Abstracts

Rationale: Recently, there has been a call for novel trial designs that reduce patient exposure to placebo. This is particularly true of studies involving patients with generalized tonic-clonic seizures, which confer a greater patient risk. One such design, “time to event”, allows patients to exit a study when it is clear that the experimental intervention (either drug or placebo) is not of benefit to them. We performed a post hoc analysis of a parallel add-on trial of perampanel in treatment-resistant PGTC seizures to assess the utility of this design. Methods: Patients aged ≥12 years with idiopathic generalized epilepsy and uncontrolled PGTC seizures, despite treatment with 1–3 antiepileptic drugs, were randomized to receive once-daily placebo or adjunctive perampanel (up to 8 mg/day), across a 17-week Double-blind Treatment Phase (4-week Titration; 13-week Maintenance). Primary efficacy outcomes were median percent change in PGTC seizure frequency per 28 days and 50% responder rates. Time from the first dose date of perampanel to the nth+1 PGTC seizure event (where n+1 = prerandomization Baseline seizure frequency per 28 days + 1) was an exploratory endpoint. Kaplan-Meier analyses with log-rank tests were performed to estimate the median time to the nth+1 PGTC seizure event and 95% confidence intervals (CI). Results: Of 164 randomized patients, 162 (placebo, n=81; perampanel, n=81) were included in the Full Analysis Set (mean age 28.4 years; 56.2% female). Compared with placebo, perampanel conferred a significantly greater median percent reduction in PGTC seizure frequency per 28 days (38.4% vs 76.5%; P < 0.0001) and 50% responder rates (39.5% vs 64.2%; P=0.0019). Time to the nth+1 PGTC seizure event from the start of study treatment was longer for the perampanel group than for the placebo group (P < 0.0001; Figure 1). The median time (95% CI) to an nth+1 PGTC seizure event was 43.0 (34.0, 51.0) days for the placebo group but could not be estimated for the perampanel group (70.0, not calculable) as < 50% of patients experienced a PGTC seizure event (Table 1). Conclusions: Consistent with primary efficacy outcomes and compared with placebo, adjunctive perampanel significantly prolonged time to prerandomization monthly seizure count in patients aged ≥12 years with PGTC seizures. Results suggest that time to the nth+1 seizure would be a useable endpoint in studies of PGTC seizures and should be considered for future trials. Funding: Eisai Inc.