Cannabis and Refractory Epilepsy

Not Medical Advice

This article is not medical advice. Refractory epilepsy is a serious neurological condition. Do not start, stop, or change anti-seizure medications based on what you read here or anywhere else online. Cannabidiol interacts with several common anti-seizure drugs (notably clobazam and valproate) and requires monitoring of liver enzymes. Work with a neurologist or epileptologist.

Plain-Language Summary

Refractory (or drug-resistant) epilepsy means seizures that don't respond to two or more appropriately chosen anti-seizure medications ^[1]. It affects roughly one-third of people with epilepsy.

Of all the cannabis-related interventions studied, only one has held up to rigorous testing: purified cannabidiol (CBD), sold as Epidiolex in the US and Epidyolex in Europe. In randomized controlled trials, it reduced seizure frequency in three specific syndromes — Dravet syndrome, Lennox-Gastaut syndrome, and seizures associated with tuberous sclerosis complex — enough to win FDA and EMA approval ^[2][3][4][5].

That is the entire 'strong evidence' column. Everything beyond it — THC, whole-plant extracts, artisanal CBD oils, CBD for adult focal epilepsy, specific cannabinoid ratios — sits somewhere between 'promising but unproven' and 'folklore.'

What Probably Works

Pharmaceutical CBD (Epidiolex), 10–20 mg/kg/day, for three syndromes:

• Dravet syndrome. A 2017 NEJM trial showed convulsive seizures dropped from a median of 12.4 to 5.9 per month on 20 mg/kg/day CBD, versus 14.9 to 14.1 on placebo ^[2]. Strong evidence
• Lennox-Gastaut syndrome. Two RCTs showed roughly 40–45% reductions in drop seizures vs. ~20% on placebo ^[3][4]. Strong evidence
• Tuberous sclerosis complex. A 2021 RCT showed seizure frequency reductions of about 48% on CBD vs. 27% on placebo ^[5]. Strong evidence

The effect is real but modest in absolute terms — most patients still have seizures. A meaningful minority (around 5%) become seizure-free during the trial period ^[2][3]. Long-term open-label data suggest the benefit persists for at least several years in responders ^[6].

Mechanism is not fully understood. CBD is not a strong agonist at CB1 or CB2 receptors; proposed targets include GPR55, TRPV1, and adenosine reuptake modulation ^[7].

What Might Work

CBD for other refractory epilepsies. Open-label studies and expanded-access programs suggest possible benefit in CDKL5 deficiency disorder, Aicardi syndrome, Doose syndrome, and various focal epilepsies ^[8]. These are uncontrolled data — no placebo, often no blinding — so effect sizes are unreliable. Weak / limited

Artisanal / full-spectrum CBD oils. Some observational data report seizure reductions, but products vary wildly in actual CBD content, THC contamination, and contaminants ^[9]. When tested, many commercial CBD products contain substantially different amounts than labeled ^[9]. If something works, you can't reliably reproduce the dose. Weak / limited

Cannabidivarin (CBDV). Preclinical data are encouraging; a Phase 2 trial in adult focal seizures did not meet its primary endpoint ^[10]. Weak / limited

The 'entourage effect' in epilepsy. A frequently cited 2018 meta-analysis suggested CBD-rich extracts outperformed purified CBD ^[11], but the comparison pooled different studies, doses, and populations and is not a controlled head-to-head test. No randomized trial has shown a whole-plant extract beats purified CBD. Disputed

What Doesn't Work or Has Weak Evidence

THC as an anti-seizure agent. Despite folklore claims, controlled evidence that THC reduces seizures in humans is essentially absent. Animal data are mixed and THC can be proconvulsant at higher doses ^[12]. Weak / limited

Specific CBD:THC ratios (20:1, 18:1, etc.). These marketing constructs are not supported by clinical trial data. They originated in dispensary practice, not pharmacology. No data

Smoked or vaporized cannabis flower for seizures. No controlled trials support flower as an anti-seizure intervention. Pediatric use of inhaled cannabis is contraindicated. No data

CBD as monotherapy. The pivotal trials all studied CBD as add-on therapy. There is no good evidence that CBD alone replaces standard anti-seizure drugs ^[2][3][4][5]. No data

What We Don't Know

• Whether CBD works in common adult epilepsies (e.g., temporal lobe epilepsy) at clinically meaningful magnitudes.
• The optimal long-term dose; tolerance and dose escalation patterns over 5+ years.
• Whether minor cannabinoids (CBDV, CBG, THCV) add anything beyond CBD.
• Why some patients are dramatic responders and most are not — no validated biomarker predicts response.
• Effects on cognition and development with long-term pediatric use beyond what trials have measured.

Comparison with Standard Treatments

For Dravet and Lennox-Gastaut, CBD is one of several add-on options. Others include:

• Stiripentol, clobazam, valproate, fenfluramine for Dravet syndrome. Fenfluramine in particular has produced larger seizure reductions than CBD in head-to-head context, though no formal RCT compares them ^[13].
• Rufinamide, clobazam, valproate, the ketogenic diet, vagus nerve stimulation, and corpus callosotomy for Lennox-Gastaut.
• Everolimus for tuberous sclerosis complex.

CBD is not a first-line replacement for these. It is an option to add when seizures remain uncontrolled, chosen partly based on prior medication history, side-effect profile, and drug interactions. The ketogenic diet remains an underused, evidence-based option for refractory pediatric epilepsy and should not be skipped over in favor of cannabinoids ^[14].

Risks and Interactions

CBD is not benign. In trials, common adverse effects included somnolence, decreased appetite, diarrhea, fatigue, and elevated liver transaminases ^[2][3][4][5]. Liver enzyme elevations were more frequent in patients also taking valproate.

Key drug interactions:

• Clobazam. CBD raises levels of the active metabolite N-desmethylclobazam via CYP2C19 inhibition. Some of CBD's apparent anti-seizure effect in early trials may partly reflect boosted clobazam exposure, though the TSC trial showed efficacy independent of clobazam ^[5][15].
• Valproate. Increased risk of hepatotoxicity when combined with CBD.
• Warfarin, tacrolimus, and other CYP substrates. Levels can shift.

Artisanal products add risks the pharmaceutical version does not: inaccurate labeling, pesticide or heavy metal contamination, and unintended THC content that can cause intoxication or, in children, hospitalization ^[9][16].

Sources

How this page was made