Frequent Cannabis Use and Epilepsy: What the Study Found

Frequent cannabis users had more convulsive seizures than nonusers in a study of 64 adults with drug-resistant epilepsy. The comparison is concerning, especially because these are seizures that can cause serious injury. It also asks a narrower question than many patients need answered: among people already admitted to a specialist epilepsy unit, how did the recorded histories of frequent users and nonusers differ?

The researchers examined records from an epilepsy monitoring unit at the University of Wisconsin, covering patients treated between 2019 and 2024. All 64 adults had drug-resistant epilepsy, meaning seizures had persisted despite trials of at least two appropriate antiseizure medications. The study comparison included 22 people whose records documented cannabis use on at least 20 days a month and 42 nonusers; people with lower levels of use were excluded. This was a look back at patients selected for specialist evaluation, rather than a trial in which people were assigned to use cannabis.

In the frequent-use group, 90.9% had experienced focal to bilateral tonic-clonic seizures, compared with 57.1% of nonusers. These seizures begin in one part of the brain and spread to involve both sides, causing convulsions. The estimated annual frequency was also higher among frequent users: an average of 3.00 such seizures versus 1.32 among nonusers. The groups differed in antiseizure medication burden as well, averaging 3.05 versus 2.21 medications. Those are comparisons between 22 and 42 patients, with substantial variation within the groups, not predictions for an individual who uses cannabis.

What the Five-Year Figure Measures

The other striking result concerns time. On average, frequent users reached the monitoring unit 5.0 years after epilepsy began, compared with 18.1 years for nonusers. The measured endpoint was admission to the monitoring unit. It was not the date on which each patient first met the criteria for drug resistance.

That distinction changes the question a reader can ask of the number. Admission depends on a clinical path as well as a disease path: a patient has to receive treatment, undergo evaluation and reach specialist care. The study measured where that path ended for this selected group. It did not establish that cannabis made medication stop working 13 years sooner. Both groups had drug-resistant epilepsy by the time researchers examined their records.

The seizure comparison deserves attention on its own terms. A greater proportion of frequent users had experienced convulsive seizures, and their estimated annual burden was higher. The study cannot tell whether cannabis contributed to that burden, whether something else associated with frequent use contributed, or whether patients with more difficult seizures were more likely to use cannabis. The records classified use over the year before admission; they cannot supply the controlled exposure history needed to sort those possibilities out.

One detail narrows a simple version of the self-medication explanation. Among 15 frequent users whose timelines could be documented, all had begun regular cannabis use before their first clinical seizure. That sequence makes it harder to explain their initial use as a response to diagnosed epilepsy. It does not show that cannabis caused their epilepsy, reveal how their use changed after seizures began, or resolve why the remaining frequent users developed severe disease. Timing helps frame a causal question; it does not answer it.

Senior author Aaron Struck has described the association as a reason to discuss cannabis use with patients, particularly when someone uses it for sleep or anxiety. He also said definitive data that it worsens epilepsy are lacking. For a patient, that leaves room for a conversation about what they use, how often, and why, without turning an observational result into an instruction to change treatment.

The Earlier Cannabis Question Was Different

Cannabis entered epilepsy treatment discussions through a different route. Controlled trials of a nearly pure cannabidiol formulation helped establish its efficacy for certain severe epilepsy syndromes, leading to US approval of a prescription CBD medicine. Cannabidiol, or CBD, and tetrahydrocannabinol, or THC, are different compounds; THC is the principal psychoactive one. A result involving frequent recreational use cannot be transferred wholesale to a defined CBD medicine.

An earlier prospective observational study of medical cannabis illustrates how much can change under the same broad label. Published in 2022, it followed 29 people aged 12 to 46 with treatment-resistant epilepsies. They received high-CBD, low-THC formulations, with CBD-to-THC ratios of 20:1 or 50:1 and a maximum of 6 milligrams of THC a day, for at least 24 weeks. Researchers compared seizure outcomes during treatment with a period before treatment. They found no significant change in convulsive seizure frequency, seizure duration, the period after a seizure, or use of rescue medication.

Both studies concern people whose seizures are difficult to control, and both examine seizure outcomes. Beyond that, the comparison has sharp limits. The 2022 researchers followed people forward while they used specified products, comparing their outcomes before and during treatment. The newer researchers looked backward through the records of admitted adults, comparing frequent users with nonusers. The newer study classified frequency of use; its reported results do not establish a THC dose or a common formulation for all 22 users. The earlier study’s small, statistically inconclusive result cannot clear recreational products of risk, just as the newer association cannot overturn trial results for purified CBD.

That history also helps explain why a question that sounds simple, “Does cannabis help epilepsy?”, keeps producing answers that resist being stacked together. A prescription CBD formulation, a high-CBD product containing some THC, and the products used by patients in a retrospective recreational-use group are different exposures. So are a randomized treatment trial, a prospective observational study and a review of specialist-unit records. If researchers want to identify a product-related risk, they need to measure the product as well as the seizures.

The newer study’s authors call for prospective research using objective measures of cannabis exposure. Such work could better establish what people consume and when, although simply following users would still leave questions about why some use cannabis and others do not. Whether reducing or stopping use changes seizures requires its own evidence. For now, the actionable detail in this study is a question to bring to an epilepsy appointment: what cannabis product is being used, at what frequency, for what purpose, and alongside which antiseizure medicines? The 22 frequent users provide a reason to ask it carefully, not an answer that fits every patient.