CBD: anti-psychotic or pro-psychotic?
Maybe cannabis is more complicated than we thought
Cannabis → psychosis is the closest psychiatry has to smoking → lung cancer. The evidence is summed up in a brilliant 2022 thread by Psych Unseen. Intravenous administration of THC (tetrahydrocannabinol, the component of cannabis thought to be most psychotogenic) causes acute psychotic symptoms in healthy volunteers. Multiple longitudinal studies show that cannabis use in childhood/adolescence has a dose-related increased risk of psychosis later in life. High potency cannabis (high THC-content) is associated with higher incidence of psychosis.
In people who have experienced a psychotic episode triggered by cannabis, 45% are diagnosed with a schizophrenia spectrum disorder within three years. The rate of schizophrenia diagnosis in people with cannabis use disorder (9%) is an order of magnitude higher than the general population (0.6%).
Obviously most people who smoke cannabis don’t develop psychosis, just as 80-90% of cigarettes smokers don’t develop lung cancer. But psychosis is such a serious outcome that I would be similarly worried by a young person smoking high potency cannabis as them smoking cigarettes.
The research evidence chimes with clinical experience. Many patients I meet with chronic psychotic disorders have a history of cannabis use. Psychotic relapse is often associated with cannabis use. Young people presenting with first episode psychosis will often be heavy users.
While THC content has been increasing over time, there are other important components of cannabis. The one which excites clinical researchers is cannabidiol, or CBD for short. Unlike THC, CBD doesn’t make people ‘high’ and in fact is thought to balance out negative effects. As today’s cannabis contains higher concentrations of THC, other cannabinoids such as CBD have been pushed down. Could this be making today’s strong cannabis strains more psychotogenic?
Promising results from small randomised trials in schizophrenia have led to the speculation that CBD could represent a new class of antipsychotic. A £16.5 million multi-centre trial led by my supervisor Philip McGuire aims to definitively answer whether CBD is an effective antipsychotic and whether it can prevent development of psychosis in young people at risk.
So the story is that THC in cannabis causes psychosis, while CBD might prevent it. And the shift towards higher THC and lower CBD might be worsening the psychotic effects of cannabis. What’s more, the harms from cannabis might eventually be outweighed by it providing a new class of antipsychotic!
It’s a nice story. But in science, we like to put nice stories to the test.
Does CBD mitigate the effect of THC?
A new study, Does cannabidiol reduce the adverse effects of cannabis in schizophrenia? A randomised, double-blind, cross-over trial provides an interesting twist on the CBD story. Before going further, I’ll admit I’m not an objective observer. Not only is the last author Philip McGuire but the first author is a friend, Edward Chesney. Perhaps even more significantly, I think the study is really cool. It asks an important clinical question and uses a clever study design to answer it.
Ed wanted to find out whether giving patients CBD prevents the psychosis-inducing effects of THC. To do this, he recruited patients with schizophrenia whose illness was known to be exacerbated by cannabis use. Patients had to be stable for three months and be using cannabis on at least a weekly basis.
He then, to put it erm bluntly, got them high on potent medical-grade inhaled cannabis (20% THC content). In case this dose did not induce acute psychotic symptoms (remember these are regular, long-term cannabis users with high tolerance) he got them back in for another visit and doubled the dose. If this didn’t result in psychotic symptoms, he invited them back to have 3x the original dose (only one patient required this super-dose).
If you think it was brave of Ed to get patients who are known to be susceptible to cannabis induced psychotic relapses and give them high strength cannabis in the lab until they become acutely psychotic, I agree with you! This study was brave. One patient developed persecutory delusions about the study team. They managed to get one patient’s blood pressure high - up to 224/78 mmHg. I’m sure at points during the experiments Ed’s blood pressure was similarly elevated.
This study wasn’t risk free and wouldn’t have been easy. That it was done at all is testament to the research team, local clinical services, regulatory authorities and most of all, the patients themselves - which the paper acknowledges.
Without the collaboration of clinicians at the South London and Maudsley NHS Foundation Trust, recruitment for the study would not have been feasible. Finally, we would like to extend our heartfelt appreciation to our participants who generously volunteered their time and efforts to the study.
Each participant had two randomly ordered experimental visits, in one they were given CBD (at 1g, a treatment dose) and in the other they received placebo. This was a proper double-blind trial, less than half of participants guessed their treatment allocation and researchers correctly guessed the treatment allocation for less than a fifth of patients. Ed wanted to see if CBD prevented THC-induced psychosis; the results were surprising, a clear effect, but not in the predicted direction.
You won’t believe what happened next
In the placebo experiments, THC resulted in PANSS scores (a standard measure of psychotic symptoms) increasing by an average of 3 points. Pre-treatment with CBD resulted in worse psychotic symptoms, an average increase of 5 points. Seven participants experienced a large increase in psychotic symptoms (>9 points) - all of these were after pre-treatment with CBD. Performance on a memory test, Hopkins Verbal Learning Test-Revised, was also worse after pre-treatment with CBD. Interestingly, CBD did not affect self-rated level of intoxication.
When I first saw these results, I assumed the CBD and placebo groups had been mislabelled during the analysis. It was no mistake though - the CBD and THC blood levels were taken at regular intervals so we can be sure which was which. We can also see the pre-treatment with CBD did not effect THC levels, or its active metabolites (shown below) so it wasn’t a simple case of THC concentrations driving the differences.
Why did CBD worsen THC-induced psychotic symptoms, contrary to everyone’s expectations? The authors don’t really give an explanation, apart from saying they don’t believe the effect is mediated through pharmacokinetic interaction with THC. Both THC and CBD have complex effects on the endocannabinoid system, including cannabinoid receptors and enzymes that act on cannabinoids, such as FAAH (fatty-acid amide hydrolase 1).
If there is a simple explanation of how THC and CBD produce psychotic symptoms, I’m not aware of it.
Taking a step back
I believe this study has advanced our understanding of cannabinoids in psychosis. Firstly, it answered the question of whether CBD prevents THC-induced psychotic symptoms in schizophrenia: no. Treating patients who have schizophrenia and co-morbid cannabis misuse with CBD is likely a non-starter. This has practical implications for people desiging larger clinical trials of CBD for schizophrenia. If I were involved in these trials (I’m not) I would not want to enrol people with schizophrenia who are regular cannabis users. Finally, it shows that simplistic explanations of cannabinoids influencing psychosis risk (THC=bad, CBD=good) are too simplistic.
Taking another step back, I’m hugely impressed and slightly envious that Ed completed a significant contribution to clinical research during his PhD. We now know that most neuroimaging or genetic research requires sample sizes that are prohibitively large for stand-alone PhDs but it is still possible to do meaningful experimental studies with sample sizes that can be realistically recruited by a dedicated and industrious PhD student.
Coincidentally, on the same day Ed’s study was published, another psychiatry registrar at South London and Maudsley and former doctoral training fellow at KCL, Luke Jelen, published an n=26 double-blind placebo controlled cross-over trial of naltrexone prior to ketamine infusion for depression in Nature Medicine. Luke showed that naltrexone attenuated ketamine’s antidepressant effect, implicating the opioid system in its mechanism of action.
If you are an ambitious clinician planning a clinical research training fellowship, you’d do worse than using these two studies as inspiration. You don’t need massive sample sizes, a large research team or multi-centre recruitment for impactful clinical research. A well-designed experimental study can provide definitive answers to important questions.
Another reminder that the randomised controlled trial continues to reign supreme.
Thanks, Tommy! I'm happy to confirm that an increase of 9+ on the PANSS positive scale is often, though not always, linked to raised blood pressure in the research team too.
Another fun fact that we didn’t mention in the published paper is that we occasionally saw reductions in positive symptoms, especially for irritability and hallucinations.
This is surprising indeed! But I'm going to suggest maybe a safer study design to follow-up: look at people who are discontinuing cannabis use and see if CBD mitigates psychotic symptoms in that context. I have always gotten worse symptoms when I try to stop smoking that I get from regular use, and there's some evidence supporting that withdrawal is an especially vulnerable time--in fact, I just went looking for papers and found one you're an author on! (Chesney et al, 2024). So it seems like this research team is familiar with the phenomenon.
("Hey, that grass is older than your country!" Or so I've been winkingly told.)