Abstracts

Rationale: Perampanel (PER), a selective, noncompetitive AMPA receptor antagonist, has been approved as adjunctive treatment for partial-onset seizures (POS). The efficacy and safety of PER has been well documented in 3 double-blind (DB), randomized, placebo-controlled phase III trials. It is not clear whether patients with partial epilepsy respond differently depending on the presence or absence of structural epileptogenic lesions. Here we present pooled PER phase III results from post-hoc analysis comparing lesional and non-lesional POS patients.Methods: Patients with refractory POS enrolled in the phase III trials were aged 12 years and receiving 1-3 concomitant AEDs. Following 6-wk baseline, patients were randomized to once-daily DB treatment (6-wk titration, 13-wk maintenance) with placebo or PER 8 or 12mg (studies 304 & 305); or with placebo or PER 2, 4, or 8mg (study 306). These study results were pooled and grouped by post-hoc assessment of etiology: structural lesions=structural brain anomalies/malformations, head injury/cranial trauma, CNS infection, vascular brain anomalies, or stroke; non-lesional=family history of epilepsy, sleep disorder, unknown or other causes). Efficacy population excluded patients from Latin American sites (N=162 patients) due to a significant treatment-by-region interaction. Efficacy was also limited to the completer population for the structural-lesion (37.3%; placebo N=133; PER N=288) and non-lesional (62.7%; placebo N=215; PER N=493) groups. Safety included full intent-to-treat population for the structural-lesion (placebo N=161; PER N=398) and non-lesional (placebo N=281; PER N=640) groups. Study endpoints: change from baseline in seizure frequency/28 days, responder rate, and treatment-emergent adverse events (TEAEs). Results reported by last dose.Results: Etiology of structural-lesion group included patients with structural brain anomalies/malformations (39.2%), head injury/cranial trauma (27.3%), CNS infection (25.2%), vascular brain anomalies (5.5%), and stroke (2.9%). Non-lesional group included patients with unknown (79.4%) or other causes (17.1%), family history of epilepsy (2.8%), and sleep disorder (0.7%). Seizure type and localization were similar for lesional and non-lesional patients. Efficacy results for patients with structural lesions vs no lesions are shown in Table 1. Median changes in seizure frequency/28 days and responder rates for placebo and PER 2, 4, 8, & 12mg treatment groups were comparable between structural-lesion and non-lesional groups. Percentages and types of TEAEs were not notably different between structural-lesion and non-lesional groups. Top 3 TEAEs for total PER by decreasing frequency were dizziness, somnolence, and headache for both groups.Conclusions: Efficacy and safety results for PER were consistent for POS patients with structural lesions and no reported lesions; both groups showed improved seizure control with PER 4, 8, & 12mg relative to placebo. PER is effective and well tolerated regardless of seizure etiology.