Cannabis research is going through an unprecedented expansion. During 2025 alone, more than 85 clinical trials on cannabis and cannabinoids were indexed in PubMed — a figure that would have been unthinkable a decade ago, when regulatory barriers made it nearly impossible to study the plant with the rigour demanded of any other medicine.
That volume of research has brought something that doesn’t always make for comfortable reading: results far more nuanced than popular enthusiasm suggests. The latest cannabis studies confirm clear benefits for a handful of specific conditions, rule them out for many others, and — most importantly — are finally mapping the risks with precision.
Here’s what the most recent evidence actually says, without hype in either direction.
Chronic low back pain: the most ambitious trial to date
The most significant study published recently is the phase 3 trial of the VER-01 preparation, which appeared in Nature Medicine. With 820 adult participants, it is one of the largest randomised trials ever conducted with a cannabis derivative.
The results were positive: the preparation achieved a significant reduction in chronic low back pain compared with placebo. What matters here isn’t only the efficacy but the sample size. Much of the earlier cannabis research rested on studies of twenty or thirty people — nowhere near enough to draw firm conclusions. A trial with 820 participants plays in an entirely different methodological league.
The work strengthens the case that medical cannabis has a genuine role in managing chronic non-cancer pain, particularly for patients who respond poorly to other options.
Treatment-resistant epilepsy: the strongest evidence we have
If there is one area where cannabis has proven its therapeutic value beyond reasonable doubt, it is epilepsy that resists conventional treatment.
Long-term extensions of the phase 3 cannabidiol trials continue to show sustained reductions in seizure frequency in patients with syndromes such as Dravet and Lennox-Gastaut. The key word is “sustained”: the effect holds over time rather than fading after a few months.
It’s no accident that this is precisely the indication where CBD has earned formal regulatory approval as a medicine in both Europe and the United States.
Cancer: modest benefits, real trade-offs
In patients with advanced cancer, a double-blind randomised trial found a modest benefit for pain, accompanied by a higher incidence of psychomimetic adverse effects — confusion, disorientation and perceptual changes.
It’s an honest and clinically useful result. A benefit exists, but it isn’t dramatic, and it carries a cost in side effects that each patient and clinician has to weigh.
Less encouraging was a pilot study on taxane-induced neuropathy, a form of nerve damage caused by chemotherapy. With only 12 participants, the trial showed worsening across every measured endpoint. The sample is far too small to be conclusive, but it points somewhere worth taking seriously: cannabis is not useful for every kind of pain.
The JAMA review that cooled the enthusiasm
In November 2025, JAMA published a review that triggered considerable debate. Its conclusion was blunt: evidence from randomised clinical trials does not support the use of cannabis or cannabinoids for most of the conditions they are promoted for.
That conclusion deserves careful reading, because it has been misrepresented in both directions. It does not say cannabis is useless — refractory epilepsy, certain types of chronic pain and chemotherapy-induced nausea all retain support. What it says is that the list of ailments cannabis is marketed for is vastly longer than the list of ailments it has evidence for.
That gap between commercial promise and scientific proof is, right now, the industry’s central problem.
Safety: liver enzymes, drug interactions and driving
The safety findings are arguably the most important development of the past year, because they dismantle the notion that CBD is a completely harmless substance.
A trial published in JAMA Internal Medicine with 201 participants found that 8 people — 5.6% of the CBD group — showed liver enzyme elevations more than three times the upper limit of normal. That’s a rate no medicine could ignore, and it makes a case for liver monitoring during prolonged use.
A significant drug interaction has also been documented: CBD raises plasma concentrations of THC and its active metabolite, 11-OH-Δ9-THC. In other words, combining the two cannabinoids does not soften THC’s effect as popular wisdom holds — it can actually intensify the body’s real exposure to it.
On road safety, driving simulator studies keep producing one uncomfortably consistent finding: there is a marked gap between perceived and actual impairment. Users tend to feel more capable of driving than they objectively are.
Minor cannabinoids: beyond THC and CBD
Research is beginning to look at compounds that were footnotes until recently.
CBN (cannabinol) is being reassessed: recent evidence positions it not just as a sleep aid but as a possible neuroprotective agent capable of mitigating oxidative stress. CBDA (cannabidiolic acid) is attracting growing interest in pharmacokinetic studies, though the evidence remains limited and it’s worth not running ahead of the data.
It’s a promising field, but an early one. The distance between “interesting laboratory results” and “useful treatment for patients” is usually measured in years.
Pharmacogenomics: why cannabis doesn’t affect everyone the same way
One of the most interesting current research lines is precision medicine applied to cannabinoids.
Evidence from 2025 and 2026 has focused on two genes: CYP2C9 and CYP2C19, which are responsible for metabolising cannabinoids in the liver. The variants a person inherits determine how quickly they process THC and CBD.
This explains something every consumer has noticed: the same dose produces radically different effects depending on the person. It isn’t a matter of tolerance or habit — it’s genetics. In future, this knowledge could allow doses to be tailored individually for therapeutic use.
A regulatory framework that finally helps
Much of this scientific progress is explained by regulatory change. In April 2026, the United States announced the rescheduling of medical marijuana as a lower-risk drug, with the explicit aim of facilitating research.
The previous classification imposed bureaucratic barriers that made research slow and expensive. Relaxing them opens the door to more trials, better funded and with larger samples. The fruits of that change will mostly be seen in the years ahead.
What we can take away
The picture emerging from the newest cannabis studies is of a plant that is neither the universal remedy some are selling, nor the useless substance others describe.
There are indications with solid evidence, such as refractory epilepsy. There are indications with growing, promising evidence, such as chronic low back pain. There are a great many indications promoted with no real backing. And there are concrete, measurable risks — hepatic, pharmacological, behind the wheel — that deserve to be taken seriously.
The best thing about this scientific moment is precisely that: for the first time, we’re in a position to answer with data rather than opinions. And that’s excellent news for cannabis advocates and sceptics alike.
Este artículo también está disponible en español: Nuevos estudios científicos sobre la marihuana.
