Children and young people with treatment-resistant Dravet syndrome on background antiseizure treatment
Studied preparation
Defined purified/pharmaceutical CBD oral solution, adjunctive use
Route
oral
Primary outcome
Change in convulsive-seizure frequency during randomized treatment
Study scope
Two distinct randomized cohorts; multiple active arms share their trial control and are not counted as separate trials
Reported finding · selected human evidence
Both selected randomized cohorts reported fewer convulsive seizures with adjunctive purified CBD than placebo. These results concern the studied treatment and population.
Two selected cohorts do not represent every Dravet study. Effects use different statistical measures and are not pooled.
2 distinct randomized cohorts in this source set. Safety Part A, overlapping open-label extension, and mouse mechanism study are not additional independent human efficacy trials.
Reported effects and harms
Estimates retain their original comparator, units and 95% confidence interval (CI). Different measures are not pooled. Study regimens identify the research; they are not treatment instructions.
199 randomized; 198 eligible for intention-to-treat analysis. The two CBD arms share one placebo group.
Both assigned CBD regimens reduced the primary seizure-frequency outcome relative to placebo.
Contrast
Measure
Estimate
95% CI
10 mg/kg/day study arm versus placebo
Model-reported reduction relative to placebo
29.8 %
8.4 to 46.2 %
20 mg/kg/day study arm versus placebo
Model-reported reduction relative to placebo
25.7 %
2.9 to 43.2 %
Harms reported
Treatment-emergent events: 56/64 in the lower-dose safety group, 62/69 in the higher-dose group, 58/65 with placebo. Two participants changed dose groups for safety analysis.
Interpretation limits
GW Research supported and participated in the study. Safety and efficacy denominators differ; higher assigned treatment is not a recommendation.
Source depth
Primary full text: Results, adverse events and Table 2 footnotes.
No laboratory context selected in this pilot source set.
Source checks and notices
The selected PubMed records were checked 2026-11-08 UTC. No other notice signals were returned for these records at that check; missing metadata signals do not prove that no notice exists. This monitor checks citation relations, not every full-text change, and never rewrites a finding automatically.
Research context, not a recommendation for a medicine, dose or retail product. Findings for the studied preparation do not establish equivalence to products in the catalog.