Medical cannabis and epilepsy: Epidiolex, the pivotal NEJM trials, and the difference between an FDA-approved pharmaceutical and a state-program product

Epilepsy occupies a unique position in the cannabis-medicine landscape. It is the only indication for which a cannabis-derived compound has earned full FDA approval (cannabidiol, marketed as Epidiolex), and the only one with three separate FDA-approved pediatric indications. It is also a qualifying condition in 44 US medical-cannabis state programs, with seizure disorders separately listed in 41. The two pathways, the pharmaceutical and the state-program, are different products, different regulatory regimes, and different evidentiary standards.

Conflating them produces real clinical confusion. A family with a child diagnosed with Dravet syndrome has a fundamentally different therapeutic option set than a family pursuing state-program cannabis for refractory adult focal epilepsy. This article walks the distinction.

Epidiolex: an FDA-approved pharmaceutical, not a state-program product

Epidiolex (cannabidiol oral solution) is the only cannabis-derived medication with full FDA approval for any pediatric or adult indication. The product is a purified cannabidiol formulation (≥ 98% CBD by weight, with negligible THC) manufactured under FDA Good Manufacturing Practice (GMP) standards, distributed through standard pharmacy channels, and covered by most prescription drug insurance plans. It is currently classified by DEA as a Schedule V controlled substance, the lowest control tier, reflecting its FDA approval and clinical safety profile.

The FDA-approved label (NDA 210365) covers three indications:

• Lennox-Gastaut syndrome (LGS): approved June 25, 2018, for patients 1 year and older. Subsequently expanded to include patients under 1 year of age in 2024.
• Dravet syndrome: approved June 25, 2018, for patients 1 year and older. Same age-range expansion in 2024.
• Tuberous sclerosis complex (TSC): approved July 31, 2020, for patients 1 year and older.

Standard adult and pediatric dosing per the FDA label is 5-20 mg/kg/day, divided into two daily doses, with titration recommended over the first week of therapy. The most common adverse effects are somnolence, sedation, lethargy, decreased appetite, and elevated liver-function tests (which require routine LFT monitoring during therapy).

Drugs@FDA's application 210365 overview records the full regulatory history, including the original approval, the supplements for the TSC indication and label expansions, and the post-marketing safety reporting.

For families navigating epilepsy treatment in any of the three approved indications, Epidiolex is a prescription drug. It does not require participation in a state medical-cannabis program. It is dispensed at standard pharmacies. It is typically covered by commercial and government insurance, including Medicaid in most states.

The pivotal clinical trials

The FDA approval of Epidiolex rested on three pivotal Phase 3 randomized controlled trials, each published in a top-tier general medical journal. Each trial enrolled patients with the specific epilepsy syndrome the approval would cover, used objective seizure-count endpoints, and was placebo-controlled and double-blinded.

Devinsky et al., NEJM 2017 (PMID 28538134): Trial of Cannabidiol for Drug-Resistant Seizures in the Dravet Syndrome

The Dravet syndrome pivotal trial enrolled 120 children and young adults with treatment-resistant Dravet syndrome. Patients were randomized to receive cannabidiol 20 mg/kg/day or matched placebo, added to their existing anti-seizure medication regimens. The primary endpoint was percentage change in convulsive seizure frequency over the 14-week treatment period.

The result: cannabidiol-treated patients experienced a median 39% reduction in convulsive seizure frequency, compared with a 13% reduction in placebo-treated patients (P=0.01). Five percent of cannabidiol-treated patients became seizure-free, compared with 0% of placebo-treated patients. Adverse events were more common in the cannabidiol arm (most often diarrhea, decreased appetite, somnolence, and elevated liver enzymes), with 8% of patients discontinuing for adverse events.

Devinsky et al., NEJM 2018 (PMID 29768152): Effect of Cannabidiol on Drop Seizures in the Lennox-Gastaut Syndrome (GWPCARE3)

The first of two LGS pivotal trials enrolled 225 patients aged 2-55 years with Lennox-Gastaut syndrome. The trial compared two cannabidiol doses (20 mg/kg/day and 10 mg/kg/day) against placebo, over 14 weeks. The primary endpoint was percentage change in monthly drop-seizure (atonic seizure) frequency.

The result: drop seizures decreased by a median 41.9% with cannabidiol 20 mg/kg/day, 37.2% with cannabidiol 10 mg/kg/day, and 17.2% with placebo. Both cannabidiol doses were superior to placebo (P less than 0.01 for both comparisons). The trial also enabled a dose-response comparison: the higher dose did not produce substantially greater seizure reduction than the lower dose but did produce more adverse events, leading to the conclusion that 10 mg/kg/day is a reasonable starting target for many patients.

Thiele et al., Lancet 2018 (PMID 29395273): Cannabidiol in patients with seizures associated with Lennox-Gastaut syndrome (GWPCARE4)

The second LGS pivotal trial enrolled 171 patients aged 2-55 years with treatment-resistant LGS. The design was similar to GWPCARE3 but compared cannabidiol 20 mg/kg/day against placebo only, over 14 weeks. The primary endpoint was percentage change in monthly drop-seizure frequency.

The result: drop seizures decreased by a median 43.9% in the cannabidiol arm versus 21.8% in the placebo arm (P=0.0135). The trial confirmed the GWPCARE3 efficacy finding and provided the second independent data set supporting the LGS indication approval.

Thiele et al., JAMA Neurology 2021 (PMID 33346789): Add-on Cannabidiol Treatment for Drug-Resistant Seizures in Tuberous Sclerosis Complex (GWPCARE6)

The TSC pivotal trial enrolled 224 patients aged 1-65 years with treatment-resistant epilepsy in tuberous sclerosis complex. The trial compared two cannabidiol doses (25 mg/kg/day and 50 mg/kg/day) against placebo over 16 weeks. The primary endpoint was percentage change in TSC-associated seizure frequency.

The result: cannabidiol 25 mg/kg/day reduced TSC-associated seizures by 48.6% from baseline, compared with 26.5% on placebo (P less than 0.001). The 50 mg/kg/day dose was numerically similar but had a higher adverse-event burden. The trial supported the 2020 FDA indication expansion to include TSC.

The combined pivotal-trial evidence is unusually strong for a cannabis-derived product. Each trial used objective seizure-count endpoints, each was conducted in a difficult-to-treat patient population where placebo response was expected to be limited, and each demonstrated statistically significant seizure reduction with cannabidiol versus placebo.

What about state-program cannabis for epilepsy?

The 44 US medical-cannabis state programs that list epilepsy as a qualifying condition (and the 41 that separately list seizure disorders) are a different category of access. The state-program cannabis pathway is:

• Northeast: Connecticut, Delaware, Maine, Maryland, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont
• South: Alabama, Arkansas, Florida, Kentucky, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, West Virginia, District of Columbia
• Midwest: Illinois, Indiana, Iowa, Minnesota, Missouri, North Dakota, Ohio, South Dakota, Wisconsin
• Mountain and West: Alaska, Arizona, California, Colorado, Hawaii, Montana, Nevada, New Mexico, Oregon, Utah, Washington

A few of these states (Texas, Georgia, Tennessee, Iowa) operate low-THC or oil-only programs that authorize CBD-rich preparations specifically for seizure indications but limit THC content and product forms substantially. Other states operate comprehensive programs that include flower, vape, and concentrate products.

The structural difference from Epidiolex:

• Product variability. State-program products are not GMP-manufactured. CBD and THC content vary between batches and between operators. Some products use the same plant material that is hemp-derived (under the 2018 Farm Bill); others are state-licensed cannabis. Contamination (pesticides, heavy metals, microbial) is a real concern in some markets.
• Dose imprecision. Epidiolex dosing is precise (5-20 mg/kg/day), titrated and monitored. State-program product dosing is patient-determined, with variable concentrations across products.
• No FDA-equivalent safety monitoring. Epidiolex carries label warnings, drug-interaction guidance, and required liver-function monitoring. State-program products carry no equivalent regulatory infrastructure.
• Cost and insurance. Epidiolex is covered by most prescription-drug plans. State-program products are not covered by insurance.

For the three Epidiolex-approved indications (LGS, Dravet, TSC), there is no compelling clinical reason to use state-program cannabis instead of FDA-approved Epidiolex. The FDA pathway is more rigorous, more consistent, more affordable for most patients, and supported by stronger evidence.

For epilepsy indications outside the three Epidiolex-approved syndromes, the situation is different. These include refractory adult focal epilepsy, post-traumatic epilepsy, autoimmune epilepsy, juvenile myoclonic epilepsy, and a long tail of less common syndromes. Epidiolex is not approved for these indications, and the off-label use is constrained by insurance coverage. State-program cannabis fills a gap in those cases, but the evidence base outside the approved indications is substantially thinner.

The CBD distinction matters

A confusion that affects almost every conversation about cannabis and epilepsy: most of the strong cannabis-epilepsy evidence is for cannabidiol (CBD) specifically, not for whole-plant cannabis or for THC.

• Cannabidiol (CBD) is the cannabinoid with the well-documented anti-seizure effect, the subject of the Epidiolex approval, and the active ingredient in most state-program "CBD for epilepsy" products.
• Tetrahydrocannabinol (THC) is the primary psychoactive cannabinoid in cannabis. THC has some emerging evidence for seizure modulation in animal models and some patient case reports, but the controlled-trial evidence for THC in epilepsy is much weaker than for CBD, and THC carries cognitive and developmental concerns in pediatric populations.
• Whole-plant cannabis contains both CBD and THC in varying ratios, plus dozens of other cannabinoids and terpenes. The whole-plant evidence base for epilepsy is essentially observational case-series, not controlled trials.

For an epilepsy patient considering cannabis, the chemotype matters. A high-CBD, low-THC product (closer to the Epidiolex composition) has stronger evidentiary backing than a high-THC product. A 1:1 CBD:THC product or a high-THC product is supported by much less evidence for seizure indication and carries more risk of adverse cognitive effects, especially in pediatric patients.

The pediatric distinction

The strongest cannabis-epilepsy evidence is in pediatric epilepsy syndromes: Dravet, LGS, TSC. These are severe, treatment-resistant childhood epilepsies where conventional anti-seizure drugs often fail, where seizure burden is high, and where the developmental impact of uncontrolled seizures is substantial. The Epidiolex pivotal trials enrolled primarily pediatric patients, and the FDA approvals cover pediatric populations from 1 year and older (with subsequent expansion to younger ages).

The pediatric medical-cannabis decision-making framework applies directly to pediatric epilepsy patients. That framework includes state-program qualification rules, the /articles/medical-cannabis-and-pediatric-patients discussion of clinical considerations, and the Jack's Law / Quintin's Law school-administration provisions. The companion pediatric-patients article covers the structural pediatric considerations in detail.

The NINDS resources for Lennox-Gastaut Syndrome and Dravet Syndrome provide non-cannabis-specific clinical-background information that families navigating these diagnoses find useful as a starting reference.

The NASEM consensus view

The 2017 NASEM consensus report on cannabis and cannabinoids classified the evidence for cannabinoids in epilepsy at the time of writing. The report noted insufficient evidence for cannabinoids as effective treatment for epilepsy in adults, but the report was published in early 2017. That was before the pivotal Epidiolex trials had been published in their final form and before the FDA approval. The evidence picture has shifted substantially since.

The 2021 Golub and Reddy review (PMID 33332006) in Advances in Experimental Medicine and Biology provides a more current synthesis specifically for refractory epilepsy. The conclusion: cannabidiol has well-documented efficacy in Dravet, LGS, and TSC; emerging evidence for additional syndromes; and a generally favorable safety profile when administered as a pharmaceutical-grade preparation under clinical supervision.

The current consensus, four years after the FDA approval and seven years after the first pivotal trials: cannabidiol is an established anti-seizure medication for three specific pediatric epilepsy syndromes. The whole-plant cannabis evidence base for epilepsy is much thinner. The off-label use of CBD or whole-plant cannabis for other epilepsy indications occupies a clinical gray area where individual-patient clinical judgment is operating without strong randomized-trial backing.

What patients and families should weigh

Five practical considerations:

1. For the three approved indications, start with Epidiolex. A family with a child diagnosed with Dravet syndrome, Lennox-Gastaut syndrome, or tuberous sclerosis complex has access to an FDA-approved, insurance-covered, pharmaceutical-grade product backed by pivotal trial evidence. There is no clinical reason to bypass that pathway in favor of state-program cannabis.
2. For off-label indications, expect a higher evidence burden. A patient pursuing state-program cannabis for an epilepsy syndrome not covered by Epidiolex is operating on weaker evidence. The clinical conversation should be honest about that fact.
3. Choose chemotype based on evidence. High-CBD, low-THC products align with the strongest evidence base. High-THC products are not the right starting place for epilepsy indication, especially in pediatric patients.
4. Drug-drug interactions are clinically significant. Cannabidiol inhibits cytochrome P450 enzymes (CYP3A4, CYP2C19) and can substantially elevate serum levels of co-administered anti-seizure medications including clobazam, valproate, and others. Liver-function monitoring is part of the standard Epidiolex protocol and is clinically appropriate for state-program patients on high-dose CBD products.
5. Continue conventional epilepsy care. Cannabis is an add-on therapy in the pivotal trials, not a replacement for established anti-seizure medications. The clinical pattern that produced the FDA approval was cannabidiol added to a patient's existing regimen, not substituted for it.

What the federal-rescheduling change means for epilepsy

The April 22, 2026 DOJ order moving cannabis from Schedule I to Schedule III did not change Epidiolex's regulatory status (which has been Schedule V since the original FDA approval) or its clinical use. For whole-plant cannabis used through state programs for epilepsy, the rescheduling has indirect implications:

• Research access for cannabis-epilepsy clinical trials is structurally easier under Schedule III, which should accelerate the pipeline of trials for off-label epilepsy indications.
• The state-program operational framework remains unchanged: state qualifying-condition lists, state physician certification, and state-licensed dispensary distribution.
• No federal prescribing pathway for whole-plant cannabis has emerged; Epidiolex remains the only cannabis-derived prescription pathway for epilepsy.

For patients tracking this: the most important near-term change for epilepsy specifically is the potential expansion of clinical-trial evidence for off-label epilepsy indications over the next several years. The clinical pathway for the three approved indications is already well-established and is unlikely to change substantively.

Related reading

• Epilepsy condition overview: diagnostic criteria, evidence summary, FAQ
• Seizure disorders condition overview: broader category of seizure conditions
• Medical cannabis and pediatric patients: full pediatric medical-cannabis framework including school-administration considerations
• How to get a medical marijuana card: the qualification process every state has in common

[Last reviewed 2026-05-24. This is informational only, not medical advice. Epilepsy treatment decisions should be made in consultation with a qualified neurologist or pediatric neurologist. Epidiolex prescribing requires a qualified prescriber and is distinct from state medical-cannabis program participation.]