We assessed efficacy and safety of perampanel (selective noncompetitive AMPA receptor antagonist) for primary generalised tonic-clonic (PGTC) seizures. Patients ≥12years with confirmed IGE; ≥3 PGTC seizures/8 weeks prior to randomization; receiving 1–3 concomitant AEDs were recruited. Trial consisted of 4–week screening; 4–8 week Baseline, 1:1 Randomization (perampanel titrated over 4 weeks to 8mg or highest tolerated dose versus placebo), 13–week Maintenance, 4–week Follow-up and Extension Phases. 164 patients were randomized; full analysis set included 81 patients each on perampanel and placebo. Median percent change in PGTC seizure frequency/28 days during Titration/Maintenance versus Baseline was –76.5% with perampanel versus –38.4% placebo; P<0.0001. 50% PGTC seizure responder rate was 64.2% with perampanel versus 39.5% placebo; P=0.0019. During Maintenance, 30.9% of perampanel patients were free of PGTC seizures versus 12.3% placebo. Treatment-emergent AEs (TEAEs) occurred in 82.7% of perampanel and 72.0% placebo patients; most common dizziness, fatigue, headache, somnolence, irritability. Serious TEAEs occurred in 6 (7.4%) perampanel and 7 (8.5%) placebo patients (one death in the perampanel group [accidental drowning; not treatment-related],one with placebo). In conclusion, adjunctive perampanel treatment up to 8mg improved seizure control in PGTC seizure patients, with almost 1/3 free of PGTC seizures during Maintenance. Perampanel was well tolerated. Study sponsor: Eisai Inc.