Abstracts

Rationale: Cannabidiol (CBD) received FDA approval for treatment of seizures associated with Lennox Gastaut syndrome (LGS) and Dravet Syndrome (DS) in 2018, and for tuberous sclerosis (TSC) in 2020. Randomized controlled trials have established short term efficacy and safety in these patient populations, but the post-marketing pattern of its effectiveness in clinical practice has not been well described.

Methods: We conducted a single-center retrospective cohort study of patients ≤ 18 years prescribed pharmaceutical CBD between 2018 and 2021. Demographic data and epilepsy characteristics were summarized via standard descriptive measures. Treatment failure was defined as CBD discontinuation, initiation of another anti-seizure therapy, or uptitration of a concurrent therapy due to poor seizure control. Time to treatment failure was analyzed using Kaplan-Meier survival curves with log-rank tests of significance. Chi-squared and Fischer’s tests were employed to analyze categorical variables. Cox proportional hazards model was used to determine independent predictors of treatment persistence.

Results: The analysis cohort included 176 patients (48% female). Mean follow-up time was 23.54 months. 73 patients had Lennox Gastaut syndrome (LGS); 18 had Dravet syndrome (DS). Treatment failure occurred in 132 patients by the end of the observation period. Probability of remaining on CBD without additional therapy was 21.21% by 12 months (SE 4%, 95% CI 15-29%) and 4.55% by 24 months (SE 2%, 95% CI 2-9%). Median time to failure was 5.48 months with mean of 7.44 months. Medan time to failure was longer in LGS patients (9.06 months) than DS patients (3.42 months) and all others (3.29 months) (p< 0.001). Median time to failure was longer in patients on clobazam at time of CBD initiation (6.72 months) compared to those not on concurrent clobazam (4.31 months) (p=0.04). Concurrent valproate had no significant effect on time to failure (median of 4.6 months vs. 5.61 months, p=0.97). The presence of “drop" seizures (tonic, atonic, tonic-clonic) was associated with longer time to failure (5.86 months vs. 3.87 months, p=0.04). In patients discontinuing CBD (n=69), the reason cited was more often lack of efficacy (36 pts, 52.17%) than lack of tolerability (10 pts, 14.49%), or both (16 pts, 23.19%). In multivariable-adjusted Cox regression, independent predictors of treatment persistence included older age at CBD start (HR 0.96, SE 0.02, p=0.04), concurrent clobazam (HR 0.65, SE 0.12, p=0.02), LGS (HR 0.51, SE 0.09, p< 0.001), and higher maximum dose/kg (HR 0.94, SE 0.02, p< 0.001). Other factors including sex, number of failed medications, “drop” seizure types, concurrent valproate, and seizure onset (focal vs generalized vs combined) were not independently associated with time to failure.