Background Perampanel (PER) is a novel antiepileptic drug that inhibits the AMPA class of glutamate receptors. It has been available in the UK since September 2012. We undertook a retrospective analysis of efficacy and tolerability of PER in 47 patients with drug-refractory epilepsy attending a regional epilepsy service in the UK. Methods Demographic and clinical data of patients with refractory epilepsy prescribed PER were collected by review of records. Efficacy, as measured by responder rates (> 50% reduction in seizure frequency), retention rates, and adverse effects, was analyzed. Results Of the 47 patients prescribed PER, 39 (87%) had focal epilepsy, four (9%) had idiopathic generalized epilepsy, 3 (6%) had symptomatic generalized epilepsy, and 1 had unclassified epilepsy. Patients were taking a median of 2 AEDs (range: 1–5) when starting on PER. The median dose of PER was 8 mg (range: 2–12 mg). Thirteen (28%) patients were classed as responders, but no patients experienced sustained seizure freedom. Twenty-one (45%) patients had withdrawn from PER during the study period, with 16 (76%) of them withdrawing due to intolerable adverse effects, 4 due to inadequate seizure control, and 1 due to the combination of both. The most frequent adverse effects requiring withdrawal from PER were behavioral reactions including suicidal ideation (n = 2), aggressive behavior (n = 2), and both (n = 1). Conclusion In our experience, PER had a retention rate of 55% and a responder rate of 28%. Psychiatric adverse effects, including suicidal ideation, were the most common reasons for withdrawal.