How Well Does CBD Work for Severe Epilepsy in the Real World?

In 107 patients with severe developmental epilepsy syndromes, 69% achieved at least a 50% seizure reduction with CBD—and patients with genetic causes responded best.

Perulli, Marco et al.·Epilepsia open·2025·Moderate EvidenceRetrospective Cohort·1 min read
RTHC-07357Retrospective CohortModerate Evidence2025RETHINKTHC RESEARCH DATABASErethinkthc.com/research

Quick Facts

Study Type
Retrospective Cohort
Evidence
Moderate Evidence
Sample
N=107
Participants
N=107 patients with developmental and epileptic encephalopathies, treated with CBD for ≥3 months, from 2020 to 2024.

What This Study Found

Clinical trials establish whether a drug works under ideal conditions. Real-world studies reveal how it performs in actual clinical practice—with diverse patients, varied medication regimens, and imperfect adherence. This study fills that gap for CBD in severe epilepsy.

One hundred seven patients with developmental and epileptic encephalopathies (DEEs) were treated with pharmaceutical CBD for at least 3 months between 2020 and 2024. The cohort included approved indications (Lennox-Gastaut, Dravet, tuberous sclerosis complex) and off-label uses in other DEEs.

At a median follow-up of 20 months, 69% achieved at least 50% seizure reduction—a strong real-world result. Twenty-one percent achieved 75% or greater reduction. These numbers are encouraging given that all patients were, by definition, treatment-resistant (having failed multiple prior medications).

Two findings stood out for personalizing treatment. First, patients with LGS, TSC, and other DEEs showed higher efficacy and retention than Dravet syndrome patients—surprising given that Dravet was one of the original FDA-approved indications. Second, genetic or unknown etiology predicted better outcomes (p = 0.011), suggesting that the underlying cause of epilepsy matters for CBD response.

The combination of CBD with valproate was associated with specific side effect considerations that required clinical management.

Key Numbers

N = 107. Median follow-up: 20 months. 69% achieved ≥50% seizure reduction. 21% achieved ≥75% reduction. LGS: 55.1%, Dravet: 16.8%, TSC: 8.4%, other DEEs: 19.6%. Genetic etiology: 56.1%. Genetic/unknown etiology associated with better outcomes (p = 0.011).

How They Did This

Retrospective study of 107 patients with DEEs treated with CBD for ≥3 months (2020–2024). Diagnoses: LGS (55.1%), Dravet (16.8%), TSC (8.4%), other DEEs (19.6%). Genetic etiology in 56.1%. Data on seizure frequency, tolerability, retention, and non-seizure outcomes. Efficacy defined as ≥50% or ≥75% seizure reduction. Statistical analysis of predictors of response.

Why This Research Matters

Real-world data like this helps clinicians set realistic expectations. The 69% responder rate is strong but also means 31% didn't achieve meaningful seizure reduction—knowing who is most likely to respond (genetic etiology, LGS/TSC) can inform treatment decisions. The off-label data in "other DEEs" also expands the evidence base beyond the three FDA-approved syndromes.

The Bigger Picture

This real-world evidence complements RTHC-00186's prescribing guide and RTHC-00165's evidence review. While RTHC-00160 and RTHC-00179 explore next-generation CBD formulations in animals, this study shows what currently available pharmaceutical CBD is achieving in clinical practice. The finding that Dravet patients responded less well than LGS/TSC is clinically important and somewhat counterintuitive given Dravet's prominence in the CBD-epilepsy narrative.

What This Study Doesn't Tell Us

Retrospective design with inherent biases (treatment decisions weren't randomized). No control group—some improvement may reflect natural seizure fluctuation. Single-center experience. The 3-month minimum treatment duration excludes early discontinuations (potentially biasing toward responders). Seizure counting relies on caregiver observation, which can be imprecise.

Questions This Raises

  • ?Why do Dravet patients show lower response rates than other DEEs in real-world use?
  • ?Could genetic testing identify CBD responders before starting treatment?
  • ?Would the off-label DEE patients have been eligible for approval trials, and should indications be expanded?

Trust & Context

Key Stat:
Evidence Grade:
Retrospective real-world cohort with 20-month median follow-up—valuable for clinical practice but limited by lack of randomization and control group.
Study Age:
Published in 2025 with data from 2020–2024, reflecting current clinical CBD use patterns.
Original Title:
Real-world efficacy and safety of cannabidiol in developmental and epileptic encephalopathies.
Published In:
Epilepsia open, 10(6), 1806-1813 (2025)Epilepsia Open is a reputable journal focusing on epilepsy research and treatment.
Database ID:
RTHC-07357

Evidence Hierarchy

Meta-Analysis / Systematic Review
Randomized Controlled Trial
Cohort / Case-ControlFollows or compares groups over time
This study
Cross-Sectional / Observational
Case Report / Animal Study

Looks back at existing records to find patterns.

What do these levels mean? →

Read More on RethinkTHC

Cite This Study

RTHC-07357·https://rethinkthc.com/research/RTHC-07357

APA

Perulli, Marco; Bianchetti, Maddalena; Pantalone, Gloria; Quintiliani, Michela; Gambardella, Maria Luigia; Picilli, Maria; Marini, Carla; Cesaroni, Elisabetta; Battaglia, Domenica Immacolata. (2025). Real-world efficacy and safety of cannabidiol in developmental and epileptic encephalopathies.. Epilepsia open, 10(6), 1806-1813. https://doi.org/10.1002/epi4.70149

MLA

Perulli, Marco, et al. "Real-world efficacy and safety of cannabidiol in developmental and epileptic encephalopathies.." Epilepsia open, 2025. https://doi.org/10.1002/epi4.70149

RethinkTHC

RethinkTHC Research Database. "Real-world efficacy and safety of cannabidiol in development..." RTHC-07357. Retrieved from https://rethinkthc.com/research/perulli-2025-realworld-efficacy-and-safety

Access the Original Study

Study data sourced from PubMed, a service of the U.S. National Library of Medicine, National Institutes of Health.

This study breakdown was produced by the RethinkTHC research team. We analyze and report published research findings without making health recommendations. All interpretations are based solely on the published abstract and study data.