Abstracts

Rationale: Lennox–Gastaut syndrome (LGS) is a childhood-onset epilepsy that causes frequent seizures of multiple types. It is one of the most drug-resistant forms of epilepsy. GW is currently developing a plant-derived pharmaceutical formulation of cannabidiol (CBD) as adjunctive treatment of seizures associated with LGS.We present an exposure-response analysis of the relation between the steady-state exposure of CBD and its two major circulating metabolites, 7-OH-CBD and 7-COOH-CBD, and the drop seizure frequency and the number of patients with adverse events (AEs), reported during two clinical trials (GWPCARE 3 and GWPCARE 4). The purpose was to determine if the efficacy (drop seizures) and/or safety (AE) signals were related to the exposure of CBD. Methods: A total of 360 LGS patients aged 2.6 to 48 years from two clinical trials (GWPCARE 3 and GWPCARE 4) were included in the analysis. The patients received one of placebo, 10 mg/kg/day, or 20 mg/kg/day CBD administered as add-on to existing AEDs. Placebo and CBD were administered in two equally divided doses for 14 weeks, including a 2-week titration period. Exposure-response plots of the drop seizure frequency change from baseline over time by dose and AUC terciles of CBD, as well as logistic regressions for the drop seizure responder rate (patients with ≥50 % reduction) and AEs were used to determine exposure-response relations. Exposure metrics were derived from a previously developed population pharmacokinetic model based on the two trials. Results: The exposure-response plots showed a separation between the change from baseline of drop seizures for active treatment compared to placebo after approximately 15 days (the titration period) for all three analytes. The two dose levels (10 and 20 mg/kg/day) appeared to be similar in terms of median drop seizure reduction, but a higher proportion of patients achieved ≥75% and 100% reduction with the 20 mg/kg/day dose. Trends towards a larger reduction in seizures with for increasing steady-state AUCs was noted for the AUC terciles. A logistic regression analysis of the drop seizure responder rate revealed a significant (p < 0.01) positive correlation with the AUC for CBD. Predicted probability of response was in the order of 80% for the highest AUCs. The logistic regressions for AEs in several cases revealed significant positive correlations between the probability of a subject having at least one AE and the AUCs of the analyte. The pattern of exposure-response for 7-OH- CBD and 7-COOH CBD followed a similar pattern to that observed for CBD. Conclusions: Drop seizure responder rate was significantly correlated with the exposure of CBD and positive correlations with AUC were also determined for several AEs. Results suggest that the observed responses are directly related to CBD exposure, and CBD could represent a viable treatment for an otherwise pharmacoresistant condition. Funding: This analysis was funded by GW Research Ltd