Introduction: Dravet syndrome (DS), a developmental and epileptic encephalopathy, presents with treatment-resistant seizures, including convulsive seizures; high convulsive seizure frequency increases risk of premature death. Cannabidiol and fenfluramine, the most recently FDA- and EMA-approved adjunctive treatments for DS, were compared using a numbers needed to treat (NNTs) analysis to quantify the number of individuals needing treatment to observe a positive outcome.

Objectives: To determine the effectiveness of fenfluramine and cannabidiol in an indirect comparison model.

Materials & Methods: Data identified from a published systematic literature review of 3 randomized controlled trials (RCTs) of DS patients comparing fenfluramine (0.4 mg/kg/d with stiripentol or 0.7 mg/kg/d without stiripentol) vs placebo and 2 RCTs comparing cannabidiol (10 or 20 mg/kg/d irrespective of clobazam) vs placebo were analyzed (Metalnsight web-based tool). Reduction in monthly convulsive seizure frequency (MCSF) from baseline of ≥50% (MCSF50) and ≥75% (MCSF75) was measured. A Bayesian network meta-analysis (NMA) model with fixed effects was conducted to estimate odds ratios of MCSF response for fenfluramine and cannabidiol vs placebo. Using an estimated placebo response rate for the network, NNTs vs placebo were calculated using NMA-adjusted MCSF rates.

Results: NMA-adjusted MCSF50 response rates were 17%, 29%, 33%, 86%, and 84% under placebo, cannabidiol 10 and 20 mg/kg/d, and fenfluramine 0.4 and 0.7 mg/kg/d, respectively; NNTs (credible interval) comparing each treatment (cannabidiol 10 and 20 mg/kg/day, fenfluramine 0.4 and 0.7 mg/kg/day) vs placebo were 8.0 (3.5, 144.3), 6.3 (3.5, 18.2), 1.5 (1.2, 2.4), and 1.5 (1.3, 1.9). NMA-adjusted MCSF75 response rates were 6%, 27%, 14%, 67%, and 66%; NNTs were 4.8 (2.3, 14.1), 11.7 (4.7, 63.6), 1.6 (1.1, 5.5), and 1.7 (1.2, 3.0).

Conclusions: Results suggest that, within an indirect comparison model, fenfluramine is more likely to achieve MCSF reduction than cannabidiol as fewer patients need to be treated for one person to benefit. These results are subject to the limitations of indirect comparison methodology and limited available data.

Supported by: UCB Pharma.