidence is mixed:
McGuire and colleagues, American Journal of Psychiatry, 2018. This was a double-blind, randomised, placebo-controlled trial at 15 hospitals in the UK, Romania and Poland. 88 patients were randomised (43 to CBD, 45 to placebo) and 83 completed. They received pharmaceutical-grade CBD at 1,000 mg/day for six weeks, added to their existing antipsychotic medication.
- PANSS positive symptoms: reduced more than with placebo (treatment difference −1.4, 95% CI −2.5 to −0.2; p = 0.019). That is a small absolute difference.
- Clinical Global Impressions–Improvement: 78.6% rated improved on CBD versus 54.6% on placebo (p = 0.018).
- Cognitive performance and overall functioning improved, but the differences fell short of statistical significance. Negative symptoms did not differ significantly.
- CBD was well tolerated, with adverse event rates similar to placebo.
Boggs and colleagues, Psychopharmacology, 2018. This six-week placebo-controlled trial gave 36 stable outpatients with chronic schizophrenia CBD at 600 mg/day on top of their antipsychotics. It found no benefit on cognition (MATRICS battery) or on psychotic symptoms (PANSS), and sedation was more common on CBD.
So the evidence on CBD is preliminary and inconsistent. It is not established. These trials used pharmaceutical-grade CBD at high doses (600–1,000 mg/day), added under medical supervision to antipsychotic treatment. They did not use CBD oils or other shop products. CBD is not an approved treatment for schizophrenia or psychosis, and nobody should use it in place of prescribed treatment.
The overall picture is neither “cannabis is safe for mental health” nor “cannabis causes schizophrenia.” It is this:
THC and CBD are different molecules, and they do not act alike in this area. THC use is associated with markedly higher odds of psychotic disorder, especially when use is daily and high-potency, in susceptible people, and in the young. CBD at high pharmaceutical doses, added to antipsychotics, has shown a modest benefit on positive symptoms in one trial and no benefit in another. It remains a research question.
The real error is treating “cannabis” as a single substance. The sceptic case makes that error in one direction, and alarmist coverage makes it in the other.
8. Who Is Actually at Risk
Four factors keep coming up, and they add to each other:
- Personal or family history of psychosis or serious mental illness.
- Adolescence: exposure while the brain is still developing carries higher risk.
- High potency: the variable that moved the odds ratio from about 3.2 to about 4.8.
- Daily use: what matters is frequency, not any single occasion.
Someone with none of these factors has a low absolute risk. Someone with all four does not, and this article is written for that group.
9. In Thailand
Cannabis flower is a controlled herb under the Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis), B.E. 2568, in force since 26 June 2025. Outside the licensed supply chain, flower may be sold only against a prescription for a medical purpose from an authorised practitioner (e.g. form PT 33). Clause 4 requires that prescription to set a quantity for no more than thirty days’ use. Online sale and advertising of flower are prohibited.
On age: the 2022 controlled-herb rules explicitly banned sales to persons under 20 and to pregnant or breastfeeding women. Clause 4 of the B.E. 2568 notification, as published, does not repeat that wording. Instead, ordinary retail is closed and sales are tied to licensees and prescriptions. The draft Cannabis Control Act was approved by the Cabinet on 22 September 2026 and sent to the House of Representatives; it remains a draft. Press summaries of the draft say it would write an explicit ban on sales to persons under 20 into statute. Whatever else one thinks of the bill, that provision maps directly onto the age-related risk described above.
If you have a family history of psychosis or bipolar disorder and have raised cannabis with a doctor, what were you told? Did the doctor discuss potency or frequency at all? The evidence says those two factors matter more than the simple yes-or-no question, so it is worth knowing how often they come up in a consultation.
References
- 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. doi:10.1016/S2215-0366(19)30048-3
- Di Forti M, Morgan C, Selten J-P, Lynskey M, Murray RM. High-potency cannabis and incident psychosis: correcting the causal assumption – authors’ reply. Lancet Psychiatry 2019;6(6):466–467. doi:10.1016/S2215-0366(19)30176-2 (the letters it answers appear in the same issue)
- Bahji A, et al. Prevalence of cannabis withdrawal symptoms among people with regular or dependent use of cannabinoids: a systematic review and meta-analysis. JAMA Network Open 2020;3(4):e202370. doi:10.1001/jamanetworkopen.2020.2370
- McGuire P, et al. Cannabidiol (CBD) as an adjunctive therapy in schizophrenia: a multicenter randomized controlled trial. American Journal of Psychiatry 2018;175(3):225–231. doi:10.1176/appi.ajp.2017.17030325
- Boggs DL, et al. The effects of cannabidiol (CBD) on cognition and symptoms in outpatients with chronic schizophrenia: a randomized placebo controlled trial. Psychopharmacology 2018;235(7):1923–1932. doi:10.1007/s00213-018-4885-9
- Leweke FM, et al. Cannabidiol enhances anandamide signaling and alleviates psychotic symptoms of schizophrenia. Translational Psychiatry 2012;2:e94. doi:10.1038/tp.2012.15
- Frisher M, Crome I, Martino O, Croft P. Assessing the impact of cannabis use on trends in diagnosed schizophrenia in the United Kingdom from 1996 to 2005. Schizophrenia Research 2009;113(2–3):123–128. doi:10.1016/j.schres.2009.05.031
- Nutt D. Estimating drug harms: a risky business? Centre for Crime and Justice Studies, October 2009 (edited transcript of the 2009 Eve Saville lecture; no DOI). Estimating drug harms: a risky business? - Drugs and Alcohol
- Notification of the Ministry of Public Health Re: Controlled Herbs (Cannabis), B.E. 2568. Department of Thai Traditional and Alternative Medicine: https://www.dtam.moph.go.th/sub-law/42864/
This article is general educational information from the Asiannabis Community, not medical advice. If you or someone you know has symptoms such as hearing voices, unusual beliefs or marked confusion, please seek medical help promptly.