Largest cannabis research review finds medicinal marijuana does not treat anxiety, depression, or PTSD

The most comprehensive review of medicinal cannabis research conducted to date has reached a conclusion that directly contradicts why most people use it. Cannabis does not effectively treat anxiety, depression, or post-traumatic stress disorder. The review, published in The Lancet Psychiatry in February 2025 and led by a team at the University of Sydney's Lambert Initiative for Cannabinoid Therapeutics, analyzed 105 randomized controlled trials involving more than 10,000 patients across multiple countries and conditions.

This matters because anxiety, depression, and PTSD are the three most common reasons people seek medicinal cannabis prescriptions in Australia, Canada, the United States, and the United Kingdom. In Australia alone, the Therapeutic Goods Administration had approved over 300,000 medicinal cannabis prescriptions by the end of 2024. A significant proportion of those prescriptions were written for exactly the conditions this review says cannabis cannot reliably treat.

What the review actually examined

The Lambert Initiative team conducted a systematic review and meta-analysis, which means they did not run new experiments. They pooled and reanalyzed data from existing randomized controlled trials, the highest quality evidence type in clinical research. The 105 trials examined outcomes for anxiety disorders, major depressive disorder, PTSD, psychosis, and attention deficit hyperactivity disorder, with cannabis products ranging from CBD-only formulations to high-THC preparations and 1:1 THC:CBD balanced products.

For anxiety, the review found that cannabinoids produced modest short-term reductions in self-reported anxiety scores compared to placebo, but those reductions did not reach clinical significance thresholds used by psychiatrists to determine whether a treatment is actually working. For depression, no statistically significant benefit over placebo was found across the pooled trial data. For PTSD, the evidence was described as insufficient to draw conclusions, because most available trials were small, poorly controlled, or used heterogeneous patient populations.

The psychosis risk finding

The review's most serious finding was not that cannabis fails to help. It was that cannabis can cause harm in a subset of users. High-THC cannabis products were associated with an increased risk of psychotic episodes in individuals with pre-existing vulnerability to psychosis, defined in the review as a personal or family history of psychotic disorders. The risk was not theoretical. The meta-analysis found that psychosis-related adverse events occurred approximately 3.2 times more often in high-THC cannabis groups than in placebo groups across the trials that tracked this outcome.

This has direct implications for prescribing. Many people with anxiety or depression also carry genetic or personal risk factors for psychosis that they may be unaware of. A GP prescribing high-THC cannabis for anxiety to a patient with an undiagnosed vulnerability is not giving a neutral treatment that simply might not work. They may be giving a treatment that makes the patient's mental health worse, in a way that requires urgent psychiatric intervention.

Why so many patients report it helps them

The gap between the clinical trial evidence and patient self-reports is real, and the review authors addressed it directly. Patient surveys consistently show that most medicinal cannabis users report improvement in their symptoms. A 2023 survey of Australian medicinal cannabis patients published in PLOS ONE found that 87 percent of respondents reported reduced anxiety and 79 percent reported improved sleep. Those numbers seem hard to reconcile with trial data showing no significant benefit over placebo.

The likely explanation is a combination of factors. Placebo response rates in psychiatric conditions are typically high, sometimes reaching 30 to 40 percent in depression trials. Cannabis also produces noticeable subjective effects, making it easy for users to feel that something is happening, which reinforces the perception of benefit. Many patients also use cannabis alongside other treatments, making it difficult to isolate what is actually producing improvement. None of this means patients are wrong about their experience. It means their experience cannot be cleanly attributed to cannabis specifically.

What the evidence does support

The review was careful not to dismiss all medicinal cannabis applications. There is consistent and reasonably strong trial evidence supporting cannabis for specific conditions. Cannabidiol has FDA approval in the United States as Epidiolex for treatment-resistant epilepsy, specifically Dravet syndrome and Lennox-Gastaut syndrome. Nabilone and dronabinol, synthetic cannabinoids, are approved for chemotherapy-induced nausea. And there is moderate evidence supporting cannabinoids for chronic neuropathic pain, though even there the effect sizes are modest.

The problem the review identifies is not with medicinal cannabis as a category. It is with the gap between conditions where the evidence is strong and conditions where patients are actually using cannabis. Pain and epilepsy represent a small fraction of medicinal cannabis prescriptions in most markets. Anxiety and depression represent the majority, which is where the evidence is weakest.

How prescribing doctors are responding

The Royal Australian and New Zealand College of Psychiatrists issued a statement following the Lancet Psychiatry publication saying the review reinforced its existing position that cannabis should not be used as a first-line or evidence-based treatment for anxiety, depression, or PTSD. The College noted that effective, evidence-backed treatments for all three conditions exist, including cognitive behavioral therapy, SSRIs, and SNRIs, and that patient access to those treatments should not be displaced by cannabis prescribing.

In Canada, where medicinal and recreational cannabis are both legal and the prescribing landscape is complex, Health Canada said it was reviewing the Lambert Initiative findings as part of its ongoing assessment of cannabis therapeutic claims. Canadian cannabis clinics, which have business models built around prescribing cannabis for mental health conditions, pushed back on the review methodology, arguing that real-world patient populations differ from controlled trial participants and that pragmatic effectiveness evidence tells a different story. The Lambert Initiative authors responded that without controlled trial evidence, there is no way to separate cannabis effects from placebo response or natural symptom variation over time.

Regulatory implications across major markets

The review is likely to affect regulatory discussions in several markets. In the UK, the Medicines and Healthcare products Regulatory Agency has been reviewing its guidance on unlicensed cannabis-based products, and the Lambert Initiative review will be cited in those deliberations. In Australia, the TGA's Office of Drug Control uses evidence reviews to inform its prescribing guidelines, and the review's authors submitted the findings directly to that office in February 2025.

The researchers have called for a moratorium on approving new medicinal cannabis products specifically marketed for anxiety, depression, or PTSD until adequately powered clinical trials can produce definitive evidence either supporting or refuting efficacy. The next large randomized controlled trial specifically designed to test cannabis for generalized anxiety disorder, being conducted by a consortium of European universities, is expected to publish results in late 2026.